 September 2003 Taming Turnover
We've disagreed with this before (The Premature Death of Buy and Hold), but still there's no denying that any portfolio will benefit from periodic review and fine tuning. Yet how much is enough? That's the real question. What we're talking about here is turnover -- how often stocks are bought and sold in a portfolio. The average equity mutual fund turns its portfolio about 1.5 times a year.
In its first three years of existence, quant Portfolio 3 averaged 157% average turnover. That's greater than the average "managed" equity fund. In itself, that wouldn't be such a big deal if the same stocks didn't seem to come and go. We documented this phenomenon in March 2002 and in July of this year made an adjustment to the portfolio's quantitative process that could potentially lower the overall turnover. So far, turnover's down, but what about performance? Change for Less ChangePortfolios 3 is based on a regression formula utilizing ten years of data from the performance of the S&P 500. This model makes no distinction between sectors, industries, or market caps, it simply seeks those stocks that score the highest in its regression formula. Its goal is to track and hopefully, outperform the S&P 500.
P3's regression formula is rerun on the 15th of every even numbered month. Since it's reconstituted six times a year, there's plenty of opportunity for turnover -- and that's what we got in its first three years. High turnover would be expected when market conditions are constantly changing, but the past three years have been one drawn out bear market. Although there were periodic rallies and selloffs, overall conditions were relatively stable -- and bad. In hindsight, a defensively positioned portfolio would have done the best. Turnover really wasn't needed. In fact, turnover may have actually hurt. With all the short-lived rallies and selloffs, it was quite easy to get whipsawed. Indeed, one could argue that's exactly what happened to P3. From its inception on July 1, 2000 through June 30, 2003, P3 lost 52.36% while the S&P declined 32.01%. All that turnover certainly didn't help. Perhaps the most concerning thing about the turnover wasn't the degree, but the fact that many of the same stocks that left one month came back the next time the portfolio was rebalanced. Often they left again on the next rebalancing. Strictly speaking, that's not really turnover, it's just churn. When we examined what was going on, we found that many of the stocks that would come and go did so for a rather arbitrary reason, one we ourselves imposed. Portfolio 3 is designed to hold the top 30 stocks in its regression formula. Originally, stocks remained in the portfolio as long as they rated 40 or better when the regression was rerun. If they fell below that rating they were removed and the next highest rated issues were added. Indeed, we found that of the 6.83 stocks exiting the portfolio every two months, 4.62 were still rated below 100. In other words, 60.1% of the turnover was the result of replacing stocks that were still in the top quintile of the S&P 500. This was an extremely tough standard and accounted for most of the churn in the portfolio. Often stocks just falling into the 40s or 50s would have to be removed, only to climb back into the top 30 in the next two months. We didn't intend to make any changes to the quant portfolios' process until they had been in existence for three full years. When that time was up in June, this was the only aspect of P3 we changed. Starting with the June 2003 formulation, stocks must fall below 100 to be removed from the portfolio. They still have to be in the top 30 in order to be added. Composition and PerformanceSo far, turnover has quieted considerably. In June, only 3 of the 30 stocks were replaced: J. P. Morgan Chase fell to 199, Advanced Micro Devices 464, and Edison International 477. On the other hand, Andrew Corporation, that had fallen to 52, was retained. That was roughly a 60% decrease, just as predicted.
In August the previously unthinkable happened: there was absolutely no turnover. Talk about lowering the average! (For the complete history of changes and current constituents of the portfolio, see Stocks of P3.) Actually the lack of turnover makes sense. If you think about it, the S&P 500 was range bound in the June-August period. Although there were daily fluctuations, it moved in a relatively tight range between 960 and 1020. This coupled with no major changes in market sentiment would speak to few, if any, changes. That's precisely what the model did. So if the reduction in turnover was the goal, it's certainly being achieved by this change. But turnover alone isn't the concern, performance is. The only reason we wanted to reduce turnover was to potentially improve performance.
Here too, things may be looking up. The nearby charts compare the performance of P3 to the S&P 500. Chart 1 covers the initial three years while Chart 2 shows the past two and a half months. Obviously you can't make too many valid conclusions in light of the time difference. There were two and a half month periods of outperformance in the first thirty-six, yet the overall period was pretty bad. In addition, Chart 1 covers the equity bear market while Chart 2 reflects a time period that was quite favorable for equities. But while it's still too early to come to any definitive conclusions, it is encouraging to see the modified P3 getting off to such a good start against the S&P 500. It's much easier to maintain a lead than to make up ground to come from behind. Search this site! Just enter you key word or words:
Get current quotes or follow your own custom portfolio,
courtesy of E-Line Financials:
|
||||||||||||||||||||||||||||||
ANY PROFESSIONAL MONEY MANAGERS will tell you that a buy-and-hold strategy won't work. Aside from the fact that they have a vested interest in promoting active money management, they'll tell you that in an ever changing market, you must be willing to buy and sell as conditions warrant.
