¹ A detailed review of the theoretical and empirical literature on Phillips-curve inflation models is given in my Economic Interest and the Politics of Macroeconomic Policy, No. C/75-14, Center for International Studies, M.I.T., Cambridge, Mass., 01 1976 Google Scholar. Copies of this monograph are available at cost from the C.I.S. Publications officer.

² Ibid.

³ See, for example, Blinder, A. and Esaki, H., “Macroeconomic Activity and Income Distribution in the Postwar U.S.” (mimeo., 11 1976)Google Scholar; Hollister, Robinson G. and Palmer, John L., “The Impact of Inflation on the Poor,” in Redistribution to the Rich and the Poor, ed. Boulding, K. E. and Pfaff, M. (Belmont, Calif.: Wadsworth, 1972), pp. 240–70Google Scholar; Metcalf, Charles E., An Econometric Model of the Income Distribution (Chicago: Markham, 1972)Google Scholar; Thurow, Lester C., “Analyzing the American Income Distribution,” American Economic Review: Papers and Proceedings, 60 (05 1970), 261–69Google Scholar; and Schultz, T., “Secular Trends and Cyclical Behavior of Income Distribution in the United States: 1944–1964,” in Six Papers on the Size Distribution of Wealth and Income, ed. Soltow, L. (New York: National Bureau of Economic Research, 1969), pp. 75–100 Google Scholar.

⁴ These hypotheses have a distinguished pedigree. They have appeared, among other places, in Hamilton, Earl J., “Prices and Progress,” Journal of Economic History, 12 (Fall 1952), 325–49CrossRefGoogle Scholar; Hansen, Alvin, “Factors Affecting the Trend of Real Wages,” American Economic Review, 15 (03 1925), 27–42 Google Scholar; Keynes, John Maynard, The General Theory of Employment, Interest and Money (New York: Harcourt, 1936)Google Scholar; Rueff, Jacques, “Nouvelle discussion sur le chomage, les salaires et les pris,” Revue d Economie Politique (1951), 761–91Google Scholar; and Weintraub, Sidney, An Approach to the Theory of Income Distribution (Westport, Conn.: Greenwood Press Inc., 1958)Google Scholar. Weintraub, , for example, has flatly asserted that “… only entrepreneurs and the actual unemployed have an unequivocal stake in maximum employment, while rentiers and the employed find their interests better served at lower levels of activity,” p. 60 Google Scholar.

⁵ Long, Clarence D., “The Illusion of Wage Rigidity: Long and Short Cycles in Wages and Labor,” Review of Economics and Statistics, 42 (05 1960), 140–51CrossRefGoogle Scholar.

⁶ Bodkin, Ronald G., “Real Wages and Cyclical Variations in Employment: A Re-Examination of the Evidence,” Canadian Journal of Economics, 2 (02 to 11 1969), 353–74CrossRefGoogle Scholar.

⁷ Bach, G. L. and Stephenson, James B., “Inflation and the Redistribution of Wealth,” Review of Economics and Statistics, 61 (02 1974), 1–13 CrossRefGoogle Scholar; Boddy, Raford and Crotty, James, “Class Conflict and Macro-Policy: The Political Business Cycle,” Review of Radical Political Economics, 7 (Spring 1975), 1–19 CrossRefGoogle Scholar; Burger, Albert, “Relative Movements in Wages and Profits,” Federal Reserve Bank of St. Louis Review, 55 (02 1973), 8–16 Google Scholar; Hibbs, “Economic Interest”; Kuh, Edwin, “Income Distribution and Employment over the Business Cycle,” in Brookings Quarterly Econometric Model of the United States, ed. Dusenberry, J. et al. (Chicago: Rand McNally, 1965), pp. 227–78Google Scholar; Hultgren, Thor, Costs, Prices, and Profits: Their Cyclical Relations (New York: National Bureau of Economic Research, 1965)Google Scholar; and OECD, Inflation, The Present Problem (Paris: OECD Publications, 1970)Google Scholar.

⁸ This has been proposed, for example, in Nordhaus, William D., “The Political Business Cycle,” Review of Economic Studies, 42 (04 1975), 169–90CrossRefGoogle Scholar.

⁹ As one White House economist reportedly put it in April of 1975 “One hundred percent of the people have been hit by inflation. Only 10 percent really worry about unemployment.” Quoted by Golden, S., “High Joblessness Expected to Persist as a Condition of U.S. through Decade,” New York Times (04 21, 1975), p. 46 Google Scholar.

