Yesterday’s post examined the results of following the 200-day SMA rule seemed to work on daily values of the Russell 1000 and 2000. The data showed that a 1-day delay in trading would have benefited the Russell 1000 timer, but hurt the Russell 2000 timer. Today’s post looks at the cost of simply delaying until the next day’s open.
Several weeks ago, I posted on “overnight returns.” This followed Falkenblog’s post on the same topic.
After posting yesterday’s analysis on the pre-cost results of applying the 200-day SMA rule to the Russell 1000 and 2000, I realized a big chunk of the effect of a delayed trade is what Falkenstein called the “overnight” return – the return while the market is closed. In his post he listed some half dozen stocks and a couple ETFs whose daily log returns were concentrated in the gap between close and open.
Any trading rule endures a delay between when your threshold is reached and when you execute. If your threshold depends on closing prices, it’s simple enough to measure the impact of not executing until the next day.
I looked at the daily Open/Close history of IWB, the iShares Russell 1000 ETF, and IWM, the iShares Russell 2000, and identified the “overnight return” whenever the 200-day SMA rule, based on the daily underlying index, would have indicated a trade. Those ETFs began trading in May 2000, and through August 2 the SMA strategy would have recommended over 100 trades each. The table below indicates whether the delay would have been advantageous to a 200-day SMA disciple, assuming he began using it on these ETFs in June 2000.
“Returns” indicate the opportunity gain/(cost) due to executing a trade at the following day’s opening price
| IWB (iShares Russell 1000 ETF) | IWM (iShares Russell 2000 ETF) |
|||
|---|---|---|---|---|
| Delay when buying | Delay when selling | Delay when buying | Delay when selling | |
| Average Daily Log Return | 0.20% | 0.04% | -0.33% | -0.05% |
| Best | 3.18% | 1.98% | 1.76% | 2.15% |
| Worst | -2.60% | -1.77% | -3.99% | -2.92% |
| St. Deviation | 1.06% | 0.97% | 1.07% | 1.07% |
| Observations | 51 | 52 | 59 | 59 |
| Returns indicate the opportunity gain/(cost) due to executing a trade at the following day's opening price | ||||
| Data encompasses daily Open and Close prices on IWB and IWM from June 1 2000 through August 2 2011 | ||||
| Source: PortfolioWizards | ||||
It’s interesting that the SMA rule appears to be beneficial when timing the Russell 1000, and a delay would have actually helped, on average. It’s actually not an implementation shortfall, it’s an implementation windfall.