Accuracy of OECD and IMF Projection

Accuracy of OECD and IMF Projection Mordechai E. Kreinin, Michigan State University This paper tests the accuracy of annual forecasts made by the OECD and the IMF of: real GDP growth rate, the GDP deflator, unemployment, and the trade balance. These projections are made for each OECD country, and cover a period of about 25 years. Neither the OECD nor the IMF succeed in forecasting cyclical turning points. But other than that, their projections appear fairly robust and certainly superior to those of a “naive” model.  2000 Society for Policy Modeling. Published by Elsevier Science Inc. Key Words: Unbiased forecast; Real GDP; DGP deflator.

1. INTRODUCTION Every year and half-year the OECD provides projections of several economic variables, published in the OECD Economic Outlook, while the IMF provides similar projections published in the IMF World Economic Outlook. Because these forecasts are used extensively by governmental and nongovernmental organizations, it is useful to examine their accuracy. The assessment provided here differs in approach from earlier assessments,1 but their purpose is similar. Because much previous work concentrated on the IMF projections (see footnote 1), the present effort is devoted mainly to those of the OECD. Whenever possible, a comparison is made between the two. But because of differences in years and countries of coverage (as contained in the IMF publicly published 1 See: Smyth, D.J. and Ash, J.C.K. “Forecasting GDP, The Rate of Inflation and the Balance of Trade: The OECD Performance”, Economic Journal, June 1975, pp. 361–364; Artis, M.J. “How Accurate Is the World Economic Outlook? A Post Mortem On ShortTerm Forecasting at the IMF,” IMF, Staff Studies for the World Economic Outlook, July 1988, pp. 1–48. Barrionuevo, J.M. “How Accurate Are the World Economic Outlook Projections?”, Staff Studies for the World Economic Outlook, December 1993, pp. 28–45; and Ibid, May 1992; and DeMasi, P. “The Difficult Art of Economic Forecasting” Finance and Development, December 1996. See also the literature cited therein. Address correspondence to M.E. Kreinin, Michigan State University, Department of Economics, East Lansing, MI 48824.

Journal of Policy Modeling 22(1):61–79 (2000)  2000 Society for Policy Modeling Published by Elsevier Science Inc.

0161-8938/00/$–see front matter PII S0161-8938(99)00009-5

62

M. E. Kreinin

version), the comparison is rather imperfect. Finally, the accuracy of projections will be compared with that of a “Naive” forecast. Of the several possible “Naive” forecasts, I have chosen to follow the book of Ecclesiastes (Chapter 1, verse 9): “That which has been is that which shall be And that which hath been done is that which shall be done And there is nothing new under the sun.”

In other words, the Naive forecasts assumes that last year’s development (e.g., growth rate or inflation) will be the forecast for the next year. OECD forecasts are generated by “country desks,” and are based on: models, the desk-official judgment, accounting checks, anticipated shocks, and the like. Each “forecasting round” lasts 3 months, and contains two to three iterations; it begins with exogenous assumptions about exchange rates and other variables. The OECD is restricted to using governmental projections of each respective public sector. Estimates are made of each country’s domestic demand and import demand, and a global model is used to check the consistency of a country’s imports with other countries’ exports. That adds to the accuracy of the trade balances projections, as shown in Section 5 below. Consistency of interest rates between countries (and that of other variables) are also checked. Reliance is made on none-OECD sources such as the IMF and the World Bank. In particular, the OECD reads the IMF projections that appear earlier. 2. REAL GDP GROWTH RATES For each year (t) the OECD Economic Outlook makes growth rate projections for every OECD member in the preceding year (t ⫺ 1), where the “forecasting round” is done during September– November; and in the middle of the year itself (t ⫺ 1/2), where the “forecasting round” is done during March–May. The actual growth rate is published in the following year (t ⫹ 1). For example, the projections for 1990 appear in 1989 and then in mid-1990, while the actual 1990 growth rates are published in 1991. To estimate the accuracy of the projections, the actual growth rate (Ga) is regressed on the “year earlier” (t ⫺ 1) projection and separately on the midyear (t ⫺ 1/2) projection: Ga ⫽ a ⫹ b Pt⫺1 ⫹ ⑀

Ga ⫽ a ⫹ b Pt⫺1/2 ⫹ ⑀

(1)