¹⁰ Katona, George, “Disputing Galbraith,” New York Times (12 22, 1974)Google Scholar, and Katona, George et al., Aspirations and Affluence (New York: McGraw-Hill, 1971)CrossRefGoogle Scholar. Many of the results from these surveys appear in annual volumes of the Survey of Consumer Finances (Ann Arbor: Survey Research Center, 1960–72)Google Scholar. Results of surveys taken before 1960 are available as mimeo reports from the SRC.

¹¹ The analyses are presented fully in the section “Public Opinion Toward Inflation and Unemployment” in Hibbs, , “Economic Interest,” pp. 24–40 Google Scholar.

¹² The class interests at stake in unemployment/inflation outcomes and policies show up in the policy positions taken by organized labor and capital as well as in the distribution of mass opinion. Throughout the postwar period, trade union spokesmen have invariably placed primary emphasis on the objective of full employment, while business elites have attached far more importance to price stability. A clear statement of labor's position is given by Goldfinger, Nat, “Full Employment: The Neglected Policy?” The American Federationist, 79 (11 1972)Google Scholar. Data on corporate thinking on the inflation and unemployment issues is presented in Silk, L. and Vogel, D., Profits and Principles: The Social and Political Thinking of American Businessmen (New York: Simon and Schuster, 1977)Google Scholar.

¹³ Kirschen, E. S. et al., Economic Policy In Our Time, Vol. I (Amsterdam: North-Holland, 1964)Google Scholar. With the exception of the balance of payments issue (the importance of which depends critically on the international economic position of a given nation), the positions attributed to the various tendances were homogeneous, across countries. For a similar scheme, see Frey, Bruno and Lau, Lawrence J., “Towards a Mathematical Model of Government Behaviour,” Zeitschrift für Nationalökonomie, 28 (1968), 355–80CrossRefGoogle Scholar.

¹⁴ Canada's New Democratic Party, a genuinely socialist party with close connections to organized labor, has exhibited increasing political vitality in recent years (capturing several provincial governments) but remains at this writing a “minor” party with little influence on national policy.

¹⁵ Actually there was one brief period of Socialist-led rule in France after 1951: Guy Mollet's government of February 1956 to May 1957. Analysis of annual data shows that unemployment was lower and inflation higher during Mollet's government (as well as during the subsequent Center-Left government of Bourges-Maunoury) than during the right-wing Gaullist governments of the late 1950s and 1960s. The Center-Left governments of the middle 1950s clearly assigned higher priority to full employment and expansion than the Gaullist regime, which pursued policies geared to disinflation and economic “stabilization.” As a result, France's location on the “international Phillips curve” has changed dramatically. (Contrast the data shown in Figure 1 to a similar display of average rates of inflation and unemployment reported by Smyth, D., “Unemployment and Inflation: A Cross-Country Analysis of the Phillips Curve,” American Economic Review, 61 (06 1971), 426–29Google Scholar, for the period 1950–1960.) Of course France's entry into the EEC in 1958 increased the importance of the external balance-of-payments constraint during the Fifth Republic. However, the deflationary policies of the Gaullist governments must be attributed to some extent to the priorities of the regime. See Maclennan, M et al., Economic Planning and Policies in Britain, France and Germany (New York: Praeger, 1968)Google Scholar.

¹⁶ Since the macroeconomic policies (and outcomes) of the 1960s were to a significant extent influenced by the performance record of the late 1940s and 1950s (especially in countries in which Social Democratic-led governments managed to maintain full employment after the war), the Socialist-Labor participation rate has been calculated over the entire postwar period (1945–69) rather than for the years 1960 to 1969 alone.

¹⁷ This has been suggested, for example, in reference to the difference in unemployment rates between North America and Western Europe, by Rees, Albert, “The Phillips Curve as a Menu for Policy Choice,” Economica, 37 (08 1970), 227–38CrossRefGoogle Scholar. Monetary policy instruments include interest rates and the supply of credit and money. Fiscal policy instruments include taxation and public spending.

¹⁸ Flanagan, Robert J., “The U.S. Phillips Curve and International Unemployment Rate Differentials,” American Economic Review, 63 (1973), 114–31Google Scholar. For additional evidence on cross-national variation in Phillips curves, see Bodkin, Ronald G. et al., Price Stability and High Employment: The Options for Canadian Economic Policy (Ottawa: Economic Council of Canada, 1967)Google Scholar.

