Objective research to aid investing decisions
Menu
Value Allocations for June 2019 (Final)
Cash TLT LQD SPY
Momentum Allocations for June 2019 (Final)
1st ETF 2nd ETF 3rd ETF

Test of Seasonal Risk Adjustment Strategy

Posted in Calendar Effects, Volatility Effects

A subscriber requested review of a strategy that seeks to exploit “Sell in May” by switching between risk-on assets during November-April and risk-off assets during May-October, with assets specified as follows:

On each portfolio switch date, assets receive equal weight with 0.25% overall penalty for trading frictions. We focus on compound annual growth rate (CAGR), maximum drawdown (MaxDD) measured at 6-month intervals and Sharpe ratio measured at 6-month intervals as key performance statistics. As benchmarks, we consider buying and holding SPY, IWM or TLT and a 60%-40% SPY-TLT portfolio rebalanced frictionlessly at the ends of April and October (60-40). Using April and October dividend-adjusted closes of SPY, IWM, PDP, TLT and SPLV as available during October 2002 (first interval with at least one risk-on and one risk-off asset) through April 2019, and contemporaneous 6-month U.S. Treasury bill (T-bill) yield as the risk-free rate, we find that:

The following table compares net performance statistics for the seasonal risk-on/risk off strategy (Risk-on/Risk-off) and the benchmarks over the full sample period. Risk-on/Risk-off achieves the highest CAGR while being competitive with 60-40 for MaxDD and Sharpe ratio.

Annual net Sharpe ratios for SPY, IWM, TLT, 60-40 and Risk-on/Risk-off for full years from the end of April 2003 through the end of April 2019 (using the average of two semiannual 6-month T-bill yields each year as the risk-free rate) are 0.59, 0.52, 0.50, 0.80 and 0.77, respectively.

Note, however, that PDP and SPLV are not available for substantial parts of the full sample period.

For perspective, we look at cumulative performances.

The following chart tracks values of $100,000 initial investments in Risk-on/Risk-off and the benchmarks over the full sample period. It appears that Risk-on/Risk-off concentrates during about 2008-2012.

As a robustness test, we look at the subperiod since all assets considered are available.

The next table compares net performance statistics for Risk-on/Risk-off and the benchmarks over the subperiod starting at the end of October 2011. SPY and therefore 60-40 are unusually strong and steady based on 6-month returns during this subperiod, such that an investor in Risk-on/Risk-off may be disappointed. In fact, Risk-on/Risk-off underperforms SPY in each April-to-April year after 2013. Performance of Risk-on/Risk-off is nevertheless good.

In summary, evidence from a short test supports some belief in the effectiveness of shifting to relatively high-risk (low-risk) investments during November-April (May-October).

Cautions regarding findings include:

  • The sample period is not long for the kind of test performed (only 16.5 years). PDP (SPLV) prices are available only for 6-month intervals starting at the end of April 2007 (October 2011) and later.
  • As noted, performance statistics are semi-annual, as though an investor using the strategy looks at the portfolio only every six months. MaxDDs measured with higher-frequency data would be deeper.
  • Findings ignore tax implications of trading, if any.

For conceptually similar analyses, see “Kaeppel’s Sector Seasonality Strategy” and “Bonds During the Off Season?”

Why not subscribe to our premium content?
It costs less than a single trading commission. Learn more here.
Daily Email Updates
Login
Research Categories
Recent Research
Popular Posts