[Analytical Perspectives]
[Other Technical Presentations]
[18. Comparison of Actual to Estimated Totals]
[From the U.S. Government Publishing Office, www.gpo.gov]
[[Page 373]]
18. COMPARISON OF ACTUAL TO ESTIMATED TOTALS
In successive budgets, the Administration publishes several estimates
of the surplus for a particular fiscal year. Initially, the year appears
as an outyear estimate at the end of the budget horizon. In each
subsequent budget, the year advances in the estimating horizon until it
becomes the ``budget year.'' One year later, the year becomes the
``current year'' then in progress, and the following year, it becomes
the just-completed ``actual year.''
The budget is legally required to compare budget year estimates of
receipts and outlays with the subsequent actual receipts and outlays for
that year.\1\ Part I of this chapter meets that requirement by comparing
the actual results for 2001 with the current services estimates shown in
the 2001 Budget published in February 2000.
---------------------------------------------------------------------------
\1\ These requirements, for receipts and ``uncontrollable outlays,''
are in 31 USC 1105(a)(18) through (20).
---------------------------------------------------------------------------
Part II of the chapter presents a broader comparison of estimates and
actuals. This part first discusses the historical record of budget year
estimates versus actuals over the last two decades. Second, it broadens
the focus to estimates made for each year of the budget horizon,
extending four years beyond the budget year. This broader focus shows
the growth in differences between estimates and the eventual actual
results as the estimates extend further into the future.
PART I: COMPARISON OF ACTUAL TO ESTIMATED TOTALS FOR 2001
This part of the chapter compares the actual receipts, outlays, and
surplus for 2001 with the current services estimates \2\ shown in the
2001 Budget published in February 2000. This part also presents a more
detailed comparison for mandatory and related programs, and reconciles
the actual receipts, outlays, and surplus totals shown here with the
figures for 2001 previously published by the Department of the Treasury.
---------------------------------------------------------------------------
\2\ The current services concept is discussed in Chapter 15: ``Current
Services Estimates.'' For mandatory programs and receipts the February
2000 current services estimate is based on laws then in place. For
discretionary programs the current services estimate is based on enacted
appropriations adjusted for inflation.
---------------------------------------------------------------------------
Receipts
Receipts in 2001 were $1,991.0 billion, which is $18.9 billion less
than the current services estimate of $2,009.9 billion in the 2001
Budget. As shown in Table 18-1, this shortfall was the net effect of
legislative and administrative changes; economic conditions that
differed from what had been expected; and technical factors that
resulted in different collection patterns and effective tax rates than
had been assumed.
Table 18-1. COMPARISON OF ACTUAL 2001 RECEIPTS WITH THE INITIAL CURRENT SERVICES ESTIMATES
(In billions of dollars)
----------------------------------------------------------------------------------------------------------------
Enacted
Feb. 2000 legislation/ Different Technical Net
estimate administrative economic factors change Actual
actions conditions
----------------------------------------------------------------------------------------------------------------
Individual income taxes................. 978.2 -40.6 17.0 39.7 16.1 994.3
Corporation income taxes................ 189.6 -31.9 8.2 -14.8 -38.5 151.1
Social insurance and retirement receipts 682.5 -0.4 6.7 5.2 11.5 694.0
Excise taxes............................ 69.4 -0.5 1.0 -3.8 -3.3 66.1
Estate and gift taxes................... 32.0 .............. 0.3 -3.9 -3.6 28.4
Customs duties.......................... 22.2 -0.7 -0.9 -1.2 -2.8 19.4
Miscellaneous receipts.................. 36.0 .............. 0.8 1.0 1.8 37.8
-----------------------------------------------------------------------
Total................................ 2,009.9 -74.2 33.0 22.3 -18.9 1,991.0
----------------------------------------------------------------------------------------------------------------
[[Page 374]]
Policy differences.--The Economic Growth and Tax Relief Reconciliation
Act of 2001 reduced 2001 receipts by $68.1 billion (see Chapter 4:
``Federal Receipts'' for a description of this Act). Other legislative
and administrative changes, including the extension of filing deadlines
for taxpayers adversely affected by the terrorist attacks of September
11, 2001, reduced 2001 receipts relative to the February 2000 current
services estimate by an additional $6.1 billion.
