Just as ZIRP (zero interest rate policy) and NIRP (negative interest rate policy) dilute bond-pricing models, the Trump rally defies equity pricing models. The stock market has soared for over 90 days consecutively without a 1% daily drop and has become a one-sided bet. Although both fundamental and technical valuations have flashed to highly unusual levels, market sentiment remains extremely positive, and indexes thrust through one new high after another. If pricing models cannot precisely capture the latest information in a precipitous shift into uncharted waters, should market participants just invest on the basis of their intuition and sentiment? Probably not. Oftentimes, one important reason why old pricing models require some fine-tuning is a sudden regime change due to a recent market event, such as the November US presidential election or the implementation of ZIRP following the financial crisis.
So what constitutes a regime change? During certain periods, asset prices in financial markets behave dramatically and persistently different. This behavior is opposite to a transitory change such as a quick jump in price. Moreover, measures of regime change go beyond just asset price. For instance, the mean, autocorrelation, and volatility of stock returns. Some regime changes, such as recessions and expansions, can happen cyclically, while others happen unexpectedly, as in 2008 and 2009. While equity markets prefer to categorize regime shifts as either high- or low-volatility environments, bond markets use monetary policy to identify regime change. Coincidentally, easing and tightening policies have historically been followed by substantially different economic cycles. As evidenced by Figure 1 (source: fred.stlouisfed.org/series/FEDFUNDS), the nine most recent US recessions (shaded areas) were mostly preceded by a period of Fed rate hikes. As we are now lifting off from ZIRP, are we going to enter into another recession? Time will tell. Another typical regime change measure is political policy. What makes the current market particularly interesting is that we are experiencing both political and monetary secular changes that make asset-pricing work even more complex than it would be under either condition alone.
Numeracy is absent at the beginning of each regime change. This is a typical challenge investment professionals face when markets experience secular changes. Rather than rely on murky narratives, econometricians have provided regime-switching models to capture the dynamics of prices and fundamental changes in financial markets. Of course, the repricing path requires time to analyze and readjust to new information. Since the consequences of regime changes do not fall within the normal distribution of past events, future pricing needs to take nonlinearity into consideration. Analysts can combine a number of conditional distributions to achieve such goal. The approximations achieved by regime-switching models can provide close insights into future returns even though the most accurate model is still unknown. In the long term, investment professionals can evaluate and improve the regime-switching model once they have obtained enough data, which is an important process for dealing with any future regime changes. However, a typical outcome from regime-switching models is that counterintuitive results may occur – and thus gut feeling may rival model building and data analysis as one era gives way to the next.
While quantitative analysis uses numerical values to run simulations and perform scenario analysis, conventional wisdom relies on past experience and future expectations. If investors seek to price in forward changes based upon the anticipation of future events, then biased expectations may play a key role in the future pricing process. We previously discussed some behavioral inefficiencies in financial markets here: cumber.com/market-volatility-etf-portfolio-4q-2016-review-market-biases/. Both confirmation bias and recency bias are likely to affect investors’ views of the near future. Academic literature has documented that investors tend to overreact to a series of good news reports and underreact to idiosyncratic announcements. This tendency leads to short-term momentum and intermediate- to long-term mean reversals. Put differently, markets that are exposed to a protracted string of good news reports attain extremely high valuations, which return to the mean eventually. This overreaction pattern is a persistent source of distortion in financial markets.
Tversky and Kahneman (1974) studied the behavioral heuristic known as representativeness. They suggest that stock market investors tend to ignore the laws of probability and the reality that there are few companies that just keep growing. Consensus opinion and “herd forecasting” may dominate, even though expectations differ among investors, especially when the narratives that drive market movements lack crucial details. In the end, however, numbers do not lie. Prices will adjust to the correct valuation levels, although the path to that restorative state is nonlinear.
At Cumberland Advisors, our quantitative analysis monitors regime changes and studies market dynamics. We also strive to exclude human behavioral inefficiencies. We believe the best investment decisions are based upon diligent work and rigorous research.
