ECONOMIC COSTS
OF POWER
LOADSHEDDING IN
PAKISTAN
Dr. Hafiz A. Pasha
Dr. Aisha Ghaus-Pasha
Wasim Saleem
INSTITUTE OF PUBLIC POLICY
BEACONHOUSE NATIONAL UNIVERSITY
1
Authors
Dr. Hafiz Ahmed Pasha
Dr. Aisha Ghaus-Pasha
Wasim Saleem
Assisted By
Ijaz Hussain
Atr-un-Nisa
Muhammad Imran
Survey Supervisors
Shafee Ahmed
Sohail Javed
Ali Hassan
Syed Mazhar Hussain
Abdul Haleem Khoso
Muhammad Nadir
Asad Ali
Izhar-ul-Hassan
Ali Sufyan
Report Composing
Muhammad Rizwan
ii
CONTENTS
ACRONYMS
EXECUTIVE SUMMARY
PART I: INTRODUCTION AND ANALYSIS OF SECONDARY DATA
CHAPTER 1
INTRODUCTION
1.1
Statement of Work
1.2
Methodology
1.3
The Outputs
1
7
8
8
9
10
CHAPTER 2
THE POWER SECTOR OF PAKISTAN
2.1
Structure and Growth of the Economy
2.2
International Comparisons
2.3
Long-Term Trends in Power Sector
2.4
Growth and Pattern of Electricity Consumption
2.5
The Supply Gap
11
CHAPTER 3
METHODOLOGY FOR QUANTIFICATION OF COST OF LOADSHEDDING
3.1
The Simple Value Added Approach
3.2
The Adjusted Value Added Approach
3.3
Marginal Cost of Unsupplied Electricity
3.4
The Value of Leisure Approach
3.5
The Consumer Surplus Approach
3.6
The Contingent Valuation (WTP) Approach
3.7
The Survey Based Approach
3.8
Results of International Studies
19
CHAPTER 4
EARLIER ESTIMATES OF COSTS OF LOAD SHEDDING IN PAKISTAN
4.1
The Pasha, Ghaus and Malik Study
4.2
The IPP Study
4.3
The PIDE Study
4.4
Conclusions
26
CHAPTER 5
COSTS OF LOADSHEDDING (FROM SECONDARY DATA)
5.1
The Simple Value Added Approach
5.2
The Adjusted Value Added Approach
5.3
The Consumer Surplus Approach
31
PART II: COST OF LOADSHEDDING TO SMALL-SCALE INDUSTRY
CHAPTER 6
INTRODUCTION
11
13
14
15
18
19
20
20
21
22
23
23
23
26
28
29
30
31
32
33
36
37
iii
CHAPTER 7
THEORETICAL FRAMEWORK
7.1
Outages and a Firm’s Behaviour
7.2
Methodology for Quantification of Cost of Outages
38
CHAPTER 8
THE SAMPLING FRAMEWORK AND ITS DISTRIBUTION
8.1
Sampling Framework
8.2
Profile of Respondants
8.3
Characteristics of Selected Units
44
CHAPTER 9
THE INCIDENCE OF LOADSHEDDING AND OUTPUT LOSSES
9.1
Incidence And Profile of Loadshedding
9.2
Extent of Total Time Lost
9.3
Seasonality of Outages
9.4
Extent of Output Lost During Outages
52
CHAPTER 10
ADJUSTMENTS TO LOADSHEDDING
10.1
Number and Types of Adjustments
10.2
Extent of Loss of Output in Outages
59
CHAPTER 11
COSTS OF OUTAGES
11.1
Total Outage Costs
11.2
Types of Outage Costs
11.3
Burden of Outage Costs
11.4
Outage Costs Per Kwh
11.5
National Estimate of Outage Costs
63
CHAPTER 12
LOAD MANAGEMENT STRATEGY: CONSUMER’S PREFERENCES
12.1
Level of Satisfaction With Current Level of Service
12.2
Preferred Changes in Timings of Loadshedding
68
CHAPTER 13
SUGGESTIONS BY THE SAMPLE UNITS
76
CHAPTER 14
CONCLUSIONS AND POLICY IMPLICATIONS
14.1
Impact of Outages
14.2
Affordability of Power Tariffs
14.3
Competitive Disadvantage of Small Units
14.4
Investment in Generators
14.5
Load Management Strategy
79
38
39
44
46
47
52
54
56
57
59
61
63
64
65
66
67
68
71
79
80
81
82
83
iv
PART III: COST OF LOADSHEDDING TO COMMERCIAL/SERVICES SECTOR
CHAPTER 15
INTRODUCTION
84
85
CHAPTER 16
THEORETICAL FRAMEWORK
16.1 Outages and a Firm’s Behavior
16.2 Methodology for Quantification of Cost of Outages
86
CHAPTER 17
THE SAMPLING FRAMEWORK AND ITS DISTRIBUTION
17.1
Sampling Framework
17.2
Profile of Respondents
17.3
Characteristics of Selected Units
92
CHAPTER 18
THE INCIDENCE OF LOADSHEDDING AND OUTPUT LOSSES
18.1
Incidence and Profile of Loadshedding
18.2
Extent of Total Time Lost
18.3
Seasonality of Outages
18.4
Extent of Output Lost During Outages
100
CHAPTER 19
ADJUSTMENTS TO LOADSHEDDING
19.1
Number and Types of Adjustments
19.2
Extent of Loss of Sales/Output in Outages
104
CHAPTER 20
OUTAGES COSTS
20.1
Total Outage Costs
20.2
Burden of Outage Costs
20.3
Outage Costs Per Kwh
20.4
National Estimate of Outage Costs
107
CHAPTER 21
LOAD MANAGEMENT STRATEGY: CONSUMER’S PREFERENCES
21.1
Level of Satisfaction with Current Level of Service
21.2
Preferred Changes in Timings of Loadshedding
111
CHAPTER 22
SUGGESTIONS BY THE SAMPLE UNITS
115
CHAPTER 23
CONCLUSIONS AND POLICY IMPLICATIONS
23.1
Impact of Outages
23.2
Affordability of Higher Tariffs
23.3
Policy Implications for Load Management
118
86
87
92
95
96
100
101
102
103
104
105
107
109
109
110
111
114
118
118
119
v
PART IV: COST OF LOADSHEDDING TO DOMESTIC/RESIDENTIAL SECTOR
CHAPTER 24
INTRODUCTION
121
122
CHAPTER 25
METHODOLOGY FOR QUANTIFICATION OF COSTS
25.1 Approaches to Quantification of Costs to Domestic Consumers
25.2 Methodology For Quantification of Outage Cost
123
CHAPTER 26
THE SAMPLING FRAMEWORK AND ITS DISTRIBUTION
26.1
Sampling Framework
26.2
Characteristics of Selected Household
129
CHAPTER 27
THE EXPERIENCE OF LOADSHEDDING
27.1
Incidence and Profile of Loadshedding
27.2
Extent of Disruption Due to Outages
135
CHAPTER 28
ADJUSTMENTS TO LOADSHEDDING
28.1
Number and Types of Adjustments
141
CHAPTER 29
OUTAGES COSTS
29.1
Total Outage Costs
29.2
Burden of Outage Costs
29.3
Outage Costs Per Kwh
29.4
National Estimate of Outage Costs
144
CHAPTER 30
LOAD MANAGEMENT STRATEGY: CONSUMER’S PREFERENCES
30.1
Level Of Satisfaction With Current Level of Service
30.2
Preferred Changes in Timings of Loadshedding
147
CHAPTER 31
SUGGESTIONS BY THE SAMPLE UNITS
150
CHAPTER 32
CONCLUSIONS AND POLICY IMPLICATIONS
32.1
Impact of Outages
32.2
Affordability
32.3
Pricing policy
32.4
Self-generation
32.5
Load Management Strategy
153
123
126
129
131
135
137
141
144
145
146
146
147
148
153
153
154
154
155
vi
PART V: COST OF LOADSHEDDING TO AGRICULTURE SECTOR
CHAPTER 33
INTRODUCTION
156
157
CHAPTER 34
METHODOLOGY
34.1 Secondary-Data Based Methodology And Estimates
34.2 Primary –Data Based Methodology
158
CHAPTER 35
THE SAMPLING FRAMEWORK AND ITS DISTRIBUTION
35.1
The Sampling Approach
162
CHAPTER 36
COST OF OUTAGES IN AGRICULTURE
36.1
Source of Irrigation
36.2
Type of Tube Wells
36.3
Incidence of Loadshedding
36.4
Reduction In Output
36.5
Cost of Outages
36.6
National Estimate of Outage Costs
164
CHAPTER 37
OTHER COSTS OF OUTAGES
37.1
Home-Based Economic Activities
37.2
Domestic Consumption of Electricity
37.3
Total Cost of Outages
167
CHAPTER 38
LOAD MANAGEMENT STRATEGY
169
38.1
38.2
169
170
Level of Satisfaction With Current Level of Service
Preferred Changes in Timings of Loadshedding
158
161
162
164
164
164
165
165
166
167
167
168
CHAPTER 39
SUGGESTIONS BY THE SAMPLE UNITS
172
CHAPTER 40
CONCLUSIONS AND POLICY IMPLICATIONS
40.1
Conclusions
40.2
Policy Implications
174
174
174
Technical Annexure
EXECUTIVE SUMMARY
Technical Annexure 1
176
PART I: INTRODUCTION AND ANALYSIS OF SECONDARY DATA
Technical Annexure 2
177
vii
Technical Annexure 3
178
Technical Annexure 4
179
PART II: COST OF LOADSHEDDING TO SMALL-SCALE INDUSTRY
Technical Annexure 5
180
PART III: COST OF LOADSHEDDING TO COMMERCIAL/SERVICES SECTOR
Technical Annexure 6
181
References
182
LIST OF TABLES
Table 1: Average Hours of Incidence of Outages 2012
2
Table 2: Use of Generators
2
Table 3: Outage Costs as Percentage of Sectoral Value Added
3
Table 4: Summary of Results From the Loadshedding Project
4
Table 5: Comparison of Outage Costs as % of Sales in Selected Asian Countries
4
Table 6: Indicators for Load Management Strategy
6
PART I: INTRODUCTION AND ANALYSIS OF SECONDARY DATA
Table 2.1: Growth Rate of Different Sectors and of the GDP
12
Table 2.2: Contribution of Different Sectors to GDP
12
Table 2.3: Indicators Of Electricity Consumption In Selected Countries (2009)
13
Table 2.4: Level and Sources of Electricity Generation and System Losses in Selected Countries
(2009)
14
Table 2.5 Long-Term Trend in Capacity and Generation in Pakistan of Electricity 1970-71 to 201011*
Table 2.6: Growth in Electricity Consumption from 2000-01 to 2010-11
15
Table 2.7: Growth Rate of Sectoral Value Added and Electricity Consumption
16
Table 2.8: Share of Different Types of Consumers in Electricity Consumption
17
Table 2.9: Growth in Electricity Consumption per Consumer
17
Table 2.10: Comparison by Sector of share in GDP and share in Electricity Consumption
18
Table 2.11: Surplus/Deficit in Demand and Supply during System Peak Hours
18
Table 3.1: Cost of Outages per Kwh to Different Types of Consumers in Various Countries
24
Table 4.1: Break-up of Outage Costs
26
Table 4.2: Cost of Planned and Unplanned Outages per Kwh in different Industries
27
Table 4.3: Pattern of Adjustment by Size of Unit
28
Table 5.1: Cost of Outages with the Adjusted Value Added Approach
32
Table 5.2: Projected GDP (at factor cost) in the Absence of Loadshedding compared to Actual GDP
33
Table 5.3: Price Elasticity of Demand for Electricity (All Consumers Combined) in Different Studies
34
16
viii
PART II: COST OF LOADSHEDDING TO SMALL-SCALE INDUSTRY
Table 8.1: National Distribution of Small Manufacturing Establishments in the Economic Census,
2005, by Province and Industry Group
Table 8.2: Distribution of Sample units by City, Province and Industry
45
Table 8.3: Distribution of Sample Analyzed by Cities
46
Table 8.4: Average Employment and Capacity Utilization by Industrial Group, 2012
47
Table 8.5: Average Sale and Operating Expenses of Sample Units, 2012
48
Table 8.6: Average Value Added, Electricity Purchased and Value Added per Kwh of Electricity of
Sample Units
Table 8.7: Operating Cost Structure of Sample Units
49
Table 8.8: Reasons Why Production Target was Not Attained
51
Table 9.1: Frequency of Loadshedding in 2012
52
Table 9.2: Percentage Distribution of Average Length of Outages, 2012
53
Table 9.3: Duration of Outages
54
Table 9.4: Proportion of Time Lost
55
Table 9.5: Seasonality in Outages
56
Table 9.6: Nature of Impact of Loadshedding
57
Table 9.7: Ranking of Disruptions Due to Outages
58
Table 10.1: Percentage of Sample Units by Number of Adjustments by Process
59
Table 10.2: Percentage of Sample units Adjusting through Various Mechanisms
60
Table 10.3: Number of Adjustments by Firms with and without Generators
61
Table 10.4: Proportion of output Loss Not Recovered
62
Table 11.1: Total Outage Costs Per Unit
64
Table 11.2: Outage Costs as Percentage of Value Added
65
Table 11.3: Outage Costs Per Kwh
67
Table 12.1: Percentage of Time DISCOs Kept to the Announced Loadshedding Schedule
68
Table 12.2: Average Time Required for Adjustment to Changes in Loadshedding Schedule
69
Table 12.3: Level of Satisfaction with Current quality of Service by DISCOs/KESC
70
Table 12.4: Additional Tariff For Better Quality of Service (with No Loadshedding)
71
Table 12.5: Perceived Outage Costs per Kwh as implied by Willingness to Pay
72
Table 12.6: Percentage Willing To Change Work Timing If off Peak Power Tariff Are Reduced
73
Table 12.7: Worst Time of The Year for Loadshedding
74
Table 12.8: Worst Day of The Week for Outages
74
Table 12.9: Information that can be provided by Distribution companies to Units
75
Table 13.1: Suggestions by Sample Units by City
77
Table 14.1: Total Costs of Electricity Consumption as a percentage of the Value of Production
80
Table 14.2: Comparison of the Impact of Loadshedding by Size of Industrial Unit
81
Table 14.3: Benefit-Cost Ratio of Self-Generation
82
46
48
ix
Table 14.4: Proposed Tariff Structure on Imported Generators
82
PART III: COST OF LOADSHEDDING TO COMMERCIAL/SERVICES SECTOR
Table 17.1: National Distribution of Commercial Establishments in the Economic Census, 2005, by
93
Province and Sector
Table 17.2: Distribution of Sample units by City, Province and Sector
94
Table 17.3: Distribution of Sample Analyzed by Cities
95
Table 17.4: Average Employment by Sector, 2012
96
Table 17.5: Average Sale and Operating Expenses of Sample Units, 2012
96
Table 17.6: Average Value Added, Electricity Purchased and Value Added per Kwh of Electricity of
97
Sample Units
Table 17.7: Operating Cost Structure of Sample Units
97
Table 17.8: Reasons Why Production Target was Not Attained
99
Table 18.1: Frequency of Loadshedding in 2012
100
Table 18.2: Percentage Distribution of Average Length of Outages, 2012
101
Table 18.3: Duration of Outages
101
Table 18.4: Proportion of Time Lost During Outages
102
Table 18.5: Seasonality in Outages
102
Table 18.6: Nature of Impact of Loadshedding
103
Table 18.7: Ranking of Disruptions Due to Outages
103
Table 19.1: Percentage of Sample Units by Number of Adjustments by Group
104
Table 19.2: Percentage of Sample units Adjusting through Various Mechanisms
105
Table 19.3: Number of Adjustments by Firms with and without Generators
105
Table 19.4: Proportion of Output Loss Not Recovered
106
Table 20.1: Total Outage Costs Per Unit
107
Table 20.2: Outage Costs as Percentage of Value Added
109
Table 20.3: Outage Costs Per Kwh
110
Table 20.4: Total Cost of Outages to the Commercial Sector
110
Table 21.1: Percentage of Time DISCOs Kept to the Announced Loadshedding Schedule
111
Table 21.2: Average Time Required for Adjustment to Changes in Loadshedding Schedule
111
Table 21.3: Level of Satisfaction with Current quality of Service by DISCOs/KESC
112
Table 21.4: Additional Tariff for Better Quality of Service (with No Loadshedding)
113
Table 21.5: Perceived Outage Costs Per Kwh as implied by Willingness to Pay
113
Table 21.6: Worst Time of The Year for Loadshedding
114
Table 21.7: Worst Day of The Week for Outages
114
Table 21.8: Information that can be provided by Distribution companies to Units
114
Table 22.1: Suggestions by Sample Units by City
116
x
Table 22.2 Suggestions by Sample Units by Sector
117
Table 23.1: Total Costs of Electricity Consumption as a percentage of the Value of Production
119
PART IV: COST OF LOADSHEDDING TO DOMESTIC/RESIDENTIAL SECTOR
Table 25.1: Outage Cost per Kwh According to the Value of Leisure Approach
123
Table 25.2: Activities most Disturbed by Loadshedding
124
Table 25.3: Percentage of Sample Households with Generator and/or UPS
125
Table 25.4: Subjective Valuation of the Outage Cost per Hour
126
Table 26.1: National Distribution of Population in the Census, 1998 by City
129
Table 26.2: Distribution of Sample by Province and by City
130
Table 26.3: Occupation of the head of the households
131
Table 26.4: Average Number of Family Members by Income Group, 2012
131
Table 26.5: Profile of ownership of Assets
133
Table 26.6: Average Monthly Expenditure, Electricity Bill and Electricity Bill as % Monthly
Expenditure Electricity of Sample Units
Table 27.1: Average Number of Times There is Loadshedding in a day
134
Table 27.2: Hours of Outages
135
Table 27.3: Percentage Distribution of Average Length of Outages, 2012
136
Table 27.4: Timing of Loadshedding
136
Table 27.5: Experience of Voltage Fluctuations
137
Table 27.6: Disruptions Due to Loadshedding
137
Table 27.7: Ranking of Disruptions Due to Outages
139
Table 27.8: Change in the Timing to make Loadshedding Less Disruptive
140
Table 28.1: Household with Generator and UPS
141
Table 28.2: Is Generator a partial or full substitute of electricity supplied publically?
141
Table 28.3: Use of Generator for Various purpose
142
Table 28.4: UPS a partial or full substitute of electricity supplied publically
142
Table 28.5: Use of UPS for Various purpose
143
Table 28.6: Various Other Adjustments made to Deal with Loadshedding
143
Table 29.1: Total Outage Cost per Residential Consumer
144
Table 29.2: Total outage Cost as % of Total Household Consumption Expenditure
145
Table 29.3: Total Outage Cost per Kwh to Residential Consumer
146
Table 29.4: National Estimate of Outage Costs to Urban Residential Consumers, 2011-12
146
Table 30.1: Power Companies Kept to Loadshedding schedule
147
Table 30.2: Level of Satisfaction with Current Quality of Service by DISCOs/KESC
147
Table 30.3: Preference for the type of Loadshedding
148
Table 30.4: Worst Time of The Year for Loadshedding
148
Table 30.5: The Worst Day of The Week for Outages
149
15
xi
Table 30.6: Preference of Loadshedding Time
149
Table 30.7: Information that can be provided by Distribution companies to consumers
149
Table 31.1: Suggestions by Sample Units by City
151
Table 31.2 Suggestions by Sample Units by Income Group
152
Table 32.1: Total Cost of Electricity Consumption Per Residential Consumer
153
Table 32.2: Present Tariff Structure on the Residential Sector
154
PART V: COST OF LOADSHEDDING TO AGRICULTURE SECTOR
Table 34.1: Distribution of Cultivated Area between Irrigated and Barani Area
158
Table 34.2: Yield of Crops in Irrigated and Barani Area of Punjab
159
Table 34.3: Number of Hours Operated by Electric and Diesel Tube wells
159
Table 34.4: Water Availability at Farm Gate due to Different Sources, 2010-11
160
Table 34.5: Electricity Consumption per Agricultural Consumer
160
Table 35.1: Sample Size and Distribution
162
Table 35.2: Comparison of the Population and Sample Size Distribution of Farms
162
Table 35.3: Distribution of Sample by Districts
163
Table 36.1: Distribution of Sample Farms by Sources of irrigation
164
Table 36.2: Incidence of Use of Electric Tube wells
164
Table 36.3: Incidence of Loadshedding during Working Hours on Farm
164
Table 36.4: Reduction in Number of Water Rounds due to Power Outages
165
Table 36.5: Extent of Production Loss due to Reduction in Water Rounds
165
Table 36.6: Loss of Value Added due to Outages
166
Table 36.7: Cost of Loadshedding in Agriculture
166
Table 37.1: Incidence of Home-based Economic Activities in Sample Farms and Nature of
Adjustment
Table 37.2: Total Outage Cost
167
Table 38.1: Level of Satisfaction with Current quality of Service by DISCOs/KESC
169
Table 38.2: How much lower Tariff for existing level of Service
170
Table 38.3: Preference of Loadshedding Time
170
Table 38.4:Preference for the type of Loadshedding
171
Table 38.5: Information that can be provided by Distribution companies
171
Table 39.1: Suggestions by Sample Units by Province
172
168
LIST OF BOXES
PART II: COST OF LOADSHEDDING TO SMALL-SCALE INDUSTRY
Box 7.1: A Numerical Example of Quantification of Net Idle Factor Cost
42
Box 8.1: Capacity Utilization and Outages
48
Box 9.1: Incidence of Outages
53
xii
Box 9.2: Production Losses During Outages
55
Box 10.1: Extent of Recovery of Output by Use of Generators
61
PART III: COST OF LOADSHEDDING TO COMMERCIAL/SERVICES SECTOR
Box 16.1: A Numerical Example of Quantification of Net Idle Factor Cost
90
LIST OF FIGURES
PART I: INTRODUCTION AND ANALYSIS OF SECONDARY DATA
Figure 3.1: Loss of Consumer Surplus Due to Outages
22
Figure 5.1: Consumer Surplus from Electricity Consumption
35
PART II: COST OF LOADSHEDDING TO SMALL-SCALE INDUSTRY
Figure 7.1: Adjustment by a Firm to Outages
38
Figure 7.2: Flow Chart Showing Costs of Outages
40
Figure 8.1: Sampling Strategy
44
Figure 8.2: National Distribution by Province and Industry Group
45
Figure 8.3: Distribution of Selected Units by Industrial Group and Process
47
Figure 8.4: Actual as % of Target Production in 2012 by Industrial Group
50
Figure 11.1: Distribution of Outage Costs
65
Figure 11.2: Outage Costs as Percentage of Value Added
66
PART III: COST OF LOADSHEDDING TO COMMERCIAL/SERVICES SECTOR
Figure 16.1: Adjustment by a Firm to Outages
86
Figure 16.2: Flow Chart Showing Costs of Outages
88
Figure 17.1: Sampling Strategy
92
Figure 17.2: National Distribution by Province and Sector
93
Figure 17.3: Distribution of Selected Units by Sector
95
Figure 17.4: Actual as Percentage of Target Sale in 2012
98
Figure 20.1: Outage Costs as Percentage of Value Added
108
PART IV: COST OF LOADSHEDDING TO DOMESTIC/RESIDENTIAL SECTOR
Figure 26.1: Sampling Strategy
129
Figure 26.2: National Distribution of Population by Province
130
Figure 26.3: Distribution of Selected Units by Income Group
132
Figure 29.1: Outage Costs as % of Annual Household Consumption Expenditure
145
xiii
ACRONYMS
CMI
Census of Manufacturing Industries
DISCO
Distribution Company
GDP
Gross Domestic Product
GENCO
Generation Company
GOP
Government of Pakistan
GST
General Sales Tax
GWH
Giga Watt Hour
HIES
Household Integrated Economic Survey
IPP
Institute of Public Policy
KESC
Karachi Electric Supply Corporation
K-PK
Khyber-Pakhtunkhwa
Kwh
Kilowatt Hour
KWH
Kilowatt Hour
LDC
Least Developed Country
LDO
Light Diesel Oil
LESCO
Lahore Electric Supply Company
MEPCO
Multan Electric Power Company
MOF
Ministry of Finance
MW
Mega Watt
NEPRA
National Electric Power Regulatory Authority
NTDC
National Transmission and Despatch Company
O&M
Operations and Maintenance
OLS
Ordinary Least Squares
PBS
Pakistan Bureau of Statistics
PEPCO
Pakistan Electric Power Company
PES
Pakistan Economic Survey
PESCO
Peshawar Electric Supply Company
PSLSMS
Pakistan Social and Living Standards Measurement Survey
SBP
State Bank of Pakistan
TV
Television
UK
United Kingdom
UPS
Un-interrupted Power Supply
USA
United States of America
USAID
United States Agency for International Development
WAPDA
Water and Power Development Authority
WDI
World Development Indicators
WTP
Willingness to Pay
xiv
EXECUTIVE SUMMARY
Since 2008 the country has witnessed quantum deterioration in some key problem areas which
attained unprecedented heights. Pakistan is in the grip of major infrastructural shortages
impeding growth in all sectors including industries, agriculture and services. The industrial
sector faces severe energy shortages both in electricity and gas.
The year 2008 witnessed a major increase in the frequency and intensity of power loadshedding
or outages generally in Pakistan and in particular in the industrial sector. A manifestation of this
problem can be seen in the large number of reports in the popular press of high incidence of
outages and protests, by not only the domestic and commercial, but also industrial consumers.
We have also seen, complaints by the various chambers of commerce and industry and other
industrial associations in the country that the level of production in a number of industries has
been reduced due to the persistence of outages which apparently have fundamentally disturbed
the normal rhythm of the production cycle in a large number of industrial units, especially in
electricity-intensive sectors like textiles, non-metallic mineral products, basic metals, leather
products, rubber and plastic products, paper and paper products, etc.
The evidence supports their claim. At the macro level, the industrial sector of Pakistan has
shown a decline of about 2.6 percent during the 2008-12. Further, at the micro level, there is
some evidence of shortages in various manufactured goods as indicated by a big double-digit
increase in prices over the last few years. Clearly, there is need to undertake research to
quantify, in more precise terms, the impact of outages on the national economy. This exercise
was last done in 1989, when the problem of power loadshedding was also significant.
In this report we estimate the economic costs of power outages in the industrial, commercial/
services, agricultural and domestic sectors. The magnitude of the cost is a basic indicator of the
benefits that could be realized from investment and improved management of the power sector.
This summary highlights the impact of loadshdedding on the national economy as a whole and
also enables a comparison among the sectors.
The analysis is, in a large part, based on a survey of 1500 small-scale industrial, commercial,
agricultural and domestic consumers of electricity in about fourteen cities in Pakistan to yield
nationally representative results.
INCIDENCE OF OUTAGES
Loadshedding has reached unprecedented levels in Pakistan and severely impacted economic
activities. Table 1 show that the annual incidence of outages1 is as higher as 2623 hours
1
including restart times
1
annually in the case of small-scale industrial units followed by 2324 hours on farms and rural
households. Domestic consumers in urban areas face outages for 1975 hours. The incidence of
outages is somewhat less at 1697 hours in
the
case
of
commercial
consumers.
Effectively 20 to 30 percent of the
available
time
is
wasted
due
EXTENT OF SELF-GENERATION
Given the high level and pervasive nature
of outages, self-generation has become a
popular adjustment by economic agents.
60
percent
of
Sector
to
loadshedding.
Almost
Table 1
Average Hours of Incidence of Outages 2012
(over 24 hours daily)
commercial
Domestic
Urban
Rural
Commercial
Industry
Agriculture
*population weighted
**including restart time
***small-scale industry
Annual Hours of
Outages
1975*
1453
2324
1697**
2623***
2324
consumers have acquired self-generation followed by 24 percent of small-scale industrial units
and 11 percent in the case of urban
Table 2
Use of Generators
consumers (see Table 2). Generators are
generally not observed yet in rural areas.
Type of Consumer
Domestic
An interesting statistics is that consumers
Urban
of power in Pakistan now self-generate
Rural
Commercial
Industry
Agriculture
n = negligible
*% with UPS
**small-scale industry
almost 3300 Gwh, equivalent to 4.5
percent of the generation by the public
supply system. In the absence of selfgeneration the impact on output would
% of Units with SelfGeneration
11
28
(30)*
n
60
24**
n
have been this much higher.
IMPACT ON SECTORAL VALUE ADDED
The impact of loadshedding on value added in different sectors is sizeable. In agriculture, the
crop output has been diminished by almost 6 percent in 2011-12. This is one of the major
factors which explain the low growth rate on average of the crop sector in agriculture after 200708 of only 2.5 percent.
The outage cost as a percentage of value added is exceptionally high at 9.5 percent in the
industrial sector. This has been perhaps the single most important factor for the plummeting of
the growth rate of large – scale manufacturing to virtually zero after 2007-08, as compared to
over 10 percent per annum in the previous four years.
2
Despite the extremely high outage costs in relation to value added of almost 13 percent in 201112 in the case of small-scale manufacturing, the GDP estimates continue to show buoyancy of
the sector with an annual growth rate of 7.5 percent. This is very unlikely and at best the growth
rate is likely to be in the range of 4 to 4.5 percent.
The magnitude of costs of loadshedding is relatively low at 4.5 percent of value added in the
commercial/services sectors (see Table 3). But this impact is also not reflected in the national
income accounts and the Community, Personal and Social Services Sector is shown to be
growing at over 7.5 percent
per annum.
Overall,
the
impact
of
loadshedding on the growth
Table 3
Outage Costs as Percentage of Sectoral Value Added
(Direct Costs)
Sector
has been of a first order
magnitude.
The
Industry
314
Large-Scale
231
2661
8.7
Small-Scale
83
653
12.7
Commercial/Services
472
10571
4.5
2007-08 and 2011-12 is just
above 2.5 percent. It is likely
to be closer to 2 percent. In
the
Agriculture
reported
growth of GDP rate between
absence
of
(Rs in Billion)
Outage Costs as
% of Value
Added
2.3
(5.7)*
9.5
Sectoral
Value
Added
3899
(1574)*
3315**
performance of the economy
National
outage
Cost
89
*only the crop sector
**excluding slaughtering
loadshedding the economy
could have shown a growth rate of almost 4 percent.
NATIONAL ESTIMATE OF OUTAGE COSTS
The study on costs of loadshedding enables estimation of national costs of outages. The only
gap in the primary data collection was large-scale manufacturing. This gap has been filled with
estimates obtained in an earlier IPP study in 2009 (see Table 4).
An international comparison of the level of outage costs in Table 5 indicates that Pakistan has
comparable levels of loadshedding with Asian countries like Afghanistan, Philippines and
Vietnam.
The national cost of loadshedding, inclusive of both direct and indirect costs2, is a
staggering Rs. 1.4 trillion in 2011-12. This is even higher than the estimate obtained earlier of
2
See technical annexure I
3
Rs 1 trillion from secondary data. The outage costs are equivalent to 7 percent of GDP in
2011-12.
Within direct costs, the highest share, 42 percent, is incurred by the commercial/services
sectors, followed by industry at 28 percent.
Table 4
Summary of Results From the Loadshedding Project
National Outage Cost
(Rs in Billion)
I. DIRECT COSTS
A. Domestic Consumers
Urban Consumers
Rural Consumers
B. Commercial Consumers
C. Industry
Large-scale*
Small-Scale
D. Agriculture
Total
II. INDIRECT COSTS***
TOTAL NATIONAL COSTS OF POWER
OUTAGES
As % of GDP
*estimated from the earlier IPP study
**weighted by level of consumption
***estimated at 37% of the costs to the productive sectors
Source: USAID Loadshedding Project IPP (2009)
240
196
44
472
314
231
83
89
1115
324
1439
Outage Cost per Kwh
(in Rupees) (in cents*)
23
24
18
68
53
54
51
29
37**
23
24
18
69
54
55
52
30
38
7.0
Table 5
Comparison of Outage Costs as % of Sales in Selected Asian Countries
Country
Afghanistan
Bhutan
Fiji
Indonesia
Kazakhstan
Mongolia
Nepal
Philippines
Samoa
Timor-Leste
Turkey
Vietnam
Pakistan
*of value added
Source: World Bank
Year
2008
2009
2009
2008
2009
2008
2008
2008
2008
2008
2008
2009
2012
Outage Costs as % of Sales
6.5
4.3
4.9
2.2
3.7
0.8
27.0
3.9
6.6
7.6
2.8
3.6
7.0*
4
OUTAGE COSTS PER KWH
Table 4 indicates that the highest outage cost per kwh is in the commercial/services sector of
Rs 68 (69c3), followed by industry at 53 Rs (54c) per kwh. Relatively low outage cost is
observed in the case of domestic consumers at Rs 23 (23c) per kwh and agricultural consumers
ar Rs 29 per kwh.
These results on sectoral ranking of outage costs are consistent with the findings in other
studies as shown in volume I.
IMPACT ON EXPORTS
The estimated impact on exports is $ 3.5 billion, equivalent to about 14 percent of the actual
level of exports in 2011-12. Agricultural exports could have been higher by $ 1 billion and
manufactured exports by $ 2.5 billion.
In the presence of outages, the exportable surpluses are lower by 10 percent in rice, 33 percent
in fruits and vegetables, 80 percent in wheat, 76 percent in cotton and 20 percent in textiles.
IMPACT ON EMPLOYMENT
The loss of output due to outages is estimated to have resulted in a loss of employment of
almost 1.8 million. 39 percent of this loss is in agriculture, 25 percent in the industrial and 36
percent in services sectors respectively.
LOAD MANAGEMENT STRATEGY
The questionnaire contains a module to solicit consumer preferences regarding timing of
loadshedding which can reduce the costs and disruptions due to the outages. These can
provide guidance to the load management strategy by DISCOs, the formulation of which should
be a priority (see Table 6). These include suggestions like proper scheduling and adherence to
loadshedding; off-peak tariffs; lower incidence on continuous process industries a policy of
shorter but more frequent outages etc.
SUGGESTIONS BY RESPONDENTS
A number of interesting suggestions emanate from the survey responses. These can largely be
categorized as relating to the following: enhancing the supply of electricity, including
construction of new dams and import of electricity; alternative sources of energy/ fuel use
including resort to solar energy, bio-gas and nuclear power; improving governance or
3
with exchange rate of 1$ = 98 Rs
5
management including dealing with corruption and theft, and; changes in pricing policy including
subsidy on power and cost reduction.
Table 6
Indicators for Load Management Strategy
Domestic
Urban
 % of Time DISCOs kept to the
Announced Schedule
 % Satisfied* with Quality of
Services by DISCO
 % Additional Willingness to Pay
for Better Services
 % Willing to Change work
Timings if off-peak are reduced
 % Indicating Summer as worst
time for loadshedding
 % preferring longer outages with
less frequency**
* rating of ‘very high’ or ‘high’
**preferred timing is 6:00 am to 12 noon
Commercial
Small-Scale
Industry
Agriculture
43
14
28
-
7
2
3
1
29
36
30
12
n.a.
-
34
-
97
96
93
-
35
-
-
49
6
PART I
INTRODUCTION AND ANALYSIS
OF SECONDARY DATA
7
CHAPTER 1
INTRODUCTION
The widespread and growing phenomenon of power loadshedding has emerged as one of the
principal supply-side constraints to growth of the economy of the Pakistan. Not only has this led
to significant losses of output, employment and exports but also during periods of high outages
there have been large-scale protests, particularly in Punjab and K-PK. The magnitude of this
problem and the urgency to resolve it has motivated Advance Engineering Associates
International, Inc., to commission the institute of Public Policy, Beaconhouse National
University, Lahore to undertake the Study on the Economic Costs of Power Loadshedding
to Pakistan.
1.1 STATEMENT OF WORK
The assessment of economic cost of load shedding to Pakistan’s economy sets out the
following tasks:

Provide an approach and methodology for quantifying cost of load shedding to
Pakistan’s economy;

Explore GDP contribution of each relevant sector of the economy including commercial,
domestic, industrial and agricultural.

Determine consumption of electricity of each of these sectors

Determine direct cost of load shedding in each of these sectors (for example loss in
business due to unavailability of electricity in the industrial sector)

Determine national cost of load shedding using secondary data, primary data, and
in‐depth interviews with representative stakeholders including but not limited to electricity
consumers in domestic, commercial, industrial, and agriculture sectors;

Determine economic costs of using alternative sources of electricity, if possible in each
sector

Determine costs associated with alternative supply of electricity primarily through captive
generation on gas and generation on oil

Provide policy implications, conclusions, and recommendations as well as explore
economic load dispatch to assist in drafting policy to optimize efficient use of electricity
and minimize costs associated with load shedding and exploring other viable options.

