Paradigm shift in Asset Management - Part 1: the buy & hold myth shattered

The models which millions of investors have used in the past to construct their portfolios are, it seems, not nearly as robust as they ought to be. 2008 was an exceptional year, and as such clearly proves the point. FUTURES looks at the key paradigms of "traditional" models and explains their limits. Part 1 focuses on the basics and examines the theory that the "buy & hold" investment strategy gives a good rate of return.

What is the probability of a coin landing heads or tails if tossed fairly and squarely? Common sense - or indeed any school kid between Tokyo and Los Angeles - would tell us 1:1, 50%, 0.5 or 1/2.

"Common sense": often little more than an approximate value.

Indeed, physics tells us that all of these answers are merely approximations. Although the chances are very slim, the coin could actually land right on its edge. After a long period of heads or tails, at some point we have to reckon with a draw. On a day-to-day basis, we could be forgiven for discounting such a rare event altogether, as with many other low-probability events. For example, what's the sense in preparing for the possible onset of winter in August if you're a Tuscan farmer? The costs involved would be higher than those caused by the potential damage.
Faced with today's challenges, wouldn't it be better to pay more attention to the improbable? We are talking about something which bestselling author Nassim Taleb refers to as "Extremistan". Extremistan is an environment where low-probability events can be highly consequential.

The Inductivist Turkey

In his book "The Black Swan", Taleb uses the allegorical tale of the turkey and the butcher attributed to philosopher Bertrand Russell: the main protagonist is the turkey, which at some point realizes he is being fed at the same time every day. The turkey finds that, on his first morning at the turkey farm, he was fed at 9 a.m. Being a good inductivist turkey he does not jump to conclusions. He waits until he has collected a large number of observations that he is being fed at a certain time. He makes these observations under a wide range of circumstances, on Wednesdays, on Thursdays, on cold days, on warm days. Finally he is satisfied that he has collected a number of observation statements to inductively infer that he is always fed at 9 a.m. However, on the morning of Christmas eve he is not fed but instead has his throat cut.
Russell's allegory demonstrates what has been known since time immemorial: we mistake the frequency of an event (inductive arguments) for proof that it will continue in the future. Even if our coin lands on either heads or tails a million times over, we can never rule out for sure that the coin will eventually land on its edge. An improbable event only has to occur once to refute the inductive argument.

Dominated by extremes

If we want to bet on heads or tails, the unlikely event of a coin landing on its edge is inconsequential. But if we were to move the goalposts and equate the consequences of the coin landing on its edge with a loss of total assets, then we'd think very carefully about playing the game at all. Oddly enough, high-risk bets are common within the financial world.
Let's take the US subprime mortgage crisis, which has its roots on the West Coast and which is a symptom of the broader global financial crisis. It set in motion the wheels of a global crisis which was based on a complex system of risk transfer. The basic premise was to create a risk-diversified portfolio tailored to individual risk-return constraints through a combination of loans to debtors of varying creditworthiness (prime, subprime, etc.). At the same time, the non-diversified risks of the specialized mortgage lenders could be fragmented and re-sold. On this basis qualified risk managers used rating agencies to put together complex products such as Mortgage Backed Securities (MBS), Asset Backed Securities (ABS) and CDOs and SIVs. With the idea that "there's a solution for everybody", banks and insurance companies pumped huge sums into these designer products. As it turned out, too many individual risks slipped through the net, having been assessed on the basis of historical trends, but their mutual dependencies left largely unresearched. As the crisis grew, product designers, rating agencies and investors were gradually hit by the realization that the coin can land on its edge after all - and at any time.
However, there were many more lessons to be learned from 2008. A whole range of other much more popular concepts which had hitherto fallen under the "common sense" umbrella reached their limitations. Some of them were based on "inductivist turkey observations" and so proved to be typically human simplifications. Others, however, had, on closer inspection, always been houses built on sand. In this and subsequent issues, we shall be focusing on the ingredients of popular Asset Management recipes and kick off here with what is probably one of the most enduring myths of the stock market.

Slung out: Unisys, the erstwhile rising star of the age of information, now reduced to a penny stock. The component was dropped from the S&P 500 in November 2008 and replaced by savings and loan holding company People‘s United Financial Inc.

