TRAINING ON FERTILIZER
DEMAND FORECASTING
Patrick Heffer, IFA
19 April 2013
Moscow, Russia
What Is IFA?
IFA in Brief
 Non-governmental organization representing the world
fertilizer industry
 About 540 members in 85 countries( including Uralkali
and many other Russian fertilizer producers)
 Based in Paris
 Organized around three standing committees:
o Technical
o Production and International Trade
o Agriculture

Main activities: market analysis, issue management/
advocacy, conference organization
IFA Agriculture Committee
Four objectives:
 Address issues facing fertilizer
demand
 Promote the efficient, balanced
and responsible use of
fertilizers
 Improve the public image of
fertilizers
 Develop reliable and authoritative
fertilizer demand forecasts
IFA Agriculture Committee
Develop reliable and authoritative fertilizer demand
forecasts
 Reports on demand forecasts twice a year
 Annual Conference (May/June)
medium-term (5-year) forecasts
 Enlarged Council Meeting (Nov/Dec)
short-term (1-year) forecasts
 Improving the forecasts
 Develop network of correspondents
 Guidelines for a crop-based approach
 Regional training programmes
IFA’s Fertilizer Demand Forecasts
Methodology and Information
Used by IFA
Current strategy
 Rely on information provided by a network of correspondents
in ~50 countries
 Heterogeneous inputs
essentially due to different methodologies used depending on
the countries
 Trend forecasts
 Government objectives
 Recommended application rates
 Econometric model (e.g. India, Brazil)
 Crop-based / expert-based model (e.g. EU, USA)
Methodology and Information
Used by IFA
Correspondents are not enough
And checks are needed
 Use additional sources of info
(reports, articles…) on:
 Economic context, weather,
policy factors
 Global agricultural situation
Fertilizer demand forecasts
are cross-checked with forecasts
on the supply side
 Check consistency between
national forecasts and the global
scenario
 To issue forecasts on the
supply/demand balances
 To ensure consistency
between the two sets
Forecasts provided by the correspondents are revised in
more than half of the countries !
They are almost always revised down.
Importance of Good National
Forecasts
National forecasts are the starting point: They provide
the foundation for developing regional and global forecasts
Objective
High margin of error
Mostly over-estimation
Small margin of error
Balanced fluctuations
around actual demand
IFA Training Programme on
Fertilizer Demand Forecasts
Objective - Improve national forecasts, which should result
in turn in better regional and global projections
 Developed guidelines for a crop-based,
expert-based forecast
 On-the-ground training programmes
Crop-Based, Expert-Based Forecast
Why Is It the Preferred Methodology?
What is a Good Forecast?
….. the most likely scenario
 Independent from:
 Commercial pressures
 Governmental objectives
 Fertilizer recommendations
 As realistic as possible
 Not too optimistic  over-investments
 Not too pessimistic  would endanger food
security
 Often more conservative than governmental targets
Types of Forecast Methodologies
 Trend analysis
 Growth rate models
 Production/trade models
 Econometric models
 Crop-based, expert-based models
Why Does IFA Recommend Using a
Crop-Based, Expert-Based Approach?