This will involve the following:

Undertake Study on Economic cost of Load Shedding to Pakistan
8

Determine, collect and analyze secondary data

Conduct cost of outage survey

Review and analyze the data generated by survey

Carry out interviews of leadership and management of WAPDA, PEPCO, DISCOs,
GENCOs, and NTDC

Carry out interviews with other stakeholders in domestic, commercial, industrial, and
agriculture sectors

Have Focus Group Discussions if needed

Provide an appropriate approach and methodology for quantifying cost of load shedding
to Pakistan’s economy

Analyze and determine national cost of load shedding

Provide policy implications for minimizing the impact of load shedding

Make presentations to USAID and GOP officials

Prepare draft and final report for the study
1.2 METHODOLOGY
The approach to quantifying the cost of power outages has been developed in Pasha, Ghaus
and Malik [1989 and 1990] based on a big study sponsored by USAID in 1987. This
methodology has been considered as pioneering in character and is extensively used
internationally.
The proposed sample sizes of different types of power consumers are given below:
Large-Scale Industry
500
Small-Scale Industry
500
Agriculture
250
Commercial
250
Domestic
500
The proposed locations of the survey are as follows:
Sindh: Karachi, Hyderabad, Sukkur
Punjab: Lahore, Faisalabad, Sialkot, Gujranwala, Multan Rawalpindi
Khyber-Pakhtunkhwa : Peshawer, Mardan, Abbottabad
Balochistan : Quetta, Turbat
Islamabad
9
The survey will be sequenced and to the extent possible local surveyors will be recruited and
trained.
It is expected that the study will be completed in nine months.
1.3 THE OUTPUTS
The report will consist of the following parts:
Part 1: Introduction and analysis of secondary data
Part 2: Costs of Loadshedding to small scale industry
Part 3: Costs of Loadshedding to commercial/service establishments
Part 4: Costs of Loadshedding to Domestic consumers
Part 5: Costs of Loadshedding to Agriculture
10
CHAPTER 2
THE POWER SECTOR OF PAKISTAN
The objective of this chapter is to highlight the key trends and facts about the economy of
Pakistan and the power sector respectively. This will highlight the contribution of different factors
to the emergence of high levels of loadshedding in the country and the impact thereof on
economic growth.
The Chapter is organized as follows: Section1 contrasts the growth rate of the economy prior to
and after the commencement of significant loadshedding in 2007-08. Section 2 presents an
inter-country comparison of access to electricity in order to highlight the relative level of
development of the power sector of Pakistan. Section 3 identifies the long term trends in the
sector, while Section 4 gives the growth in and pattern of electricity consumption. Finally,
Section 5 quantifies the supply gap in the provision of power and the implied level of
loadshedding.
2.1 STRUCTURE AND GROWTH OF THE ECONOMY
The economy exhibited considerable dynamism during the earlier years of the last decade,
especially from 2002-03 to 2006-07. But from 2007-08 onwards the growth rate has declined
sharply up to 2010-11 from almost 6% annually to below 3% as shown in Table 2.1. In 2011-12
there has been a modest recovery.
The plummeting of the GDP growth rate coincides with the emergence of loadshedding. But
while this supply-side factor has constrained growth, it is not the only factor. There was a major
financial crisis globally in 2008 and Pakistan’s exports were adversely affected. Subsequently,
with the intensification of war on terror, security conditions have deteriorated and heightened
perceptions of risk have led to a steep fall in private investment, which has virtually hit an all
time low level. On the top of all this, the devastating floods of 2010 led to large losses of output,
especially in agriculture. Overall, it is clear that in the absence of power load shedding the
economy would not have achieved 6% growth, but in the presence of problem the fall in growth
rate has been greater.
Table 2.1 shows that the sharpest decline in the growth rate has been in the industrial sector
from 7 to 2.5 percent. Industry is a major consumer of electricity. The only sector which has
shown high growth is the government, especially in current expenditure.
11
Table 2.1
Growth Rate of Different Sectors and of the GDP
(Rs in Billion at 1999-2000 prices)
2000-01
903.5
2007-08
2011-12
1148.9
1269.5
(3.5) *
(2.5)
Industry
865.2
1387.1
1531.2
(7.0)
(2.5)
Commercial/Services
1638.2
2526.4
2829.0
(6.4)
(2.9)
Government
225.1
320.6
398.9
(5.2)
(5.6)
GDP(fc)
3632.0
5383.0
6028.6
(5.8)
(2.9)
*Figures in brackets are annual growth rates from 2000-01 to 2007-08 and from 2007-08 to 2011-12
respectively
Source: PES
Agriculture
The
contribution
of
different
sectors to the GDP is indicated
in Table 2.2. The process of
Table 2.2
Contribution of Different Sectors to GDP
(at constant prices of 1999-2000)
(%)
2000-01
24.9
11.8
11.9
1.2
23.8
2.4
10.3
2007-08
21.3
9.5
11.2
0.6
25.8
2.6
13.4
2011-12
21.1
8.8
11.6
0.6
25.4
2.4
11.9
5.4
5.8
6.7
communications and wholesale
Electricity and Gas
Construction
Commercial/Services
Transport and Communications
Whole sale & Retail Trade
Finance & Insurance
Ownership of Dwellings
2.9
2.4
45.1
11.6
17.9
3.1
3.2
1.6
2.4
47.0
10.0
17.4
6.3
2.7
2.2
2.2
46.9
9.6
17.1
4.8
2.7
and retail trade. A comparison of
Social and Personal Services
9.3
10.6
12.6
6.2
6.2
100.0
6.0
6.0
100.0
6.6
6.7
100.0
structural change which was
observed in the period of high
growth
has
been
largely
arrested as shown in Table 2.2.
Industry now accounts for about
one-fourth of the economy, while
the
agricultural
sector
contributes just over one-fifth to
the GDP. The largest part of the
economy is accounted for by
commercial/services
especially
in
activities,
transport
and
the sectoral contribution to the
GDP with the share in power
consumption is made later.
Agriculture
Crops
Livestock
Other
Industry
Mining
Large-Scale Manufacturing
Small-Scale
Manufacturing**
Government
Public Admin & Defence
Total GDP
Source: PES, MOF
12
2.2 INTERNATIONAL COMPARISONS
A comparison is made of key indicators of the power sector in different Asian countries, both in
South Asia and East Asia. There is apparently a strong correlation between per capita electricity
consumption and the level of development. Pakistan has both relatively low energy consumption
and per capita income as compared to India and Sri Lanka in South Asia and virtually all East
Asian countries.
The somewhat early stage of development of the power sector in Pakistan is also indicated by
the low share of population with access to electricity at 62% as compared to 66% in India, 77%
in Sri Lanka, 90% in Philippines and 99% in Malaysia.
It is also observed that the energy intensity in production increases as development takes place.
This is confirmed by Table 2.3 which shows that the GDP per Kwh is $ 2.3 in China as
compared to $ 11.5 in an LDC like Nepal. Pakistan is at an intermediate level at $ 5.3.
Table 2.3
Indicators Of Electricity Consumption In Selected Countries
(2009)
Per capita GDP,
PPP (at constant
2005 prices)
Country
Bangladesh
1419
China
6206
India
2813
Indonesia
3696
Malaysia
12526
Nepal
1048
Pakistan
2357
Philipines
3364
Sri Lanka
4301
Thailand
7160
Vietnam
2721
Source: World Bank, WDI
% of Population
with Access to
Electricity
Electricity
Consumption per
Capita (Kwh)
GDP per Kwh ($)
41.0
99.4
66.3
64.5
99.4
43.6
62.4
89.7
76.6
99.3
97.6
251.6
2631.4
570.9
590.2
3613.5
91.0
449.3
593.5
412.9
2044.8
917.6
5.64
2.36
4.93
6.26
3.47
11.52
5.25
5.67
10.42
3.50
2.96
The level and sources of electricity generation and magnitude of system losses (in transmission
and distribution) are presented in Table 2.4. There is a wide variation among countries in
sources of power. Countries which rely most on coal include China, India and Indonesia.
Hydroelectricity is the dominant source in Nepal, which given its location in proximity to the
Himalayas has enormous untapped potential for hydro-electricity, like Pakistan.
13
Table 2.4
Level and Sources of Electricity Generation and System Losses
in Selected Countries
(2009)
Per Capita
Electricity
Production (Kwh)
Country
Bangladesh
257.6
China
2776.0
India
744.7
Indonesia
654.9
Malaysia
3759.7
Nepal
106.1
Pakistan
559.2
Philipines
675.3
Sri Lanka
483.3
Thailand
2159.9
Vietnam
967.3
Source: World Bank, WDI.
System
Losses
(%)
Sources of Electricity
Coal
(1)
1.7
78.8
68.6
41.8
30.9
0.0
0.1
26.6
0.0
19.9
18.0
Hydro
(2)
4.1
16.7
11.9
7.3
6.3
99.6
29.4
15.8
39.5
4.8
36.0
Gas
(3)
89.4
1.4
12.4
22.1
60.7
0.0
29.4
32.1
0.0
70.7
43.4
Nuclear
(4)
0.0
1.9
2.1
0.0
0.0
0.0
3.0
0.0
0.0
0.0
0.0
Oil
(5)
4.8
0.4
2.9
22.8
2.0
0.4
38.0
8.7
60.3
0.5
2.5
2.3
5.2
23.3
9.9
3.9
14.2
19.7
12.1
14.6
5.3
5.1
Gas is the principal source in Bangladesh, Malaysia, Philippines, Thailand, and Vietnam until
recently, Pakistan. Thermal power has emerged as a major source in Sri Lanka and Pakistan.
The pattern of sources of electricity of Pakistan exposes the country to two major risks.
Dependence on gas at a time when reserves are depleting implies severe constraints in the
medium to long run and the need to switch to other sources. The growing reliance on thermal
power exposes the economy to ‘oil price shocks’.
Turning to system losses, a key indicator of efficiency, both India and Pakistan do poorly at 23%
and 20% respectively. Even an LDC like Bangladesh performs much better with only marginal
losses. Similarly, other countries like China, Malaysia, Thailand and Vietnam have been
successful in containing system losses.
2.3 LONG-TERM TRENDS IN POWER SECTOR
The growth in installed capacity and generation of electricity in Pakistan is presented in Table
2.5 since 1970-71. The former has been more than doubling every decade up to 2000-01, with
annual growth rate over 7%. It is only during the last decade that the rate of expansion in
capacity has substantially slowed down to 3% per annum. In the initial years of the decade there
was significant excess capacity, due to the hump in investment by the IPPs in the mid-to late90s. But adequate provisions were not made to cater for the future growth in demand.
14
The growth in electricity generation was rapid in the 70s and 80s. in particular, the
commissioning of the Tarbela Dam in the early 80s enabled a quantum jump in supplies at low
cost. Consequently, electricity generation grew annually by almost 10%. During the 90s as the
growth rate of the economy slowed down, demand for electricity was not so buoyant and the
rate of increase annually in power generation declined to 5%. During the last decade, this has
fallen further to only 3%.
An index of capacity utilization4 is constructed in Table 2.5. The rate of capacity utilization
exceeded 100% by 1990-91 and the loadshedding which occurred in a significant way in the
mid-to-late-80s can be attributed to a shortage of capacity. It was during this period that the first
study in Pakistan on costs of loadshedding was undertaken by Pasha, Ghaus and Malik [1989],
with support from USAID. As opposed to this, the upsurge in loadshedding once again since
2007-08 can be attributed primarily to a lack of full capacity utilization arising from lack of
adequate maintenance of older plants and liquidity problems due to the ballooning of circular
debt.
2.4 GROWTH AND PATTERN OF ELECTRICITY CONSUMPTION
The growth in electricity consumption by type of consumer during the last decade is presented
in Table 2.6. The analysis is broken up into two sub-periods, the years prior to commencement
of significant loadshedding in 2007-08 and the years thereafter. In the latter period, the overall
growth rate in power consumption has fallen to 2% only due to progressively higher levels of
loadshedding. A comparison of Table 2.5 with Table 2.6 also indicates that system losses in
transmission and distribution were 29% in 2000-01 and 19% in 2010-11.
Table 2.5
Long-Term Trend in Capacity and Generation in Pakistan of Electricity 1970-71 to 2010-11*
Installed
Capacity (MW)
Annual
Growth Rate
(%)
Electricity
Generation
(GWH)
Annual
Growth Rate
(%)
1970-71
1862
7202
1980-81
4105
8.2
16062
8.4
1990-91
8356
7.4
41042
9.8
2000-01
17498
7.7
68117
5.2
2010-11
22447
2.5
94653
3.3
*Figures for 2011-12 were not finalized at the time of preparation of this report.
Source: Handbook of Statistics, SBP, Pakistan Economic Survey, MOF, GOP
4
Index of
Capacity
Utilization (%)
81
82
102
81
88
300 days operation with 16 hours daily.
15
Table 2.6
Growth in Electricity Consumption from 2000-01 to 2010-11
(GWH)
Domestic
Industrial
Commercial
Agricultural
Others*
Total
2000-01
22765
14349
1774
4924
3773
48585
2007-08
33704
20129
5572
8472
4923
73400
2010-11
35885
21207
5782
8971
5254
77099
2000-01 to 2007-08
5.8
5.4
10.5
8.1
6.7
6.1
2007-08 to 2010-11
2.1
0.8
0.5
1.9
2.2
1.6
2001-01 to 2010-11
4.7
4.0
7.6
6.2
3.4
4.7
Growth Rate (%)
*
mostly government, street lights and traction
Source: PES
The fastest growth in consumption during the last decade is observed in the case of commercial
consumers followed by agriculture. Consumption by industrial consumers has grown less rapidly
at 4% per annum and by less than 1% since 2007-08. The power shortage is clearly a major
factor limiting industrial growth currently.
Table 2.7 highlights the relationship between growth in sectoral value added and growth in
electricity consumption. The rise in the rate of response of production to an increase in supply of
electricity after 2007-08 indicates that efforts are being made to increase energy efficiency and
invest in self-generation of power to at least partially sustain levels of production.
Table 2.7
Growth Rate of Sectoral Value Added and Electricity Consumption
(%)
Agriculture
Growth
Rate of
Value
Added
3.5
2000-01 to 2007-08
Growth Rate
Elasticity*
of Electricity of
Consumption
8.1
0.432
Growth
Rate of
Value
Added
2.3
2007-08 to 2010-11
Growth Rate
Elasticity*
of Electricity
Consumption
1.9
1.211
Industry
7.0
5.4
1.296
2.2
0.8
2.750
Commercial
6.4
10.
0.610
2.4
0.5
4.800
Government
5.8
6.7
0.866
6.6
1.9
3.473
Source: Derived
*Growth rate of value divided by growth rate of electricity consumption.
16
The share of different types of consumers in power consumption is given in Table 2.8. The
largest category continues to be domestic consumers, with a share of almost 47% in 2010-11.
The next group is industrial consumers whose share has fallen from about 30% in 2000-01 to
28% in 2011-12. As such, these two groups of consumers account for almost three-fourths of
the consumption of electricity in the country. The remainder is shared between agricultural
consumers (12%), commercial consumers (7%) and others (7%).
The impact of load shedding in recent years is clearly demonstrated in Table 2.9 by the fall in
energy consumption per consumer after 2007-08 in the case of all consumers, with the largest
decline in the case of industrial consumers of over 8%.
Table 2.8
Share of Different Types of Consumers in Electricity Consumption
(%)
Domestic
Industrial
Commercial
Agricultural
Others*
Total
2000-01
46.9
29.5
5.7
10.1
7.8
100.0
2007-08
45.9
28.2
7.6
11.5
6.7
100.0
2010-11
46.4
27.5
7.5
11.6
6.8
100.0
Year
Source: NEPRA, State of Industry Reports
Table 2.9
Growth in Electricity Consumption per Consumer
(%)
Growth rate (%)
Type
2000-01 to 2007-08
2007-08 to 2010-11
Domestic
1.7
-6.0
Industrial
-1.2
-8.1
Commercial
21.8
-3.9
Agricultural
-0.7
-7.2
Others
20.6
-7.5
Total
0.5
-6.6
Source: NEPRA, State of Industry Report
A comparison is also made for each sector between the share in the GDP and the share in
electricity consumption in table 2.10. The ratio in the table is a measure of the electricityintensity of a sector. This is the highest for industry followed by government. The least
electricity-intensive sector is commerce.
17
Table 2.10
Comparison by Sector of share in GDP and share in Electricity Consumption*
(%)
2011
Sector
Share in GDP
(1)
Industry
25.4
Share electricity
consumption*
(2)
51.4
Commerce
46.7
14.0
0.30
Agriculture
21.2
21.7
1.02
other**
6.7
12.8
1.91
*
Ratio
(2)÷(1)
2.02
excluding domestic consumers
**
mostly government
Source: derived
2.5 THE SUPPLY GAP
The surplus/deficit between demand and supply during system peak hours for NTDC and KESC
combined is given Table 2.11. The supply gap was 1912 MW in 2007 which has risen to 5656
MW, equivalent to 27 % of demand5. It is important to note that in 2010-11 NEPRA reports the
generation capability as only 69% of the installed capacity. The former has remained, more or
less, unchanged since 2007.
Table 2.11
Surplus/Deficit in Demand and Supply during System* Peak Hours
2007
Generation
Capacity
15575
2008
14707
19281
-4574
24
2009
16050
20304
-4254
26
2010
15144
21029
-5885
28
2011
15430
21086
-5656
27
Demand
Supply-Gap
%
17487
-1912
11
*
NTDC and KESC combined
Source: NEPRA, State of Industry Report
According to NEPRA, the highest incidence of outages regionally is in the area served by
MEPCO, PESCO and LESCO6. The least outages are in IESCO. Most areas of Punjab and
Khyper-Pakhtunkhwa are more vulnerable to load shedding.
5
6
This gap is based on little or no growth in demand in 2011.
The distribution companies in Multan, Peshawar and Lahore designated areas.
18
CHAPTER 3
METHODOLOGY FOR QUANTIFICATION OF COST OF LOADSHEDDING
Various approaches have been developed in the literature for quantification of the cost incurred
by different types of consumers as a result of power outages. These approaches vary greatly in
terms of data requirements and level of complexity. The Chapter starts with the simple value
added approach and ends with the full-blown survey based and contingent valuation
approaches.
3.1. THE SIMPLE VALUE ADDED APPROACH
A relatively high estimate of the cost of loadshedding is as follows:
Vi = Value added by sector i in absence of Loadshedding
Ei = Electricity consumption in the absence of loadshedding
Then the cost Ci, of loadshedding is give by
𝐶𝑖 =
𝑉𝑖
𝑙 ……………………………………………………….(1)
𝐸𝑖 𝑖
Where 𝑙𝑖 is the quantum of electricity not supplied due to outages. Summing across sectors,
the total cost of loadshedding is given by
𝐶 = ∑𝑛𝑖=1
𝑉𝑖
𝑙 ……………………………..………………….(2)
𝐸𝑖 𝑖
Where n is the number of sectors.
This approach can be applied on the production sectors of the economy, viz, agriculture,
industry and commerce, but not to domestic consumption of electricity.
The reasons why this approach leads to a high estimate of the cost of Loadshedding are as
follows:
(i)
It does not distinguish between the average and marginal productivity of the
electricity input, that is, there could be some economies of scale in the use of energy.
(ii)
It assumes that output lost is proportional to the extent of electricity not supplied and
the firms do not make adjustments to recover at least part of the output.
As opposed to the above, an approach that yields a low estimate is one which focuses only the
wage cost, on the assumption that the idle factor during outages is labor. As such, in this case
19
𝐶𝑖 =
𝑊𝑖
𝑙 ……………..…………………………………………..(3)
𝐸𝑖 𝑖
Where W i is the wage bill.
3.2. THE ADJUSTED VALUE ADDED APPROACH
This approach postulates the marginal cost of unsupplied elasticity is different from the average
cost as given in (1) above. Accordingly,
𝜕𝑉𝑖
𝜕𝐸𝑖
=𝛽
𝑉𝑖
𝐸𝑖
𝛽>0
………………….…………………. (4)
𝛽 is estimated on the basis of the historical relationship between value added and electricity
consumption. Generally, it is observed that 𝛽 <1.
However, the value added approaches suffer from the defect that they do not allow for spoilage
costs arising from damage to materials that takes place at the time when the outage occurs,
especially if there is no prior notice.
3.3 MARGINAL COST OF UNSUPPLIED ELECTRICITY
It has been argued by Bental [1982] that by observing a firms behavior with respect to the
acquisition of own generating power, the marginal cost of unsupplied electric energy may be
inferred. A competitive risk-neutral firm equates, at the margin, the cost of generating a Kwh on
its own to the expected gain due to that Kwh. This expected gain is also the expected loss from
the marginal Kwh which is not supplied by the utility. Therefore, the marginal cost of generating
its own power may serve as an estimate of the marginal outage cost.
The cost to a firm of generating its own power consists of the two elements. The first part is the
yearly capacity cost of the generator. This can be represented as follows:
K(c)
= annual capital cost (depreciation + interest cost) of a generator with capacity in kva
In addition,
VC
= variable cost per Kwh, consisting mainly of fuel cost
l
= hours of outages
The marginal cost, MC of self-generation per Kwh is given by
𝜕𝐾(𝑐)
𝑀𝐶 =
𝜕𝑐
+ 𝑣𝑐 ……………………………….. (5)
On the assumption that the MC is constant, the total cost, TC, of Loadshedding is given by
20
… … ………………………………….. (6)
𝑇𝐶 = 𝑀𝐶. 𝑙
This approach may not lead to proper estimates in the following cases:
(i) Presence of economies/ diseconomies of scale in the capital cost of generators such
that
𝜕𝐾(𝑐)
𝜕𝑐
is not constant.
(ii) Imperfections in the capital market whereby firms, especially the smaller ones, are
unable to borrow for acquisition of a generator.
(iii) In Pakistan previous surveys of firms, for example by the Institute of Public Policy [2009],
indicated that not all units have self-generation. This implies that the marginal cost of
outages is lower than the marginal cost of a generator. For such units, this method
cannot, therefore be applied.
3.4. THE VALUE OF LEISURE APPROCH
M. Munasingha [1980] has proposed a novel approach for evaluating the cost of outages to
residential consumers, as the value of leisure foregone. According to this approach, the
principal outage cost imposed on a household is the loss of leisure during the evening hours
when electricity is essential. During the day time there is sufficient slack in the execution of
household activities that are interrupted by the outage, such as cooking or cleaning, to permit
rescheduling of these activities without causing much inconvenience.
As such, the monetary value of this lost leisure is equal to income earning rate on the basis of
consumers’ labor–leisure choice. Munashinghe accordingly computes the cost per Kwh of
unsupplied electricity as
𝐶=
𝑦
… … ……………………………………(6)
𝑘
Where y is the hourly income and k the normal level of electricity consumed per hour in the
absence of outages. Therefore, the total cost of outages to residential consumer is, C, where
𝐶=
𝑦
𝑘
.𝑙
… … …………………………………..(7)
A principal practical advantage of this method of estimating outage costs for residential
consumers is that it relies on the relatively easy-to-obtain data. But for proper application of this
method it is essential to have the levels of electricity consumption by households at different
income levels.
Other problems with this approach include the following:
21
(i) It assumes that the income earner in the household has flexible working hours so that
he/she can effectively exercise his/her labor-leisure choice. This may be true in the case
of self-employed persons. But for wage earners who work fixed hours, the marginal
value of leisure is unlikely to be equal to the income rate per hour. As such, some
authors have preferred to apply this approach by assuming that the value of leisure is
only a fraction of income.
(ii) It ignores the presence of household economic activities like cottage industry or
sewing/embroidery work by women, especially in lower income households. This is
sometimes the case in Pakistan. Such, activities may not readily be rescheduled in the
presence of outages, especially if they are of long durations. As such, in these cases the
cost of outages must include the value of lost output.
(iii) Outages, especially when accompanied with voltage fluctuations, can damage homebased appliances like TV, refrigerator, air-conditioner, freezer, etc. Cost has to be
incurred to repair the damage. These are equivalent to spoilage costs and should be
included in the cost of loadshedding.
3.5. THE CONSUMER SURPLUS APPROACH
This is relatively popular approach and has been applied by Sanghvi [1982]. The demand
curve for electricity captures the willingness to pay for the service and the consumer surplus
of electricity supply is represented by the area between the demand and supply curves. The
loss of consumer surplus due to supply interruptions is represented by the shaded area,
ABE, in Figure 3.1 below.
Figure 3.1
Loss of Consumer Surplus Due to Outages
22
The prime magnitude required for application of this approach is the price elasticity of demand,
which is not possible to measure in the presence of outages. Also, given a non-linear schedule
of power tariffs, as is the case with residential consumers in Pakistan, the magnitude of the
consumer surplus lost due to outages becomes difficult to quantify. Further, if AB is large then
the consumer may be able to reduce the loss by investing in self-generation. This becomes
more attractive the larger the amount of electricity not supplied.
3.6. THE CONTINGENT VALUATION (WTP) APPROACH
This approach involves asking consumers their willingness to pay for more reliable supplies of
power. For example, the question could be as follows:
If the incidence of outages is reduced to half its present level, how much more would you
be willing to pay on your monthly electricity bill?
An alternative approach is to ask the following question:
If level of outages were to double, what reduction in your monthly electricity bill would
you consider to be fair?
The contingent valuation approach is prone to giving biased estimates as it is based on
subjective responses. It is likely that in response to the first question the consumer understates
his willingness to pay for improved service, while he may overstate the compensation that he
would like to receive for deterioration in the reliability of supply.
3.7. THE SURVEY BASED APPROACH
The most comprehensive approach to quantify the cost of outages is to undertake a random
survey of affected consumers. This enables explicit and direct determination of different
components of outage costs including the spoilage cost, idle factor cost and adjustment cots.
However, the survey based approach is more costly than approaches which rely largely on
secondary data. Also, the possibility of a bias cannot be ruled out by the respondents who may
exaggerate the costs in order to attract greater attention to the problem of loadshedding.
3.8. RESULTS OF INTERNATIONAL STUDIES
The results on estimates of outage costs per Kwh in studies done in other countries are
presented in Table 3.1. The principal conclusions are as follows:
(i)
Much of the interest in deriving the cost of outages has been in the USA, primarily in the
decade of the 70s. Other countries where estimates have been made include the UK,
New Zealand, Canada, Finland and Taiwan. There are some studies on developing
countries like Pakistan, Egypt, Nigeria, Chile and Zimbabwe.
23
(ii) The focus of the studies has been mostly on the cost to residential, commercial and
industrial consumers. There are few studies on agricultural consumers.
(iii) The survey-based approach has been used primarily in the cases of industrial and
commercial consumers, whereas indirect proxies have been used for domestic
consumers.
(iv) Among different types of consumers, the highest outages cost per Kwh appears to be in
the case of commercial consumers, followed by industrial consumers. These costs
emerge as relatively small for domestic and agricultural consumers.
(v) Within a consumer category, like industrial or commercial consumers, the highest cost
estimate is yielded by the survey-based approach. For the residential sector, the largest
cost is produced by the value of leisure approach. As highlighted earlier, this probably
represents a significant overstatement.
(vi) Outage costs per Kwh are generally a large multiple of the power tariff in public supply.
Table 3.1
Cost of Outages per Kwh to Different Types of Consumers in Various Countries
Approach
RESIDENTIAL SECTOR
Value of Leisure
WTP Survey
Standby generator cost
Consumer Surplus Loss
Estimated by Survey
INDUSTRIAL SECTOR
Estimated by Survey
Cost of Wages Lost
Value Added Lost
COMMERCIAL SECTOR
Cost of Wages Lost
Estimated by Survey
AGRICULTURE
Estimated by Survey
Authors
Average
outage cost
per Kwh ($)
Sheppard[1965:UK]; Lundberg[1970:UK]; Yaborf of
[1980:USA];
Munasinghe
[1977:Brazil];
Turner
[1977:New Zealand]
Lundberg[1970:Sweden]; Ontario Hydro [1977]; Finnish
Power Producers Council [1978]; Yaborff[1980:USA];
Faucett [1979:USA]; Systems[1980,USA]
Sanghvi[1980,USA]
Sanghvi[1980,USA]
Balducci[2002,USA]
2.10
Swedish
Electric
Council
[1969];
Modern
Manufacturing[1969:USA];
Gannon/IEEE[1971:USA];
IEEE[1973:USA]; UNIPEDE [1970 Sweden]; Jackson&
Salvage[1974:UK];
Ontario
Hydro
[1977];
Sanghvi[1980,USA]; Balducci[2002,USA];
Sheppard[1965:UK]; Turner [1977 New Zealand]
Lundberg[1970:UK];
Taiwan
Power
Co[1975]
Yaborf[1980:USA];
4.25
Sheppard[1965:UK]; Turner [1977 New Zealand]
Yaborf[1980:USA];
Lundberg[1970:Sweden];Patton[1975:USA];
Congressional Research Service [1979:USA]; Ontario
Hydro [1978]; Finnish Power Producers Council [1979];
Balducci[2002,USA]
3.30
Balducci[2002,USA]
0.65
1.04
2.25
0.58
0.75
2.21
3.46
6.65
24
At this stage, some conclusions may tentatively be drawn about the cost of Loadshedding in
Pakistan from the above findings. First, these costs are likely to be limited by the share (47%) of
domestic consumers in electricity consumption, where outage costs appear to be low. Second,
approaches based on secondary data may yield lower magnitude of costs as opposed to
primary data obtained from a survey of units.
25
CHAPTER 4
EARLIER ESTIMATES OF COSTS OF LOADSHEDDING IN PAKISTAN
The previous Chapter has highlighted the various approaches that can be used for deriving the
cost of outages. This Chapter summarizes the results of various studies in Pakistan undertaken
earlier to quantify the costs of load shedding.
4.1 THE PASHA, GHAUS AND MALIK STUDY
This study was undertaken in 1987 when load shedding first emerged as a problem, with
support from USAID. It was a pioneering study undertaken in the context of a developing
country and was subsequently published in the prestigious international journal, Energy
Economics. This study is extensively quoted in the literature.
The study involved an in-depth survey of 843 units, randomly selected from different locations in
Pakistan. Table 4.1 reveals the distribution of costs due to power outages, with differentiation
between planned and unplanned outages.
Table 4.1
Break-up of Outage Costs
(% of units)
Planned Outages
Unplanned Outages
Direct Cost
87.8
89.1
Spoilage Cost
54.8
54.7
Idle Factor Cost
33.0
34.4
Indirect Cost
12.6
10.1
Labor Related Costa
3.2
5.3
Costb
8.4
5.8
Timing Related Costc
1.0
0.0
100.0
100.0
Capital Related
Total Cost
a
b
c
Cost of additional overtime or shifts
Cost of generators or more intensive operation of machinery
Cost of changes in shift timings or working days
Bulk of the costs was in the form of direct costs, including spoilage and idle factor costs.
Adjustment (indirect) costs were limited, especially since, at the time of the survey, only 12% of
the units had acquired self-generation capacity7 . Other adjustments included overtime work by
7
The marginal cost of self-generation was 250% of the power tariff
26
26% of units, working additional shifts by 8% and changing shift timings or working days by 2%
respectively.
The estimated cost of loadshedding per Kwh in different industries is given in Table 4.2 A
number of key conclusions emerge from the Table. First, the outage costs due to unplanned
outages were substantially higher by 75 percent in relation to costs associated with planned
outages. Second, outage costs in continuous process industries were five to six times higher
than in batch-making industries. Third, there is substantial variation in outage costs among
industries, ranging from a minimum of 24 cents to 185 cents per Kwh.
Table 4.2
Cost of Planned and Unplanned Outages per Kwh in different Industries
Planned Outages
Cents/KWH
Unplanned Outages
Food, Beverages and Tobacco
54
164
Textiles
23
29
Wearing Apparel &Footwear
62
70
Chemicals
57
102
Wood and Paper
69
178
Non-Metallic Mineral Products
23
24
Metal and Metal Products
35
72
Machinery and Equipment
139
185
Other Industries
29
54
Batch-making
20
38
Continuous Process
119
200
Total
36
63
Industry Group
By Process
Overall, it appears that in 1987 the average cost of outages to industrial consumers in Pakistan
was approximately 50 cents per Kwh. This is significantly lower than estimates for developed
countries, highlighted in Table 3.1.
Based on the same data set, the authors published another paper on the differentiated impact
(by size of unit in terms of employment) of load shedding. The survey revealed that the
incidence of load shedding was higher for small units (employing up to 10 workers) at 230 hours
27
annually as compared to 179 hours for medium sized units (employing more than 10 workers up
to 50 workers) and 103 hours in the case of large units (employing more than 50 workers).
The pattern of adjustment also appeared to vary by size of firm, as shown in Table 4.3. Only 2%
of small units had acquired a generator as compared to 22 percent in the case of large units.
Similarly, while 18% of small units increased overtime work, over 31% of large firms opted for
this form of response to loadshedding. The extent of recovery of lost output was the highest at
63% by large units.
Table 4.3
Pattern of Adjustment by Size of Unit
(%)
Small
Medium
Large
Utilising Capacity more Intensively
26
29
22
Working Overtime
18
30
31
Working Additional Shifts
8
7
6
Changing Production Techniques
21
19
11
Buying Generators
6
9
22
Changing Shift Timings
2
2
0
Changing Working Days
2
3
1
Making Some Adjustment
69
79
87
Extent of Recovery of Output
35
63
40
