The Mean Reversion Way of Making Money
(October 16, 2005)
Dear Subscribers and Readers,
We switched from a neutral position to a 25% short position in our DJIA Timing System on the morning of July 14th at DJIA 10,616. As of Friday at the close, the Dow Industrials stood at 10,287.34 - giving us a still respectable gain of 328.66 points (although with the market now so oversold, anything can happen).
In last week's commentary, I stated that despite our short-term bearishness, not every sector is created equal - and that has been especially true over the last two years. Furthermore, we argued that financials, commodities, and the homebuilders were still the most vulnerable, and that investors should continue to avoid those sectors despite the fact that they were getting increasingly more oversold by the day. Despite the action of last week, this author still believes that this is the case.
In this week's commentary, I would like to return to the overbought/oversold theme, and specifically, the fact that this author would only commit on the long side (in this kind of trendless market) provided that it only becomes sufficiently oversold. Readers who want to review our oversold/overbought indicators can review last week's commentary as well as our Thursday morning's commentary.
Okay, I bet some of our readers are now saying: Not again! We are getting bored of your overbought/oversold technical indicators. Well, I am going to expand my thoughts - and at the same time, help our readers (you) expand yours. Specifically, I want to discuss the three "tried-and-true" ways of making money in investing and also what this author prefers - not because it is an "easier" way to make money in the markets, per se, but because it suits my own preferences and psychology. Over the long-run, each investor will need to "kill his or her own snakes" and come to understand his own psychology and find one's best style - an investing or trading style that is unique to himself. Finally, I will try to weave this altogether by stating how the "oversold/overbought" theme fits together with this author's preferred investing style.
As I have mentioned previously in our commentaries and in our MarketThoughts discussion forum, there are essentially three distinct strategies that investors can make money in the financial markets over the long-run. This has been illustrated in a previous post of mine in our discussion forum. I am going to quote this post of mine word-for-word:
Throughout financial history, there has been three effective ways to make money. They are, not in particular order:
- The carry trade. This is where one borrows and leverages in a market and buys in another. An arbitrage is a carry trade, for example. This carry trade works very well at the beginning of a theme - that is, until everyone and his neighbor has adopted it. The yen carry trade that blew up in 1998 was one great example of the end of a carry trade.
- Momentum "investing." Everyone should be familar with this trade, as I am sure virtually all our readers have gone through the 1990s technology bubble.
- Mean reversion trade. This is where one buys an undervalued stock or commodity and wait for it to return to its true value. Warren Buffett is known for adopting this strategy to perfection (although he has done other trades at various points in his career).
Out of the three effective ways to make money that I have just mentioned, this author prefers the third way. Momentum trading causes me to lose sleep, and the carry trade just does not appeal to me. In this current market, the next trade is to buy a hugely oversold situation, and this author will not hesitate to jump in on the long side when the right time comes along.
Over the long-run, betting on mean reversion in an asset class or a commodity is a sure-way to make outsized returns, provided one is not leveraged and one does not grow to be too big and illiquid (similar to what happened with Long-Term Capital Management). Please note that when I mention "mean reversion trades" I am talking about longer-term trades that can last for weeks or months, such as betting on the valuation cycle of an asset class or the business cycle. Betting on cycles is an interesting thing, since at some point, there will always be a reversion to the mean. The trick is to remain solvent while one is waiting for such a turn. Particular (and recent) examples of such trades include buying crude oil and energy stocks in the Fall of 1998, selling technology stocks in spring 2000, and buying gold and unhedged gold-mining stocks in the Fall of 2000.
