Market timing: comparing instruments (SPY, value and beta)

I will use Georg Vrba’s recession model to compare timing versus Buy & Hold for the S&P500, low & high beta and value investing:

blog figs

The buy/sell dates are taken from the COMP-IBH model (above).

Data sources:

Beta series: Falkenstein

Value series: French data library

Results (1970 – 2012) CAGR:

  • The timing model increases returns in all cases, although marginally for low beta.
  • Return on cash during time out of the market is not included therefore timing results are a conservative estimate.
  • Small value is the clear out-performer, both for timing and Buy & Hold.  $1 returns almost $1000 over 42 years (30% of that time in cash).

blog figs

  COMP-IBH Buy & hold
SPX 11.2% 7.4%
Low beta 11.8% 11.4%
High beta 14.1% 6.6%
Small value 17.7% 15.8%

15 thoughts on “Market timing: comparing instruments (SPY, value and beta)

  1. Just an observation but it seems like a lot of the recession signals and IBH signals may be curve fit. While reading up on it I saw they changed IBH to attempt to get out of the market before the ’87 crash and there is a note on the COMP graph that states some of the levels of COMP may differ from previous releases due to revisions of COMP’s components. I’m not suggesting this strategy doesn’t work, however, these attempts at optimization and applying them to history surely artificially improves results.

    • I can assure you there is no curve fitting and our recession indicators BCI and COMP and buy and hold indicators as IBH, MAC , GOLD and SILVER Coppock are always calculated using the latest economic data series (ig FRED, S&P500 etc) and third party indicators as ECRI’s WLI or the Conference Board LEI. Data series like employment data, housing data are revised and these revisions necessitate a revision in our COMP or BCI in the same manor as ECRI’s WLI and Conference Board LEI are revised.

      The art of mathematical modelling is to analyse the history, find models that maximise returns with the hope to that future returns are also maximised, but there is no guarantee – but what I can guarantee it is that these mathematical models are better unbiased decision makers than throwing darts for decision making.

      • Chad and Anton,

        Thank you for your participation and discussion. I understand a concern of recession models is that there are few instances therefore curve fitting may be possible. I was introduced to recession forecasting by Jeff Miller’s blog and there is fantastic information there:

        Jeff is also a big fan of the Vrbas and RecessionAlert and tracks the models weekly: highly recommended.

        Over the next few posts, I have some interesting results without forecasting.


  2. Anton,

    I’m not questioning the it is a very useful indicator that I will use in my decision-making process, my only concern is that when I look at past indicators on the chart, are they different than what would have actually been generated in real-time given that there are revisions to the model that weren’t in the model 20 years ago?

    • Our mathematical model do not get revised, the same mathematical model is applied in our series from 1969 to the present. Revision are in the financial data and are mostly small revisions that do not signal a new or a different trend.

      It would be foolish of me to say that the real-time series as experienced at a any time in the past would exactly match the published data of the model, but it would be pretty close and the trends would not be changed

      The models give a probability, and should the models indicate a recessionary trend, we would not say “a recession is 20 weeks away” but rather “the probability of a recession in 20 weeks is high” in effect it could be 10 weeks or there could be turn around, However, historically there are levels in our indicators from which there is no return.

  3. Also I suppose my main question is, how many false signals would have been given in the past that don’t show up on the chart given revisions in the data? I’m just trying to conclude whether this is applicable in real-time or if the historical charts would not be reflected the same way as it happens. I’m not being critical here, I’m genuinely interested in seeing if I can add this into my investing playbook.

    • In my opinion we would not see false signals but rather a timing shift of the signal, i.e a week or two earlier or later, however with an average lead of 20 weeks ample time to act still remains.

      Instead of a hard go/no-go signal we have developed a comparative and probabilistic view, comparing our BCI indicator to previous tracks to recession, leaving it to the investor to judge when he wants to exit the market. This method we have not released yet. Contact me using the contact page of and I will send you more details of this method.

      A cautious investor will not act on one indicator only and will not switch his stocks to cash at one particular point in time, instead he will spread his investment and gradually exit and re-enter the market.

      And always remember: Past performance is no guarantee for future performance

      • Anton – thanks again for your participation, it is very valuable and I will certainly follow imarketsignals.


  4. Pingback: Combining value and momentum (HT Asness) | RRSP Strategy

    • As far as I know, they are not shown. The monthly average change of those stocks is listed in the spreadsheet.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s