Economic differences.--Differences between the economic assumptions
upon which the current services estimates were made and actual economic
performance accounted for a net increase in 2001 receipts of $33.0
billion.\3\ Higher-than-anticipated wages and salaries and other sources
of personal income were in large part responsible for the increases in
individual income taxes and social insurance and retirement receipts of
$17.0 billion and $6.7 billion, respectively. Increased corporation
income taxes, attributable to higher-than-expected corporate profits,
increased 2001 receipts by an additional $8.2 billion relative to the
February 2000 estimate. Higher-than-estimated levels of gross domestic
product (GDP), which affect excise taxes, and higher-than-expected
interest rates, which affect deposits of earnings by the Federal Reserve
(miscellaneous receipts), increased receipts above the budget estimates
by an additional $1.0 billion and $0.8 billion, respectively. Customs
duties were $0.9 billion below the budget estimate, reflecting lower-
than-expected imports.
---------------------------------------------------------------------------
\3\ Changes in economic assumptions between the 2001 and 2002 Budgets,
reflecting improvements in the economic outlook over that period,
increased the estimate of receipts by $54.2 billion. This improvement
was offset by weaker-than-expected economic outcomes subsequent to the
2002 budget, which reduced receipts by $21.2 billion.
---------------------------------------------------------------------------
Technical reestimates.--Technical factors increased 2001 receipts a
net $22.3 billion above the February 2000 current services estimate.
This net increase was attributable to higher-than-anticipated
collections of individual income taxes, social insurance and retirement
receipts, and miscellaneous receipts, which were partially offset by
lower-than-anticipated collections of corporation income taxes and other
sources of receipts. Higher effective tax rates on personal income than
estimated in February 2000, and the effect of the stock market on
capital gains, were primarily responsible for the net increase in
individual income taxes of $39.7 billion. Higher-than-expected
collections of payroll taxes, attributable in large part to adjustments
for prior year receipts, partially offset by lower-than-expected
unemployment insurance receipts, increased social insurance and
retirement receipts a net $5.2 billion above the budget estimate.
Different collection patterns and effective tax rates than assumed in
February 2000 were primarily responsible for the lower-than-anticipated
collections of corporation income taxes of $14.8 billion.
Outlays
Outlays for 2001 were $1,863.9 billion. This was $25.2 billion more
than the $1,838.8 billion current services estimate in the 2001 Budget
(February 2000).
Table 18-2 distributes the $25.2 billion net increase in outlays among
discretionary and mandatory programs and net interest.\4\ The table also
makes rough estimates according to three reasons for the changes:
policy; economic conditions; and technical estimating differences, a
residual.
---------------------------------------------------------------------------
\4\ Discretionary programs are controlled by annual appropriations,
while mandatory programs are generally controlled by authorizing
legislation. Mandatory programs are mostly formula benefit or
entitlement programs with permanent spending authority that depend on
eligibility criteria, benefit levels, and other factors.
Table 18-2. COMPARISON OF ACTUAL 2001 OUTLAYS WITH THE INITIAL CURRENT SERVICES ESTIMATES
(In billions of dollars)
----------------------------------------------------------------------------------------------------------------
Current Changes
Services -----------------------------------------
(Feb. Total Actual \1\
2000) Policy Economic Technical changes
----------------------------------------------------------------------------------------------------------------
Discretionary:...................................
Defense....................................... 295.2 7.6 ........ 3.4 10.9 306.1
Nondefense.................................... 340.3 10.4 ........ -7.5 2.9 343.3
--------------------------------------------------------------
Subtotal, discretionary..................... 635.5 18.0 ........ -4.1 13.8 649.3
Mandatory:
Social Security............................... 422.2 5.5 3.4 -1.7 7.2 429.4
Other programs................................ 573.0 15.2 1.9 -11.0 6.0 579.0
--------------------------------------------------------------
Subtotal, mandatory......................... 995.2 20.6 5.2 -12.7 13.2 1,008.4
Net interest..................................... 208.1 3.1 -4.4 -0.6 -1.9 206.2
--------------------------------------------------------------
Total outlays................................... 1,838.8 41.7 0.9 -17.4 25.2 1,863.9
----------------------------------------------------------------------------------------------------------------
\1\ Actuals do not reflect the accrual funding of Federal retiree costs in the 2003 Budget.
Policy changes are the result of legislative actions that change
spending levels, primarily through higher or lower appropriations or
changes in authorizing legislation. For 2001, policy changes increased
outlays an estimated $41.7 billion relative to the initial current
services estimates.
