Risk Control as a Return Enhancer
IN PRINT ARCHIVE CIR Winter 1999
|Risk Control as a Return Enhancer: A paradigm shift?|
|Maybe...but which kind?|
|by Eric Innes|
We often hear the words "paradigm shift" these days. But, which kind? Some suggest that we've entered a new era, in which the growth prospects of the economy are so steady and reliable that a fundamental revaluation in the capital markets is not only taking place, but is sustainable long-term.
We think we may be facing a paradigm shift of a different sort, in which institutional investors reshape their return expectations for the future to reflect the lofty valuation levels that we see today, and in which institutional investors come to accept the notion that sustained double-digit returns are not plausible over the coming 10 or 20 years. The change in plausible long-term capital market returns is sufficient to merit a fundamental reevaluation of asset allocation policy and of actuarial return assumptions, with all that this may imply.
The past 23 years have been spectacular - a dollar invested in the S&P in 1975, with reinvestment of dividend income, would today be worth over $40. This represents a huge appreciation in the market, dramatically beyond economic, earnings or dividend growth. Bonds have also been impressive, so that balanced portfolios have delivered mid-teens returns for 23 years.
What about the next 23 years? Are we going to see another forty-fold appreciation of every dollar invested in stocks over the next 23 years? Such returns are not believable from today's valuation levels.
Reflections on Long-Term Returns
Why should something unsustainable in the long term be a reasonable expectation in the short term? Common sense suggests that there's a bit of a mismatch here. Let's take a more serious look at this issue.
Lessons from Recent History
Finally, real dividends historically have grown at 1% per annum. This surprises a lot of people; it's lower than most people think. But keep in mind that real dividends don't just grow with the economy; economic growth consists of growth of existing enterprises as well as the introduction of new enterprises, and dilution of existing holdings with secondary equity offered, all of which means that the growth in real dividends is necessarily slower than the growth of the economy, and in fact historically has averaged 1%. We might hypothesize that today, with dividend payout ratios at very low levels by historic standards, we might prospectively double the historical norms. So, we might see 2% real increases in dividends for the relatively long term, over the next decade or two.
If we start with 8.2% real returns over the past 72 years, we must first take away the 1.8% real return that came solely as a consequence of changes in valuation levels; so we are down to a 6.4% real yield. If you take away the 3.8% difference between the starting yield in 1925 and your current yield in 1998, you're down to 2.6%. The basic message is simple: much of the real returns that we've seen in the past are not replicable because we can examine the components of returns and ask, "what is a reasonable expectation?" The answers are disappointing.
Conversely, if we start at 1.6% dividend yield, and have a 2% real increase in dividends, that gets us to a 3.6% real return. Not 8%, not even 5%, but 3.6%, plus or minus whatever valuation level changes may occur. Can valuation changes make up the difference and get you back up to an 8% total real return? Of course they can. Over the next 23 years, if dividend yields fell from 1.6% to 0.15%, real returns on equities would match the past 23 years. Over a 72-year span starting this year, if yields fell to just over one basis point (1.5 bp, actually), you would get back up to an 8% real return.
We can also subtract the bond yield to get an expectation with regard to risk premium of stocks relative to bonds. In 1925, bonds gave us a 3.7% yield. With inflation expectations at the time hovering near-zero, this would also be the expected real yield of that era. But bond yields went up, so bond valuations went down by 0.7% per year, giving us a risk premium of stocks relative to bonds of 5.2%. Plugging today's real yields into the same set of expectations, we come up with a prospective stock/bond risk premium in the 0%-2% range.
This is awful; it can't possibly be right. Stocks have never delivered so little for an extended period of time, right? Not true. In 1801, an investor in stocks would have earned less than an investor in bonds for the next 71 years. So for those who think that 72-year span covered by Ibbotson is sufficient to shape long-term expectations, history reminds us that times change.
