This paper is a critical review of and a reader's guide to a collection of papers by Robert E. Lucas, Jr. about fruitful ways of using general equilibrium theories to understand measured economic aggregates. These beautifully written and wisely argued papers integrated macroeconomics, microeconomics, finance, and econometrics in ways that restructured big parts of macroeconomic research. (JEL A31, E00, E13, E50)

I. Introduction, 212. — II. An income-expenditures model, 214. — III. Conclusions, 224.

I. Introduction, 127. — II. A loanable funds model, 130. — III. Empirical results, 132. — IV. Macauley's criticism reconsidered, 138. — V. Conclusions, 139.

Monetary policy can be constrained by fiscal policy if fiscal deficits grow large enough to require monetization of government debt. That fact implies that the administrative independence of central banks does not by itself imply that monetary policy is independent of the fiscal decisions of governments. This essay describes limitations, possibilities, and suitable goals for monetary policy within the existing pattern of institutional responsibilities. The economic limitations of what can be achieved by monetary policy are summarized in six propositions developed in the paper.

Martin Feldstein and Otto Eckstein (1970) have set out and estimated a model of interest rate determination which they claim represents an integration of Keynes's liquidity preference theory with Irving Fisher's theory of the impact of expected inflation on interest rates. They achieve their integration by first inverting a Keynesian demand function for real balances, solving it for the nominal rate of interest. Then to incorporate Fisher's effect, they simply add to the right side of this equation a distributed lag in current and past actual rates of inflation, the same proxy for expected inflation that Irving used. Feldstein and Eckstein interpret sizable and statistically significant estimated coefficients on the elements of that distributed lag as confirming the presence of an effect of anticipated inflation on the interest rate. It is questionable whether Keynes's and Fisher's theories stand in need of any integration at all, since they are in principle compatible in the first place. The two theories are on very different footings. Keynes's liquidity preference theory is a theory about one particular structural equation relating real money balances, income, and the nominal rate of interest. On the other hand, Fisher's theory is one about how the whole economy is put together; that is, it is a statement about the reduced form equation for the nominal interest rate. In Fisher's theory, an exogenous increase in the anticipated rate of inflation is asserted to work its way through the economy in such a fashion that the nominal interest rate rises by the amount of the increase in anticipated inflation.1 In this note, I suggest that Feldstein and Eckstein's equation does not successfully synthesize Keynes and Fisher. Furthermore, I suggest that Feldstein and Eckstein's econometric procedure is not a good one for estimating the dimensions of the Fisher effect. In particular, the effect may be present in full force but still not be detected by Feldstein and Eckstein's procedure. On the other hand, it is possible to construct examples of economies in which there is really no effect but in which Feldstein and Eckstein's test would point to the presence of one. Finally, I show that in a model that includes both Keynes's liquidity preference schedule and a reduced form for the interest rate like the one posited by Fisher, Feldstein and Eckstein's equation is not statistically identifiable.

ECONOMISTS have recently begun to devote an increasing amount of attention to the relationships among interest rates on various instruments. Rates on instruments which differ with respect to maturity alone, all other features supposedly being held constant, have in particular received a great deal of attention since the publication of Meiselman's [7] work in 1962. While the term structure has certainly received the most concern, import-ant contributions have been made to the broader problem of studying relationships among rates on bonds which differ with respect to features other than maturity.' An especially important aspect of this broader area of research lies in examining the relationships between rates on government and corporate bonds. Knowledge of the characteristics of these relationships is important for an understanding of the paths through which monetary policy affects yields on corporate bonds, and hence, perhaps, expenditures on investment. This paper presents the results of an examination of the relationships among several interest rates on government and corporate bonds for the period January 1951 through December 1960. Monthly data are used, and series for the rates on three-month treasury bills, one, two, three, four, five, ten, and twentyyear government bonds, commercial paper, and Moody's Aaa's and Baa's are studied.2 This list represents an array of instruments ranging over a broad maturity spectrum and featuring various levels of quality. Tools of spectral and cross-spectral analysis are used to study the behavior of the series and the relationships among the series at various important components of oscillation.3 In addition to calculating the standard statistics associated with the spectrum and cross-spectrum, the relatively new tool of complex demodulation is employed to study the seasonal behavior of selected rates.4

where ut is a random disturbance with zero mean. Koyck pointed out that subtracting Xyt-i from (1) produced the equation yt = A'+xyt_i+ bxt+ut' (2) where A' = A(1 -X) and ut' = ut -Xut1. Thus instead of being forced to deal with the model in its distributed lag form (1), which involves the seemingly intractable task of estimating a relation with an infinite number of explanatory variables from a finite amount of data, we can estimate the parameters of the autoregressive form of the model given in equation (2). However, this apparent simplification is purchased only at a cost, for consistent estimation of relation (2) requires that we face several estimation problems associated with equations in which lagged dependent variables appear as explanatory variables. While ordinary least squares estimates of the parameters of (2) are consistent provided that the disturbances ut' are serially independent and follow a distribution which satisfies the assumptions of the central limit theorem, even in this case a small sample bias exists. If the disturbances are serially dependent, an asymptotic bias exists.' Moreover, the transformation from (1) to (2) has changed both the variance and the serial correlations of the disturbances. Hence, if the disturbances in (1) are serially independent, those in (2) are necessarily autocorrelated, which means that applying least squares to equation (2) yields inconsistent estimates of the parameters. In addition to ordinary least squares (OLS), several techniques for estimating such distributed lag relations are available. Generally these techniques have been recommended on the basis of their desirable asymptotic properties. However, for economists, who are forced to work in a world where data are scarce, asymptotic properties are frequently of little relevance. What is more often required is knowledge of the properties of the estimators in small samples. Unfortunately, it has proved difficult to investigate these properties analytically. In the absence of such results, sampling or Monte Carlo experiments provide an alternative, if less elegant, source of information. Accordingly, this paper presents the results of a Monte Carlo study of several lag estimators under conditions in which the disturbances ut' of relation (2) are serially correlated. In addition to ordinary least squares, the following five methods were studied. 1) Two Stage Regression (TSLS): This is an application of Leviatan's instrumental variable approach. Leviatan [12] has suggested that xti1 be used as an instrument for yt-i in estimating relation (2). In order to increase the efficiency of the technique, we employed a linear combination of lagged x's as the instrument. The linear combination was determined by first estimating the equation