Policy changes increased discretionary outlays by $18.0 billion,
because outlays from final appropriations were above the initial current
services estimates. Defense discretionary outlays increased by $7.6
billion and nondefense discretionary outlays increased by $10.4 billion.
Policy changes increased mandatory outlays by $20.6 billion above
current law. The largest changes
[[Page 375]]
were an increase of $10.8 billion for agricultural programs, mostly for
emergency assistance; an increase of $5.5 billion for Social Security
benefit payments from repealing the Social Security earnings test and
correcting a Consumer Price Index error; and an increase of $4.8 billion
for expansion of Medicare benefits.
Economic conditions that differed from those forecast in February 2000
resulted in a net increase in outlays of $0.9 billion. Outlays for
mandatory programs increased an estimated $5.2 billion, largely due to a
higher-than-expected rise in prices, and consequently in cost-of-living-
adjustments (COLAs), which increased outlays for Social Security benefit
payments, refundable Earned Income Tax Credits, Medicare, and Federal
employee retirement benefits. Lower-than-expected unemployment
contributed to reduced outlays for Food Stamps, which partially offset
the outlay increases resulting from higher-than-expected COLAs. The
increased outlays for mandatory programs were largely offset by a
decrease of $4.4 billion in net interest due to decreased borrowing
requirements that resulted from the effect of economic factors on
receipts.
Technical estimating differences and other changes resulted in a net
decrease in outlays of $17.4 billion. Outlays for discretionary programs
decreased an estimated $4.1 billion. Outlays for mandatory programs
decreased an estimated $12.7 billion, largely due to lower-than-
anticipated outlays for Medicare. Technical changes result from changes
in such factors as the number of beneficiaries for entitlement programs,
crop conditions, or other factors not associated with policy changes or
economic conditions.
Table 18-3. COMPARISON OF THE ACTUAL 2001 SURPLUS WITH THE INITIAL CURRENT SERVICES SURPLUS ESTIMATE
(In billions of dollars)
----------------------------------------------------------------------------------------------------------------
Current Changes
Services -----------------------------------------
(Feb. Total Actual
2000) Policy Economic Technical changes
----------------------------------------------------------------------------------------------------------------
Receipts........................................... 2,009.9 -74.2 33.0 22.3 -18.9 1,991.0
Outlays............................................ 1,838.8 41.7 0.9 -17.4 25.2 1,863.9
------------------------------------------------------------
Surplus........................................ 171.2 -115.9 32.1 39.7 -44.1 127.1
----------------------------------------------------------------------------------------------------------------
Note: Surplus changes are receipts minus outlays. For these changes, a plus indicates an increase in the
surplus.
Surplus
The preceding two sections discussed the differences between the
initial current services estimates and the actual receipts and outlays
for 2001. This section combines these effects to show the net impact of
these differences on the surplus.
As shown in Table 18-3, the 2001 current services surplus was
initially estimated to be $171.2 billion. The actual surplus was $127.1
billion, which was a $44.1 billion decrease from the initial estimate.
Receipts were $18.9 billion less than the initial estimate, and outlays
were $25.2 billion more. The table shows the distribution of the changes
according to the categories in the preceding two sections.
The net effect of policy changes for receipts and outlays reduced the
surplus by $115.9 billion. Economic conditions that differed from the
initial assumptions in February 2000 accounted for an estimated $32.1
billion increase in the surplus. This was the combined effect of an
increase in receipts of $33.0 billion and an increase in outlays of $0.9
billion. Technical factors increased the surplus by an estimated $39.7
billion. This was due to an increase in receipts of $22.3 billion and a
decrease in outlays of $17.4 billion for technical estimating reasons.
Comparison of the Actual and Estimated Outlays for Mandatory and Related
Programs for 2001
This section compares the original 2001 outlay estimates for mandatory
and related programs under current law in the 2001 Budget (February
2000) with the actual outlays. Major examples of these programs include
Social Security and Medicare benefits for the elderly, agricultural
price support payments to farmers, and deposit insurance for banks and
thrift institutions. This category also includes net interest outlays
and undistributed offsetting receipts.
A number of factors may cause differences between the amounts
estimated in the budget and the actual outlays. For example, legislation
may change benefit rates or coverage; the actual number of beneficiaries
may differ from the number estimated; or economic conditions (such as
inflation or interest rates) may differ from what was assumed in making
the original estimates.
