HIGH CASH
FLOW - RELATIVE STRENGTH PORTFOLIOS
Fidelity’s Peter Lynch has been quoted as saying that share price follows earnings. We agree with this statement, but we also believe that share price follows cash flow better. This presentation demonstrates that the use of cash flow as a selection criteria combined with the concept of relative strength leads to superior performance.
— Inflation Versus Disinflation
Theme investing is an approach that attempts to identify and profit from the onset of a longer-term economic or financial trend. For example, the inflationary environment of the 1970s caught Wall Street by surprise. Asset-rich stocks such as gold and energy shares outperformed, and value investing was the most profitable approach. This period ended in 1980, bringing in a period that favored growth analysis. This technique led many investors to place funds in technology stocks, thereby generating very high returns. Theme investing seeks to identify and capitalize upon a trend which will remain in force for some time. The inflationary psychology lasted for more than a decade; the disinflation theme is now in its twentieth year. Thus, the portfolio manager can let his profits run and not be concerned about high valuations or over-weighting of stocks in his portfolio, if the manager understands the implications of theme investing.
— Identifying Disinflation Hedges by Cash Flow Generation
The rationale behind this cash flow approach is that inflation hedges, companies with physical assets such as oil or real estate, were the leaders in the 1970's. With the onset of disinflation, it was reasonable to assume that stocks that are the reverse of companies with physical assets would be the leaders. Paper assets are the opposite of hard assets. So, high cash flow generators were designated to be this opposite group. Since each dollar of cash is more valuable (or is depreciating less rapidly) when the prices of physicals decrease, the cash generators will likely rise during disinflation.
Cashflow analysis has advantages over traditional earnings-per-share analysis. First, net income represents only about a third of gross cash flow. Second, earnings are subject to accounting distortions. And, research by Merrill Lynch’s quantitative unit demonstrates that the ability of analysts to predict earnings is limited. Their analysis reveals that the projection of continuing low growth (less than 5 percent) for slow-growing companies tends to be most accurate. These estimates were accurate over a five-year period about 75 percent of the time. The projection of continuing high growth (more than 20 percent) for the fastest-growing companies tends to be second most accurate. These estimates were accurate over 5 years about 40 percent of the time. In the first case, the stocks tend to under-perform. In the second, the stocks tend to outperform, but the stocks of those companies that fail to meet projections (60 percent) suffer severely. In fact, 24 percent of the estimates were dramatically inaccurate.
The point is that investors pay for real economic performance as measured by cashflow and not necessarily for accounting performance as measured by earnings per share. A unique cashflow measurement is utilized to screen companies. The list is then sorted to determine the top 100 companies on this basis. The accumulated cash will likely be put to one of these four uses by the companies:
The 100 highest cashflow generators were entered into a portfolio reporting system at the beginning of each year. At the end of the year, the portfolio was sold and re-invested on the first of the new year in the highest cashflow generators on that date. The performance versus 3 benchmarks was calculated and is summarized below. This test assumed that an equal dollar amount was placed in each stock from the beginning of the year to yearend.
| Year | Hi cash flow | S&P 500 | DJIA | NASDAQ INDUST. |
| 1985 | 52.3% | 26.3% | 27.7% | 26.6% |
| 1986 | 12.7 | 14.6 | 22.6 | 5.8 |
| 1987 | 3.8 | 2.0 | 2.3 | -3.0 |
| 1988 | 32.2 | 12.4 | 11.9 | 11.8 |
| 1989 | 46.7 | 27.3 | 27.0 | 18.2 |
| 1990 | 10.4 | -6.7 | -4.3 | -9.4 |
| 1991 | 87.0 | 26.3 | 20.3 | 64.8 |
| 1992 | 19.8 | 4.5 | 4.2 | 8.4 |
| 1993 | 20.3 | 7.06 | 13.64 | 11.16 |
| 1994 | 1.41 | -1.54 | 2.25 | -6.46 |
| 1995 | 29.98 | 34.11 | 33.45 | 27.97 |
| 1996 | 26.35 | 20.26 | 26.01 | 15.03 |
| 1997 | 29.48 | 31.01 | 22.64 | 10.04 |
| 1998 | 25.74 | 26.67 | 16.10 | 39.63 |
| 1999 | 25.32 | 19.53 | 25.22 | 71.67 |
| 2000 | 8.69 | -10.14 | -6.18 | -33.76 |
| 2001 | 19.41 | -13.04 | -7.10 | -21.05 |
| 2002 | -13.96 | -23.37 | -16.76 | -31.53 |
| 2003 | 58.1 | 26.4 | 25.3 | 55.7 |
| 2004 | 19.0 | 3.2 | 9.0 | 8.6 |
— Comparison of Buy and Hold Strategies:
$1,000,000 invested in each average in January 1985 would have grown to these amounts at December 31's close:
|
The cashflow portfolio outperformed the S&P in 16 out of the 20 years under review.