Overall, it is clear that power loadshedding had a larger adverse impact on small and mediumsized units as compared to large units. This is attributable not only to the higher incidence of
outages but also to the relative lack of flexibility to make adjustments with small units.
The study also derived the impact of the economy-wide multiplier, that is, the second round
consequences of outages in terms of impact on aggregate demand. The multiplier had a
magnitude of 1.37, imply that the economy-wide impact was 37% higher than the cost incurred
directly by production units.
4.2 THE IPP STUDY
This study was undertaken in 2009, when concerns started being voiced about the alarming rise
in loadshedding in the country, especially in Punjab. The research study was carried out by the
same team, Pasha and Ghaus, which had carried out the 1987 study. However, the sample was
28
substantially lower, due to budget constraints, at 65 and limited only to large–scale industrial
units in the major cities of Punjab.
The survey revealed that the level of load shedding were substantially higher at over five times
the level prevailing in the mid-80s. It was estimated at an average of 3.8 hours per day, with the
peak level being observed during the months of December to March. The sample firms reported
that over 20% of the time that they could have been in production was lost due to loadshedding.
The incidence of adjustments was greater than in 1987 due to the high level of loadshedding.
75% of the firms were engaged in self-generation; 18% worked more overtime and 15% worked
additional shifts.
The value added loss was considerably larger for units without self-generation. The overall cost
of loadshedding to industry was estimated at Rs 157 billion equivalent to 9% of national value
added. Inclusive of the multiplier of effect the cost was Rs 230 billion.
The implied outage cost per Kwh was estimated at approximately at 50 cents. This is close to
the 1987 estimate, despite higher incidence of load shedding and more frequent adjustments.
4.3 THE PIDE STUDY
This study was undertaken by R. Siddiqui, H.H.Jalil, M.Nasir, W.S. Malik and M. Khalid of PIDE
and published in 2011. A survey was undertaken of 339 firms in four cities of Punjab. It
excluded small units (employing 10 or less workers).
The study estimated that the sample units on average lost 3.3 hours per day. This is close to the
2009 estimate by IPP. Almost 25% of the firms lost more than 5 hours a day.
In terms of adjustments, this survey also revealed a high frequency, with 76% of the units
having opted for stand-by generators. 69% of sample firms reported delays in delivery of orders.
The methodology used for quantifying the cost of loadshedding consisted primarily of deriving
the idle factor cost of labour only and excluding spoilage or other adjustment costs. As such, the
cost is understated. The cost is estimated on the basis of assumed length of shift at 8, 10, or 12
hours.
The resulting loss in value of production is estimated at Rs 400 billion for Punjab with an 8 hour
shift and Rs 267 billion with 12 hour shift. Conversion to value added and blowing up the sample
for Pakistan as a whole, the estimate of cost of loadshedding is Rs 176 billion for the large-scale
29
manufacturing sector, equivalent to 12% of the total national value added by the sector. As
such, this estimate is somewhat higher than that by IPP in 2009.
4.4 CONCLUSIONS
The above studies enable the following conclusions:
i.
Levels of loadshedding have risen to unprecedented levels in recent years and on
average firms have experienced outages between 20 to 25% of their working hours.
ii.
Adjustments of various types have been made by a high proportion of firms, including
investment in self-generation by over three-fourths of the units.
iii.
The costs of loadshedding to industry have reached a high level, equivalent to 10% or
more of national sectoral value added or almost 2% of the GDP. These losses have
been accompanied by significant declines in profitability, employment and exports.
iv.
The average cost of outages to industry is about 50 cents per Kwh. This is high in
relation to the existing level of power tariffs. But it is low by international standards.
30
CHAPTER 5
COSTS OF LOADSHEDDING (FROM SECONDARY DATA)
Chapter 3 has described the various approaches that have been developed for estimation of the
economic cost of loadshedding. Chapter 4 then has highlighted the results of earlier studies on
quantification of the costs loadshedding in Pakistan by application of some of these approaches.
This chapter extends the analyses by applying more approaches, especially those that can be
implemented with secondary data. In addition, an attempt is made to up-date the estimates of
the cost of loadshedding upto 2010-11.
5.1 THE SIMPLE VALUE ADDED APPROACH
The estimated value added per Kwh (at constant prices of 1999-2000) is presented for different
years in Table 5.1 As the electricity-intensity of the economy has been rising, this ratio has
shown a downward trend. For the economy as whole, the value added per Kwh has declined by
almost 2% annually over the last forty years, such that the magnitude in 2010-11 is less than
half of the value in 1971-72. It is interesting to note that, contrary to the overall trend, the value
added per Kwh in industry has risen, albeit slowly. The implication is that the industrial structure
of Pakistan is moving in the direction of less energy-intensity, in contrast to the trend observed
in most developing countries.
The table also demonstrates, however, that industry is more electricity-intensive than agriculture
and commerce. Accordingly, the simple value added approach will imply that the outage cost
per Kwh is the highest in commerce/services, followed by agriculture and the lowest in industry.
The value added per Kwh of 2000-01 is used for this analysis. During this year there was
significant excess capacity in the power sector and there was hardly any loadshedding. This is
projected up to 2010-11 on the basis of the growth rate between 1971-72 and 2000-01. The
approach produces very high estimates of outage cost per Kwh of $ 2.13 in the case of industry,
$4.44 in agriculture and as much as $ 20.65 in commerce/services. This compares with the
outage cost per Kwh of $ 0.50 in earlier studies for Pakistan. Given the estimates of cost per
Kwh and the quantity of electricity not supplied, based on the incidence of loadshedding at 20%
of working hours, the total cost of outages works out to a whopping Rs 1.6 trillion. This estimate
is beyond reasonable expectations and is rejected outright.
31
5.2 THE ADJUSTED VALUE ADDED APPROACH
This approach has been described in the Section 3.2 of Chapter 3. The magnitude of β is
estimated by econometric analysis on historical data for the period when there were no
significant outages (see Technical Annex II).
Table 5.1
Cost of Outages with the Adjusted Value Added Approach
β
Outage cost per Kwh
(Rs)
Quantity of
electricity Not
Supplied
Cost of
Outages (Rs
in Billion)
Sector
With 20 % time loss
Agriculture
0.050
Industry
0.126
Commerce/Services
0.062
18
(0.22) *
22.8
(0.27)
108.8
(1.28)
2243
5302
2759
10304
Total
42
(0.5) **
121
(1.4)
300
(3.5)
463
(5.4)
* in
**
US $
in billion US $
With the adjusted approach the total cost of loadshedding falls very sharply to Rs 463 billion or
$ 5.4 billion. However, these estimates are at the other extreme and look too low although they
are consistent with the international evidence that cost per Kwh is highest in commerce,
followed by industry and agriculture. For example, the outage cost per Kwh to industry is
derived as 27 cents, compared to 50 cents in other earlier studies. As such, this approach does
not appear also to be giving credible estimates.
A new variant of the value added approach has been developed for this study for quantification
of outage costs which focuses on the relationship between development of infrastructure and
economic growth. The former includes electricity generation. The relationship is estimated of
GDP per capita growth rate with through on OLS regression the following explanatory variables:
Agricultural Growth Rate
= AGR
Private Investment as % of GDP
= PRI
Growth Rate of Electricity Generation per Capita
= EGE
Growth Rate of Water Availability (in agriculture)
= WAV
32
The results of the econometric analysis are presented in the Technical Annex III. It appears that
a 1% increase in electricity generation per capita leads to a 0.167% increase in per capita GDP.
The next step in the analysis is to compare the actual growth rate of electricity generation per
capita from 2007-08 with the growth rate historically during the period when there was minimal
loadshedding. Given the coefficient of 0.167 the impact on the GDP per capita of a slowdown in
the process of electricity generation, which has led to outages, can be determined. The
divergence in growth rates is presented in Table 5.2.
Table 5.2
Projected GDP (at factor cost) in the Absence of Loadshedding compared to Actual GDP
(Rs in Billion)
2007-
Actual per
Capita
Growth Rate
of Electricity
Generation
(%)
-5.0
Historical
Per Capita
Growth Rate
of Electricity
Generation
(%)
4.0
Difference
(%)
Projected
GDP per
capita
growth
rate (%)
Actual
GDP at
current
prices
Projected
GDP at
current
prices
Cost of
outages
-9.0
Actual
per
Capita
GDP
growth
rate (%)
1.6
3.1
9922
10067
135
-6.1
4.0
-10.1
-0.4
1.3
12110
12490
380
1.9
4.0
-2.1
1.0
1.4
14034
14471
437
-3.0
4.0
-7.0
0.9
2.1
17093
17910
817
08
200809
200910
201011
The difference in the projected and actual GDP is the economic cost of loadshedding. It has
risen exponentially according to this approach from Rs 135 billion in 2007-08 to Rs 817 billion ($
9.6 billion) by 2010-11. It is likely that by 2011-12 the economic cost of loadshedding to the
national economy has exceeded Rs ONE TRILLION. This is approaching 5% of the GDP.
The implied outage cost per Kwh across the economy in 2010-11 is Rs 79.3 or 93 cents.
5.3 THE CONSUMER SURPLUS APPROACH
Application of the consumer surplus approach requires knowledge of the price elasticity of
demand. Table 5.4 presents the estimates of this elasticity for a number of developing countries.
The short-run elasticity is taken for the analysis. It appears that the elasticity is generally low,
especially in countries like Turkey, Israel, South Africa, etc. The estimate of -0.13 by Jamil and
Ahmed [2010] for Pakistan is used for the analysis.
33
The other assumption that has to be made relates to the demand that would have been
observed in the absence of outage at the existing power tariff. As highlighted in Chapter 4,
consumers have been exposed to loadshedding for 20 to 25 % of their working hours. As such,
it is assumed that unconstrained demand would have been at least 25% higher in comparison to
the observed level.
Table 5.3
Price Elasticity of Demand for Electricity (All Consumers Combined) in Different
Studies
Author
Country
Price Elasticity (short-run)
1
Al Faris [2002]
GCC Country
-0.11
2
Bose & Shula [1999]
India
-0.34
3
Narayan [2007]
Seven Countries
-0.06
4
Zchariadis and Pashaurtidou [2006]
Cyprus
-0.02
5
Lin [2003]
China
0.18
6
Erkan [2007]
Turkey
-0.04
7
Beensock [1999]
Israel
-0.01
8
Ziramba [2008]
South Africa
-0.02
9
Jamil and Ahmed [2010]
Pakistan
-0.13
The existing effective average tariff for all consumers combined is about Rs 9 per Kwh. Inclusive
of the GST, withholding income tax and FAC raises it to almost Rs 13 per Kwh. This is the price
used for the analysis.
The consumer surplus corresponds to the shaded area in Fig 5.1.
The methodology for estimating the consumer surplus is given in Technical Annex-IV.
The estimated magnitude of the consumer surplus is Rs 837 billion ($ 9.8 billion) in 2010-11.
This is very close to the cost of loadshedding derived from a variant of the value added
approach of Rs 817 billion ($ 9.6 billion). Therefore, secondary data based approaches to
quantification of economic cost of loadshedding yield an estimate of above Rs 800 billion ($ 9.4
billion) in 2010-11 and it is likely that this cost has exceeded Rs 1 Trillion ($ 11.1 billion) by
2011-12.
34
Figure 5.1
Consumer Surplus from Electricity Consumption
35
PART II
COST OF LOADSHEDDING TO
SMALL-SCALE INDUSTRY
36
CHAPTER 6
INTRODUCTION
This part of the report presents the findings on costs of loadshedding to small-scale industrial
units in Pakistan, quantified on the basis of data obtained from a nationwide survey of such
units.
The report is organized in nine chapters. Chapter 7 presents the methodology used for
qualification of costs due to outages. Chapter 8 describes the survey including the sampling
methodology and assessment of the quality of data collected, given the complex technical
nature of the survey. Subsequent Chapters up to Chapter 12 present the magnitudes of key
parameters like the relevant characteristics of the responding units, incidence of outages, level
and pattern of adjustments and the magnitude of different outages costs. Chapter 13 highlights
the suggestions by sample units for reduction in incidence and costs of outages.
Chapter 14 gives a summary of the principal findings and the major policy implications emerging
from the research. It is clear from the results that small-scale industry has faced severe
disruptions due to the high and growing incidence of loadshedding. As such, the economic
return of reducing outages and of facilitating the process of adjustment to these outages is very
high. This could contribute to some revival of the economy and reduce unemployment.
Thanks are due to the sample units for responding to a questionnaire, which was complex and
difficult to administer. Thanks are also due to the survey team which travelled all over the
country and sometimes found itself in a difficult law and order situation, especially in Karachi.
The main text of the report is presented in a non-technical manner.
Technical analyses are included in various Boxes and Technical Annexes.
Any defects which remain are of course, the responsibility of the authors.
37
CHAPTER 7
THEORETICAL FRAMEWORK
7.1 OUTAGES AND A FIRM’S BEHAVIOUR
The behavior of a firm in the presence of frequent and persistent outages has been modelled in
the Technical Annex V. The firm is assumed to be operating in a competitive environment, given
the smallness of its size; and pursues profit maximization. The results derived from this
theoretical framework are as follows:
i.
If outages are seen as, more or less, permanent in nature then the optimal size of the firm
is lower than in the absence of outages. In particular, there is a tendency to shed some
labor.
ii.
The likelihood that the firm will make adjustments to recover some of the lost output
depends on the following:
a) The extent to which the market situation is favorable for the firm
b) The electricity-intensity of the production process; the lower the intensity the
greater the likelihood that the firm will make an adjustment
c) The lower the costs of adjustment.
d) The larger the outage and the level of expectation that this will continue.
Figure 7.1
Adjustment by a Firm to Outages
38
Based on the results from the theoretical analysis we present in Figure 7.1 visually the change
in the equilibrium of a firm in the presence of outages
Type I firm initially experiences outages which reduce production from 𝑄0 − 𝑄1 . At 𝑄1 the gap
between price and marginal cost is AB. AB is larger the bigger the outage. The marginal cost
curve of adjustment by the firm is given by XY. If XY is too high, then the firm makes no
adjustment.
In the case of type II firm, Y lies between A and B. As such, the firm makes an adjustment and
the new output level is 𝑄𝑒 . Type I firm reduces output by (𝑄0 − 𝑄1 ) while the loss of output of
Type II firm is (𝑄0 − 𝑄𝑒 ). Also, the fall in profit of Type I firm due to outages is ABC. In the case
of Type II firm it is BYZC.
The above theoretical framework is used to develop the methodology for quantifying the cost of
outages.
7.2. METHODOLOGY FOR QUANTIFICATION OF COST OF OUTAGES
We have indicated above that in the presence of persistent outages firms will be inclined to
make adjustments. The extent and nature of adjustments will depend on a number of factors
which have also been identified above. The methodology used for quantifying the cost of
outages is based largely on that developed by Pasha, Ghaus and Mallik [1989]. Fig 7.2 presents
a flow chart for identifying the different types of outage costs.
There are two types of costs associated with outages. The first type is direct costs which
consist primarily of the value of lost production and spoilage costs. The second type is
adjustment costs. The particular mechanisms chosen for recovering some or all of the output
lost will be based on cost minimisation considerations. Accordingly, a firm will opt for a type of
adjustment as long as it is cheaper than other options. Therefore, firms may opt for multiple
adjustments, especially when the size of the outage is large.
Direct Costs
In order to derive the magnitude of direct costs, we designate the following:
𝑛𝑖 = number of times of occurrence of outage daily on average of duration i. i = 1, 2, 3, 4, 58
𝜖𝑖 = proportion of output lost during an outage of duration i
𝛾𝑖 = restart time after an outage of duration i.
8
The durations are 0-1/2 hr; ½ hr to 1 hr; 1 hr to 2 hrs; 2 hrs to 3 hrs; 3 hrs and above.
39
Figure 7.2
Flow Chart Showing Costs of Outages
OUTAGES
PLANT
SHUTDOWN
SPOILAGE COST
RESTART TIME
OUTPUT
RECOVERED
OUTPUT LOST
PERMANENTLY
TOTAL DIRECT
COSTS
ADJUSTMENTS
USE OF
GENERATORS
OVER TIME
EXTRA SHIFTS
NET IDLE
FACTOR COST
OTHERS
TOTAL INDIRECT (ADJUSTMENTS COSTS)
TOTAL OUTAGE
COST
40
The total number of outages during the year is given by
NOUT = ∑5𝑖=1 𝑛𝑖 × 365
..…………………………………………………...(2)
The total time lost due to outage is
TOUT = ∑5𝑖=1( 𝑛𝑖 [𝑑𝑖 + 𝛾𝑖 ] × 365)
……..………………………………………………(3)
Where 𝑑𝑖 is the duration of the outage.
The potential extent of output loss due to outages is given by
LOUT = ∑5𝑖=1 𝑛𝑖 [𝑑𝑖 + 𝛾𝑖 ] 𝜖𝑖 × 365
.……………………………………………………(5)
But the firm may not be operating throughout the year and for 24 hours each day. Therefore, if
H is the normal hours worked during the year, the actual output lost is given by
….…………………………………………………(5)
𝐻
ACOUT = 𝐿𝑂𝑈𝑇. 8760
And the value of this loss is as follows
VOUT = ACOUT.V
.……………………………………………………(6)
Where V is the value added by the firm per hour.
However, the firm may recover some of the output lost through adjustments. If 𝜆 is the extent of
output recovered then we have the expression for the net idle factor cost, NIFC, as follows:
NIFC = (1-λ) VOUT
……………………………………………………(7)
Box 7.1 gives a simple numerical example for calculation of the net idle factor cost incurred by a
sample unit due to outages on the basis of data obtained from the survey.
The other part of direct costs is spoilage costs. We represent
𝑆𝑖 = spoilage cost (in rupees) in each outage of duration i
Then the spoilage cost, SPC, is derived as follows:
𝐻
SPC = ∑𝑛𝑖=1 𝑛𝑖 𝑠𝑖 . 365 × 8760
……….…………………………………………….(8)
And the total direct costs of outages are
TDC = NIFC + SPC
…………………………………………………………(9)
41
Box 7.1
A Numerical Example of Quantification of Net Idle Factor Cost
Suppose a firm experiences outages of duration of one hour six times a day (over the 24 hours) and the
restart time after the outage is half an hour.
Then
NOUT = 6 × 365 = 2190
TOUT = 6 × 365× 1.5 =3285
If the proportion of output lost is 50%. Then
LOUT = 3285 × 0.5 = 1642.5
The firm normally operates one shift (of 8 hours) for 300 days. Then
H = 8 × 300 = 2400
And, ACOUT = 1642.5 ×
2400
8760
= 450
If the value added per hour is Rs 1000, then
VOUT = 450 × 1000 = 450,000Rs
and if the proportion of output recovered through adjustments is 60%, then λ = 0.6 and NIFC is given by
NIFC = 450,000 × (1-0.6) = 180,000 Rs
Adjustment Costs
Generators Cost
A key response by industrial units, which is being observed in Pakistan, is the resort to own
sources of energy supply through investment in generators. This is motivated by the high and
rising incidence of outages since 2007 and the growing recognition that power loadshedding
(along with gas shortages) could persist over the next many years.
In practice, the extent of substitution of the conventional power source (through DISCOs)
depends on the energy-intensity of operations (as derived in the theoretical framework above),
on the possibility of making other cheaper adjustments and the cost of capital for acquisition of
generators. The latter is likely to be relatively high for small-scale units.
The survey of units has indicated, first, whether a unit has a generator or not, second, the
capital cost of the generator, third, monthly running cost of fuel for operating the generator and
fourth, other costs (including labor, repairs and maintenance cost, etc.) on a quarterly basis. As
such, we designate the following:
𝐾𝑔 = capital cost of generator
foc = fuel operating cost per month
opc = other operating costs quarterly
This leads to the estimate of the overall annual generator cost, GENCO, as follows:
GENCO = (𝜏 + 𝛿)𝐾𝑔 + 12(foc) +4(opc)
……...……………………………………………(10)
42
Where, 𝜏 is the cost of capital and 𝛿 is the rate of depreciation. The combined value of 𝜏 𝑎𝑛𝑑 𝛿
is taken as 0.32.
However, the use of generators implies savings in costs of power supply from the local DISCO.
Therefore, the net cost, NGENCO, is given by
NGENCO = GENCO – k (TOUT) × (ADJG) ×tf .
𝐻
………..………………………………(11)
8760
Where
K = electricity consumption per hour in Kwh
TOUT = total hours lost as derived in equation (3)
ADJG = extent of adjustment by use of generators
tf = tariff per Kwh of the DISCO.
Other Adjustments
Other adjustments include the following and are more in the nature of short-run adjustments
when loadshedding is seen as temporary in nature:
i.
More intensive utilization of existing plant and machinery during times when there is no
loadshedding
ii.
Overtime or additional shifts to make up for at least part of the output loss
iii.
Changes in working days or timings.
The survey reveals that the majority of firms have not made significant adjustments of the above
type and the costs associated with these adjustments are not large. The methodology used for
quantifying these costs has been taken from Pasha, Ghaus and Mallik [1989]. They are
represented by OTC.
Overall, the total adjustment cost, TAJCO, is derived as
TAJCO = NGENCO + OTC
……………………………………………………….(12)
And the total outage out cost, TOUTCO, as follows:
TOUTCO = TDC + TAJCO
……………………………………………………….(13)
The magnitudes of the different components of outage costs are presented in Chapter 6, by
location, industrial classification and type of process of the sample units.
43
CHAPTER 8
THE SAMPLING FRAMEWORK AND ITS DISTRIBUTION
8.1 SAMPLING FRAMEWORK
The primary instrument of data collection was a survey on a pre-designed and tested
questionnaire of a stratified (by city, firm and industry group size) national random sample of
industrial units (see Figure 8.1).
Figure 8.1
Sampling Strategy
Primary Sampling Unit
Industrial Units in Cities
Secondary Sampling Unit
Small Scale Units in Cities
Tertiary Sampling Unit
Various Small Scale Units in Industry Groups
Individual industrial units in various Industry
Groups
The population of the industrial units was obtained from the Economic Census 2005, published
by the Pakistan Bureau of Statistics (PBS). This source was preferred to the Census of
Manufacturing Industries (CMI), which only covers large and medium-scale units (with
employment of 10 workers and above) while the focus of this study is on the small-scale
industrial sector. The national distribution of small manufacturing establishments by province
and industry group is presented in Table 8.1 and Figure 8.2. The derived sample distribution by
city and industrial group is presented in Table 8.2.
Once the sample distribution across cities and industrial groups was finalized, the individual
sample units were selected from the population of units obtained from the Provincial Directories
of Industries prepared by the respective Provincial Labour Departments. For each unit, the
Provincial Directories give name, address, and year of registration and level of employment.
The last variable enables the selection of small units.
The questionnaire administered on the sample respondents contains five modules: basic
information on production/sales/employment/costs; incidence of outages; costs of outages;
adjustment to outages, and, preferred load management practices. Though the questionnaire
44
was structured, the last question was open-ended asking the respondents to make suggestions
to reduce the costs of loadshedding. This provides the respondent’s perspective on actions to
counter the problem.
Table 8.1
National Distribution of Small Manufacturing Establishments* in the Economic
Census, 2005 by Province and Industry Group
Distribution of Manufacturing Establishment*
Percentage
Punjab
68.4
Sindh
13.9
K-PK
15.9
Balochistan
1.4
Islamabad
0.4
Total
100
Distribution of National Employment in Small-Scale Units (less than 10 employees)
Industry
%
Food, Beverages
20.9
Textiles, Wearing Apparel & Leather
43.4
Wood & Wood Products (including furniture)
11.1
Paper & Paper products
1.7
Chemicals
0.9
Non-Metallic Mineral Products
2.2
Basic Metal Products
0.7
Fabricated Metal Products, Machinery & Equipment
10.0
Others
9.1
Total
100
*with employment below 10 workers.
Figure 8.2
National Distribution by Province and Industry Group
By Province
By Industry Group
Balochistan
1%
KPK
16%
Others
15%
Fabricated
Metal
Products
10%
Sindh
14%
Punjab
69 %
Wood
and
Wood
Products
11 %
Textiles,
Wearing
Apparel
and
Leather
43%
Food,
Beverage
s and
Tobacco
21%
45
Table 8.2.
Distribution of Sample units by City, Province and Industry
Province
Punjab
Sindh
KPK
Balochistan
Cities
Food
Beverages
and
Tobacco
Lahore
23
Textiles,
Wearing
Apparel
and
Leather
46
Faisalabad
14
Sialkot
Wood &
Wood
Products
Fabricated
Metal
Products
Others
Total
14
12
19
114
29
9
7
12
72
6
12
4
3
5
30
Gujranwala
5
10
3
2
4
24
Multan
6
12
4
3
5
30
Rawalpindi/Islamabad
8
14
4
5
8
39
Total
62
123
38
33
54
310
Karachi
13
26
8
7
12
66
Hyderabad
4
8
2
2
4
20
Sukkur
3
6
2
1
3
14
Total
20
40
12
10
18
100
Peshawar
13
29
10
7
11
70
Total
13
29
10
7
11
70
Quetta
5
8
-
-
7
20
Total
Total
5
8
0
0
7
20
100
200
60
50
90
500
The survey was successfully administered on
500 units as targeted. Following the process of
edit and consistency checking of the completed
questionnaires, 412 units, over 82 percent of
the sample, have been included in the analysis.
8.2 PROFILE OF RESPONDANTS
Distribution of selected units for analysis by city
is given in Table 8.3. 63 percent of the sample
units are in the province of Punjab, given the
concentration of small-scale industry in Punjab,
while about 20 percent are in Sindh. From the
Table 8.3.
Distribution of Sample Analysed by Cities
Cities
Lahore
Faisalabad
Gujranwala
Multan
Sialkot
Rawalpindi / Islamabad
Karachi
Hyderabad
Sukkur
Peshawar
Quetta
Total
Numbers
104
61
16
29
23
27
54
16
11
58
13
412
%
25.2
14.8
3.9
7.0
5.6
6.6
13.1
3.9
2.7
14.1
3.1
100.0
remaining 17 percent, 14 percent are in Khyber-Pakhtunkhwa (K-PK) and 3 percent in
Balochistan. The distribution by industry group and nature of production process is given in
Figure 8.3. The presence of continuous-process technology in small units is low.
46
Figure 8.3
Distribution of Selected Units by Industrial Group and Process
Industry Group
Nature of Production Process
Food
beverages
and
Tobacco
19%
Other
industries
18%
9%
Fabricated
Metal
Products
10%
Textile
wearing
Apparel
and
leather
42%
Wood and
Wood
products
11%
91%
Continuous Process
Batch-Process
8.3 CHARACTERISTICS OF SELECTED UNITS
Basic Information
Respondent units from among
the
batch-making
industries
generally operated a one shift
production cycle. The Second
Table 8.4.
Average Employment and Capacity Utilization by
Industrial Group, 2012
Employment
(No)
Capacity
Utilization
(%)
Food Beverages and Tobacco
7.1
82.3
Textile Wearing Apparel and Leather
8.1
80.7
Wood and Wood products
6.7
76.7
Fabricated Metal Products
6.4
76.6
Other industries
7.0
81.6
Total
7.4
80.3
Industries
shift was generally operated by
industrial
textile,
units
wearing
belonging
apparel
to
and
leather and food, beverages and
tobacco
industries
group.
Continuous process industries,
on an average, worked 293 days
a year. 15 percent of sample units worked 315 days while another 20% of sample units worked
330-335 days a year. The sample units on an average employed just over 7 persons and
utilized 80 percent of their capacity (see Table 8.4). Econometric analysis of the determinants of
the rate capacity utilization has been undertaken in Box 8.1. The results indicate that the
incidence of outages has a significant negative impact on the rate of capacity utilization.
47
Box 8.1
Capacity Utilization and Outages
The following variables are designated for each unit:
CAPU = Rate of capacity utilization (%)
DIND1 = 1 for food, beverages and tobacco
= 0 otherwise
DIND2 = 1 for textiles, apparel and leather
= 0 otherwise
DIND3 = 1 for wood, and wood products
= 0 otherwise
DIND4 = 1 for fabricated, metal products
= 0 otherwise
PTOUT= proportion of time lost due to outages (%).
The results a of the OLS regression with CAPU as the dependent variable are as follows:
CAPU =
84.34
-5.19DIND3
-4.38DIND4
-0.06PTOUT
(58.3)*
(-3.18)*
(-2.55)*
(-1.93)**
R2 =0.06, SEE =8.43, Degrees of Freedom = 394, F = 4.6*
*Significant at 5% level
** Significant at 10% level
The results indicate a lower rate of capacity utilization in the wood and wood products and the fabricated
metal products industry. Also, the incidence of outages has a significant negative impact on the
rate of capacity utilization.
a only the significant variables are highlighted.
Sales and Value Added
The average sales of
the
respondent
Table 8.5
Average Sale and Operating Expenses of
Sample Units 2012
(Rs in Thousands)
units
chosen for analysis
in
2012 is projected at Rs
5.5 million (see Table
8.5)
demonstrating
growth of
almost
Their
Total
Food beverages and Tobacco
6888
4246
Operating
Expenses
as % of
Sales
62
Textile wearing Apparel and leather
5877
3352
66
Wood and Wood products
6497
4336
67
Fabricated Metal Products
5151
3488
68
Others industries
4431
2904
65
Total
5472
3565
65
a
15
percent over the 2011
level.
Industrial Group
Total
Operating
Expenses
operating
expenses average Rs
3.6 million, implying an
operating profit of Rs
1.9 million. It appears that only 14 percent of the sample units are export-oriented, mostly
belonging to textiles, wearing apparel and leather industry.
48
The
average
value
Table 8.6
Average Value Added, Electricity Purchased and Value Added
per Kwh of Electricity of Sample Units
added by sample units
in 2012 is projected at
Value Added
(Thousands)
Electricity
Consumed
(Kwh)
3139
22730
Value
Added
Per
Kwh
(Rs.)
138.1
1930
19801
97.5
2426
12598
192.6
Fabricated Metal Products
2053
19190
107.0
Other industries
2094
19169
109.2
Total
2259
19384
116.5
Rs 2.2 billion, highest
Industrial Group
being in food, beverage
and tobacco industry
Food beverages and Tobacco
followed by wood and
wood
products
Table
units
8.6).
over
sample
have,
average,
Textile wearing Apparel
leather
Wood and Wood products
(see
on
an
purchased
19
thousand
and
kilowatt hours (Kwh) of electricity annually from the public distribution companies. The average
value added per Kwh is Rs 117. This is the first estimate of the outage cost per Kwh. As
highlighted earlier, it overstates the magnitude because it does not incorporate the impact of
adjustments.
Operating Costs
Turning next to operating costs, as mentioned above, average annual operating cost of sample
units is Rs 3.6 million. Out of this, the highest proportion, (56 percent), is spent on purchase of
raw materials, followed by wages (25 percent) (See Table 8.7). Electricity costs purchased from
the distribution companies and self-generation combined account for 15 percent of the costs.
Table 8.7
Operating Cost Structure of Sample Units
(%)
Industrial Group
Food beverages and
Tobacco
Textile
wearing
Apparel and leather
Wood and Wood
products
Fabricated
Metal
Products
Others industries
Total
Total
Operating
Cost
(Rs. In
Thousands
)
4246
(Percentage of Operating Cost)
Wages/
Salaries
Raw
Material
Repairs/
Maintenance
Cost of
Electricity
Cost of Self
Generation
Others
24.9
55.6
3.0
10.8
4.5
1.2
3352
24.8
55.2
3.0
10.1
4.4
2.5
4336
21.8
62.2
3.4
7.4
3.7
1.4
3488
20.6
60.5
3.6
9.4
3.5
2.5
2904
27.4
51.1
4.1
11.3
4.0
2.1
3565
24.6
55.8
3.3
10.1
4.2
2.0
49
Attainment of Production targets
Sample units, on an average, have been able to achieve 78 percent of their production target,
as shown in figure 8.4. The highest rate of target achievement is by food, beverages and
tobacco industry at 85%, followed by wood and wood product industry at 78 percent.
Figure 8.4
Actual as % of Target Production in 2012 by Industrial Group
Industrial Group
(%)
85.2
78.8
77.4
78.5
76.9
73.0
Food beverages
and Tobacco
Textile wearing
Apparel and
leather
Wood and Wood Fabricated Metal Others industries
products
Products
Total
By City
84.1
79.5
81.1
78.6
84.7
82.1
69.9
76.4
78.8
71.3
71.5
78.5
Interestingly, sample units in Punjab were able to achieve a higher proportion of the target
compared to the other cities, possibly because the targets were already modified somewhat to
allow for the presence of loadshedding.
50
When enquired as to why the production target was not attained, the principal reasons given are
high level of power outages, law and order situation (especially in Karachi) and market related
factors. (see Table 8.8).
Table 8.8
Reasons Why Production Target was Not Attained
(%)
Power outages
24
Textiles
wearing
Apparel
and
Leather
19
Law and order
25
19
16
19
20
High inflation & market
competition
Shortage or high cost or raw
materials
Other reasons
18
20
20
25
19
14
20
20
17
28
18
22
25
19
13
Total
100
100
100
100
