—Stephen Covey

Currently, the starting point for investing in the industry is to assign various investments to differing buckets or asset classes, and then optimize the sizing of each bucket based on the average performance of the asset class, with the goal of reaching some “efficient frontier.” First the investment committee or consultant decides how many and which buckets to name. Then they decide which bucket each investment belongs to. Some can be easy, but others not so much. Should this investment be called distressed debt, or is it high-yield? Does the high-yield manager fit into the fixed income class or alternative class? Is international equity separate from domestic equity or are they both part of global equity or is it an alternative? Does an equity mutual fund of natural resource commodities heavily biased to international names fit into the alternative box, or the international equity box, or is it a new class—natural resources? Where do you put MLPs and REITs? After you decide on which bucket each belongs to, the next decision is how much of each bucket you want. Into the machine goes these several decisions, and like the Great Wizard of Oz out comes the answer. The next step is then to find filler for each bucket. The actual investment manager or investment used to fill the bucket is often the one who is on the consultant’s list or platform, and not necessarily the “best” or even “good” for your portfolio.

Asset allocation like this is probably responsible for more failed performance than any other element of investing. One cannot separate allocation from manager selection and risk control. Asset allocation is the result of the portfolio-building process, not the start. This is crucial to successful investing.

Deciding that a certain percentage of your portfolio needs to be in any particular asset class is an exercise in hubris. Who can know in advance which managers one can find, or the state of risk? Does one add in any old manager just to satisfy an allocation percentage pulled from space? Where did that percentage come from—someone else’s allocation? What risk did that fund take? What were their goals? A manager that fits the portfolio is more important than mere participation in some ill-defined asset class like “alternative” or “emerging market” or even “large-cap value.” Building an iron box in advance and forcing your portfolio to fit that box would make Procrustes proud.

One state teachers’ pension fund was convinced that they needed a certain percentage allocation to private equity. They were told it would take several years to find the right (diversified) funds for this $300 million allocation, because not all the best were in the market at the time. They were also cautioned about the “J-curve” effect. They ignored the advice and committed the entire amount inside of a couple of months. They were a little big for that bed and so were severely chopped off, losing a significant amount of money.

The key to controlling a portfolio is diversification; a simple idea, but not so simple to execute. In fact, building a diversified portfolio is difficult because it requires deep thought and an understanding of the investment. This is the area that separates the middling investment officers from the outstanding ones. The average investment officer, average consultant, or average investment committee member thinks (because they have been told far too often), that diversification is a simple effort and that many differently named asset classes are all that is needed. The real truth is that asset classes don’t provide diversification; they are used by academics only because they are easy to name, and the data is easy to get.

Rethinking the process

So how can a foundation or endowment allocate assets to control their risk? To start, they would look to the end point, their goal. If they don’t have a destination, they won’t get there except by chance. If return is not already stated as a percentage, you will need to convert your goal into a percentage. This becomes not your target return, but your minimum return. Anything more is wonderful and anything less is a failure.

Starting the process, not unreasonably, starts with the grist in your mill—the diverse package of managers you have found. Using a field of potential investments as a core and then simulating (Monte Carlo) builds a distribution of potential returns. You can start with a single manager or a small bloc of managers. Selecting these managers en bloc is required only when starting this process for the first time, or occasionally to test the portfolio. Once the fund is established, each newly suggested investment is modeled into the portfolio to see if the probabilities improve or not. The fund can test whether or not replacing a manager is helpful or if making a change to governance helps, or it can test a new type of asset or investment to see if it makes a difference, and if so, how much.

Businesslike decisions can be made. Implementing socially responsible investing rules is a business decision. Do social investing rules help or not? If they don’t, how much in terms of risk will it cost to use them anyway? How much will it cost to dedicate a portion of the fund to short-term bonds in order to pay for a particular need? For example: one foundation wanted to exclude sin stocks. After evaluating, they discovered that decision was worth $1,000,000 annually in potential returns, so since that million would feed and house many children in the orphanages the foundation supported, they decided that it was just fine that smokers and drinkers supported these activities. A note here: this was done with the managers available at that time, the market at that time, and within their governance set. It may or may not be different now and with different governance. A second example: one hospital made large changes to their portfolio at the end of a bad year in the market—against advice—which cost them $87 million in the first year. An “expensive good night’s sleep,” to quote someone there. They did not test beforehand.

