Alternative Asset Allocation: Tricks and Treats

The trick to adding alternative assets to your asset allocation is to utilize a rigorous process. This avoids the scary scenario where you have the wrong assets in your allocation or the wrong weightings on the assets. With the right approach, your treat is getting the optimal asset allocation relative to your risk profile and your needs.

Secrets from the crypt

The first step in looking at whether to add alternative assets to your allocation is to calculate correlations. You can build a less risky (lower volatility of return) allocation by combining assets whose returns are not highly correlated. A classic example is the standard 60% stock/40% bond portfolio. If stock prices go down, bond prices normally don’t go down as much. Prices of safer bonds such as those issued by the Government of Canada may even go up. If stocks go down significantly you can rebalance back by selling some of the bonds and buying stocks. Table 1 shows the correlations of quarterly returns of bonds, stocks, and two alternative assets (private equity and private debt)[1]. I used desmoothed returns on the alternative assets as described in Alternative Reality: Return and Risk.

Table 1

What do these numbers mean? Looking at the Canadian Stocks column, the correlation between Canadian stocks and Canadian bonds is a chilling 7%. That means the direction and magnitude of Canadian bond returns have almost no relationship to the returns on Canadian stocks. On the other hand, Canadian stocks are highly correlated with other stocks. The correlation is a ghoulish 75% with U.S. stocks and 72% with international stocks. That means about three quarters of the time the returns are in sync. The alternative assets are less in sync. The correlation between Canadian stocks and private equity drops to 49% and falls further to 25% with private debt.

Eerie Expectations

The next step in the asset allocation process is to create forecasts for the return and risk of the assets. Chart 1 below is an example of a conservative forecast for 10 year returns and risk (standard deviation of return). There is no magic to the return and risk numbers. They are only an example of one forecast among many that can be considered. The labels identify the asset and the forecast return. The returns are conservative, particularly for the U.S., given the higher valuation levels (stocks prices to earnings) for equity markets. The standard deviation is based on the last ten years of quarterly returns for bonds and stocks and the desmoothed returns for the alternatives.

Chart 1

Mystical Math

The final step is to create the asset allocation with the well-known method (at least to finance geeks) of mean-variance optimization from Modern Portfolio Theory. The mean refers to the expected returns. These returns are assumed to be the mean of a normal distribution of the expected returns for each asset. The variance is the distribution or volatility of returns around that mean. Standard deviation, the risk measure I am using, is the square root of the variance. While we won’t get into the debate here, assuming returns have a normal distribution is a simplifying assumption.

I used the mean-variance optimization process to assign weights to the assets. Given that the assets have various degrees of correlation, the optimizer can derive the asset allocation with the lowest risk for a given level of return, or the highest return for a given level of risk. It does this by combining assets with lower correlations into the portfolio if their expected return versus their risk is attractive.

Otherworldly Optimization

The challenge with alternatives is that we don’t know their true return volatility. The reported returns are based on estimated valuations, which artificially smooth returns and therefore standard deviations. To address this, I use the standard deviations derived from their desmoothed returns.

Even after desmoothing some alternatives have low volatilities compared to their estimated returns. If I did an unconstrained optimization, I would get an unrealistically high allocation to alternatives. I say unrealistic because there are other risks, relative to bonds and stocks, which are not captured by standard deviation alone. Table 2 shows the risks of alternative asset funds that are not present in typical bond and stock funds.

Table 2

More realistic optimization results are achieved by adding constraints, such as a maximum weight in an asset. One way to address the uncertain volatilities of alternatives and their unique risks is to use a “risk bucket” approach. Given an investor’s risk profile and their income and liquidity needs, they can fall into different ranges or buckets in terms of their target allocation to alternatives. I say target because most alternative funds require a capital commitment up front. That commitment is drawn down over a period of years as the fund makes its investments. There is uncertainty about how long it will take to invest the commitment and when the fund will make distributions. This means that the actual percentage allocation to alternatives will usually be higher or lower than the target.

Spine Tingling Scenarios

Chart 2 shows the results of the mean-variance optimization process, using the data from Table 1 and Chart 1. Each curve plots the expected return and risk for various asset allocations with 0%, 15%, or 30% alternatives. To get more realistic results, I added a constraint that the maximum amount in Canadian, U.S. or international stocks could not exceed 40% of total stocks.

The orange curve plots the asset allocations with the maximum return for a given level of risk (standard deviation), assuming no alternative assets. The labels for each dot show the combined total in Canadian, U.S. and international stocks (the remainder would be in Canadian bonds) and the expected return for that asset allocation. The blue curve shows the effect of adding a risk bucket of 15% alternatives assets (private equity and private debt) and the green curve is for 30% alternatives. The labels for each dot on the blue and green curves show the expected returns at that asset allocation.

Chart 2

Given the assumptions and constraints, asset allocations with alternatives have more expected return for a given level of risk or lower risk for a given level of expected return. As an example, let’s look at the 60% equity risk portfolio near the middle of the orange curve. The optimizer has chosen an asset allocation of 60% in stocks (24% Canadian, 12% U.S. and 24% international) and 40% in bonds. The expected return is 4.1%, and the risk is 7.3%. As the blue line shows, if we add 15% alternatives to the allocation, we get an expected return of 4.8% for the same risk of 7.3%. The allocation is 50% stocks (24%  Canadian, 2% U.S. and 24% International), 35% bonds, 10% private equity and 5% private debt. Going to the green line with 30% alternatives, the expected return goes up to 5.3% with an allocation of 44% stocks (22% Canadian and 22% international), 26% bonds, 11% private equity and 19% private debt for a risk of 7.3%.

Sweet Summary

The complexity of alternative assets requires more analysis to determine what weight they should have in your asset allocation. One way to do this is through an optimization process. While it isn’t a perfect methodology, it does allow you to see what asset allocations look like for various levels of risk. You can then choose an allocation that aligns with your risk profile and liquidity requirements. This gives you the treat of potentially higher expected returns with the same risk, or lower risk with the same expected return, compared to your current asset allocation.

Invest Wisely,

Dave Schaffner, CFA

Principal, Wayfairer Capital Management Ltd.

[1] Quarterly returns over the ten years 2011-2020 are for the following indices: FTSE TMX Canada Universe Bond, S&P TSX Capped Composite, Russell 3000, MSCI EAFE, PitchBook North American Private Equity and PitchBook North American Private Debt.