— Practical Considerations
There are several practical rules that I have developed for the actual management of such a fund due to the nature of the companies selected:
1. There are many building stocks on the list because they pay no dividends and have no plant spending. Because these stocks are so interest-sensitive, the stocks will not be purchased if the interest-rate environment is unfavorable.
2. There are many technology stocks. Due to the short product life cycles, these stocks move very quickly. One must use some timing to get in and out early. Buy only when share price momentum comes out of a base and accelerates.
3. There are many issues like Playboy Enterprises that have good cash flows, but no revenue growth. Investors generally are not going to pay up for these stocks unless the company shows sales growth. Practically speaking, these stocks are not buy candidates.
— Knowing the Strengths and the Weaknesses
As with all strategies, this one has its Achilles heel, for which I do have remedies. A return of an inflationary environment will likely cause this group to underperform. This occurred in the first quarter of 1999. The solution is to monitor upticks in interest rates, commodities, and other inflation hedges as 'early warning signals' to consider reducing exposure to these stocks. The next section features a useful tool for such changes. It is especially important to watch the relative strength of the inflation hedge groups to detect a change in group leadership.
The second potential weakness is that high cashflow generators will likely suffer more during market corrections than will the average stock. The reason lies in the fact that these stocks run up during bullish phases. When the corrections hit, stocks that have run up sharply can pull back dramatically. The solution is to employ market timing. If such a pullback can be anticipated, the manager can raise cash beforehand. The next section concerns a cyclical approach to market timing.
— Improving the Results With a Relative Strength Screen
Richard Bernstein, Merrill Lynch’s Chief Strategist rated over forty stock selection techniques from 1987 through 2004. This test was conducted by selecting the top 50 stocks in each category on a monthly basis. For example, the top 50 relative strength stocks were chosen, held for one month, and then sold.The ratings of the methods are listed in the table below:
|
Technique |
17-Year Compound Annual Return |
|
Most Active |
+17.0% |
|
High Relative Strength |
+16.7% |
|
Low PEG Ratio |
+16.3% |
|
Low Price/Free Cashflow |
+15.7% |
|
S&P 500 |
+9.4% |
Only one stock selection methodology, Most Active, exceeded the return generated by high relative strength by returning +17.0 percent. Most active stocks are defined as stocks that have the highest monthly trading volume. This measure is highly inconsistent, and pushed its way into first place only because of the volatility and trading activity during and after the 2000 bubble. We do not consider this selection method as being sound nor does it generate consistent results, and thereby consider relative strength as being the number one methodology. Also note that low price/free cashflow ranked highly. Although this is not the cash flow method that we use, it is reassuring to see its high ranking.
The thesis that stock analysis and portfolio analysis can be improved with relative strength is supported by research. Professors Philip Brown and Ray Ball (Journal of Accounting Research, August 1968) found that when a company’s reported earnings were better than forecasted, the share price rose in the 12 months prior to the announcement and tended to fall in the 6 months after the announcement. Manown Kisor and Van Messner studied trends in earnings in the Financial Analysts Journal in January of 1969. They did not find predictability in earnings trends. Instead, they found that stocks that advance in price substantially ahead of the market for 6 months tend to have above average appreciation potential in the following 6 months. Thus, share price appreciation tends to precede favorable earnings, and stocks that perform well over 6 months tend to do so in the ensuing 6 months.
