Ricardo Mestre

1
Joint EU-OECD Workshop on International
Development of Business and Consumer tendency
SurveysUse of Business and Consumer pinion
Survey Data in the ECB
Ricardo Mestre
Brussels, 14-15 November 2005
2
1. Structure of Presentation

• Actual use of surveys in the ECB
• Type of surveys in use
• Type of data used in the analysis
• Type of analysis performed
• Survey information that could be of help

3
1. Actual Use of Surveys in the ECB

• Conjuncture analysis
• Inflation indicators (e.g., long-term inflation
  expectations)
• Cyclical indicators
• Indicators for forecasts
• Information is timely and exhaustive
• Research analysis
• Rationality of expectations
• Historical information consistent and long

4
2. Type of Surveys Used in Analysis

• European Commission Consumer survey, Industry
  survey, Construction survey, Retail trade survey,
  Service sector survey
• Survey of Professional Forecasters
• Purchasing Managers
• Economic Consensus Forecasts
• Country-level surveys NBB, Ifo, ZEW, INSEE, etc.

5
3. Type of Data Used in Analysis

• Confidence indicators
• Consumer confidence
• Other
• Balance of opinions
• Quantified qualitative questions
• In special, for inflation
• Raw questions
• Used for analysis based on factor-extraction
  techniques

6
3.a Confidence Indicators
Industrial production and industrial confidence
7
3.a Confidence Indicators
Private final consumption and consumer confidence
8
3.a Confidence Indicators
Retail trade and retail trade confidence
9
3.b Quantified Questions
10
3.c Raw Questions
11
3.c Raw Questions
12
4. Type of Analysis Performed

• Current analysis
• Surveys directed to firms capture the cycle
• Consumer survey
• Long-term inflation expectations
• Forecasts based on indicators
• Surveys provide useful indicators for simple
  forecasts
• Surveys add useful information to factor-based
  forecasts

13
4. Type of Analysis Performed (II)

• Forecasts are generated
• Via simple bridge equations with a small number
  of indicators
• Using large data sets (including survey
  information) to derive dynamic factors

14
4. Bridge-Equation-Based Forecasts

• Bridge equations
• Simple OLS univariate regressions
• Indicators must be timely
• Common indicators
• Surveys EC Economic Sentiment
• Others Ind. Prod., OECD CLI, etc.

15
4. Factor-Based Forecasts

• Factors derived from large dataset
• n variables (xt) , r factors (Ft), q dynamic
  shocks (ut), ngtgtrgtq
• The methodology allows for end-of-sample missing
  data (early exploitation of timely data)

16
4.b SPF Inflation Expectations
17
4.c Indicator-Based Forecasts
Euro area GDP growth forecasts
18
4.d Factor-Based Forecasts
Updates of the signal for 2004qQ2, from April 04
to July 04, GDP growth
Standard error of the common and idiosyncratic
components
19
4.d Factor-Based Forecasts
Updates of the std deviation of the nowcast for
2004qQ2, From April 04 to July 04, GDP growth
20
5. Further Survey Data of Use

• Better coverage of prices (e.g., retail trade and
  service sectors)
• Broader sectoral coverage (e.g., services)
• Ambiguity of some questions (e.g., price
  questions for consumers)
• Time-consistency of data must be preserved!

21
(No Transcript)