As I mentioned above, this author prefers the third way - the mean reversion trade - of making money, primarily because it suits this author's style but also because of the inherent dangers in the "carry trade" and "momentum investing." Let me elaborate. First of all, a carry trade involves selling short one financial instrument (that is very liquid) and buying another financial instrument - with the condition that the instrument you borrow costs less than the returns you will reap from the instrument that you buy. The yen carry trade that lasted from 1995 to 1998 was one such example. Speculators borrowed yen at rock-bottom interest rates and bought US dollar-denominated assets. One could've put the money in a bank account and get very decent returns - with the additional "oomph" provided by a rising U.S. dollar during that period. Such a trade, however, requires a significant amount of leverage to make decent returns and ultimately results in a lot of tears for a particular party - that is, the party that fails to find a chair at the end of the "musical chairs" game. Julian Robertson of the Tiger fund was just such a casualty in the end of the yen carry trade. In a particular bad day for him on October 7, 1998, the Tiger Fund lost a whopping $2 billion - a day when the yen gained more than 8% against the U.S. dollar during that day. Second of all - with the assets of the hedge fund industry now approaching $1.2 trillion, it is just much more difficult trying to find such trades near the beginning of the trend. The hedge funds have cheap access to capital, as well as dozens of analysts and a significant amount of computer power to execute these trades. The retail or part-time investor just does not stand a chance.
As for the momentum trade, one just needs to be nimble and quick on his feet in order to successfully execute such a trade from start to finish. One small mistake or mistiming, and this kind of trade can blow up in your face very quickly. Moreover, a typical momentum trade is usually accompanied by bravado and a "concept story" - and for the informed investor like most of our readers, it is just psychologically difficult to hold on to these positions primarily because they are so speculative. Finally, this author is intellectually curious and usually likes to study whole industries and countries before I would make a trade - and such intense analysis does not add any value to an investor that focuses on momentum investing. That is - at the end of the day, even if this author makes money, momentum trading does not add any value to my repertoire of knowledge, and thus, momentum investing (even if successful) does not satisfy me.
But Henry, you just mentioned that the investing world is now much more efficient. How could we expect to make outsized returns going forward, and how could we expect to find mean reversion trades, if that is the case? That is a very valid statement. Even as many formerly obscure markets are now available for the American public to invest in, more investment vehicles and capital have sprung up to take advantage of those formerly obscure opportunities. The world of finance is continuing to become a more efficient environment - with the hedge funds replacing the investment bankers and the mutual fund managers as the new "masters of the universe," so to speak. If one looks at the history of finance - starting with the Medici family in Italy in the 13th century, progressing through to the Rothschilds, the Warburgs, the Barings, the Morgans, the Jardine Flemings in Asia and then the investment bankers and the leveraged buyout funds of the 1980s and the VCs of the 1990s - and the recent changes in the marketplace, it is not too difficult to see. Today, the hedge fund industry has assets of $1.2 trillion, is attracting the top talent and is still growing exponentially. This is projected to grow to $2 trillion by 2008. With leverage of 4x and a reasonable assumption of 4x turnover of capital within a year, the amount of "efficiency capital" available to hedge funds in a given year will be somewhere around $32 trillion in a few years time - or 3x the U.S. GDP. This proliferation of the most flexible investment vehicle (with 10,000 different players) that men has ever known will ensure near-efficiency going forward.
Given that the hedge funds are much more flexible and more rigorous with their strategies and research than the investment bankers and VCs ever were, I wouldn't be surprised to see in a few years time, for example, a bunch of hedge funds taking control of General Motors, spinning off GMAC and then chopping GM into different pieces and then selling off those pieces to various (foreign) auto companies. This is very similar to the dilemma facing Volkswagen right now. Other recent hedge fund "meddlings" include their roles in the Verizon-MCI merger, the lobbying of Phelps Dodge's management to buy back their own shares, and the ouster of the Deutsche Bourse's Chief Executive when a London based hedge fund blocked his proposal to acquire the London Stock Exchange earlier this year. Going forward, hedge funds will be a significant source of capital - not unsimilar to the powers that were available to the Morgans, the Rothschilds, and the Barings in the 19th century.
