One further wrinkle. Part of what makes market timing so hard is that it requires at least two timing decisions: when to sell and when to re-buy. For our simulation we used three ‘reasonable’ re-buying strategies, namely:
• Clients always reinvest at the beginning of the next quarter.
• Clients wait a bit longer to reinvest, re-buying not at the beginning of the next quarter but the beginning of the following quarter (or the next if there is a subsequent 10 per cent drop rule breach).
• Clients follow a similar strategy when buying as when selling – they follow the markets closely and wait until the index has bounced by 10 per cent from its lows before reinvesting (that is, once there is superficial evidence of a recovery).
Winners and losers
To be clear, the period under review was not an ideal cycle to be invested as it assumes investing just before the bursting of the tech bubble in 2000 and includes the credit crunch sell-off in 2008.
Nevertheless, clients who remained invested would have seen their portfolios more than double over the period after factoring in dividends and share price growth. Annualised, this was 4.6 per cent a year.
Those that panicked and sold out after every 10 per cent drop would all have done considerably worse. Sellers who reinvested quickest, at the beginning of the next quarter, would have done least worst – up about 70 per cent over the period (3.2 per cent a year).
Those that hesitated longer, reinvesting only at the beginning of the second quarter after a 10 per cent drop out, would have experienced poorer performance – up roughly 30 per cent overall, about 1.7 per cent a year.
But worst of all would have been the performance of those that only reinvested after seeing a 10 per cent bounce. The whipsawing volatility of the credit crunch would have been particularly cruel to these investors who would have experienced periods of pulling out after a big drop, missing the 10 per cent bounce and reinvesting just in time to experience another 10 per cent drop. They would have been up only 9 per cent over the period – 0.5 per cent a year.
The above results are pretty scary and do demonstrate just how damaging panic selling can be. But a note of warning about over-interpreting the results: changing the underlying index, the start date of the analysis and reporting period dates (say tax year end quarters instead of calendar quarters, for example) does alter the results and although in most cases market timing strategies underperformed it is possible to get some to outperform by choosing the inputs carefully enough.