¹⁹ Of course leftist governments have not been equally effective in this regard. For example, British Labour governments have been much less imaginative in developing macroeconomic policy (and have pursued a more centrist political strategy) than Swedish Social Democratic administrations. See the perceptive comparative analysis in Martin, Andrew, The Politics of Economic Policy in the U.S.: A Tentative View from a Comparative Perspective (Beverly Hills: Sage Professional Paper in Comparative Politics, 1973)Google Scholar. The best treatment in English of the archetypal Swedish model is probably Lindbeck, A., Swedish Economic Policy (Berkeley: University of California Press, 1974)CrossRefGoogle Scholar.

²⁰ See, for example, Alford, Robert, Party and Society (Chicago: Rand-McNally, 1963)Google Scholar.

²¹ For an argument that organized labor and the Democratic party in the United States are interpenetrated in a way that is at least partially equivalent to Socialist party-labor union alliances in much of Westem Europe, see Greenstone, J. D., Labor in American Politics (New York: Alfred A. Knopf, 1969)Google Scholar.

²² See Box, G. E. P. and Jenkins, G. M., Time Series Analysis; Forecasting and Control (San Francisco: Holden-Day, 1970), part IIIGoogle Scholar; and Box, G. E. P. and Tiao, G. C., “Intervention Analysis with Applications to Economic and Environmental Problems,” Journal of the American Statistical Association, 70 (03 1975), 70–79 CrossRefGoogle Scholar. The scheme of Box, Jenkins, and Tiao is contrasted with the conventional structural equation approach in Hibbs, Douglas A. Jr., “On Analyzing the Effects of Policy Interventions: Box-Jenkins and Box-Tiao vs. Structural Equation Models,” in Sociological Methodology 1977, ed. Heise, D. (San Francisco: Jossey Bass, 1977), pp. 137–79Google Scholar.

²³ The cyclical or seasonal component of the model is not represented explicitly by the ARMA terms of eq. (1).

²⁴ The one quarter lag on G_t may be too short, especially for the United States. However, since the intervention function allows U to respond gradually to shifts in G, this is not an important problem.

²⁵ The ARMA model building process is systematically reviewed in Hibbs, “On Analyzing Policy Interventions,” and developed in great detail by Box and Jenkins, Time Series Analysis.

²⁶ Sample autocorrelations are simply the correlations between observations separated k periods in time and are given by:

Thus r₁ denotes the correlation between U_t and U_t—1 ; r ₂ denotes the correlation between U _t and U_t— ; and so on.

²⁷ The British unemployment data (wholly unemployed as a percentage of the civilian labor force) were, obtained from the Ministry of Labour Gazette, various issues. In view of the unprecedented exogenously imposed economic crisis facing advanced industrial societies since 1973, the time series analyses are intentionally not taken beyond the fourth quarter of 1972.

²⁸ For example, it is estimated that the earnings-related benefits increased the unemployment income of a typical married male worker with two children from about 40 percent to 60 percent of average employment income. See OECD, Manpower Policy in the United Kingdom (Paris: OECD Publications, 1970)Google Scholar.

²⁹ Unfortunately the picture is complicated by the fact that a number of other macroeconomic policy changes were implemented during the 1965–67 period. These policy changes are reviewed by Bowers, et al., in “The Change in the Relationship Between Unemployment and Earnings Increases: A Review of Some Possible Explanations,” National Institute Economic Review (11 1970), 44–63 Google Scholar. However, the survey-based analysis of MacKay, D. and Reid, G. in “Redundancy, Unemployment and Manpower Policy,” Economic Journal (12 1972), 1256–72CrossRefGoogle Scholar, leaves little doubt that the new compensation law had a significant effect on the duration (and thus the rate) of unemployment. Also see the discussion by Feldstein, M., “The Economics of the New Unemployment,” Public Interest, 33 (Fall 1973), 3–42 Google Scholar.

³⁰ The models in this section were estimated with Kent D. Wall's ERSF program, which provides Full Information Maximum Likelihood estimates of Rational Distributed Lag Structural Form equations. Details are given in Wall, , “FIML Estimation of Rational Distributed Lag Structural Form Models,” Working Paper No. 77 (Cambridge: National Bureau of Economic Research, Inc., 03, 1975)CrossRefGoogle Scholar.