But I digress. Even though the world's financial markets have gotten more efficient as time goes on, mean reversion trades have still invariably "popped up" every now and then. They come into being when three forces come together:
- Group think and a herd mentality: The universal belief that an asset class has just got to go up or down forever - resulting in a huge overvaluation (relative to historical average) of the current asset class in favor. The most recent example has been the bottoming of the U.S. Dollar in January this year. Not surprisingly, the U.S. dollar bottomed at about the same time a front-page bearish article on the U.S. dollar appeared on Newsweek.
- The ability for retail investors to participate in such a trade - and the widespread publicity this trade has "enjoyed" from the mass media. The U.S. dollar trade again fitted this bill.
- The Federal Reserve stepping in and starts to fight the current prevalent trade idea, although in longer-term trades, they are typically early. Again, coming back to the U.S. dollar example: By June 2004, the Fed has started hiking its Fed Funds rate once again - which is usually a precursor to a rising dollar. The opposite was true before the technology bubble popped in March 2000: The Fed has already started its hikes a year prior.
Going forward, there is no doubt that we will find more of these trades going forward. Even as the world has become a more efficient place, the list of trades runs long: Buying energy and semiconductor stocks in Fall 1998, selling (or shorting) technology stocks in March 2000, buying unhedged gold stocks and selling the U.S. dollar in late 2000, buying utility and technology stocks in late 2002, and buying the U.S. Dollar Index in December of last year. Today, if anything is coming close to a mean reversion trade, it is selling commodities (especially copper and crude oil), homebuilding stocks (although this asset class is now oversold on a short-term basis), and various emerging market equities. Being successful in executing a mean reversion trade involves a contrarian mentality and the ability to think independently. To avoid being "killed" if one turns out to be incorrect, one always should be open to all opinions and to posssess the ability to filter those opinions appropriately. Denial has no place in investing. For example, denial ran rampant even in the couple of days prior to the Northwest and Delphi bankruptcies - and all those long investors were burned thinking going long in those two stocks may be potential mean reversion "plays." Even Steve Cohen of SAC Capital was not immune; although my bet is that he minimized his losses as quickly as he could (SAC capital "got its revenge" by making at least $15 million on the latest Refco short trade).
This author has always believed the most money is made on the long side - and so the next trade would be going long the U.S. stock market in a substantial way. Hence the obsession with watching our overbought/oversold indicators as well as other signs of a major rally going forward. Even though many longer-term pieces of our puzzle (such as Consumer Confidence, a Lowry's 90% downside day, etc.) are failing into place, it is not obvious to me that a reversal and subsequent rally is imminent in the near future. As many of our subscribers should know, going long - even if you are only a mere few months before a major bottom - can be very painful, indeed. Of course, this author does not have the ability to call a market bottom to the day, but I would try to make that probability as high as possible. For now, the odds are against the market having already made a major bottom, as "evident" in our commentary last Thursday morning. Finally, there is one more crucial element in making a mean-reversion trade on the long side: You have to be reasonably confident that there will be a mean reversion! Northwest and Delphi were prime examples where there were no mean reversions. Commodities, however, are a different story. Beaten-down names of great companies also may be a good fit. A prime example is Philip Morris back in 2000. As for particular industries, there is now no doubt that manufacturing companies will not stand a chance in this globalized world of ours. The industries where Americans have a competitive advantage and will continue to have one for the foreseeable future is information technology and biotechnology. You want to short a stock? Short an airline or GM, but don't short Google. Over the long-run, shorting Google (even though it is overvalued on a P/E basis) is most likely a money-losing venture. I am reminded of investors at the turn of the 19th century who were still infatuated with railroad stocks; and at the same time, who showed a disdain for industrial stocks, even though they were about to overtake the railroads in earnings power and growth. The following quote on page 81 from the book "House of Morgan" is particularly interesting:
"So long as markets were local, industry seldom required large-scale financing, and there was a Wall Street and City bias against manufacturing as small-time businessmen. The Morgans had been mostly associated with railroad securities (As late as 1911, the second Baron Revelstoke of Barings could snobbishly protest, "I confess that personally I have a horror of all industries companies.") Now, as the great merger wave gathered pace, the focus of elite Wall Street banks shifted from railroads to industrial trusts."