A coefficient b ⫽ 1 and a ⫽ 0 would indicate an unbiased

ACCURACY OF OECD AND IMF PROJECTION

63

forecast. Regressions are estimated in a time series for each country, cross-sectionally for each year across countries, and for all years and countries combined. Individual country errors can offset each other in the second and third approaches. Table 1 presents the regression results for all country–year combinations. The midyear OECD projections are more accurate than the year-earlier forecasts, but in neither case do the coefficient 1 and the constant zero fall within the 95 percent of their respective confidence intervals. On the other hand, the projections are superior to the results of the “Naive” model. Similarly, the IMF projections appear superior to its corresponding “Naive” model. But the OECD and IMF forecasts are not comparable because they differ both in the number of countries and years of coverage. More important are the individual country projections. Those made by the OECD for the eight largest countries are shown in Table 2. For seven of the countries (all except Italy) the yearearlier projections are unbiased in a sense that the dual criterion is met: the coefficient 1 and the constant 0 fall within their respective 95-percent confidence intervals. The U.S. projections appear “best” in terms of the statistical properties. In general, the midyear projections are superior to the year-earlier ones, as might be expected. In the following smaller countries the coefficient ⫽ 1 and the constant ⫽ 0 falls within their respective 95-percent confidence intervals: Austria, Belgium, Denmark, Finland, Greece, The Netherlands, Portugal, Spain, Sweden, Switzerland, and Turkey. Neither condition is met in Norway. Only one of the conditions (coefficient ⫽ 1) is met in Iceland, Ireland, and Luxemburg. In the cross-sectional regressions (not shown) both conditions are met in 15 out of 27 years, but there is no discernible improvement over time. Also the midyear projections do not represent a marked improvement over the year-earlier ones. That these projections are far superior to the forecast by the Naive model is shown in Table 3. In none of the eight “large” countries is the dual criterions met. Nor is it met in any of the smaller countries. Using the same criteria for the cross-sectional data covering the years 1962–94, in 18 years neither of the two criteria is met; in 8 years both criteria are met; in 5 years only the constant ⫽ 0 criterion is met; and in 1 year only the coefficient ⫽ 1 criterion is met. Table 4 shows the IMF projections of growth rates for the G-7 and other OECD countries. With only 10 observations, the

0.7 0.9 0.4 2.0 0.7 0.6 0.4 1.3

Coefficient Constant Coefficient Constant Coefficient Constant Coefficient Constant

t⫺1 Coef.

5.3 5.5

3.1 1.0

13.5 14.7

15.1 5.8

t Stat.

0.2

0.1

0.2

0.3

R2 N

t ⫺ 1/2 Coef.

0.6 0.9

(i.i) Naive Model (same as last year) 0.3–0.6 112 0.9–1.8

(B) IMF: (i) Projections 0.1–1.1 80 ⫺0.6–1.9

(i.i) Naive Model (same as last year) 0.3–0.47 893 1.7–2.3

(A) OECD: (i.) Projections 0.6–0.8 611 0.9 0.6–1.2 0.6

95% Conf. interv.

9.0 4.7

27.9 5.1

t Stat.

Table 1: Real GDP Growth Rates Projections: Time Series and Cross-Section Combined

0.4

0.5

R2

0.5–0.75 0.5–1.3

0.83–0.96 0.4–0.9

95% Conf. interv.

112

644

N

64 M. E. Kreinin

Australia

Italy

Canada

United Kingdom

France

Germany

Japan

United States

Country Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant

0.9 0.0 0.8 0.6 1.0 0.1 0.8 0.5 0.9 0.4 1.4 ⫺1.0 0.6 1.4 0.6 1.0

t⫺1 Coef. 7.0 0.0 4.5 0.6 3.7 0.2 4.6 0.8 3.7 0.6 5.3 ⫺1.1 3.6 2.2 2.0 1.0

t Stat.

0.1

0.3

0.5

0.3

0.4

0.3

0.4

0.7

R2 0.7–1.2 ⫺0.9–0.9 0.4–1.1 ⫺1.4–2.7 0.4–1.6 ⫺1.5–1.8 0.4–1.2 ⫺0.7–1.7 0.4–1.4 ⫺0.8–1.6 0.8–1.9 ⫺2.9–0.8 0.3–1.0 0.1–2.6 ⫺0.0–1.2 ⫺1.1–3.3

95% Conf. interv.

