An estimated dynamic stochastic general equilibrium model of the euro area
Summary (5 min read)
1. Introduction
- In this paper the authors present and estimate a stochastic dynamic general equilibrium (DSGE) model for the euro area.
- 2 Several results of their analysis are worth highlighting.
- Fourth, regarding the relative contribution of the various shocks to the empirical dynamics of the macro economic time series in the euro area, the authors find that the labour supply and the monetary policy shock are the two most important structural shocks driving variations in euro area output.
- In Section 3, we, first, discuss the estimation methodology, then, present the main results and, finally, compare the empirical performance of the estimated DSGE model with that of various VARs.
2. A DSGE model for the euro area
- In this section the authors derive and present the linearised DSGE model that they will estimate in Section 3.
- Households maximise a utility function with two arguments (goods and leisure (or labour)) over an infinite life horizon.
- This model is a version of the model considered in Kollmann (1997) and features monopolistic competition in both the goods and labour markets.
- In addition, several features of CEE (2001) are introduced.
- The marginal costs depend on wages and the rental rate of capital.
2.1 The household sector
- Households differ in that they supply a differentiated type of labour.
- So, each household has a monopoly power over the supply of its labour.
- Equation (2) above also contains two preference shocks: ε.
- The external habit stock is assumed to be proportional to aggregate past consumption: (3) ChH tt 1−= 5 Habit depends on lagged aggregate consumption which is unaffected by any one agent's decisions.
2.1.2 Labour supply decisions and the wage setting equation
- Households act as price-setters in the labour market.
- Following Kollmann (1997) and Erceg, Henderson and Levin (2000), the authors assume that wages can only be optimally adjusted after some random “wage-change signal” is received.
- The probability that a particular household can change its nominal wage in period t is constant and equal to wξ−1 .
- When 0=wγ , there is no indexation and the wages that can not be re-optimised remain constant.
2.1.3 Investment and capital accumulation
- Finally, households own the capital stock, a homogenous factor of production, which they rent out to the firm-producers of intermediate goods at a given rental rate of ktr .
- An infinite supply elasticity limits the increase in marginal costs and prices following an expansion of output in a model with sticky prices, which helps to generate real persistence of monetary shocks.
- 11 This specification of the costs is preferable above a specification with costs in terms of a higher depreciation rate (see King and Rebelo, 2000; or Greenwood, Hercowitz, and Huffman, 1988; Dejong, Ingram and Whiteman, 2000) because the costs are expressed in terms of consumption goods and not in terms of capital goods.
- 8 NBB WORKING PAPER No.35 - October 2002 Equation (16) describes the optimal dynamic behaviour of investment.
- As the rental rate increases it becomes more profitable to use the capital stock more intensively up to the point were the extra gains match the extra output costs.
2.2 Technologies and firms
- The final good is used for consumption and investment by the households.
- There is monopolistic competition in the markets for intermediate goods: each intermediate good is produced by a single firm.
2.2.1 Final-good sector
- Recently, Woodford (2000) has illustrated how this assumption can be relaxed in a model with sticky prices and adjustment costs in investment.
- The hypothesis has important consequences for the estimation of the degree of price stickiness.
- With capital specific to the firm, firms will be more reluctant to change the price of their good as the resulting demand response will have a much stronger impact on the marginal cost of production.
- The assumption of capital mobility across firms therefore biases the estimated degree of price stickiness upwards.
- NBB WORKING PAPER No. 35 - October 2002 9 where jty denotes the quantity of domestic intermediate good of type j that is used in final goods production, at date t. tp,λ is a stochastic parameter which determines the time-varying mark-up in the goods market.
2.2.2 Intermediate goods producers
- (22) α α− = 1 ~ , , tj k t tjt Kr LW Equation (22) implies that the capital-labour ratio will be identical across intermediate goods producers and equal to the aggregate capital-labour ratio, also known as Cost minimisation implies.
- This implies that the marginal cost, too, is independent of the intermediate good produced.
- As in Calvo (1983), firms are not allowed to change their prices unless they receive a random “price-change signal”.
