Top

Empirical Economics

Published in:

23-08-2016

Macroeconomic and credit forecasts during the Greek crisis using Bayesian VARs

Author: Dimitrios P. Louzis

Published in: Empirical Economics | Issue 2/2017

Activate our intelligent search to find suitable subject content or patents.

search-config

AI-assisted search

Off

Abstract

We investigate the ability of small- and medium-scale Bayesian VARs (BVARs) to produce accurate macroeconomic (output and inflation) and credit (loans and lending rate) out-of-sample forecasts during the latest Greek crisis. We implement recently proposed Bayesian shrinkage techniques based on Bayesian hierarchical modeling, and we evaluate the information content of forty-two (42) monthly macroeconomic and financial variables in terms of point and density forecasting. Alternative competing models employed in the study include Bayesian autoregressions (BARs) and time-varying parameter VARs with stochastic volatility, among others. The empirical results reveal that, overall, medium-scale BVARs enriched with economy-wide variables can considerably and consistently improve short-term inflation forecasts. The information content of financial variables, on the other hand, proves to be beneficial for the lending rate density forecasts across forecasting horizons. Both of the above-mentioned results are robust to alternative specification choices, while for the rest of the variables smaller-scale BVARs, or even univariate BARs, produce superior forecasts. Finally, we find that the popular, data-driven, shrinkage methods produce, on average, inferior forecasts compared to the theoretically grounded method considered here.

previous article Sudden stops and output: an empirical Markov switching analysis

next article The housing market and excess monetary liquidity in China

Dont have a licence yet? Then find out more about our products and how to get one now:

Springer Professional "Wirtschaft"

Online-Abonnement

Mit Springer Professional "Wirtschaft" erhalten Sie Zugriff auf:

über 67.000 Bücher
über 340 Zeitschriften

aus folgenden Fachgebieten:

Bauwesen + Immobilien
Business IT + Informatik
Finance + Banking
Management + Führung
Marketing + Vertrieb
Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

inform now

Springer Professional "Wirtschaft+Technik"

Online-Abonnement

Mit Springer Professional "Wirtschaft+Technik" erhalten Sie Zugriff auf:

über 102.000 Bücher
über 537 Zeitschriften

aus folgenden Fachgebieten:

Automobil + Motoren
Bauwesen + Immobilien
Business IT + Informatik
Elektrotechnik + Elektronik
Energie + Nachhaltigkeit
Finance + Banking
Management + Führung
Marketing + Vertrieb
Maschinenbau + Werkstoffe
Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

inform now

Available only for authorised users

Alternatively, factor methods have also been used in studies that involve a large number of variables (e.g., see Stock and Watson, 2002, 2006 among others).

The authors also show that models with more than 20 variables can hardly improve forecasting performance.

Hyperparameter is a coefficient that is used to parameterize a prior distribution.

See Provopoulos (2014) for the origins of the Greek crisis and Gibson et al. (2014a) for an overview of the crisis in the Euro area.

Successive recession may be attributed to reduced internal demand (due to increased unemployment and reduced salaries) and weak dynamics in exports (due to firms’ reduced ability to access credit channels) (Kiriakidis and Kargas 2013).

The FSSI is a composite systemic stress index that applies the insights from the standard portfolio theory to summarize stress measures of different market segments into an aggregate index. The key feature of the FSSI is the time-varying cross-correlations among the stress measures, which form the mechanism that captures the systemic nature of stress. For an alternative version of the FSSI, see also Louzis and Vouldis (2012).

Macroeconomic variables include the industrial production index and the consumers price index, while credit variables include the outstanding amount of total loans granted to the private sector and the corresponding lending rate (see Sect. 3 and Table 7 for details).

See also the discussion in Kazanas and Tzavalis (2014).

For a recent example of the marginal approach, see Caraiani (2014).

The term conjugate prior is used to define the prior that comes from the same family of distributions along with the posterior.

