Regime determination is usually based on macroeconomic indicators like GDP and CPI. The problem with this approach is that markets largely overlook current economic conditions. Major market moves are driven instead by expectations of future economic conditions. This means that markets often show flat returns during economic strength and recoveries during periods when the economy is still in recession.
I’ve spent over 10 years developing a proprietary model delivering regime classifications with strong predictive power spanning multiple countries and asset classes.
How does it work?
The model calculates a numerical score for a large sample of economic data. These scores are then aggregated into seven indicators; Composite, Employment, Housing, Growth, Inflation, Leading and Surprise. The Growth and Inflation indicators are used to determine the economic regime. Approximately 30 variables are calculated for each economic indicator using non-parametric statistical methods in addition to proprietary formulas. These variables are then combined into a single score based on weighting factors. Using these scores allows us to quantify the movements in economic indicators.
Successful Regime based Asset Allocation doesn’t require a crystal ball. Regime Switching allows us to know when the future is going to change.
The chart above provides a clear example of the differences between the regime model and a traditional macro indicator. I’ve zoomed in on the last US recession occurring from 2007 – 2009 (classified retrospectively by the NBER). The red shading indicates the period of recession formally defined as 2 negative quarters of GDP. The green shading indicates periods where the US economy entered Regime 4 (Low Growth / Deflation).
When the Regime switches into 4, the future is likely to entail periods of declining market returns if not a full recession. In the chart above, regime 4 switching occurs from August 2007, approximately 16 months before the official GDP figure indicated a formal recession. The Regime 4 periods also occurred before the start of the SP500 drawdown of > 50%.
Thousands of data points, 4 Regimes
Each day, there are approximately 50 economic data releases. The model runs multiple times per day, performing over 3000 calculations to compute the scores that form the basis for the indicators and regimes for 15 countries;
- New Zealand
That’s a lot of number crunching, all distilled down into simple, actionable data.