The buy & hold legend

The legend has it that equities virtually always render higher returns for those prepared to wait long enough compared with low-risk forms of investment such as government bonds. Investors need only to put a collection of stocks in a securities account and wait 10 or 20 years before cashing them in at a big profit. As the conventional marketing blurb would have us believe, buy & hold is a long-term investment strategy based on the concept that in the long run financial markets always give a good rate of return.
In its most basic form, the legend can be taken with a pinch of salt and unraveled without the help of any statistics at all. There are demonstrably very few exceptional companies which have been extraordinarily successful over a given period of time and even fewer have been able to sustain that level of success over a period of several years. Clearly the vast majority of companies are somewhere between mediocrity and bordering on insolvency. Lastly, there is a not insignificant number of companies which have gone bankrupt, ceased trading or have been taken over.
The story is no different for corporations. After all, there is no established law which only allows exceptional companies to go public.
The typical benchmarks for success and failure within the equity market appear to tell us exactly the opposite. Equity indices, known as the "Top 30" (DAX) or "Top 500" (S&P 500) and all other known "benchmarks" which are found every day in the pages of financial daily newspapers, in the reportages of financial TV and radio and rooted in thousands of websites as a representation of the stock market, are anything but  representative of being exceptional. They exclusively comprise those companies which have already shown flashes of brilliance or at least demonstrated above-average growth.

Survival of the fittest

Even if such companies approach the end of their existence as market leaders, there is no ramification for the relevant index, even in cases where an organization's exit is as quick and spectacular as erstwhile US investment banks Bear Stearns or Lehman Brothers. They are delisted, i.e. removed from the index, without so much as a by-your-leave. The gap is then filled by younger, more successful or even formerly listed companies which have managed to claw their way back to the top of the ladder after a period in the doldrums. Therefore, indices are not the ideal instrument for painting the big picture in terms of stock market trends. They are much more a gauge for proponents of Darwinian Economics; indices represent the value of the fittest companies at any one time. In the financial research world, this kind of distortion is referred to as "survivorship bias". Survivorship bias is the tendency for failed companies to be excluded from performance studies because they no longer exist. Those companies which perform badly (or which are no longer quite as fit) are dropped, thus masking poor performance.

Comings and goings on the DAX

Let us consider the DAX index, a stock performance index comprising the 30 most actively traded German Blue Chip stocks on the Frankfurt Stock Exchange. Of the original index values from 1988, only 16 were still in existence at the start of 2009, and if we were to take mergers and acquisitions into account, Daimler Chrysler, for example, then we would be left with just 13 companies. In 2008 alone, three of the 30 DAX companies, or one tenth, lost their places to other components. Mining company K+S replaced the poorly-performing TUI conglomerate. Hypo Real Estate was replaced by Salzgitter and tire manufacturer Continental by the more resilient branded consumer products company Beiersdorf (band aids beat car tires). Founded just 21 years ago, the DAX is one of the youngest equity indices. As far as the age-old Dow Jones Industrial Average is concerned, only one company remains since the index's launch in 1896 (General Electric) and the S&P 500 has witnessed around 290 changes since 2000. In 2008 alone, 37 components were dropped from the index.
The structure of an index, which reflects the performance of the largest companies (in terms of order book volume and market capitalization) within an industry group at any one given time, is not based on any mischievous ulterior motives, but is essentially based on logic and sensibility. What is untenable, however, is the widely held view that stock indices accurately reflect the trends of the vast majority of corporations. This is a premise of which the founders of modern portfolio theory were also aware of, which recommended an alignment of the weighting of individual stocks within a portfolio with the current market capitalization (i.e. the same strategy employed by capitalization-weighted indices). Only those investors who choose to follow this advice, which is no problem nowadays thanks to easily accessible and inexpensive index instruments such as ETFs, can therefore bank on equities giving a good rate of return in the long run.

Long-term yields on the S&P 500

To see whether the buy & hold legend has held true in the past, at least in terms of strongly index-oriented buying patterns, let us take a look at one of the stalwart indices. In its current form, the S&P 500 has been around since 1957 and data generated using backward projection goes back to 1928. The index reflects the success and failure of the USA's 500 exchange-listed heavyweights and represents around 75% of the total capitalization of the US stock market. Assuming an investor continually adjusted his portfolio in line with this index, starting in 1957, the result would then be, before expenses, a return of 15 dollars and 75 cents from every dollar invested (as of end February 2009). This equates to an annual return of roughly 5.4%.
An investment horizon of more than 50 years is, however, exceptional. It is more usual to look at expected yields from a 10 to 20-year period.
An analysis of this period reveals highly disparate results. Since 1950, the S&P 500 shows the bandwidth for annualized returns of -5.1% to +16.8% for an investment horizon of 10 years and factoring in the respective end-of-month values. Given an investment horizon of ten years, we have a bandwidth between an overall loss of around 41% of the capital (from every dollar invested at the end of February 1999 only 59 cents were left by 2009) and an excellent profit (every dollar invested would have grown to 3.70 between August 1990 and 2000).

