# Optimality on the short-run Phillips curve revisited.

Introduction

Golden, et al. (1987) noted that in post World War II elections in which an incumbent president won by a decisive margin, the inflation rate was lower than the unemployment rate by 2 to 4 percent.(1) To explain this phenomenon the authors proposed a misery index (the inflation rate plus the unemployment rate) criterion model. In minimizing the misery index, optimization requires that the unemployment rate exceeds the inflation rate. This was based on the assumption that voters are myopic in choosing between inflation and unemployment in the sense that they "ignore the long-run consequences of the chosen short-run inflation/unemployment trade-off." Why is this so? The authors conjectured that while both unemployment and inflation generate social and private costs, "the latter hurts a broader proportion of the population."

While it is true that in some countries the inflation rate rarely exceeds the unemployment rate, other countries have experienced the opposite. In Latin American countries, for example, the inflation rate has sometimes been in triple digits. The model presented by Golden, et al. apparently does not apply to these countries. This paper attempts to modify their model to incorporate the selection of high inflation/low unemployment by policy makers in Latin American countries. The key lies in the respective weights given to unemployment and inflation used to construct a weighted misery index. The next section presents this weighted misery index model.

Model and Results

The model consists of a weighted misery index:

WM = |w.sub.i~i + |w.sub.u~u (|w.sub.i~|is greater than or equal to~1, |w.sub.u~|is greater than or equal to~1) (1)

and the short-run Phillips Curve:

i = a + b/u (b|is greater than~0|is greater than~a) (2)

The weighted misery index (WM) is defined as the weighted sum of the inflation (i) and unemployment (u) rates. |w.sub.i~ and |w.sub.u~ are the weights set by the policy maker for the inflation and unemployment rates, respectively.(2) The weights reflect the preferences of the policy maker and their absolute size reflects the intensity of preferences. The short-run Phillips Curve is downward sloping with parameters a and b. The negativity of parameter a indicates that extremely high unemployment is associated with deflation. As the current model is a generalization of Golden, et al., we maintain all their behavioral assumptions except that these pertain to the policy maker here rather than to the voters. This is necessary for the generalized model to be applicable to countries where democracy might not yet prevail.

Minimizing WM with respect to either u or i yields an identical relationship between the optimal solutions of the unemployment rate (u*) and the inflation rate (i*):

u* = (|w.sub.i~/|w.sub.u~)(i* - a) (3)

As far as the policy maker's preferences are concerned, the choice of weighting scheme for the unemployment and inflation rates has three possibilities: (i) |w.sub.i~ |is greater than~ |w.sub.u~ |is greater than or equal to~l, (ii) |w.sub.u~|is greater than~|w.sub.i~|is greater than or equal to~l, and (iii) |w.sub.i~=|w.sub.u~=1. The optimal solution for options (i) and (iii) results in a rate of inflation lower than the rate of unemployment. However option (ii) may yield the opposite result.(3)

The model presented by Golden, et al. is a case in which the American policy maker views inflation and unemployment as equally important problems, i.e., option (iii) |w.sub.i~ = |w.sub.u~ = 1. The results of Gallup opinion polls suggest that this is a justifiable categorization of the American policy maker's preferences (Dornbusch and Fischer, 1987, p. 576).(4)

The prevailing phenomenon of high inflation/relatively low unemployment in Latin American countries can be explained by the policy option (ii), i,e., |w.sub.u~|is greater than~|w.sub.i~|is greater than or equal to~1. For example, Brazil during the military regime, i.e., 1964-80, followed a more or less consistent strategy that prized rapid economic growth (Roett, 1984, p. 167; Skidmore, 1988, p. 208).(5) The result was an average annual rate of inflation of 31.3% for the period 1965-80 (World Bank, 1988, Table 1),(6) while the unemployment rate was no more than 3% (International Labor Office, 1988, Table 9). On the other hand, West Germany is a good example of policy option (i), i.e., |w.sub.i~|is greater than~|w.sub.u~|is greater than or equal to~1. It has been documented that German policy makers showed a stronger desire to combat inflation (rather than unemployment) than their counterparts in other European countries (Bismut and Kroger, 1985, p. 377).(7)

Conclusion

Weighting the misery index in accordance with the policy maker's preferences for inflation versus unemployment generalizes the model so that it can be applied to a broad range of countries. This generalized model includes that of Golden, et al. as a special case and can also be used to explain the experience of Latin American countries where the double/triple digit inflation rate has been far above the unemployment rate.

