We’ve conducted many studies on market declines and recoveries. For example, since World War II, there have been 58 pullbacks (5-10 percent decline), 19 corrections (10-20 percent decline), and 12 Bear markets (+20 percent declines). For each, the average duration was 1, 5 and 14 months, and the average recovery was 2, 4 and 25 months, respectively.
The challenge when evaluating historical data, is determining if a 5-10 percent decline will turn into a 10-20 percent decline. We've taken this type of data and tried to design trading models to take advantage of it. Unfortunately, stock market returns are not normally distributed—although they are modeled as such—or serially correlated. In fact, they are random in the short run, so designing a strict formulaic trading strategy around declines has not been proven to yield consistent results.
Interestingly, when investing new money, we use a similar type of decision making process. That’s because, history has shown no statistically significant pattern as to whether to invest a lump sum or average in.
Of course, there is a benefit to being aware of these historical patterns and having a general awareness of where we are in a decline or uptrend. So we have incorporated tolerance ranges—both percentage-based and dollar-based—for rebalancing and tax loss harvesting around asset classes.