Transcript P7.2.6

Indicators to measure cyclical
movements
Julian Chow
United Nations Statistics Division
Scope of chapter 5
• Describe conventional approaches that
define business cycle
• Address specific statistical and data issues
in cyclical movement measurement
• Focus on the phase of data compilation
Key points
• Reference variables (real statistical variables
versus latent variables)
• Advantages and drawbacks of referring to
alternative cyclical definition
• The usefulness of simultaneous monitoring of
classical and growth cycle
• Limits of detrending techniques (e.g. instabilities
of current estimates)
• combining cyclical estimates (subjective versus
statistical aggregation)
Outline of chapter 5
• The choice of target variables
• The choice of the reference cycle (classical, growth or
acceleration cycle)
• Filtering techniques to achieve noise minimization and a
proper estimation of trend-cycle component
• Detrending methods: Parametric versus non parametric
(univariate versus multivariate)
• Aggregation of individual signals: choice of the weights
versus combining forecasts
• Multivariate detrending methods and filtering
• Some examples
The choice of target variables
• Real statistical variables versus latent variables
• Choice of variables that are used as components of
composite business cycle composite indexes
• Subject to the agreement, the section may
recommend a harmonized set of component
indicators that can be used for the construction of
business cycle composite indicators
The choice of the reference cycle
•
The section aims to clarify the kind of
business cycle people are referring to.
1. Classical cycle
2. Growth cycle
3. Acceleration cycle
• Usefulness of simultaneous monitoring of
classical and growth cycle
Filtering techniques and
detrending methods
Two parts:
• Filtering techniques to achieve noise
minimization and a proper estimation of
trend-cycle component
• Trend-cycle decomposition to obtain the
cyclical component
Detrending methods
•
Non-parametric methods
1.
First differencing filter
2.
Henderson filter
3.
Phase Average Trend Method (NBER)
•
Parametric methods
1.
Beveridge-Nelson decomposition (1981)
2.
Unobserved component model (Havery 1985)
3.
Stock and Watson decomposition (1986)
4.
Hodrick-Prescott filter (1997)
5.
Baxter-King filter (1999).
6.
Christiano-Fitzgerald filter (1999)
Aggregation of individual signals
•
Describes the methodology of different types of weighting and
aggregation methods that use to combine individual component
series into composite indicators.
1. Equal weights
2. Standardization factors (Conference Board)
3. Weights based on principal component analysis and factor
analysis
4. Regression analysis
5. Unobserved component models
6. Weights based on public/expert opinion
•
The underlying assumptions, advantage and disadvantage of each
method will be discussed.
Multivariate detrending methods
and filtering
•
•
Rationale of multivariate detrending methods
Methodology of different types of multivariate
detrending procedure. The proposed methods are
listed as follows
1. Multivariate Hordrick and Prescott Filter
(Laxton and Telow 1992)
2. Multivarate Beveridge and Nelson
decomposition (Evans and Reichlin 1994)
3. Multivariate unobserved component
decomposition (Harvey 1985)
Discussion 1
• Is it possible to recommend a common approach
to define business cycle in the handbook?
• Should indicators be based on a common set of
basis statistics?
• Should the compilation based on a standard
methodology or countries should choose among a
set of recommended approaches by selecting the
one giving the best performance with respect to
specificities of each country?
Discussion 2
• What is our opinion on the dilemma between
subjective approach and statistical approach in the
process of variable selection and weight definition
to construct composite indicators?
• Please recommend methods, such as filtering
techniques, detrending and aggregation method,
that are appropriate to be included in this chapter.
Discussion 3
• Please suggest any other key points needed
to be addressed in this chapter.
• Please recommend documents/papers that
can be used as basis for the eventual text in
this chapter.