INCOME POVERTY AND STATE SUPPORT IN TRINIDAD AND TOBAGO:

advertisement
DEPARTMENT OF ECONOMICS
CONFERENCE ON THE ECONOMY (COTE)
October 8th – 9th 2008
INCOME POVERTY AND STATE
SUPPORT IN TRINIDAD AND TOBAGO:
THE CASE FOR ALIGNMENT
Karl Theodore and Ewan Scott
Labour Market and Poverty Studies Unit
Department of Economics
UWI
October 9, 2008
OBJECTIVE
The primary objective is threefold.
• First we would like to determine whether the
poverty situation in Trinidad and Tobago has
trended upwards or downwards since it was
last officially measured in 2005.
• Second we would like to determine the extent
of the support given to low income
households by the state.
• Finally, putting the two sets of results
together, we will open the case for developing
a new poverty measure for Trinidad and
Tobago.
2
CONTEXT
• The buoyancy of the economy is without
question: per capita real income doubling
between 1998 and 2008 (99.6%)
• Continuation of an upward trend for past decade
is prima facie evidence that generally living
conditions in the country are improving.
• Evidence from other places shows that even
when average incomes are increasing there might
be no change whatsoever in the poverty rate
• Same for T&T?
3
Quotable quotes
For the USA we read that
• ”the inflation-adjusted incomes of median households
rose by 1.3% in 2007 from the previous year, from
$49,568 in 2006 to $50,233 in 2007 (2007 dollars),
while the overall poverty rate increased slightly, from
12.3% to 12.5%.” Jared Bernstein(2008)
For New York City in particular, it was found that
• “median income rose slightly from 2005 to 2006, but
the number of people living in poverty also increased.
The poverty rate remained stubbornly resistant at
about one person in five…” Sam Roberts, New York Times, August 28,
2007,
4
Study Approach
• Poverty estimates for Trinidad and Tobago have
traditionally come from the Household Budget
Surveys (HBS) done by the CSO or, more recently,
on Surveys of Living Conditions (SLC).
• Since up-to-date HBS data are not available and
since the last SLC was done in 2005, the attempt
to determine poverty trend since 2005 will be
based on the CSSP data produced by the CSO.
• Thanks to CSO for providing us with the datasets
which made the poverty estimation possible.
5
LIMITATIONS
There are at least five sources of overestimation of poverty levels
using the CSSP data:
• To the extent that some households may have no members
employed and hence no employment income, at the outset such
households will be categorized as poor.
• Income is thought to have a larger transitory component than
consumption making it a less reliable indicator of welfare
potential;
• Income sources are usually more diverse than consumption,
highlighting the need for a wider range of survey questions, and
making income a more difficult variable to measure;
• It is a widely accepted claim that incomes tend to be underreported by respondents in a survey.
• In addition, there may be instances of non-reporting of
employment income by members in some households thereby
lowering the total income for such households.
6
METHODOLOGY
• Using the CSSP data, the procedure for arriving at an estimate of
the poverty level for 2008 is necessarily different from that used
in respect of the SLC data.
• Given the absence of consumption expenditure data in the CSSP,
and thus inability to compute Engel coefficients, computation of
the poverty line is an adaptation. The 2008 poverty line was
arrived at by inflating the 2005 SLC poverty line by the year-onyear changes in the CPI. By this procedure, the 2008 poverty line
was found to be TT$815 per adult per month.
• As a basis for comparison, CSSP poverty estimates were
computed by applying the SLC 2005 poverty line of $665 to the
2005 CSSP data.
• A household was deemed to be poor if its per capita
employment income in adult equivalence fell below the poverty
line.
7
RESULTS
• Shown in the diagram are the two different sets of poverty
estimates for Trinidad and Tobago over the period 1997 to
2008: CSSP in blue and SLC in red. The data are shown in the
Table below.