Table 2: OECD Projections of Real GDP Growth Rates–Time Series

22

26

26

26

26

26

26

27

N 0.8 0.4 0.8 0.8 0.9 0.6 0.8 0.7 1.1 0.4 1.0 0.2 0.9 0.7 0.9 0.3

t ⫺ 1/2 Coef. 10.0 1.4 8.0 1.3 9.0 1.7 8.8 2.1 10.1 1.3 7.5 0.5 7.0 1.5 5.0 0.5

t Stat.

0.5

0.6

0.7

0.8

0.7

0.7

0.7

0.8

R2

0.6–0.9 ⫺0.2–1.0 0.6–1.0 ⫺0.5–2.1 0.7–1.1 ⫺0.1–1.2 0.6–1.0 0.0–1.3 0.9–1.3 ⫺0.2–0.9 0.7–1.3 ⫺0.8–1.3 0.6–1.2 ⫺0.2–1.6 0.5–1.2 ⫺1.0–1.6

95% Conf. interv.

23

28

28

28

28

28

28

27

N

ACCURACY OF OECD AND IMF PROJECTION 65

66

M. E. Kreinin

Table 3: OECD Projections of Real GDP Growth Rate–Time Series Naive Model

Country United States Japan Germany France United Kingdom Canada Italy Australia

Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant

t⫺1 Coef.

t Stat.

0.3 2.1 0.5 3.2 0.3 2.0 0.6 1.2 0.3 1.5 0.4 2.3 0.3 2.1 0.3 2.8

1.6 3.2 2.9 2.9 1.9 3.1 4.5 2.2 1.9 2.9 2.5 3.1 2.2 3.1 1.7 4.0

R2 0.0 0.2 0.1 0.4 0.1 0.1 0.1 0.1

95% Conf. interv. ⫺0.1–0.6 0.8–3.4 0.1–0.8 1.0–5.5 ⫺0.0–0.7 0.7–3.3 0.3–0.9 0.1–2.4 0.0–0.7 0.4–2.6 0.1–0.7 0.8–3.9 0.0–0.7 0.7–3.5 ⫺0.1–0.6 1.4–4.3

N 34 34 33 33 33 33 33 33

statistical properties are rather poor, but they improve considerably in the midyear projections. Yet the dual criterion is met in most cases. In the cross-sectional regressions, with only eight observations, the dual criterion is met in 10 out of 15 years. These results are superior to forecast by the Naive model shown in Table 5. Where the OECD and IMF projections do poorly is in forecasting the turning points. Table 6 shows poor statistical properties, and coefficients at variance with expectations. The only possible exception is in the IMF midyear forecast for the downturns and upturns combined. 3. INFLATION Projections of the GDP deflator (cross-section and time series combined) are shown in Table 7. The OECD projections easily meet the dual criterion, and have desirable statistical properties. The midyear projections are superior to the year-earlier ones. But the Naive model also appears to give unbiased (though less accurate) results. Similar observations may be made about the IMF projections; although, again, a comparison between the two

Other OECD

Italy

Canada

United Kingdom

France

Germany

Japan

United States

Country Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant

1.4 ⫺1.4 0.5 1.6 0.3 1.8 ⫺0.8 4.0 2.2 ⫺3.1 1.4 ⫺1.7 0.5 0.8 ⫺0.0 2.5

t⫺1 Coef. 2.4 ⫺0.8 0.7 0.7 0.4 0.8 ⫺1.0 2.3 1.9 ⫺1.1 2.0 ⫺0.8 1.2 0.7 ⫺0.0 0.8

t Stat.

⫺0.1

0.0

0.2

0.2

0.0

⫺0.1

⫺0.1

0.4

R2 0.1–2.7 ⫺5.2–2.4 ⫺1.1–2.1 ⫺4.0–7.3 ⫺1.7–2.4 ⫺3.3–6.9 ⫺2.5–0.9 ⫺0.1–8.1 ⫺0.4–4.9 ⫺9.8–3.6 ⫺0.2–3.0 ⫺6.8–3.4 ⫺0.5–1.5 ⫺2.0–3.6 ⫺3.0–3.0 ⫺4.7–9.7

95% Conf. interv.

