III Choice of Potential Growth Rate

Ahsan Mansur, Richard Haas, and Peter Heller
Published Date:
May 1986
  • ShareShare
Show Summary Details

Both the fiscal impulse and structural balance measures are sensitive to the assumed rate of potential GDP growth. The gap between potential and actual output directly enters into the calculation of the cyclically neutral balance, and the level of potential output is used in estimating the structural balance. The sensitivity of the fiscal impulse measure to the gap is illustrated in Table 3. Currently, the Fund’s estimates of potential output are developed by the individual country desk economists and represent an amalgam of estimates provided by the authorities and staff estimates. Country-specific criteria are important, particularly the extent to which a sustainable balance of payments may constrain potential output. Yet the underlying concept of the current measures supplied by the staff working on individual countries may vary from one country to another. In some instances, it may refer to a peak-to-peak value attainable only with increasing inflation; in others, it may be based on a natural-growth-rate notion.

Table 3.Seven Major Industrial Countries: Sensitivity of Estimates of Central Government Fiscal Impulse to Potential Growth Rate Assumptions, 1977–85(As percentages of GDP)
CountryAssumed Average Potential Growth Rate197719781979198019811982198319841985
Scenario A4.751.100.88−0.84−0.57−1.341.010.51−0.12
Scenario B3.751.281.07−0.66−0.37−
Scenario C2.751.471.25−0.47−0.18−0.971.500.910.23
United States
Scenario A4.00−0.05−0.20−0.990.13−−0.26−0.36
Scenario B3.000.16−0.790.350.160.522.03−0.07−0.15
Scenario C2.000.380.21−0.590.560.360.762.230.110.04
Scenario A6.20−0.05−0.030.81−0.24−0.55−0.90−0.70−0.48−0.94
Scenario B5.−0.01−0.31−0.64−0.45−0.23−0.68
Scenario C4.200.430.451.280.23−0.07−0.40−0.210.01−0.45
Scenario A4.50−0.561.66−1.05−0.900.73−0.21−0.57−0.60−0.48
Scenario B3.50−0.351.87−0.85−0.680.960.02−0.32−0.34−0.23
Scenario C2.50−0.142.09−0.64−0.471.190.24−0.09−0.10−0.01
Germany, Federal Republic of
Scenario A3.70−0.66−0.17−0.29−0.68−1.01−2.10−0.45−0.53−0.39
Scenario B2.70−0.370.12−0.38−0.69−1.75−0.11−0.21−0.07
Scenario C1.70−0.070.420.29−0.09−0.38−1.420.200.080.21
Scenario A4.50−1.144.92−3.36−0.460.160.17−0.59−0.58
Scenario B3.50−0.775.30−2.98−0.070.590.65−0.07−0.09
Scenario C2.50−0.385.70−2.590.311.011.100.420.35
United Kingdom
Scenario A3.70−2.722.360.15−2.77−2.90−1.662.03−1.62−0.96
Scenario B2.70−2.352.740.53−2.36−2.46−1.252.43−1.22−0.54
Scenario C1.70−1.963.120.90−1.95−2.03−0.852.80−0.85−0.17
Source: Fund staff estimates as of February 29, 1984.
Source: Fund staff estimates as of February 29, 1984.

There are several alternative ways to develop consistent measures of potential output. One can use estimated aggregate production functions with suitably adjusted estimates of the levels of the different factors of production. The drawbacks to this approach are not only the substantial computational burden but also the difficult problems of measuring the capital stock in each country.

A second approach would center on the relationship between a single factor and the hypothetical level of output. The authorities in the United States and Canada focus on the labor force and changes in labor productivity to generate “high-employment” and average levels of output, respectively. The German authorities focus on the capital stock as a factor of production in an analogous manner. While easier to implement than a production function-based technique, the single-factor approach still involves a good deal of computational effort and is more restrictive.

A third approach, used by the OECD, is to derive the productive potential from trend regression analyses on output (when official estimates are not available). In France, a “balance of payments constraint” is taken into account by the OECD to allow for their assumption that growth rates are linked to the growth of productive potential elsewhere.

A fourth approach begins by identifying a medium-term attainable level of output for some future target year and seeks a potential growth rate that will eliminate the gap by the target year. A constant growth trend can then be estimated from the base year to this target year, which is then assumed to be the constant growth rate of potential real output over the whole period. Actual and predicted GDP deflators are then applied to generate the nominal values required for the calculation of the fiscal impulse measure. This concept of an attainable output growth rate is similar to the “middle-expansion trend” approach suggested by de Leeuw and Holloway (1983).18

Figure 1 shows a stylized example with 1987 as the target year: YalternativeIp reflects this fourth approach. In the figure, the time path of Ycurrentp reflects the potential growth rate in actual use and indicates that the potential growth rate estimates have varied over the period. A disadvantage of this fourth approach is that exogenous economic disturbances (such as the “oil shock”) may have reduced potential output at some point after the base year. Estimation of a trend line from the base year would therefore underestimate the relevant trend potential growth rate. This could lead both to distortions in any retrospective appraisal of past fiscal policies and in the evaluation of the current fiscal impulse.


