by Martin Zweig
The right-hand column of table 5 shows the returns on the S&P 500 during each “extremely bullish” period. The S&P appreciated 8% or more in thirteen of the seventeen spans, while only two endured moderate losses.
TABLE 5
PERFORMANCE OF FED INDICATOR WHEN RATED EXTREMELY
BULLISH:
January 24, 1958 to March 20,1996
Extremely Bullish Period
Return on ZYPI
Return on S&P 500
1/2/58 to 10/17/58
+40.6%
+27.6%
6/10/60 to 6/10/61
+20.5%
+14.1%
4/7/67 to 10/7/67
+16.1%
+9.1%
8/30/68 to 12/18/68
+12.0%
+8.2%
11/3/70 to 7/16/71
+22.9%
+17.7%
11/19/71 to 6/17/72
+12.7%
+18.0%
11/28/74 to 12/6/75
+26.5%
+24.4%
12/24/75 to 6/24/76
+26.2%
+16.0%
12/17/76 to 5/19/77
+5.2%
-4.2%
5/22/80 to 11/14/80
+24.8%
+25.8%
9/21/81 to 6/4/82
-1.3%
-6.1%
7/19/82 to 5/19/83
+64.0%
+46.3%
11/21/84 to 11/17/85
+21.0%
+20.4%
4/18/86 to 1/10/87
+.3%
+6.7%
12/4/90 to 6/21/92
+36.9%
+23.6%
7/2/92 to 8/17/92
+1.4%
+2.3%
7/6/95 to present
+13.2%
+17.3%
Total years buKsh = 12
$10,000 investment = $190,906 $107,262
Annualized return = +28.0% +22.0%
The returns in the “extremely bullish” modes are so good that it would pay for a patient and risk-averse investor to stay completely out of the stock market at any time the Fed Indicator rated anything less. Since 1958 the “extremely bullish” zone was in effect for a cumulative total of just 12 years. Suppose one had invested in the broad market (ZUPI) only in those years and then gone into cash equivalents in the remaining 26.2 years at an average yield of 6%. Assuming no taxes or dividends, a $10,000 investment would have grown to $887,824, an annualized gain of 12.5%. That would have clobbered the buy-and-hold return of 6.4% and its final value of $99,382. Moreover, the fortunate investor would have had no money at risk 69% of the time. A conservative investor can’t ask for much better than that, even though in some periods, such as 1962-65 and in 1978, he would have watched from the sidelines as the market moved upward. Of course, during the grim days of 1969–70 or 1973–74 he would have slept comfortably while stocks were being ravaged.
Clearly, investment strategy should never be determined merely on the basis of one indicator. But the results found here strongly argue against “fighting the Fed.”
INSTALLMENT DEBT INDICATOR
Loan demand has an important effect on interest rates. When demand for loans rises excessively, it puts upward pressure on rates. When it drops dramatically, it works to lower interest rates.
There are several major sources of loan demand, including federal, state, and local government borrowings; corporate borrowings both in the short-term money markets (commercial paper and bank loans) and in the longer-term bond markets; mortgage debt; and consumer installment debt. The latter figure has maintained one of the best records at calling the shots for the stock market. Also, since it is reported only once a month, it’s a very simple tool to use. So, let’s try it.
I personally use a rather complicated approach in dealing with the consumer installment debt numbers. But the idea of this book is to make it easy for the “weekend investor” to make decisions. Moreover, even a very simplified model using installment debt works very well.
First, the monthly total of such debt is released by the Federal Reserve around midmonth for the month ended about six weeks earlier. In other words, the data for, say, September would come out around November 15 or so. The delay in getting the numbers is not that important since our concern here is only with the major trend, one that changes very slowly. The installment figures are reported in most major newspapers, including The New York Times and The Wall Street Journal. You can also be placed on the mailing list for the release itself by writing to the Federal Reserve Board in Washington, D.C. Ask for Federal Reserve Statistical Release G.19.
The figures are reported both on a seasonally adjusted basis and a non-seasonally adjusted basis. Use the latter … the non-seasonally adjusted number.
GRAPH G
Ned Davis Research
Take that total for the month and divide by the total for the same month a year ago. Then subtract 1.000. That leaves you with the percentage change in installment debt on a year-to-year basis. When done this way, you don’t need the seasonal adjustment since you are comparing January to January, February to February, etc.