Table 4: IMF Projections of Real GDP Growth Rates–Time Series

10

10

10

10

10

10

10

10

N 1.2 ⫺0.3 1.4 ⫺1.1 0.2 1.4 0.8 0.7 1.1 0.6 1.4 ⫺0.6 0.9 0.4 1.1 0.1

t ⫺ 1/2 Coef. 9.8 ⫺0.8 7.3 ⫺1.7 2.0 2.6 2.3 1.2 7.9 1.7 5.9 ⫺1.0 3.1 0.6 2.1 0.1

t Stat.

0.2

0.4

0.7

0.8

0.2

0.2

0.8

0.9

R2

0.9–1.5 ⫺1.0–0.4 1.0–1.8 ⫺2.5–0.3 0.0–0.4 0.2–2.5 0.0–1.5 ⫺0.6–1.9 0.8–1.5 ⫺0.2–1.2 0.9–1.9 ⫺2.0–0.8 0.3–1.5 ⫺0.9–1.6 ⫺0.2–2.2 ⫺2.0–2.2

95% Conf. interv.

14

14

14

14

14

14

14

14

N

ACCURACY OF OECD AND IMF PROJECTION 67

68

M. E. Kreinin

Table 5: IMF Data on Real GDP Growth Rate–Time Series Naive Model

Country United States Japan Germany France United Kingdom Canada Italy Other Europe

Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant

t⫺1 Coef.

t Stat.

0.2 2.0 0.7 0.8 0.3 1.4 0.3 1.4 0.6 1.1 0.3 1.9 0.4 1.0 0.5 1.1

0.8 2.5 2.7 0.8 1.3 1.9 1.1 2.2 3.4 2.0 1.0 1.9 1.7 1.8 2.1 1.7

R2 0.0 0.3 0.0 0.0 0.4 0.0 0.1 0.2

95% Conf. interv. ⫺0.4–0.8 0.2–3.8 0.1–1.2 ⫺1.3–3.0 ⫺0.2–0.9 ⫺0.2–3.0 ⫺0.3–0.9 0.0–2.8 0.2–1.0 ⫺0.1–2.3 ⫺0.3–0.9 ⫺0.3–4.0 ⫺0.1–0.9 ⫺0.2–2.3 ⫺0.0–1.1 ⫺0.3–2.4

N 14 14 14 14 14 14 14 14

sets of forecasts is impossible by reason of different years and countries of coverage. Table 8 shows that the OECD inflation projections for individual countries meet the dual criterion (coefficient ⫽ 1 and constant ⫽ 0 within the respective 95-percent confidence intervals) and have desirable statistical properties. The United States and France exhibit the most robust results. Indeed, the dual criterion is met also in all but two (Greece and Luxembourg) of the small OECD countries, not shown in the tables. The midyear projections constitute an improvement over the year-earlier ones. In the crosssectional regressions for the years 1975–94, the dual criterion was met in 14 years, while at least one criterion was met in 5 years. However, as shown in Table 9, the Naive model gives equally unbiased results, for both the large and the small (not shown) OECD countries. In its cross-sectional regressions for 1973–94 the dual criterion was met in 12 years, and at least one of them was not met in 10 years. Very similar results are obtained in the case of the IMF projections. Again, the forecasts are “best” for the United States and France; they are more reliable in midyear than a year earlier (Table 10); the dual criterion was met in the cross-sectional data

Downturns and upturns combined

Downturns and upturns combined (B) IMF Downturns only

(A) OECD Downturns only 0.0 ⫺1.5 0.2 0.5 0.7 ⫺2.4 ⫺0.0 0.4

Coef. Constant Coef. Constant Coef. Constant Coef. Constant

t⫺1 Coef.

1.9 ⫺3.1 ⫺0.0 0.3

0.3 ⫺4.1 1.1 1.5

t Stat.

0.0

0.2

0.0

0.0

R2

⫺0.0–1.4 ⫺3.8–0.9 ⫺1.2–1.2 ⫺2.3–3.0

⫺0.2–0.3 ⫺2.1–0.8 ⫺0.1–0.5 ⫺0.1–1.1

95% Conf. interv.

14

8

119

59

N

0.3 ⫺1.0 1.3 ⫺0.0

0.4 ⫺1.3 0.7 0.2

t ⫺ 1/2 Coef.

2.0 ⫺4.2 5.6 ⫺0.2

3.3 ⫺4.6 8.0 0.8

t Stat.