Notes

1. These years included 1956, 1964, 1972, and 1984. In 1980, when the inflation rate was far above the unemployment rate, the incumbent president lost by a landslide. All these years are consistent with the misery index criterion model proposed by Golden, et al. (1987). The unemployment rate and the inflation rate (defined as the annual change in the consumer price index, following the Golden, et al.'s paper) for those years are respectively, 4.1%/1.5% in 1956, 5.2%/1.3% in 1964, 5.6%/3.2% in 1972, 7.2%/13.5% in 1980, and 7.5%/4.3% in 1984. (Source: Economic Report of the President 1989, Tables B-31 and B-55).

2. The reason that the weights used in this paper are at least equal to one instead of being between 0 and 1 is to match the original definition of misery index by Okun (see Golden, et al., note 1). However, this choice will not affect the conclusion of this paper.

3. There is a chance that, even with the choice of |w.sub.u~|is greater than~|w.sub.i~|greater than or equal to~1, the optimal result of the choice might still be u*|is greater than~i* if i* |is less than~ -a|w.sub.i~/(|w.sub.u~-|w.sub.i~). The reason for this possibility is the presence of parameter a. The absolute order of magnitude of this parameter can only be decided empirically.

4. According to Dornbusch and Fischer (1987, p. 576), in virtually every Gallup opinion poll between 1973 and 1983, more than 50 percent of respondents rate either inflation or unemployment as the most serious problem facing the U.S. economy. This is consistent with the assumption that U.S. voters see inflation and unemployment as equally important problems, and, therefore, that the policy maker takes the same position.

5. Brazil was chosen as an example to illustrate policy option (ii) because its two decade-long military rule followed a more or less consistent policy strategy: favoring economic growth. The resulting high inflation/low unemployment phenomenon prevailing in those years has been documented as the consequence of conscious choices on the part of policy makers in Brazil (e.g., Roett, 1984; Skidmore, 1988). Some might argue that favoring economic growth is not equivalent to favoring unemployment unless wages are not perfectly flexible (Quandt and Rosen, 1988, p. 9). In light of the fact that indexation or prefixed indexation of wages is a common income policy in Brazil (e.g., Skidmore, 1988, p. 206, and p. 230), one may conclude that the assumption of equivalency of favoring economic growth to favoring unemployment is plausible in the Brazilian case. It should be also noted that the military rule in Brazil ended in 1985. However, the first half of the 1980s was not considered here due to the fact that Brazil after 1982 was subject to the austerity program prescribed by the International Monetary Fund because of Brazil's international debt crisis (Skidmore, 1988, p. 237).

6. The inflation rate was measured by the World Bank as the annual rate of change in the Gross Domestic Product Deflator.

7. Bismut and Kroger (1985, p. 377) indicated in their recent econometric study on the dilemmas of economic policy (including the trade-off relationship between unemployment and inflation) in France and Germany that, during the 1975-1979 period, "An annual rate of unemployment of 5% corresponds to a rate of inflation of 11% in France and 4% in Germany. These figures are reasonably consistent with the historical record."

References

Council of Economic Advisors, Economic Report of the President 1989, Washington, D.C.: Printing Office of Government, 1989.

Dornbusch, R. and Fischer, S. Macroeconomics, 4th ed., New York: McGraw-Hill, 1987, p. 576.

Golden, J. M., Orescovich, R. and Ostafin, D. "Optimality on the Short-Run Phillips Curve: A "Misery Index" Criterion, A Note." The American Economist, 31 (Fall 1987), p. 72.

International Labor Office, Yearbook of Labor Statistics 1988, Geneva: International Labor Office, 1988.

Quandt, R. E. and Rosen, H. S. The Conflict Between Equilibrium and Disequilibrium Theories, Kalamazoo, MI: W. E. Upjohn Institute, 1988.

Roett, R. Brazil Politics in a Patrimonial Society, 3rd ed., New York: Praeger, 1984.

Skidmore, T. E. The Politics of Military Rule in Brazil, 1964-85. New York: Oxford University Press, 1988.

World Bank, World Development Report 1988, New York: Oxford University Press, 1988.

Bismut, C. and Kroger, J. "The dilemmas of economic policy in France and Germany: trade-offs between inflation, unemployment, and the current account." in G. De Menil and U. Westphal, eds., Stabilization Policy in France and the Federal Republic of Germany, Amsterdam: North-Holland, 1985, pp. 303-347.

Drexel University. I am greatly indebted to an anonymous referee for his/her suggestions for better illustrations of my generalized model. Nonetheless, all the remaining errors, if any, are mine.