مدل های تعادل عمومی برآورد شده برای ارزیابی سیاست های پولی در آمریکا و اروپا

In this paper we derive a general equilibrium model based on optimising behaviour, which also implies a data consistent framework for monetary policy analysis. Specifically, our model accounts for nominal inertia in both price and wage setting as well for habits in consumption. Using US and European data from 1970 to 1998 our parameter estimates reveal that (i) price contracts last for 8 months and 13 months in the US and Euro-area, respectively; (ii) wage contracts have a length of 7 months and 1.75 years in the US and Europe, respectively; (iii) the extent of backward-looking behaviour in price setting is statistically significant in both economies with 41% of price contracts in the US and 28% in the Euro-area set according to a simple rule-of-thumb; (iv) backward-looking wage setting is only present in Europe with 17% of contracts set in a backward-looking manner; and (v) similar habits effects are present in both European and US consumption. Finally, we simulate the effects of monetary policy by considering the impact of a 1 point increase in nominal interest rates for one quarter. Our parameter estimates imply that there is a relatively muted inflationary response to interest rate increases in Europe (price inflation falls by -0.08%-0.08% in Europe and 0.11% in the US) and there is a correspondingly large output response (-0.2%(-0.2% in the US and -0.6%-0.6% in Europe).

Despite the reputation of macroeconomics as a subject plagued by numerous ideological disputes, some authors are now arguing that the field is developing a new consensus (see, for example, Woodford, 2003, p. 6; Goodfriend and King, 1997). The latter authors have gone as far as dubbing this emerging view, ‘the New Neo-Classical Synthesis’ (NNCS). Essentially, the NNCS extends the optimising behaviour underlying the real-business cycle (RBC) literature to include the frictions considered by New Keynesian economists in the 1980s (see Mankiw and Romer (1991) for a collection of influential papers). As a result the NNCS can use the insights of RBC theory to explain equilibrium output, while, at the same time, explaining deviations of actual output from equilibrium as a result of stickiness in the adjustment of prices and wages. The NNCS paradigm has been employed in numerous academic studies of monetary policy (see, for example, Rotemberg and Woodford, 1997; Clarida et al., 1999, Taylor, 1999 and Erceg et al., 2000). Moreover, the NNCS has come to dominate policy evaluation in central banks throughout the world (see Taylor, 1999) and has even been suggested as the most relevant macroeconomic framework to teach to undergraduate students (see Taylor, 2000; Romer, 2000). However, despite the emergence of a dominant theoretical framework for the analysis of monetary policy, it is still crucial in designing an optimal monetary policy to identify the quantitative importance of the economic mechanisms underpinning this benchmark framework. The optimal response to shocks will be dependent upon the implicit trade-off between output and inflation variability as well as the responses of economic agents to interest rate changes. In VAR-based studies of the monetary policy transmission mechanism it is difficult to identify the economic mechanisms underpinning any observed responses to monetary policy shocks. Indeed McCallum (2001) argues that in the VAR literature there has been an overemphasis of the unsystematic component of interest rate movements and that structural models which are invariant to policy changes are the only means of analysing the quantitatively more important systematic element of policy. While in cross country comparisons of the transmission mechanism based on the VAR approach, it is also difficult to assess the extent to which any differences in the macroeconomic responses to policy shocks are statistically significant and which structural differences account for any observed differences. Additionally, while larger scale macroeconometric models do offer the opportunity to examine systematic policy and structural differences between economies, any observed differences in the monetary policy transmission mechanism may be due to different modelling assumptions as much as structural differences between economies (for a discussion of this point, in the context of such an exercise, see McAdam and Morgan (2001)). In this paper we attempt we build a structural model in the spirit of the NNCS models used to analyse monetary policy, but with sufficient frictions and inertia to enable it to fit the data. The benchmark NNCS model typically consists of a demand side based on an Euler equation describing the intertemporal consumption decisions of a representative consumers. Monetary policy is given a role through the adoption of nominal inertia in the form of Calvo (1983) contracts giving rise to a New Keynesian Phillips Curve (NKPC) representation of the supply-side. The model is then closed through some specification of monetary policy often in the form of an interest rate rule. The benchmark model has often been criticised on empirical grounds (see, for example, Mankiw and Reis, 2002), largely for failing to capture the true nature of the output-inflation trade-off by adopting a NKPC and by being too forward-looking to reflect the inertia often thought to be part of the transmission mechanism. In order to address these concerns we undertake several extensions to this benchmark. Firstly, since there is clear evidence that the output gap is a poor proxy for marginal costs in the NKPC (see, for example, Galí and Gertler (1999), Galí et al. (2001) for a re-evaluation of the NKPC which accounts for this), we allow for nominal wage inertia and habits effects in consumption/labour supply to break the proportionality between the output gap and marginal costs typically found in NNCS models. Additionally, allowing for habits effects in consumption and backward-looking price- and wage-setting behaviour means that the estimation can capture any potential inertia in the monetary policy transmission mechanism. The contribution of this paper is then to jointly estimate this model within a systems-framework using time-series data for both the US and Europe to obtain estimates of the structural parameters of the model. While it is interesting to examine each of these economies individually,1 our joint estimation will allow us to assess the size and source of any differences in the monetary transmission mechanisms across the two economies. Our econometric work further allows us to obtain estimates of the degree of nominal inertia in both wages and prices in the two economies, along with the extent to which economic agents using backward-looking rules-of-thumb, as opposed to intertemporal optimisation, in setting wages and prices. We also attempt to quantify the responsiveness of output to interest rates, by estimating the intertemporal elasticity of substitution. Finally, our econometric work provides an estimate of our consumers’ rate of time preference, and the extent to which their past levels of consumption affect the utility they derive from current consumption and leisure activities. We find that there is more price stickiness in Europe than in the US, with prices taking on average 13 and 8 months, respectively, to adjust. Europe is also significantly more sticky in wage adjustment with wage contracts lasting for 7 months and 1.75 years in the US and Europe, respectively. With the exception of wage setting in the US, there is evidence of a statistically significant level of backward-looking behaviour. US price setters are more likely to use a backward-looking rule-of-thumb than their European counterparts, possibly reflecting the costs of obtaining information necessary to reset prices optimally when that price is unlikely to remain in place for long. In contrast, wage setters in Europe are more backward looking than in the US. This may reflect the existence of wage-indexation mechanisms in Europe. We also found strong real interest rate effects in both the US and Europe, which were strongest in the US, but not statistically significantly so. We had to impose the degree of habit persistence in Europe, but estimated a plausible degree of habits effects in the US. In pairwise comparisons of these parameters across the two economies, all parameters aside from the discount factor and the intertemporal elasticity of substitution were found to be statistically different. These significant differences in estimated parameters also imply differences in the monetary policy transmission mechanism across the two economies. The relatively high degree of wage and price stickiness in Europe implies that the inflation response to a 1 point increase in nominal interest rates for one quarter is relatively muted in Europe with a peak fall in inflation of -0.08%-0.08%, compared with -0.11%-0.11% in the US. As a result the output consequences of the monetary policy shock are higher in the Euro-area with a peak output response to the monetary policy shock of -0.6%-0.6% in contrast to -0.2%-0.2% in the US. The plan of the rest of the paper is as follows. Section 2 derives the general equilibrium model. Section 3 then discusses the estimation strategy, results and simulates the two economies in the face of a monetary shock. Section 4 concludes.