 Can be used on a consistent basis across all the
countries and regions
 Yields much more accurate forecasts than the trend
analysis, growth rate models and production/trade
models
 Does not require large databases and knowledge in
econometrics as econometric models. Also, more
accurate than econometric models
 Explains where changes in demand are anticipated
to come from
The Four Stages of
a Crop-Based, Expert-Based Forecast
The Four Stages
Our Goal: “To Arrive at a Realistic, Objective,
and Defensible Forecast”
Main Steps:
I. Create a Historical Database or ‘Base Year’
II. Develop a Qualitative Scenario: Outlook
Conditions
III. Prepare the Quantitative Forecast
IV. Validate the Forecast
I. The Historical Database
Overview
Consists of Data for 3 Forecast Components:
1. Area Planted to Major Crops
2. Percent of Planted Area Fertilized by Nutrient
and Crop Type
3. Average Application Rates of Nutrients by
Crop Type
I. The Historical Database
Data Collection
1. Area Planted to Major Crops
Identify the Major Nutrient Consuming Crops
Develop a Database:
• Major Crops
• Other Crops
Potential Problems
• Only Harvested Area Data is Available
• Data Only Available for Some/Few Crops
Nutrient Use – Major U.S. Crops
U.S. Nutrient Use by Crop, FY2004/05
50
45
44
43
41
40
68%
76%
NITROGEN
PHOSPHATE
35
73%
POTASH
25
19
20
17
16
14
15
10
6
5
5
5
5
2
Cotton
Soybeans
Wheat
Corn
Cotton
Soybeans
Wheat
Corn
Cotton
Soybeans
Wheat
0
Corn
percent
30
Acres Planted – Major U.S. Crops
100,000
90,000
80,000
70,000
Thousand Acres
60,000
50,000
40,000
30,000
20,000
10,000
0
1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008
Corn
Wheat
Soybeans
Cotton
Share of Total Fertilizer Consumption
by Crop Type in 2010 in Russia
wheat
+ barley
+ maize
+ sunflower
+ sugar beet
= 75-80%
Source: IFA
I. The Historical Database
Data Collection
2. Percent of Planted Area Fertilized
• Identify Area Fertilized for Each Major Crop
• Express as Percent of Planted Area
3. Average Application Rates
• Determine Amount of Each Nutrient Applied to
Crop
• Express in kg/ha
Potential Problems
• Only Some Data Exists
1984
N
P2O5
K2O
2004
2002
2000
1998
1996
1994
1992
1990
1988
1986
1982
1980
1978
1976
1974
1972
1970
1968
1966
1964
Percent
Percent of Corn Acres Treated
by Nutrient
100
90
80
70
60
50
40
30
20
10
0
1984
N
P2O5
K2O
2004
2002
2000
1998
1996
1994
1992
1990
1988
1986
1982
1980
1978
1976
1974
1972
1970
1968
1966
1964
lbs. per treated acre
Application Rate per Treated
Corn Acre
140
120
100
80
60
40
20
0
1984
N
P2O5
K2O
2004
2002
2000
1998
1996
1994
1992
1990
1988
1986
1982
1980
1978
1976
1974
1972
1970
1968
1966
1964
lbs. per acre
Average Nutrient Application
Rate - Corn
140
120
100
80
60
40
20
0
II. The Qualitative Scenario
“What Will the Future Look Like?”
Develop Assumptions About Conditions That Will
Impact Crop Acreage and Nutrient Application
Consider
• Domestic and World Economies
• Crop Market Conditions
• Fertilizer Market Conditions
• The Regulatory Environment
• Infrastructure Investments or Limitations
Avoid
• Government/Industry Targets
III. The Quantitative Forecast
Applying Your Qualitative Conditions
“For Each Crop, Consider Impact of Stage II
Assumptions on Each Forecast Component”:
Area Planted
Impacted by:
• Crop Prices
• Weather
• Government Policies
• Changes to Demand
• Infrastructure
III. The Quantitative Forecast
Applying Your Qualitative Conditions
“For Each Crop, Consider Impact of Stage II
Assumptions on Each Forecast Component”:
Percent of Area Fertilized and Application Rates
Impacted by:
• Crop Prices and Yields
• Fertilizer Prices and Availability
• Soil Type, Moisture, and Weather
• Policy Changes
IV. Forecast Validation
Recall Our Goal: “To Arrive at a Realistic,
Objective, and Defensible Forecast”
The Final Step: Verify That the Forecast Meets These
Criteria
1. Cross Check Results
2. Seek Out Supporting Data
3. Compare to Other Methodologies
• Trend Forecast
• Econometric Model if Available
Once Comfortable With the Forecast, Explain Results