As an aside, the bears will point to GM as "proof" that America is going down the drain; but history argues otherwise. The disappearance of GM as a major American industrial company will be replaced many times over by the rise of companies such as Google, eBay, Yahoo, Amgen, and Genentech. Let there be no mistake: The reason why GE has survived and has continued to prosper is because of its ability to reinvent itself constantly. Today, GE provides a variety of goods and services such has financial services, mass media, and medical equipment. With the world moving at an ever-quicker pace, it pays to be vigilant. Just witness the collapse of Refco in less than a week. Even the collapse of Long-Term Capital Management took a few months in 1998. Enron handily beat that record - collapsing in approximately a month in 2001. And now, there's Refco.
Okay, now that we have the main subject of our commentary out of the way, let's discuss the shorter-term action of the stock markets. This week, I would like to discuss, in particular, the homebuilding sector. With the ECRI Future Inflation Gauge now predicting the highest inflationary pressures in five years, it is natural to believe that the long bond will decline going forward (and thus increasing yields)- unless the U.S. economy or the world economy enters a recession sometime next year. In actuality, the latter scenario isn't too "out of this world" as any slowdown in the emerging markets may cause a "flight to quality" scenario where U.S. or foreign citizens sell their emerging market instruments and use the proceeds to buy U.S. Treasuries instead. This scenario occurred during the Asian Crisis in 1997 - accelerating during the Summer of 1998 with the Russian default, the Brazilian, and finally, the LTCM crisis. Many commentators have flip-flopped on the issue of where the 30-year Treasury is going to head going forward. In light of the difficulty in predicting the future short-term direction of the 30-year Treasuries, it now seems the prediction for lower rates back in our March 27, 2005 commentary wasn't a bad one at all - as we stated that rising rates were not a given at that time (the 30-year yielded 4.84% at the close on March 24th and subsequently declined over 60 basis points over the next couple of months).
In that commentary, I mentioned that both the Japanese and Chinese governments have no incentive to dump Treasuries en masse, and that continues to be the case today. The bears arguing for a sky-high interest rate precisely because of this possibility continue to be disappointed. This is important, since it is obvious that lower mortgage rates have substantially allowed homeowners to afford more expensive homes over the last few years. I will discuss a little bit more on the long bond on Thursday's commentary. For now, let's just discuss the most recent action of the homebuilding stocks, and what we can possibly expect going forward.
Readers who do not have a long trading history or who have not studied the history of the homebuilding stocks tend to forget that the homebuilding industry is a very cyclical sector. The relatively low PEs (relative to the S&P 500, that is) that you are seeing today may actually skyrocket when the industry turns (where earnings can literally disappear overnight). Please note that I am now discussing the homebuilding industry independent of the direction of the U.S. long bond going forward. With the speculative forth in the real estate markets and with the proliferation of the adjustable-rate mortgage markets, this author believes that we are due for at least a mid-cycle slowdown (if not a top) in the homebuilding industry and in housing price appreciation in the most speculative areas of the United States. Sure, the homebuilding stocks are very oversold in the short-run - but despite the most recent decline in the most popular homebuilding stocks, they are still only at the most recent April to May 2005 lows - suggesting that over the period of the next 12 to 24 months, prices still have a lot of room to fall, assuming that we are now in a mid-cycle slowdown. Following is a chart we have shown before - a chart comparing the current "bubble" in the homebuilding stocks with bubbles in the past:
The MarketThoughts U.S. Homebuilders Index is a price-weighted index of the five largest (by market capitalization) homebuilders in the United States. Like the above chart mentioned - should the homebuilding stocks ultimately turn into a full-blown asset bubble, we are already overdue for at least a mid-cycle correction here. As evident from the chart - even though the Homebuilders Index have fallen by nearly 25% since the July high, there is still a lot of fall if the bubble in homebuilding stocks fall the path of Gold in the 1970s, the Nikkei in the 1980s, and the NASDAQ Composite in the 1990s. Readers should also note that this is the most optimistic scenario, since there is also a good chance that the housing market - and subsequently, the homebuilding stocks - has already made a major top here.