Table 6: Projections of Real GDP Growth-Turning Points: Cross-Section and Time Series Combined

0.6

0.2

0.3

0.1

R2

0.0–0.7 ⫺1.5–0.5 0.7–1.4 ⫺0.6–0.5

0.1–0.6 ⫺1.8–0.7 0.6–0.9 ⫺0.2–0.6

95% Conf. interv.

26

15

121

61

N

ACCURACY OF OECD AND IMF PROJECTION 69

1.2 ⫺0.3 0.9 0.4 0.9 0.6 0.8 0.5

Coef. Constant Coef. Constant Coef. Constant Coef. Constant

t⫺1 Coef.

25.9 2.5

8.8 1.3

48.6 1.3

65.0 ⫺1.5

t Stat.

0.9

0.6

0.8

0.9

R2 N

t ⫺ 1/2 Coef.

(i.i) Naive Model (same as last year) 0.7–0.8 111 0.1–0.9

(B) IMF: (i) Projections 0.7–1.1 80 0.9 ⫺0.3–1.4 0.2

(i.i) OECD Naive Model (same as last year) 0.9–1.0 608 ⫺0.2–0.9

(A) OECD: (i.) Projections 1.1–1.2 535 1.0 ⫺0.7–1.0 0.1

95% Conf. interv.

Table 7: Projections of GDP Deflator: Cross Section and Time Series Combined

30.6 0.8

69.5 0.9

t Stat.

0.9

0.93

R2

0.9–1.0 ⫺0.3–0.7

1.0–1.1 ⫺0.2–0.4

95% Conf. interv.

112

563

N

70 M. E. Kreinin

Australia

Italy

Canada

United Kingdom

France

Germany

Japan

United States

Country Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant

0.92 0.5 0.98 0.2 1.1 ⫺0.1 0.97 0.4 1.0 0.9 1.1 ⫺0.5 0.94 2.0 0.93 0.8

t⫺1 Coef. 8.2 0.7 4.4 0.2 7.5 ⫺0.1 15.2 0.9 8.1 0.7 6.4 ⫺0.5 10.1 1.9 5.8 0.6

t Stat.

0.6

0.8

0.6

0.7

0.9

0.7

0.5

0.75

R2

Table 8: OED Projections of GAP Deflator–Time Series

0.7–1.2 ⫺0.9–1.8 0.5–1.4 ⫺2.1–2.5 0.8–1.4 ⫺1.2–1.1 0.8–1.1 ⫺0.6–1.5 0.8–1.3 ⫺1.6–3.4 0.7–1.4 ⫺2.8–1.8 0.7–1.1 ⫺0.2–4.2 0.6–1.3 ⫺2.1–3.6

95% Conf. interv.

20

23

23

23

23

23

23

23

N 1.1 ⫺0.1 0.94 0.1 1.0 ⫺0.1 0.95 0.4 0.98 0.5 1.1 ⫺0.5 0.9 1.5 0.93 0.5

t ⫺ 1/2 Coef. 17.8 ⫺0.3 16.1 0.3 16.1 ⫺0.2 20.8 1.1 25.6 1.2 12.3 ⫺0.8 22.2 3.1 9.5 0.6

t Stat.

0.8

0.95

0.9

0.96

0.95

0.9

0.9

0.9

R2

0.9–1.2 ⫺0.8–0.6 0.8–1.1 ⫺0.6–0.9 ⫺0.9–1.2 ⫺0.6–0.5 0.9–1.0 ⫺0.3–1.1 0.9–1.1 ⫺0.3–1.3 0.9–1.3 ⫺1.6–0.7 0.8–1.0 0.5–2.5 0.7–1.1 ⫺1.3–2.4

95% Conf. interv.

20

26

26

26

26

26

26

26

N

ACCURACY OF OECD AND IMF PROJECTION 71

72

M. E. Kreinin

Table 9: OED Projections of GAP Deflator Naive Model

Country United States Japan Germany France United Kingdom Canada Italy Australia

Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant

t⫺1 Coef.

t Stat.