Table 18-4 shows the differences between the actual outlays for these
programs in 2001 and the amounts originally estimated in the 2001
Budget, based on laws in effect at that time. Actual outlays for
mandatory spending and net interest in 2001 were $1,214.6 billion, which
was $11.3 billion more than the initial estimate of $1,203.3 billion,
based on existing law in February 2000.
[[Page 376]]
Table 18-4. COMPARISON OF ACTUAL AND ESTIMATED OUTLAYS FOR MANDATORY AND RELATED PROGRAMS UNDER CURRENT LAW
(In billions of dollars)
----------------------------------------------------------------------------------------------------------------
2001
-----------------------------------------
Feb. 2000
estimate \1\ Actual \2\ Change
----------------------------------------------------------------------------------------------------------------
Mandatory outlays:
Human resources programs:
Education, training, employment, and social services............. 10.7 2.9 -7.8
Health:
Medicaid....................................................... 124.2 129.4 5.2
Other.......................................................... 8.1 9.7 1.6
-----------------------------------------
Total health................................................. 132.3 139.1 6.8
Medicare......................................................... 218.3 214.1 -4.2
Income security:
Retirement and disability...................................... 85.2 86.2 1.0
Unemployment compensation...................................... 24.6 27.9 3.3
Food and nutrition assistance.................................. 31.4 29.4 -2.0
Other.......................................................... 80.6 82.1 1.5
-----------------------------------------
Total, income security....................................... 221.9 225.6 3.8
Social security.................................................. 422.2 429.4 7.2
Veterans benefits and services:
Income security for veterans................................... 24.3 22.5 -1.8
Other.......................................................... 1.6 0.2 -1.4
-----------------------------------------
Total veterans benefits and services......................... 25.9 22.6 -3.3
-----------------------------------------
Total mandatory human resources programs..................... 1,031.2 1,033.7 2.5
Other functions:
Agriculture...................................................... 14.3 21.3 7.0
Mortgage Credit.................................................. -4.5 -* 4.5
Deposit insurance................................................ -1.5 -1.4 0.1
Other functions.................................................. 1.5 1.9 0.4
-----------------------------------------
Total, other functions......................................... 9.7 21.7 12.0
-----------------------------------------
Undistributed offsetting receipts:
Employer share, employee retirement.............................. -38.5 -38.8 -0.3
Rents and royalties on the outer continental shelf............... -3.7 -7.2 -3.5
Other undistributed offsetting receipts.......................... -3.6 -1.0 2.5
-----------------------------------------
Total undistributed offsetting receipts........................ -45.7 -47.0 -1.3
-----------------------------------------
Total, mandatory............................................. 995.2 1,008.4 13.2
Net interest:.........................................................
Interest on Treasury debt securities (gross)....................... 359.3 359.5 0.2
Interest received by trust funds................................... -141.9 -144.1 -2.2
Other interest..................................................... -9.4 -9.2 0.2
-----------------------------------------
Total net interest........................................... 208.1 206.2 -1.9
-----------------------------------------
Total outlays for mandatory and net interest................. 1,203.3 1,214.6 11.3
----------------------------------------------------------------------------------------------------------------
* $50 million or less.
\1\ Estimates reflect the function shift for foster care and adoption assistance in the 2003 Budget.
\2\ Actuals do not reflect the accrual funding of Federal retiree costs in the 2003 Budget.
Actual outlays for mandatory human resources programs were $1,033.7
billion, $2.5 billion more than originally estimated. This increase was
the net effect of legislative action, differences between actual and
assumed economic conditions, differences between the anticipated and
actual number of beneficiaries, and other technical differences.
Outlays for other functions were $12.0 billion more than originally
estimated, largely due to increases of $7.0 billion for agriculture
programs and $4.5 billion for mortgage credit programs. Undistributed
offsetting receipts were $1.3 billion more than expected.
Outlays for net interest were $206.2 billion, or $1.9 billion less
than the original estimate. This decrease was the net effect of changes
in interest rates from those initially assumed, changes in borrowing
requirements due to differences in surpluses, and technical factors.
[[Page 377]]
Reconciliation of Differences with Amounts Published by Treasury for
2001
Table 18-5 provides a reconciliation of the receipts, outlays, and
surplus totals published by the Department of the Treasury in the
September 2001 Monthly Treasury Statement and those published in this
budget. The Department of the Treasury made adjustments to the estimates
for the U.S. Government Annual Report, which increased receipts by $727
million and outlays by $870 million. Additional adjustments for this
budget increased receipts by $100 million and outlays by $17 million.