A major change in long-term institutional investor expectations is necessary and this is the paradigm shift that we think is actually likely. The institutional investors will gradually, in time, reshape their own expectations from 5+% real returns as "the norm" for the typical balanced portfolio, to an expectation of 2% to 4% real returns as "the norm" for a reasonable balanced portfolio. This is a dramatic change, which must happen, in time! The implications for asset allocation policy and for actuarial return assumptions may be significant.
What is Portfolio Wealth?
What, then, is "wealth" for the endowment, foundation or going concern defined benefit pension plan that seeks to serve a perpetual obligation? Here, the answer has to do with the anticipated purchasing power as a perpetuity of the portfolio. If markets are high and the prospective real returns are only 3% (as one might well argue for equities today), the purchasing power of the portfolio is not necessarily very large relative to the nominal assets in the portfolio. Conversely, if the market has recently tumbled and the prospective real returns are in the 6% range or higher (as was the case in 1982 and 1974), then the purchasing power of the portfolio can actually be surprisingly large relative to the portfolio's nominal value.
In short, portfolio wealth is not simply assets. Nor is it the simple tally of assets less liabilities that we are so accustomed to. That simplistic definition of wealth might be termed "current" wealth. In most categories of investors, whether individuals, endowments, foundations, or pensions, the far more important measure of wealth is the size of the real income stream that the assets can purchase. This measure of portfolio wealth is only tangentially related to the size of the current net worth. To the extent that the current net worth rises or falls with the whims of the capital markets, the real income stream that the portfolio can generate often barely moves. At the nadir of the Great Depression, for an investor with a 50/50 stock/bond mix, the asset value of the portfolio was down 56%, but the real income generated by the portfolio was down only 8%, during the greatest bear market in the United States of the past two hundred years!
This is a long preamble to introduce the idea that modern risk control technology, combined with derivative structures, can provide the best opportunity to enhance the true value of a pool of assets.
The Concept of Portable Alpha
The portability of alpha suggests a simple and intriguing strategy. Hire the managers whom you believe offer the largest alpha and/or the greatest likelihood of a positive alpha. The manager selection decision is made without any regard to which asset class the manager invests in. Futures are then used to "transport" this alpha into whatever mix of assets you really want to hold on an asset allocation basis. In a 50/50 stock/bond portfolio, if you think your stock managers have skill and your bond managers do not, why settle for only half of the potential alpha? By investing with the equity managers and "transporting" half of the alpha to the bond markets, you can wind up with your intended mix and with twice the alpha.
For example, suppose you want to hold the MSCI World Index, but the only managers whose skills you truly respect manage (1) an Australian market neutral (long/short) electric utilities strategy, and (2) a French bond portfolio. The former has no market exposure, because the strategy is market neutral, but does have currency exposure. The latter has French bond and French franc exposure. If we hire both managers, short the Australian dollar, French franc and French bond futures to an appropriate extent, and purchase stock index futures in an optimized basket of the available stock futures in the MSCI World Index, along with their corresponding currencies, we have achieved the goal. The two superstar managers produce their alpha, which you have simply added on top of a passive MSCI World Index return.
The preceding example was deliberately farfetched. However, we are seeing more and more examples of conventional "Portable Alpha" programs. Two-thirds of our U.S. Market Neutral Equity clients "equitized" their program into the S&P500. Simple market neutral programs produce a return that closely resembles Treasury bills plus two alphas. By buying S&P futures, we wind up with S&P returns plus the same two alphas. The alpha is "transported" into the S&P500.
The basic message here is fairly simple. "Portable alpha" programs are already more than a concept, they are a live reality. They are no longer experiments on the "radical fringe" of the institutional investing world. It makes a great deal of sense to look for the managers whom you trust to produce reliable material alpha. Separately, make the asset allocation decision as to what mix of asset classes you want to hold. Use global or domestic futures to transport the alpha of your active managers into the markets that you truly want to hold. This is a common sense expedient to wring the most out of markets that are no longer priced to provide any realistic expectation of long-term double-digit returns. It is not realistic to hope for double-digit returns unless one can add material and reliable alpha to your basic asset allocation decisions.
Eric Innes is Chairman and CIO of YMG Capital Management Inc.