strsubcodechain Composite Leading Indicators - Data and Methods
Sources and Definitions
 
Search OECD Statistics
    Sources, Definitions       & Public Data (Home)
    Business Cycle 
      Analysis Database
    Key Economic
      Indicators
    Real-Time &       Revisions Database
    MEI Home
    Contact Us
      
[Download Public Data][Version française] | [Printable Version]
Subjects > Composite Leading Indicators

|Australia | Austria | Belgium | Brazil | Canada | Chile | China | Czech Republic | Denmark | Estonia | Finland | France | Germany | Greece | Hungary | India | Indonesia | Ireland | Israel | Italy | Japan | Korea | Luxembourg | Mexico | Netherlands | New Zealand | Norway | Poland | Portugal | Russian Federation | Slovak Republic | Slovenia | South Africa | Spain | Sweden | Switzerland | Turkey | United Kingdom | United States | OECD - Total | Major Seven | OECD - Europe |



Subject: Component series


.LOCN.......
Key statistical concept
The component series are time series which exhibit leading relationship with a reference series at the turning points. Their seasonally adjusted or raw form is combined into a CLI.


OECD CLI > Component series

.LOCO.......
Key statistical concept

The OECD CLIs are composite indicators with components that:

  • measure early stages of production,
  • respond rapidly to changes in economic activity,
  • are sensitive to expectations of future activity or
  • are control variables that measure policy stance.

All components are passed through a series of filters before aggregation (seasonal adjustment, outlier detection, trend-removal, smoothing and normalisation). The composite indicator is constructed to preserve the leading properties of the components, but at the same time have more stable lead times, and have fewer missed or extra turning-points with respect to the business cycle than the components alone.




Subject: Leading Indicators OECD > Composite indicators > Total leading indicator > Amplitude adjusted

.LOCSTL04.......
Key statistical concept
For the CLI, the trend is eliminated from all component series included in the CLI by dividing the original component series by their trends and then aggregating to give the ratio-to-trend CLI. This series is then adjusted to ensure that its cyclical amplitude on average agrees with that of the de-trended reference series to form the amplitude adjusted CLI.

Estimation
The amplitude adjusted CLI is derived by adjusting first the mean of the detrended CLI to unity; and then adjusting its long-term cyclical amplitude to be equal to that of the detrended reference series.


Subject: Leading Indicators OECD > Composite indicators > Total leading indicator > Trend restored

.LOCSTL05.......
Key statistical concept
The trend restored series of the CLI is an alternative way to present the OECD CLI in which the long-term trend of the reference series has been incorporated.

The OECD uses the year-on-year growth rate of the CLI as its preferred pointer to possible turning points, as year-on-year growth rate is less volatile and provides earlier and clearer signals for future turning points than the CLI itself.

Estimation
This form of the CLI is calculated by multiplying the amplitude-adjusted CLI by the trend of the reference series to obtain the trend restored CLI. This trend-restoration enables direct comparison with the reference series.

The year-on-year growth rate of a CLI for a given month t is {[X(t)/X(t-12)]-1}*100, where X(t) is a CLI for month t (f.e. January 1997) and X(t-12) is a CLI for month t-12 (f.e. January 1996).


OECD CLI Amplitude Adjusted

.LOLITOAA.......
Key statistical concept

The CLI amplitude adjusted is the most straightforward way to present the CLI. The CLI is the average of de-trended and smoothed and normalised component series. The amplitude adjusted CLI rescales this ‘averaged' CLI to match the amplitudes of the de-trended reference series. This form allows for "output-gap" type interpretations.

The formula for calculating the amplitude adjusted CLI is:

CLIAAt=(CLINormt-100)*StDev(RefSeriesRT) + 100

Where CLIAAt is the amplitude adjusted CLI, CLINormt is the normalized CLI and StDev(RefSeriesRT) is the standard deviation of the de-trended reference series.





Subject: Amplitude adjusted
Measure: Quantum (non-additive or stock figures) SA


.LOLITOAA.STSA......
Unit of measure used

IDX





Subject: Normalised
Measure: Quantum (non-additive or stock figures) SA


.LOLITONO.STSA......
Unit of measure used

IDX




OECD CLI Trend Restored

.LOLITOTR.......
Key statistical concept

The trend restored CLI is the product of the trend of the reference series and the amplitude adjusted CLI. This form of the CLI facilitates analyses of classic business cycles. The trend restoration alters the position of peaks and troughs: in general peaks occur later and troughs occur earlier in the trend restored series than in the amplitude adjusted series.

The formula for calculating the trend restored CLI is:

CLITRt=(CLIAAt*RefSeriesTDt)/100

where CLITRt is the trend restored CLI, CLIAAt the amplitude adjusted CLI and RefSeriesTDt is the trend of the reference series.

Besides the levels of the series, we also present/calculate the 12 months rate of change of the trend restored CLI ; fluctuations of these series are comparable with the growth rate of the reference series and turning points of the real GDP.





Subject: Trend restored
Measure: Growth rate same period previous year SA


.LOLITOTR.GYSA......
Unit of measure used
PC



Subject: Trend restored
Measure: Quantum (non-additive or stock figures) SA


.LOLITOTR.STSA......
Unit of measure used

IDX




Subject: Leading Indicators OECD > Reference series

.LORE.......
Key statistical concept
A reference series is a time series which is used to identify the turning points chronologies in OECD CLI system. For the majority of countries the index of industrial production covering all industry sectors (IIP) excluding construction is used as a proxy measure (or reference series) for GDP.


OECD CLI - Reference series

.LORS.......
Key statistical concept
An OECD Composite Leading Indicator, as the name suggests, is constructed from a small number of economic time series that have similar cyclical fluctuations to those of the business cycle, and moreover have a tendency to turn earlier than the business cycle. The business cycle is typically represented by movements in GDP around its long term trend.

Up till March 2012, the OECD system of composite leading indicators has used the index of industrial production (IIP) as a reference series, which is available on a monthly basis and has also, historically at least, displayed strong co-movements with GDP.

In 2011 however the OECD has investigated whether methods could be applied to generate monthly estimates of GDP based on the official quarterly estimates.  This investigation has demonstrated that it is feasible to do so, whilst also continuing to provide high quality results. From April 2012 therefore the OECD has switched to using GDP as the reference, ceasing to rely on the IIP as an intermediate target.  The change means that the timeliness of the reference series decreases by approximately two months but the timeliness of the CLI is not affected and the change brings improved clarity and interpretability of the CLIs; as the CLI-based projections are directly readable as GDP projections. More detailed information explaining the changeover from the IIP to GDP as the reference series is available at: http://www.oecd.org/std/leading-indicators/49985449.pdf

The GDP index derived from chained volume estimates in US dollars converted using 2005 Purchasing Power Parities (PPPs) of GDP and are linearly interpolated and smoothed.





Subject: Normalised
Measure: Quantum (non-additive or stock figures) SA


.LORSGPNO.STSA......
Unit of measure used

IDX





Subject: Original series
Measure: Growth rate same period previous year SA


.LORSGPOR.GYSA......
Unit of measure used
PC



Subject: Original series
Measure: Index publication base SA


.LORSGPOR.IXOBSA......
Unit of measure used
IDX

Reference period
2010



Subject: Ratio to trend
Measure: Quantum (non-additive or stock figures) SA


.LORSGPRT.STSA......
Unit of measure used

IDX





Subject: Trend
Measure: Quantum (non-additive or stock figures) SA


.LORSGPTD.STSA......
Unit of measure used

IDX