Simulation is not a black box approach. Will it give you precisely accurate numbers? Don’t count on it. What it will do is get you close to the magnitude of the change and accurately give the direction of the change. Is there much difference in saying that there is a 95% vs. 96% chance of success? No, but there is a big difference if the numbers are 90% vs. 97%. It is a big flag if they move either up or down with whatever change was contemplated? In the hospital example above, the predicted cost, had they tested beforehand, was$60 million, less than the $87 million it actually cost. Either way, an expensive decision. Would they have made the same decision if they had examined the cost in advance? No one can ever know for sure.

We like a step-by-step approach because you can monitor the process and see where changes are coming from. Assuming you are starting from scratch, that process would look like this:

Core

To make it easier and quicker, we generally start with a small select group of listed equity managers. They are generally domestic, but on occasion may include a global listed equity manager(s) as well; you can start with just one single manager; the process is exactly the same.

Once you have identified your base group of managers, you can simulate possible returns and get a cumulative probability set or a distribution of returns. This set of probabilities tells us not what will happen but what the chances of a certain return or more happening will be, on average. This is a lot like rolling a pair of dice; we don’t know exactly the next roll, but do know that the probability of rolling more than a three is high (in fact, it is 33 out of 36, or 91.7%). Once we have this return set with the managers, we can simply choose the return we are interested in, our minimum required return, (for a foundation this might equal spending + CPI + maybe something extra), and read the probability of achieving that return or better. Changing managers or changing allocated amounts will change this probability, which will in turn change our risk of failure. You can try out various options to find the optimal one. It will be the one with the highest chance of success and lowest chance for failure. If you started with a few managers, you may change them out one at a time with different potential managers to see which would be the better set. This approach gives you a core or base to your portfolio with a certain level of risk.

Absolute

Next, begin adding other managers and investments that tend to skew the distribution further or make it more peaked, in order to continue to lower the risk of failure (increase the chance of success). Always, always ensure that the managers add diversification to the portfolio. Remember, diversify along three axes: (1) information stream, (2) style of conversion of information to return, and (3) risks. For this phase, you must find investments that have a more regular return and stray little from their mean, like fixed income, or investments that act like fixed income. These investments may have lower returns than your average, maybe even lower than your required return (though high-return options are certainly welcome and can sometimes be found). The steadiness of these investments helps especially during those times that are outsized on the low side, but then may hurt the portfolio when returns are outsized on the high side. You can think of this as giving up return to buy insurance. This makes sense as long as the added investment(s) lower the probability of failure. There is no sense in adding a large contingent of bonds (or any other investment) to the portfolio if the chance of reaching the required return goes down. Keep adding more and more of that particular manager, pushing down the risk, until the risk starts to rise. You are done; the investment or manager is properly sized.

This is the biggest fault of pension plans (especially public ones). They like to add very low returns in order to get very low volatility. The fund needs to make a certain required absolute return in order to pay benefits, and they end up carrying so much fixed income or low returning hedge funds in order to ensure low volatility, they never get to the required return, which is a failure.

One state pension plan has a statutory requirement of 30% for fixed income. How can legislators know in advance the right amount of bonds? With ten-year bonds in the 2-3% range, right now they should not be surprised that they can’t meet their obligations. What do they tell employees? “You can’t have your retirement payment, but look. . .no volatility!”

Enhanced

The next set of managers or investments to add are the long-ball hitters, the expanded or enhanced return generators. Add these like you did the steady eddies and pull the return distribution back up, increasing returns but still decreasing the probability of failure. Add as much as you can until you can no longer increase the chances for a win.

Yes, these investments are much more risky individually, but adding a bit into the mix can, and does, reduce the risk of failure.

This approach gives you a method of evaluating the “goodness” of a manager or allocation. Good for you, that is good for the portfolio, means this manager or this investment or this allocation is a positive influence on the portfolio by increasing the chance for success. At exactly the same time as you are evaluating managers, you are sizing the allocation and adjusting risk in the portfolio. A great manager may well have great track record, but if he does not help the portfolio, why would you include him?

A special note here: using this approach means using managers with skewed and peaked return distributions, such that there are many more good things happening in the portfolio than bad. This most often eliminates CTAs (commodity traders) because they have many more bad returns than good, even if the occasional great return pulls up their average returns. This is often their marketing cry: that they are not correlated with other investments, and using them will lower volatility. We are maximizing the probability of success, not minimizing volatility; checks are not written in deviations, standard or not.

