Navigation path

Page navigation

Research areas

EEMC Software

EEMC develops and contributes to the improvement of several software packages dedicated to econometric analysis:

  • PROGRAM GAP implements a bivariate model as proposed by K.Kuttner (1994)
  • QUEST III: the DYNARE implementation of peer-reviewed versions of the European Commission's QUEST III model.
  • DMM: Fortran program for Bayesian analysis of dynamic mixture models.
  • DYNARE: sensitivity analysis package for DYNARE.
  • BUSY, a software for business cycle analysis.
  • FLASH: rapid estimates of Quarterly National Account main aggregates.
  • GLUEWIN: it implements the Generalized Likelihood Uncertainty Estimation (GLUE) methodology (Beven and Binley, 1992).
  • SS-ANOVA-R: software to estimates of multivariate Smoothing Spline ANOVA models with Recursive algorithms.


Output gap is the key variable of the cyclical adjustment of EU Member States budget balances. Following an ECOFIN 2002 decision, ECFIN measures output gap through a Cobb-Douglas production function that relates the gap to the cyclical components of unemployment and of total factor productivity (TFP).

PROGRAM GAP is built for the trend-cycle decomposition of unemployment and TFP. The model implemented is similar to the one proposed by K.Kuttner (1994) in "Estimating Output as a Latent Variable", Journal of Business Economic and Statistics, 12, 3, 361-368, with several extensions. Version 4.2 includes a Bayesian module.

Download the GAP Manual.

To download the GAP software please send an email request to:
alessandro.rossi [AT]
christophe.planas [AT]

Quest III

We provide here the DYNARE implementation of the versions of the European Commission's QUEST III model described in the papers:

1) "QUEST III: An estimated open-economy DSGE model of the Euro area with fiscal and monetary policy", Economic Modelling, 26 (2009), 222-233.

QUEST III is a DSGE model for an open economy estimated on Euro area data using Bayesian estimation techniques. The model features nominal and real frictions, as well as financial frictions in the form of liquidity-constrained households. The model incorporates active monetary and fiscal policy rules (for government consumption, investment, transfers and wage taxes) and can be used to analyze the effectiveness of stabilization policies. To capture the unit root character of macroeconomic time series we allow for a stochastic trend in TFP, but instead of filtering data prior to estimation, the model is estimated in growth rates and stationary nominal ratios.

2) "The recent boom–bust cycle: The relative contribution of capital flows, credit supply and asset bubbles", European Economic Review, 55 (2011), 86-406. Download

We use an estimated open economy DSGE model with financial frictions for the US and the rest of the world to evaluate various competing explanations about the recent boom–bust cycle. We find that the savings glut hypothesis is insufficient for explaining all aspects of the boom in the US. Relatively strong TFP growth and expansionary monetary policy are also not able to explain fully the volatility of corporate and in particular residential investment. We identify bubbles in the stock and housing market as crucial. Concerning the downturn in 2008/2009, the fall in house prices and residential investment only plays a minor role. Mortgage defaults have more explanatory power, especially in a specification of the model with a segregated equity market. Finally, the bursting of the stock-market bubble was at least as important in this recession as in 2001. Because of various negative shocks hitting the economy at the same time in 2008/2009 and continued positive technology growth, not only the real interest rate declined but inflation fell rapidly and left insufficient room for monetary policy to play a similar stabilizing role as in previous recessions.

New developments May 2009

To download the QUEST III software please send an email request to:
marco.ratto [AT]


DMM is a Fortran program for Bayesian analysis of dynamic mixture models which delivers posterior samples of the unobserved state vector, of the discrete latent variables, and of model parameters using Markov Chain Monte Carlo (MCMC) simulations. Besides computational efficiency, DMM has several attractive features: the endogenous series can be univariate or multivariate, stationary or non-stationary, with some possibly missing observations, and they may be linked to some exogenous variables. The manual describes the methodology implemented and illustrates by examples how to use the program.

Download the DMM Manual.