In order to improve returns, a screen was employed beginning in 1996. The group was screened for relative strength at the beginning of each quarter from Q1 1996 to the present. The proprietary methodology eliminates any stock that gave negative relative strength signals for the previous 6 months. (Previous tests reveal that there is little difference between use of a 6-or 12-month time horizon. This process usually eliminates about 55% of the stocks being screened. On average, it eliminated 51% of the stocks during the study period.) If the relative strength measure of the stock was higher at the beginning of the quarter than it was 6 months earlier, then it was included in the portfolio.
The purpose of this screen is to eliminate stocks that either have temporarily run their course, or to drop stocks that are simply beginning to fail. In the first situation, the positive fundamentals have been recognized and exploited. The stock is beginning a period of consolidation that will not contribute relative strength to the portfolio. In the second case, the fundamentals have deteriorated and informed investors are selling the stock. Note that stocks frequently give signals in relative terms before they do so in absolute terms.
— Portfolio Results
The attached sheet shows the results. It was assumed that each stock was equally weighted in the portfolio. The screened group outperformed the S&P 500 in 29 of 37 quarters. The first three columns depict the returns.
Because the list contains stocks from the S&P, NASDAQ, and the AMEX, a second composite benchmark was created. This was constructed in the following way. If 50% of the stocks selected were NYSE issues, the S&P return was multiplied by 0.50. If 40% of the stocks were from the NASDAQ, then 40% of the NASDAQ's return for the quarter was calculated. Assuming that the remaining stocks were AMEX shares, the same proportion of that index would be determined. These amounts were then added to create a composite return. This number and the difference from the portfolio are listed in columns 5 and 6. The portfolio outperformed this composite in 30 of 37 quarters.
Looking at the returns on an annualized basis from January 1, 1996 through March 31, 2005:
S&P Annual Return: 7.5%
Portfolio Annual Return: 28.5%
Enhancing These Returns
Accurate overall market timing can also enhance these returns. This group of stocks will usually come down more than the overall market during corrections. The attached quarterly returns show that the margin of outperformance is usually greatest in the second and the fourth quarters, and it is least in the first and the third quarters. Market timing can be utilized to reduce exposure at key market junctures.
1996-2004 QUARTERLY RETURNS OF CASH FLOW/RELATIVE STRENGTH PORTFOLIO — Hit Ratio