An alternative approach to impose the N-IW prior is to use a fictitious prior data set as in Banbura et al. (2010).

Since the prior mean of \(\varSigma \) is defined as \(\left( {d-n-1} \right) ^{-1}\varSigma \), \(d=n+2\) is the minimum value that guarantees its existence.

The Minnesota prior was originally developed by Litterman (1979, (1980).

Hierarchical modeling refers to models that place priors on hyperparameters (i.e., hyperpriors), and thus, one more step is added to the prior structure in a hierarchical manner.

See Giannone et al. (2015) and references therein for a list of studies using these methods.

We use the industrial production index as a proxy for the aggregate output as in Caraiani (2014), since it is the only index of economic activity for Greece available at a monthly frequency. The industrial production index is also one of the main determinants of the Greek GDP (Kiriakidis and Kargas 2013).

We use the Euro OverNight Index Average (EONIA) as a proxy for the monetary policy stance in the Euro zone. The EONIA rate is defined as the average of overnight rates for unsecured interbank lending in Euro area. The Governing council of the European Central Bank (ECB) determines the range of fluctuation of the EONIA rate, which is given by the following range: [deposit facility rate, marginal lending facility rate]. The EONIA rate is considered an efficient proxy of the monetary policy stance in the Euro area, as compared to other money market instruments, such as the Euribor, the overnight interest swap (OIS) rate on EONIA rate or the repo rates (see the discussion in Ciccarelli et al. 2010; Gerlach and Lewis 2014).

As already mentioned there are also data availability issues, which pose limitations to this kind of analysis.

Estimation of the various constant parameter BVAR models is based on the MATLAB code kindly provided by Domenico Giannone in his homepage: http://homepages.ulb.ac.be/~dgiannon/. For the estimation of the TVP-VAR we use the MATLAB code kindly provided by Haroon Mumtaz in his homepage: https://sites.google.com/site/hmumtaz77/.

Recall that the baseline model uses the variables of Table 7.

In practice we set \(\lambda =10^{-4}\) and \(\lambda =10^{4}\) for the dogmatic and flat priors, respectively, as in Berg and Henzel (2015).

This is analogous to the use of principal components implemented by Banbura et al. (2010, p. 80).

The only exception is the TVP-VAR which is estimated recursively.

The description of the evaluation metrics as well as the presentation of the results follows closely the contributions of Carriero et al. (2015a, (2015b).

This is equivalent to testing whether relative RMSFE \(< 1\) and \(\Delta \) SCORE \(> 0\), respectively.

I would like to thank an anonymous reviewer for proposing to incorporate the AR model into the analysis.

Particularly, the BVAR + FSSI outranks the benchmark in 33 % of the cases, while the \(\hbox {Medium}^{\mathrm{Fin}}\)-BVAR in 17 % of the cases across variables and forecasting horizons.

The average \(\lambda \) estimates for \(\hbox {Medium}^{\mathrm{Ec}}\)-BVAR, \(\hbox {Medium}^{\mathrm{Fin}}\)-BVAR are 0.21 and 0.20 for the CKM method and 0.27 and 0.26 for the GLP method.

Other studies that implement TVP-VARs in forecasting exercises use long data sets with quarterly frequencies that cover time periods that range between 20 and 60 years (e.g., see D’Agostino et al. 2013; Barnet et al. 2014; Carriero et al. 2015a).

We use the first seventy (70) observations of our sample as an in-sample period.

The method implies rolling or recursive estimation. For \(\tau =1\) (i.e., the first point in the out-of-sample period) for which \(\lambda _{m}^{*} (\tau +1)\) is not available we set \(\lambda _{m}^{*} (1)=0.2\), which is a benchmark value used by many researchers (see Carriero et al. 2015b and references therein).