To download the DMM software please send an email request to:
alessandro.rossi [AT]
christophe.planas [AT]



EEMC staff are members of the DYNARE Team and contributes to the development of DYNARE.

A Global Sensitivity Analysis Toolbox for DSGE models is freely available (Documentation). The Toolbox runs under the DYNARE environment, and is available as a separate package for DYNARE versions v4.0.3, v4.1 and v4.2.1.

NOTE: as of DYNARE v4.3 the Sensitivity Toolbox is included in the official DYNARE distribution!

Access to official Dynare Site

To download the Dynare software please send an email request to:
marco.ratto [AT]



BUSY is a software for business cycle analysis, developed and partially funded within the IST line of the 5th Framework Programme. Two main appraoches are implemented: the traditional NBER approach and the more recent dynamic factor models. See BUSY Manual for a detailed description.

Download BUSY Manual, Version 4.1, release June 2003.

To download the BUSY software please send an email request to:
alessandro.rossi [AT]
christophe.planas [AT]


A shared-cost action within the information society programme CPA8-research and technological development.

The objective of FLASH is to produce early estimates of main Quarterly National Accounts aggregates of the European Union. The target delay is 40-45 days after the end of the reference period. The project supplies a coherent system able to help in conducting short-term economic analysis and in tacking monetary policy decisions, while the shortcomings of the delay of availability of the official quarterly figures is avoided.

To download the FLASH software please send an email request to:
alessandro.rossi [AT]
christophe.planas [AT]



GLUEWIN is a code designed for analysing the output of Monte Carlo runs when empirical observations of the model output are available and implements the combination of GSA and GLUE (Beven and Binley, 1992) methodologies.

In the last decade, a method based on the concept of Bayesian Inference for uncertainty estimation, has been used in hydrology as the Generalised Likelihood Uncertainty Estimation Technique (GLUE) (Beven and Binley, 1992). The GLUE technique is as an extension of the Generalised Sensitivity Analysis methodology, which has now come to be called Regional Sensitivity Analysis (RSA), by R.C Spear and G.M. Hornberger. GLUE has been developed from an acceptance of the possible equifinality of models, i.e. different sets of model factors/structures, lumped under the term 'input factors' in this work, may be equally likely as simulators of the real system. It works with multiple sets of factors, typically via Monte Carlo sampling, and applies likelihood measures to estimate the predictive uncertainty of the model. Model realisations are weighted and ranked on a likelihood scale via conditioning on observations and the weights are used to formulate a cumulative distribution of predictions. Applying the RSA terminology, model structures/factor sets with almost zero likelihood can be classified as non-behavioural and rejected.
The GLUE procedure is based upon making a large number of runs of a given model with different sets of factor values, chosen randomly form specified distributions. Different sets of initial, boundary conditions or model structures can also be considered. On a basis of comparing predicted and observed responses, each set of factor values is assigned a likelihood of being a simulator of the system.
Rescaling of the likelihood measures such that the sum of all the likelihood values equals 1 yields a distribution function for the factor sets. From this, the uncertainty estimation can be performed, by computing the model output cumulative distribution, together with prediction quantiles.

Beven K.J., Binley A., The Future of Distributed Models: Model Calibration and Uncertainty Prediction, Hydrological Processes, 6, 279-298, 1992.
Spear R.C. and Hornberger G.M., Eutrophication in Peel Inlet, II. Ideintification of Critical Uncertainties via Generalised Sensitivity Analysis, Water Research, 14, 43-49, 1980.

To download the GLUEWIN software please send an email request to:
marco.ratto [AT]



SS-ANOVA-R: MATLAB Toolbox for the estimation of Smoothing Spline ANOVA models with Recursive algorithms.

The internal recursive SDR (State-Dependent Regression) routines used in this toolbox have been developed by Lancaster University in conjunction with the JRC. The SDR routine is a special purpose version of the full SDP algorithm (State-Dependent Parameter), developed by Prof. Peter Young and co-workers at Lancaster University (CAPTAIN Toolbox).

To download the SS-ANOVA-R software please send an email request to:
marco.ratto [AT]
andrea.pagano [AT]