The aim of this study is to contribute to the measurement and analysis of errors in economists' predictions of changes in aggregate income, output, and the price level. Small sample studies of forecasts can be instructive, but their limitations must be recognized. Compilation of consistent forecast records extending over longer periods of tine is necessary to establish a reasonably reliable base for assessments of forecasting behavior and. performance. Thus the historical record of post-World War II forecasts assembled in the 1960's by the NBER is here extended and updated.
While most people would agree that familial responsibilities affect women's labor market behavior and wages, surprisingly little is known about how this process operates. Past investigations of women's wages have generally relied on data sets designed for other purposes, and, as a result, theoretically important determinants of women's wages may be measured imprecisely or omitted entirely from analyses. Familial responsibilities influence women's labor market behavior in at least two distinct ways. First, many women will withdraw entirely from labor market activities to bear and/or raise children. Not only does this reduce the total amounts of work experience and job tenure women acquire, but Jacob Mincer and Solomon Polachek further argue that women's human capital (work skills) will depreciate during such withdrawals and that such withdrawals will affect the timing of women's investments in on-the-job training. Second, women who choose to work may adjust their labor market activities to meet family responsibilities in ways which reduce productivity and hence wages. For instance, women with home responsibilities might restrict job locations or schedules, or might take off extra time from work to care for sick children. The 1976 Panel Study of Income Dynamics (PSID) is well suited for exploring female wages. The PSID is a longitudinal study of 5,000 families which began in 1968. In 1976, male heads of household, female heads of household, and wives were asked to provide detailed information on earnings, education, work history, absenteeism, and self-imposed restrictions on job hours and job location. These data are used to describe women's patterns of work history and labor force attachment, to specify the determinants of women's wages, and to investigate the wage gaps between white men and each group of women.
In an article in this Review, Christopher Sims presented an innovative statistical technique to determine the direction of causality, then applied this methodology to money and nominal income in the United States. He concluded: main empirical finding is that the hypothesis that causality is unidirectional from money to income agrees with the postwar U.S. data, whereas the hypothesis that causality is unidirectional from income to money is rejected (p. 540). In a more recent paper in this Review, David Williams, C. A. E. Goodhart, and D. H. Gowland applied Sims' statistical methodology to the United Kingdom and concluded: found for the U.K. some evidence of unidirectional causality running from nominal incomes to money but also some evidence of unidirectional causality running from money to prices. Taken together, this evidence suggests, perhaps, a more complicated causal relationship between money and incomes in which both are determined simultaneously (p. 423). Furthermore, Williams, Goodhart, and Gowland suggest some general possibilities for the differences between the United States and the United Kingdom, and they are careful to note that: Because of the various differences in context the finding that in the United Kingdom the relationship between money and income appears different from that found by Sims for the United States in no way casts any doubt on the validity of Sims' own (p. 417). The purpose of this paper is to present a concise model which draws together the findings of Sims for the United States and Williams, Goodhart, and Gowland for the United Kingdom. To accomplish this, a fixed exchange rate system is modeled in which one country, the United States, serves as the primary reserve currency country, while other countries, the United Kingdom in this case, hold a substantial portion of their international reserves denominated in terms of the reserve currency.' Particular attention is paid to the asymmetrical nature of the system with respect to money's influence on nominal income and vice versa. Indeed, the ability of the reserve currency country to create international reserve assets plays the primary role in explaining the asymmetrical nature of the system and the empirical results of Sims, and Williams, Goodhart, and Gowland. The model is couched in a world in which asset reallocations are viewed as adjustments toward maintaining general equilibrium. This equilibrium is based on a stable set of preferences regarding the structure of individual portfolios, broadly defined in terms of holdings of real consumption goods, real interest bearing financial assets, real money balances, and leisure time. The assumption of equilibrium conditions in all markets allows attention to focus directly on the money market to isolate the process of portfolio adjustment in international markets. As in similar models based on the monetary approach to *Economists, Chase Manhattan Bank, N.A. The views expressed in this paper are solely our own and do not necessarily represent those of the Chase Manhattan Bank. We wish to thank David T. King, the managing editor of this Review, J. Richard Zecher, J. E. Tanner, Walton T. Wilford, C. A. E. Goodhart, and Marc A. Miles for their comments on earlier drafts. 1One may note that for the Commonwealth countries the British pound acted as a reserve currency. However, the pound's relative world influence vis-avis the U.S. dollar was small during the Bretton Woods period.
The author uses the life-cycle growth model to clarify the implications of government involvement in capital accumulation, arguing that a long-run trade-off between consumption possibilities is critical to long-run optimality. The long-run capital/labor ratio is shown to determine the amount that each member of a given generation will consume in each period of his lifetime. With the option of redistributing income between generations, the optimal path of a centrally planned economy is less-restrictive. This is not true in the case of government activities financed by debt, which suggests the government's desired role at present is to provide a mechanism for redistributing income between the younger and older generations.
It is common to assume that expectations about future inflation influence economic behavior. The behavior of an economic system over time is determined in part by the manner in which these expectations are formed. Theoretical and empirical investigations frequently utilize a hypothesis that agents forecast future inflation rates primarily on the basis of past inflation rates: static, adaptive, extrapolative, and regressive expectations adjustment schemes are examples of such forecasting rules.1 But such rules generally ignore the fact that individuals observe prices, not inflation rates. A change in observed prices can result from general price inflation, but it may also result from transitory shocks to the price level or observation error. Similarly, apparent changes in the general inflation rate may be purely transitory in nature or may signify the beginning of a trend. A model of inflation expectations should admit the possibility of these different sources of price change and embody the natural human tendency to extrapolate seeming trends. This paper examines the extent to which agents' reported price expectations are consistent with such naive forecasting. A multilevel adaptive expectations model is developed that takes observed prices as the information available to agents. On the basis of these prices, individuals revise their beliefs about not only the price level but also the underlying inflation rate and trend in the inflation rate. When fitted to the Livingston survey data the model appears to provide a unified explanation of price expectations in the United States from 1947 to 1975. This is contrasted with earlier investigations suggesting that expectations formation differed significantly in the periods before and after 1960.