This approach gives the investment committee the ability to make knowledgeable businesslike decisions. They can evaluate the cost in terms of risk or dollars in making or not making a certain investment. One could look at the cost of using or not using hedge funds or international stocks or almost anything.

Event Risks

The next step is to look for outside risks, those known or unknown elements that, if they happen, would have a big negative effect on the portfolio. Then find investments that insure against that risk. (The word “hedge” used to be used, but now it conjures up different images, mostly of excess fees). For example: in a geo-politically unstable world, oil might be good insurance (so, perhaps, is the dollar). In an environment with rapidly growing economies, investments in emerging countries and natural resources are a good hedge to events. In an era of inflation, real estate is often a good bet, but so are natural resources. The threat of down markets points to investments that are not marked to market frequently, such as private equity or perhaps even cash.

As you identify these event risks, look to your portfolio to see if you already have some investments that will backstop or insure against that risk. If not, add a manager or investment that will. Using the same approach as above, you can see if there is a cost in risk or potential return and balance it against the insurance you hope to get from the investment. You may find that increasing your risk is necessary to insure against whatever event is of concern.

Liquidity

The next step is to layer on any liquidity requirements, spending, benefit payments, and the like. It is important to remember that the entire equity portfolio is liquid; you do not need to match income or dividends with expenses (we are investors, not accountants). Trying to match income and expenses is not a best practice, since it reduces your total return. It is not uncommon to think, since equity markets are both up and down, that in the future you may need money when the market might be down, and therefore cash needs to be set aside now. That is not always the case. If you are earning, for example, 10% annually in the portfolio, even if the portfolio is down several percentage points, the resultant return may still be better than a money market over the same period of time. The comparison is of total dollars earned over the time period. One technique used heavily by corporate America, but not often by foundations or endowments, is the letter of credit.

The endowment model so often discussed includes the notion of selling liquidity. If your fund invests too heavily in private deals and has a liquidity crisis during a big sell-off of the equity markets, which happened in 2008, a standby source of short-term funds is most welcome and demonstrates superior prior planning. Why foundation CFOs and treasurers don’t use letters of credit more often has always been a wonder to me; they should certainly be considered.

Tilt

As things change, or we think they might change, in the short run we could tilt the portfolio slightly to take advantage of the opportunity, or to avoid the risk. As we artificially increase some or decrease other investments, we can go back to the simulation process to measure any increased risk, which helps us gauge just how much tilt to put into place. An important note to remember: a tilt need not be made, and in fact we recommend few and generally only small tilts at that.

This is partly philosophical. If you were so good at picking (anything), you would do it and forgo a well-balanced or managed portfolio. You are not, and therefore should hold a position near the managed portfolio. This core or baseline portfolio is the answer to the question, “What do we do when we don’t know what to do?”

The Whole

Taken step by step, this seems like a linear process—one is done, then the next, but in fact, all are done together, or more accurately, a little of one then a little of the next, and so on. This recursive rework of the portfolio is to find the optimal set of managers and allocations. Remember that, while this process seems incredibly precise, it is just not accurate. Therefore your “allocations” are often round numbers, and sometimes tied to minimum investment amounts.

The portfolio is now “allocated,” at least for the moment, until something changes.

Conclusion

Gather the best managers you can find who have the return distributions that you need, and then, using simulation, allocate among them to find a sufficiently high return at a low level of risk. Then find asset classes and investments that will address event risks and opportunities and add them in, using the same simulation method. What matters is the return needed at a desired risk level, not whether or not a popular asset class is included or even if all classes are included. This is important: Not all asset classes or even the popular ones or newest ones need to be in your portfolio; just the ones that help it. You are solving a problem, not sampling every dish on the menu.

Allocation is under continuous review as conditions change in the economy or markets, and is entirely driven by your needs and perceived future risks. Allocations are (or should be) adjusted to control risk, and risk is forecasted (ex-ante), not reported (ex-post), as with standard deviation. Rebalancing is continual, but should not be done without thought. Letting winners run for a while is not a bad thing for the portfolio (the extent to which you let them run is a policy and governance issue) and always maintaining a stable risk profile is imperative.

When you are finished, the results are a distribution with your minimum required return marked, along with the probability of underperforming. In the end, this is the number you want to control. It represents the risk in the portfolio. Figure 16-1 is a real life example.