100
Reason for Not Attaining
Targets
Food
Beverages
and
Tobacco
Wood
and
Wood
Products
Fabricated
Metal
Products
Other
Industries
18
21
20
51
CHAPTER 9
THE INCIDENCE OF LOADSHEDDING AND OUTPUT LOSSES
This Chapter quantifies the incidence of loadshedding and the resulting output losses as
revealed by the national sample of small-scale industries units.
9.1 INCIDENCE AND PROFILE OF LOADSHEDDING
The costs of loadshedding will, to a large extent,
Table 9.1
Frequency of Loadshedding in 2012
depend on the frequency and duration of
outages. The incidence of loadshedding is given
in Table 9.1. Overall, the average number of
outages in Pakistan in 2012 is estimated at
14119. Highest number of times outages have
occurred in Lahore at 1874, followed by Sukkur
at 1795, Gujranwala at 1683, and Multan at
1561. Clearly, the average incidence is lower in
Sindh at of 1267 times a year at about 10%
below the national average.
Industry-wise,
the
highest
incidence
was
experienced by the textile, wearing apparel and
leather
industry
(1495)
followed
by
other
industries (1398) and food, beverages and
tobacco industries (1365). Box 9.1 gives the
results of econometric analysis of the incidence
of outages. The results confirm that this is
By Province/City1
Location
Punjab
Lahore
Faisalabad
Other Cities
Sindh
Karachi
Other Cities
KPK
Balochistan
Total
By Industrial Group
Industrial Group
Food beverages and Tobacco
Textile wearing Apparel and leather
Wood and Wood products
Fabricated Metal Products
Others industries
Total
By Process
Nature of Production
Continuous Process
Batch-Process
Total
Average
1499
1874
1047
1379
1267
1095
1596
1239
1348
1411
1365
1495
1289
1296
1398
1411
1956
1355
1411
1
significantly higher in the cities of Punjab and
Balochistan. Also, the incidence of outages is
Other cities are the smaller cities. They have been
combined in view of the smallness of samples in each
location.
higher in continuous- process units, although outage costs are likely to be higher for such units.
The distribution of outages by duration is given in Table 9.2. Generally, the outage duration is 1
to 3 hours a day. The highest number of outages occurs for 1 to 2 hours a day at 51 percent,
followed by outages of 2-3 hours a day at 32 percent. 11 percent of outages have duration of
half to one hour while 5 percent of outages last for over four hours. There is a divergence in the
provincial patterns. In Punjab, outages mostly last for 1-2 and 2-3 hours. In Karachi almost half
of the outages last for over two hours. This pattern of outages is likely to have significant
implications for the costs of outages as will be discussed in Chapter 11.
9
from January to September 2012. The annual incidence was estimated by multiplying by 1.33.
52
Box 9.1
Incidence of Outages
The following variables are designated for each unit:
DLOCi
=1 for the i-th city
= 0 otherwise
DPROD
= 1 if continuous-process unit
= 0 otherwise.
TOUT
= Total incidence of outages (in hrs)
The results a of the OLS regression with TOUT as the dependent variable are as follows:
TOUT =1983
+290DLAH
+676.6MUL
+ 374.0DPES +431.3RWP
+771.3DSIA
(16.3)* (2.09)* (3.6)*
(2.40)*
(2.22)*
(3.79)*
1490.1DQUE
+1070.1DGUJ +316.1DPROD
(5.85)*
(4.6)*
(2.16)*
R2 =0.17, SEE =811, Degree of Freedom = 407, F = 6.89*
*Significant at 5% level
The results indicate that the incidence of outages is significantly higher in the cities of Punjab, Peshawer
and Quetta. Also, the incidence of outages is higher in continuous-process units, although outage costs
are likely to be higher for such units.
a
only the significant variables are highlighted.
Table 9.2
Percentage Distribution Average Length of Outages, 2012
(%)
By City
Location
Punjab
Lahore
Faisalabad
Other Cities
Sindh
Karachi
Other Cities
KPK
Balochistan
Total
By Industrial Group
Food beverages and Tobacco
Textile wearing Apparel and leather
Wood and Wood products
Fabricated Metal Products
Others industries
Total
By Process
Continuous Process
Batch-Process
Total
Less than ½ hr
1/2 – 1hr
1-2 hrs
2-3 hrs
More than 4 hrs
0
0
0
1
3
4
1
0
0
1
15
33
3
3
8
8
10
2
0
11
53
60
44
31
54
39
92
51
0
51
29
5
51
30
35
49
15
28
92
32
2
2
2
2
0
0
0
20
8
5
0
1
2
3
0
1
12
9
4
18
16
11
45
55
63
52
41
51
37
31
26
23
38
32
6
4
4
5
5
5
3
1
1
19
10
11
54
51
51
24
33
32
1
5
5
53
The overall duration of outages, which includes
both the time lost due to an outage and the restart
time (time lost in restarting work following an
outage), is presented in Table 9.3. The total hours,
on an average, lost per annum due to loadshedding
are estimated at 2623 for the sample units. The
highest number of hours lost is in Quetta. Overall in
Punjab, the average number of hours lost per
annum is 2666, in Sindh 2200, K-PK 2777 and
Balochistan, 3733. These durations are for 24hrs a
day for 365 days a year. Clearly, the actual total
time lost depends on working hours during the
year.
Other industries, textiles and wearing apparel and
leather industry take the brunt, losing respectively
2789 and 2727 hours. Again, the incidence is
higher in continuous process (2995 hours) than in
batch-making (2583) industries.
Table 9.3
Duration of Outages
(Outage + Restart Time) [Hours]
By City
Cities
Punjab
Lahore
Faisalabad
Other Cities
Sindh
Karachi
Other Cities
KPK
Balochistan
Total
By Industrial Group
Industrial Group
Food beverages and Tobacco
Textile wearing Apparel and
leather
Wood and Wood products
Fabricated Metal Products
Others industries
Total
By Process
Nature of Production
Continuous Process
Batch-Process
Total
Average
2666
2740
2372
2769
2200
2128
2325
2777
3733
2623
2488
2727
2354
2438
2789
2623
2996
2585
2623
Box 9.2 undertakes econometric analysis of the extent of output losses during outages. The
results confirm that production losses are lower in some cities of Punjab and Peshawer and
Quetta. Also, we have the important finding that outages of longer duration imply higher losses.
Further, the higher the percentage of time the distribution companies keep to the preannounced loadshedding schedule, units are better prepared at these times and suffer lower
losses.
9.2 EXTENT OF TOTAL TIME LOST
The proportion of production time lost is given in Table 9.4. Overall, small scale units in
Pakistan, on an average, are likely to lose 30% of their production time in 2012 due to
loadshedding. The highest, 43%, is lost in Quetta, and the least, 24 percent, in Karachi. Other
industries, and textiles, wearing and apparel industries lose over 30 percent of their production
time, while the loss for continuous process industries is over one-thirds.
54
Table 9.4
Proportion of Time Lost
during outages
(%)
By City
Cities
Average
Punjab
Lahore
Faisalabad
Other Cities
Sindh
Karachi
Other Cities
KPK
Balochistan
Total
30
31
27
32
25
24
27
32
43
30
By Industrial Group
Food beverages and Tobacco
28
Textile wearing Apparel and leather
31
Wood and Wood products
27
Fabricated Metal Products
28
Others industries
32
Total
30
By Process
Continuous Process
34
Batch-Process
29
Total
30
Box 9.2
Production Losses During Outages
The following variables are designated for each unit:
LOSSOUT
= proportion of output lost during outages (prior to recovery)
LOUTN
= proportion of outages of duration of more than 2 hrs
ANALD = percentage of time DISCO/ KESC kept to announced loadshedding
The results a of the OLS regression with LOSSOUT as the dependent variable are as follows:
LOSSOUT =
0.64 - 0.16DLAH
-0.36MUL
- 0.21DPES
+0.10RWP
-0.12DSIA
(24.94)* (-5.91)*
(-6.84)*
(-7.4)*
(2.75)*
(-3.30)*
-0.21DSUK
-0.21DFB
-0.18DQUE
+0.3x10-3LOUTN
-0.06DIND1
(-4.30)*
(-7.03)*
(-3.78)*
(2.05)*
(-2.53)*
-0.04DIND2
-0.06DIND4
-0.75x10-3ANALD
(-2.01)*
(-1.98)*
(-2.23)*
R2 =0.43, SEE =0.13, Degree of Freedom = 348, F = 14.4*
*Significant at 5% level
The results indicate that production losses are lower in some cities of Punjab, Peshawer and Quetta.
Also, we have the important finding that outages of longer duration imply higher losses. Further, the
higher the percentage of time the distribution companies keep to the announced loadshedding schedule,
units are better prepared at these times and suffer lower losses.
a only the significant variables are highlighted.
55
9.3 SEASONALITY OF OUTAGES
A significant seasonality in the incidence of loadshedding emerges from the data (see Table
9.5). The peak loadshedding months are July (accounting for almost 16% of the hours of
loadshedding) closely followed by June. August and May also emerge as high incidence
months, each accounting for over 14% of loadshedding hours. The pattern appears to be similar
for all four provinces and industrial groups.
Table 9.5
Seasonality in Outages
(% of Outage)
By City
Cities
March
April
May
June
July
August
September
Punjab
13
13
14
16
16
15
13
Lahore
13
14
15
15
15
14
12
Faisalabad
14
14
14
14
14
14
14
Other Cities
12
13
14
16
17
15
13
Sindh
13
14
14
15
16
14
13
Karachi
14
14
14
16
15
13
14
Other Cities
13
14
14
15
16
14
13
KPK
10
11
14
17
18
16
12
Balochistan
14
14
15
15
15
13
14
Total
13
13
15
16
16
15
13
March
April
May
June
July
August
September
Food beverages and Tobacco
13
14
15
16
16
14
13
Textile wearing Apparel and leather
13
14
15
16
16
15
13
Wood and Wood products
12
13
14
16
16
15
13
Fabricated Metal Products
13
13
14
16
16
15
13
Others industries
13
13
14
15
16
15
14
Total
13
13
15
16
16
15
13
March
April
May
June
July
August
September
Continuous Process
13
14
15
15
16
14
13
Batch-Process
13
13
14
16
16
15
13
Total
13
13
15
16
16
15
13
By Industrial Group
Industrial Group
By Process
Nature of Production
56
9.4 EXTENT OF OUTPUT LOST DURING OUTAGES
Loadshedding leads to a complete shutdown for 31 percent of sample units, with highest
proportion being in Sindh. However, for 61 percent of the firm it results in partial shutdown. (see
Table 9.6).
Table 9.6
Nature of Impact of Loadshedding
(%)
Complete
Shutdown
Partial
Shutdown
No Impact
By City
Punjab
28
65
8
Lahore
20
80
0
Faisalabad
10
82
8
other Cities
53
48
45
Sindh
51
41
9
Karachi
59
31
9
other Cities
56
65
18
KPK
19
74
7
Baluchistan
38
62
0
Total
31
61
8
By Industrial Group
Food beverages and Tobacco
32
53
15
Textile wearing Apparel and leather
30
65
5
Wood and Wood products
33
52
15
Fabricated Metal Products
30
68
3
Others industries
35
63
3
Total
31
61
8
Continuous Process
39
61
0
Batch-Process
30
61
8
Total
31
61
8
By Process
In the questionnaire the respondants were asked to rank ‘what is most disruptive about an
outage’. For 40% of the respondants loss of/inconvenience to customers was the most
disruptive consequence of loadshedding, for 24% it was equipment shut down and for, 20%, idle
labor (see Table 9.7). Losses of product, breakdown of production process and spoilage costs
were cited as other disruptions.
57
Table 9.7
Ranking of Disruptions Due to Outages
(%)
Equipment
Shut down
Labor will
be idle
Product
will be
lost
Loss of /
Inconvenien
ce to
Customers
Break
down of
producti
on
process
High
spoilage
cost of
raw
material
42
40
34
48
21
24
15
60
38
40
7
14
2
1
0
0
0
10
8
6
0
1
0
0
1
0
4
0
0
0
63
25
15
23
20
13
7
14
3
Total
By City
Punjab
Lahore
Faisalabad
Other Cities
Sindh
Karachi
Other Cities
KPK
Balochistan
Total
19
11
25
24
53
61
37
9
15
24
24
17
38
22
15
9
26
12
23
20
8
16
2
4
10
6
19
9
15
9
By Industrial Group
Food beverages
and Tobacco
23
15
11
42
8
1
19
Textile wearing
Apparel and leather
24
18
9
40
8
1
42
28
30
7
35
0
0
11
15
23
13
50
0
0
10
27
24
5
39
5
0
18
Total
24
20
9
40
6
0
Continuous Process
37
8
5
32
18
0
9
Batch-Process
23
22
9
41
5
1
91
Total
24
20
9
40
6
0
Wood and Wood
products
Fabricated Metal
Products
Others industries
By Process
58
CHAPTER 10
ADJUSTMENTS TO LOADSHEDDING
This chapter focuses on the types of adjustments that small firms make to outages in Pakistan.
The extent of output that is not recovered following the adjustments is also quantified.
10.1 NUMBER AND TYPES OF ADJUSTMENTS
Table
10.1
presents
the
estimates
of
frequency
of
different types of adjustments by
small
units.
It
appears
Table 10.1
Percentage of Sample Units by Number of Adjustments
by Process
that
almost one fourths of the firms in
Continuous
Process
BatchMaking
Total
the sample are unable to make
No Adjustment
5
24
22
any form of adjustment. 46%
One Adjustment
50
45
46
make
Two Adjustments
29
20
21
Three or More Adjustments
16
11
11
Total
100
100
100
one
adjustment,
21%
make two types of adjustment
while 11% are implementing
three or more types of adjustments. Continuous process units are more likely to make
adjustments in the face of potentially higher losses.
The frequency of different types of adjustments is given in Table 10.2. The highest frequency of
adjustment is in the form of self-generation. For the national sample, this is 68%. It is the
highest in Punjab, follow by K-PK, and the lowest in Balochistan. On average generators are
able to substitute for 48% of the public source. Econometric analysis has been undertaken in
Box 10.1 of the extent of recovery of output lost by use of generators. The results indicate, first,
that higher incidence of outages leads to greater resort to generators, second, units with lower
electricity-intensity are more likely to use generators and, third, continuous- process units tend
to have a higher presence of generators.
Beyond the use of generators, the next most frequent form of adjustment is the working of
overtime by 29% of the units, as shown in Table 10.2. Approximately 11% of the firms have
opted for changing shift timings. The labor-related adjustments are more commonly observed in
Punjab.
59
Table 10.2.
Percentage of Sample units Adjusting through Various Mechanisms
(%)
Buying or
Operating
Existing
Generator
Working
Overtime
Increasing
Intensity
of
Machinery
Operation
Changing
Shift
Timings
Changing
Working
Days
Working
Additional
Shifts
Punjab
18.8
23.8
8.5
41.9
6.5
0.4
Lahore
10.6
17.3
16.3
40.4
14.4
1.0
Faisalabad
24.6
37.7
1.6
34.4
1.6
0.0
Other Cities
24.2
22.1
4.2
48.4
1.1
0.0
Sindh
53.1
14.8
9.9
21.0
0.0
1.2
Karachi
61.1
9.3
5.6
24.1
0.0
0.0
Other Cities
37.0
25.9
18.5
14.8
0.0
3.7
KPK
Baluchistan
Total
8.6
15.4
24.0
60.3
38.5
40.3
10.3
7.7
5.8
0.0
0.0
0.5
Food beverages and
Tobacco
Textile wearing Apparel
and leather
Wood
and
Wood
products
Fabricated
Metal
Products
Others industries
21.2
16.9
32.1
47.1
32.1
31.0
45.1
45.1
31.9
27.9
36.3
23.2
11.0
15.2
13.3
12.8
8.9
30.9
8.1
8.8
3.8
6.2
5.7
0.1
14.6
14.0
19.0
6.2
16.9
15.0
Total
68.2
28.9
7.8
11.4
3.2
11.4
Continuous Process
11.4
12.6
15.6
21.3
14.9
15.4
Batch-Process
88.6
87.4
84.4
78.7
85.1
84.6
Total
68.2
28.9
7.8
11.4
11.4
3.2
By City
12.1
8.6
23.1
15.4
20.4
9.0
By Industrial Group
By Process
60
Box 10.1
Extent of Recovery of Output by Use of Generators
The following variables are designated for each unit:
BEXGP
= Extent of adjustment by use of Generator
DOUT
= Total incidence of outages and restart times (in hours) annually
TPE
= Ratio of Value Added to Electricity purchased (Rs per Kwh)
ANALD = Per cent of time DISCO/KESC kept to announced loadshedding
DPROD
= 1 if continuous-process unit
= 0 otherwise.
The results a of the OLS regression with BEXGP as the dependent variable are as follows:
BEXGP =
10.57 + 0.15x10-2 DOUT +
73.64 TPE +
0.08 ANALD
+8.49 DPROD
(3.76)*
(1.89)**
(2.21)*
(2.17)*
(2.72)*
R2 =0.05, SEE = 17.97, Degree of Freedom = 396, F = 3.75*
*Significant at 5% level
** Significant at 10% level
The results indicate, first, higher incidence of outages leads to greater resort to generators, second, units
with lower electricity-intensity are more likely to use generators and, third, continuous-process units tend
to have a higher presence of generators.
a only
the significant variables are highlighted.
The question is whether self-generation is the sole adjustment mechanism or if it accompanied
by other adjustments. This analysis is presented in Table 10.3. It appears that almost 70% of
the units without generators make no adjustments at all while 26% of the units with generators
make another adjustment.
Table 10.3
Number of Adjustments by Firms with and without Generators
(%)
Numbers of Adjustments
None
Units without Generators
70
Units with Generators*
-
One
17
59
Two
10
26
Three or more
3
15
Total
100
100
*including the one adjustments of use of generator.
10.2 EXTENT OF LOSS OF OUTPUT IN OUTAGES
Table 10.4 highlights the extent of the permanent loss of output which is not recovered through
the various adjustments. Overall, it is 9% nationally, 8% in Punjab, 9% in Sindh, and K-PK and
12% in Balochistan. These losses are a key indicator of the magnitude of net idle factor costs.
61
Table 10.4
Proportion of output Loss Not Recovered
(%)
By City
Location
Average
Punjab
Lahore
Faisalabad
Other cities
8
9
6
8
9
Sindh
Karachi
Other cities
9
9
KPK
Balochistan
9
12
Total
9
By Industrial Group
Industrial Group
Food beverages and Tobacco
Average
7
Textile wearing Apparel and leather
8
Wood and Wood products
8
Fabricated Metal Products
9
Others industries
Total
11
9
By Process
Nature of Production
Continuous Process
Batch-Process
Total
Average
8
9
9
62
CHAPTER 11
COSTS OF OUTAGES
The objective of this chapter is to present the estimated magnitudes of different types of costs
associated with outages. As identified in chapter 7, these include direct costs which consist of
net idle factor costs and spoilage costs and indirect or adjustments costs which include
generator costs and costs of other types of adjustments like overtime, additional shifts, etc.
Section 1 of the chapter presents the total outage costs by location (province), industrial
classification and process used by sample units. Section 2 derives the cost per Kwh of load
shedding. Finally, by blowing up the sample, the magnitude of outage costs to the small-scale
industrial sector of Pakistan is derived.
11.1
TOTAL OUTAGE COSTS
Given the high frequency of load shedding in Pakistan today it appears that the total outage cost
even to small-scale industrial units of the country is high in absolute terms. Table 11.1 shows
that the outage cost approaches to Rs. 300,000 per unit on average in the sample units or close
to $3092 per annum10.
The variation in outage costs annually per unit is substantial depending upon the nature of the
process, that is, batch-making or continuous process. The latter type of units are especially
vulnerable to disruptions from outages. As such, the average cost for a continuous process unit
is Rs. 426,000 or $4392 per annum. This is 52% higher than the cost incurred due to load
shedding by batch-making units.
Among industries, the highest outage cost for small-scale units is in the food, beverages and
tobacco industry at Rs 369,000 or $3804 per annum, almost 26% above the average. Outage
costs are also somewhat higher in the case of textiles and apparel industry and lower in the
wood and wood products and fabricated metal products industries.
Turning to location, the outage cost per unit is substantially higher in Sindh province, especially
in the city of Karachi, at over Rs 500,000 or $ 5155 annually. Units in Punjab have lower outage
costs on average of Rs 235,000 or $ 2422 annually. The question is why in presence of
apparently higher incidence of outages in Punjab (See Table 11.1) the total outage cost is
lower. This is due the higher magnitude of net idle factor costs due to a lower proportion of
10
At the exchange rate of Rs 97= 1 US $
63
recovery of output in the presence of a troubled law and order situation11. Also, the generator
costs are higher because of greater electricity-intensity of operations by units in Sindh12.
Further, there is evidence that outages that occur in Karachi are of a longer duration each time
and unanticipated outages occur more frequently.
Table 11.1
Total Outage Costs Per Unit
(000 Rs)
Net Idle
Factor Cots
Spoilage
Costs
Generator
Costs
Other
Costs
Total
Outage
Costs
145
305
171
40
10
41
45
178
44
5
11
9
235
504
265
252
41
69
7
369
139
39
103
8
289
190
160
206
28
19
29
28
38
41
6
10
5
252
227
281
165
183
181
(62)
76
30
34
(12)
179
60
71
(24)
6
7
7
(2)
426
280
293
(100)
LOCATION
Punjab
Sindh
K-PK + Baluchistan
INDUSTRY
Food, beverages and tobacco
Textiles and wearing Apparel
and Leather
Wood and Wood Products
Fabricated Metal Products
Other Industries
PROCESS
Continuous Process
Batch-Making
TOTAL
(%)
11.2 TYPES OF OUTAGE COSTS
Table 11.1 also clearly establishes the dominance of net idle factor costs at 62% in total outage
costs. This reflects the constraints which small units face in making adjustments, especially in
the presence of difficult law and order situation in some parts of the country, like Karachi, which
reduces the ability to make labor- related adjustments. Consequently, the loss of output and the
implied net idle factor costs are relatively high.
11
As shown in Table 54, while the proportion of lost output, after allowing for adjustments, is 5% in Punjab, it is
almost double, at over 9% in Sindh.
64
The second major cost item is generator costs, with a share in total outage costs of 24%. This is
a reflection of the fact that almost 68 % of the sample units have installed generators and this
has become the primary form of adjustment to outages.
Spoilage costs constitute 12% of the outage costs. As expected, these costs are relatively high
in continuous –process units. Other adjustment costs are minor and account for only 2% of
outage costs.
Figure 11.1
Distribution of Outage Costs
2%
12%
24%
62%
others
idle factor costs
generator costs
spoilage costs
11.3 BURDEN OF OUTAGE COSTS
The burden of outage costs as a percentage of the value added is given in Table 11.2. It
appears that the burden is quite high at almost 13% of the value added. Much of this is a
decline in profits. It is the highest at over 17% in the case of continuous process units, in the
textiles and wearing apparel industry and in Sindh (especially Karachi).
Table 11.2
Outage Costs as Percentage of Value Added
(000 Rs)
LOCATION
Punjab
Sindh
K-PK + Baluchistan
INDUSTRY
Food, beverages and tobacco
Textiles and wearing Apparel and Leather
Wood and Wood Products
Fabricated Metal Products
Other Industries
Sample
Size (No)
Total Outage
Cost
Value
Added
Outage Cost
as % of
Value
Added
260
81
71
235
504
265
1978
3392
2245
11.9
14.9
11.8
79
172
46
40
75
369
289
252
227
281
3153
1971
2529
2027
2198
11.7
14.7
10.0
11.2
12.8
65
PROCESS
Continuous Process
Batch-Making
TOTAL
38
374
426
280
293
2460
2291
2307
17.3
12.2
12.7
Figure 11.2
Outage Costs as Percentage of Value Added
By Province
20
15
10
5
14.9
11.9
11.8
0
Punjab
Punjab
Sindh
Sindh
k-Pk + Balochistan
k-Pk + Balochistan
By Process
20
15
10
17.3
12.2
5
0
Continuous Process
Batch Making
Continuous Process
Batch Making
By Industry
20
15
10
5
11.7
14.7
10
11.2
12.8
Wood & Wood
Products
Fabricated Metal
Products
Others
0
Food, Beverages & Textiles, Apparel &
Tobacco
Leather
Food, Beverages & Tobacco
Textiles, Apparel & Leather
Fabricated Metal Products
Others
Wood & Wood Products
11.4 OUTAGE COSTS PER KWH
66
Table 11.3 indicates that the outage cost per Kwh in the case of small-scale industrial units is
about Rs 50 per Kwh. At about 52 cents per Kwh, this is similar to earlier estimates for Pakistan
as highlighted in Part I of the report. In line with the above results, the outage cost per Kwh is
the highest in Sindh. Although the overall outage cost is the highest for continuous process
units, it is lower per Kwh because of a higher quantity of electricity not provided.
Table 11.3
Outage Costs Per Kwh
Total Outage Cost
(000 Rs)
Electricity not
Provided during
Outages
(000 Kwh)
Outage Cost per Kwh
(Rs)
235
504
265
5.5
7.3
5.0
42.7
69.0
53.0
369
289
6.0
6.2
61.5
46.6
252
227
281
3.7
6.6
5.3
68.1
34.4
53.0
426
280
293
9.8
5.4
5.8
43.5
51.9
50.5
Location
Punjab
Sindh
K-PK + Baluchistan
Industry
Food, beverages and tobacco
Textiles and wearing Apparel
and Leather
Wood and Wood Products
Fabricated Metal Products
Other Industries
Process
Continuous Process
Batch-Making
Total
Overall, the survey of small-scale industrial units reveals that the absolute magnitude of outage
costs is high. There have been significant output losses, as reflected by the large net idle factor
costs. The use of generators has enabled significant recovery of output but at a cost more than
twice the tariff paid for supply from a DISCO. Losses in profitability have been proportionally
greater approaching 36% or more of the profit that could have been made in the absence of
outages.
11.5 NATIONAL ESTIMATE OF OUTAGE COSTS
For the 412 sample of small-scale industrial units the total outage cost is estimated at Rs 121
million on a value added of Rs 950 million. The only contempory information available as of
2011-12 on small-scale industry is the value added in the latest Pakistan Economic Survey.
This is estimated at Rs 653 billion. We use the value added as the blow up factor from the
sample to the national estimate as follows: National Estimate of Outage Costs (= (.121/0.950) *
653 Billion) equals Rs = 83.2 Billion Rs
67
Therefore, the bottom line of the research is that in 2011-12, the total cost of outages to
the small-scale industrial sector of Pakistan is projected at Rs 83 billion. This is equivalent
to $885 million, at the exchange rate of Rs 97 per US dollar.
CHAPTER 12
LOAD MANAGEMENT STRATEGY: CONSUMER’S PREFERENCES
The questionnaire contains a module to solicit consumer preferences regarding timing of
loadshedding which can reduce the costs and disruptions due to the outages. These can
provide guidance to the load management strategy by DISCOs, the formulation of which should
be a priority since loadshedding is likely to persist over the next few years.
12.1 LEVEL OF SATISFACTION WITH
CURRENT LEVEL OF SERVICE
Only 28 percent of sample firms
indicated that DISCOs kept to the
announced loadshedding schedule (see
Table
12.1).
The
percentage
is
Table 12.1
Percentage of Time DISCOs Kept to the Announced
Loadshedding Schedule
City
Average
Punjab
35.8
Lahore
35.8
Faisalabad
44.3
generally higher for cities in Punjab.
Other Cities
30
Performance of DISCOs in Sindh and
Sindh
7.8
Balochistan in this respect is particularly
Karachi
6.9
other Cities
9.5
KPK
26.7
weak,
with
effectively
no
prior
scheduling. This has had a significant
Balochistan
2.4
impact on the costs of loadshedding as
Total
28
highlighted in Chapter 6. The time
By Industrial Group
required for small-scale units to adjust
Industrial Group
to
changes
in
the
loadshedding
Food beverages and Tobacco
25.8
Textile wearing Apparel and leather
30.3
schedule is 1.9 hours on an average.
Wood and Wood products
26.6
The time is much higher for continuous
Fabricated Metal Products
25.9
process industry (2.6 hours) than for
Others Industries
26.8
Total
28.0
batch making units (1.8 hours). (see
Table 12.2).
The survey teams enquired from the
Nature of Production
Continuous Process
21.4
Batch-Process
28.6
Total
28.0
respondents if they were satisfied with
the current level of service by the DISCOs/KESC. More than half of the respondents ranked
their satisfaction level as very low while one-third ranked it as low (see Table 12.3).
68
Table 12.2
Average Time Required for Adjustment to Changes in Loadshedding Schedule
(hrs)
By City
City
Average
Punjab
2.2
Lahore
3.1
Faisalabad
1.8
Other Cities
1.5
Sindh
0.9
Karachi
0.6
Other Cities
1.6
KPK
1.8
Balochistan
1.7
Total
1.9
By Industrial Group
Industrial Group
Food beverages and Tobacco
2.1
Textile wearing Apparel and leather
1.8
Wood and Wood products
1.3
Fabricated Metal Products
1.8
Others industries
2.2
Total
1.9
By Process
Nature of Production
Continuous Process
2.6
Batch-Process
1.8
Total
1.9
69
Table 12.3
Level of Satisfaction with Current quality of Service by DISCOs/KESC
(%)
z
Very high
High
Medium
Low
Very Low
By City
Punjab
1
0
6
35
58
Lahore
0
0
7
53
40
Faisalabad
0
0
0
16
84
Other Cities
2
0
8
27
62
Sindh
5
7
14
31
43
Karachi
2
11
15
35
37
other Cities
11
0
11
22
56
KPK
0
3
41
28
28
Balochistan
0
0
0
8
92
Total
1
2
12
32
52
By Industrial Group
Food
beverages
and
Tobacco
Textile wearing Apparel and
leather
Wood and Wood products
3
4
11
34
48
1
1
10
39
50
7
4
15
9
65
Fabricated Metal Products
0
3
28
15
55
Others industries
0
1
8
39
52
Total
2
2
12
32
52
By Process
Continuous Process
0
0
5
26
68
Batch-Process
2
2
13
33
51
Total
2
2
12
32
52
The sample firms were also asked how much higher tariff they are willing to pay for better
quality and reliability of power service–essentially with no loadshedding. This provides the first
estimate of the respondent’s perception of the cost of loadshedding. On an average,
70
respondents are willing to pay an extra 30% for uninterrupted power supply as revealed by
Table 12.4. The premium for better service is highest around 48% in Karachi.
Fabricated metal product and textile, wearing apparel and leather industry it appears, have
indicated their willingness to pay
the highest additional premium for
better quality service.
Table 12.4
Additional Tariff For Better Quality of Service (with No
Loadshedding)
(%)
By City
Translated
into
the
subjective
valuation of the outage cost per
Location
Average
Punjab
26.2
hour, the average for the sample
Lahore
27.8
units is 21 Rs per Kwh.
Faisalabad
17.8
highlighted in Part 1 of the report,
Other Cities
29.8
there is a tendency for the WTP to
Sindh
42.5
be understated. The willingness to
Karachi
46.9
pay is highest, at 52 Rs per Kwh
Other Cities
33.7
in Karachi.
KPK
35.9
Balochistan
4.0
Total
30.0
As
12.2 PREFERRED CHANGES IN
TIMINGS OF LOADSHEDDING
One-thirds of the sample units are
willing to change work timings if
off-
peak
power
tariffs
are
reduced. The willingness appears
to
be
higher
in
the
smaller
provinces of Balochistan, K-PK
and Sindh. (see Table 12.5). Also,
food,
beverages and tobacco,
By Industrial Group
Food beverages and Tobacco
25.1
Textile wearing Apparel and leather
34.0
Wood and Wood products
30.6
Fabricated Metal Products
34.4
Others industries
23.5
Total
30.0
By Process
Continuous Process
29.7
wood and wood products are
Batch-Process
30.1
industries which are more willing
Total
30.0
to change work timings if off peak
tariffs are reduced.
71
Table 12.5
Perceived Outage Costs per Kwh as implied by Willingness To Pay
Rs
Location
Average
Punjab
16
Lahore
18
Faisalabad
10
Other Cities
16
Sindh
42
Karachi
52
Other Cities
22
KPK
21
Balochistan
11
Total
21
By Industrial Group
Food beverages and Tobacco
14
Textile wearing Apparel and leather
23
Wood and Wood products
26
Fabricated Metal Products
29
Others industries
17
Total
21
By Process
Continuous Process
17
Batch-Process
22
Total
21
Over 93 percent of the sample firms reported summertime as the worst season for loadshedding
(see Table 12.7). Winter time is the second worst season for loadshedding. Interestingly,
textiles, wearing apparel and leather and wood and wood products are dominant industries
categorizing wintertime as bad season primarily because they are unable to meet their orders,
especially international orders. This is principally the case with export-oriented industries.
72
Table 12.6
Percentage Willing To Change Work Timing If off Peak Power Tariffs Are Reduced
(%)
Yes
No
Total
Punjab
27
73
100
Lahore
25
75
100
Faisalabad
34
66
100
Other Cities
23
77
100
Sindh
30
70
100
Karachi
13
87
100
Other Cities
63
37
100
KPK
62
38
100
Balochistan
100
0
100
Total
34
66
100
Food beverages and Tobacco
47
53
100
Textile wearing Apparel and leather
27
73
100
Wood and Wood products
43
57
100
Fabricated Metal Products
30
70
100
Others industries
36
64
100
Total
34
66
100
Continuous Process
45
55
100
Batch-Process
33
67
100
Total
34
66
100
By City
By Industrial Group
By Process
73
Table 12.7
Worst Time of The Year for Loadshedding
Summer
Punjab
Lahore
Faisalabad
Other Cities
Sindh
Karachi
Other Cities
KPK
Balochistan
Total
Spring
By City
92
87
98
94
93
91
96
98
100
93
2
3
2
1
2
2
4
0
0
2
Rank
Winter
Fall
Total
6
11
0
5
4
6
0
2
0
5
0
0
0
0
1
2
0
0
0
0
100
100
100
100
100
100
100
100
100
100
By Industrial Group
Food beverages and Tobacco
96
3
1
0
100
Textile wearing Apparel and leather
Wood and Wood products
Fabricated Metal Products
Others industries
Total
91
89
93
97
93
1
4
0
1
2
8
7
8
0
5
0
0
0
1
0
100
100
100
100
100
15.8
3.7
4.9
0.0
.3
.2
100
100
100
Continuous Process
Batch-Process
Total
By Process
78.9
5.3
94.7
1.3
93.2
1.7
The questionnaire also contained a question
regarding the worst day of the week for
outages. While 30 percent of the respondents
Table 12.8
The Worst Day of The Week for Outages
Frequency
Percentage
Sunday
52
12.6
said all days are bad, over one-thirds said
Monday
141
34.2
Monday is the worst day. Sunday was the
Tuesday
13
3.2
worst day for about 13 percent of the
Wednesday
11
2.7
Thursday
19
4.6
respondents. The principal reasons cited for
Friday
27
6.6
this is related to the fact that Monday is the
Saturday
23
5.6
start of a work week and an outage disturbs
All days
126
30.6
the working environment. Sunday, of course,
Total
412
100.0
is family/rest day and loadshedding disturbs it. Around 21 percent of the respondents indicate
that will be helpful if the power companies provided more information relating to the scheduling
of the outage; methods to lower the electricity- intensity of operations; methods to minimize loss,
74
and; how to work more efficiently under the
circumstances (see Table 12.9). Clearly,
these should be focused upon in the load
management strategy of the distribution
companies.
Table 12.9
Information that can be provided by
Distribution companies to Units
Percentage
Work more efficiently
33.0
Information about outage
42.0
Time table for load shedding
45.5
Awareness about outage required
14.8
Consumption minimize method
9.1
Minimize line losses
4.5
Others
2.3
75
CHAPTER 13
SUGGESTIONS BY THE SAMPLE UNITS
The questionnaire at the end solicited the respondent’s views/ suggestions to help handle the
loadshedding problem in the country. Specifically, the open ended question asked for
“suggestions to reduce the costs of loadshedding to the small scale industry”. A number of
interesting suggestions emanate from the survey responses. These can be categorized as
relating to the following:

Enhancing the supply of electricity

Alternative sources of energy/ fuel use

Improving governance or management

Changes in pricing policy
Enhancing the Supply of Electricity: The highest numbers of respondents, 45% are of the
view that new dams, including Kala Bagh Dam, should be constructed to permanently enhance
the production of electricity in the country at lower cost (see Table 13.1). This suggestion
dominates the response not only from the sample units located in Punjab, but is also significant
in the case of Peshawar and Karachi. Over 30% of respondents are also of the view that
electricity should be imported. Building new power plants, installing gas pipeline from Iran to
avoid gas shortages are the other significant suggestions to enhance the supply of electricity
and curb power shortages in Pakistan. Responses are more or less, similar across industry
groups. (See Table 13.2)
Alternative sources of Energy Fuel for Energy: A number of suggestions have been given
regarding resort to alternative energy and fuel sources by the respondents. One quarter of the
respondents suggested the use of different methods of electricity generation, while 19 percent
specifically suggested the use of coal for electricity generation. Close to 9 percent of the sample
units suggested introduction of solar energy systems (particularly by the food, and beverage
industry), while there was also some mention of bio- gas and wind energy.
Improving Governance/Management. The most dominant recommendation in this category is
to curb corruption, with over a quarter of respondents emphasizing it. Minimization of electricity
theft has been suggested of 24 percent of the respondents. Only 3 percent of the sample units
proposed Privatization of the DISCOs.
76
Table 13.1
Suggestions by Sample Units by City
(% of Respondents)
Reasons
Lahore
Enhancing Supply of Electricity
Gas Pipe line from Iran to
11
avoid gas shortage
Faisalabad
Gujranwala
5
44
Sialkot
Rawalpindi /
Islamabad
Karachi
Hyderabad
Sukkur
Peshawar
Quetta
Total
7
17
7
31
0
0
9
31
13
Multan
Import Electricity
16
75
25
45
52
22
31
13
45
19
8
33
Construct
new
Dams
(including Kala Bagh Dam)
49
48
63
34
83
67
24
0
9
53
31
45
Use rental power system
3
0
0
0
4
0
4
0
0
5
15
3
Build new power plants
24
7
13
45
22
33
9
6
45
31
15
22
39
25
3
26
30
17
19
18
10
46
19
5
13
14
57
48
30
13
18
24
23
25
0
0
0
0
0
0
0
0
0
2
0
0
18
2
6
24
0
4
2
0
9
3
15
9
Introduce Wind Energy
0
0
0
0
0
0
0
0
0
2
0
0
Use of Petrol
0
0
0
0
0
0
0
0
0
0
8
0
Privatize Electric department
1
0
0
0
0
0
9
0
0
3
31
3
Need Honest Employees
1
0
0
0
0
0
0
6
0
3
8
1
Minimize electric theft
19
36
19
3
30
4
44
25
18
21
23
24
Stop Corruption
21
31
31
7
17
15
20
81
9
38
8
25
Minimize line losses
1
0
0
0
0
0
0
56
0
0
0
2
1
0
0
0
0
0
0
19
0
0
0
1
4
0
0
0
0
0
0
0
0
0
0
1
Government give subsidy on
electricity
8
38
69
28
22
11
28
6
9
7
85
22
Reduce price at source
10
7
0
0
35
26
26
13
0
3
23
12
Total
25
15
4
7
6
7
13
4
3
14
3
100
Alternative Energy Fuel Sources
Use
Coal
for
electric
8
generation
Use different method of
30
electric generation
Bio Gas system
Introduce
System
solar
energy
Governance/Management
Give awareness to people
use of electricity and avoid un
nec
Knowledge about minimize
usage
Pricing Policy
77
Minimizing line losses and the need for honest employees were also emphasized by the sample
units.
Other suggestions relate to educating the electricity consumers, regarding efficient energy use,
and proper use of electricity so that unnecessary consumption is avoided.
Pricing Policy Around one-fifth of the sample units requested for subsidy for electricity from
the government, while 12% of the respondents suggested that the price (at source) should be
reduced through economizing on costs.
To conclude, the top five suggestions emanating from the respondents of the survey are as
following:
First: Construct Dams
Second: Import Electricity
Third: Stop Corruption
Fourth: Use Different Methods of Electricity Generation
Fifth: Minimize Electricity theft.
78
CHAPTER 14
CONCLUSIONS AND POLICY IMPLICATIONS
We have highlighted in previous Chapters the principal findings on the incidence and costs of
power loadshedding to small-scale industrial units of Pakistan. In this concluding Chapter we
derive the key policy implications, starting with estimates of the multi-dimensional impact of
outages on small firms in the manufacturing sector.
14.1 IMPACT OF OUTAGES
The estimated impact of outages on small-scale units from the survey are summarised below:
(i) Outages, on the average, occur for 30 per cent of the time available during normal working
hours. During these outages about 60 per cent of the output is lost, prior to adjustments.
Following adjustments, the permanent loss of output due to outages is 9 percent. This
implies an over 36 per cent reduction in profitability. The outage cost per Kwh works out as
Rs 51 per Kwh (53 cents per Kwh). This is similar to the magnitude derived by earlier
studies in Pakistan.
(ii) Total employment in the small-scale industrial sector, according to the Labor Force Survey
for 2010-11 by the PBS, was about 3.7million. With a lower output of 9 per cent and an
employment elasticity of 0.8, the employment level in small-scale manufacturing in the
absence of outages would have been higher by about 266,000.
(iii) The impact on exports is fortunately not very large because only a minor proportion of small
industrial units are engaged directly in exporting.
(iv) According to the Pakistan Economic Survey, the annual growth rate of value added in
small-scale manufacturing has been assumed to be as high as 7.5 per cent during the last
five years, when the overall GDP growth rate has on average been only 3 per cent. The
survey reveals that the growth rate of this sector is likely to be between 3.5 per cent and
4.5 per cent. With a share of 5% of the sector in the GDP, it appears that the GDP of
Pakistan in 2011-12 has probably been overstated by about Rs 128 billion. Accordingly, the
revised estimate of the size of the small-scale manufacturing is Rs 525 billion as compared
to the estimate in PES of Rs 653 billion.
(v) Within small-scale industry, the cost of outages appears to be the highest in Sindh
province; in textiles, apparel and leather industry and in continuous process units as shown
in Table 14.1. Gives the relatively low level of output, the outage cost as a percentage of
the value of production emerges the highest in Balochistan.
79
14.2 AFFORDABILITY OF POWER TARIFFS
The total costs of electricity consumption, that is, the costs of public supply and of outage
costs, as a percentage of the value of production are given in Table 14.1. On average, these
costs aggregate to 12 per cent, with a share of outage costs in these costs of 44 per cent.
According to the Census of Manufacturing (CMI) of 2005-06, the share of electricity cost in
value of production, for industry as a whole, was about 3 per cent only, during the period
when there were no outages. As such, the costs of energy have gone up sharply both due to
escalation in tariffs and in the presence of outages.
It is clear that small-scale units now have very limited affordability of further enhancement in
tariffs. Already, profits are down by over 36 per cent in relation to the potential profits in the
absence of outages, a further reduction in profits could lead to the closure of a large number
of small units and displacement of many more workers.
Table 14.1
Total Costs of Electricity Consumption as a percentage of the Value of Production
(000 Rs)
Value of
Production
Electric Cost
A.
B.
C.
LOCATION
Punjab
Lahore
Faisalabad
Other Cities
Sindh
Karachi
Other Cities
K-PK
Balochistan
INDUSTRY
Food, Beverages &
Tobacco
Textiles & Apparel
Wood & Wood Product
Fabricated Metal Products
Others
PROCESS
Continuous Process
Batch-Making
TOTAL
Electricity Cost as Per
Cent of Value of
Production
of public
supply
Total Cost
of Outages
Total
Cost
(at current
prices)
312
399
259
252
602
590
626
307
217
227
245
195
227
519
418
721
321
88
539
644
454
479
1121
1008
1347
628
305
4359
3649
5354
4497
8927
6749
13285
6333
1178
12.4
17.6
8.5
10.7
12.6
14.9
10.1
9.9
25.9
461
376
837
7224
11.6
381
269
306
322
294
246
224
267
675
515
530
589
4922
6592
4621
4448
13.7
7.8
11.5
13.2
609
341
366
414
281
293
1023
622
659
6256
5351
5435
16.3
11.6
12.1
80
14.3 COMPETITIVE DISADVANTAGE OF SMALL UNITS
The high incidence of outages has placed small-scale units at an even greater competitive
disadvantage with respect to large-scale units in the same industry. This is highlighted in Table
14.2 which compares the impact of laodshedding by size of unit. While the relevant magnitudes
for small-scale units are taken from the survey carried out as part of this study, the
corresponding numbers for large-scale units are extracted from IPP (2009).
Table 14.2
Comparison of the Impact of Loadshedding by Size of Industrial Unit
Cost of Loadshedding
Per cent of time lost due to loadshedding a (%)
Per cent of output lost that is recovered (%)
Per cent of output lost (%)
Cost of loadshedding as per cent of Value Added (%)
Fall in profitability (%)
Incidence of Loadshedding and Adjustments thereof
Annual hours of Loadshedding (hrs)
Per cent of units with Generators (%)
Per cent of units making some adjustment (%)
a including restart times
b n.e. = not estimated
c excluding multiplier effects
*from IPP Annual Review of 2009
Large-Scale
Industrial Sector
(2008)*
Small-Scale
Industrial
Sector
(2012)
20
60
6
9
n.e.b
30
60
9
13
36
2023
75
84
2623
68
78
It appears that not only is the incidence of outages higher in the case of small units but also the
extent of recovery of output through adjustments is less. Consequently, the burden of outage
costs as a percentage of value added is higher.
This is probably one of the reasons why a large proportion of public limited companies continue
to enjoy high level of corporate profits, at a time when the economy is growing slowly. This has
resulted in a boom in the stock market.
The equity implications of displacement of production by small-scale units are serious. Not only
does this imply less inclusive growth but also the emergence of quasi-monopolistic elements in
the economy. Perhaps the time has come to consider ‘life-line’ charges for small-scale industrial
units of the type offered to domestic consumers.
81
14.4 INVESTMENT IN GENERATORS
The
cost-benefit
ratio
of
investment
in
Table 14.3
Benefit-Cost Ratio* of Self-Generation
Benefit-Cost
Ratio
generators is high. For small-scale units, this
ratio is high at 3.7 as given in Table 14.3. But
despite this high ratio, one-thirds of the units
have not invested in self-generation and the
extent of substitution of the conventional
power source by those who have generators is
only about one half. Therefore, the scope for
self-generation is about three times the
INDUSTRY
Food, beverages and tobacco
Textiles and Apparel
Wood and Wood Products
Fabricated Metal Products
Other Industries
PROCESS
Continuous-Process
Batch-Making
Total
*
2.2
4.6
4.8
3.0
2.3
4.3
3.6
3.7
present level.
This probably a reflection of inability of small firms to self-finance the acquisition of generators
(with appropriate capacity) and/or because of imperfections in the capital market, whereby small
firms have limited access to bank credit and face high interest rates in the informal sector. As
such, there is a strong case for increasing the access to institutional credit through the offering
of a special package for purchase of small generators. International agencies could provide
special support in this area.
It needs to be emphasised that this is not an efficient solution. There are significant economies
of scale in the use of generators. A multitude of small generators would imply higher costs of
adjustment to outages. But this
outcome is better than the high
idle factor costs that would result
in the absence of such an
adjustment.
At the minimum, it is suggested
that the present tariff structure on
imported generators be reduced.
The
precise
nature
of
Table 14.4
Proposed Tariff Structure on Imported Generators
(Percent)
Description
Generating Sets (diesel
HS Code
or semi-diesel engines)
with:
8502.1110
Capacity < 5 KVA
8502.1120
5 KVA to 20 KVA
8502.1130
20 KVA to 50 KVA
Source: Customs Tariff,2012-13, FBR
Existing
Tariff
Proposed
Tariff
0
20
20
0
5
10
the
proposal is given in table 14.4.
82
It appears that in the presence of extremely high and persistent levels of loadshedding and
inadequate response by the power sector, second best solutions may have to be adopted to
reduce the associated outage costs.
14.5 LOAD MANAGEMENT STRATEGY
The survey of small-scale industrial units has provided valuable insights on the optimal load
management strategy to minimize the costs of outages as follows:
a) Priority must be given for uninterrupted supply to continuous process units where outage
costs are higher.
b) For the same level of outages (in terms of time lost), outages of shorter duration impose
lower costs and long outages should be avoided.
c) There is a case for reducing off-peak tariffs to facilitate changes in timing daily of
production activities, especially by firms which have low labor costs in relation to costs of
power consumption.
d) Seasonal variation in tariffs may also be contemplated so that lower tariffs are charged in
the winter season, when electricity demand currently is relatively low.
e)
Distribution companies may consider the setting up of extension services, especially for
small firms, to enable them to find ways to conserve energy and to make the right kinds of
adjustments to outages.
In addition, the respondants have made a host of suggestions, which have been listed in
Chapter 7.
In conclusion, the survey of small-scale units has revealed the difficult situation that these
entities face in the presence of the high level of loadshedding. As such, along with the strategy
to reduce outage costs in the face of a given quantum of loadshedding, efforts have to be made
to reduce sharply the level of loadshedding. The returns to the economy of reducing outages
are very large. If the marginal cost of providing electricity through the public system is Rs 12 per
Kwh and the saving in outage cost is Rs 51 per Kwh, as estimated from the survey, then the
benefit cost ratio is as high as 4.2.
Measures of the type recommended above will have to be adopted on a top priority basis if the
interests of thousands of small entrepreneurs and millions of workers are to be protected. Their
voices are seldom heard currently in the corridors of power.
83
PART III
COST OF LOADSHEDDING TO
COMMERCIAL/SERVICES
SECTOR
84
CHAPTER 15
INTRODUCTION
This part of the report presents the findings on costs of loadshedding to commercial/services
establishments in Pakistan, quantified on the basis of data obtained from a nationwide survey of
such units.
The report is organized in nine chapters. Chapter 16 presents the methodology used for
qualification of costs due to outages. Chapter 17 describes the survey including the sampling
methodology and assessment of the quality of data collected, given the complex technical
nature of the survey. Subsequent Chapters up to Chapter 21 present the magnitudes of key
parameters like the relevant characteristics of the responding units, incidence of outages, level
and pattern of adjustments and the magnitude of different outages costs. Chapter 22 highlights
the suggestions by sample units for reduction in incidence and costs of outages.
Chapter 23 gives a summary of the principal findings and the major policy implications emerging
from the research. It is clear from the results that commercial sector has faced severe
disruptions due to the high and growing incidence of loadshedding. As such, the economic
return of reducing outages and of facilitating the process of adjustment to these outages is very
high. This could contribute to some revival of the economy and reduce unemployment.
Thanks are due to the sample units for responding to a questionnaire, which was complex and
difficult to administer. Thanks are also due to the survey team which travelled all over the
country and sometimes found itself in a difficult law and order situation, especially in Karachi.
The main text of the report is presented in a non-technical manner. Technical analyses are
included in the Technical Annexes.
Any defects which remain are of course, the responsibility of the authors.
85
CHAPTER 16
THEORETICAL FRAMEWORK
16.1 OUTAGES AND A FIRM’S BEHAVIOUR
The behavior of a firm in the presence of frequent and persistent outages has been modeled in
the Technical Annex VI. The firm is assumed to be operating in a competitive environment,
given the smallness of its size; and pursues profit maximization. The results derived from this
theoretical framework are as follows:
iii.
If outages are seen as, more or less, permanent in nature then the optimal size of the firm
is lower than in the absence of outages. In particular, there is a tendency to shed some
labor.
iv.
The likelihood that the firm will make adjustments to recover some of the lost output
depends on the following:
e) The extent to which the market situation is favorable for the firm
f)
The electricity-intensity of the business; the lower the intensity the greater the
likelihood that the firm will make an adjustment
g) The lower the costs of adjustment.
h) The larger the outage and the level of expectation that this will continue.
Based on the results from the theoretical analysis we present in Figure 16.1 visually the change
in the equilibrium of a firm in the presence of outages
Figure 16.1
Adjustment by a Firm to Outages
86
Type I firm initially experiences outages which reduce production from 𝑄0 − 𝑄1 . At 𝑄1 the gap
between price and marginal cost is AB. AB is larger the bigger the outage. The marginal cost
curve of adjustment by the firm is given by XY. If XY is too high, then the firm makes no
adjustment.
In the case of type II firm, Y lies between A and B. As such, the firm makes an adjustment and
the new output level is 𝑄𝑒 . Type I firm reduces output by (𝑄0 − 𝑄1 ) while the loss of output of
Type II firm is (𝑄0 − 𝑄𝑒 ). Also, the fall in profit of Type I firm due to outages is ABC. In the case
of Type II firm it is BYZC.
The above theoretical framework is used to develop the methodology for quantifying the cost of
outages.
16.2 METHODOLOGY FOR QUANTIFICATION OF COST OF OUTAGES
We have indicated above that in the presence of persistent outages firms will be inclined to
make adjustments. The extent and nature of adjustments will depend on a number of factors
which have also been identified above. The methodology used for quantifying the cost of
outages is based largely on that developed by Pasha, Ghaus and Mallik [1989]. Fig 16.2
presents a flow chart for identifying the different types of outage costs.
There are two types of costs associated with outages. The first type is direct costs which
consist primarily of the value of lost production and spoilage costs. The second type is
adjustment costs. The particular mechanisms chosen for recovering some or all of the output
lost will be based on cost minimization considerations. Accordingly, a firm will opt for a type of
adjustment as long as it is cheaper than other options. Therefore, firms may opt for multiple
adjustments, especially when the size of the outage is large.
Direct Costs
In order to derive the magnitude of direct costs, we designate the following:
𝑛𝑖 = number of times of occurrence of outage daily on average of duration i. i = 1, 2, 3, 4, 513
𝜖𝑖 = proportion of output lost during an outage of duration i
𝛾𝑖 = restart time after an outage of duration i.
13
The durations are 0-1/2 hr; ½ hr to 1 hr; 1 hr to 2 hrs; 2 hrs to 3 hrs; 3 hrs and above.
87
Figure 16.2
Flow Chart Showing Costs of Outages
OUTAGES
SHUTDOWN
SPOILAGE COST
RESTART TIME
OUTPUT
RECOVERED
OUTPUT LOST
PERMANENTLY
TOTAL DIRECT
COSTS
ADJUSTMENTS
USE OF
GENERATORS
OVER TIME
EXTRA SHIFTS
NET IDLE
FACTOR COST
OTHERS
TOTAL INDIRECT (ADJUSTMENTS COSTS)
TOTAL OUTAGE
COST
88
The total number of outages during the year is given by
NOUT = ∑5𝑖=1 𝑛𝑖 × 365
..…………………………………………………...(2)
The total time lost due to outage is
TOUT = ∑5𝑖=1( 𝑛𝑖 [𝑑𝑖 + 𝛾𝑖 ] × 365)
……..………………………………………………(3)
Where 𝑑𝑖 is the duration of the outage.
The potential extent of output loss due to outages is given by
LOUT = ∑5𝑖=1 𝑛𝑖 [𝑑𝑖 + 𝛾𝑖 ] 𝜖𝑖 × 365
.……………………………………………………(5)
But the firm may not be operating throughout the year and for 24 hours each day. Therefore, if
H is the normal hours worked during the year, the actual output lost is given by
….…………………………………………………(5)
𝐻
ACOUT = 𝐿𝑂𝑈𝑇. 8760
And the value of this loss is as follows
VOUT = ACOUT.V
.……………………………………………………(6)
Where V is the value added by the firm per hour.
However, the firm may recover some of the output lost through adjustments. If 𝜆 is the extent of
output recovered then we have the expression for the net idle factor cost, NIFC, as follows:
NIFC = (1-λ) VOUT
……………………………………………………(7)
Box 16.1 gives a simple numerical example for calculation of the net idle factor cost incurred by
a sample unit due to outages on the basis of data obtained from the survey.
The other part of direct costs is spoilage costs. We represent
𝑆𝑖 = spoilage cost (in rupees) in each outage of duration i
Then the spoilage cost, SPC, is derived as follows:
𝐻
SPC = ∑𝑛𝑖=1 𝑛𝑖 𝑠𝑖 . 365 × 8760
……….…………………………………………….(8)
And the total direct costs of outages are
TDC = NIFC + SPC
…………………………………………………………(9)
89
Box 16.1
A Numerical Example of Quantification of Net Idle Factor Cost
Suppose a firm experiences outages of duration of one hour six times a day (over the 24 hours) and the
restart time after the outage is half an hour.
Then
NOUT = 6 × 365 = 2190
TOUT = 6 × 365× 1.5 =3285
If the proportion of output lost is 50%. Then
LOUT = 3285 × 0.5 = 1642.5
The firm normally operates one shift (of 8 hours) for 300 days. Then
H = 8 × 300 = 2400
And, ACOUT = 1642.5 ×
2400
8760
= 450
If the value added per hour is Rs 1000, then
VOUT = 450 × 1000 = 450,000Rs
and if the proportion of output recovered through adjustments is 60%, then λ = 0.6 and NIFC is given by
NIFC = 450,000 × (1-0.6) = 180,000 Rs
Adjustment Costs
Generators Cost
A key response by commercial units, which is being observed in Pakistan, is the resort to own
sources of energy supply through investment in generators. This is motivated by the high and
rising incidence of outages since 2007 and the growing recognition that power loadshedding
(along with gas shortages) could persist over the next many years.
In practice, the extent of substitution of the conventional power source (through DISCOs)
depends on the energy-intensity of operations (as derived in the theoretical framework above),
on the possibility of making other cheaper adjustments and the cost of capital for acquisition of
generators. The latter is likely to be relatively high for small commercial units.
The survey of units has indicated, first, whether a unit has a generator or not, second, the
capital cost of the generator, third, monthly running cost of fuel for operating the generator and
fourth, other costs (including labor, repairs and maintenance cost, etc.) on a quarterly basis. As
such, we designate the following:
𝐾𝑔 = capital cost of generator
FOC = fuel operating cost per month
OPC = other operating costs quarterly
This leads to the estimate of the overall annual generator cost, GENCO, as follows:
GENCO = (𝜏 + 𝛿)𝐾𝑔 + 12(FOC) +4(OPC)
……...……………………………………………(10)
90
Where, 𝜏 is the cost of capital and 𝛿 is the rate of depreciation. The combined value of 𝜏 𝑎𝑛𝑑 𝛿
is taken as 0.32.
However, the use of generators implies savings in costs of power supply from the local DISCO.
Therefore, the net cost, NGENCO, is given by
NGENCO = GENCO – k (TOUT) × (ADJG) ×tf .
𝐻
………..………………………………(11)
8760
Where
K = electricity consumption per hour in Kwh
TOUT = total hours lost as derived in equation (3)
ADJG = extent of adjustment by use of generators
tf = tariff per Kwh of the DISCO.
Other Adjustments
Other adjustments include the following and are more in the nature of short-run adjustments
when loadshedding is seen as temporary in nature:
iv.
More intensive utilization of existing plant and machinery during times when there is no
loadshedding
v.
Overtime or additional shifts to make up for at least part of the output loss
vi.
Changes in working days or timings.
The survey reveals that the majority of establishments have not made significant adjustments of
the above type and the costs associated with these adjustments are not large. The methodology
used for quantifying these costs has been taken from Pasha, Ghaus and Mallik [1989]. They
are represented by OTC.
Overall, the total adjustment cost, TAJCO, is derived as
TAJCO = NGENCO + OTC
……………………………………………………….(12)
And the total outage out cost, TOUTCO, as follows:
TOUTCO = TDC + TAJCO
……………………………………………………….(13)
The magnitudes of the different components of outage costs are presented in Chapter 6, by
location and type of business of the sample units.
91
CHAPTER 17
THE SAMPLING FRAMEWORK AND ITS DISTRIBUTION
17.1 SAMPLING FRAMEWORK
The primary instrument of data collection was a survey on a pre-designed and tested
questionnaire of a stratified (by city, type of business) national random sample of commercial
units (see Figure 17.1).
Figure 17.1
Sampling Strategy
Primary Sampling Unit
Commercial Units in Province
Secondary Sampling Unit
Commercial Units in Cities
Tertiary Sampling Unit
Various Types of Commercial Units
Individual Commercial units in various Groups
The population of the establishments was obtained from the Economic Census 2005,
published by the Pakistan Bureau of Statistics (PBS). The national distribution of commercial
establishments by province and type of business is presented in Table 17.1 and Figure 17.2.
The derived sample distribution by city and type of business is presented in Table 17.2.
Once the sample distribution across cities and type of business groups was finalized, the
individual sample units were selected from upper, medium and lower income commercial
centers. The individual unit was selected through random walk.
The questionnaire administered on the sample respondents contains five modules: basic
information on sales/employment/costs; incidence of outages; costs of outages; adjustment to
outages, and, preferred load management practices. Though the questionnaire was structured,
the last question was open-ended asking the respondents to make suggestions to reduce the
costs of loadshedding. This provides the respondent’s perspective on actions to counter the
problem.
92
Table 17.1
National Distribution of Commercial Establishments in the Economic Census, 2005 by
Province and Sector
Distribution of Establishment*
Percentage
64.6
18.9
13.5
2.3
0.6
100
Punjab
Sindh
K-PK
Balochistan
Islamabad
Total
Distribution of National Employment in Commercial Sector
%
67.3
2.2
2.1
28.4
100.0
Wholesale & Retail Trade Establishments Restaurants & Hotels
Transport & Communications
Financing, Insurance, Real Estate & Business Services
Community, Social and Personal Services
Total
Figure 17.2
National Distribution by Province and Sector
By Province
By Sector
Balochist
an
2%
N.W.F.P
14%
Sindh
19%
Punjab
65%
Financing,
Insurance,
Real Estate
& Business
Services
2%
Transport
&
Communic
ations
2%
Communit
y, Social
and
Personal
Services
29%
Wholesale
& Retail
Trade
Establishm
ents
Restaurant
s & Hotels
67%
93
Table 17.2
Distribution of Sample units by City, Province and Sector
Provinces
Punjab
Sindh
KPK
Balochistan
Total
Cities
Wholesale &
Retail Trade
Establishments
Restaurants
& Hotels
Transport &
Communications
Community,
Social and
Personal
Services
Total
3
Financing,
Insurance, Real
Estate &
Business
Services
4
Lahore
22
5
9
43
Faisalabad
12
3
1
2
5
23
Sialkot
3
1
0
1
1
6
Gujranwala
6
1
1
1
2
11
Multan
9
2
1
2
3
17
Rawalpindi/Islamabad
16
5
1
4
8
34
Total
68
17
7
14
28
134
Karachi
18
4
5
6
9
42
Hyderabad
4
1
1
1
2
9
Sukkur
2
1
0
1
1
5
Total
24
6
6
8
12
56
Peshawar
12
2
3
3
4
24
Mardan
3
1
1
1
1
7
Abbottabad/Bannu
3
1
1
1
1
7
Total
18
4
5
5
6
38
Quetta
10
3
2
3
4
22
Total
10
3
2
3
4
22
120
30
20
30
50
250
94
The survey was successfully administered on 250
units as targeted. Following the process of edit
and consistency checking of the completed
questionnaires, 241 units, over 96 percent of the
sample, have been included in the analysis.
17.2 PROFILE OF RESPONDENTS
Distribution of selected units for analysis by city is
given in Table 17.3. 54 percent of the sample
units are in the province of Punjab, while about 23
percent are in Sindh. From the remaining, 15
percent are in Khyber-Pakhtunkhwa (K-PK) and 8
Table 17.3.
Distribution of Sample Analyzed by
Cities
Cities
Lahore
Faisalabad
Gujranwala
Multan
Sialkot
Rawalpindi / Islamabad
Karachi
Hyderabad
Sukkur
Peshawar
Mardan
Abbotabad
Quetta
Total
Numbers
%
43
21
9
14
5
37
43
8
4
25
6
6
20
241
17.8
8.7
3.7
5.8
2.1
15.4
17.8
3.3
1.7
10.4
2.5
2.5
8.3
100.0
percent in Balochistan. The distribution by type of business is given in Figure 17.3. 53 percent of
the sample respondents were from wholesale and retail sector while 15 percent each belonged