| Portfolio % | S&P % | Difference | Composite % | Difference | No.Stocks | Hit Ratio-S&P | Comp. | |
| Q1 1996 | 6.76 | 4.8 | 1.96 | 4.91 | 1.85 | 49 | 53% | 53% |
| Q2 1996 | 10.56 | 3.89 | 6.67 | 5.52 | 5.04 | 48 | 63% | 58% |
| Q3 1996 | 11.06 | 2.49 | 8.57 | 1.16 | 9.90 | 47 | 43% | 72% |
| Q4 1996 | 8.88 | 7.77 | 1.11 | 3.97 | 4.91 | 42 | 48% | 57% |
| Q1 1997 | 1.27 | 2.21 | -0.94 | -3.08 | 4.35 | 38 | 50% | 66% |
| Q2 1997 | 18.21 | 16.91 | 1.3 | 16.07 | 2.14 | 32 | 44% | 44% |
| Q3 1997 | 21.19 | 7.02 | 14.17 | 11.31 | 9.88 | 27 | 74% | 56% |
| Q4 1997 | 5.88 | 2.44 | 3.44 | -2.75 | 8.63 | 47 | 62% | 72% |
| Q1 1998 | 18.71 | 13.53 | 5.18 | 12.39 | 6.32 | 50 | 60% | 62% |
| Q2 1998 | 7.94 | 2.91 | 5.03 | 0.43 | 7.51 | 39 | 62% | 64% |
| Q3 1998 | -13.22 | -10.3 | -2.92 | -15.24 | 2.02 | 38 | 42% | 45% |
| Q4 1998 | 22.05 | 20.86 | 1.19 | 22.49 | -0.44 | 22 | 50% | 41% |
| Q1 1999 | 2.12 | 4.65 | -2.53 | 5.65 | -3.53 | 32 | 53% | 39% |
| Q2 1999 | 12.1 | 6.71 | 5.39 | 12.34 | -0.24 | 35 | 60% | 49% |
| Q3 1999 | 2.06 | -6.56 | 8.62 | 3.87 | -1.81 | 35 | 69% | 49% |
| Q4 1999 | 31.46 | 14.54 | 16.92 | 28.32 | 3.14 | 39 | 54% | 54% |
| Q1 2000 | 14.52 | 2 | 12.52 | 13.07 | 1.45 | 42 | 62% | 45% |
| Q2 2000 | -0.33 | -2.93 | 2.6 | -9.04 | 8.71 | 44 | 55% | 66% |
| Q3 2000 | 2.6 | -1.24 | 3.84 | -1.9 | 4.50 | 49 | 51% | 57% |
| Q4 2000 | 13.33 | -8.09 | 21.42 | -14.8 | 28.13 | 43 | 79% | 84% |
| Q1 2001 | -5.35 | -12.11 | 6.76 | -13.4 | 8.05 | 39 | 74% | 74% |
| Q2 2001 | 12.98 | 5.52 | 7.46 | 11.48 | 1.50 | 53 | 60% | 53% |
| Q3 2001 | -17.91 | -14.99 | -2.92 | -20.3 | 2.39 | 64 | 52% | 56% |
| Q4 2001 | 23.46 | 10.29 | 13.17 | 17.47 | 5.99 | 50 | 50% | 54% |
| Q1 2002 | 4.22 | -0.06 | 4.28 | 0.59 | 3.63 | 62 | 66% | 65% |
| Q2 2002 | -7.33 | -13.73 | 6.4 | -12.71 | 5.38 | 63 | 60% | 56% |
| Q3 2002 | -13.92 | -17.63 | 3.71 | -19 | 5.08 | 54 | 44% | 54% |
| Q4 2002 | 3.49 | 7.92 | -4.43 | 8.7 | -5.21 | 47 | 38% | 32% |
| Q1 2003 | -2.74 | -3.6 | 0.86 | -2.1 | -0.64 | 54 | 52% | 52% |
| Q2 2003 | 24.2 | 14.9 | 9.3 | 18.5 | 5.70 | 51 | 63% | 57% |
| Q3 2003 | 11.9 | 2.2 | 9.7 | 5.9 | 6.00 | 61 | 62% | 51% |
| Q4 2003 | 16.9 | 11.6 | 5.3 | 13.4 | 3.50 | 80 | 60% | 53% |
| Q1 2004 |
8.9 |
1.29 | 7.61 | 2.62 | 6.28 | 71 | 59% | 55% |
| Q2 2004 |
0.64 |
1.3 | -0.66 | 2.05 | -1.41 | 68 | 50% | 47% |
| Q3 2004 | -5.16 | -2.3 | -2.86 | -4.22 |
0.94 |
55 | 49% |
51% |
| Q4 2004 | 13.25 | 8.73 | 4.52 | 11.54 |
1.71 |
39 |
56% |
46% |
| Q1 2005 | -3.58 | -2.59 | -0.99 | -4.48 | 0.90 |
56 |
45% | 54% |
Annualized Return=28.5%
S&P Return=7.5%
|
DIFFERENCE VERSUS S&P BY QUARTER |
|||
|
Q1 |
34.71 |
10 |
3.47 |
|
Q2 |
36.03 |
9 |
4.00 |
|
Q3 |
39.91 |
9 |
4.43 |
|
Q4 |
62.64 |
9 |
6.96 |
Updated 08/21//2005 dfc