Let’s try an example. Take out your calculator and work along with me. Suppose on November 15, 1984, you get a release showing that non-seasonally adjusted consumer installment debt was $450,131 billion at the end of September 1984. The G.19 release will also show the non-seasonally adjusted debt for September 1983, which we’ll assume was $375,246 billion. Now, divide the former by the latter and you get 1.200 (rounded upward). Subtract 1.000 and you are left with .200, which when converted from decimals equals +20.0%. In other words, in the one year ending September 1984, consumer installment debt rose by 20.0%.
The year-to-year percentage change in installment debt is the only calculation you have to make on this indicator. It will take you a few seconds a month.
Graph G (pp. 62–63) shows installment debt plotted on the year-to-year basis just described. Above it is the Standard & Poor’s 500 Index. It is apparent that an expansion in installment debt tends to be bearish, as it was in late 1968, 1972, and late 1976, and moderately negative from mid-1983 to mid-1984. Conversely, when the trend of such debt plunges, it’s bullish for stocks, as it was in late 1966, 1970, late 1974, and 1980.
The important question is, just how much of a year-to-year change in installment debt is needed to signify a bullish or a bearish condition for stocks? It appears that 9% is the key level. At the least, the 9% mark offers an easy method for generating good signals.
Rules:
A buy signal is given when the year-to-year change in installment debt has been falling and drops to under 9%. A sell signal comes when the year-to-year change has been rising and hits 9% or more. That’s it. Example: Table 6 shows a three-year history of consumer installment debt from 1974 to 1976. Column 1 gives the actual month of the data. Column 2 shows the approximate day you would have gotten the figures from the government. Recall, there is about a six-week lag. Column 3 provides the monthly total of installment debt (in billions of dollars). Column 4 shows the year-to-year change.
The year-to-year change was falling in 1974, finally breaking under 9% in October when it struck 8.2%. You would have gotten that information about 6 weeks later, on December 13, the date of the buy signal. Afterward the series kept falling until mid-1975, when it bottomed at 1.6%. From there it rose again until the key 9% level was topped in June 1976 at a reading of 9.2%, That tripped a sell signal effective six weeks later, on August 13.
Table 7 shows the performance of the Installment Debt Indicator from 1951 to 1996 vs. the Zweig Unweighted Price Index. It has given only twelve buy signals and eleven sell signals. A $10,000 investment in the ZUPI during only the buy periods would have grown to $107,046, an annualized return of 12.1%. Buy-and-hold over that period returned only 6.2% a year. In the sell modes a $10,000 investment would have shown little change with an annualized gain of 1.2%.
Note that not all the signals were good
ones. The worst was the sell in October 1983, although stocks went lower for the next ten months. However, the indicator continued in its negative mode until March 1987—during which period prices rose sharply. Stocks then rose for three years before hitting another bear market, during 1990. Nonetheless, the Installment Debt Indicator stayed correctly bearish during just about all of the two worst bear markets since the Depression, 1969–70 and 1973–74. It also caught most of the major bull advances of the past few decades.
Table 8 shows how the Installment Debt Indicator performed when tested against the Standard & Poor’s 500 Index. The buy modes produced annualized gains of 10.6%, well above the buy-and-hold return of 7.8%. The sell modes showed a small annualized gain of 5.4% … but that is still substantially inferior to the buy-and-hold results. One would have been much better off, and at much less risk, staying out of the market during the sell periods and keeping the money in Treasury bills or the like.