20-year window on the S&P 500: from savings book levels to jackpot returns

Over a 20-year period, the historical bandwidth for annualized returns is between 2.4% and 14.4%. At any rate, the loss of capital would have been smaller for this investment horizon. Any investor fortunate enough to purchase S&P stocks in October 1980 and to repeat that success through to 2000 would have hit the jackpot and multiplied his capital (before expenses) by almost fifteen. The worst return would have come from an investment made in March 1962, with total returns of around 59% for the entire period equating to something like savings account levels. As of the end of February 2009, the returns achievable over the past 20 years stood at an annualized 4.15% or an absolute 125% which, given the risk, is pretty meager. Talking of risk, how do stocks compare in the long run with other more volatile forms of investment?

Stocks versus bonds

The younger an investor, the greater the proportion of stocks in an investment portfolio comprising stocks and bonds should be. This general rule of thumb is regularly touted as the gospel truth. It is also based on the assumption that stocks return higher yields than bonds over a longer period of time.
Let us test the probability using the S&P 500 and 10-year US government notes based on the following model:
1.) The equity investment is made in such a way that the portfolio always corresponds to the benchmark. For the purposes of our assumption, we shall analyze all possible 10 and 20-year periods since April 1953 on a monthly basis (the maximum possible period for which data are available for both the S&P 500 and US bonds).
2.) The S&P 500 is not a Total Return Index, where it is assumed that dividend distributions are reinvested in the index. In the interests of fairness, we shall assume that the investor buys and holds a bond with a 10-year maturity whose return corresponds exactly to the return of a 10-year US government bond as of the time of investment. The annual coupon payments are not reinvested.
3.) Over a 20-year period, for which there is no uninterrupted time frame for bonds, we shall assume there are two successive ten-year bonds (e.g. January 1953 and 1963), which are calculated according to the aforementioned model. The capital and returns from the first bond are invested in the next one at the end of the first ten-year period. In many cases the bond's performance naturally appears to be poorer than would be the case, as for example, if we were to take a (e.g. 30-year) bond with a remaining time to maturity of exactly 20 years for our comparison (but such a bond would not always be available in the past).
4.) Expenses, securities account fees and taxes are not factored in.
5.) Because the downside risk of stock investments is much higher than for fixed-interest bonds, they should offer a risk premium in the form of a reasonable outperformance. This will be measured separately.

Historical out/underperformance (annualized in each case) on the S&P 500 compared with US government bonds over a 10-year (dark blue) and 20-year (light blue) investment horizon. Example: in the last 20 years (02/1989 to 02/2009 - last value on the "20-Y" curve) the S&P 500 has returned a 0.64% lower yield per year than government bonds.

The results

Across all 551 ten-year periods between 1953 and 1999, the stock investment showed an average annualized outperformance of 2.2% compared with the fictitious ten-year government bond. The highest outperformance in the period October 1990 to 2000 was above the 10% threshold. Overall, stocks outperformed bonds in fewer than three quarters (73.5%) of all possible 120-month periods. Or to put it another way, in the past, the S&P 500 returned higher yields than government bonds in almost three out of four cases.
For the 1960s and 70s as well as the recent past, the remaining 26.5% of cases (146 ten-year periods) look pretty miserable. At present (February 1999-2009), the underperformance of the S&P 500 over a ten-year period has matched the current all-time low of a drop of over 9% compared with bond yields (per year!). Or to put it in concrete terms, whereas each dollar increased to 1.30 in the case of our fictitious bond, each dollar dropped to 59 cents on the S&P in the same period.

20 years: no better

Incredibly, the performance figures for an investment horizon of 20 years are even worse, which is a clear indication that holding on to equities for a longer period of time does not necessarily increase the chances of bigger returns. In a buy & hold scenario of 20 years, only 71% of equities produced greater returns than our (fictitious) bonds. In a historical 20-year window, equities show an average annualized outperformance of 1.8%. The best result would have come from an investment made precisely in June 1979, experiencing the entire bull cycle of the 1990s and avoiding the decline after 2000 (7.4% risk premium compared with our bonds). The worst result would have come from the period between December 1959 and 1979 (-1.9% underperformance annualized).