³¹ The predicted level data are obtained by summing the predicted four-quarter difference series, i.e.,

The summation operator Σ is the inverse of the difference operator (1–L) in the same way that integration is the inverse of differentiation in continuous time problems.

³² The lag 4 autocorrelation is of course significant and therefore the model might be improved by specifying where the ν _t are . Since the k=4 autocorrelation was essentially induced by the seasonal differencing (which overcompensates for the four-quarter seasonal dependency), and we are primarily interested in predicting the level unemployment series, modification of the model in this way is not advantageous.

³³ Meyers, R. J., “The Unemployment Problem: What We Can Learn from European Experience,” in Measuring Employment and Unemployment by the Joint Economic Committee of the U.S. Congress (Washington, D.C.: Government Printing Office, 1963)Google Scholar.

³⁴ Equation (8) is nearly identical to the model developed by C. R. Nelson for quarterly U.S. unemployment data over the period 1948(1)–1966(4). Nelson's model, incidentally, outperformed the MIT-FRB-Penn econometric model in short-term forecasting experiments. “The Predictive Performance of the FRB-MIT-PENN Model of the U.S. Economy,” American Economic Review (1972), 902–17Google Scholar.

³⁵ Since the unemployment data did not exhibit a trend over the observation period, all variables were deviated from their means and the model was estimated without a constant term. θ₀ therefore does not appear in Table 3.

³⁶ The t ratio of = 1.23 and of = 1.60; both are insignificant at the .05 level. Computation of the implied dynamic response of the unemployment rate to American involvement in the Korean and Vietnamese civil wars is therefore problematic. Robert Solow has suggested to me that since the effects of both the war term and the administration term work through the actual tax, expenditure, and monetary actions of the government, the model might be better specified by constraining δ₁ = δ₂. However, estimates obtained by imposing this constraint did not alter the results reported in Table 3 and graphed below in Figure 12 appreciably: the war coefficient remained insignificant, =.969, and = –.091.

³⁷ The negative residual autocorrelation at k = 8 (r8〈ât〉 = –.253) indicates that there is a modest, negative two-year (8 quarter) dependency between U.S. unemployment rates. This is compatible with the political-electoral business cycle argument of Nordhaus, Tufte and others, in which unemployment tends to fall before Presidential elections and to rise thereafter in response to administration efforts to engineer favorable economic conditions just prior to elections and to postpone austerity measures until after elections are safely over. [Nordhaus, , “The Political Business Cycle;” and Tufte, Edward, Elections and Economics: Macroeconomics Under Conditions of Political Competition (Princeton, N.J.: Princeton University Press, forthcoming)Google Scholar.] If this pattern was strong and more or less uniform across four-year presidential administrations, we should observe a sizable negative autocorrelation at k = 8 (two-year intervals) and a positive autocorrelation at k = 16 (four-year intervals). Although the focus of this study is on long-run patterns in macroeconomic policies and outcomes that distinguish left- and right-wing regimes, attempts were made to build an electoral unemployment cycle of this sort into the model. However, elaborations of the model along these lines did not yield significant results.

³⁸ This is readily confirmed by evaluating the expression over the index i for fixed G_t. Unlike the British results, which implied convergence to steady state after only 16 quarters, the U.S. steady state is not reached until the index i is taken to well over 100 quarters.

³⁹ The results graphed in Figure 12 were obtained in the same way as described earlier for the British case, i.e., by simulating

for G_t held at +1 and then –1 over regimes of 32 quarters (8 years).

⁴⁰ See, for example, Eckstein, O., “Economic Policy in the United States,” in Economic Policy in Our Time, Vol. II, ed. Eckstein, O. et al. (Amsterdam, North Holland, 1964), pp. 1–88 Google Scholar; and especially Stein, H., The Fiscal Revolution in America (Chicago: University of Chicago Press, 1969), Ch. 11-14Google Scholar.

⁴¹ Many analysts argue that Eisenhower's fiscal policies not only did little to combat the economic contradictions of the period but were also important causes of the 1957–58 and 1960–61 recessions. See Lewis, W., Federal Fiscal Policy in the Postwar Recessions (Washington, D.C.: The Brookings Institution, 1962)Google Scholar; and Stein, The Fiscal Revolution in America.

⁴² Eisenhower, D., Waging Peace, 1956–61 (Garden City, N.Y.: Doubleday, 1965), pp. 461, 462 Google Scholar.

⁴³ Cited in Dulles, F. R., Labor in America (New York: Crowell 1966), p. 394 Google Scholar.