Regime Shift Identification using Using EMA Technical Indicators

Regime Shift Identification using EMA Technical Indicators - Kyle E. Lamb

Problem Statement:

Can we use Regime Shift modeling to create higher risk adjusted returns in our portfolios?

Financial markets can change their behavior rather abruptly at times. The behavior of stock prices throughout different periods in time can be seemingly random and difficult to predict. Many people attempt to trade the market with a goal of creating excess returns in their portfolios, yet most of them fail tremendously. When creating financial models for the stock market, we will find that the mean and variance of stock prices can vary dramatically from period to period. It is important for an investor to adjust their strategy based on different regimes that a market reside in.

In a Nutshell:

This project will attempt to use technical regime shifts to define periods in time where the market is more volatile and/or subject to lower returns based on the specific regime that the market lies in. I defined the model using a technical indicator, called the exponential moving average, and simply create regimes based on whether price is above or below the moving average.

Data Visualization:

I will be using data from the SPY for this study. The SPY is an ETF meant to track the price movements of the S&P500 index. As this index is used as the benchmark for the entire market, it will be a valid dataset to determine overall market trends and regimes. As for the threshold indicator, I will be using the 200 day EMA. If price is below the 200 EMA, that will be defined as a bear regime. If price is above the 200 day EMA, it will be classified as a bull regime. The plot below represents the EMA, where the bear regime is indicated in red and the bull in blue.

For this project, we will not be discussing hedging during bear regimes, that is a separate topic for another time. In this strategy, we will simply be in the market during the bull regimes and out of the market during the bear regimes. Our goal is to decrease volatility and hopefully increase the Sharpe ratio while maintaining similar returns compared to the benchmark. The model uses the strategies defined above, the following plots display the performance of the model over a 20 year period.

Strategy Visualization:

As we can see from the performance report, we have successfully decreased volatility while maintaining close to the same returns. It seems that this strategy was effective in avoiding the volatility and drawdowns that came with the 2000 dot com bubble, as well as the 2008 financial crisis. Due to the speed of the Covid-19 recession, the strategy was not able to completely avoid the losses, which inevitably caused the largest drawdown in the strategies PnL curve.

Conclusion:

It is possible that this strategy could be beneficial. Although it did not produce higher returns, it could assist to avoid the large market drops that typically come with recessions. One possible strategy to consider could be that due to the decreased volatility, which lies in each regime, it could be possible to take on extra leverage such that your overall risk would be equivalent to that of the benchmarks. Overall, I would not consider this a exceptional investing strategy, but it could be used to minimize the risk on your portfolios while preserving similar returns.