Monday, December 3, 2007

Regime-Switching Models


Sea change, a fundamental, profound change, a regime-switch in modeling terms, not a gradual process as in Shakespeare's The Tempest where Ariel sang:
Full fathom five thy father lies
Of his bones are coral made
Those are pearls that were his eyes
Nothing of him that doth fade
But doth suffer a sea-change
Into something rich and strange.
Your Sunny Day and Rainy Day Screens is an auto regime-switching model. * for reference on Sunny Day and Rainy Day screens see http://valuengine.blogspot.com/2007/10/sunny-day-and-rainy-day-screens.html

James D. Hamilton of the Department of Economics, University of California, San Diego is one of the pioneers in work on regime-switching models. To explain "regime-switching models", we first quote from the introduction of his paper " Regime-Switching Models" (May, 2005) prepared for the Palgrave Dictionary of Economics.
"Many economic time series occasionally exhibit dramatic breaks in their behavior, associated with events such as financial crises or abrupt changes in government policy. Of particular interest to economists is the apparent tendency of many economic variables to behave quite differently during economic downturns, when underutilization of factors of production rather than their long-run tendency to grow governs economic dynamics . Abrupt changes are also a prevalent feature of financial data, ..... calculations for how such abrupt changes in fundamentals should show up in asset prices."
We need not delve into the heavy math, but essentially a regime-switching model is one that can 'operate' in more than one mode. In our context, we have seen that the performance of each of the ValuEngine models [Valuation model, Forecast model, Rating model] depends on the kind of mode [or mood] the market is in. When the market is optimistic, investors place less emphasis on Valuations and more on Growth, and vice versa when the market is pessimistic. If Valuations and Growth were put on a scale from zero to 100, with 100% emphasis on Valuations= zero and 100% emphasis on Growth=100, then at any point in time, the market would be somewhere between zero to 100. A Regime-Switching model with two regimes [Valuations and Growth] could incorporate a threshold that would determine what mode it should operate in. Let us suppose that the Sunny Day screen emphasizes the ValuEngine Forecast[growth] model , and the Rainy Day screen emphasizes the Valuation model. [The Rating model is on overall quality and factors in both Valuations and Growth], then if you have a portfolio whose component stocks are from both the Sunny Day screen and the Rainy Day screen, you would have an automatic regime-switching model. Because the number of stocks that the screening criteria of each screen would output would automatically reflect the mode the market is in. Therefore, if the market is emphasizing Valuations (Rainy), your portfolio [a combination of the Sunny Day and Rainy Day screens] would automatically be biased towards stocks with good valuations. And the same goes for when the market is emphasizing Growth [Sunny]. In fact, your portfolio has become an automatic regime-switching model that is smoother than ordinary regime-switching models that abruptly switch between two regimes in an all or nothing, yes or no, black or white way triggered by a threshold value. [the number of stocks in your portfolio would vary as you rebalance your portfolio each month] You have managed to operate in the real world of in-betweens and shades of gray. * if a stock appears twice [i.e.in both the Sunny Day screen as well as the Rainy Day screen] then give it double-weighting. When the market's mood changes, your screen would automatically adjust by changing the ratio of number of stocks in each screen.

No comments:

Post a Comment