iM-Best10: A Portfolio Management System for High Returns
Note from dshort: Anton and Georg Vrba subsequently addressed the question of survivorship bias in Survivorship Bias: neither Myth nor Fact.
The S&P 500 has notched up significant gains over the last year. But looking a bit further back then the performance of the stock market is dismal. An investment in an S&P 500 Index fund made 13 and 14 years ago has only produced negative real returns to date. However, one can get much better returns from the S&P 500 by using a ranking system, and then periodically rebalance one's portfolio to hold only the 10 highest ranked stocks of the S&P 500 at any time. This method would have produced average annual returns of about 35% over the last 14 years.
So how does this get done?
To find stocks which may be undervalued all 500 stocks of the S&P 500 index are ranked weekly according to the following parameters, with the highest rank obtainable being 100:
Valuation (measured as market capitalization, debt and cash relative to earnings before interest, taxes, depreciation & amortization, future cash flow and projected earnings), Efficiency (measured as future cash flow relative to total assets), Financial Strength (measured as future cash flow relative to total debt), Short Interest (being the short interest ratio), Trend (measured as the stock price relative to a moving average of the price).
Stocks are sold from the portfolio once their rank falls below 97 and replaced by the highest ranked stocks at that time. The model does not restrict buying to stocks which are ranked 97 or higher.
The model assumes stocks to be bought and sold at the next day's opening price after a signal is generated. Transaction costs accounting for brokerage fees and slippage, together amounting to about 0.35% of the transaction value of each trade, are taken into account. Taxes are assumed to be deferred, as for retirement accounts.
The trading performance is shown in Figure 1 further down. The annualized return over the last 14 years and 5 months was 35.3%, and the maximum draw-down was 64% versus 55% for SPY. Annualized returns for periods of 2 to 14 years were all higher than 28%; they are listed in Table 2.
For an investment of $100,000 made on Jan-2-1999 and holding about 10 stocks at any time, there were over the last 14 years 1,609 realized trades, 1,069 of them winners and 540 losers, as shown below; the figures include transaction costs of $599,272.
Further details are given in the Appendix.
Also one can see from the general slope of the ratio graph at the bottom of the chart that Best10 always outperformed SPY from 1999 onwards, producing 46 times the value which one would have had from a 1999 investment in SPY.
Performance during recession periods
During the period from Jan-1999 to May-2013 there were two recessions. Returns can usually be increased, and drawdowns decreased, by placing the portfolio's funds, when not in the stock market, into treasury bond funds. My COMP-IBH system provides information when to exit the market and when to re-enter it, and the system's buy and sell signals are shown in Figure 1 as the vertical green and red lines, respectively.
This strategy would have increased performance considerably, because Treasuries gained greatly during the last recession. $100,000 would have grown to $12,776,122 (about 75 times more than SPY) from 1999 to May 2013, providing an annualized return of 40% and a maximum drawdown of 25%, all as shown in Figure 1.
However, it is noteworthy that during the years 2001 and 2002 Best10 did not follow the market lower, and performed as well as Treasuries over that time. Only during the great recession did Best10 lose value.
Annual performance from January to December ranged from a maximum of 140% for 2009 to a minimum of -37% for 2008, as can be seen in Figure 2. (This is the return if one had bought stocks on the first day of the year, then traded according to the model's signals, and sold all the holdings on the last day of the year.)
A Fund's performance is better measured by calculating the terminal values of hypothetical equal annually recurring investments, than by the average annual percentage returns for a few time-periods. This simulates returns from saving, which is a continuous process. It also highlights whether a fund has consistently performed well, or whether a good return measured from inception is perhaps only due to an initial period of exceptional returns, never to be repeated later.
Table 1 in the Appendix shows the futility of having invested in an S&P 500 index fund on an ongoing basis for the last 14 years. Assuming an investment of $1.00 at the end of May of every year, from 1999 to 2012, one would have invested a total of $14 cumulatively by the end. Summing the 14 terminal values, this strategy would have netted this dollar-per-year investor the sum of $22.03 by May 31, 2013, and accounting for inflation this would only have amounted to $17.70. Retirement accounts reflect this poor performance, and it is not surprising that the elderly have for some time now accounted for the highest growth rate in the labor force; obviously this is because they can't afford to retire.
Making the same $1.00 annual investments in a portfolio managed according to the Best10 trading system would have provided a terminal value of $235.23, as shown in Table 2 in the Appendix. That is about 11 times of what one would have had from an S&P 500 index fund. Also one can see that the lowest annualized returns over all the various time periods was 28.1% for the period 2008-13.
Here is a portfolio management system which does require frequent trading (about once a week), but which would have provided huge returns. There is no art of hindsight involved because only stocks from the S&P 500 are used, and a simple algorithm is consistently applied to the stock selection process. This is not a complex, high risk system which could easily break down. Additionally, selecting only from large and mid-cap stocks from the S&P 500 assures liquidity, even if the portfolio value is large.
More particulars, including a listing of all the transactions since 1999, can be found on our website's "Systems" page where one can also follow the portfolios' progress over time as new buy and sell signals emerge. The performance simulations were done on a web-based platform which allows back-testing to the beginning of 1999.
The composition of the S&P 500 differs over the years. The results shown assume that the pool of stocks to select from were the current stocks of the S&P 500. For example, since 2009 there were 29 stocks added and 29 removed. A back-test of this model using the composition of the S&P 500 of 2009 would have provided a CAGR of 32% instead of 35% since 1999. Similarly, if one used the S&P 500 of 1999 one would get a different return.
The intent of this analysis is not to provide an accurate number for the return, but to show that by holding a few of the highest ranked stocks of a pool of company one would always outperform a constant investment in all the companies by a wide margin.
Georg Vrba is a professional engineer who has been a consulting engineer for many years. In his opinion, mathematical models provide better guidance to market direction than financial "experts." He has developed financial models for the stock market, the bond market, yield curve, gold, silver and recession prediction, all published in Advisor Perspectives. The models are updated weekly at http://imarketsignals.com/.