8
Poverty estimates
CSSP data
(quarters, 2,3)
1997
2005
2008*
CSSP Headcount
Index
%
29.25
22.95
18.00
SLC Headcount
Index
%
24
16.7
-
•For 2008 only first quarter data are available
9
COMMENT
• What we see in the diagram is that although the CSSP
poverty estimate is consistently higher than the SLC
estimate, as expected, the trend in both estimates is
almost exactly the same between 1997 and 2005. In
both cases poverty is shown to be falling at almost
the same rate. We note that according to the CSSP, in
the period 2005 to 2008 poverty fell at a faster rate
than before. If the SLC estimate for 2008 follows the
same pattern we will expect the 2008 SLC estimate,
when it is produced, not only to be lower than its 2005
counterpart of 16.7% but also to show a speeding up in
the decline of poverty.
10
STATE SUPPORT OF LOW INCOME
HOUSEHOLDS
• In the Table below we present a summary of the
monthly value of the support that is made available
to the poorer households of the country. Since the
data on some grants and subsidies were not available
in a form to be readily converted into monthly values
these grants and subsidies have been omitted. The
figures presented are therefore underestimates of
the true level of support. The information is
presented on the basis of the composition of lowincome households in the country. For present
purposes we identify five household categories.
11
Composition of Five 5-person Households (Selected
from SLC Combinations)
• Household Type 1: 1 adult (unemployed)
4 children (2 primary; 2 secondary)
• Household Type 2: 1 adult (employed) 4 children
(2 primary; 2 secondary, including 1
disabled)
• Household Type 3: 2 adults (employed)
•
3 children (2 primary; 1 secondary)
• Household Type 4: 3 adults (1 employed; 1 unemployed; 1
elderly) 2 children (1 primary; 1 secondary)
• Household Type 5: 4 adults (1 employed; 1 unemployed; 1
elderly; 1 disabled) 1 child (1 secondary)
12
MONEY VALUE OF BENEFITS (MONTHLY) FROM SELECTED SUPPORT
INSTRUMENTS, BY SELECTED HOUSEHOLD COMBINATIONS (1)
2008 POVERTY LINE = $815
No.
1.
2.
3.
4.
5.
6.
7.
Support
Instruments
Selected
Grants &
Subsidies:
Education
Grant
Textbook
Rental/Loan
School
Transportation
Schools
Nutrition
Prog.
CDAP
Food Debit
Card
Old Age
Pension /
Senior
Citizens’
Grant
H/Hold 1
H/Hold 2
H/Hold 3
H/Hold 4
H/Hold 5
$0
0<$ 500
< $1,500
<$ 1,000
< $1,000
560
560
420
280
140
696
696
696
348
232
756
756
567
378
189
1,304
1,304
652
652
326
52.30
52.30
52.30
104.60
52.30
400
400
400
400
400
-
-
-
1,650
1,650
13
MONEY VALUE OF BENEFITS (MONTHLY) FROM SELECTED SUPPORT
INSTRUMENTS, BY SELECTED HOUSEHOLD COMBINATIONS (2)
2008 POVERTY LINE = $815
No.
Support
Instruments
10.
11.
Type 1 benefit
9.
H/Hold 2
H/Hold 3
H/Hold 4
H/Hold 5
Monthly Income ($TT)
Selected
Grants &
Subsidies:
Public
Assistance
Prog
Disability
Assistance Pro
Fuel Subsidy
8.
H/Hold 1
Total
monthly
benefits
($TT)
ANNUAL
TOTAL
SUPPORT
($TT)
$0
$0< 500
<$ 1,500
< $1,000
< $1,000
650
650
-
650
650
-
1,100
-
-
1,100
36
36
36
36
36
800
800
800
800
800
5,254.3
6,354.3
3,623.3
5,298.6
5,575.3
63,051.6
76,251.6
43,479.6
63,583.2
66,903.6
14
CONCLUDING COMMENT
Three conclusions seem to be warranted:
1. In spite of its upward bias the CSSP poverty estimate
faithfully mirrors the downward trend in the SLC
estimate
2. The money value of available income support for lowincome households vastly exceeds the poverty line
3. Traditional poverty estimates which exclude the
impact of the income support will tend to overstate
the poverty levels in the country. The prima facie case
for a new measure of poverty in T&T seems to be a
strong one.
15
Download