Two "Total Return" indices compared

Professional investors would have serious reservations about the approximation model used for our historical comparison by utilizing the S&P 500 and the fictitious bonds. They would object that no investor in his right mind would hold on to low interest-bearing stocks through to the end of their maturity if he had the opportunity to invest in higher interest-bearing equities. Other factors in our simple comparison might also give the "hypothetical" buy & hold stockholder the edge over the fictitious bond investor, as for example, the fact that there were long periods where a reinvestment of high-coupon bonds would have returned significantly higher yields than comparatively meager dividends.
Total Return indices would provide a more valid comparison, reflecting the behavior of a "rational" investor more accurately. They invest dividend distributions as well as coupon payments and also reflect fluctuations in the prices of stocks when the level of interest rates changes. They include the German DAX as well as its counterpart for German government bonds, the REX Performance Index. Both indices have a real-time history dating back to 1988. There are back projections for the DAX dating back to 1959, but unfortunately there is nothing for the REX prior to 1988. Therefore, we can only draw conclusions from the more recent past and have sufficient data for the 10-year window only.
The figures paint a very similar picture in regard to the longer-term comparison between the S&P 500 and the US government bonds:

This comparison also reveals a 3/4 chance of higher returns from stocks than bonds in all of the 10-year periods in question. The average risk premium was 2.83% (= annualized outperformance).
Over the entire period of a little more than 22 years, a fleeting glance at the chart reveals that the outperformance of the stock portfolio represented by the DAX comes at a considerable risk. Although the DAX and REX are relatively close towards the end of the axis, the stockholder experiences slight peaks and deep troughs along the way, compared with a gentle upward curve for the bond holder. Or to put it into more concrete terms, the total yield of the DAX was 284%, or 6.6% annualized with an annualized volatility (annual fluctuation range) of 22%. The overall performance of the DAX was 262%, resulting in an annualized yield just 30 basis points below the DAX (6.3%). By contrast, the volatility of the REX at just 3.3% was much lower.
At present the bond holder who began to align his portfolio to the REX ten years ago would have been over 7% per year better off than the holder of a portfolio aligned with the DAX.

Conclusion: invest wisely

The figures prove that the buy & hold myth is one of the many rules of thumb you cannot rely on. It may have held true at least for the two index-based investment strategies presented here in most of the historical analyses, but in only three out of four cases. As demonstrated by the turkey allegory, it is not possible to base prognoses on past observations. We cannot even make forecasts about the markets under examination using the historical probabilities we have calculated. And our calculations are useless in terms of estimating potential returns on other markets for where even less performance data is available.
Added to which is the fact that our model, when applied in practice, would probably see stocks come off worse. Before index funds with very low fees became available (i.e. the 1990s), a private investor could only have accurately reproduced an index of "survivors" at considerably higher cost than for a comparatively simple bond portfolio. In reality, a stock investor would therefore do well to estimate the risk even higher than our two tests suggest.

And now for some good advice on the subject of buy & hold:

1.) Buy & hold strategies on the stock market have always been something of a gamble. Even a long-term investment in stocks and a strict alignment with an index is no guarantee of playing it safe. So the same applies to this kind of investment as to all other assets with risk; limit your losses using well-placed stops (stop buy/stop loss defined at the time of purchase). 2.) Hedge your long-term holdings. If you are planning to employ a buy & hold strategy, the high downside risk can only be offset by a similarly dimensioned counterweight in the overall portfolio. Such an instrument should be negatively correlated in downward stock market trends and yet still be capable of a positive earnings expectation over time. Managed Futures would fit the bill ideally (e.g. FTC Futures Fund Classic: +77.8% three-year performance between February 2006 and 2009). 3.) Greater security through an absolute return strategy. Any investor hoping to significantly lower the risk of drawn-out and severe bear markets and still participate on upward markets would be well advised to follow absolute return strategies rather than the traditional buy & hold investment strategy. Here, stocks are held during upward markets but returns can still be generated during downward markets. Which is why FTC offers the FTC Gideon 1 stock fund, which offers strict risk management of the "long" stock portfolio in combination with a market-neutral long/short portfolio which, even in longer downward markets, is still capable of returning yields.