Introducing the Euro-Sting: Short Term Indicator of Euro Area Growth

Abstract

We set out a model to compute short-term forecasts of the euro area GDP growth in real-time. To allow for forecast evaluation, we construct a real-time data set that changes for each vintage date and includes the exact information that was available at the time of each forecast. With this data set, we show that our simple factor model algorithm, which uses a clear, easy-to-replicate methodology, is able to forecast the euro area GDP growth as well as professional forecasters who can combine the best forecasting tools with the possibility of incorporating their own judgement. In this context, we provide examples showing how data revisions and data availability a¤ect point forecasts and forecast uncertainty.Keywords: Business Cycles, Output Growth, Time Series. JEL Classi…cation: E32, C22, E27We th a n k th e e d ito r a n d th e tw o re fe re e s fo r e x tre m e ly va lu a b le c o m m e nts w h ich h ave g re a tly im p rove d th e c o nte nts o f o u r p a p e r. We a lso th a n k Fra n c is D ie b o ld , J a n J a c o b s, S im o n va n N o rd e n , K e n Wa llis a n d th e p a rtic ip a nts a t th e 5 th Wo rk sh o p o n Fo re c a stin g Te ch n iq u e s, a t th e 5 th C o llo q u iu m o n M o d e rn To o ls fo r B u sin e ss C y c le A n a ly sis, a n d a t th e inte rn a l se m in a r se rie s o f B a n c o d e E sp a ñ a , th e E u ro p e a n C e ntra l B a n k , C E M F I, th e U n ive rsity C a rlo s I I I, th e U n ive rsity o f A lic a nte a n d th e U n ive rsity o f N ava rra fo r h e lp fu l c o m m e nts a n d su g g e stio n s. We re c e ive d e x tre m e ly va lu a b le re se a rch a ssista n c e fro m C a m ilo U llo a . We a re a lso g ra te fu l to th e E u ro A re a U n it o f th e B a n k o f S p a in fo r a ll th e p re lim in a ry m e e tin g s th a t le a d to th is p a p e r. F in a lly, w e a re th a n k fu l to J u a n A y u so fo r h is c le ve r id e a c o n c e rn in g th e n a m e o f th e in d ic a to r. A ny re m a in in g e rro rs a re o u r ow n re sp o n sib ility. P a rt o f th e p a p e r w a s w ritte n w h e n th e …rst a u th o r w a s v isitin g th e B a n k o f S p a in . M a x im o C a m a ch o th a n k s th e …n a n c ia l su p p o rt o f th e p ro je c t S E J 2 0 0 6 -1 5 1 7 2 . T h e v ie w s in th is p a p e r a re th o se o f th e a u th o rs a n d d o n o t re p re se nt th e v ie w s o f th e B a n k o f S p a in o r th e E u ro S y ste m .

Introducing the Euro Area‐wide Leading Indicator (ALI): Real‐Time Signals of Turning Points in the Growth Cycle from 2007 to 2011

“…Euro area applications are the monthly published 'Eurocoin' (Altissimo et al, 2001) and 'New Eurocoin', which are designed to track a monthly 'smoothed' estimate of real GDP growth (Altissimo et al, 2010). Dynamic factor models have also been applied for GDP forecasting (see, for example, Angelini et al, 2008;Camacho and Perez-Quiros, 2008) complementing bridge equation models (Diron, 2008;Hahn and Skudelny, 2008). Markov switching models provide probabilities for the different stages of the economic cycle (Clements and Krolzig, 1998; for euro area applications, see Krolzig, 2001, andArtis et al, 2004).…”

Maximum Likelihood Estimation of Factor Models on Datasets With Arbitrary Pattern of Missing Data

“…15 The nowcasting framework of Giannone et al (2008) and of many similar applications has been based on large datasets. However, some recent studies have favoured the adoption of small and medium-sized models (see, for example, Camacho and Perez-Quiros, 2010;Frale et al, 2011). M. BAŃBURA AND M. MODUGNO The number in parentheses reflects the 'shift' with respect to the reference quarter.…”

Spain‐sting: Spain Short‐term Indicator of Growth*

“…1 Remarkably, with this specification we get to explain 79 per cent of the variance of GDP growth with the evolution of the common factor. In the second extension of the Stock-Watson set of indicators, we follow the procedure described in Camacho and Perez Quiros (2010) which is based on including more variables into the model whenever they increase the variance of GDP explained by the common factor. Contrary to standard techniques, more explanatory variables do not always increase the variance of GDP explained by the model.…”

“…As expected, the loading factors are positive and statistically significant which reinforces the standard view that the indicators are procyclical. In addition, as in Banbura and Rünstler (2007) or Camacho and Perez Quiros (2010), the empirical results show that a suitable treatment of publication lags may lead some indicators to provide important sources of information in predicting the GDP beyond the information provided in the in-sample estimates of the loading factors.…”

“…On the other hand we have the small-scale factor models which rely on previous reasonable pre-screenings of the series. These specifications, which have recently been applied by Camacho and Perez Quiros (2010) and by Frale et al (2008) to euro-area data, are estimated by using strict factor models.…”