Winning On Wall Street

It would be preferable if you could be invested in the market when it is a raging bull and to be sitting on the sidelines when the market has a bad case of the bears. This was part of the investment strategy of Martin Zweig, who in his 1986 book, Winning on Wall Street described a system for deciding when you should be fully invested and when not. Zweig advises that you should never hold stocks, even of the best companies, in a bear market, since even they can disappoint.

The question is does the Zweig system still work? The banking financial crisis of 2008 was certainly considered by most economists to be the most serious since the crisis that started the Great Depression. Although the system was written to determine the condition of the US stock market, it is still of interest to investors in other countries. Not only do most other stock markets in the developed world follow the US stock markets but the policies of other central banks such as the Bank of England are unlikely to be very different from that of the Federal Reserve as they will be facing the same conditions and will follow the same policies.


Zweig's Market Timing System

Martin Zweig used a market timing system based largely on the relationship between interest rates and share prices. A simplified version for use by individual investors was described in his book Winning on Wall Street.

Can You Time the Market?

Many successful professional investors do not believe market timing systems work and do not use them. There are extra costs involved and tax implications if you buy and sell your holdings frequently. Investment funds need to be in the market at all times and unlike individuals cannot in a short period sell a large part of their holdings and go into cash. There are extra costs involved and tax implications if you buy and sell your holdings frequently.

Martin Zweig ran an investment newsletter, the Zweig Forecast and a number of mutual and hedge funds which made him a multi-millionaire. He was an academic with a Ph.D. in finance and lectured at Iona College and Baruch College in the USA. Zweig had a disciplined number-crunching approach to market investments and invented the put-call ratio market-sentiment indicator.

Martin Zweig authored the book Winning on Wall Street in 1986 which set out a simplified version of his approach used in the Zweig Forecast that was designed for ordinary investors to use. Zweig’s system finds market conditions that statistically give a higher probability of gains. The model for finding favorable market periods is finally combined with a share picking system designed to find shares that are likely to rise in price. Risk minimization and loss limitation are crucial to his strategy.

Zweig appeared regularly on PBS television's Wall $treet Week with Louis Rukeyser to promote his investment activities. A visibly worried Zweig stated on October 16, 1987 during this live TV program, that "I haven’t been looking for a bear market per se. I've been, really, in my own mind, looking for a crash". He went on to say "I don’t look for a long bear market here; I only look for a brief decline, but a vicious one".

By that night, the market had already been reeling and to many investors it looked like a buying opportunity. Zweig's put-call ratio, a measure of bearish versus bullish bets indicative of the nerves of the crowd, showed signs of a potential panic. Zweig had used options to short the market and advised the readers of the Zweig Forecast to do the same. The Dow Jones Industrial Average plunged a record 22.6% in the next session, in what came to be called Black Monday.

Trying to Identify the Causes for Major Market Volatility Movements

Actually finding the cause of a major market movement is difficult. Jeremy J. Seigel Stocks for the Long Run (2008) says that in the vast majority of cases, major market movements are not accompanied by any news that explains why prices change. The record 22.6% one-day fall in the stock market on October 19, 1987 is not associated with any one readily identifiable news event.This suggests that you cannot identify the causes of most major market moves but Zweig gets around this problem with his model by simply using such major market movements as market turning points in the general trend of the market.

From 1885 -2008, the Dow Jones Industrial Average has had 126 days when it has changed by 5% or more. But of these, Seigel argues that only 30 major moves can be identified with a specific world political or economic event, such as wars, political changes, or governmental policy shifts. So less than one in four major market moves can be clearly linked to a specific world event. Different reasons for a market crash have been given by different market observers.

