Constructing a historical Economic Sentiment Indicator

1
Constructing a historical Economic Sentiment
Indicator
European Commission Directorate General Economic
and Financial Affairs

• Biau Olivier,
• Borg Reuben,
• Smyth Christopher
• Joint EU-OECD WORKSHOP ON
• BUSINESS AND CONSUMER OPINION SURVEYS
• Brussels, 12 October 2009

2
Overview

• Main interest
• EC Economic Sentiment Indicator
• - how did it perform in the current recession
  compared with previous recessions?
• Publically available data for the ESI starts in
  1985
• Attempt is to replicate same process for
  construction of ESI, but going further back
  pre-1985

3
Participating institutes

• IFO (Germany)
• INSEE (France)
• ISAE (Italy)
• CBI (United Kingdom)
• CBS (Netherlands)
• NBB (Belgium)
• SD (Denmark)

4
A description of the data set available pre-1985

• Complete raw monthly data for the following

5
Coverage - country

• Economic weight of included countries in euro
  area and EU aggregates
• Industry DE, FR, IT, NL, BE ( 79 of euro
  area)
• DE, FR, IT, NL, BE, DK, UK ( 74 of EU)
• Construction DE, IT, BE ( 42 of euro area)
• DE, IT, BE, DK ( 32 of EU)

6
Coverage - question

• Questions
• Industry
• 2 order books
• 4 stocks
• 5 expectations
• Construction
• 3 order books
• 4 employment expectations
• Same composition of questions as for the
  Industrial and Construction confidence indicators
  ensures continuity with present ESI
• Methodology documented in European Economy
  Special Report
• The Joint Harmonised EU Programme of Business and
  Consumer Surveys

7
Comparing the Proxy with the Actual euro area ESI
very close representation
8
Comparing the Proxy with the Actual EU ESI
very close representation
9
Constructing the historical ESI - splicing
10
Comparison with historical euro area real gdp
11
An analysis of specific questions stocks the
culprit
12
Conclusions

• Proxy-ESI (for the limited sample) gives a good
  representation of the actual ESI post-1985
• Difference between proxy and actual is mainly due
  to the other sectors (services, consumers and
  retail)
• This permits us to be reasonably confident in
  splicing the historical series in 1985 with the
  actual ESI (no visible break-in series)
• Excessive stock levels during the current
  recession were still deemed to be less than 1970s
  crises
• Importance of stocks in the composition of the
  ESI stock cycle moderated during the Great
  Moderation period