Adolfson M, Linde J, Villani M (2007) Forecasting performance of an open economy DSGE model. Econom Rev 26:289–328CrossRef

Amisano G, Giacomini R (2007) Comparing density forecasts via weighted likelihood ratio tests. J Bus Econ Stat 25:177–190CrossRef

Banbura M, Giannone D, Reichlin L (2010) Large Bayesian vector auto regressions. J Appl Econom 25:71–92CrossRef

Bank of Greece, 2014 (2013) Stress test of the Greek banking sector. Bank of Greece March 2014

Barnet A, Mumtaz H, Theodoridis K (2014) Forecasting UK GDP growth and inflation under structural change. A comparison of models with time-varying parameters. Int J Forecast 30:129–143CrossRef

Berg TO, Henzel SR (2015) Point and density forecasts for the euro area using Bayesian VARs. Int J Forecast 31:1067–1095CrossRef

Blanchard O, Leigh D (2013) Growth forecast errors and fiscal multipliers. IMF working paper WP/13/1

Caraiani P (2014) Do money and financial variables help forecasting output in emerging European economies? Empir Econ 46:743–763CrossRef

Carriero A, Kapetanios G, Marcellino M (2009) Forecasting exchange rates with large Bayesian VAR. Int J Forecast 25:400–417CrossRef

Carriero A, Kapetanios G, Marcellino M (2012) Forecasting government bond yields with large Bayesian vector autoregressions. J Bank Finance 36:2026–2047CrossRef

Carriero A, Mumtaz H, Theophilopoulou A (2015a) Macroeconomic information, structural change, and the prediction of fiscal aggregates. Int J Forecast 31:325–348CrossRef

Carriero A, Clarck TE, Marcellino M (2015b) Bayesian VARs: specification choices and forecast accuracy. J Appl Econom 30:46–73CrossRef

Ciccarelli M, Maddaloni A, Peydro J-L (2010) Trusting the bankers: a new look at the credit channel of monetary policy. ECB working papers series no 1228, July 2010

Clark T, McCracken M (2013) Advances in forecast evaluation. In: Elliot G, Timmermann A (eds) Handbook of economic forecasting, vol 2(B). Elsevier, Amsterdam, pp 1107–1201

D’Agostino A, Gambetti L, Giannone D (2013) Macroeconomic forecasting and structural change. J Appl Econom 28:82–101CrossRef

De Mol C, Giannone D, Reichlin L (2008) Forecasting using a large number of predictors: is Bayesian regression a valid alternative to principal components? J Econom 146:318–328CrossRef

Diebold FX, Mariano RS (1995) Comparing predictive accuracy. J Bus Econ Stat 13:253–263

Doan T, Litterman R, Sims C (1984) Forecasting and conditional projection using realistic prior distributions. Econom Rev 3:1–144CrossRef

Faust J, Wright JH (2009) Comparing Greenbook and reduced form forecasts using a large realtime dataset. J Bus Econ Stat 27:468–479CrossRef

Geweke J, Amisano G (2010) Comparing and evaluating Bayesian predictive distributions of asset returns. Int J Forecast 26:216–230CrossRef

Giannone D, Lenza M, Momferatou D (2014) Short-term inflation projections: a Bayesian vector autoregressive approach. Int J Forecast 30:635–644CrossRef

Giannone D, Lenza M, Primiceri G (2015) Prior selection for vector autoregressions. Rev Econ Stat 97:436–451CrossRef

Gibson HD, Hall SG, Tavlas GS (2012) The Greek financial crisis: growing imbalances and sovereign spreads. J Int Money Finance 31(3):498–516CrossRef

Gibson HD, Palivos T, Tavlas GS (2014a) The crisis in the Euro area: an analytic overview. J Macroecon 39:233–239CrossRef

Gibson HD, Hall SG, Tavlas GS (2014b) Fundamentally wrong: market pricing of sovereigns and the Greek financial crisis. J Macroecon 39:405–419CrossRef

Gerlach S, Lewis J (2014) Zero lower bound, ECB interest rate policy and the financial crisis. Empir Econ 46(3):865–886CrossRef