Concluding Remarks
Maintaining Your Forecast
1. Update Your Stage II Assumptions Regularly
2. Anticipate/Explain Structural Changes
• Historical Data
• Forecast
3. Compare Results With Others
4. Rely On and Trust Your Own Expertise
Your Forecast Will Improve Over Time
• Progressive Increase of Crop Coverage
• Progressive Improvement of Data Quality
Crop-Based, Expert-Based Forecast
in Summary
Our goal: “to arrive at a realistic, objective,
and defensible forecast”
3 variables
Main steps:
1. Area planted to major
I. Create a historical database
crops
or ‘base year’
2. Percent of planted
area fertilized by
II. Develop a qualitative scenario:
nutrient and crop
outlook conditions
type
III. Prepare the quantitative
3. Average application
forecast
rates of nutrients by
crop type
IV. Validate the forecast
Outlook for World and
Regional Fertilizer Demand
Note regarding data presented:
Short-term outlook of December 2012
Medium-term Outlook of June 2012
World Cereal Production
and Utilization (Mt)
Balance forecast
at the end of the
2012/13 campaign:
 FAO: -30 Mt
 USDA: -45 Mt
 IGC: -45 Mt
Source: FAO
World Cereal Stock-to-Use Ratio
Anticipated ending stock evolution in
2012/13:
Wheat:
 -23 Mt (-12%)
 Lowest S/U ratio since 2007/08;
2nd lowest since 1980
 Major exporters stock-todisappearance: 13.9%
Coarse grains:
 -13 Mt (-8%)
 Lowest S/U ratio since 1980
 Major exporters stock-todisappearance: 8.9%;
below 6% in the US
Source: FAO
Rice:
 Increase for 6th consecutive year
Relative Evolution of
Agricultural Commodity Prices
Relative Price Evolution from January 2006 to October 2012
Rice
Wheat
Maize
Maize
Soybean
Soybean
Wheat
Rice
Sources: Financial Times and IMF
Global Fertilizer Demand
Short-term Forecasts (Mt nutrients)
Change
2011/12
2012/13
2013/14
N
+3.0%
+1.5%
+1.5%
P2O5
+1.2%
-2.7%
+3.5%
K2O
+2.1%
+0.1%
+4.5%
Total
+2.4%
+0.3%
+2.4%
Source: IFA Agriculture
Reminder on reference years:
~1/2 of the world market: January X  December X
~1/6 of the world market: April X  March X+1
~2/6 of the world market: July X  June X+1
Global Fertilizer Demand
Short-term Forecasts (Mt nutrients)
Variation in 2011/12
Variation in 2012/13
Variation in 2013/14
0.0
K20
0.6
1.3
P205
-1.1 0.5
1.4
N
3.2
-2
-1
0
1
1.6
2
3
4
1.6
5
6
Source: IFA Agriculture
7
Regional N Fertilizer Demand
Short-term Forecast (Mt N)
East Asia
South Asia
North America
Lat. Am. & Carib.
W. & C. Europe
E. Eur. & C. Asia
Africa
Variation in 2011/12
Variation in 2012/13
Variation in 2013/14
West Asia
Oceania
-1
0
1
2
Source: IFA Agriculture
3
Regional P Fertilizer Demand
Short-term Forecast (Mt P2O5)
East Asia
South Asia
North America
Lat. Am. & Carib.
W. & C. Europe
E. Eur. & C. Asia
Africa
Variation in 2011/12
Variation in 2012/13
Variation in 2013/14
West Asia
Oceania
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
Source: IFA Agriculture
Regional K Fertilizer Demand
Short-term Forecast (Mt K2O)
East Asia
South Asia
North America
Lat. Am. & Carib.
W. & C. Europe
E. Eur. & C. Asia
Africa
Variation in 2011/12
Variation in 2012/13
Variation in 2013/14
West Asia
Oceania
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
Source: IFA Agriculture
2013/14 vs. 2007/08 (Mt nutrients)
N
P2O5
K2O
East Asia
South Asia
North America
Lat. Am. & Carib.
W. & C. Europe
E. Eur. & C. Asia
Africa
West Asia
Oceania
-3 -2 -1
49% of global
net increase
0
1
2
3
4
5
6
7
8
9 10 11
Source: IFA Agriculture
Medium-Term Outlook for
Agricultural Commodity Prices
World Nominal Prices (US$/t)
Source: OECD-FAO
Global Fertilizer Demand
Medium-term Outlook (Mt nutrients)
Base year
Average Annual Change
Base Year  2016/17
N
+1.5% p.a.
P2O5
+2.3% p.a.
K2O
+3.7% p.a.
Total
+2.1% p.a.
Source: IFA Agriculture
Regional Fertilizer Consumption
Medium-term Outlook (Mt nutrients)
East Asia
30%
South Asia
24%
North America
Lat. Am. & Carib.
21%
W. & C. Europe
E. Eur. & C. Asia
6%
West Asia
Av. 2009/10 to 2011/12
Africa
Variation in 2016/17
Oceania
0
10
20
30
40
50
60
70
Source: IFA Agriculture
80
for more information:
www.fertilizer.org