The major changes were for student loan reserve funds, highway trust
fund revenues, the Federal Communications Commission universal service
fund, and the transactions of the United Mine Workers of America benefit
funds.
Table 18-5. RECONCILIATION OF FINAL AMOUNTS FOR 2001
(In millions of dollars)
----------------------------------------------------------------------------------------------------------------
Receipts Outlays Surplus
----------------------------------------------------------------------------------------------------------------
Totals published by Treasury (September 2001 MTS)............... 1,990,203 1,863,039 127,165
Miscellaneous Treasury adjustments............................ 727 870 -144
-----------------------------------------------
Totals published by Treasury in U.S. Government Annual Report... 1,990,930 1,863,909 127,021
United Mine Workers of America benefit funds.................. 150 150 ..............
Federal Communications Commission Universal Service Fund...... 109 87 22
Highway trust fund............................................ -164 .............. -164
Student loan reserve fund..................................... .............. -231 231
Other......................................................... 5 11 -6
-----------------------------------------------
Total adjustments, net.......................................... 100 17 83
-----------------------------------------------
Totals in the budget............................................ 1,991,030 1,863,926 127,104
MEMORANDUM:
-----------------------------------------------
Total change since year-end statement......................... 827 887 -61
----------------------------------------------------------------------------------------------------------------
Part II: HISTORICAL COMPARISON OF ACTUAL TO ESTIMATED SURPLUSES
This part of the chapter compares actual surpluses to estimated
surpluses over the last two decades. The first section compares the
estimate for the budget year of each budget with the subsequent actual
surplus. The second section extends the comparison to the estimated
surpluses for each year of the budget window--that is, for the current
year through the fourth year following the budget year. This part
concludes with some observations on the historical record of surplus
estimates versus the subsequent actual surpluses.
Historical Comparison of Actual to Estimated Surpluses for the Budget
Year
Table 18-6 compares the estimated and actual surpluses or deficits
since the deficit estimated for 1982 in the 1982 Budget. The estimated
surpluses or deficits here for each budget include the Administration's
policy proposals. Therefore, the estimated surplus for 2001 differs from
that shown in Table 18-3, which is on a current services basis. Earlier
comparisons of actual and estimated surpluses were on a policy basis, so
for consistency the figures in Table 18-6 are on this basis.
On average, the estimates for the budget year underestimated actual
surpluses (or overestimated actual deficits) by $5 billion over the
twenty-year period. Policy outcomes that differed from the original
proposals reduced the surplus by an average of $16 billion. Differences
between economic assumptions and actual economic performance reduced the
surplus an average of $1 billion. Differences due to these two factors
were more than offset by technical revisions, which increased the
surplus an average of $23 billion.
The relatively small average difference between actual and estimated
surpluses conceals a wide variation in the differences from budget to
budget. The differences ranged from a $206 billion overestimate to a
$190 billion underestimate. The $206 billion overestimate, in the 1991
Budget, was due to the combination of weaker-than-expected economic
performance (largely the effect of the 1990-91 recession) and
unexpectedly large outlays to resolve the savings and loan crisis. The
$190 billion underestimate of the surplus, in the 1998 Budget, stemmed
largely from stronger-than-expected economic growth and a surge in
individual income tax collections beyond that accounted for by economic
factors.