Seigel identifies monetary policy, as been the biggest single driver of these massive market outbreaks of euphoria or fear. Four out of the five largest moves in the stock market over the past century for which there is a clearly identifiable cause, can be directly associated with changes in monetary policy. War is usually the biggest market mover as it gives the greatest level of uncertainty. The market drop on September 17, 2001, was more than twice the 3.5% drop that occurred on the day following the attack on Pearl Harbor, and it was a greater one day decline than any decline during a period when the United States was officially at war.

The Theory

Zweig used a combination of economic monetary statistics together with large stock market swings that convert into points that indicate their influence and the direction they are taking the US stock market. Zweig examined the relationship between stock market action and just about every conceivable economic or market indicator. These studies showed Zweig that monetary conditions exert an enormous influence on stock prices. Interest rates and Federal Reserve policy is the dominant factor in determining the stock market’s direction and once established these trends lasted on average two to three years. A rising trend in interest rates is bearish for stocks and conversely a falling trend is bullish. The reason is that falling interest rates reduce competition on stocks from other investments, especially short-term instruments such as Treasury bills, certificates of deposit or money market funds. When interest rates fall then stocks and the dividends they give look more attractive to investors. It cost corporations less to borrow when interest rates fall and this reduces a major expense to them, especially for heavy borrowers such as airlines, public utilities or savings and loans. As expenses fall, profits rise and Wall Street reacts to future earnings going up, investors are willing to pay more for shares.

The Prime Rate Model

The first monetary factor that Zweig considered was the prime rate. This is the interest that banks charge major corporations that are unlikely to go bankrupt and are the banks best customers. The less-credit worthy is the borrower then the more the banks will charge above the prime rate. Zweig believed that changes in interest rates generally lead to changes in the stock market and as the prime rate moves a little behind other interest rates they often mark just that point when stocks finally respond to changes in the rate.

Zweig set 8% as a high interest and the buy signals are cut if the rate was below 8% and if the rate is higher on the second cut or a full 1% cut. The sell signal is any rise below 1% if the rate is 8% or higher. If it is less on the second rise or a full 1% increase. These signals when compared with the S&P 500 index shows a gain of 15.5% against a buy and hold strategy of just 7.9%. Between 1954 and 1995 the S&P 500 went up eighteen times out of twenty-two signals a success rate of 81%. On the sell signals it only gave 57% successful signals. Since 1995, the Prime Rate Indicator has not done so well with one success out of three buy signals and the two failures gave substantial losses. The sell signal was correct only once out of three with the two fails giving substantial gains. The problem with calculating the Prime Rate Indicator is that different sources gave different dates when the prime rate changed and even different rates. I used the Wall Street Journal as suggested by Zweig but still found some differences.

Fed Indicator

Federal Reserve SymbolThe Federal Reserve is the US central bank and was set up to insure the stability of the US banking system. The Fed's main role is to conduct monetary policy by influencing the monetary and credit conditions in the economy in pursuit of maximum employment stable prices and moderate long-term interest rates. The Fed's main policy instruments are setting the discount rate, the federal funds rate and reserve requirements. The discount rate is an interest rate a central bank charges depository institutions that borrow reserves from it. The federal funds rate is the interest rate at which private depository institutions (mostly banks) lend balances (federal funds) at the Federal Reserve to other depository institutions, usually overnight. It is the interest rate banks charge each other for loans. The reserve requirement (or required reserve ratio) is a bank regulation that sets the minimum reserves each bank must hold to customer deposits and notes. It can be used as a tool of monetary policy but as this would cause immediate liquidity problems for banks with low excess reserves this is not normally done.

