Forecast Combinations

Aiolfi, Marco; Capistrán, Carlos; Timmermann, Allan

doi:10.2139/ssrn.1609530

Cited by 27 publications

(25 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…First, the optimal combination of alternative forecasts is often taken as a simple and effective way to improve and robustify the forecasting performance of individual models; see, for example, the seminal contribution of Bates and Granger (1969). As a result, forecast combinations are now in widespread use in central banks, among private sector forecasters, and in academic studies; see Aiolfi et al (2010) for a comprehensive survey on forecast and nowcast combination. Second, the analysis of factor models by the method of principal components is widely understood and frequently applied in the social sciences.…”

Section: Introductionmentioning

confidence: 99%

The Information in Systemic Risk Rankings

2015

View full text Add to dashboard Cite

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in AbstractWe propose to pool alternative systemic risk rankings for financial institutions using the method of principal components. The resulting overall ranking is less affected by estimation uncertainty and model risk. We apply our methodology to disentangle the common signal and the idiosyncratic components from a selection of key systemic risk rankings that are recently proposed. We use a sample of 113 listed financial sector firms in the European Union over the period 2002-2013. The implied ranking from the principal components is less volatile than most individual risk rankings and leads to less turnover among the top ranked institutions. We also find that price-based rankings and fundamentals based rankings deviated substantially and for a prolonged time in the period leading up to the financial crisis. We test the adequacy of our newly pooled systemic risk ranking by relating it to credit default swap premia.

show abstract

Section: Introductionmentioning

confidence: 99%

The Information in Systemic Risk Rankings

2015

View full text Add to dashboard Cite

show abstract

“…BVAR Averaging with Equal Weights Recent studies argue that averaging forecasts is a simple though successful method to handle a large dataset and improve out-of-sample forecast accuracy (see, e.g., Clark and McCracken, 2010;Aiolfi, Capistrán, and Timmermann, 2011;Barnett, Mumtaz, and Theodoridis, 2012;Henzel and Mayr, 2013;Wolters, 2014, among others).…”

mentioning

confidence: 99%

Point and density forecasts for the euro area using Bayesian VARs

Berg

Henzel

2015

International Journal of Forecasting

View full text Add to dashboard Cite

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Recent articles suggest that a Bayesian vector autoregression (BVAR) with shrinkage is a good forecast device even when the number of variables is large. In this paper we evaluate different variants of the BVAR with respect to their forecast accuracy for euro area real GDP growth and HICP inflation. We consider BVAR averaging, Bayesian factor augmented VARs (BFAVARs), and large BVARs, which differ in the way information is condensed and shrinkage is implemented. We find that: (a) large BVARs produce accurate point forecasts but show a poor performance when the entire density is considered; (b) BVAR averaging shows the opposite pattern; (c) BFAVARs perform well under both evaluation criteria; (d) choosing the degree of shrinkage optimally does not improve forecast accuracy; (e) all variants except the large BVAR tend to be well calibrated for inflation but poorly calibrated for real GDP growth; (f) these findings are robust to several features of the forecast experiment. Terms of use: Documents inJEL-Code: C110, C520, C530, E370.

show abstract

“…Combining may also improve robustness, as shown by Hendry and Clements () and Aiolfi et al . (). Whether judgmental forecasts by analysts embody valuable additional information can be more formally investigated by running encompassing tests vs. the DFM forecasts.…”

Section: Resultsmentioning

confidence: 97%

“…The evidence in Aiolfi, Capistrán and Timmermann (), Jansen et al . () and Liebermann () suggests that subjective forecasts by private sector analysts often embody valuable information that sophisticated mechanical forecasting procedures fail to pick up.…”

Section: Introductionmentioning

confidence: 91%

Combining Model‐Based Near‐Term GDP Forecasts and Judgmental Forecasts: A Real‐Time Exercise for the G7 Countries

Jansen

Winter

2018

Oxf Bull Econ Stat

View full text Add to dashboard Cite

We investigate the effects of combining model‐based near‐term GDP forecasts and judgmental (quarterly) forecasts by professional analysts (Consensus forecasts) in a real‐time setting for the G7 countries over the years 1999–2013. Model‐based forecasts are produced by a dynamic factor model (DFM). We consider as combination schemes the weighted average and the linear combination. Combining with subjective information delivers sizable gains in forecasting ability of statistical models for all countries except Japan, even when subjective forecasts are somewhat dated. Accuracy gains are much more pronounced in the volatile period after 2008 due to a marked improvement in predictive power of Consensus forecasts relative to the DFM. A possible explanation is that mechanical models may be more vulnerable to extreme observations in estimation samples. Consensus forecasts are superior at the moment of publication for most countries since 2008. For some countries forecast combinations can improve upon Consensus forecasts in between the quarterly release dates of the Consensus survey.

show abstract

Forecast Combinations

Cited by 27 publications

References 32 publications

The Information in Systemic Risk Rankings

The Information in Systemic Risk Rankings

Point and density forecasts for the euro area using Bayesian VARs

Combining Model‐Based Near‐Term GDP Forecasts and Judgmental Forecasts: A Real‐Time Exercise for the G7 Countries

Contact Info

Product

Resources

About