[[Page 378]]
Table 18-6. COMPARISON OF ACTUAL AND ESTIMATED SURPLUSES SINCE 1982
(In billions of dollars)
----------------------------------------------------------------------------------------------------------------
Surplus Differences due to
or ----------------------------------
deficit (- Actual
) Total surplus
Budget estimated Enacted Economic Technical difference or
for legislation factors factors deficit(-
budget )
year \1\
----------------------------------------------------------------------------------------------------------------
1982........................................ -62 15 -70 -11 -66 -128
1983........................................ -107 -12 -67 -22 -101 -208
1984........................................ -203 -21 38 -* 17 -185
1985........................................ -195 -12 -17 12 -17 -212
1986........................................ -180 -8 -27 -7 -41 -221
1987........................................ -144 2 -16 8 -6 -150
1988........................................ -111 -9 -19 -16 -44 -155
1989........................................ -130 -22 10 -11 -23 -152
1990........................................ -91 -21 -31 -79 -131 -221
1991........................................ -63 21 -85 -143 -206 -269
1992........................................ -281 -36 -21 48 -10 -290
1993........................................ -350 -8 -13 115 95 -255
1994........................................ -264 -8 16 52 61 -203
1995........................................ -165 -18 1 18 1 -164
1996........................................ -197 6 53 30 89 -108
1997........................................ -140 1 -4 121 118 -22
1998........................................ -121 -9 48 151 190 69
1999........................................ 10 -22 56 81 115 125
2000........................................ 117 -42 88 74 119 236
2001........................................ 184 -129 32 40 -57 127
-------------------------------------------------------------------
Average..................................... -16 -1 23 5
Absolute average \2\........................ 21 36 52 75
Standard deviation.......................... 31 45 68 97
----------------------------------------------------------------------------------------------------------------
* $500 million or less.
\1\ Surplus or deficit estimate includes the effect of the budget's policy proposals.
\2\ Absolute average is the average without regard to sign.
Because the average surplus difference obscures the degree of under-
and overestimation in the historical data, a more appropriate statistic
to measure the magnitude of the differences is the average absolute
difference. This statistic measures the difference without regard to
whether it was an under- or overestimate. Since 1982, the average
absolute difference has been $75 billion.
Another measure of variability is the standard deviation. This
statistic measures the dispersion of the data around the average value.
The standard deviation of the surplus differences since 1982 is $97
billion. Like the average absolute difference, this measure illustrates
the high degree of variation in the difference between estimates and
actual surpluses.
Five-Year Comparison of Actual to Estimated Surpluses
The substantial differences between actual surpluses and the budget
year estimates made less than two years earlier raises questions about
the degree of variability for estimates of years beyond the budget year.
Table 18-7 shows the summary statistics for the surplus differences for
the current year (CY), budget year (BY), and the four succeeding years
(BY+1 through BY+4). These are the years that are required to be
estimated in the budget by the Budget Enforcement Act.
On average, the budget estimates since 1982 understated the surplus in
the current year and the budget year, by $19 billion and $5 billion
respectively. The budget estimates overstated the surplus in the years
following, by amounts growing from $4 billion for BY+1 to $26 billion
for BY+4. While these results suggest a slight tendency to overestimate
surpluses toward the end of the budget horizon, the averages are not
statistically different from zero in light of the high variation in the
data.
The average absolute difference between estimated and actual surpluses
grows dramatically over the six years from CY through BY+4, from $47
billion in the current year to $75 billion for the budget year, to $205
billion for BY+4. While under- and overestimates of the surplus have
historically tended to average out, the absolute size of the under- or
overestimates grows as the estimates extend further into the future. The
standard deviation of the surplus differences shows the same pattern.
The standard deviation grows from $60 billion for current year estimates
to $97 billion for the budget year estimates and continues to increase
steadily as the estimates extend further out, reaching $241 billion for
BY+4.
The estimates of variability in the difference between estimated and
actual surpluses can be used to construct a range of uncertainty around
a given set of surplus estimates. Statistically, if these differences
are normally distributed, the actual surplus will be within a
[[Page 379]]
range of two standard deviations above or below the estimate about 90%
of the time. Chart 18-1 shows this range of uncertainty applied to the
surplus estimates in this budget. This chart illustrates that unforeseen
economic developments, policy outcomes, or other factors could give rise
to large swings in the surplus estimates.
Table 18-7. DIFFERENCES BETWEEN ACTUAL AND ESTIMATED SURPLUSES FOR FIVE-YEAR BUDGET ESTIMATES SINCE 1982
(In billions of dollars)
----------------------------------------------------------------------------------------------------------------
Estimate for budget year plus
Current Budget -----------------------------------
Measure year year One Two Three Four
estimate estimate year years years years
(BY+1) (BY+2) (BY+3) (BY+4)
----------------------------------------------------------------------------------------------------------------
Average difference \1\.................................. 19 5 -4 -12 -28 -26
Average absolute difference \2\......................... 47 75 109 147 179 205
Standard deviation...................................... 60 97 142 178 207 241
----------------------------------------------------------------------------------------------------------------
\1\ A positive figure represents an underestimate of the surplus or an overestimate of the deficit.
\2\ Average absolute difference is the average difference without regard to sign.