The mid-cycle correction argument gets stronger when one takes a look at the following chart - a quarterly chart showing the amount of real estate held by households and nonprofits as a percentage of their total assets vs. the quarterly change (not annualized) in the OFHEO House Price Index:
This data runs from the second quarter of 1975 to the second quarter of 2005. The quarterly data for the OFHEO House Price Index is old data, but the data for the amount of real estate assets held by households and nonprofits is relatively new - as it was just released a couple of weeks ago along with the latest Flow of Funds report. First, a look at the OFHEO House Price Index shows a quarterly increase of 3.2%, which ranks as one of the highest ever readings. This is especially significant given that the reading during the third quarter of 2004 came in at the highest reading ever - which is doubly amazing when one takes into account the fact that this data is not inflation adjusted. That is, the real, inflation-adjusted appreciation in the OFHEO House Price Index is actually several times higher than the readings in the late 1970s - since the late 1970s was a period of high inflation. The percentage of household assets currently in real estate confirms the speculative nature in the housing markets, as this percentage again just made a record high - coming in at 30.23% and surpassing the 30% level for the first time in U.S. history. The U.S. homeowner has made a huge levered bet on a U.S. dollar-denominated, long-duration asset. Is this a good bet? History suggests that it is a resounding "no." Unless the prices of homebuilding stocks crash in the next two weeks, this author will continue to avoid homebuilding stocks (on the long side) for the foreseeable future.
Okay, this commentary is getting long already, so I will try to end this by now discussing the most recent action in the U.S. stock market. To sum up, this author is still not satisfied with the current conditions in the stock market. Sure, it is oversold per some of our technical indicators. For example, the Rydex Cash Flow Ratio has further increased on Friday from a reading of 1.08 to 1.09 - which is actually a six-year high. The ten-day moving average of the equity put/call ratio is also now at 0.70 - comparable to the readings that we got in April earlier this year (although this reading "sold off" to as high as 0.85 in August 2004). Interestingly, however, the ten-day moving average of the NYSE ARMS Index actually decreased from 1.19 to 1.16 from Wednesday to Friday at the close - definitely not an oversold reading by any means. Moreover, breadth indicators aside, we still have not seen a major sell-off in any of the major stock market indices so far, and historically, the market does not make a sustainable bottom until we have at least seen some panic selling in the major market indices.
Can the market bottom here? Of course, anything can happen - especially in the financial markets (for readers who don't believe this, just look at the collapse of Refco in a period of less than a week). The market could very well take off here without a major sell-off in the major market indices of a higher NYSE ARMS reading, but history suggests that it won't be a sustainable one. Moreover, given the onset of earnings season and the inability for GE to rally off of its most recent earnings report, this author would definitely not commit on the long side just yet.
We now turn to the chart showing the most recent daily action of the Dow Industrials vs. the Dow Transports. Again, the most significant news of last week from a Dow Theory point of view is this: The refusal of the Dow Transports to decline below its September 20th low is a bullish non-confirmation of the Dow Industrials. The Dow Transports almost confirmed on the downside on Thursday at the close, but it was not to be. The relief rally on Friday sealed it for the bulls again. If the Dow Transports does not decline below its September 20th low within the week or so, then this author will most probably cover our 25% short position in our DJIA Timing System:
Even though the Dow Transports was the weaker index, the fact that it again did not confirm on the downside is a warning sign for the bulls, and should continued to be watched going forward. Of course, with earnings season now upon us, anything can happen. The "cushion" of 57 points is really nothing if UPS, say, misses its earnings or guides lower going forward. However, this author would not hesitate covering our 25% short in our DJIA Timing System should the Dow Transports failed to confirm on the downside within the next week or so. For now, however, this trade still looks okay - but with the market this oversold per some of our indicators, it may not be prudent to stay short for much longer.