0.8 1.2 0.7 1.3 0.6 1.6 0.9 0.6 0.7 2.6 0.8 0.9 0.9 1.1 0.9 0.4

5.8 1.5 4.2 1.4 4.2 2.4 9.5 0.8 4.7 1.8 6.6 1.1 10.3 1.1 7.5 0.3

R2 0.6 0.4 0.4 0.8 0.5 0.6 0.8 0.7

95% Conf. interv. 0.5–1.0 ⫺0.4–2.8 0.3–1.0 ⫺0.6–3.2 0.3–0.9 0.2–3.0 0.7–1.1 ⫺0.9–2.1 0.4–1.0 ⫺0.4–5.7 0.6–1.1 ⫺0.8–2.6 0.7–1.1 ⫺1.0–3.2 0.7–1.2 ⫺1.9–2.7

N 27 27 27 27 27 27 27 22

in 9 out of 10 years; and the Naive forecasts gives equally unbiased results (Table 11). 4. UNEMPLOYMENT Table 12 shows the OECD unemployment projections, where the U.S. results are the “best” of the eight countries, but the dual criterion is met in six of them.2 The midyear forecast shows a clear improvement over the year-earlier one. However, only in half of the smaller OECD countries was the dual criterion met. In the cross-sectional analysis, the dual criterion was met in 25 out of 27 years. But the Naive model shows equally good if not better results for the eight countries as well as for the small OECD countries (Table 13). Similar observations can be made about the IMF projections (not shown; but available from the author upon request). 5. TRADE BALANCES Only the OECD and not the IMF provide projections of trade balances. Table 14 shows the combined cross-section and time 2

The time series and cross-section combined are available from the author upon request.

Other OECD

Itlay

Canada

United Kingdom

France

Germany

Japan

United States

Country Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant

1.2 ⫺1.1 0.2 1.0 1.1 ⫺0.1 1.1 ⫺0.5 1.5 ⫺1.9 0.7 0.5 0.6 2.7 0.1 4.8

t⫺1 Coef. 4.6 ⫺1.2 0.3 1.0 3.2 ⫺0.1 7.1 ⫺0.9 3.1 ⫺0.9 1.6 0.3 2.8 2.0 0.2 2.5

t Stat.

⫺0.1

0.4

0.2

0.5

0.8

0.5

⫺0.1

0.7

R2

95% Conf. Interv. 0.6–1.8 ⫺3.4–1.1 ⫺1.1–1.5 ⫺1.1–3.0 0.3–2.0 ⫺2.5–2.3 0.8–1.5 ⫺1.7–0.7 0.4–2.6 ⫺7.0–3.2 ⫺0.3–1.6 ⫺2.9–3.9 0.1–1.1 ⫺0.4–5.8 ⫺1.0–1.1 0.4–9.3

Table 10: IMF Projections of the GAP Deflator–Time Series

10

10

10

10

10

10

10

10

N 1.0 0.1 0.3 0.9 0.9 ⫺0.0 0.9 0.3 1.0 0.1 0.8 0.0 1.0 0.2 0.7 2.2

t ⫺ 1/2 Coef. 21.4 0.4 2.7 2.7 6.0 ⫺0.0 18.0 0.8 14.6 0.2 6.0 0.0 16.3 0.3 4.9 2.4

t Stat.

0.6

0.95

0.7

0.94

0.96

0.7

0.3

0.97

R2

0.9–1.1 ⫺0.4–0.6 0.1–0.6 0.2–1.6 0.6–1.3 ⫺1.3–1.3 0.8–1.1 ⫺0.5–1.1 0.8–1.1 ⫺1.0–1.2 0.5–1.1 ⫺1.6–1.7 0.9–1.1 ⫺1.3–1.7 0.4–1.1 0.2–4.2

95% Conf. interv.

14

14

14

14

14

14

14

14

N

ACCURACY OF OECD AND IMF PROJECTION 73

74

M. E. Kreinin

Table 11: IMF Data on Real GDP Growth Rate–Time Series Naive Model

Country United States Japan Germany France United Kingdom Canada Italy Other Europe

Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant

t⫺1 Coef.

t Stat.

