چارچوب اقتصاد سنجی کلان کینزی برای تجزیه و تحلیل قواعد سیاست پولی

In the framework of a Keynesian based monetary macro model, we study the implications of targeting monetary aggregates or targeting the interest rate as two alternative monetary policy rules. Whereas the former targets the inflation rate indirectly, through the control of the money supply, the latter, also called the Taylor rule, implies direct inflation targeting. Our monetary macro model exhibits: asset market clearing, disequilibrium in the product and labor markets, sluggish price and quantity adjustments, two Phillips curves for the wage and price dynamics and expectations formation which represents a combination of adaptive and forward looking behavior. The parameters of different model variants are estimated partly through single equation and partly through subsystem estimations for US quarterly time series data 1960.1–1995.1. With the estimated parameters system simulations for the two monetary policy rules are performed. Although the two rules have slightly different stability properties, we show that discretionary monetary policy, i.e. policy that responds to the state of macro variables, has stabilizing effects. We also show that our model with either of the two rules generate in terms of impulse response functions roughly the same responses to shocks as one obtains from standard VAR studies.

Recently, in macroeconomics the quantitative study of monetary policy rules has been undertaken in a variety of frameworks. Such frameworks are, for example, the large-scale macroeconometric models (Fair, 1984 and the contributions collected in Taylor, 1999), the VAR (Bernanke and Blinder, 1992 and Sims, 2000) and the optimization based approach (Rotemberg and Woodford, 1999 and Christiano and Gust, 1999). Usually two alternative monetary policy rules have been considered, namely the monetary authority (1) targeting monetary aggregates or (2) targeting the interest rate. The former implies an indirect and the latter a direct inflation targeting. The latter rule originates in Taylor (1993) and has also been called the Taylor rule.1 As has been shown historically, most central banks of OECD countries switched during the 1980s from the policy of controlling monetary aggregates to targeting inflation rates through controlling short-term interest rates.2 The second type of monetary policy rule, the Taylor rule, has recently been given much attention and has extensively been evaluated in the context of macroeconometric frameworks, see Taylor (1999). This paper employs a small scale Keynesian integrated macromodel to evaluate the above monetary rules of central banks. Our approach is novel in the sense that we employ a consistently formulated and complete Keynesian macroeconometric framework to study monetary policy issues. The Keynesian model presented and estimated here exhibits along the lines of Flaschel et al. (1997) asset market clearing, disequilibrium in product and labor market, sluggish price and quantity adjustments, two Phillips curves for the wage and price dynamics and expectations formulation which represents a combination of adaptive and forward looking behavior. Moreover, as in Chiarella and Flaschel (2000), the current paper also includes real growth, inflationary dynamics and inventory adjustment. As to the historical tradition, on the demand side it is Keynesian, it makes use of Kaldor’s distribution theory, uses the asset market structure as in Sargent’s (1987) Keynesian model, employs Malinvaud’s (1980) investment theory, and a Metzler type inventory adjustment process and uses an expectations mechanism which is forward and backward looking.3 The model’s dynamic features for the two policy regimes are explored for certain parameter constellations. The general dynamic behavior of our system can be analytically studied locally but the global behavior has to be inferred from numerical simulations. For the model with money supply rule it is indicated that for a certain range of parameter constellations interesting dynamics, for example, persistent cycles, may arise. On the other hand, the Taylor rule appears to add further stabilizing forces to this type of model, since it counteracts the destabilizing Mundell effect of inflationary expectations and thus brings more stability into the macro model. In order to match the model with the US macroeconomic time series data, we estimate key parameters through single equation or subsystem estimations using US quarterly data from 1960.1 to 1995.1. In the estimation of the parameters for the wage–price dynamics and for the inventory dynamics as well as investment and consumption functions expectations variables appear which are not observables. We can, however, transform the equations to be estimated and estimate the adjustment speeds involved in the expectations dynamics. Those estimations are undertaken with two stage least square (2SLS). We want to remark that this kind of estimation strategy can also be found in recent literature on macro estimations for large systems with many parameters. Note that, since we are interested here in developing a model that replicates the empirical effects of policy actions, we explore less to what extent our model improves the forecast of particular time series data but rather whether our model can match some times series properties of the data. Our econometric method resembles the method that has been used in the calibration literature, see, for example, the work by Rotemberg and Woodford (1999). In the last step then, for our parameter estimates, we explore the stability properties of our two policy rules and study the question whether the impulse–response functions of our model variants match those of the data. Since both policy rules are defined here as feedback rules, we find that they generate less instability than compared with studies that employ only exogenous policy shocks, for example, auto regressive processes for the monetary policy.4 This means that discretionary monetary policy that is following some feedback rule will be stabilizing. This is a property that many Keynesian models have predicted. Moreover, our model is able to replicate well-known stylized facts obtained, for example, from VAR studies of macroeconomic variables. The remainder of the paper is organized as follows. Section 2 gives a broad overview on the various feedback structures of the model. Section 3 introduces the small scale integrated monetary macromodel. Section 4 studies the steady state and the dynamics of the model, in intensive form. In Section 5, we describe our econometric estimation strategy and report results from our estimations. Section 6 evaluates our results and Section 7 concludes the paper. The appendices provide the notations used.

In the paper, we have chosen a Keynesian based macroeconometric framework for studying macrodynamics and monetary policy. In our framework disequilibrium is allowed in the product and labor markets whereas the financial markets are always cleared. There are sluggish price and quantity adjustments and expectations formation represents a combination of adaptive and forward looking behavior. We consider two monetary policy rules. These policy rules are (1) the money supply rule and (2) the interest rate targeting by the monetary authority. We demonstrate the implication of those policy rules for a monetary macromodel of Keynesian type, and study how the private sector behaves under those alternative policy rules. We estimate the parameters of the model employing US macroeconomic time series data from 1960.1 to 1995.1. Based on the estimation of the parameters, obtained partly from subsystems partly from single equations, we study, using simulations, the dynamic properties for economies which employ either the money supply or the Taylor rule. As we could show with respect to volatility of the macroeconomic variables the model with the Taylor rule seems to perform better in the sense that it gives rise to a faster convergence of macroeconomic variables. We also show that discretionary monetary policy that responds to the state of macroeconomic variables appear to be stabilizing — at least if the policy is pursued with sufficient strength. This is contrary to what one obtains from purely exogenous policy shocks. Moreover, the impulse response functions for our two model variants show roughly the same features as empirical impulse response functions based on VAR studies show. Our results thus show that our disequilibrium model can compete with currently widely used equilibrium macro models. Of course, more empirical work needs to be done in order to confirm or evaluate the findings of this paper. Yet, AS–AD disequilibrium models that include a treatment of income distribution, the role of aggregate demand and economic growth, have not yet been discussed in the theoretical and applied literature to a sufficient degree and thus deserve more attention than they have received so far.