The Use of Recession Indicators in Stock Market Timing

February 19th, 2013

by Georg Vrba

Advisor Perspectives welcomes guest contributions. The views presented here do not necessarily represent those of Advisor Perspectives.

In previous articles I showed that stock market timing models, such as IBH and MAC, can provide better returns than what one could get from a buy-and-hold strategy. However, investors with long-time horizons can also do well by only exiting the stock market at beginnings of recessions and returning to the market when the market is recovering. Recession indicators such as COMP are useful in identifying recession starts. To re-enter the market one can successfully use the buy signals from the IBH model. This strategy would have provided a compound annual return of 15.5% from 1980 to the end of 2012.

COMP is a recession indicator derived from the composite data of three freely available U.S. economic indices, hence its name COMP. (See appendix A for the indices and their components.) The indicator was constructed from 1968 onwards over a period which included seven recessions. COMP oscillates about zero; when it moves below zero it identifies recession periods as shown in figure 1.

My recession indicator evaluation system gave it a score of 0.97, the highest score of all the recession indicators tested. Also using likelihood ratios I found that COMP provided a high 85% probability that a recession exists when it is in recession territory (below zero) and also a low 2.5% probability for the economy to be in recession when it does not signal recession. (See appendix B for details and comparison with other indicators.)

In the following table are the leads to recession starts and ends from COMP; a negative lead denoting that COMP was late to identify the beginning of a recession, or late to signal the recession end. COMP never signaled a recession when there was not one.

Although COMP provides good signals to exit the stock market it is not useful in identifying the entries back into the stock market. Typically, by the time the business cycle has bottomed the stock market has normally already gained a lot from its recession low. To re-enter the market one can use my IBH model which provides good buy signals after the sell signals from COMP. Figure 2 shows all the signals from the COMP–IBH system superimposed on the S&P 500 graph.

Below are the relevant signal dates obtained from the COMP-IBH and COMP-MAC systems with corresponding levels of the S&P 500. One can see that, except for the 1980 recession, it was always advantageous to exit the market after a COMP sell signal and to return at the date of the first occurring buy signal from either the IBH or MAC model. (Shaded green cells in the second column of the two tables below.)

The compound annual returns for an investment in the S&P500 using the COMP–IBH system are shown below for the periods 1966-2012 and 1980-2012, without interest and dividends, with interest and no dividends, and with interest and dividends. Returns are also shown for Buy-and-Hold for the same periods, with and without dividends.

Avoiding the stock market during recessions is a good strategy as is evident from the tables above. The COMP-IBH system produced respectable returns from the stock market over the longer term which should be quite acceptable to investors who do not want to time the market unnecessarily. However, COMP provided no exit signal prior to the October 1987 market crash, because a recession was not imminent then.

Also investors should be concerned about the dubious recession calls from financial analysts. It is best to ignore the many recession calls from those pundits and to rather follow the unbiased signals from a good recession indicator. The current level of COMP is far away from a recession signal as shown in figure 1 below.

Appendix A – Components of COMP

1. Aruoba-Diebold-Scotti business conditions index (ADS)

The underlying (seasonally adjusted) economic indicators of ADS are:

  1. Weekly initial jobless claims,
  2. Monthly payroll employment,
  3. Industrial production,
  4. Personal income less transfer payments,
  5. Manufacturing and trade sales; and
  6. Quarterly real GDP.
2. Growth rate of ECRI's Weekly Leading Index (WLI)

In 1999 ECRI published the components of the WLI (seasonally adjusted, where applicable):

  1. Initial claims for unemployment insurance,
  2. Money supply (M2 plus),
  3. Industrial materials price index,
  4. Mortgage loan applications,
  5. 10 year treasury bond yield / BAA corporate bond yield
  6. BAA corporate bond yield; and
  7. Stock price index.
3. Growth rate of the Conference Board Leading Economic Index® (LEI)

The underlying (seasonally adjusted, where applicable) economic indicators of LEI are:

  1. Average weekly hours, manufacturing,
  2. Average weekly initial claims for unemployment insurance,
  3. Manufacturer's new orders, consumer goods and materials,
  4. ISM® new order index,
  5. Manufacturer's new orders, nondefense capital goods excl. aircraft
  6. Building permits, new private housing units,
  7. Stock prices, 500 common stocks,
  8. Leading Credit IndexTM
    • 2-year Swap Spread
    • LIBOR 3 month less 3 month Treasury-Bill yield spread
    • Debit balances at margin account at broker dealer
    • AAII Investor Sentiment Bullish (%) less Bearish (%)
    • Senior Loan Officers C&I loan survey – Bank tightening Credit to large and medium firms
    • Total Finance: Liabilities – Security Repurchase
  9. Interest rate spread, 10-year Treasury bonds less federal funds,
  10. Average consumer expectations for business conditions.

Appendix B

Using likelyhood ratios I found that there are only a few leading recession indicators which provide high probabilities that a recession exists when the indicators are in recession territory and also provide low probabilities for the economy to be in recession when the indicators do not signal recession. They are shown in the table below. The scores from my recession indicator evaluation system are also shown in the last line of the table.

  • The weekly leading SuperIndex is maintained by
  • The CB-LEIg +2.35% is the six months smoothed annualized growth rate of the recently revised Conference Board Leading Economic Index for the United States with 2.35% added to the growth rate to make zero the recession trigger line.

Georg Vrba is a professional engineer who has been a consulting engineer for many years. In his opinion, mathematical models provide better guidance to market direction than financial "experts." He has developed financial models for the stock market, the bond market, the yield curve, gold, silver, and for the recession indicator COMP, all published in Advisor Perspectives. The models are updated weekly. If you are interested to receive theses updates at no cost, send email request to

Website by the Boston Web Company