0.7 0.8 0.5 0.6 0.6 1.2 0.9 ⫺0.2 0.5 2.4 0.7 0.7 0.8 0.7 0.9 ⫺0.0

5.5 1.1 2.0 1.3 2.5 1.5 12.3 ⫺0.4 5.6 3.5 5.9 1.3 8.5 0.6 6.4 ⫺0.0

R2 0.7 0.2 0.3 0.9 0.7 0.7 0.8 0.8

95% Conf. interv. 0.4–1.0 ⫺0.7–2.2 ⫺0.4–1.1 ⫺0.1–1.7 0.1–1.1 ⫺0.5–3.0 0.7–1.1 ⫺1.3–0.9 0.3–0.7 0.9–3.8 0.4–0.9 ⫺0.5–1.9 0.6–1.0 ⫺1.7–3.1 0.6–1.3 ⫺2.2–2.2

N 14 14 14 14 14 14 14 14

series regressions. The trade balances forecast is shown to be unbiased, with robust statistical properties, and superior to that of the Naive model. The midyear projection is even better. As shown in Tables 15 and 16, the unbiasness exists in the G-5 countries in the time series analysis, and for 11 of the small OECD countries. In the cross-sectional analysis the dual criterion was met in 21 out of 27 years (1967–94). In the Naive model, the dual criterion is met in the G-7 countries and in all the small countries except Turkey. On the other hand, in the cross-sectional analysis the dual criterion was met in only 10 years. 6. CONCLUSION Efforts by the OECD and IMF to provide forecasts of crucial variables are clearly warranted. Although their accuracy and degree of unbiasness vary greatly, and neither organization succeeds in forecasting the turning points, in most other cases they offer an unbiased forecast that improves on a “naive” idea that “the projection for next year will be the same as last year’s performance.” This is particularly true for the most crucial variables: the growth rate and inflation.

Australia

Italy

Canada

United Kingdom

France

Germany

Japan

United States

Country Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant

0.9 0.4 0.6 1.0 0.4 4.4 1.0 ⫺0.1 0.5 3.9 0.7 3.0 0.1 9.5 0.8 1.4

t⫺1 Coef. 8.9 0.5 3.2 2.0 3.0 4.1 5.2 ⫺0.0 2.4 1.8 3.9 1.7 0.7 4.2 3.8 0.8

t Stat.

0.6

⫺0.0

0.5

0.3

0.7

0.4

0.4

0.9

R2

Table 12: OECD Unemployment Projections–Time Series

0.7–1.2 ⫺1.3–2.0 0.2–1.0 ⫺0.1–2.0 0.1–0.7 2.0–6.8 0.6–1.4 ⫺4.3–4.2 0.0–1.0 ⫺1.0–8.8 0.3–1.1 ⫺0.9–7.0 ⫺0.3–0.6 4.5–14.5 0.3–1.3 ⫺2.8–5.7

95% Conf. interv.

11

13

13

13

13

13

13

13

N 0.9 0.4 1.0 0.0 0.6 3.2 1.0 ⫺0.2 0.8 1.1 0.9 1.0 0.9 1.0 1.0 0.1

t ⫺ 1/2 Coef. 18.1 1.0 11.3 0.0 4.6 3.5 11.0 ⫺0.2 8.0 1.1 10.9 1.2 6.7 0.7 16.4 0.2

t Stat.

0.96

0.8

0.9

0.8

0.9

0.6

0.9

0.96

R2

0.8–1.0 ⫺0.4–1.2 0.8–1.2 ⫺0.5–0.5 0.3–0.8 1.2–5.3 0.8–1.2 ⫺2.1–1.8 0.6–1.0 ⫺1.1–3.3 0.7–1.1 ⫺0.8–2.8 0.6–1.2 ⫺2.2–4.2 0.8–1.1 ⫺1.0–1.3

95% Conf. interv.

12

14

14

14

14

14

14

14

N

ACCURACY OF OECD AND IMF PROJECTION 75

76

M. E. Kreinin

Table 13: OECD Projection of Unemployment–Time Series Naive Model

Country United States Japan Germany France United Kingdom Canada Italy Australia

Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant

t⫺1 Coef.

t Stat.

0.7 1.8 0.9 0.2 1 0.3 1 0.4 0.9 0.7 0.9 1.4 0.9 0.6 0.9 0.7

6.0 2.3 13.0 1.1 18.6 1.0 29.9 1.7 13.5 1.5 10.0 2.0 22.4 1.5 13.4 1.5

R2 0.6 0.9 0.9 0.9 0.9 0.8 0.9 0.9

95% Conf. interv. 0.5–0.9 0.2–3.4 0.8–1.1 ⫺0.1–0.5 0.8–1.1 ⫺0.3–0.9 0.9–1.1 ⫺0.1–0.9 0.7–1.1 ⫺0.3–1.7 0.7–1.0 ⫺0.1–2.8 0.9–1.0 ⫺0.2–1.3 0.8–1.1 ⫺0.2–1.6

N 27 27 27 27 27 27 27 27

1.0 0.1 1.04 0.94

Coef. Constant Coef. Constant

t⫺1 Coef.