In this paper we derive and estimate general equilibrium models for the evaluation of monetary policy which are similar in structure to the benchmark model of the NNCS (Goodfriend and King, 1997). Our work extends this model in several ways. We first allow for the possibility that it is not only prices, but also wages which are sticky. Second we model the extent to which some economic agents use backward-looking rules of thumb when setting wages and/or prices, while others attempt to maximise lifetime utility or discounted profits. Finally, we introduce habits into consumption to allow for inertia in output. Our estimates of these models produce a number of interesting results. For example, we find that there is more price stickiness in Europe than in the US, with prices taking on average 13 and 8 months, respectively, to adjust. Europe is also significantly more sticky in wage adjustment with wage contracts lasting for 7 months and 1.75 years in the US and Europe, respectively. With the exception of wage setting in the US, there is evidence of a statistically significant level of backward-looking behaviour. US price setters are more likely to use a rule-of-thumb than their European counterparts. In contrast, wage setters in Europe are more backward looking than in the US. This may reflect the existence of wage-indexation mechanisms in Europe. We also found strong real interest rate effects in both the US and Europe, which were greatest in the US, but not statistically significantly so. In pairwise comparisons of these parameters across the two economies, all parameters aside from the discount factor and the intertemporal elasticity of substitution were found to be statistically different. Finally we considered the response of each economy to a monetary policy shock in the form of a temporary rise in nominal interest rates. Here our results supported our finding that there are significant differences in the structure of the European and US economies which imply that monetary shocks yield quite distinct real and nominal macroeconomic outcomes across countries. Given the finding by Erceg et al. (2000), that the design of optimal monetary policy should minimise the variance in the relatively sticky nominal magnitude, our estimation and simulation results suggest policy makers in Europe would do well to concentrate on offsetting the distortions associated with staggered wages whereas US policy makers should continue to concentrate on distortions due to sticky prices.