Zweig’s studies showed that the Fed has the greatest impact on the stock market. Zweig suggests that as the discount rate and federal funds rate serve basically the same function it is only necessary to follow the discount rate. InvestStrat has calculated both and averaged them and this gives some differences to the signal and the more complex model tends to give a sell signal earlier. For the recent period when the new policy tool of Quantitative Easing has been used these changes have been calculated in the same way using the rounded average of the three measures. The Zweig model uses the target rate set by the Federal Open Market Committee rather than the actual effective rate. Zweig clearly meant for changes to reserve retirements to pick up on policy change announcements rather than actual changes in the same way as federal funds rate and discount rate. Zweig suggests that there are few changes to reserve requirements and says the last change from when he was writing in 1986 was during the autumn of 1981. However, reserve requirements have only occasionally in the past been used by the Fed in this way although until its policy revision at the end of 2011 had frequently tweaked them. Therefore only major changes that have altered required reserves by at least US$1 billion have been used to cause a change in the reserve requirements model which seems to fit Zweig’s intention.

The power of the Fed's policy tools has changed over time. Originally when the Fed created new bank reserves, the banks lent them out. These loans were spent, and the proceeds were deposited at other banks as new checking accounts.  However, in the 1970s this relationship started to end with the emergence of money market funds, which required no reserve requirements.  Reserve requirements in the early 1990s, were dropped to zero on savings deposits, CDs, and Eurocurrency deposits. Reserve requirements in the US now apply only to transactions deposits which are essentially checking accounts.  Commercial and consumer loans are normally no longer funded by such bank deposits but through by issuing large denomination CDs.  Canada abolished its reserve requirement in 1992. See this article titled Why the Federal Reserve is Irrelevant.

The rules for calculating the Fed Indicator are to grade each instrument and then combine the scores. An increase is bearish and a hike gives minus one point that wipes out any positive points. This remains for six months before its affect fades away. Moves by the Fed have a greater impact on stock prices than the negative effect created by tightening moves. An initial cut wipes out a rise in that component and gives two points. After six months one point is discarded and the remaining point fades away after a year has passed. Zweig tested the Fed Indicator from 1936, when reserve requirements were first introduced, until 1957. Once the rules were formulated he tested it from 1958 to 1996, the era when institutional trading came to dominate the stock market. The results for this later period did not differ significantly from the earlier days. The extremely bullish mode of the Fed Indicator showed a 22% annualized gain with the S&P 500. The neutral indicator gave a 4% a year gain and the moderately bearish indicator a loss of -0.3% a year. The extremely bearish rating lost 3.4% per annum. Zweig warns that monetary indicators have their greatest impact in bullish markets and only a moderate impact in bearish conditions.

How has the rating done more recently? Both the federal funds rate and discount rate gave the Fed Indicator a sell signal from December 1986 and this was increased in September 1987 long before the crash of October 19th, 1987. However, the signal for September 2008 was neutral before changing to extremely bullish in October 2008, as the Fed had cut interest rates in an attempt to avoid the financial crisis.

The long extremely bullish indication period from December 20, 1991 to May 31, 1992 gave a return of 7% on the S&P 500, whilst the bearish indication period between August 24, 1999 and October 31, 2000 gave a decline of -22%. The indicator does not always work so well as the January 31, 2001 – April 30, 2002 period which had an extremely strong bullish indication but gave a fall of -21% on the S&P 500. What the Fed Indicator sees as a largely bullish period between September 18, 2007 and August 31, 2009 returned a fall of -32.8%. The Fed Indicator by itself is not necessarily a good indicator of a change in market direction.

Installment Debt Indicator

The Installment Debt Indicator models changes in loan demand which has an important effect on interest rates. When demand for loans rises it puts upward pressure on rates and when it drops it puts pressure to lower rates. The main problem with the Indicator is that the Fed releases its figures of such debt called G19 some six weeks later. The most recent figures are projected. The percentage change in installment debt on a year-to-year basis is calculated. Zweig's studies led him to believe that when changes in the installment debt drops below 9% it is bullish and when above 9% it is giving a sell signal. Zweig tests shows that buy signals give annualized returns of 10.6% whilst sell moves give a gain of 5.4%. There has only been one long period when installment debt was above 9% which was between March 1994 and January 1997. Installment debt fell from May 2008, long before the financial crisis became evident and interestingly from March 2009 to March 2011 it was giving a negative figure.