60.6 1.9

62.3 0.2

t Stat.

0.9

0.9

R2 (i) Projections 522

N

(ii) Naive Model 1.01–1.07 572 ⫺0.33–1.92

0.96–1.02 ⫺0.9–1.2

95% Conf. interv.

1.0 ⫺0.8

t ⫺ 1/2 Coef.

85.9 ⫺1.6

t Stat.

Table 14: OECD Projections of Trade Balances: Cross-Section and Time Series Combined

0.95

R2

0.9–1.0 ⫺1.8–0.2

95% Conf. interv.

527

N

ACCURACY OF OECD AND IMF PROJECTION 77

Australia

Italy

Canada

United Kingdom

France

Germany

Japan

United States

Country Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant

1.0 ⫺4.9 1.0 2.6 0.8 5.4 0.7 ⫺1.4 0.9 ⫺1.5 0.5 2.6 0.5 0.9 0.3 ⫺0.3

t⫺1 Coef. 18.6 ⫺1.1 15.3 0.6 7.8 1.4 5.1 ⫺1.4 10 ⫺1.0 4.3 2.4 4.2 0.5 1.3 ⫺0.6

t Stat.

0.0

0.4

0.4

0.8

0.5

0.7

0.9

0.9

R2

95% Conf. interv. 0.9–1.1 ⫺13–4 0.9–1.1 ⫺5.9–11.1 0.6–1.0 ⫺2.5–13.3 0.4–0.9 ⫺3.4–0.6 0.7–1.1 ⫺4.4–1.5 0.3–0.8 0.3–5.0 0.3–0.8 ⫺2.5–4.3 ⫺0.1–0.8 ⫺1.2–0.6

Table 15: OECD Projections of Trade Balances: Time Series

19

26

26

26

26

26

26

26

N 1.0 ⫺1.1 1.0 0.8 0.9 1.2 0.9 ⫺1.1 0.9 ⫺0.8 0.7 1.8 0.9 0.9 0.6 ⫺0.5

t ⫺ 1/2 Coef. 34 ⫺0.5 32.2 0.4 13.5 0.5 8.2 ⫺1.6 8.9 ⫺0.5 7.4 2.3 7.9 0.9 3.5 ⫺1.4

t Stat.

0.4

0.7

0.7

0.7

0.7

0.9

0.97

0.98

R2

0.97–1.1 ⫺6.0–3.6 0.93–1.05 ⫺3.2–4.8 0.7–1.0 ⫺3.7–6.3 0.7–1.1 ⫺2.5–0.3 0.7–1.1 ⫺4.0–2.4 0.5–0.9 0.2–3.4 0.7–1.1 ⫺1.3–3.2 0.3–1.0 ⫺1.2–0.3

95% Conf. interv.

18

28

28

28

28

23

23

23

N

78 M. E. Kreinin

ACCURACY OF OECD AND IMF PROJECTION

79

Table 16: OECD Projections of Trade Balances: Time Series Naive Model

Country United States Japan Germany France United Kingdom Canada Italy Australia

Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant Coef. Constant

t⫺1 Coef.

t Stat.

1.0 ⫺6.2 1.0 3.5 0.8 5.8 0.8 0.5 0.9 ⫺1.9 0.8 1.4 0.9 1.2 0.4 ⫺0.1

15.1 ⫺1.2 17.0 1.0 7.8 1.6 4.9 0.5 8.7 ⫺1.2 7.3 1.7 5.1 0.8 1.5 ⫺0.2

R2 0.9 0.9 0.7 0.5 0.7 0.7 0.5 0.1

95% Conf. Interv. 0.9–1.1 ⫺1.6–4.0 0.9–1.2 ⫺3.7–10.8 0.6–1.1 ⫺1.6–13.1 0.4–1.1 ⫺2.7–1.6 0.7–1.1 ⫺5.0–1.2 0.6–1.1 ⫺0.3–3.1 0.6–1.3 ⫺1.7–4.1 ⫺0.1–0.9 ⫺0.9–0.8

N 28 28 28 28 28 28 28 20