- With sticky prices the mark-up becomes variable over time when the economy is hit by exogenous shocks.
2.3 Market equilibrium
- The final goods market is in equilibrium if production equals demand by households for consumption and investment and by the government: 15 Erceg, Henderson and Levin (2000) use indexation to the average steady state inflation rate.
- The capital rental market is in equilibrium when the demand for capital by the intermediate goods producers equals the supply by the households.
- The interest rate is determined by a reaction function that describes monetary policy decisions.
2.4 The linearised model
- For the empirical analysis of section 3 the authors linearise the model equations described above around the non-stochastic steady state.
- Note that in this case the interest elasticity of consumption depends not only on the intertemporal elasticity of substitution, but also on the habit persistence parameter.
- Similarly, partial indexation of nominal wages results in the following real wage equation: 16 This is the only shock that is not directly related to the structure of the economy.
- For such alternative interpretations of this equity premium shock and an analysis of optimal monetary policy in the presence of such shocks, see Dupor (2001).
- Finally, the model is closed by adding the following empirical monetary policy reaction function: (36) R t p 1t1t p tty1tt p 1t1tYt1tt1tt )).
3 Estimation results
- First, discuss how the authors estimate the structural parameters and the processes governing the ten structural shocks.
- Next, the authors present the main estimation results.
- Finally, the authors compare the empirical performance of the estimated DSGE model with a number of a-theoretical VARs.
3.1 Estimation methodology
- There are various ways of estimating or calibrating the parameters of a linearised DSGE model.
- In the final step, the parameters are estimated by maximising the likelihood function.
- 16 NBB WORKING PAPER No.35 - October 2002 econometric interpretation of DSGE models is not necessarily less stringent than the strong interpretation: in spite of the focus on a restricted set of moments, the model is assumed to account for all aspects of the observed data series and these aspects are used in calculating the moments of interest.
- The standard errors were set so that the domain covers a reasonable range of parameter values.
- A wide range of calibrations has been used for the inverse elasticity of labour supply.
3.2 Parameter estimates
- In addition to the prior distribution, Table 1 reports two sets of results regarding the parameter estimates.
- The second set reports the 5, 50 and 95 percentile of the posterior distribution of the parameters obtained through the MetropolisHastings sampling algorithm.
- The persistent shocks are estimated to have an autoregressive parameter which lies between 0.82 (for the productivity shock) and 0.95 for the government spending shock.
- Only when they assume decreasing returns to scale and an upward-sloping marginal cost curve, Gali, Gertler and LopezSalido (2000) estimate a more reasonable degree of price stickiness that is comparable with what the authors estimate for wages.
- One needs to be careful when making such comparisons, as their model features external habit formation which turns out to be significant.
3.3.1 Comparing the estimated DSGE model with VARs
- The discussion in the previous section shows that the model is able to deliver reasonable and significant estimates of the model parameters.
- The authors analyse how well their estimated model does compared to a-theoretical VAR models estimated on the same data set.
- Given these samples of the posterior distribution, Geweke (1998) also proposes different methods to calculate the marginal likelihood necessary for model comparison (a method for importance sampling and for MH algorithm, a method for the Gibbs sampler, and the modified harmonic mean that works for all sampling methods).
- This method applies a standard correction to the posterior evaluation at the posterior mode to approximate the marginal likelihood.
- The marginal likelihood of the DSGE model is larger than that of the VAR(2) and VAR(3) model and very close to that of the VAR(1) model.
3.3.2 Comparison of empirical and model-based cross-covariances
- Traditionally DSGE models are validated by comparing the model-based variances and covariances with those in the data.
- Graph 2 summarises the results of this exercise.
- The error bands are quite large, indicating that there is a large amount of uncertainty surrounding the model-based crosscovariances.
- In particular, the crosscorrelations with the interest rate do not seem to be fully satisfactory.
4.1 Impulse response analysis
- Note that these impulse responses are obtained with the estimated monetary policy reaction function.
- The real interest rate rises reflecting the fact that the wage mark-up shock creates a trade-off between inflation and output gap stabilisation.