to the restaurant and hotel sector and community, social and personal services.
Figure 17.3
Distribution of Selected Units by Sector
Community, Social and
personal Services
15%
Finance, Insurance, Real
Estate & Business Services
10%
Transport and
communication
7%
Whole sale Retail Trade
Establishments
53%
Restaurant and Hotel
15%
95
17.3 CHARACTERISTICS OF SELECTED UNITS
Basic Information
Establishments,
on
an
Table 17.4.
Average
Employment
by Sector, 2012
average, worked 316 days a
year,
with
hotels
and
restaurants working 331 days
a year and transport and
communication working 320
days annually. The sample
units on an average employed
Group
Whole sale Retail Trade Establishments
Restaurant and Hotel
Transport and communication
Finance, Insurance, Real Estate & Business
Services
Community, Social and personal Services
Total
Employment
(No)
3.8
11.0
10.5
4.2
5.5
5.6
about 6 persons, the average employment being 11 in hotels and restaurants and transport and
communications (see Table 17.4).
Sales and Value Added
The average sales of the respondent units chosen for analysis in 2012 is projected at Rs 4.7
million (see Table 17.5) demonstrating a low growth of almost 7 percent over the 2011 level.
Their operating expenses average Rs 3.3 million, implying an operating profit of Rs 1.4 million.
Table 17.5.
Average Sale and Operating Expenses of
Sample Units 2012
Industrial Group
Total
(Rs in Thousands)
Operating
Total
Expenses
Operating
as % of
Expenses
Sales
3086
75
Whole sale Retail Trade Establishments
4124
Restaurant and Hotel
6427
4790
75
Transport and communication
9138
6042
66
Finance, Insurance, Real Estate & Business Services
3622
1933
53
Community, Social and personal Services
3839
2529
66
Total
4720
3341
71
The average value added by sample units in 2012 is estimated at Rs 2.1 million, highest being
in transport and communication followed by restaurants and hotel (see Table 17.6). Sample
units have, on an average, purchased almost 10 thousand kilowatt hours (Kwh) of electricity
annually from the public distribution companies. Value added per Kwh of electricity consumed is
96
Rs 214. This is the first estimate of the cost of loadshedding, not adjusted for any recovery of
value added lost.
Table 17.6
Average Value Added, Electricity Purchased and Value Added per Kwh of Electricity of
Sample Units
Value Added
(Thousands)
Electricity
Consumed
(Kwh)
Value Added
Per Kwh (Rs.)
Whole sale Retail Trade Establishments
1675
10269
163
Restaurant and Hotel
2853
11761
243
Transport and communication
5458
14403
379
Finance, Insurance, Real Estate & Business Services
1742
6427
271
Community, Social and personal Services
1695
7365
230
Total
2125
9949
214
Industrial Group
Operating Costs
Turning next to operating costs, as mentioned above, average annual operating cost of sample
units is Rs 3.3 million. Out of this, the highest proportion, (74 percent), is spent on purchase of
raw materials, followed by wages (17 percent) (See Table17.7). Electricity costs purchased from
the distribution companies and self-generation combined account for 6 percent of the costs.
Table 17.7
Operating Cost Structure of Sample Units
(%)
Industrial Group
Whole sale Retail
Trade
Establishments
Restaurant and
Hotel
Transport and
communication
Finance,
Insurance, Real
Estate & Business
Services
Community, Social
and personal
Services
Total
Total
Operating
Cost
(Rs. In
Thousands)
(Percentage of Operating Cost)
Wages/
Salaries
Raw
Material
Repairs/
Maintenance
Cost of
Electricity
Cost of Self
Generation
Others
3086
11.9
79.8
1.4
3.9
1.6
1.3
4790
18.5
70.9
1.8
4.6
2.5
1.6
6042
34.8
59.0
2.4
2.3
1.9
1.8
1933
17.7
73.1
1.9
3.6
2.7
1.1
2529
20.4
71.6
2.0
3.4
2.0
0.7
3341
17.5
74.0
1.7
3.8
1.9
1.3
97
Attainment of Production Targets
Sample units, on an average, have been able to achieve 73 percent of their sales target, as
shown in figure 17.4. The highest rate of target achievement is by community, personal and
social services at 82 percent, followed by restaurants and hotels at 74 percent.
Interestingly, inter-city differences are prominent in the achievement of sales targets. While
Faisalabad, Gujranwala, Mardan and Quetta were able to achieve over 79 percent of their
target, possibly because the targets were already modified somewhat to allow for the presence
of loadshedding, Karachi, Rawalpindi/Islamabad and Multan’s achievement of target sales was
among the lowest.
Figure 17.4
Actual as Percentage of Target Sale in 2012
(%)
Sector
71.2
74.3
72.7
82.1
66.5
73.0
By City
79.4 78.9 79.4
78.9
75.4 79.0 79.6 69.8 72.2
73.0
72.8 74.0
68.2 65.0
98
When enquired as to why the sales target was not attained, the principal reasons given are bad
law and order situation, high level of power outages, and market related factors, principally high
inflation. (see Table 17.8).
Table 17.8
Reasons Why Production Target was Not Attained*
(%)
Reason for
Not Attaining
Targets
Whole sale Retail
Trade
Establishments
Restaurant and
Hotel
Transport
and
communica
tion
Power outages
Law and order
83
79
29
Finance,
Insurance,
Real Estate
& Business
Services
29
81
79
65
100
92
73
83
46
55
47
50
44
48
16
12
18
8
20
15
66
52
53
38
72
60
High inflation &
market
competition
Shortage or
high cost or
raw materials
Other reasons
Community,
Social and
personal
Services
88
Total
*accounts for multiple responses
99
CHAPTER 18
THE INCIDENCE OF LOADSHEDDING AND OUTPUT LOSSES
This Chapter quantifies the incidence of loadshedding and the resulting output losses as
revealed by the national sample of commercial/services establishments.
18.1 INCIDENCE AND PROFILE OF LOADSHEDDING
The costs of loadshedding will, to a large extent,
depend on the frequency and duration of
outages. The incidence of loadshedding is given
Table 18.1
Frequency of Loadshedding in 2012
By Province
Location
in Table 18.1. Overall, the average number of
Average
Punjab
2189
Sindh
673
989 . Highest number of times outages have
KPK
947
occurred in Punjab at 2189, followed by
Balochistan
1077
Balochistan at 1077, and K-PK, 947. Clearly, the
Total
989
outages in Pakistan in 2012 is estimated at
14
average incidence is lower in Sindh, of 673
times a year at over 30% below the national
average.
Sector-wise,
the
highest
incidence
was
experienced by the wholesale and retail trade
By Sector
Whole sale Retail Trade Establishments
1078
Restaurant and Hotel
834
Transport and communication
880
Finance, Insurance, Real Estate &
905
Business Services
Community, Social and personal
establishments (1078) followed by community,
Services
social and personal services (933) and finance,
Total
933
989
insurance, real estate and business services
(905).
The distribution of outages by duration is given in Table 18.2. Generally, the outage duration is
1 to 3 hours a day. The highest number of outages occurs for 1 to 2 hours a day (61 percent),
followed by outages of 2-3 hours a day (23 percent). 11 percent of outages have duration of half
to one hour while 3 percent of outages last for over four hours. There is a divergence in the
provincial patterns. In Punjab, over 80 percent of the outages last for 1-2 and 2-3 hours. In
Sindh almost three-fourths of the outages last for 1-2 hours. In Balochistan 80 percent of the
outages are of 2-3 hours. This pattern of outages is likely to have significant implications for the
costs of outages as will be discussed in Chapter 20.
14
from January to September 2012. The annual incidence was estimated by multiplying by 1.33.
100
Table 18.2
Percentage Distribution of Average Length of Outages, 2012
(%)
By Province
Location
Punjab
Sindh
KPK
Balochistan
Total
Less than ½ hr
1/2 – 1hr
1-2 hrs
2-3 hrs
More than 4 hrs
0
2
0
0
0
15
14
0
0
11
66
73
51
20
61
15
8
43
80
23
4
0
5
0
3
0
15
57
24
4
0
0
6
0
56
71
29
29
6
0
0
4
75
21
0
3
11
69
14
0
0
11
61
23
3
By Sector
Whole sale Retail Trade
Establishments
Restaurant and Hotel
Transport and communication
Finance, Insurance, Real Estate &
Business Services
Community, Social and personal
Services
Total
The overall duration of outages, which includes
both the time lost due to an outage and the
restart time (time lost in restarting work following
an outage), is presented in Table 18.3. The total
hours, on an average, lost per annum due to
loadshedding are estimated at 1697 for the
sample units. The highest number of hours lost
are in Balochistan and K-PK . Overall in Punjab,
the average number of hours lost per annum is
1743, in Sindh 979, K-PK 2025 and Balochistan,
2618. These durations are for 24hrs a day for
365 days a year. Clearly, the actual total time
Table 18.3
Duration of Outages
(Outage + Restart Time) [Hours]
By Province
Location
Punjab
Sindh
KPK
Balochistan
Total
Average
1743
979
2025
2618
1697
By Sector
Whole sale Retail Trade Establishments
Restaurant and Hotel
Transport and communication
Finance, Insurance, Real Estate &
Business Services
Community, Social and personal
Services
Total
1796
1531
1635
1550
1633
1697
lost depends on working hours during the year. Wholesale, retail trade, transport and
communications and community, social and personal services establishments take the brunt,
losing respectively 1796 and around 1635 hours.
18.2 EXTENT OF TOTAL TIME LOST
The proportion of production time lost is given in Table 18.4. Overall, commercial
establishments in Pakistan, on an average, are likely to lose 19 percent of their working time in
2012 due to loadshedding. The highest, 30 percent, is lost in Balochistan, and the least, 11
percent, in Sindh. Wholesale, retail trade, transport and communications and community, social
101
and personal services establishments lose about 19-20 percent of their production time, while
the minimum loss is for restaurants and hotels, at 17 percent.
Table 18.4
Proportion of Time Lost during outages
(%)
By Province
Cities
Average
Punjab
Sindh
KPK
Balochistan
Total
20
11
23
30
19
By Sector
Whole sale Retail Trade Establishments
Restaurant and Hotel
Transport and communication
Finance, Insurance, Real Estate & Business Services
Community, Social and personal Services
Total
20
17
19
18
19
19
18.3 SEASONALITY OF OUTAGES
A significant seasonality in the incidence of loadshedding emerges from the data (see Table
18.5). The peak loadshedding months are July (accounting for almost 16% of the hours of
loadshedding) closely followed by June and August. May also emerges as a high incidence
month, accounting for over 14% of loadshedding hours. The pattern appears to be similar for all
four provinces and business establishments.
Table 18.5
Seasonality in Outages
(% of Outage)
Province
By Province
March April May
June
July
August
September
14
14
14
14
14
15
14
17
15
15
16
15
18
16
16
15
15
15
15
15
14
14
13
14
13
13
13
13
14
14
15
16
16
16
16
16
16
15
15
14
13
13
12
13
14
14
15
16
15
14
13
14
14
15
16
15
13
13
13
14
15
16
15
13
Punjab
Sindh
KPK
Balochistan
Total
13
14
11
13
13
Whole sale Retail Trade Establishments
Restaurant and Hotel
Transport and communication
Finance, Insurance, Real Estate &
Business Services
Community, Social and personal Services
13
13
13
Total
14
14
12
14
13
By Sector
102
18.4 EXTENT OF OUTPUT LOST DURING OUTAGES
Loadshedding leads to a complete shutdown for 10 percent of sample units, with highest
proportion being in Punjab. However, for 54 percent of the firm it results in partial shutdown.
(see Table 18.6).
Table 18.6
Nature of Impact of Loadshedding
(%)
Complete
Shutdown
Partial
Shutdown
No Impact
By Province
Punjab
Sindh
KPK
Balochistan
Total
15
6
3
0
10
50
41
73
80
54
35
53
24
20
36
By Sector
Whole sale Retail Trade Establishments
Restaurant and Hotel
Transport and communication
Finance, Insurance, Real Estate & Business Services
Community, Social and personal Services
Total
11
0
6
13
14
10
53
71
41
50
53
54
36
29
53
38
33
36
In the questionnaire the respondents were asked to rank ‘what is most disruptive about an
outage’. For about 37 percent of the respondents, loss of sales was the most disruptive
consequence of loadshedding, for 30% it was establishment/equipment shut down and for, 26
percent, idle labor (see Table 18.7). Loss of customers was cited as the other disruption.
Table 18.7
Ranking of Disruptions Due to Outages
(%)
Equipment
Shut down
Labor
will be
idle
Sales
will be
lost
Loss of
Customer
s
Total
27
22
26
30
20
26
33
31
30
46
55
37
9
0
0
14
0
7
100
100
100
100
100
100
29
15
29
21
28
26
34
41
53
46
28
37
10
12
0
4
0
7
100
100
100
100
100
100
By Province
Punjab
Sindh
Karachi
KPK
Balochistan
Total
30
47
44
11
25
30
By Sector
Whole sale Retail Trade Establishments
Restaurant and Hotel
Transport and communication
Finance, Insurance, Real Estate & Business Services
Community, Social and personal Services
Total
28
32
18
29
44
30
103
CHAPTER 19
ADJUSTMENTS TO LOADSHEDDING
This chapter focuses on the types of adjustments that firms make to outages in Pakistan. The
extent of sales/output that is not recovered following the adjustments is also quantified.
19.1 NUMBER AND TYPES OF ADJUSTMENTS
Table 19.1 presents the estimates of frequency of different types of adjustments by
commercial/service establishments. It appears that almost 39 percent of the firms in the sample
are unable to make any form of adjustment. 57 percent make one adjustment, 2.5 percent make
two types of adjustment while less than 2 percent are implementing three or more types of
adjustments.
Table 19.1
Percentage of Sample Units by Number of Adjustments by Group
No
Adjustment
By Province
Punjab
Sindh
KPK
Balochistan
By Sector
Whole sale Retail Trade
Establishments
One
Adjustment
Two
Adjustments
Three or More
Adjustments
Total
34.9
33.3
48.6
60.0
59.7
60.8
51.4
40.0
3.1
3.9
0.0
0.0
2.3
2.0
0.0
0.0
100
100
100
100
46.8
50
1.6
1.6
100
Restaurant and Hotel
Transport and communication
Finance, Insurance, Real
Estate & Business Services
Community, Social and
personal Services
20.6
35.3
73.5
64.7
5.9
0
0
0
100
100
37.5
50
4.2
8.3
100
30.6
66.7
2.8
0
100
Total
38.8
57
2.5
1.7
100
The frequency of different types of adjustments is given in Table 19.2. The highest frequency of
adjustment is in the form of self-generation. For the national sample, this is 60 percent. It is the
highest in Punjab, follow by Sindh, and the lowest in Balochistan. On average generators are
able to substitute for 58 percent of the public source.
Beyond the use of generators, some firms adjust by working overtime and some by changing
shift timings. These adjustments are practiced in Punjab and Sindh while no adjustment, other
than the purchase of generator, is tried in K-PK and Balochistan.
104
Clearly for the commercial/services sector the mechanism to reduce the impact of loadshedding
is primarily limited to acquiring self- generation capability. As shown in Table 19.3, either the
establishment runs a generator and recovers his lost time/sales or it makes no adjustment at all.
Table 19.2
Percentage of Sample units Adjusting through Various Mechanisms
(%)
Buying or
Operating
Existing
Generator
Working
Overtime
Changing
Shift
Timings
Changing
Working
Days
Punjab
64
5
3
0
Sindh
63
8
0
4
KPK
51
0
0
0
Balochistan
40
0
0
0
Total
60
5
2
1
Whole sale Retail Trade Establishments
52
4
1
2
Restaurant and Hotel
79
3
3
0
Transport and communication
65
0
0
0
Finance, Insurance, Real Estate & Business Services
63
13
8
0
Community, Social and personal Services
67
6
0
0
Total
60
5
2
1
By Province
By Sector
Table 19.3
Number of Adjustments by Firms with and without Generators
(%)
Numbers of Adjustments
None
Units without Generators
Units with Generators*
96.8
0.0
One
2.1
93.7
Two
1.1
3.5
Three or more
0.0
2.8
Total
100
100
*including the one adjustments of use of generator.
19.2 EXTENT OF LOSS OF SALES/ OUTPUT IN OUTAGES
Table 19.4 highlights the extent of the permanent loss of sales/output which is not recovered
through the various adjustments. Overall, it is over 4 percent nationally, 4 percent in Punjab, 2
percent in Sindh, 6 percent in K-PK and about 8 percent in Balochistan. These losses are a key
indicator of the magnitude of net idle factor costs.
105
The losses of wholesale,
retail
trade
and
community,
social
and
personal
services
establishments are 4-5 percent while losses of restaurants and hotels are the least, around 3
percent.
Table 19.4
Proportion of output Loss Not Recovered
(%)
By City
Location
Average
Punjab
Sindh
KPK
Balochistan
Total
4.1
2.1
6.2
7.7
4.3
By Sector
Industrial Group
Average
Whole sale Retail Trade Establishments
4.9
Restaurant and Hotel
3.1
Transport and communication
3.5
Finance, Insurance, Real Estate & Business Services
3.5
Community, Social and personal Services
Total
4.2
4.3
106
CHAPTER 20
OUTAGE COSTS
The objective of this chapter is to present the estimated magnitudes of different types of costs
associated with outages. As identified in chapter 17, these include direct costs which consist of
net idle factor costs and spoilage costs and indirect or adjustments costs which include
generator costs and costs of other types of adjustments like overtime, additional shifts, etc.
Section 1 of the chapter presents the total outage costs by location (province), and sector.
Section 2 derives the cost per Kwh of load shedding. Finally, by blowing up the sample, the
magnitude of outage costs to the commercial sector of Pakistan is derived.
20.1 TOTAL OUTAGE COSTS
Given the high frequency of outages, the outage costs are high in absolute terms even for small
shops and other commercial establishments. Table 20.1 shows that the outage costs per
sample unit are above Rs 100,000 on average.
Table 20.1
Total Outage Costs per Unit
(Rs in 000s)
Costs
Location
Punjab
Sindh
K-Pk
Balochistan
Sector
Wholesale & Retail Trade
Restaurants & Hotels
Transport & Communications
Finance and Insurance
Community, Social and
Personal services
Total
Share(%)
Sample
Size
Net Idle
Factor Cost
Spoilage
Cost
Generator
Costs
Total
Outage
Costs
129
55
37
20
76
24
84
84
10
3
7
8
36
25
39
13
122
52
130
105
128
35
17
25
36
67
57
155
41
47
10
9
3
2
24
50
51
21
44
101
116
209
62
93
241
66
(63)
7
(7)
32
(30)
105
(100)
Outage costs per unit are the highest in the cities of Punjab and K-PK at Rs 130,000 and
122,000 respectively. They are the lowest in Karachi and other cities of Sindh at Rs 52,000, due
primarily to substantially lower net idle factor costs.
107
Among sectors, the highest outage cost is observed in the case of transport and
communications at an average of Rs 209,000, followed by restaurants and hotels at Rs
116,000. Units operating in the financial sector have the lowest cost of Rs 62,000.
Overall, the dominant component in outage costs is idle factor cost with a share of 63%. Next in
importance are generator costs at 30%. Spoilage costs account for only 7 percent of the total
outage cost.
Chart 20.1
Outage Costs as Percentage of Value Added
By Location
14
Punjab
Sindh
K-Pk
Balochistan
12
10
8
11.5
6
4
5.4
6.2
3.5
2
0
Punjab
Sindh
K-Pk
Balochistan
By Sector
8
7
6
5
4
3
2
1
0
7.2
4.7
Wholesale & Retail
Trade
Restaurants &
Hotels
5.1
4.1
3
Transport &
Communications
Finance and
Insurance
Wholesale & Retail Trade
Restaurants & Hotels
Transport & Communications
Finance and Insurance
Community, Social
and Personal
services
Community, Social and Personal services
108
20.2 BURDEN OF OUTAGE COSTS
The burden of outage costs in relation to the value added is given in Table 20.2. Overall, for the
sample units, outage costs are 5.4 percent of the value added. Within the provinces, the highest
burden is in Balochistan, due particularly to a relatively value added per unit.
Within sectors, wholesale and retail trade establishments have the highest incidence of outage
costs on value added at over 7 percent. This is the lowest in units operating in the financial
sector.
Table 20.2
Outage Costs as Percentage of Value Added
(Rs in 000)
Sample Size
(No)
Total
Outage
Cost
Value
Added
Outage Cost as
% of Value
Added
Punjab
129
122
2252
5.4
Sindh
55
52
1503
3.5
K-Pk
37
130
2106
6.2
Balochistan
20
105
917
11.5
Wholesale & Retail Trade
128
101
1404
7.2
Restaurants & Hotels
35
116
2472
4.7
Transport & Communications
17
209
5062
4.1
Finance and Insurance
25
62
2062
3.0
Community, Social and Personal
services
Total
36
93
1824
5.1
241
105
1948
5.4
Location
Sector
20.3 OUTAGE COST PER KWH
Table 20.3 indicates that the outage cost per Kwh is approximately Rs 68(70 cents). This is 33
percent higher than the outage cost per Kwh to small-scale industry. As such, this is consistent
with findings of studies in other countries, as highlighted in Part I of the Report.
The outage cost per Kwh in commercial/service establishments is the highest in K-PK at Rs 97
($1). Among sectors, the highest cost per Kwh is observed in the case of units from transport
and communications at Rs 102 (105 cents). The lowest cost is incurred per Kwh in the case of
wholesale and retail trade establishments at Rs 61 (63 cents).
109
20.4 NATIONAL ESTIMATE OF OUTAGE COSTS
Data on value added in different service sectors is available in the Pakistan Economic Survey
for the latest year, 2011-12. As such, the ratio of the outage cost to the value added in each
sector is used for blowing-up to arrive at the national estimate of outage costs in the commercial
sector15 of the economy. These estimates are presented in Table 20.4.
Table 20.3
Outage Cost per Kwh
Location
Punjab
Sindh
K-PK
Balochistan
Sector
Wholesale & Retail Trade
Restaurants & Hotels
Transport & Comm
Finance & Insurance
Community, Social and Personal Services
Total
Sample
Size (No)
Total
Outage Cost
(Rs in 000)
Electricity
not
Provided
(000 Kwh)
Outage Cost
per Kwh
(Rs)
129
55
37
20
122
52
130
105
1.76
0.73
1.34
2.82
69.3
71.2
97.0
37.2
128
35
17
25
36
101
116
209
62
93
1.67
1.69
2.05
0.97
1.16
60.5
68.6
101.9
63.9
80.2
241
105
1.55
67.7
Table 20.4
Total Cost of Outages to the Commercial Sector
Sector
Wholesale & Retail Trade
Restaurants & Hotels
Transport & Communication
Finance & Insurance
Community, Social and Personal Services
Total
Outage Cost as
% of Value
Added (%)
Value Added
2011-12 (Rs in
Billion)
Outage Cost
(Rs in Billion)
7.2
4.7
4.1
3.0
5.1
3181
393
2477
460
2134
229
18
102
14
109
472
The largest cost of outages is in wholesale and retail trade, which is also the largest sector in
terms of value added. The other sectors which have high costs are community, social and
personal services and transport and communications. Overall, the total cost of outages in the
commercial/service sectors is estimated in 2011-12 at Rs 472 billion, equivalent to almost
2.4% of the GDP.
15
Public administration and defense has been excluded from the analysis because the value added in this sector
consists primarily of wages and salaries of employees. This is taken as the measure of value added. These are likely
to be unaffected by outages, especially in the absence of adjustments. Also, spoilage costs are minimal.
110
CHAPTER 21
LOAD MANAGEMENT STRATEGY: CONSUMER’S PREFERENCES
The questionnaire contains a module to
solicit consumer preferences regarding
timing of loadshedding which can reduce
the costs and disruptions due to the
outages. These can provide guidance to
the
load
management
strategy
by
DISCOs, the formulation of which should
be a priority since loadshedding is likely
to persist over the next few years.
21.1 LEVEL OF SATISFACTION WITH
CURRENT LEVEL OF SERVICE
Only 14 percent of sample firms
Table 21.1
Percentage of Time DISCOs Kept to the Announced
Loadshedding Schedule
By Province
Average
Punjab
Sindh
KPK
Balochistan
Total
20
5
11
4
14
By Sector
Whole sale Retail Trade Establishments
15
Restaurant and Hotel
12
Transport and communication
Finance, Insurance, Real Estate &
Business Services
Community, Social and personal Services
5
17
Total
14
14
indicated that DISCOs kept to the
announced loadshedding schedule (see
Table
21.1).
The
percentage
is
significantly higher for cities in Punjab.
Table 21.2
Average Time Required for Adjustment to Changes in
Loadshedding Schedule
(%)
Performance of DISCOs in Sindh and
Balochistan in this respect is particularly
weak,
with
effectively
no
By Province
Average
prior
scheduling. This has had a significant
Punjab
1
impact on the costs of loadshedding.
Sindh
0
KPK
1
The time required for establishments to
Balochistan
1
adjust to changes in the loadshedding
Total
1
schedule is 1 hour on an average. (See
By Sector
Table 21.2).
Whole sale Retail Trade Establishments
1
Restaurant and Hotel
1
Transport and communication
Finance, Insurance, Real Estate & Business
Services
Community, Social and personal Services
0
Total
1
The survey teams enquired from the
respondents if they were satisfied with
the current level of service by the
DISCOs/KESC. More than half of the
1
3
respondents ranked their satisfaction level as very low while about 30 percent ranked it as low
(see Table 21.3).
111
Table 21.3
Level of Satisfaction with Current quality of Service by DISCOs/KESC
(%)
Very high
High
Medium
Low
Very Low
By Province
Punjab
1
0
11
31
57
Sindh
2
6
25
33
33
KPK
3
0
6
22
69
Balochistan
0
0
0
20
80
Total
1
1
12
29
56
By Sector
Whole sale Retail Trade
Establishments
2
1
14
29
55
Restaurant and Hotel
0
0
3
27
70
Transport and communication
0
0
0
35
65
Finance, Insurance, Real
Estate & Business Services
0
4
21
17
58
Community, Social and
personal Services
3
3
17
39
39
Total
1
1
12
29
56
The sample firms were also asked how much higher tariff they are willing to pay for better
quality and reliability of power service–essentially with no loadshedding. This provides the first
estimate of the respondent’s perception of the cost of loadshedding. On an average,
respondents are willing to pay an extra 35 percent for uninterrupted power supply as revealed
by Table 21.4. The premium for better service is highest around 45 percent in K-PK.
Wholesale, retail trade and transport and communications units, it appears, have indicated their
willingness to pay the highest additional premium for better quality service.
Translated into the subjective valuation of the outage cost per hour, the average for the sample
units is Rs 38 per Kwh. The willingness to pay is highest, at Rs 73 per Kwh in Sindh (See Table
21.5).
112
Table 21.4
Additional Tariff For Better Quality of Service (with No Loadshedding)
(%)
By Province
Average
Punjab
34.3
Sindh
36.2
KPK
45.4
Balochistan
25.3
Total
35.7
By Sector
Whole sale Retail Trade Establishments
38.5
Restaurant and Hotel
33.1
Transport and communication
34.4
Finance, Insurance, Real Estate & Business Services
30.4
Community, Social and personal Services
32.2
Total
35.7
Table 21.5
Perceived Outage Costs per Kwh as implied by Willingness To Pay
Rs
Location
Average
Punjab
29.1
Sindh
73.0
KPK
32.6
Balochistan
12.2
Total
37.7
By Sector
Whole sale Retail Trade Establishments
38.4
Restaurant and Hotel
31.8
Transport and communication
31.1
Finance, Insurance, Real Estate & Business Services
35.6
Community, Social and personal Services
45.3
Total
37.7
113
21.2 PREFERRED CHANGES IN TIMINGS OF LOADSHEDDING
About 96 percent of the sample firms reported summertime as the worst season for
loadshedding (see Table 21.6). Winter time is the second worst season for loadshedding.
Interestingly, transport and communications is the dominant industry categorizing wintertime as
bad season.
Table 21.6
Worst Time of The Year for Loadshedding
Summer
Spring
By Province
Punjab
Sindh
KPK
Balochistan
Total
97
94
97
95
96
0
2
3
0
1
Rank
Winter
Fall
Total
3
4
0
0
3
0
0
0
5
0
100
100
100
100
100
By Industrial Group
Whole sale Retail Trade Establishments
Restaurant and Hotel
Transport and communication
Finance, Insurance, Real Estate &
Business Services
Community, Social and personal
Services
Total
98
94
88
2
0
0
0
6
12
0
0
0
100
100
100
96
0
0
4
100
94
0
6
0
100
96
1
3
0
100
The questionnaire also contained a question
regarding the worst day of the week for
outages. While 35 percent of the respondents
said all days are bad, about one-fourths said
Monday is the worst day. Sunday was the
worst day for about 15 percent of the
respondents. The principal reasons cited for
this is related to the fact that Monday is the
Table 21.7
The Worst Day of The Week for Outages
Frequency
Percentage
36
57
3
2
14
27
15
83
237
15.2
24.1
1.3
.8
5.9
11.4
6.3
35.0
100.0
Sunday
Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
All days equal
Total
start of a work week and an outage disturbs
the working environment. Sunday, of course,
is family/rest day and loadshedding disturbs it.
Table 21.8
Information that can be provided by
Distribution companies to Units
Around 20 percent of the respondents indicate
Percentage
that will be helpful if the power companies
Save electricity
48.9
provided more information relating to the
Information about outage
38.3
Time table for load shedding
23.4
Awareness about outage required
2.1
Others
17.0
methods to save electricity, information about
outages and the scheduling of the outage
(see Table 21.8). Clearly, these should be
focused upon in the load management strategy of the distribution companies.
114
CHAPTER 22
SUGGESTIONS BY THE SAMPLE UNITS
The questionnaire at the end solicited the respondent’s views/ suggestions to help handle the
loadshedding problem in the country. Specifically, the open ended question asked for
“suggestions to reduce the costs of loadshedding”. A number of interesting suggestions
emanate from the survey responses. These can be categorized as relating to the following:

Enhancing the supply of electricity

Alternative sources of energy/ fuel use

Improving governance or management

Changes in pricing policy
Enhancing the Supply of Electricity: About half of the respondents are of the view that new
dams, including Kala Bagh Dam, should be constructed to permanently enhance the supply of
electricity in the country at low costs (see Table 22.1). This suggestion dominates the response
not only from the sample units located in Punjab, but is also significant in the case of Peshawar
and Quetta. One quarter of the respondents also think that new power plants should be built
while close to a fifth of respondents each are also of the view that electricity should be imported
and gas pipeline from Iran should be installed to avoid gas shortages. Responses are more or
less, similar across type of commercial activities. (see Table 22.2)
Alternative sources of Energy Fuel for Energy: A number of suggestions have been given
regarding resort to alternative energy and fuel sources by the respondents. Over one-fifths of
the respondents suggested the use of different methods of electricity generation, while 20
percent specifically suggested the use of coal for electricity generation. Close to 7 percent of the
sample units suggested introduction of solar energy systems (particularly by restaurants and
hotels, finance, insurance, real estate and business services and transport and communications
establishments).
Improving Governance/Management. The most dominant recommendation in this category is
to minimization of electricity theft, with over a quarter of respondents emphasizing it. Curbing of
corruption has been suggested by 17 percent of the respondents. Need for honest employees,
minimization of line losses and awareness creation for proper use of electricity along with
privatization of the DISCOs were also cited as possible mechanisms to lower loadshedding
costs.
115
Lahore
Enhancing Supply of Electricity
Gas Pipe line from Iran to avoid gas shortage
Import Electricity
Construct new Dams (including Kala-Bagh Dam)
Use rental power system
Build new power plants
Alternative Energy Fuel Sources
Use Coal for electric generation
Table 22.1
Suggestions by Sample Units by City
(% of Respondents)
Faisal Gujranw Mult Sialk Rawal Karachi
abad ala
an
ot
pindi /
Islama
bad
Hyderab
ad
Peshaw
ar
Mard
an
Abbotab
ad
Quet
ta
Total
14
14
63
7
40
14
95
67
0
10
22
11
56
0
67
0
14
43
0
50
0
0
40
0
0
27
3
62
0
24
30
23
21
12
19
13
0
0
0
0
8
8
64
0
24
67
50
50
0
0
17
0
67
0
0
15
15
45
0
5
19
20
50
3
24
Use different method of electric generation
Bio Gas system
Introduce solar energy System
12
37
0
12
29
10
0
0
44
78
0
0
7
0
0
14
0
60
0
0
8
14
0
5
44
9
2
2
13
25
0
13
16
24
0
20
0
0
0
17
0
17
0
0
20
30
0
0
20
22
0
7
Introduce wind energy
0
0
0
0
0
0
0
0
0
0
17
0
0
Governance/Management
Privatize Electric department
0
0
0
0
0
0
2
0
4
0
0
0
1
Need Honest Employees
0
0
0
0
0
16
5
0
0
0
0
0
3
Minimize electric theft
28
14
11
14
20
24
35
38
16
0
50
40
26
Stop Corruption
19
19
0
14
0
16
7
63
28
0
17
20
17
Minimize line losses
0
0
0
0
0
0
0
0
0
0
3
0
0
0
63
38
0
0
0
0
0
17
0
0
3
2
26
21
18
19
10
9
11
0
4
14
0
6
0
60
2
5
46
16
23
30
18
0
13
3
4
4
11
0
0
3
0
17
3
40
5
8
16
20
Give awareness to people use of electricity
Pricing Policy
Government give subsidy on electricity
Reduce price at source
Total
116
Table 22.2
Suggestions by Sample Units by Sector
(% of Respondents)
Whole
sale
Retail
Trade
Establis
hments
Restaur
ant and
Hotel
Transport
and
communic
ation
Finance,
Insurance,
Real Estate
& Business
Services
Community,
Social and
personal
Services
21
18
18
13
17
19
19
18
29
17
25
20
44
59
41
71
50
50
Use rental power system
2
0
0
4
11
3
Build new power plants
24
21
29
21
25
24
21
24
18
17
17
20
24
29
6
8
25
22
1
5
1
0
18
0
0
12
0
0
13
0
0
0
0
0
7
0
0
3
29
16
3
6
6
18
18
0
0
0
12
29
6
0
4
33
33
4
0
3
22
3
0
1
3
26
17
3
2
3
0
0
3
2
17
6
12
8
31
16
21
53
9
14
12
7
17
10
33
15
20
Enhancing Supply of Electricity
Gas Pipe line from Iran to avoid
gas shortage
Import Electricity
Construct new Dams (including
Kala Bagh Dam)
Total
Alternative Energy Fuel Sources
Use Coal for electric generation
Use different method of electric
generation
Bio Gas system
Introduce solar energy System
Introduce wind energy
Governance/Management
Privatize Electric department
Need Honest Employees
Minimize electric theft
Stop Corruption
Minimize line losses
Give awareness to people use of
electricity
Pricing Policy
Government give subsidy on
electricity
Reduce price at source
Total
Pricing Policy Around one-fifth of the sample units suggested that the price (at source) should
be reduced through economizing on costs while 16 percent requested for subsidy for electricity
from the government.
To conclude, the top five suggestions emanating from the respondents of the survey are as
following:
First: Construct Dams
Second: Minimize Electricity theft.
Third: Build New Power Plants
Fourth: Use Different Methods of Electricity Generation
Fifth: Import Electricity and Reduce Prices at Source (Cost Minimization)
117
CHAPTER 23
CONCLUSIONS AND POLICY IMPLICATIONS
We have highlighted in previous Chapters the principal findings on the incidence of outages in
commercial/services sectors of the economy. In this concluding Chapter we derive the key
policy implications, starting with estimates of the multi-dimensional impact of power
loadshedding on service establishments in the country.
23.1 IMPACT OF OUTAGES
The estimated impact of outages is as follows:
(i)
Outages, on the average, occur for 19 percent of the time available during operating
hours. The proportion of business/sales/output lost permanently is 4 percent. This
implies an over 15 percent fall in profitability.
(ii)
The outage cost per Kwh works out at Rs 68 per Kwh (70cents). This is 33 percent
higher than the corresponding cost to small-scale industry.
(iii)
The employment in services, according to the Labor Force Survey of 2010-11 of
PBS, is 17.6 million. With a lower output of 4 percent and an employment elasticity of
0.6, the employment level would have been higher in the absence of outages by
about 422,000.
(iv)
Within services, the cost of outages appears to be the highest in Punjab and K-PK
and in establishments in transport and communications and wholesale and retail
trade.
23.2 AFFORDABILITY OF HIGHER TARIFFS
The total costs of electricity consumption, that is, the costs of public supply and of outages
costs, as a percentage of the value of production are given in Table 23.1.
On average, these costs aggregate to 6 percent of the value of production, with the highest
percentage in Balochistan (due primarily to low value of production/sales) and in the case of
wholesale and retail trade establishments. The corresponding percentage in the case of smallscale industrial sector is 12 percent. Therefore, the affordability of higher tariffs is somewhat
higher in services. But the likelihood of organized large-scale protests by this sector is greater in
response to a hike in tariffs, in the presence of poor quality of service by DISCOs/KESC.
118
Table 23.1
Total Costs of Electricity Consumption as a percentage of the Value of Production
Value of
Production/Sales*
Electricity Cost
of public
supply
Location
Punjab
220
Sindh
161
K-PK
145
Balochistan
165
Sector
Wholesale & Retail
171
Trade
Restaurants & Hotels
334
Transport & Comm
253
Finance & Insurance
121
Community, Social and
135
Personal Services
Total
190
*Assumed at 2.5 times of value added
(Rs in 000s)
Electricity Cost as
% of the Value of
Production/Sales
Total
outage
cost
Total
Cost
122
52
130
105
342
213
275
270
5630
3757
5265
2293
6.1
5.7
5.2
11.8
101
272
3510
7.7
116
209
62
93
450
462
183
228
6180
12655
5155
4560
7.3
3.7
3.5
5.0
105
295
4870
6.0
23.3 POLICY IMPLICATIONS FOR LOAD MANAGEMENT
(1) 60 percent of the Sample Units have acquired generators, mostly of small capacity (<5KVA).
There is need to extend the sales tax exemption to small generators, beyond the elimination
of import duty, in view of the high benefit-cost ratio of investment in generators of over 2.4:1.
(2) On the average, units are willing to pay 36 percent higher tariffs for reliable power supply
(with minimal outages). The highest willingness to pay is in Sindh, followed by Punjab. The
subjective valuation, of outage costs per Kwh on average is Rs 37.7 Kwh.
Therefore, the subjective valuation is 55 percent of the actual estimated outage cost per
Kwh. There is need to undertake cost-benefit analysis of improving reliability of supply. This
likely to be very high at over 5:1.
(3) The sample units report that the DISCOs/KESC adhered to the announced loadshedding
schedule only 14 percent of the time. The performance appears to be the worst of KESC
and HESCO. On the average, commercial sector establishments require one hour to adjust
to a change in the schedule of outages. Clearly, much more discipline has to be exercised
by DISCOs/KESC in adherence to pre-announced schedule of outages.
119
(4) The worst season for outages is summer, as reported by 96 percent of the units. The two
days of the week which are considered the worst for outages are Monday and Friday, as
reported by 68% and 17 percent of the units respectively. Seasonal and day-to-day variation
in tariffs may be considered.
(5) The level of satisfaction with the service provided is ‘very low’ in the case of 56 percent of
the units and ‘low’ according to 29 percent of the units. This highlights the virtually total loss
of confidence of consumers with the power distribution system.
(6) The principal suggestions by responding units for reducing the incidence of loadshedding
can be classified in four categories as follows:
I.
II.
III.
IV.
Better Management of the Power Sector
Units*
 Minimize electricity theft
 Stop corruption
 Honest employees be hired
 Minimize line losses
 Give awareness to people for efficient use
Expand Capacity
 Construct new dams (including Dam)
 Build new power plants
 Import electricity
 Gas pipeline from Iran
 Use rental power
Develop Alternative Sources
 Use coal for electricity generation
 Use different method for electricity generation
 Introduce solar energy
 Use biogas
Pricing Policy
 Reduce price at source
 Give subsidy
% of
18
16
3
1
2
29**
19
15
15
2
19
16
3
1
9
9
*The total adds up to more than 100 due to multiple responses by units.
**36% of the responses were from outside Punjab.
It is interesting to note the relatively higher incidence of suggestions for improving the
management of the power sector and the development of alternative sources of energy. Also,
the highest response relates to the construction of new dams.
120
PART IV
COST OF LOADSHEDDING TO
DOMESTIC/RESIDENTIAL
SECTOR
121
CHAPTER 24
INTRODUCTION
This part of the report presents the findings on costs of loadshedding to the domestic/residential
sector in Pakistan, quantified on the basis of data obtained from a nationwide survey of
households.
The report is organized in nine chapters. Chapter 25 presents the methodology used for
qualification of costs due to outages. Chapter 26 describes the survey including the sampling
methodology. Subsequent Chapters up to Chapter 30 present the magnitudes of key
parameters like the relevant characteristics of the responding households, experience of
outages, level and pattern of adjustments and the magnitude of different outages costs. Chapter
31 highlights the suggestions by sample households for reduction in incidence and costs of
outages.
Chapter 32 gives a summary of the principal findings and the major policy implications emerging
from the research. It is clear from the results that households have faced severe disruptions due
to the high and growing incidence of loadshedding. These have led to mass protests on streets
resulting in disruption of other economic activities including those of the commercial/industrial
sectors. As such, the economic return of reducing outages and of facilitating the process of
adjustment to these outages is high.
Thanks are due to the sample households for responding to a questionnaire, which was
complex and difficult to administer. Thanks are also due to the survey team which travelled all
over the country and sometimes found itself in a difficult law and order situation, especially in
Karachi.
The main text of the report is presented in a non-technical manner. Technical analyses are
included in the Technical Annexes.
Any defects which remain are of course, the responsibility of the authors.
122
CHAPTER 25
METHODOLOGY FOR QUANTIFICATION OF COSTS
The approach to quantification of outage costs in the case of domestic consumers used by
various studies referred to in part I of the Report is one of the following:
25.1 APPROACHES TO QUANTIFICATION OF COSTS TO DOMESTIC CONSUMERS
i.
Value of Leisure: This approach uses the income per Kwh as a measure of the outage
cost on the assumption that the value of leisure corresponds to income.
ii.
Standby Generator Cost: The adjustment to outages by domestic consumers is
assumed to be primarily in the form of self-generation. As such, the cost of using the
generator is used as the measure of outage cost.
iii.
Willingness to Pay: In this approach consumers are asked the magnitude of higher
tariff that they are willing to pay for reliable public supply (with minimal outages).
Only one study by Balducci [2002] in USA has adopted the survey approach to quantification of
the outage cost to consumers.
We review each of the above approaches to quantification of outage costs to domestic
consumers in light of the data obtained from the survey of 500 households.
VALUE OF LEISURE APPROACH:
Munasinghe [1980]16 had first
suggested this approach based on
the observation that outages during
the day normally do not affect the
performance of household activities
like cooking, washing, laundry, etc,.
which can be performed at other
times when there are no outages.
As such, he argued that the only
activity
which
is
affected
is
watching TV and other forms of
leisure in the evening. Therefore,
the outage cost corresponds to the
Table 25.1
Outage Cost per Kwh
according to the Value of Leisure Approach*
Group
Income**
Electricity
Outage
(Rs
per per hour Consumption per
Cost per
Month)
hour***(Kwh)
Kwh (Rs)
0-15000
67.5
0.9
75
15001-35000
144.8
1.5
97
35001-70000
295.5
3.3
90
Above 70000
612.6
5.7
107
Total
218.3
2.4
91
*Y = income per hour worked based on 8 hours a day for 22
days a month.
Kwh = normal power consumption per hour (in public supply)
**Proxied by consumption expenditure, which is assumed to
correspond to permanent income
***On the assumption that electricity is consumed 16 hours a
day. The consumption of electricity in the evenings is assumed
to be three times the daily average.
16
“Costs Incurred by Residential Electricity Consumers Due to Power Failures”, Mohan Munasinghe, Journal of
Consumer Research, Vol. 6, March 1980.
123
value of leisure, which he proxies by income.
The estimated outage cost per Kwh for domestic consumers based on this approach is derived
from the Survey as Rs 91 per Kwh in Table 25.1. This is higher than estimates obtained of
outage cost per Kwh for small-scale industry and commercial consumers. Therefore, the
Munasinghe approach yields very high estimates.
There is another way of examining the validity of assumptions made by Munasinghe.
Respondents were asked which activities are disrupted most in the household by loadshedding.
The frequency of different responses is given in Table 25.2.
Leisure is reported by only 2
percent
of
the
Table 25.2
Activities most Disturbed by Loadshedding
sample
households as the activity
most
disturbed
by
loadshedding.
activities
are
Other
of
greater
importance to households,
including
studies
cooling/heating,
of
children
and
Cooling/heating
Studies (home work) of children
Preparation for work/school
Regular household work (cooking, cleaning, etc.)
Shortage of water
Income generating activities (home based)
Social Activities
Entertainment, leisure
Total
% of sample units
24.4
18.2
17.4
14.6
13.0
8.2
2.2
2.0
100.0
preparation for work/school
reported 24 percent, 18 percent and 17 percent respectively as the principal activity affected by
outages. Therefore, the Munasinghe hypothesis that leisure is the activity most disrupted is not
borne out by the data obtained from households in Pakistan.
It is our view that the Munasinghe approach has a developed country bias. It cannot be applied
in the context of low-to-middle income countries like Pakistan. A significant and new finding is
the impact of outages on children, either in terms of the ability to undertake studies
(homework) or in preparation to go to school.
GENERATOR COST APPROACH
This approach is based on the assumption that the principal form of adjustment to outages by
households is the acquisition of a generator and/or a UPS (Uninterrupted Power Supply). As
such, the cost of self-generation corresponds to the outage cost.
The question that arises is if a household does not have a generator/UPS then is the outage
cost zero? Clearly, this is not the case.
124
It is likely that there are outage costs, especially in terms of the monetized value of the utility lost
due to disturbance to some household activities, but these costs may not be large enough to
justify the resort to self-generation.
Table 25.3 gives the percentage of households by level of consumption expenditure with
generator and/or UPS. Overall, 28 percent of the households have a generator and 30 percent
have
UPS.
Poorer
households generally are
Table 25.3
Percentage of Sample Households with Generator and/or UPS
% of Sample Households
unable to self-generate
electricity.
majority
However,
of
the
households in the upper
most income group have
Level of monthly consumption
expenditure (Rs)
0-15,000
15,001-35,000
35,001-70,000
70,001 and above
Total
With Generator
With UPS
2
17
45
75
28
4
26
47
43
30
made arrangements for
alternative sources of power at the time of loadshedding.
Given the high percentage of households which do not have self-generation the issue is one of
quantifying the cost of outages in the case of such households.
WILLINGNESS TO PAY
The willingness to pay approach provides the basis for determining the subjective valuation by
households of the cost of outages to them. There is, of course, the likelihood of a ‘free rider’
problem here. A household may understate its willingness to pay on the expectation that other
households may reveal a high enough WTP to justify investment in improving the reliability of
the power system.
Table 25.4 indicates the outage cost per hour as implied by the WTP. This can be estimated as
follows:
𝑊𝑇𝑃 𝐴𝐸𝐵
SOCKW = ( 100 ) 𝐸𝑁𝑆
……………………………………………………………………………… (1)
Where,
SOCKW = subjective valuation by household of the outage cost per Kwh
WTP = % higher tariff that the household is willing to pay for improved reliability of
power supply (with minimal outages)
AEB = Annual electricity bill paid to the DISCO/KESC
ENS = electricity not supplied in the outages.
125
Table 25.4
Subjective Valuation of the Outage Cost per Hour
Monthly
Expenditure Group
Willingness
to Pay
Annual
Electricity
Bill
Electricity not Supplied
(Rs)
(%)
(Rs)
(Kwh)
Subjective
Valuation by
Household of
Outage Cost per
Hour
(Rs per Kwh)
0-15000
15001-35000
35001-70000
70001 and above
Total
30.3
28.7
28.3
31.8
29.2
15330
28836
65094
130590
46734
479
732
1599
4299
1289
9.70
11.31
11.52
9.66
10.59
It is interesting to note that while the subjective valuation of the outage cost per hour is
somewhat low at below Rs 11 per Kwh, it is higher for households belonging to the ‘middle
class’.
25.2 METHODOLOGY FOR QUANTIFICATION OF OUTAGE COST
The methodology for quantification of outage cost to domestic consumers is qualitatively
different from that used in the case of small-scale industry and commercial consumers. The
basic reason for this is that there is no notion of ‘output’ in the case of a household17, which is
more of a consuming unit. As such, outages impact on the level of utility/quality of life of a
household.
The exposure to outages daily is given by DLOUT where
𝐷 = ∑𝑛𝑖=1 𝑛𝑖 𝑑𝑖
………………………………………..
(1)
Where 𝑛𝑖 = number of outages of duration 𝑑𝑖 , i = 1, …..n.
The normal level of electricity consumption per hour is given
e=
(𝐾𝑤ℎ1 + 𝐾𝑤ℎ2) …………………………………………………
(2)
8760−365𝐷
Where, Kwh1 = electricity purchased from the distribution company during summer months
Kwh2 = electricity purchased from the distribution company during winter months
The normal consumption of electricity during times when there are no outages depends upon
the number of electrical appliances at home. As such,
17
With the exception of households which engage in some economic activity at home.
*The 𝛽𝑗 is estimated by OLS regression across the sample households of electricity consumption per hour
with ownership of different types of appliances.
126
𝑒 = 𝛽𝑜 + ∑𝑚
𝑗=1 𝛽𝑗 𝐴𝑗
……………………………………………….
(3)
Where, 𝛽𝑗 *= electricity consumption by appliance j, where j =1,2,3,……..,m.
𝐴𝑗 = number of appliance j
𝛽𝑜 = basic electricity consumption (e.g. for lighting).
Depending upon the nature of use of particular appliances the share of electricity consumed in
different activities like heating/cooling, household functions, entertainment/leisure is derived.
That is
∑𝑟𝑘=1 𝑊𝑘 =1
……………………………………………….
(4)
Where 𝑊𝑘 = share in electricity consumption of activity k, k=1,2,……,r.
If a household has a generator then the sample household has reported if a particular activity
can be performed during the outages in the presence of a generator, and
𝑃𝑘1 = 1 if activity k can be performed during the outage.
𝑃𝑘1 = 0 if activity k cannot be performed during the outage.
Then the extent of substitution, S, by the generator of public supply during outages is given by
S1 where
𝑆1 = ∑𝑟𝑘=1 𝑊𝑘 𝑃𝑘1
………….. (5)
Similarly, the extent of substitution by a household which has a UPS can be derived
𝑆2 = ∑𝑟𝑘=1 𝑊𝑘 𝑃𝑘2
………….. (6)
It may, of course, be noted that in the case of household which has neither a generator nor an
UPS, S1=0, S2=0.
For a household which has a generator the costs of operation have been obtained as
𝐺𝑐 = 𝐾(𝑖 + 𝛿) + 12𝑓 + 4(𝑚 + 𝑜) − 𝑇 ………. (7)
Where, K = capital cost, I = annual interest rate, 𝛿 = annual rate of depreciation, f = monthly fuel
cost, m = quarterly maintenance costs, o = quarterly other costs, T = savings in terms of
payment to the utility.
Similarly, the cost of a UPS can be derived as Gu . In this case T = 0 because the UPS stores
electricity obtained at the time when there are no outages.
127
There are also other costs arising from the outages, including spoilage cost, SPC, damage to
appliances, DAC and miscellaneous costs, MC.
The last part of the methodology relates to the valuation of costs arising from disturbance of
activities which cannot be performed or only partially performed during the outages either
because of the absence of self-generation or because of only partial substitution by
generator/UPS.
These costs are subjective in nature in terms of a loss of utility and are, therefore, not observed.
We use the willingness-to-pay (WTP) as a measure of the subjective costs and apply this
magnitude to the part of the electricity consumption which is not substituted by self-generation
during outages. As such,
𝑀𝑈𝑇𝐿 = 𝑊𝑇𝑃(𝐵1 + 𝐵2 )(1 − 𝑆1 − 𝑆2 )
…………….(8)
Where, WTP = extent of higher tariff that household is willing to pay for better quality of service
(with minimal outages)
B1 = electricity bill of the distribution company during summer months
B2 = electricity bill of the distribution company during winter months
The overall outage costs to the household, OTC, is given by
𝑂𝑇𝐶 = 𝐺𝑐 + 𝐺𝑢 + 𝑆𝑃𝐶 + 𝐷𝐴𝐶 + 𝑀𝐶 + 𝑀𝑈𝑇𝐿
……………….(9)
In the case of a household with no self-generation capacity
𝑂𝑇𝐶 = 𝑆𝑃𝐶 + 𝐷𝐴𝐶 + 𝑀𝐶 + 𝑀𝑈𝑇𝐿
Where, 𝑀𝑈𝑇𝐿 = 𝑊𝑇𝑃(𝐵1 + 𝐵2 )
This methodology is new and has not been used yet in other studies
128
CHAPTER 26
THE SAMPLING FRAMEWORK AND ITS DISTRIBUTION
26.1 SAMPLING FRAMEWORK
The primary instrument of data collection was a survey on a pre-designed and tested
questionnaire of a stratified (by province and by city) national random sample of households
(see Figure 26.1).
Figure 26.1
Sampling Strategy
Primary Sampling Unit
Population in Provinces
Secondary Sampling Unit
Population in Cities in Provinces
Tertiary Sampling Unit
Localities by Income Groups in Cities
Individual Household in Different Income
Localities
The provincial population was obtained from the Census
Report while city-wise population was obtained from the
Development Statistics of the provinces, published by the
Provincial Governments. The national distribution of
population by province is presented in Figure 26.1 and city
wise population in Table 26.1. The derived sample
distribution by city is presented in Table 26.2.
Once the sample distribution across cities was finalized,
upper, medium and lower income residential localities were
selected within each city for survey. Individual household
within a locality was selected through random walk
procedure.
The
questionnaire
administered
on
the
sample
Table 26.1.
National Distribution of
Population in the Census, 1998
by City
Cities
Lahore
Faisalabad
Gujranwala
Multan
Sialkot
Rawalpindi
Islamabad
Karachi
Hyderabad
Sukkur
Peshawar
Mardan
Abbotabad
Quetta
Percentage
4.8
4.1
2.6
2.4
2.1
2.5
0.6
0.7
2.2
0.7
1.2
1.5
1.1
0.6
respondents contains five modules: basic information on households; experience of
129
loadshedding; adjustment to outages; costs of outages/ brown outages (voltage fluctuations);
and, preferred load management practices. Though the questionnaire was structured, the last
question was open-ended asking the respondents to make suggestions to reduce the costs of
loadshedding. This provides the respondent’s perspective on actions to counter the problem.
Figure 26.2
National Distribution of Population by Province
BALOCHISTAN
5%
ISLAMABAD
1%
KPK
14%
SINDH
23%
PUNJAB
57%
Table 26.2.
Distribution of Sample by Province and by City
Provinces
Punjab
Sindh
KPK
Balochistan
Total
Cities
Lahore
Faisalabad
Sialkot
Gujranwala
Multan
Rawalpindi/Islamabad
Total
Karachi
Hyderabad
Sukkur
Total
Peshawar
Mardan
Abbotabad/Bannu
Total
Quetta
Total
Numbers
96
51
13
26
38
61
285
80
20
10
110
50
13
12
75
30
30
500
Percentage
19
10
3
5
8
12
57
16
4
2
22
10
3
2
15
6
6
100
130
The survey was successfully administered on 500 households as targeted. 57 percent of the
sample household units are in the province of Punjab while about 22 percent are in Sindh. From
the remaining 33 percent, 15 percent are in Khyber-Pakhtunkhwa (K-PK) and 6 percent in
Balochistan.
26.2 CHARACTERISTICS OF SELECTED HOUSEHOLDS
Over 30 percent of the sample head of households were in
Table 26.3.
Occupation of the Head of the
Households
Percentage
Occupation
business while 18 percent were employed in the private
jobs. 11 percent were in government jobs (See Table 26.3).
The average family size of the sample households is 7
Business
Private job
Government job
Teacher
Retired person
Engineer
Driver
Others
Total
persons, being the highest in Balochistan (See Table 26.4).
The average number of children in the household is 2.
Also, other than the lowest income group, there is an
average of two earners per households. 18 percent of
sample households had a member working from home.
The
distribution
30.6
18.4
11.4
5.6
4.8
4.2
3.8
21.2
100
of
sample households by
Table 26.4.
Average Number of Family Members by Income Group, 2012
income group is given in
percent
of
the
households
have
permanent
income,
monthly
monthly
proxied
by
consumption
expenditure, of upto Rs.
15000, 36 percent have
income
between
Rs
15000 to Rs. 35000, 35
percent
have
income
Average
Number of
Adults
Average
Number of
Children
Number of
Earning
Members in
Household
6
6
9
10
7
4
5
6
5
5
2
2
3
5
2
2
2
2
3
2
6
4
2
1
2
7
5
3
7
7
7
5
5
5
2
2
2
Average
Family
Size
Figure 26.3. About 21
By Province
Punjab
Sindh
KPK
Balochistan
Total
By Income Group
Upto 15000
1500135000
3500170000
70001 +
Total
2
2
2
between Rs. 35000 to
Rs.70000 while 8 percent have income above 70000 per month. The overall average monthly
income of sample households is Rs. 38429.
131
Figure 26.3
Distribution of Selected Households by Income Group
70001 +
8%
Upto 15000
21%
35001-70000
35%
15001-35000
36%
Almost 86 percent of the sample households owned their home, with 4 rooms on an average.
Lower income households lived in 2 room houses.
The profile of ownership of assets is given in Table 26.5. Since these consumer durables
operate on electricity, the demand for electricity in the household depends on the ownership of
such assets, some durables being more electricity—intensive than others. 72 and 61 percent of
lower income households own televisions and washing machines, which indicates that they also
have a significant demand of electricity. However the more electricity consuming appliances is
owned by upper-middle and upper income households. Multiple ownership of ACs, TVs, DVDs,
fridges and heaters emerges from the survey.
Consequently, sample units, on an average, are spending almost Rs. 7800 a month on
electricity (see Table 26.6). This is equivalent to 20 percent of their monthly expenditure. The
average monthly expenditure on electricity for low income families is Rs. 2500 per month
increasing to Rs. 21000 for the upper income households. The highest burden of the electricity
bill appears to be on the lowest income group at 21.5 percent of monthly expenditure, declining
somewhat to 20.2 percent for the upper income households.
132
Table 26.5
Profile of ownership of Assets
(%)
CAR
TV
Air
Condit
ioner
Micro
wave/
Oven
DVD
players
Radio
Fan
Fridge
Deep
Freezer
Electric
Heater
Washing
machine
Internet
Computer
Punjab
34
86
36
44
30
27
100
76
21
34
79
43
54
Sindh
35
93
26
40
30
15
100
94
29
5
93
45
58
KPK
43
89
48
44
24
21
100
95
29
20
96
49
57
Balcohistan
50
97
23
57
50
20
100
97
23
37
100
60
80
By Province
By Income Group
Upto 15000
2
72
1
5
12
14
100
58
5
7
61
8
18
15001-35000
16
89
8
24
18
24
100
88
9
23
89
31
47
35001-70000
67
97
71
78
44
26
100
94
44
34
95
75
85
70001 +
88
98
93
88
73
30
100
93
60
50
95
80
88
133
Table 26.6
Average Monthly Expenditure, Electricity Bill and Electricity Bill as % Monthly
Expenditure Electricity of Sample Units
Average Monthly
Expenditure
Electricity
Consumed
(Rs)
Electricity bill as monthly
expenditure (%)
Punjab
41051
7319
17.8
Sindh
33482
8371
25.0
KPK
36973
7333
19.8
Balochistan
35300
11247
31.9
Total
38429
7788
20.3
Upto 15000
11882
2555
21.5
15001-35000
25489
4806
18.9
35001-70000
52034
10849
20.9
70001 +
107825
21765
20.2
Total
38429
7788
20.3
By Province
By Income Group
134
CHAPTER 27
THE EXPERIENCE OF LOADSHEDDING
This Chapter discusses the incidence of loadshedding and the disruptions leading to costs and
to utility losses of households.
27.1 INCIDENCE AND PROFILE OF LOADSHEDDING
The costs of loadshedding, to a large extent,
depend on the frequency and duration of
outages. The incidence of loadshedding is
given in Table 27.1. Overall, on an average
outages occurred 5 times a day in Pakistan in
2012, highest being in Punjab, 6 times.
Households, on an average did not have
electricity
supply
from
power
distribution
companies for 1453 hours in 2012. The
highest loadshedding has occurred in Punjab
at 1683, followed by K-PK, 1216. Clearly, the
Table 27.1
Average number of times there is
loadshedding in a day
By Province
Location
Punjab
Sindh
KPK
Baluchistan
Total
By Income Group
Upto 15000
15001-35000
35001-70000
70001 +
Total
Average
6
3
4
4
5
5
4
5
5
5
average incidence is lower in Sindh and
Balochistan.
The distribution of outages by duration is given
in Table 27.3. The highest number of outages
occurs for 1 to 2 hours a day (70 percent),
followed by outages of one-half to an hour a
day (25 percent). 3 percent of outages each
have duration of half to one hour and for over
two hours. There is some divergence in the
provincial patterns. In Punjab, 64 percent of
Table 27.2
Hours of Outages
By Province
Location
Punjab
Sindh
KPK
Balochistan
Total
By Income Group
Upto 15000
15001-35000
35001-70000
70001 +
Total
Average
1683
1123
1216
1069
1453
1498
1394
1430
1702
1453
the outages last for 1-2 hours while in Sindh
almost 85 percent of the outages were for that duration. In Balochistan 8 percent of the outages
are of over 2 hours while this duration of outage was not reported in the survey in Sindh.
135
Table 27.3
Percentage Distribution of Average Length of Outages, 2012
(%)
By Province
Less than ½ hr
1/2 – 1hr
1-2 hrs
More than 2 hrs
Punjab
3
29
64
3
Sindh
2
12
85
0
KPK
0
20
78
2
Balochistan
0
22
68
8
Total
3
25
70
3
Upto 15000
3
18
76
3
15001-35000
1
25
73
2
35001-70000
3
28
66
3
70001 +
8
26
63
3
Total
3
25
70
3
Location
By Income Group
The pattern appears to
vary
across
Table 27.4
Timing of Loadshedding
different
income localities. Over
one-third
of
income
households
experienced outages of
up to 1 hour while this
proportion
for
lower
income households is
one-fifths. As compared
to this, 69 percent of
upper
(%)
upper
income
households experienced
outages
exceeding
hour
while
1
6 am - 12
noon
12 noon 6 pm
6 pm midnights
Midnights
- 6 am
Total
By Province
Punjab
41
Sindh
35
KPK
36
Balochistan
30
Total
39
By Income Group
28
31
27
28
28
12
28
23
14
17
19
6
15
28
16
100
100
100
100
100
Upto 15000
39
26
17
19
100
37
29
18
17
100
41
38
39
29
30
28
15
17
17
15
15
16
100
100
100
1500135000
3500170000
70001 +
Total
this
percentage for lower income households is 80.
136
Half
of
the
respondents
Table 27.5
Experience of Voltage Fluctuations
indicate that the pattern of
(%)
loadshedding typically varies
Power Fluctuations
on a daily basis while the
other half does not. Currently,
39 percent of the households
experienced loadshedding in
morning
hours
while
percent
experience
28
it
in
afternoons (See Table 27.4).
Loadshedding in evenings and
nights were experienced by
one-third
of
the
Frequent Fluctuations
Yes
No
Total
Yes
No
Total
By Province
Punjab
Sindh
KPK
Balochistan
By Income Group
68
41
57
93
32
59
43
7
100
100
100
100
41
76
58
71
59
24
42
29
100
100
100
100
Upto 15000
15001-35000
35001-70000
70001 +
Total
63
68
59
48
62
37
32
41
53
38
100
100
100
100
100
72
54
38
32
51
28
46
62
68
49
100
100
100
100
100
sample
Table 27.6.
Disruptions Due to Loadshedding
households.
(%)
The respondents were asked
if they experienced brown
Very
high
High
Medium
Low
Very
Low
outages (voltage fluctuations)
By Province
and
Punjab
Sindh
KPK
53
46
64
17
32
17
23
16
15
2
4
1
5
2
3
Balochistan
By Income group
63
10
0
23
3
Upto 15000
15001-35000
35001-70000
70001 +
60
52
52
50
54
18
21
21
18
20
14
17
23
20
19
6
4
2
5
4
2
5
3
8
4
whether
frequent.
these
62
respondents
were
percent
indicated
of
that
they have brown outs while 51
percent reported
them
to
be
Table
27.5).
frequent(See
Some
inter-
Total
provincial differences also emerge from the survey. Brown outs are more of a phenomenon in
Balochistan than in Sindh, However, those who experience voltage fluctuations in Sindh, have it
frequently.
27.2 EXTENT OF DISRUPTION DUE TO OUTAGES
To understand the nature of loadshedding cost on households, the respondents were asked
how disruptive loadshedding was. Table 27.6 shows that three-fourth of the sample households
think that outages are highly disruptive. The disruptions are higher for the lower and middle
income households as they are unable to make adjustments to reduce the costs of
137
loadshedding. 78 percent of lower income, as compared to 67 percent of upper income
households rated loadshedding highly disruptive.
The nature of disruption, already identified in chapter 2, are elaborated in Table 27.7. While
children’s school preparations/home work are ranked as important disruptions in Punjab, lack of
cooling/heating is by far the most important disruption in Sindh according to the survey.
Shortage of water and children’s studies emerge as important disruptions in K-PK while lack of
cooling/heating bothers the sample households of Balochistan the most.
The importance of disruptions also varies somewhat across income groups. Top three
disruptions of loadshedding for different income groups are: for lower income groups resultant
shortage of water( due to inability to pump water) ,no cooling and children’s studies; for lower
middle income group no cooling/heating, children’s studies and preparation for school/work; for
upper middle income group no cooling/heating, regular household chores and preparation for
school/work; and for upper income group no cooling/heating, preparation for school/work and
children’s studies.
43 percent of sample households are of the view that change in loadshedding timing can make
loadshedding less disruptive (See Table 27.8). Sample preferences with respect to
loadshedding timing are given in Chapter 30.
138
Table 27.7
Ranking of Disruptions Due to Outages
(%)
Preparation
for
work/school
Studies
(home work
of children)
Income
generating
activities
work
Regular
household
work
Entertainment/
Leisure
Social
Activities
(visits to/ of
friends, etc
No
cooling/
heating
Shortage
of water
Total
Punjab
19
21
9
12
3
3
20
13
100
Sindh
9
16
11
13
1
2
44
5
100
Karachi
25
7
3
28
0
0
17
20
100
KPK
17
27
0
17
3
0
10
27
100
Balochistan
17
18
8
15
2
2
24
13
100
By Income Group
Upto 15000
12
18
13
11
2
2
19
23
100
By Province
1500135000
3500170000
70001 +
17
22
8
9
2
0
27
16
100
21
16
6
23
1
2
25
6
100
20
13
8
10
10
13
23
5
100
Total
17
18
8
15
2
2
24
13
100
139
Table 27.8
Change in the Timing to make Loadshedding Less Disruptive
(%)
By Province
Punjab
Sindh
KPK
Balochistan
By Income Group
Upto 15000
15001-35000
35001-70000
70001 +
Total
Yes
43
45
32
73
No
57
55
68
27
Total
100
100
100
100
56
44
39
28
43
44
56
61
73
57
100
100
100
100
100
140
CHAPTER 28
ADJUSTMENTS TO LOADSHEDDING
This chapter focuses on the types of adjustments that households make to outages in Pakistan.
28.1 NUMBER AND TYPES OF ADJUSTMENTS
As identified in Chapter 2, households have adapted to loadshedding through behavioral
changes. An important adjustment is acquisition of back-up power supply devices like
generators and UPSs. 28 percent of the
sample
households
have
acquired
generators (see Table 28.1).
Table 28.1
Household with Generator and UPS
(%)
The pattern differs by province and income
By Province
groups. A lower proportion of households
Location
in Punjab have purchased generators, 23
percent, as compared to 41 percent in KPK, and about one-third in Sindh and
Balochistan.
Expectedly,
purchase
of
generator differs with income. While about
three-fourths of upper income households
have generators, this proportion is lower
for middle and low income households.
Punjab
Sindh
KPK
Balochistan
By Income Group
Residents
with
Generators
23
32
41
33
Residents
with UPS
2
17
45
75
28
4
26
47
43
30
Upto 15000
15001-35000
35001-70000
70001 +
Total
35
21
27
23
Overall, small capacity generators have been acquired, of average capacity of 3.5 kva.
Consequently they only partially substitute
for power supply from public utility for 91
percent of the sample with self generator
(See Table 28.2).
However, for 14 percent of the sample
households in Sindh, 8 percent in Punjab
and 6 percent in K-Pk, generators fully
substitute for public electricity supply.
Generators are largely able to smooth the
disruption in children’s studies, cooling
(principally through fans) and some social
Table 28.2
Is Generator a Partial or Full Substitute of
Electricity Supplied Publically?
(%)
By Province
Partial
Full
Total
Punjab
Sindh
KPK
Balochistan
By Income Group
92
86
94
100
8
14
6
0
100
100
100
100
Upto 15000
15001-35000
35001-70000
70001 +
Total
100
100
88
90
91
0
0
12
10
9
100
100
100
100
100
activities while they are less able to ensure continuation of leisure/entertainment and regular
household work See Table 28.3)
141
Table 28.3
Use of Generator for Various purposes
Leisure/Ente
rtainment
Yes
No
Cooling/Hea
ting
Yes
No
Social
Activities
Yes
No
Homebased/Economic
Yes
No
Children
Study
YES
No
By Province
Punjab
58
42
74
26
69
31
65
35
89
11
Sindh
49
51
69
31
63
37
54
46
91
9
KPK
16
84
55
45
35
65
42
58
68
32
Balochistan
30
70
60
40
70
30
30
70
80
20
Upto 15000
100
0
100
0
100
0
100
0
100
0
15001-35000
48
52
65
35
61
39
42
58
81
19
35001-70000
37
63
62
38
54
46
53
47
85
15
70001 +
57
43
83
17
73
27
70
30
87
13
Total
45
55
67
33
60
40
55
45
84
16
By Income Group
Households rely more on UPSs in Punjab, 35 percent, than in the other provinces where more
households have acquired generators (See Table 28.1). Affordability and perhaps the pattern of
loadshedding (with somewhat more outages of an hour duration) may explain this pattern. Also,
a higher proportion of lower and
lower middle households have
UPSs,
about
30
percent,
as
compared to generators at 19
percent.
UPSs
are
electricity for 94 percent of the
sample households (See Table
28.4). They are partially fulfilling
leisure/entertainment
cooling
requirements,
By Province
partial
substitute for public supply of
the
Table 28.4
UPS a partial or full substitute of electricity supplied
publically (%)
and
social
activities, household chore needs
Punjab
Sindh
KPK
Balochistan
By Income Group
Upto 15000
15001-35000
35001-70000
70001 +
Total
Partial
Full
Total
97
96
75
100
3
4
25
0
100
100
100
100
100
96
91
100
94
0
4
9
0
6
100
100
100
100
100
for about 35-45 percent of the sample households (See Table 28.5).
142
Table 28.5
Use of UPS for Various purposes
(%)
Leisure/
Entertainment
Yes
No
By Province
Punjab
Sindh
KPK
Balochistan
By Income Group
Upto 15000
15001-35000
35001-70000
70001 +
Total
Work/Economic
Activities
Yes
No
Cooling/Heating
Yes
No
Social Activities
Yes
No
42
35
10
14
58
65
90
86
31
70
60
0
69
26
40
100
42
39
70
29
58
61
30
71
37
48
10
43
63
52
90
57
50
33
34
47
36
50
67
66
53
64
25
41
35
59
40
75
57
65
41
60
50
41
46
47
45
50
59
54
53
55
25
39
32
47
36
75
61
68
53
64
Beyond the use of generators and UPSs, some households adjust by shifting timings of various
activities to avoid loadshedding. These adjustments are presented in Table 28.6. 77 percent of
the sample households reported shifting timings of studies and regular household work while 45
percent said that they shifted the time of economic activities because of loadshedding. Over and
above these, households have made purchases of battery operated appliances like emergency
lights/fans to minimize the impact of loadshedding. These practices are in all provinces and in
all income groups.
Table 28.6
Various Other Adjustments made to Deal with Loadshedding
(%)
Changed the
timing of study
Changed the
timing of regular
household work
Changed the timing of
economic activities
Bought battery
operated
electrical
appliances
By Province
Punjab
Sindh
KPK
Balochistan
By Income Group
Upto 15000
15001-35000
35001-70000
70001 +
Total
Yes
81
70
70
80
No
19
30
30
20
Yes
78
72
81
73
No
22
28
19
27
Yes
47
28
68
43
No
53
72
32
57
Yes
36
67
42
80
No
64
33
58
20
79
85
76
40
77
21
15
24
60
23
81
83
74
50
77
19
17
26
50
23
47
49
44
33
45
53
51
56
68
55
45
58
41
25
46
55
42
59
75
54
143
CHAPTER 29
OUTAGE COSTS
The objective of this chapter is to present the estimated magnitudes of different types of costs
associated with outages. As identified in chapter 25, these include direct costs which consist of
spoilage costs and indirect or adjustments costs which include generator costs and UPS costs.
Section 1 of the chapter presents the total outage costs by location (province), and income
expenditure group. Section 2 derives the cost per Kwh of load shedding. Finally, by blowing up
the sample, the magnitude of outage costs to the residential sector of Pakistan is derived.
29.1 TOTAL OUTAGE COSTS
Table 29.1 shows that the total outage cost on average to each residential consumer is almost
31,000 Rs per annum. The variation in outage costs is not very large among Provinces, ranging
from about Rs 29,200 per consumer in Punjab to Rs 34,100 in K-PK.
Table 29.1
Total Outage Cost per Residential Consumer
(Rs)
Monetization
of Utility Loss
By Province
Punjab
Sindh
K-PK
Balochistan
By Income Group(Rs)
0 – 15000
15001 – 35000
35001 – 70000
70001 and above
Total
Share (%)
7355
7626
4954
3530
Cost of Self-Generation
Other Costs
Total Outage
Cost
3864
2054
2037
2573
6747
6075
8104
5235
29229
33317
34059
29458
Generator
Cost
UPS Cost
11263
17562
18964
18120
3828
290
400
4262
8780
5655
6380
2734
6749
21518
9544
8193
6824
22
22370
50900
14215
46
4831
4550
3114
10
7053
4549
6712
22
43798
75192
30865
100
Outage costs rise sharply by consumption (income) level of a consumer. For households with
monthly consumption expenditure of upto Rs 15000, the outage cost annually is Rs 8800. For
the highest expenditure group of households the cost rises to Rs 75200.
144
Overall, for the sample as a whole, the largest component of outage costs is self-generation
costs at 56 per cent. Monetization of utility loss and other costs (spoilage costs, income
foregone in household economic activity, etc. each account for 22 per cent.
For lower income households, the share of monetization of utility loss is higher at 44 per cent
because a low proportion of such households have either a generator or an UPS. As opposed to
this, the share of self-generation costs for the highest expenditure households is high at 74 per
cent.
29.2 BURDEN OF OUTAGE COSTS
The burden of outage costs as
a
percentage
of
total
consumption expenditure by a
Table 29.2
Total Outage Cost as Percentage of Total Household
Consumption Expenditure
(000Rs)
household is given in Table
29.2.
highest
It
appears
burden
that
the
on
the
is
‘middle class’ living in the cities
of Pakistan. It is 7 per cent for
such households as compared
0 – 15000
15001 – 35000
35001 – 70000
70001 and above
Total
Annual
Outage
Cost
8.8
21.5
43.8
75.2
30.9
Annual
Consumption
Expenditure
142.5
305.9
627.6
1293.9
461.1
Outage Costs %
of Consumption
Expenditure
6.2
7.0
7.0
5.8
6.7
to 6.2 for low income households and 5.8 per cent for the richest households.
Figure 29.1
Outage Costs as % of Annual Household Consumption Expenditure
7.0
7.0
6.2
0 – 15000
5.8
15001 – 35000
35001 – 70000
70001 and above
145
29.3 OUTAGE COST PER KWH
Table 29.3 indicates the
total outage cost per Kwh
for residential consumers on
Table 29.3
Total Outage Cost per Kwh to Residential Consumer
(Rs)
average is close to Rs 24
(25 cents) per Kwh. This is
substantially lower than the
outage cost to small-scale
industry
and
commercial
consumers of Rs 51 (53
cents) and Rs 68 (70 cents)
respectively.
The highest outage cost per
By Location
Punjab
Sindh
K-PK
Balochistan
By Income Group
0 – 15000
15001 – 35000
35001 – 70000
70001 and above
Total
Total Outage
Costs
Electricity not
provided
(Kwh)
Outage Cost
per Kwh (Rs)
29229
33317
34059
29458
1655
830
865
1474
17.66
40.14
39.37
20.00
8780
21518
43798
75192
30865
479
732
1599
4299
1289
19.32
29.40
27.39
17.49
23.94 (25 c)
Kwh is observed in Sindh at Rs 40 (42 cents) per Kwh, while the lowest cost is in Punjab at Rs
18 (19 cents) per Kwh. In line with the pattern observed in figure 29.1 the outage cost per Kwh
is the highest for the `middle class` at Rs Rs. 27 (28 cents)- Rs 29 (30 cents).
29.4 NATIONAL ESTIMATE OF OUTAGE COSTS
Blowing-up of the sample to
arrive at a national estimate
Table 29.4
National Estimate of Outage Costs to Urban Residential
requires, first, estimation of
Consumers, 2011-12
the number of urban
households in the country. Monthly Total
Number of
Outage Cost Total Outage
Households
per
Cost
According to the PES the Consumption
Expenditure
(000s)a
Household
(Rs billion)
population of Pakistan in Group(Rs)
(Rs)
0
–
15000
5014
8780
44.0
2011-12 is 180.7 million, out
15001 – 35000
4360
21518
93.8
of which 37.4 percent is 35001 – 70000
763
43798
33.4
327
75192
24.6
located in the urban areas. 70001 and above
Total
10464b
195.8
The average household size a adjusted on the basis of distribution in the HIES, 2010-11
is given in the latest HIES of b 10.9 million households in urban areas with 98 percent of households
having access to electricity according to PSLSMS, 2010-11
the PBS at 6.19. This
implies that there are 10.9 million urban households in the country.
Second, there is need to determine the distribution of urban households by level of monthly
consumption expenditure. This has also been derived from the HIES and is presented in Table
29.4.
Overall, the total outage cost to residential consumers in the urban areas of Pakistan is
Rs 195.8 Billion in 2011-12.
146
CHAPTER 30
LOAD MANAGEMENT STRATEGY: CONSUMER’S PREFERENCES
The questionnaire contains a module to solicit consumer preferences regarding timing of
loadshedding which can reduce the costs and disruptions due to the outages. These can
provide guidance to the load management strategy by DISCOs, the formulation of which should
be a priority since loadshedding is likely to persist over the next few years.
30.1 LEVEL OF SATISFACTION WITH CURRENT LEVEL OF SERVICE
Only 43 percent of sample firms indicated that DISCOs kept to the announced loadshedding
schedule (see Table 30.1).
Table 30.1
respondents if they were satisfied with
Power Companies Kept to Loadshedding Schedule
(%)
By Province
the current level of service by the
Location
DISCOs/KESC.
Punjab
Sindh
KPK
Balochistan
The survey teams enquired from the
43 percent
of the
respondents ranked their satisfaction
Yes
level as very low while over one-third
By Income Group
ranked it as low (see Table 30.2).
Upto 15000
15001-35000
35001-70000
70001 +
Total
Clearly,
the
consumers’
level
of
satisfaction with the public distribution
No
Total
33
75
45
10
67
25
55
90
100
100
100
100
40
44
44
35
43
60
56
56
65
57
100
100
100
100
100
companies is very low.
The
questionnaire
solicited
the
preferred type of load shedding from
the sample households. Specifically,
they
were
asked:
what
type
of
Table 30.2
Level of Satisfaction with Current Quality of
Service by DISCOs/KESC
(%)
loadshedding is preferred- longer each
time but fewer outages or shorter each
time but more outages. Table 30.3
shows that 65 percent of the consumers
prefer the latter. Residential consumers
in Sindh and Punjab clearly prefer
shorter though more outages for the
same
total
time
of
outages.
The
Very
high
High
Medium
Low
Very
Low
1
6
0
3
5
6
4
0
10
24
25
0
39
27
31
23
44
36
40
73
2
2
2
10
2
4
3
6
13
5
9
17
16
18
15
34
35
38
18
34
51
43
39
43
43
By Province
Punjab
Sindh
KPK
Balochistan
By Income Group
Upto 15000
15001-35000
35001-70000
70001 +
Total
147
preference is for the opposite in K-
Table 30.3.
PK, that is, longer each time but
fewer
outages,
while
Preference for the type of Loadshedding
(%)
the
preference between the two types
of
loadshedding
is
evenly
30.2 PREFERRED CHANGES
IN
TIMINGS
OF
LOADSHEDDING
About 97 percent of the sample
households reported summer time
the
worst
season
Shorter each
time but more
outages
Total
34
15
64
50
66
85
36
50
100
100
100
100
40
32
34
45
35
60
68
66
55
65
100
100
100
100
100
By Province
distributed in Balochistan.
as
Longer each
time but
fewer
outages
for
Punjab
Sindh
KPK
Balochistan
By Income Group
Upto 15000
15001-35000
35001-70000
70001 +
Total
loadshedding (see Table 30.4).
Winter time is the second worst season for loadshedding.
Table 30.4
Worst Time of The Year for Loadshedding
Summer
Spring
Rank
Winter
Fall
Total
93
97
92
100
97
1
0
3
0
1
4
1
5
0
1
0
1
0
0
1
100
100
100
100
100
Upto 15000
97
1
1
1
100
15001-35000
97
1
2
0
35001-70000
70001 +
Total
90
95
97
2
0
1
7
2
1
1
3
1
100
100
100
100
By Province
Punjab
Sindh
KPK
Balochistan
Total
By Industrial Group
The questionnaire also contained a question regarding the worst day of the week for outages.
While 27 percent of the respondents said all days are bad, about one-thirds said Sunday is the
worst day and Friday was the worst day for about 20 percent of the respondents. Friday, of
course, is the prayer day.
Over one-third of the sample households preferred loadshedding in the first half of the day, that
is, between 6:00 am-12noon. 27 percent of the respondents preferred loadshedding in the
second half (12noon-6:00pm) while close to one-fifth each preferred it to be during 6:00pmmidnight and mid-night-6:00am (see Table 30.6).
148
Around 31 percent of the respondents
indicate that it will be helpful if the power
companies provided more information
Table 30.5
The Worst Day of The Week for Outages
(see Table 30.7). Clearly, these should be
Sunday
Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
All days equal
focused upon in the load management
Total
relating to the methods to save electricity
while
about
information
one
about
quarter
outages
said
that
and
the
scheduling of the outage will be useful
Frequency
165
68
1
2
4
99
24
137
Percentage
33.0
13.6
0.2
0.4
0.8
19.8
4.8
27.4
500
100.0
strategy of the distribution companies.
Table 30.6
Preference of Loadshedding Time
(%)
Rank
6am-12 Noon
12Noon-6:00pm
6 pm-Midnights
Midnight-6am
Total
31
29
47
47
34
26
45
8
20
27
22
12
24
17
20
21
15
21
17
19
100
100
100
100
100
Upto 15000
21
34
25
21
100
15001-35000
32
44
35
34
24
24
33
27
21
16
20
20
23
16
13
19
100
100
100
100
By Province
Punjab
Sindh
KPK
Balochistan
Total
By Income Group
35001-70000
70001 +
Total
Table 30.7
Information that can be provided by Distribution companies to consumers
Percentage
Save electricity
31.2
Information about outage
25.5
Time table for load shedding
26.1
Awareness about outage required
14.6
Others
2.5
149
CHAPTER 31
SUGGESTIONS BY THE SAMPLE UNITS
The questionnaire at the end solicited the respondent’s views/ suggestions to help handle the
loadshedding problem in the country. Specifically, the open ended question asked for
“suggestions to reduce the costs of loadshedding”. A number of interesting suggestions
emanate from the survey responses. These can be categorized as relating to the following:

Enhancing the supply of electricity

Alternative sources of energy/ fuel use

Improving governance or management

Changes in pricing policy
Enhancing the Supply of Electricity: About 43 percent of the respondents are of the view
that new dams, including Kala Bagh Dam, should be constructed to permanently enhance the
supply of electricity in the country at low costs (see Table 31.1). This suggestion dominates the
response not only from the sample units located in Punjab, but is also significant in the case of
Peshawar. Over 27 percent of the respondents also think that new power plants should be built
while close to a fifth of respondents are also of the view that electricity should be imported.
Responses are more or less, similar across income groups. (See Table 31.2)
Alternative sources of Energy Fuel for Energy: A number of suggestions have been given
regarding resort to alternative energy and fuel sources by the respondents. Over one-fifths of
the respondents suggested the use of different methods of electricity generation, while 14
percent specifically suggested the use of coal for electricity generation. Close to 8 percent of the
sample units suggested introduction of solar energy systems (particularly by respondents in
Multan and Mardan).
Improving Governance/Management. The dominant recommendations in this category are to
minimization of electricity theft and to stop corruption, with 17 percent of respondents each
emphasizing it. Need for honest employees, minimization of line losses and awareness creation
for proper use of electricity along with privatization of the DISCOs were also cited as possible
mechanisms to lower loadshedding costs.
150
Table 31.1
Suggestions by Sample Units by City
(% of Respondents)
Reasons
Lahore
Faisalabad
Gujranwala
Multan
Sialkot
Rawalpindi
/Islamabad
Karachi
Hyderabad
Sukkr
Peshawar
Mardan
Abbotabad
Quetta
Total
Enhancing Supply of Electricity
Gas Pipe line from
Iran to avoid gas
shortage
Import Electricity
Construct new
Dams (including
Kala Bagh Dam)
Use rental power
system
Build new power
plants
9
8
8
3
0
18
25
5
0
20
92
0
13
15
16
69
46
3
0
10
31
10
0
12
23
0
20
22
59
53
19
63
31
57
18
25
0
40
77
58
27
43
8
0
0
3
0
10
10
0
0
2
0
0
0
5
27
24
27
18
31
51
20
30
30
22
0
8
40
27
Alternative Energy Fuel Sources
Use Coal for
electric generation
Use different
method of electric
generation
Bio Gas system
Introduce solar
energy System
13
10
35
3
0
16
24
5
0
16
23
0
10
14
36
14
19
29
46
20
14
40
20
24
15
25
3
23
0
0
0
0
0
0
1
0
0
0
0
0
0
0
15
2
15
24
0
7
0
10
0
6
23
8
3
8
Governance/Management
Privatize Electric
department
Need Honest
Employees
Minimize electric
theft
Stop Corruption
Minimize line
losses
Give awareness to
people use of
electricity
1
0
0
0
0
3
5
0
0
0
0
0
3
2
0
10
15
5
0
8
3
0
0
10
15
0
0
5
10
8
12
50
15
16
15
30
40
10
0
25
30
17
16
6
23
8
8
20
4
60
50
30
0
25
17
17
1
0
0
3
0
3
0
10
0
0
0
8
0
1
1
0
0
11
0
0
0
0
0
0
0
0
0
1
5
12
12
5
0
11
28
5
0
8
0
8
43
13
4
0
0
13
54
18
8
30
0
10
0
0
13
10
Pricing Policy
Government give
subsidy on
electricity
Reduce price at
source
151
Pricing Policy Around 13 percent of the sample units requested for subsidy for electricity from
the government while 10 percent suggested that the price (at source) should be reduced
through economizing on costs.
Table 31.2
Suggestions by Sample Units by Income Group
( % of Respondents)
Reasons
Enhancing Supply of Electricity
Gas Pipe line from Iran to avoid gas shortage
Import Electricity
Construct new Dams (including Kala Bagh Dam)
Use rental power system
Build new power plants
Alternative Energy Fuel Sources
Use Coal for electric generation
Use different method of electric generation
Bio Gas system
Introduce solar energy System
Governance/Management
Privatize Electric department
Need Honest Employees
Minimize electric theft
Stop Corruption
Minimize line losses
Give awareness to people use of electricity
Pricing Policy
Government give subsidy on electricity
Reduce price at source
Upto 15000
1500135000
3500170000
70001
+
Total
8
19
38
3
28
15
21
39
3
32
20
25
47
6
21
10
23
60
10
30
15
22
43
5
27
8
26
0
7
11
19
1
9
18
24
0
6
28
25
0
18
14
23
0
8
1
3
22
17
0
3
1
7
17
21
2
1
1
5
15
14
2
0
0
3
18
8
0
0
2
5
17
17
1
1
13
12
13
10
11
7
15
10
13
10
To conclude, the top five suggestions emanating from the respondents of the survey are
as following:
First: Construct Dams
Second: Build New Power Plants
Third: Use Different Methods of Electricity Generation
Fourth: Import Electricity
Fifth: Minimize Theft and Stop Corruption
152
CHAPTER 32
CONCLUSIONS AND POLICY IMPLICATIONS
We have highlighted in the previous chapters the principal findings on the incidence of outages
in the residential sector. In this concluding chapter we derive the key policy implications.
32.1 IMPACT OF OUTAGES
The estimated impact of outages on households is as follows:
(i) Outages on the average occur almost five times a day for 17% of the time. The highest
incidence is in Punjab at 1683 hours annually, 16% above the national average. The
lowest incidence is in Sindh at 23% below the national average.
(ii) Outages are disruptive most of heating/cooling, household activities, preparation for
work/study (especially by children) and any home-based economic activity.
(iii) The outage cost per Kwh works out as Rs 24(25c). This is 53% less than the cost to smallscale industry and 65% less than the cost faced by the commercial sector. These results
are consistent with the findings of the other studies.
32.2 AFFORDABILITY
Table 32.1 presents the total cost of electricity consumption to household at different levels of
total consumption expenditure (proxy for income). Overall, this is estimated at close to 17%. A
striking finding is that the cost is the lowest for the upper most income group.
In the pre-loadshedding period, in 2005-06, according to the HIES, the share of electricity cost
in total consumption expenditure was 5% on average for urban households. Following the high
levels of loadshedding this share has jumped up by over three times.
Table 32.1
Total Cost of Electricity Consumption Per Residential Consumer
Monthly Expenditure
Group(Rs)
0-15000
15001- 35000
35001-70000
Annual
Electricity Cost
of Public
Total Outage
Supply
Cost
15.3
8.8
28.8
21.5
65.1
43.8
142.5
305.9
627.6
(Rs in 000)
Total Electricity
Cost as % 0f
Consumption
Expenditure
16.9
16.4
17.4
Annual
Consumption
Expenditures
70001 and above
130.6
75.2
1293.9
15.9
Total
46.7
30.9
461.1
16.8
153
It is clear that the high share of expenditure on electricity is cutting into consumption of food,
clothing and basic services (like education and health), especially by the low income groups. As,
such an indirect impact of the high level of loadshedding in the country is the reduction in
nutrition levels, particularly of children. Along with impact on preparation for school and home
work, the impact of outages on children needs to be more strongly highlighted.
Overall, limits of affordability to power tariffs have been reached by bulk of the households and
the scope for further enhancement in tariffs is very limited.
32.3 PRICING POLICY
There are concomitant implications of the above findings in affordability on the power tariff
structure for the residential sector. The present structure is given in Table 32.2, excluding taxes
and other charges.
The average tariff for different
levels of electricity billing (in Kwh)
is given in figure 32.2.
Given the regressive burden of
electricity costs, as shown in Table
32.2, there is need to make the
tariffs structure more progressive
in a revenue-neutral way. In line
with
these
considerations
Table 32.2
Present Tariff Structure on the Residential Sector
(Rs)
Up to 50 units
Actual
Per Kwh
Proposed
Per Kwh
2.00
2.00
For consumption exceeding 50 units
1 – 100 units
5.79
101 – 300 units
8.11
4.50
7.50
301 – 700 units
12.33
13.00
Above 700 units
15.07
17.50
the
proposed tariff structure is also given in Table 32.2. Beyond 300 units it is proposed to enhance
incremental tariffs and reduce them before this level of consumption.
32.4 SELF-GENERATION
The prevalence of self-generation is relatively low among residential consumers. 28% have
generators and 30% have UPS. Resort to self-generation is the highest is Sindh and K-PK and
among consumers in the highest income category.
The average capacity of generators in use is under 3.5 KVA. The proposal for eliminating the
GST on small generators and UPS is justified in this case also, as for commercial consumers.
154
32.5 LOAD MANAGEMENT STRATEGY
Based on responses by the sample households, the following proposals are presented for
reducing the level of outage costs:
(i)
The majority, 65%, of respondents prefer, given the total duration of loadshedding,
shorter though more frequent outages. Higher duration of a typical outage is one of
the main reasons why outages costs are higher in Karachi, despite lower incidence
of outages.
(ii)
Bulk of the loadshedding is in the morning from 6:00 am to 9:00 am. This creates
disturbance in preparation for work/school and heating during winters. Over 43% of
sample households report that changing loadshedding times to later in the day would
be less disruptive, especially to low income households.
(iii)
The worst time in year for load shedding is summer and worst day are Sunday,
Monday and Friday. To the extent there is scope, the pattern of loadshedding needs
to be adjusted accordingly.
(iv)
There has been a clear vote of non-confidence against the services provided by the
power sector. 43% rate the quality of services as ‘very low’ and 35% as ‘low’.
Distribution companies, in particular, will have to work very hard to rehabilitate their
image.
(v)
A series of recommendations have been made for reducing the costs of
loadshedding, as follows,
Construct New Dams
43%
Build New Power Plants
27%
Import Electricity
22%
Minimize Electricity Theft
17%
Stop Corruption
17%
Use Coal
14%
Gas Pipeline From Iran
15%
Subsidy
13%
Reduce Price
10%
Solar Energy
8%
Therefore the largest responses relate to enhancement in electricity supply and to improved
management of power sector.
155
PART V
COST OF LOADSHEDDING TO
AGRICULTURE SECTOR
156
CHAPTER 33
INTRODUCTION
This part of the study presents the findings on costs of loadshedding to the agricultural sector of
Pakistan and to rural households.
The report is organized in eight chapters. Chapter 34 presents the methodology used for
quantification of costs due to outages. An estimate is also made of the cost of loadshedding on
the basis of secondary data. Chapter 35 describes the survey including the sampling
methodology. Chapter 36 gives the magnitude of the parameters like the incidence of outages,
impact on water supply, loss of output, etc., leading to an estimate of the cost of outages in
agriculture.
Chapter 37 quantifies the cost of outages in home-based economic activities and in domestic
consumption. Chapter 38 presents proposals for an improved load management strategy while
Chapter 39 highlights the suggestions made by sample units. Finally, Chapter 40 summarizes
the key findings and the policy implications which emerge from the research.
157
CHAPTER 34
METHODOLOGY
34.1 SECONDARY-DATA BASED METHODOLOGY AND ESTIMATES
Electricity consumers in agriculture are owners or renters of tube wells operated by electric
motors, both public and private. Consumption for other purposes is shown in other categories of
consumers.
Given secondary data availability, it is possible to get first estimate of the power outage costs in
agriculture. These costs are designated as OC and derived on the following basis:
…………………………………………… (1)
OC = q . t . e. ℓ .V
Where
q = share of crop output from irrigated area in Pakistan
t = share of irrigated area that is irrigated by tube well
e = share of water from tube wells extracted by electric tube wells
ℓ = share of time lost due to loadshedding
V = share value added in the crop sector, both major and minor crops (in the absence
of loadshedding)
The magnitudes of the above variables are derived below.
Estimate of q: the latest Agricultural Census of 2010 gives an estimate of the extent to which
the cultivated area is irrigated, as follows:
Table 34.1
Distribution of Cultivated Area between Irrigated and Barani Area
(000 Hectares)
Irrigated
Punjab
% by
Province
65
9425
(86)a
Sindh
2925
20
(95)
K-PK
1111
8
(62)
Balochistan
1016
7
(72)
Pakistan
14477
100
(84)
Source: Agricultural Census, 2010
Barani
Province
Total
1516
(14)
168
(5)
691
(38)
397
(28)
2771
(16)
55
10941
(100)
3093
(100)
1802
(100)
1413
(100)
17248
(100)
6
25
14
100
% by
Province
63
18
10
8
100
158
In order to derive the share of crop output of irrigated land, it is also necessary to know the
difference in yield of individual crops between irrigated and barani area. This exercise has been
undertaken for Punjab, where over 70 percent of the tube wells are located. Results are given in
Table 34.2. It is estimated that the value added per acre in irrigated area is about 2.5 times the
value added per acre in barani area. As such, about 93 percent of the value added of crops in
Pakistan is generated in the irrigated areas of the country. Therefore, q = 0.93.
Table 34.2
Yield of Crops* in Irrigated and Barani Area of Punjab
Yield (tonnes per hectare)
Irrigated
Barani
Ratio
Wheat
2.95
1.32
2.23
Maize
8.24
2.43
3.39
Gram
0.47
0.47
1.00
*Other crops like rice, sugarcane, cotton, vegetables and fruits are mostly grown on irrigated land.
Source: Punjab Development Statistics.
Estimate of t and e: The quantum of water provided by electric and diesel tube wells
depends on their rates of capacity utilization. The average hours operated by each type of tube
well are given in Table 34.3. It appears that although the number of electric tube wells is less,
they are operated more intensively.
Table 34.3
Number of Hours Operated by Electric and Diesel Tube wells
Unit
BY OWNER
Number of Tube wells
’000
Hours per tube wells
Hrs
Total hours of use
Million hrs
BY RENTING
Number of Tube wells
’000
Hours per tube wells
Hrs
Total hours of Renting
Million hrs
Total Hours
Share
%
Source: Agricultural Statistics Year Book
Electric
Diesel
Total
161
1104
177.7
776
625
485.0
937
48
597
28.7
206.4
28
142
316
44.9
529.9
72
662.7
190
73.6
736.3
100.0
As shown in Table 34.3, the share of electric tube wells in water extracted by tube wells is 28
percent.
159
Next, the share in water availability from different sources at the farm-gate is given in Table
34.4.
Table 34.4
Water Availability at Farm Gate due to Different Sources, 2010-11
Total Water Available
(MAF)
Surface Water
86.95
From Electric Tube
wells
(MAF)
------
Diesel Tube wells
(MAF)
Ground Water
50.20
20.47
29.73
Public Tube wells
8.91
8.91
--------
Private Tube wells
41.30
11.56
29.7
Total*
137.15
20.47
29.73
------
*Other sources make a very small contribution
**28% as derived from table
Source: Agricultural Statistics Year Book
Therefore, from Table 34.4 we have that t = 0.37 and e = 0.41
Estimate of 𝓵: The electricity consumption per agricultural consumer is given in Table 34.5.
There has been fall after 2009-10 from 23365 Kwh to 17714 Kwh in 2011-12, a fall of 24
percent. As such, the magnitude of ℓ is assumed to 0.24.
Table 34.5
Electricity Consumption per Agricultural Consumer
Growth Rate
(%)
2007-08
Agricultural Consumers
(Tube well)
18130
2008-09
23146
17.7
2009-10
23365
10.9
2010-11
20077
-14.1
2011-12
1714
-11.9
*Excluding Balochistan, with pre-dominantly deep water public tube wells
Source: NEPRA, State of Industry Report, 2010
Estimation of V: Net of the impact of loadshedding, the value added in the crop sector of
Pakistan in 2011-12 is Rs 1574 billion. Given equation (1) we have from that q = 0.93, t = 0.37,
e = 0.41 and ℓ = 0.24. This implies that the cost of loadshedding to the agricultural sector
of the country is Rs 55 billion. This represents a production loss of 3.5 percent in the case of
crops.
160
34.2 Primary–Data Based Methodology
The cost of outages is estimated for a farm which owns/rents an electric tube well, by first
designating the following variables:
WK = % reduction in number of rounds in kharif season due to power outages
WR = % reduction in number of rounds in Rabi season due to power outages
OK = % loss of output due to reduction in number of rounds in Kharif
OR = % loss of output due to reduction in number of rounds in rabi
Then,
OK = fk (wK) and OR = fR (wR)
…………………………………………… (1)
The cost of outages in the case of a farm with no stand-by generator is given by
CO = OK VK + OR VR
…………………………………………… (2)
Where,
VK = value added by the farm from crops during Kharif season (in the absence of outages)
VR = value added by the farm from crops during Rabi season (in absence of outages)
Virtually all the farms in the sample did not have a stand-by generator. Therefore, the operative
methodology is given by (2).
161
CHAPTER 35
THE SAMPLING FRAMEWORK AND ITS DISTRIBUTION
Implementation of the methodology for estimation of the cost of power outages in agriculture
requires the collection of data from a sample of farmers in Pakistan. The approach adopted to
selection of the sample and the resulting sample distribution is described below.
35.1 THE SAMPLING APPROACH
The limitation of funds for the study restricted the national sample size to only 250 farmers.
Random sampling would have severely limited the number of farmers with electric tube wells.
Therefore, a purposive sampling approach was adopted. An attempt was made to ensure that at
least one-third of the farmers sampled had electric tube wells. This involved more search for the
appropriate farmers in the villages covered during the survey. The resulting sample is given in
Table 35.1.
Following edit checks, 240 sample farms have been analyzed to determine costs of outages. A
comparison of the population and sample size distribution is given in Table 35.2.
Table 35.1
Sample Size and Distribution
Number of sample
Farms
Province
Punjab
125
Sindh
60
K-PK
40
Balochistan
25
Size of Farm (Cropped Area)
0-10
18
10-25
70
25-50
87
50-100
53
100+
22
Total
250
% of Sample
Number of Farms with
electric Tube wells
% Sample
Farms
50
24
16
10
43
23
13
15
34
38
32
60
7
28
35
21
9
100
1
24
36
27
6
94
6
34
41
51
27
38
Table 35.2
Comparison of the Population and Sample Size Distribution of Farms
(%)
Farm Size (Cropped Area)
0-10
10-25
25-50
50-100
100+
Total
Population Distribution
Sample Distribution
78
18
3
0.8
0.2
100.0
1
23
40
30
6
100
162
Therefore, the sample is concentrated in medium-sized and large farms who are more likely to
own tube wells.
The distribution of the original sample by districts within a province is given in Table 35.3.
Surveyors were instructed to select villages which were near or for away from the district road
network.
Table 35.3
Distribution of Sample by Districts
Province/District
Punjab
Lahore
Faisalabad
Sialkot
Gujranwala
Multan
Rawalpindi
Sindh
Karachi
Hyderabad
Sample Size
125
40
20
10
10
30
15
60
10
25
Province/District
Sukkur
K-PK
Peshawer
Mardan
Abbottabad
Bannu
Balochistan
Quetta
Total
Sample Size
25
40
20
10
6
4
25
25
250
163
CHAPTER 36
COST OF OUTAGES IN AGRICULTURE
The primary cost of outages in agriculture is the impact on output of crops due to reduction in
water availability arising from a reduction in the number of hours operated by electric tube wells.
The methodology of estimating accordingly the cost of outages is given in chapter 2 and is
applied below.
36.1 SOURCE OF IRRIGATION
For the total sample of 240 farms, the distribution of sources of irrigation is given below:
The analysis has not been conducted
by province because of the small
number of tube well owners in the
provinces
of
Sindh,
K-PK
Table 36.1
Distribution of Sample Farms by Sources of
irrigation
(%)
and
Balochistan. The share of tube wells is
higher because of the purposive
nature of the sample chosen.
Canal
Tube well + Canal Tube well
Other
Total
Kharif
47
38
15
100
Rabi
46
38
16
100
36.2 TYPE OF TUBE WELLS
Table 36.2 gives the presence of
owning or renting of electric tube wells
in farms of different sizes. The incidence
of use of electric tube wells is the
highest in the case of farms between 25
and 100 acres of cropped area. This
group of farmers is generally considered
as ‘progressive’ in terms of use of inputs
and agricultural practices.
36.3 INCIDENCE OF LOADSHEDDING
The incidence of loadshedding during
working hours on farm is given in Table
36.3. It appears that farms using electric
tube wells lose about 27 percent of the
working hours in outages.
Table 36.2
Incidence of Use of Electric Tube wells
Cropped Area
(Acres)
Sample Size
0-10
10-25
25-50
50-100
100 +
Total
18
65
82
53
22
240
Using Tube
wells
(Electric)
1
21
36
27
6
91
(%)
6
32
44
51
27
38
Table 36.3
Incidence of Loadshedding during Working
Hours on Farm
Cropped Area
(Acres)
0-10
10-25
25-50
50-100
100 +
Total
(farms with
electric tube
wells)
674
664
669
901
1011
716
(hours
annually)
2441
2240
2702
3182
4012
2693
28
30
25
28
25
27
164
The incidence of loadshedding appears to be the highest in Balochistan al almost 62 percent,
followed by the other three provinces at close to 23 percent.
36.4 REDUCTION IN OUTPUT
The reduction in number of water rounds and the consequent loss of output in farms using
electric tube wells is given in Tables 4.4 and 4.5 respectively.
The extent of production losses correspond closely to the incidence of loadshedding as hardly
any sample farm has a standby generator/motor.
Table 36.4
Reduction in Number of Water Rounds due to Power Outages
(Farms using Electric Tube wells)
Farm Cropped
Area (Acres)
Average
Number of
Rounds
(actual)
0-10
10-25
25-50
50-100
100 +
Total
4.00
6.29
5.11
5.37
5.00
5.44
Kharif
Reduction
in Number
of Rounds
due to
Outages
(actual)
2.00
2.71
1.94
1.81
1.83
2.08
%
(Reduction)
Average
Number of
Rounds
(actual)
33.3
30.1
27.5
25.2
26.8
27.7
9.00
8.19
7.97
8.44
10.83
8.36
Rabi
Reduction
in Number
of Rounds
due to
Outages
(actual)
2.00
3.62
2.83
3.11
5.17
3.24
%
(Reduction)
18.2
30.6
26.2
26.9
32.3
27.9
Table 36.5
Extent of Production Loss* due to Reduction in Water Rounds
(Farms using Electric Tube wells)
(%)
Farm Cropped (Acres)
0-10
10-25
25-50
50-100
100 +
Total
Kharif
30.0
28.6
26.1
25.3
31.6
27.0
Rabi
30.0
25.6
23.4
23.6
23.8
24.1
36.5 COST OF OUTAGES
The resulting loss of value added from cultivation of crops, both major and minor, is given in
Table 36.6. There does not appear to be much variation by from size in the percentage of crop
value added lost due to outages.
165
Table 36.6
Loss of Value Added due to Outages (Farms using Electric Tube wells)
(000 Rs per farm)
Farm Cropped
(Acres)
0-10
10-25
25-50
50-100
100 +
Total
Loss of Value Added
37
113
89
146
264
122
Value Added (in the
absence of outages)
153
453
386
620
1023
510
% Loss
24
25
23
24
26
24
It appears that almost one-fourth of crop output is lost by farms using electric tube wells. This is
a high percentage of loss. On the average, the outage cost per acre of cropped area is
estimated at approximately Rs 2650.
36.6 NATIONAL ESTIMATE OF OUTAGE COSTS
The sample has been adjusted in Table 36.7 to reflect the population distribution of farms by
size on the basis of the Agricultural Census of 2010. It is estimated that there are approximately
1 million farms in the country which either own or rent an electric tube well. This implies that on
average over four farmers use an electric tube well.
Table 36.7
Cost of Loadshedding in Agriculture
Cropped Area*
(Acres)
Number of
Farms
(000)
% using
Electric Tube
wells
Number using
Electric Tube
well (000)
Loadshedding
Cost per Tube
well
(’000)
37
113
89
146
264
0-10
6446
6
387
10-25
1240
32
397
25-50
372
44
164
50-100
165
51
84
100 +
41
27
11
Total
8264
1043
*On average farm size is 75 % of cropped area (both seasons combines)
Source: Agricultural census, 2010
Loadshedding
Cost (Rs in
Billion)
14.3
44.9
14.6
12.2
2.9
88.9
Almost 75 percent of farms using tube wells are relatively small, with cropped area upto 25
acres, and account for two-third of the national cost of outages in the agricultural sector.
Overall, the national outage in agriculture cost is Rs 89 billion in 2011-12. This is
significantly higher than the magnitude estimated from secondary data of Rs 55 billion. The
estimated amount of electricity not supplied is 3066 Kwh. This implies from the former estimate
that the outage cost per Kwh is Rs 29 (29 cent).
166
CHAPTER 37
OTHER COSTS OF OUTAGES
The survey carried out of farmers has also enabled determination of outages costs in homebased activities and in domestic consumption of electricity.
37.1 HOME-BASED ECONOMIC ACTIVITIES
The incidence of home-based economic
activities in the sample households is given
in Table 37.1 along with the nature of
adjustment of these activities to outages. It
appears that the only activity which is
significant impacted by loadshedding is
milking of animals.
The
estimated
value
added
in
milk
production on average per farm is Rs
10,000 per month. Sample farms have
reported that they lost 17 percent of their
Table 37.1
Incidence of Home-based Economic Activities
in Sample Farms and Nature of Adjustment
(with use of electricity)
Activity
% of
% making alternative
household
arrangements*
Stitching
11
60
Embroidery
5
65
Milking
33
46
Grinding
2
100
Crushing
10
100
Churning
7
100
Poultry
9
100
Others
2
100
*Primarily changing timing of activities
milk output due to outages. Given there
are 83 million farms in the country, the cost of outages in milking activity, COML, is given by
COML = 0.33 × 0.54 × 0.17 × 10000 ×12 × 8.3
million Rs
COML = 30172 million rupees
That is,
COML ≅ 30 billion rupees
37.2 DOMESTIC CONSUMPTION OF ELECTRICITY
The survey of households in the rural areas of Pakistan revealed that there is little or no use of
generators. As such, we have relied on the willingness to-pay approach to estimate the costs of
loadshedding in domestic consumption.
On the average, the sample households revealed that they are willing to pay an extra 50
percent to avoid outages.
The estimated number of rural households in Pakistan in 2011-12 is domestic consumption in
the country. This implies that the average annual electric consumption per rural household is
436 Kwh annually. At this level, the power tariff is approximately Rs --- per Kwh. As such the
cost of outages in domestic consumption by households is given by CODC where
167
CODC = 0.27 × 436 × 15.4 × 8.00 × = 14314 million Rs
That is,
CODC ≅ Rs 14 billion
This is based on the incidence of
outages at 27 percent of the time
and that 88 percent of rural
households
electricity.
have
access
to
Table 37.2
Total Outage Cost
(Rs in Billion)
Cost of Outages in:
89
Agriculture
Home-based economic activities
30
Domestic consumption
14
Total
133
37.3 TOTAL COST OF OUTAGES
In summary, the total cost of power outages in the rural areas of Pakistan are given in Table
37.2. The outage costs aggregate to Rs 133 billion in the rural areas of the country.
168
CHAPTER 38
LOAD MANAGEMENT STRATEGY
The questionnaire contains a module to solicit consumer preferences regarding timing of
loadshedding which can reduce the costs and disruptions due to the outages. These can
provide guidance to the load management strategy by DISCOs, the formulation of which should
be a priority since loadshedding is likely to persist over the next few years.
38.1 LEVEL OF SATISFACTION WITH CURRENT LEVEL OF SERVICE
The survey teams enquired from the respondents if they were satisfied with the current level of
service by the DISCOs/KESC. Almost two-thirds of the respondents ranked their satisfaction
level as very low while more than one-quarter ranked it as low (see Table 38.1) while almost
one-third of the respondents in Sindh appear moderately satisfied with the service level, the
proportion is much lower in the other provinces. Also, there are indication of a decline with the
level of satisfaction with the economic well-being of the respondent.
Table 38.1
Level of Satisfaction with Current quality of Service by DISCOs/KESC
(%)
By Province
Punjab
Sindh
KPK
Balochistan
Total
By Farm Size
0-10
10-25
25-50
50-100
100 +
Total
High
Medium
Low
Very Low
0
0
3
0
0
4
32
0
0
9
26
44
15
16
27
70
24
83
84
64
6
0
0
0
0
0
6
9
6
13
9
9
6
22
27
32
50
27
83
69
67
55
41
64
The sample for respondents was also asked how much lower tariff should they be paying for the
existing level of service, with loadshedding. This provides the first estimate of the respondent’s
perception of the cost of loadshedding. On an average, respondents felt that they should pay 28
percent less than what they currently are for the existing quality of service with interrupted
power supply as revealed by Table 38.2. The highest reduction is suggested in Punjab, 38
percent, and the large landlords, 40 percent.
169
Table 38.2
How much lower Tariff for existing level of Service
Percentage
By Province
Punjab
Sindh
KPK
Balochistan
Total
By Farm Size
0-10
10-25
25-50
50-100
100 +
Total
38.4
21.3
15.5
12.0
28.3
16.4
24.6
30.3
29.0
40.2
28.3
38.2 PREFERRED CHANGES IN TIMINGS OF LOADSHEDDING
One-thirds of the sample units prefer midnight to 6:00 am for loadshedding, probably because
though it impacts sleep but does not affect the economic activities. Loadshedding during 6 pm
to midnight is preferred by 28 percent of the respondents while 22 percent prefer loadshedding
in morning hours. The preference pattern in the smaller provinces appears different from Punjab
and Sindh. In Balochistan and K-PK, loadshedding unambiguously is preferred in the morning
hours (see Table 38.3). Also, preferences differ with the economic well being. While
Table 38.3
Preference of Loadshedding Time
(%)
Rank
6am-12 Noon
12Noon-6:00pm
6 pm-Midnights
Midnight-6am
Total
Punjab
Sindh
KPK
Balochistan
Total
6
12
45
84
22
By Province
19
14
3
4
14
By Farm Size
0-10
10-25
25-50
50-100
100 +
50
17
11
22
100
43
17
11
29
100
15
10
33
40
100
4
11
38
47
100
9
23
55
14
Total
22
14
28
35
36
32
15
4
28
37
42
38
8
35
100
100
100
100
100
100
170
respondents cultivating smaller area prefer loadshedding in morning hours, respondents with
large cultivated areas prefer it in evenings or night.
Respondents were equally divided
Table 38.4.
Preference for the Type of Loadshedding
when enquired about the preferred
type
of
loadshedding.
(%)
Half
preferred longer but fewer outages
each
time
and
half
shorter
each
outages
(see
Table
appears
that
in
Punjab
longer
but
Balochistan
time
preferred
but
more
38.4).
It
and
Punjab
Sindh
KPK
Balochistan
fewer
Longer each
Shorter each
time but
time but more
fewer
outages
outages
By Province
63
37
38
62
18
83
52
48
By Farm Size
Total
100
100
100
100
44
56
100
43
57
100
56
44
100
53
47
100
sized cultivators prefer longer but
0-10
10-25
25-50
50-100
100 +
36
64
100
fewer outages, the small and large
Total
49
51
100
outages are preferred while in
Sindh and K-PK the opposite
holds. Likewise, while medium
cultivators prefer shorter but more frequent outages.
Around one-third of the respondents indicate that it will be helpful if the power companies
provided more information relating to the scheduling of the outage; methods to lower the
electricity- intensity of
operations;
monthly
fixation of schedule of
Table 38.5
Information that can be provided by Distribution companies
Percentage
Information regarding loadshedding schedule
65
loadshedding to better
Use media to minimize consumption at peak timing
61
manage the outages
Fixed schedule for one month at least for better management
40
and how to work more
Awareness about outage adjustment system
7
Use energy saver instead of heavy bulbs
2
efficiently under
the
circumstances (see Table 38.5). Clearly, these should be focused upon in the load management
strategy of the distribution companies.
171
CHAPTER 39
SUGGESTIONS BY THE SAMPLE UNITS
The questionnaire at the end solicited the respondent’s views/ suggestions to help handle the
loadshedding problem in the country. Specifically, the open ended question asked for
“suggestions to reduce the costs of loadshedding”. A number of interesting suggestions
emanate from the survey responses. These can be categorized as relating to the following:

Enhancing the supply of electricity

Alternative sources of energy/ fuel use

Improving governance or management

Changes in pricing policy, and

Alternate sources of water
Enhancing the Supply of Electricity: The highest numbers of respondents, 15 percent are of
the view that new dams, including Kala Bagh Dam, should be constructed to permanently
enhance the production of electricity in the country at lower cost (see Table 39.1). Building new
power plants is the other significant suggestion to curb power shortages in Pakistan.
Table 39.1
Suggestions by Sample Units by Province
(% of Respondents)
Reasons
Punjab
Sindh
KPK
Balochistan
Total
Enhancing Supply of Electricity
Construct new Dams (including Kala Bagh Dam)
Build new power plants
Alternative Energy Fuel Sources
Use nuclear technology for generation
Introduce solar energy System
Introduce Wind Energy
Governance/Management
Privatize Electric department
Minimize electric theft
Stop Corruption
Pricing Policy
Government give subsidy on electricity
Fixed minimum price for water
Give subsidy on water
Alternative Supply of Water
Need proper canal water system
Provide minimum water for irrigation
20
17
12
6
5
0
16
0
15
10
17
45
24
8
50
4
0
58
0
8
56
8
11
49
14
10
1
5
0
0
12
0
0
13
0
0
8
5
0
8
42
19
31
24
10
24
25
15
35
48
8
32
36
15
30
22
28
26
38
40
40
32
20
27
31
172
Alternative sources of Energy Fuel for Energy: A number of suggestions have been given
regarding resort to alternative energy and fuel sources by the respondents. Almost half the
respondents suggested the introduction of solar energy systems while 14 percent and 11
percent respectively suggested the use of wind energy and nuclear energy.
Improving Governance/Management. Recommendations in this category include: privatization
of DISCOs; curb corruption, and; minimization of electricity theft.
Pricing Policy: A number of recommendations emanate in this category. Over one-third of the
respondents requested for subsidy for electricity from the government, while 30 percent of the
respondents suggested that there should be a subsidy on water. 15 percent suggested a fixed
minimum price for irrigation water.
Alternative Supply of Water: Over one-quarter of the respondents recorded the request for
proper supply of timely canal water so that their dependence on tube wells is correspondingly
reduced. Almost 31 percent have also demanded minimum water for cultivation purposes.
To conclude, the top five suggestions emanating from the respondents of the survey are as
following:
First: Introduce solar energy
Second: Government subsidy on electricity
Third: Ensure minimum water for irrigation
Fourth: Subsidy on water
Fifth: Ensure proper supply of canal water.
173
CHAPTER 40
CONCLUSIONS AND POLICY IMPLICATIONS
40.1 CONCLUSIONS
The research has yielded the following results:
i.
Application of a methodology based on secondary data yields an estimate of the costs of
power loadshedding in agriculture of Rs 55 billion, equivalent to a loss in value added by
crops of 3.5 percent.
ii.
The survey of 250 farmers indicates that 27 percent of the working hours on farm are
lost due to outages in the case of farms owning/renting electric tube wells.
iii.
The cost of outages, as quantified from the survey is Rs 89 billion, equivalent to over 5.5
percent of crop value added. This is significantly higher than the estimate from
secondary data.
iv.
The outage cost is estimated at Rs 29 (29 cents) per Kwh in agriculture
v.
The home-based economic activity impacted significantly by outages is milking of
animals. The estimated cost is Rs 30 billion. The cost of outages in terms of domestic
consumption in rural areas is Rs 14 billion.
vi.
Overall, the total cost of power loadshedding in rural areas of Pakistan in 2011-12 is Rs
133 billion.
40.2 POLICY IMPLICATIONS
A number of policy implications emerge from the research as follows:
Development of Alternate Technology: A large number of respondents have emphasized on
the access to solar and wind energy for operating tube wells. This makes eminent sense given
the dispersal of villages, especially in Balochistan. As an incentive all equipment for solar or
wind energy should be both duty free and exempted from GST.
Improvement of Irrigation System: Many farmers have highlighted the need for improvement
in canal water supplies in order to reduce the dependence on underground water. Provincial
governments must attach greater priority to regular and proper O & M of the irrigation network
and lining of canals to reduce water losses.
Substitution by Diesel Tube wells: In 2007-08, the share of electric tube wells in private tube
wells was 14 percent, with the remainder, 86 percent, being operated by diesel. This has
increased to 17 percent by 2010-11 despite the high incidence of outages. The reason for this is
the upsurge in the price of LDO which was 60 percent of the price of motor spirit in 2008. Now
174
the price is almost the same. It is necessary to reduce the price of the LDO to make the
operation of diesel tube wells more economic.
Incentive for Self-Generation: It has been observed in the survey that hardly any farm has
self-generation. As recommended elsewhere the purchase of small generators should also be
exempted from GST.
Moratorium on Village Electrification: Despite the severe shortage in availability of electricity
in rural areas, the program of village electrification continues unabated. Between 2007-08 and
2010-11 power has been supplied to an additional 36,606 villages/sub-villages. This is
exacerbating the problem of loadshedding.
175
TECHNICAL ANNEXURE
TECHNICAL ANNEXURE I
Multiplier Effects: it has been mentioned earlier that a change in the value added in the
productive sectors has secondary or multiplier effects on the rest of the economy. Therefore,
outages in the productive sector by reducing the value added in that sector cause a decline in
the level of economic activity in other sectors of the national economy. The generic form of the
macro-model used to derive the short-run magnitude of the multiplier is as follows:
Y1 = level of value added in a productive sector
Y2 = level of value added in other sectors of the economy
C = total consumption expenditure
I = total gross fixed capital formation
The structural equations of the model are as follows:
𝑌 = 𝑌1 + 𝑌2
……………………………..(1)
𝐶 = 𝛼0 + 𝛼1 𝑌 + 𝑎2 𝑍1
……………………………..(2)
𝐼 = 𝛽0 + 𝛽1 (𝑌 − 𝑌−1 ) + 𝛽2 𝑍2 ……………………………..(3)
𝑌2 = 𝛾0 + 𝛾1 𝐶 + 𝛾2 𝐼 + 𝛾3 𝑍3
……………………………..(4)
Where Z1, Z2 and Z3 are sets of exogenous variables and Y-1 is the level of GDP in the previous
period.
Comparative statics of the model yields
𝜕𝑌
𝜕𝑌1
= 1−𝛾
1
1 𝛼1 −𝛾2 𝛽1
……………….…………..(5)
The expression on the right hand side of equation (5) is the multiplier. It indicates the extent of
change in the GDP for a one rupee change in the value added by the industrial sector. A similar
result is derived for other sectors from the IPP Macro Economic Model.
The resulting estimate of the multiplier is 1.34. In other words, inclusion of a multiplier effect
increase indirectly the productive sectors outage costs by as much as 34 percent.
176
TECHNICAL ANNEXURE-II
Econometric analysis has been undertaken to determine by sector the relationship between
electricity consumption and value added:
Agriculture
AGVAL = Agricultural value added (at constant prices)
AGELC= Electricity consumption in agriculture
Results of the OLS regression are as follows:
Ln(AGVAL)
=
0.601
+0.05Ln(AGELC)
(1.84)**
(1.876)**
+0.927Ln(AGVAL -1)
(23.46)*
R2 = 0.994, D-W = 2.68, Degrees of Freedom= 37
*Significant at 5 per cent level
** Significant at 10 per cent level
Industry
INVAL = Industrial value added (at constant prices)
INELC = Electricity consumption in industry
Results of the OLS regression are as follows:
Ln(INVAL)
=
0.702
+0.126 Ln(INELC)
+0.864 Ln(INVAL -1)
(3.434)*
(2.374)*
(16.964)*
R2 = 0.998, D-W = 1.468, Degrees of Freedom= 37
*significant at 5 per cent level
Services
SEVAL = Services value added (at constant prices)
SEELC = Electricity consumption by services
Results of the OLS regression are as follows:
LN(SEVAL)
=
0.844
+0.062 LN(SEELC)
+0.910 LN(SEVAL -1)
(2.969)*
(2.038)*
(25.132)*
R2 = 0.998, D-W = 2.027, Degrees of Freedom = 37
*significant at 5 per cent level
177
TECHNICAL ANNEXURE-III
Econometric analysis has been undertaken to determine the relationship between GDP growth
per capita and growth of infrastructure (including power generation) and basic services (like
education)
The following variables are designated:
GGDP = Growth rate of real GDP per capita
GAG = growth rate of agriculture
r = nominal interest rate
GSCH = growth rate of mean years of schooling
GWAT = growth rate of water availability (in agriculture)
Results of the OLS regression are as follows:
GGDP = 2.946
(2.447)*
+ 0.32 GAG - 0.252r
(6.357)*
(-2.746)*
+0.304 GSCH
(2.530)*
+0.167 GELE
(5.193)*
+0.325 GWAT
(4.592)*
R2 = 0.726, D-W = 2.646, Degrees of freedom= 36
*Significant at 5 per cent level
The analysis is undertaken for the period, 1975-76 to 2010-11.
The Wu-Hausman test was performed to test for endogeniety.
The equation passed the test.
It appears that a 1 per cent increase in electricity generation per capita raises the GDP per
capita growth rate by 0.167 per cent.
178
TECHNICAL ANNEXURE IV
ANALYSIS OF WILLINGNESS TO PAY
(Aggregate)
The demand function given by the function
𝐴 1
𝑞 = 𝐴𝑝−∝  p = (𝑞 )∝
Suppose that in the absence of outages the demand (unconstrained) would have been 25%
higher
Therefore,
102799 = 𝐴(13.00)−0.13
Where the price elasticity of demand is taken as -0.13, as derived by Jamil and Ahmed [2010]
From (2)
𝐴 = 102799(13.00)0.13 = 102799(1.396) = 143507
When 𝑞 = 77099
𝑝0.13 =
143507
= 119.00
77099
Area under the demand curve is given by
102799 𝐴
Area = ∫77099 (𝑞 )7.692 𝑑𝑞
102799
Area = 4.642 × 1039 ∫77099 𝑞 −7.962 𝑑𝑞
Area =
4.642×1039
[−(77099)−6.692
6.692
+ (102799)−6.692 ]
Area = − 0.694 × 1039 [−1.976 × 10−33 + 2.882 × 10−34 ]
Area = − 0.694 × 106 [−1.976 + 0.288]
Area = 0.694 × 106 × 1.688 = 1.171 × 106
Consumer Surplus = 1.171 × 106 − 0.334 × 106
Therefore, Consumer Surplus = . 837 × 1012 𝑅𝑠 because consumption is in Gwh.
As such, CS = Rs 837 billion.
179
TECHNICAL ANNEXURE V
FIRM BEHAVIOR IN THE PRESENCE OF OUTAGES
We make the assumption that the firm is ‘small’ and competitive Input and Output Markets. We
designate the following variables:
p = exogenously given price of output
w = wage rate
r = cost of capital
The firm is a profit maximising agent
We also have variables related to outages as follows:
𝜀 = proportion of time lost due to outages
𝜃 = proportion recovery through adjustments
CA = Cost of Adjustments
m = consumption of electricity per unit of time
𝜋 = pf(K,L) [1 – 𝜀 + 𝜃𝜀 ] – wL – rK- CA (𝜃 ∈ 𝑚)
B
𝜕𝐹
𝜕𝜋
𝜕𝐾
=p
𝜕𝜋
𝜕𝐿
=p
𝜕𝐾
………………………………………………………………………..(1)
.B – r = 0
𝜕𝐹
𝜕𝐾
………………………………………………………………………….(2)
.B – w = 0
q = f(K,L)B
𝜕𝜋
𝜕𝜃
= p.q.𝜀 -
= pq -
𝜕𝐴𝐶
𝜕𝐾𝑊
Therefore,
𝜕𝐴𝐶
𝜕𝐾𝑊
. (m 𝜀) = 0
.𝑚 = 0
…………………………………………………...(3)
𝑞
𝜕𝐴𝐶
𝑚
𝜕𝐾𝑊
p.[ ] -
=0
As such, the probability of
i.
ii.
iii.
𝜕𝜋
𝜕𝜃𝜃=0
> 0 depends on the following factors:
How high p is. That is the extent to which the market situation is favorable for the firm
The electricity intensity of the firm, the lower the intensity the greater the likelihood of the
firm making an adjustment
The lower the marginal costs of adjustments by the firm to recover output lost.
It is also likely that in the long run the firm uses less capital and less labor if B < 1.
180
TECHNICAL ANNEXURE VI
FIRM BEHAVIOR IN THE PRESENCE OF OUTAGES
We make the assumption that the firm is ‘small’ and competitive Input and Output Markets. We
designate the following variables:
p = exogenously given price of output
w = wage rate
r = cost of capital
The firm is a profit maximising agent
We also have variables related to outages as follows:
𝜀 = proportion of time lost due to outages
𝜃 = proportion recovery through adjustments
CA = Cost of Adjustments
m = consumption of electricity per unit of time
𝜋 = pf(K,L) [1 – 𝜀 + 𝜃𝜀 ] – wL – rK- CA (𝜃 ∈ 𝑚)
B
𝜕𝜋
𝜕𝐾
=p
𝜕𝐹
𝜕𝐾
𝜕𝜋
𝜕𝐿
=p
……………………………………………..(1)
.B – r = 0
𝜕𝐹
𝜕𝐾
…….……………………………………….(2)
.B – w = 0
q = f(K,L)B
𝜕𝜋
𝜕𝜃
= p.q.𝜀 -
= pq -
𝜕𝐴𝐶
𝜕𝐾𝑊
Therefore,
𝜕𝐴𝐶
𝜕𝐾𝑊
. (m 𝜀) = 0
.𝑚 = 0
………….…………………………………...(3)
𝑞
𝜕𝐴𝐶
𝑚
𝜕𝐾𝑊
p.[ ] -
=0
As such, the probability of
i.
ii.
iii.
𝜕𝜋
𝜕𝜃𝜃=0
> 0 depends on the following factors:
How high p is. That is the extent to which the market situation is favorable for the firm
The electricity intensity of the firm, the lower the intensity the greater the likelihood of the
firm making an adjustment
The lower the marginal costs of adjustments by the firm to recover output lost.
It is also likely that in the long run the firm uses less capital and less labor if B < 1.
181
REFERENCES
1. AF Adenikinju, ‘Analysis of the Cost of infrastructure failures in developing countries: The
case of the electricity sector in Nigeria.’
2. AP Sanghvi, ‘Economic Cost of Electricity supply interruptions: US and foreign experience’,
Energy Economics, 1982.
3. B.Bental, ‘A simple method for evaluating the marginal cost of unsupplied electricity’, The
Bell Journal Economics, 1982, JSTOR
4. Beenstock, M., E. Goldin and Y. Haitovsky, ‘The Cost of Power Outages in the Business
and Public Sectors in Israel: Revealed Preferences vs Subjective Valuation’, The Energy
Journal, 1997, Volume 18, Number 2.
5. CKW Train, ‘The cost of electric power interruptions to commercial firms’, The Energy
Journal, 1988.
6. Douglas W.Caves, Joseph A.Herriges and Robert J. Windle, ‘Customer’ Demand for
services reliability in the electric power industry: A Synthesis of the Outage Cost Literature’,
April 1990, Bulletin of Economic Research, Vol 42, Issue 2, pages 79-121.
7. F. Jamil and E. Ahmed’, ‘The Relationship between Electricity Consumption, Electricity
Prices and GDP in Pakistan’ Energy Policy, Vol 138, Issue 10, 2010 pg 6016 - 6025.
8. Hafiz A. Pasha, Aisha Ghaus and Salman Malik, ‘The Economic Cost of Power Outages in
the Industrial Sector of Pakistan’, Energy Economics, Vol 11, Issue 4, October 1989, pg
301-318.
9. Hafiz A. Pasha, Aisha Ghaus and Salman Malik’, ‘The Differential Impact of Outages in the
Industrial sector of Pakisan, Pakistan Journal of Applied Economics, 1990.
10. M. Breenstock, ‘Generators and the cost of electricity outages’, Energy Economics, 1991,
Vol 13, Issues 4, October, Pages 283-289.
11. M.J.Doane, R.S.Hartman’, ‘Household preferences for interruptible rate options and the
revealed value of services reliability, The Energy Journal, 1988.
12. M.Munasinghe’, ‘Costs incurred by residential electric consumers due to power failures’,
Journal of Consumer Research, 1980, JSTOR.
13. MAB Hegazy, ‘The economic costs of electricity shortages: a case study of Egypt’, The
Energy Journal, 1988.
14. N.Kaseke, ‘Cost of Power Outages in the Zimbab wean Agricultural Sectoral’, Journal of
Strategic Studies, 2011, Vol 2, No 1, pg 1-31
15. OA Unchendu, ‘Economic Cost of electricity ontages: Evidence from a sample study of
industrial and commercial firms in the Logos Area of Nigeria’, CBN Economic and Financial
Review, 1993.
182
16. R.Jyoti and A.Ozbafli’, ‘The Opportunity Cost of Electricity Outages and Privatisation of SubStations in Nepal’, 2006.
17. Rehana Siddiqui, H.H. Jamil, M. Nasir, W.S. Malik and M. Khalid’, ‘The Cost of un served
Energy: Evidence from Selected Industrial Cities of Pakistan’, PIDE Working Paper No 75,
2011.
18. RW Gilmer’, ‘The Cost of residential electric power outage’, ‘The Energy Journal, 1988.
19. Sebastian Bernstein and Renata Agurto, ‘Use of outage cost for electricity pricing in Chile’,
Utilities Policy, Vol 2, Issues 4, October 1992, Pages 299-302 (website: agriculture
outages).
20. U.B.Akuru and O.L.Okoro, ‘Economic Implications of Constant Power Outages on SMEs in
Nigeria’.
183