A fifth approach, which is a variant of the fourth, is to estimate the output which could be reasonably attained in the current year under suitable macroeconomic policies and without an acceleration of inflationary pressures. This assumes a reasonable utilization of the available capital stock and a rate of unemployment reflecting current frictional and structural unemployment rates. Potential output would then be assumed to grow at a constant rate from the base-year output level to this estimated output level attainable in the current year, and then to grow at a constant rate. For purposes of illustration, the 1984 and 1985 values of the potential growth rate have been set equal to the Fund desk economists’ projections. This approach has the net effect of lowering the level of the potential output path over the medium term, relative to the current procedure, but not of reducing the growth rate of potential output. This is shown in Figure 1 as YalternativeIIp.

Table 4 provides estimates of the potential growth rates actually used in the Fund’s World Economic Outlook exercise and the estimates implied by using these last two alternative techniques. The 1987 values employed were the medium-term scenario output levels used in the preparation of the World Economic Outlook published in April 1984.19 These last two alternatives typically generate lower implied potential output growth rates than the estimates provided by the Fund economists working on particular countries. For Japan, the United States, the Federal Republic of Germany, Canada, and France, one also observes higher potential growth rates over 1978–83 when using the fifth approach than when using the fourth. In any given year, an increase in the potential growth rate reduces the cyclically neutral balance Bn, thereby yielding a smaller fiscal impulse.

Table 4.Seven Major Industrial Countries: Current and Alternative Real Potential Output Growth Rates, 1978–851(In percent)
Current measure3.253.002.752.752.752.752.752.75
Alternative I2.352.352.352.352.352.352.352.35
Alternative II2.452.452.452.452.452.452.752.75
United States
Current measure3.002.752.752.752.752.752.752.75
Alternative I2.472.472.472.472.472.472.472.47
Alternative II2.502.502.502.502.502.502.752.75
Current measure5.004.504.504.304.
Alternative I3.673.673.673.673.673.673.673.67
Alternative II4.764.764.764.764.764.764.004.00
Current measure2.502.502.502.502.502.502.502.50
Alternative I1.731.731.731.731.731.731.731.73
Alternative II2.
Germany, Federal Republic of
Current measure2.602.602.602.302.
Alternative I1.701.701.701.701.701.701.701.70
Alternative II2.302.302.302.302.302.302.002.00
Current measure3.503.503.503.503.503.003.003.00
Alternative I1.691.691.691.691.691.691.691.69
Alternative II
United Kingdom
Current measure2.
Alternative I1.
Alternative II0.980.980.980.980.980.982.002.00

Current measure refers to the estimates of potential growth rates used in the Fund’s April 1984 World Economic Outlook exercise. The methodology underlying Alternatives I and II is described in the text. OECD estimates were those provided as of November 1983.

Current measure refers to the estimates of potential growth rates used in the Fund’s April 1984 World Economic Outlook exercise. The methodology underlying Alternatives I and II is described in the text. OECD estimates were those provided as of November 1983.

Conceptually, one should note some of the issues and problems associated with a change to either of the latter two approaches to measuring potential output. First, the fourth approach implies a change that renders the entire potential output series more sensitive to the output projections of the medium-term scenario of a given World Economic Outlook exercise, and thus is implicitly sensitive to the underlying assumptions on fiscal policy. Second, one needs to evaluate what a reasonable growth path for “cyclically neutral” expenditure is. Is it more likely to be based on a smooth potential growth rate over a long period that is related to attainable output, or to be based on a variable growth rate that is related to potential, rather than realistically attainable, output? Third, revisions in the target year may have significant implications for the resulting growth path.

Finally, it should be stressed that the estimate of the potential growth rate obtained by fitting a line between the full-capacity base year and a point several years hence, when (it is assumed) full capacity will again prevail (the fourth approach above), will lead to implausibly low estimates of potential output growth in the future if a once-and-for-all occurrence in the past has lowered the level of potential output. In this case, the effect of a past drop in potential output should be allowed to lower the historical growth rate but not the future growth rate. If the future growth rate of potential output is biased downward, as well it might be, then so will be the cyclical adjustment. Consequently, there will be an upward bias in the fiscal impulse. This can lead to differences in the resulting fiscal impulse measures. In some cases, this may distort the perceived character of fiscal policy. For example, if a country were to pursue a consciously contractionary fiscal policy over a number of years, this might lead ultimately to a downward revision in the attainable output level, and restrospectively lead to an unwarranted re-evaluation of the fiscal policy of past years and a false conclusion that it had been expansionary.

    Other Resources Citing This Publication