- Qualitatively similar impulse responses are derived following a temporary negative equity premium shock (Graph 9), but in this case the effects on output, employment and investment are much more short-lived and the resulting effects on real wages, the marginal cost and prices much more limited.
- This leads to a hump-shaped fall in output, consumption and investment.
4.2 Variance decomposition
- The contribution of each of the structural shocks to the forecast error variance of the endogenous variables at various horizons (short run: 1 year; medium run: 2.5 years and long run: 25 years) is reported in Table 3.
- Beyond the very short-term horizon, output variations are driven primarily by the labour supply and the monetary policy shocks.
- The price and wage mark-up shocks do not seem to matter for output variability.
- Empirically inflation is a quite volatile process.
- Somewhat surprisingly, other shocks together typically account for less than 15% of the variance in inflation.
4.3 Historical decomposition
- Graphs 13 and 14 summarise the historical contribution of the various structural shocks to output and inflation developments in the euro area.
- This decomposition is based on their best estimates of the various shocks.
- While obviously such a decomposition must be treated with caution, it helps in understanding how the estimated model interprets specific movements in the observed data and therefore can shed some light on its plausibility.
- Finally, the run-up in inflation in the late 1980s and early 1990s is attributed to the various “supply” and “demand” shocks.
- While loose monetary policy contributed to offsetting the fall in output due to negative supply and demand shocks in the 1970s, it contributed very little to output variations in the 1980s and 1990s.
5 Output and interest rate gaps: an application
- In a simple benchmark New-Keynesian model with only nominal price rigidities and no “mark-up” shocks, Woodford (2002) has pointed out that optimal monetary policy will be able to replicate the flexible price equilibrium, thereby restoring the first best.
- As these shocks give rise to inefficient variations in the flexible-price-and-wage level of output, one can argue that monetary authorities should not accommodate such variations and instead try to keep output at its efficient level.
- A positive labour supply shock has very similar effects on output and the natural real interest rate (Graph 16).
- From 1982 onward potential output has gradually risen to a higher level with a dip in the early 1990s.
- 42 The authors gap differs from the gap that is calculated in Gali, Gertler and Lopez-Salido (2002) and Neiss and Nelson (2001) in the sense that those papers implicitly assume that there are no mark-up shocks.
6 Conclusions
- Recently a new generation of small-scale monetary business cycle models generally referred to as New-Keynesian or New Neoclassical Synthesis models have been developed (Goodfriend and King (1997), Rotemberg and Woodford (1997) and Clarida, Gali and Gertler (1999)).
- Clearly, all “non-monetary” shocks will potentially affect output and the real rate in a flexible price and wage economy.
- Overall, the results presented in this paper show that an estimated version of the DSGE model with sticky prices and wages can be used for monetary policy analysis in an empirically plausible set-up.
- Also a further examination and identification of the various structural shocks would be interesting.
Did you find this useful? Give us your feedback
Citations
3,155 citations
3,115 citations
2,382 citations
1,798 citations
1,542 citations
References
8,580 citations
8,414 citations
5,728 citations
3,990 citations
2,711 citations
Related Papers (5)
Frequently Asked Questions (5)
Q2. What is the effect of a positive preference shock on investment?
Graph 7 shows that a positive preference shock, while increasing consumption and output significantly, has a significant negative crowding-out effect on investment.
Q3. What is the argument that monetary authorities should not accommodate such variations?
As these shocks give rise to inefficient variations in the flexible-price-and-wage level of output, one can argue that monetary authorities should not accommodate such variations and instead try to keep output at its efficient level.
Q4. What are the parameters that can be estimated from the mean of the observable variables?
Most of these parameters can be directly related to the steady-state values of the state variables and could therefore be estimated from the means of the observable variables (or linear combinations of them).
Q5. What is the degree of stickiness of the price and wage curves?
Only when they assume decreasing returns to scale and an upward-sloping marginal cost curve, Gali, Gertler and LopezSalido (2000) estimate a more reasonable degree of price stickiness that is comparable with what the authors estimate for wages.