Gupta R, Kabundi A (2010) Forecasting macroeconomic variables in a small open economy: a comparison between small- and large-scale models. J Forecast 29:168–185CrossRef

Harvey D, Leybourne S, Newbold P (1997) Testing the equality of prediction mean squared errors. Int J Forecast 13:281–291CrossRef

Hristov N, Hulsewig O, Wollmershauser T (2012) Loan supply shocks during the financial crisis: evidence for the Euro area. J Int Money Finance 31:569–592CrossRef

Kadiyala K, Karlsson S (1997) Numerical methods for estimation and inference in Bayesian VAR models. J Appl Econom 12:99–132CrossRef

Karlsson S (2013) Forecasting with Bayesian vector autoregression. Handb Econ Forecast 2:791–897CrossRef

Kazanas T, Tzavalis E (2014) Comment on “Fundamentally wrong: market mispricing of sovereigns and the Greek financial crisis”. J Macroecon 39:420–423CrossRef

Kiriakidis M, Kargas A (2013) Greek GDP forecast estimates. Appl Econ Lett 20:767–772CrossRef

Koop G, Korobilis D (2010) Bayesian multivariate time series methods for empirical macroeconomics. Foun Trends Econom 3:267–358CrossRef

Koop G (2013) Forecasting with medium and large Bayesian VARs. J Appl Econom 28:177–203CrossRef

Litterman RB (1979) Techniques of forecasting using vector autoregressions. Federal Reserve of Minneapolis working paper 115

Litterman RB (1980) A Bayesian procedure for forecasting with vector autoregressions. Working paper, Massachusetts Institute of Technology, Department of Economics

Litterman RB (1986) Forecasting with Bayesian vector autoregressions: five years of experience. J Bus Econ Stat 4:25–38

Louzis DP, Vouldis AT (2012) A methodology for constructing a financial systemic stress index: an application to Greece. Econ Model 29:1228–1241CrossRef

Louzis DP, Vouldis AT (2013) A financial systemic stress index for Greece. ECB working paper series no 1563, July 2013

Louzis DP (2015) Measuring spillover effects in Euro area financial markets: a disaggregate approach. Empir Econ 49:1367–1400CrossRef

Marcellino M (2008) A linear benchmark for forecasting GDP growth and inflation? J Forecast 27:305–340CrossRef

Provopoulos G (2014) The Greek economy and banking system: recent developments and the way forward. J Macroecon 39:240–249CrossRef

Sims C (1980) Macroeconomics and reality. Econometrica 48:1–48CrossRef

Sims C (1993) A nine-variable probabilistic macroeconomic forecasting model. In: Business cycles, indicators and forecasting, nber chapters. National Bureau of Economic Research, Inc. pp 179–212

Title: Macroeconomic and credit forecasts during the Greek crisis using Bayesian VARs
Author: Dimitrios P. Louzis
Publication date: 23-08-2016
Publisher: Springer Berlin Heidelberg
Published in: Empirical Economics / Issue 2/2017
Print ISSN: 0377-7332
Electronic ISSN: 1435-8921
DOI: https://doi.org/10.1007/s00181-016-1128-y

Springer Professional

Abstract

Please log in to get access to your license.

Dont have a licence yet? Then find out more about our products and how to get one now:

Springer Professional "Wirtschaft"

Springer Professional "Wirtschaft+Technik"

Other articles of this Issue 2/2017

Are state–local government expenditures converging? New evidence based on sequential unit root tests

The impact of targeting policy on spouses’ demand for public goods, labor supplies and sharing rule

Hurdle models of repayment behaviour in personal loan contracts

Exchange rate volatility and bilateral exports of Malaysia to Singapore, China, Japan, the USA and Korea

Investigation of institutional changes in the UK housing market using structural break tests and time-varying parameter models

Markups and bargaining power in tradable and non-tradable sectors