Policy-ineffectiveness proposition

Abstract: A heated debate has arisen over what Modigliani has dubbed the Macro Rational Expections (MRE) hypothesis. This hypothesis embodies two component hypotheses: 1) rational expectations and 2) short-run neutrality -- i.e., that anticipated changes in aggregate demand will have already been taken into account in economic agents' behavior and will thus evoke no output or employment response. Together these component hypotheses imply that deterministic feedback policy rules will have no effect on business cycle fluctuations. The irrelevance of these types of policy rules is inconsistent with much previous macro theorizing as well as with the views of policymakers. It is thus an extremely controversial proposition which requires a wide range of empirical research. This paper is a sequel to a previous paper by the author. That paper developed a methodology for testing the MRE hypothesis and found that anticipated money growth does matter to the business cycle. This paper extends the analyses to cases where the rate of nominal GNP growth or the inflation rate, rather than money growth, is the aggregate demand variable. The empirical results are also negative on the MRE hypothesis and its corresponding policy ineffectiveness proposition.

Abstract: T HE combination of the natural rate of unemployment and the rational expectations hypotheses in macroeconomic models results in the now well-known conclusion that anticipated short-run monetary stabilization policies do not influence real economic variables. This policy ineffectiveness proposition has been named the Macro Rational Expectations (MRE) hypothesis by Modigliani (1977). As a result, the implicit policy prescription is that a price level stabilization policy should be pursued. The empirical validity of this conclusion is perhaps the central issue in modern stabilization theory. Barro (1977, 1978, 1981), Barro and Rush (1980), and Leiderman (1980) have presented empirical evidence for the United States supporting this proposition. However, a number of recent studies have claimed that previous empirical results which support the finding that anticipated policy is neutral are flawed. Corrections of these problems often result in conflicting conclusions (see Small (1979), Gordon (1979), and Mishkin (1982)). Given the important implications of the MRE hypothesis, it is curious that almost all formal testing has been focused upon the United States. While Barro has suggested that a cross-country study would be useful, as far as we know, only Darby (1980) attempts to test whether anticipated changes in monetary policy are neutral across countries. Unfortunately, his results suffer from problems caused by using generated regressors. In addition, he makes no attempt to test the individual implications of the MRE hypothesis which encompasses the propositions that expectations are rational (lad anticipated monetary policy does not matter. In this paper, we employ the methodology developed by Mishkin to test the MRE hypothesis in six countries-Canada, Germany, Italy, Japan, the United Kingdom, and the United States. The paper is divided into three major sections and a conclusion. Section II briefly reviews Mishkin's testing method. In section III the procedure employed to specify the money growth equation is discussed and these equations specified. Section IV presents and discusses the results of applying the method introduced in section II. In the conclusion, the major findings are restated, possible limitations discussed, and extensions suggested.

Abstract: RECENT WORK BY Mishkin (1982a, 1982b) and Gordon (1982) tests the Lucas-Sargent-Wallace (henceforth LSW) policy ineffectiveness proposition with econometric methods. This work generally finds support for the alternative proposition that there is some short-term influence of anticipated demand policy or anticipated money growth on output. As McCallum (1982) has pointed out, one difficulty with these studies is they presume movements in output and employment are best represented as stationary fluctuations around deterministic trends, in constrast to evidence given by Nelson and Plosser (1982), which suggests these variables are nonstationary processes that have no tendency to return to a deterministic path. If these series are nonstationary processes, then the distribution theory implicit in the hypothesis tests of the Gordon and Mishkin studies does not apply and their results are open to question. Also, as McCallum (1982, p. 4) points out, "the methods of overcoming the observational equivalence difficulty (Sargent 1976) are not entirely satisfying" in these studies. In effect, Gordon and Mishkin are forced to make some untested restrictions in order to distinguish the anticipated from the unanticipated components

Abstract: One of the most controversial macroeconomic developments of the last decade has been the rise of the so-called 'new classical' (NC) approach to macroeconomic theory and policy. Lucas (I972), Sargent and Wallace (I975), and Barro (I 976) have presented models which combine the Friedman (i968)-Phelps (I970) natural rate theory, with assumptions of market clearing and rational expectations. As is well known, these models deliver the striking result that the mean values of real variables cannot be affected by the government's monetary policy.' The publication of the theoretical work establishing the impotency result led to much empirical testing of the NC models. One of the most successful approaches from the point of view of NC advocates is contained in a series of papers by Barro (I977, I978, I98I), Barro and Rush (I980), and Rush (I986). Recently, Pesaran (I982) has questioned Barro's results and claimed that once an 'implausible assumption' is discarded, Barro's work can be rejected in favour of a Keynesian alternative. Because Pesaran has put forth an explicit counter-model to Barro's work, his results are potentially quite important. This paper examines in detail Pesaran's work and conclusions.