The Monetary Model Indicator

The Monetary Model is simply calculated by adding up the scores from the Prime Rate, Fed and Installment Debt Indicators. It gives a score of between 0-8. A six triggers a buy signal which stays in effect until a two is reached which triggers a sell signal. Zweig had eleven buys and ten sells between 1954 and 1986. The buy period gave annualized gains of 15% whilst a buy and hold strategy would have given 6.5%, ignoring dividends. The point Zweig makes is that if the investor went into money market instruments in bearish periods he would make a substantially higher return. Monetary indicators generally have their greatest impact on the bullish side and only moderate impact on the bearish side.

The Four Per Cent Model Indicator

New York Stock Exchane buildingThe Four Per Cent Model Indicator acts as a momentum measurement. Zweig’s studies showed him that strong market bursts give high gains and he adopted this model by Ned Davis into his system. The model requires a wide representation of changes in the US stock market and uses the Value Line data. The weekly closing figures are plotted and the percentage change from the previous week is calculated. The triggers are 4% of change up or down in a week. The buy or sell signal continues until a change occurs. Zweig shows that between 1978 and 1996 there were 61 buy signals and of these 30 or 49% were profitable. However, these 30 profitable buys produced average profits of 14.1% per trade. The 30 losing trades lost only 3.5%. So the profits overwhelm the losses on the poor signals and gave an annualized gain of 16.2% from 1966 to 1996. Zweig warns that because only weekly closing prices are used there is the possibility of greater losses.

The Super Model

The Monetary Model and the Momentum Model are then combined into the final Super Model. The Monetary Model can run from 0 to 8 points and to this is added 0 points when the four per cent model is bearish or two points when the model is bullish. When the Super Model gives 6 points or more it is bullish and when it falls to 3 points or less it gives a sell signal. Zweig suggests that the investor could go fully invested above certain levels and three-quarters or half invested at other levels. The weekly data is converted into monthly to give a representation of what happened during a month from the weekly signals.

The Super Model has given fourteen sell signals and fourteen buy signals between 1966 and 1996 against the Standard & Poor 500 index, twelve of the buy signals were profitable and with the losers giving -5.3% in 1980-81 and -3.2% in 1989. Eight of the fourteen sell signals led to market declines, with two of the five failures showing only very small advances. The annualized return on the buy signals was 14.5% against 6.8% for a buy and hold strategy. The S&P 500 declined by 4.5% per year during sell periods. This means that an investor following the signals who invested US$10,000 would have increased it to US$490,919 an annualized gain of 13.9%. A buy and hold strategy on the S&P 500 would have turned that US$10,000 into US$222,816 a per annum return of only 10%. The model works even better if you could invest in a wider range of shares such as the Value Line Index and this would give around an 18% annualized return compared with just 5% for a buy and hold strategy.

The Super Model for the period between 1996 to 2013 gave five buy and four sell signals. Three of the buy signals were profitable and the losers gave -1.5% between June 2 - June 30, 2000 and -15% between November 3, 2000 - December 31, 2004. Three of the four bearish signals led to market declines with the failure giving a market advance of 4.8% for the period September 1, 1999 – March 10, 2000. The two long buy periods gave 80.6% for the start of the period in January 1997 – September 1, 1999 and 27.7% for December 31, 2004 to the end of the period in March 2013, despite the financial crisis.

Zweig’s Super Model is not an extrapolation of current factors into the future but tells the investor what the situation of the market is currently. This is useful after a change in a market factor such as a raise in interest rates changing a bull market into a bear market. This gives the investor a handle on what the future trend is likely to be and how much he should be invested in the market. As the Federal Reserve knows that uncertainly can cause major market volatility it likes to signal its policy changes long before they happen. The investor can test the likely policy change in the model and so make portfolio changes even before these policies are put into effect.


Zweig's Market Timing System Part 2