Introduction section TRILL: The TReasury ItaLian Leading Indicator

TRILL: The TReasury ItaLian Leading Indicator

TRILL is a monthly indicator developed in year 2009 by the Direction I of the Department of Treasury, for monitoring the state of the Italian economy and computing short-term forecasts.[1] The indicator results from the combination of three models, which rely on recent research on mixed-frequency factor models.[2]

Mixed-frequency models are particularly useful in short-term forecasting, as they allow an optimal use of ragged-edge datasets, which contain indicators released with different delay and observed at different time frequency. This fact occurs often in the practice of macroeconomic forecasts, and thus these models represent a significant progress with respect to other approaches commonly used by institutions. Moreover, the decomposition of GDP data in monthly figures results in a coincident indicator of the business cycle that is coherent with official quarterly GDP data released by ISTAT.

The peculiarity of TRILL consists in combining the multivariate approach of mixed-frequency factor models with an innovative polynomial structure aiming at employing efficiently the leading properties of short-term indicators. Italian quarterly GDP is indeed temporally disaggregated and forecasted on the basis of different possible lags of monthly indicators. TRILL is innovative not only methodologically but also for the data chosen. Indeed, TRILL employs variables generally not included in the practice of short-term forecasting, such as truck traffic, electricity consumption, paper and paperboard production and loans to the business sector.
 

[1]
The methodology is documented in Frale, C. and Monteforte, L. (2010),
"FaMIDAS: A mixed frequency factor model with MIDAS structure" (PDF 3,32 MB), MEF and Department of the Treasury working paper series, n. 3/2010.
[2]
Cfr.  Frale, C., M. Marcellino, G. L. Mazzi, T. Proietti, (2010), “Survey data as Coincident or Leading Indicators”, Journal of Forecasting 29, special issue: advances in Business Cycle Analysis and Forecasting, Jan-Mar 2010 Frale, C., M. Marcellino, G. L. Mazzi, T. Proietti, (2010), “EUROMIND: A Monthly Indicator of the Euro Area Economic Conditions”, forthcoming in the Journal of the Royal Statistical Society series A.

Related Content