Our most popular sentiment charts this week is going the right direction for the bulls, as they are starting to become pretty oversold. Let's start with the Bulls-Bears% Differential readings in the American Association of Individual Investors Survey vs. the Dow Industrials. The latest weekly reading is declined from last week's 23% (which was very high given the weak performance of the stock market) to this week's negative 9%. The ten-week moving average is now at 4.4%:
Both the weekly and the 10-week readings are now becoming very oversold, but please keep in mind that we have seen much more oversold readings earlier this year in April, as well as back in March 2003. We did not really get a sustainable rally after the oversold conditions in April. My guess is that the current bullish sentiment in the AAII survey would need to get more oversold before we will see a sustainable bottom. At the minimum, this author would like to see the 10-week reading sell off to below the 0% level before committing on the long side. This does not preclude us from covering our 25% short position and shifting to a completely neutral position, however - as it may be prudent to try to protect one's profits on the short side instead of "going for the grand slam." As a rule, one shouldn't "go for the grand slam" on the short side unless one is either reasonably confident that the company one is shorting will go bankrupt or if one is certain the market is entering a new cyclical bear market.
The Bulls-Bears% Differential in the Investors Intelligence Survey is again "in gear" as the weekly reading declined from 21.7% to a highly oversold reading of 16.6% in the latest week - the most oversold reading since early May of this year. The four-week moving average declined from 25.9% to 23.4%:
Bulls and bears alike should continue to be careful here. Optimally, this author would like to see a still-lower reading just up ahead, with the resultant four-week moving average declining below 20% when all is said and done. Of course, nothing ever works out perfectly, but like I have mentioned many times before, this author would like to see as many factors lined up on our side before committing on the long side. The Investors Intelligence Survey is just one piece of the puzzle.
As for the Market Vane's Bullish Consensus, we finally saw some significant action in the latest week, as the Market Vane's Bullish Consensus declined from a reading of 62% to 58% - the most oversold weekly reading since late August 2004. Is this a sufficiently oversold reading? This author would still argue "no," as the four-week moving average is "only" at 61.8% - which is still relatively high compared to the 58.0% reading we received in late August 2004:
Recent history has shown that the market cannot enjoy a sustainable rally without the Market Vane's Bullish Consensus declining to at least a reading at 50% or lower - no matter how oversold the AAII or the Investors Intelligence Survey became. Therefore, this author would like to see the Market Vane's Bullish Consensus declined to precisely that level before committing on the long side. A 50% reading would make it the most oversold reading since August 2003.
Conclusion: For readers who are sincerely interested in learning about the markets and want to succeed in the long-run, the mean reversion trade has been one of the three distinctive strategies for making outsized returns throughout financial history. While this author would argue that both the carry trade and the momentum trade would be made more difficult by an ever-more efficient market going forward, this is not the case with mean-reversion trades. Moreover, a mean reversion trade is usually conservative in nature - and also allows one to reap a substantial amount of knowledge in both the financial markets and the industry/asset class one is dealing with. However, reaping profits through the mean reversion trade is not for everyone, as it is equally important to learn about one's psychology and tolerance as it is to make correct calls in the stock and financial markets.
At this point, commodities and homebuilding stocks continue to be a "sell." The general market, however, is getting too oversold for my liking - and if the Dow Transports does not decline below its September 20th closing low within the next week or so, this author will most probably cover our 25% short position in our DJIA Timing System, and go completely neutral. That does not mean we will commit on the long side, however, as this author is still waiting for a more oversold condition in the general market before doing so. We will continue to keep our readers up-to-date throughout the week.
Henry K. To, CFA