TABLE 6
CALCULATING THE INSTALLMENT DEBT INDICATOR
Month of Data
Date Data Received
Consumer Installment Debt ($billions)
Year-to-year Change in Consumer Installment Debt
1974
January
3/18/74
$145.55
+14.9%
February
4/18/74
145.29
+14.0%
March
5/15/74
145.02
+12.6%
April
6/14/74
146.27
+12.2%
May
7/16/74
148.13
+11.5%
June
8/15/74
149.91
+10.7%
July
9/13/74
151.36
+10.1%
August
10/15/74
153.71
+9.9%
September
11/14/74
154.47
+9.4%
October
12/13/74
154.51
BUY+8.2%
November
1/17/75
154.36
+6.9%
December
2/14/75
155.38
+6.1%
1975
January
3/18/75
153.36
+5.4%
February
4/18/75
152.40
+4.9%
March
5/15/75
151.10
+4.2%
April
6/13/75
151.12
+3.3%
May
7/16/75
151.41
+2.2%
June
8/14/75
152.64
+1,8%
July
9/15/75
154.52
+2.1%
August
10/15/75
156.20
+1.6%
September
11/13/75
157.45
+1.9%
October
12/15/75
158.19
+2.5%
November
1/19/76
159.22
+3.1%
December
2/13/76
162.24
+4.4%
1976
January
3/17/76
160.82
+4.9%
February
4/15/76
160.40
+5.2%
March
5/13/76
160.73
+6.4%
April
6/16/76
162.33
+7.4%
May
7/16/76
164.10
+8.4%
June
8/13/76
166.66
SELL+9.2%
July
9/15/76
168.67
+9.2%
August
10/14/76
171.16
+9.6%
September
11/16/76
172.92
+9.8%
October
12/16/76
173.93
+10.0%
November
1/17/77
175.33
+10.1%
December
2/14/77
178.78
+10.2%
TABLE 7
INSTALLMENT DEBT INDICATOR
VS. ZWEIG UNWEIGHTED PRICE INDEX: 1951 TO 1996
BUY SIGNALS SELL SIGNALS
Date
ZUPI
%Change
Date
ZUPI
% Change
9/14/51
33.60
-4.1
7/10/52
32.21
+11.0
6/16/54
35.75
+41.5
6/16/55
50.58
+8.0
4/17/57
54.64
+30.5
9/15/59
71.31
+18.5
3/17/61
84.47
-16.3
10/15/62
70.70
+53.8
11/16/66
108.75
+59.2
12/31/68
173.15
-44.1
5/14/70
96.78
+34.2
3/16/72
130.26
-63.5
12/13/74
47.49
+79.7
8/13/75
85.36
+37.6
6/16/80
117.46
+78.0
10/13/83
209.04
+54.2
3/20/87
322.33
-21.5
4/15/88
252.98
+4.4
10/21/88
264.07
+2.9
3/20/89
271.78
+6.0
11/15/89
288.22
-1.9
12/18/89
282.62
+0.2
1/18/90
283.08
+33.1
7/12/94
376.70
+27.7
3/20/96 *
$10,000 becomes: $107,046 $13,376
Annuafized return = +121% +1.2%
Buy-and-hold return = +6.2% per year
MONETARY MODEL
Thus far I’ve taken you through the simple calculations on three important monetary indicators—Prime Rate, the Fed, and Installment Debt. The next step is to combine them into a model. A model may sound like some sort of fancy mathematical word. Don’t let that bother you. All model means in this case is that we give each of our indicators a numerical score, then combine them to get a composite reading on monetary conditions. Once we do that, we’ll develop rules to make buy and sell decisions.
The First indicator we developed is the Prime Rate Indicator. When the prime rate gives a buy signal according to our rules (see pages 43–46), give it 2 model points. When it gives a sell signal, accord it zero points. Table 9, on page 71, is a worksheet that shows the grading for the prime rate and the other indicators from the end of 1979 to the end of 1988. Each time any one of the three indicators changed, an entry was made on the worksheet.
Now, look at the Prime Rate column. As of December 31, 1979 (our starting point for the worksheet), the prime rate was on a buy signal (which had been given on December 7, 1979). Thus, it carried a score of 2. On February 19, 1980, the prime rate went bearish and the score for it fell to zero. Then on May 1, 1980, the indicator gave a buy signal and our rating went back up to
2. If you note the buy and sell signals for the prime rate that were presented in table 1, you’ll see that each is recognized and given a score of either 2 or zero in the worksheet in table 9.
TABLE 8
INSTALLMENT DEBT INDICATOR VS. STANDARD & POOR’S 500 INDEX: 1951 to 1996
BUY SIGNALS SELL SIGNALS
Date
S&P
% Change
Date
SAP
% Change
9/14/51
2169
+4.7
7/10/52
24.81
+17.0
6/16/54
29.04
+37.6
6/16/55
39.96
+12.8
4/17/57
45.08
+25.7
9/15/59
56.68
+14.0
3/17/61
64.60
-11.3
10/15/62
57.27
+43.8
11/16/66
82.37
+26.1
12/31/68
103.86
-27.4
5/14/70
75.44
