Last update 31 January 2006
All files are in PDF format (size ranges from 150kb to 1.3 Mb)
|Forecasting corporate investment: An indicator based on
revisions in the French investment survey, October 2005|
The quarterly survey of investment in industry, conducted by the National Institute for Statistics and Economic Studies (INSEE) is a prime source of information concerning short-term evolutions in productive investment, making it possible to estimate these evolutions at an early stage and with considerable precision.
However, the annual nature of the questions posed makes it is difficult to use the results for forecasting on a quarterly basis. This article proposes a quarterly indicator based on revisions in industrial firms’ expectations regarding their investment. The proposed indicator measures the adjustments to investment figures made during the year in response to changes of a short-term nature. It turns out to be closely correlated with quarterly evolutions in firms’ investment as measured in the national accounts. Moreover, it is available roughly three months before the publication of the initial quarterly national accounts figures.
As the distributions examined fail to verify the classic normality hypothesis (thick tails and heavy concentrations at zero) it is necessary to apply an estimation method that is robust to extreme revisions. Taking into account also the presence of heteroscedasticity, the method known as "Quasi-generalised M-estimators" was applied.
| A comparison on the company level of Manufacturing
Business Sentiment Survey data and Realized Turnover, September 2005|
Using a classification method developed in this paper, the quality of qualitative survey data of the manufacturing industry at micro-economic level is investigated. For single companies, recent opinions on recent production developments are compared to quantitative results of industrial turnover. The results show that 57.6% of the analyzed companies give useful qualitative answers for calculating meaningful balance statistics such as producers’ confidence. The level of agreement between quantitative and qualitative data for companies with seasonal patterns in turnover on average is 10.6%-points higher than for companies without seasonal patterns.
| Using the results of qualitative surveys in quantitative
analysis, September 2005|
The answers to qualitative questions put to economic operators can be integrated in standard macro-economic analysis by using a "quantification" procedure chosen among the probabilistic approach, the regression methods or the latent factor approach. The first one is the most commonly used. It is based on the assumption that the respondents reply that the value of the reference variable x can be described by a certain statement (e.g.: x stays stable) if it lies between two known thresholds (e.g.: ±5% around its initial value). A number of quantified indicators may be derived by assuming a special functional form for the frequency distribution of opinions and expectations of respondents about x. According to the regression approach, the respondents attach to each qualitative answer a reference value of x, which can be estimated by regressing an available quantitative measure of x against the time series (or longitudinal samples) of percentages of people who gave each qualitative answer. Finally, one can assume that every percentage of answers is driven by a single common "latent factor", which can be estimated by applying to the percentages of answers the standard tools of multivariate statistics. All of the three approaches include as a special case the "balance statistic", adopted very often in empirical analyses. Both the theoretical analysis and the empirical evidence on the relative merits of various methods are mixed. In general, no single quantification technique clearly outperforms the others, at least in preliminary analysis. However, when the quantified indicators must be included as explanatory variables in standard econometric models, the regression approach seems the most suitable and natural one. On its turn, the latent factor approach provides a profitable alternative to the regression method in exploratory analyses and when multicollinearity and degrees of freedom of estimates become severe constraints.
|A Monthly indicator of the Business Climate in the
French Service Industry, July 2005|
Since September 2004, Insee has published the results of its business survey in the services sector on a monthly basis together with a synthetic indicator. This indicator is extracted from both monthly and quarterly balances of opinion derived from the survey. The methodological framework is provided by dynamic factor analysis. Indeed, it is flexible enough to deal with both multi frequency series and changes in frequencies. This model has a state-space representation and can be estimated by a Kalman filter. Alternative models have been explored and yield very similar indexes, which emphasizes the robustness of the Insee indicator.
This index can be obtained at a disaggregated level and provides an economic outlook of the services sector. Thus, it confirms the resurgence of activity in the services sector in France since mid-2003. More precisely, the activity appeared to be rather bumpy in 2004 and tends to decelerate in the beginning of 2005.
| Business Survey Data: Do They Help in Forecasting the
Macro Economy?, June 2003|
This paper examines whether data from business tendency surveys are useful for forecasting the macro economy (GDP, unemployment, price and wage inflation, interest rates, exchange-rate changes etc.) in the short run. The starting point is a so-called dynamic factor model (DFM), which is used both as a framework for dimension reduction in forecasting and as a procedure for filtering out unimportant idiosyncratic noise in the underlying survey data. In this way, it is possible to model a rather large number of noise-reduced survey variables in a parsimoniously parameterised vector autoregression (VAR). To assess the forecasting performance of the procedure, comparisons are made with VARs that either use the survey variables directly, are based on macro variables only, or use other popular summary indices of economic activity. A revised version of this paper has been published in the Journal of International Forecasting, No 21, 2000".
| Forecasting Euro Area Industrial Production Using (mostly)
Business Surveys Data, March 2003|
In this paper we propose a relatively simple procedure to predict Euro-zone industrial production using mostly data derived from the business surveys of the three major economies within the European Monetary Union (France, Germany, and Italy). The basic idea is that of estimating business cyclical indicators to be used as predictors for the industrial production in France and Germany; as far as Italy is concerned, forecasts are produced using a model that in the recent past proved to be able to produce accurate forecasts up to six months ahead. In order to derive quantitative predictors from the business surveys data and to aggregate the nation-wide forecast into the Euro-zone forecast, we propose using an approach based on dynamic factors and unobserved components models. The resulting forecasts are accurate up to six steps ahead.
|An Indicator of Economic Sentiment For The Italian
Economy, October 2002|
The long and sustained expansion of the nineties has generated, especially in the US, widespread rumours about the "death of the cycle". Nevertheless, towards the end of the last decade, it became clear that fluctuations of economic activity were far from being extinct. This has contributed greatly to a renewed interest among economists for the elaboration of statistical indicators capable of tracking and, if possible, anticipating the cyclical features of the economy. The aim of this paper is to build such an aggregate composite indicator for the Italian Economy, based on the ISAE surveys on households and those on the manufacturing, retail and construction sector. The first step of the analysis consists in using a dynamic factor model to extract a "common factor" from the different series of each survey, which may be interpreted as a composite confidence indicator. We then evaluate, for each survey, its in-sample and out-of sample properties, comparing them with those of the usual ISAE-EC Confidence indicators. Finally, we use again the dynamic factor model to build, from the sectoral Composite Indicator (CI), a Composite Aggregate Indicator (CAI) for the Italian economy, and test its ability in tracking the cyclical features of Italian aggregate GDP.
In case of problems with downloading these files, you can contact us.