Diversification is not an end unto itself. More and more diversification is not necessarily helpful at all. The thing is, concentration adds to return and diversification removes risk, but both only to a point. So which is it that you need, a more concentrated or a more diversified portfolio? This is a question without an answer. There is no science to balancing these two—just art. As we will see in the chapter on allocation process, we can determine if an added manager reduces risk or not and by how much. We also can see what effect the new investment may have on the expected return and the right (good) side of the return distribution, but when is enough enough? The easy answer is that it depends. Certainly once a new investment stops subtracting risk or adding return, you have reached enough. Before that point however, it will only be your sense that your efforts toward risk and return have balanced one another.

In a general sense, as you or your investment officer add unconnected investments, you are increasing the probability of success but also reducing the returns that are possible. High average returns are part of reducing risk, and do provide a buffer for those inevitable down drafts. Also, part of your effort to reduce risk should be focused on the specific events you can forecast and the investments that have a high probability of impacting them. This part is less science and, again, more art. Are you forecasting inflation, geo-political disruptions, resurgence of particular markets, strong growth in particular ­segments, bursting of market bubbles, or a change in political direction? What about a change in your need for cash, perhaps due to a big project or new grant program? Once these events are identified, the question then becomes, What investment will best push the portfolio in the direction you want, or do you have enough already in the portfolio? If not, more investments will be required—more “diversification,” if you like.

The discussion of what constitutes diversification is an “inside baseball” discussion. Academics and most consultants tout “classes” of assets, while some CIOs and a few consultants like the “buckets” approach. Both depend on correlation coefficient calculations, which become their weak link. Correlations are a weak reed upon which to base decisions. The story is that you want investments that are correlated such that when one investment is up, the other is down. Hogwash is the nicest term we can use here. You already know that when markets are strongly down, all correlations become one; in other words, they all go down together. Those that live on correlations must then believe that when markets go up, they want part of their portfolio to go down, so that when markets go down all of their investments can go down as well, and the correlation stays low. If you buy the correlation theory, what you really want is a correlation of one, so when the markets go up so do all your investments, because when the markets are down you know all your investments will be down too. Following that line of thought, we would just ignore correlation.

Too often, people equate correlation with connectivity. You can talk about the correlation of the markets with ladies’ skirt length, and with who won a sports game, or a particular year in a President’s term, but no one really believes that skirt lengths or the number of years a President has been in office determine the market.

What we are trying to do when we diversify is to control the connectivity between our investments, or find investments that are not connected at all. The way to do that is not by assigning different names to asset classes or different buckets of investments. It is to focus on those areas that are or may be common to each investment. The three biggest elements or vectors on which to diversify are (1) information stream, (2) execution and (3) unique risks.

Let’s take three private equity funds that all fit into the private equity bucket of the alternative asset class and explore just what we mean by these three diversification vectors. The three are a growth equity fund in China, an energy fund in Brazil, and a micro LBO fund specializing in fast food restaurants in the southeastern U.S. These were actual funds in an actual portfolio. We are using these three because the differences are easy to see, but the principles apply to every manager in your portfolio.

Information stream is the how and where does the manager get information. Research, sure, but what research, what markets, and what sub-parts of the market? Do they have something extra, some secret sauce or some lagniappe? These three example funds are all private equity, but one fishes from consumer-based growing companies with revenue in China, another uses only guaranteed contracts for energy in Brazil, and the last fishes from small franchise restaurants one or two at a time. A change in the U.S. has little connection with any of them. Take-or-pay energy contracts in Brazil have no connection to non-export consumer businesses in China, and going out to Taco Bell has nothing to do with China or Brazil, and a limited amount to do with the state of the economy in the U.S.

Even in the small-cap value asset class, two managers can use different sources of information; one can focus on IPO and mutual insurance conversions and the other on turnaround or underperforming names. The information stream will include all the data they use to track names in their space. One micro-cap manager added to his normal research through a massive Rolodex of individuals that he could call to get information—not inside, just public information that was not followed widely by most analysts.

Execution is the process the manager uses to convert the information into return. Back to our special three: The China group plots growth and focuses on the ability of the company to IPO or sell to a strategic buyer in a short time, the Brazil group on the ability to build and then contract for long-term delivery of electricity, and the Taco Bell fund rolls up those small franchises into one and relies on the economy of scale. Some equity managers use momentum and some use a discount cash flow or dividend approach.

Unique risks are just that: what will cause a particular manager to underperform? The unique risks of the three private equity managers are easier to see, but unique risks exist for domestic managers as well. One may be sensitive to interest rates, another to a regional economy and a third to regulatory approval or changes. The list is almost endless (unfortunately). The fewer shared risks the better the diversification.