Asset Allocation in a Crisis: Ways to Reduce Risk and Add Return

By Brian J. Jacobsen, Ph.D., CFA, CFP^®

Brian J. Jacobsen, Ph.D., CFA, CFP^®, is a capital markets strategist with Wells Fargo Funds Management, LLC in Menomonee Falls, Wisconsin. Contact him at Brian.Jacobsen@wellsfargo.com .


The market crisis of 2007 and 2008 shattered many investors' notions of the efficiency and stability of markets. It also likely shattered a lot of investors' confidence in asset allocation schemes. Some advisers thought they were following sound strategies, only to have severely adverse results. In our view, the financial crisis of 2007 and 2008 created a crisis in asset allocation. However, we believe that it was the implementation, and not the essence, of asset allocation that failed. 

In our assessment, the three main culprits were: 

1. confusing static for strategic asset allocation;
2. using historical instead of prospective estimates as inputs into an asset allocation optimization; or,
3. assuming risk management is all about size and style while ignoring sectors, countries, liquidity, and other factors.

This is neither the first nor the last time that a market dislocation will occur, which is why it's so important for advisers to follow a sound asset allocation strategy to help reduce risk and add value to their clients' portfolios.

Fundamentals of Asset Allocation: Ways to Reduce Risk and Add Return

When evaluating the merits of asset allocation, it is important to recognize that not only does the initial allocation matter, but it also matters how-and how often-an investor changes the allocation. It would be very easy to demonstrate how asset allocation outperforms a market index, such as the S&P 500, but the market experience of 2007 and 2008 was so extreme that many investors may have lost all confidence in stocks (The S&P 500 went from a close of 1,561.80 on October 12, 2007 to a close of 676.53 on March 9, 2009.) Rather than take a simple approach of showing how asset allocation might have limited losses in such a scenario, we will describe how even qualitative approaches to dynamic asset allocation can benefit investors in a crisis situation.

In any asset allocation plan, the first step is to determine the investment time horizon. This will be influenced not only by the investors' ultimate goal, such as retirement, but also by the amount of cash on hand or the ability to meet funding needs. For a person in retirement, a spending account-typically in the form of money market account holdings-is a needed source of funding for immediate and incipient expenses. For someone with other sources of income-for example, a job in which a periodic paycheck is the source of immediate and near-term funding-the same amount of cash in an account may not be necessary, but the source of income is the analog to the retiree's spending account. An investor with a job will also likely want to hold cash balances to account for any possible job loss, but that is to serve as a form of insurance instead of as part of an investment strategy.

For retirees, the amount of cash is often taken as an exogenous variable-sometimes kept at a minimum to keep the client "fully invested," but we believe that it is important to manage the proportion of a portfolio held in cash. The cash buffer serves as the measure of the minimum length of time the investor can wait for markets to rebound after a downturn before needing to liquidate equity and fixed-income holdings to replenish the cash buffer.

Investors with additional sources of income can be fully invested, but an adviser must still set an expectation to what a typical minimum holding period should be in order to keep clients from panicking over shorter-term market fluctuations. If a client has six months of expenses in cash or sources of funding, then that investor needs to be concerned about returns and risks that are over a six-month horizon. If an investor has one month of cash on hand or sources of funding for only one month of expenses (e.g., one month from retirement without any additional cash savings), then that necessitates a focus on shorter-term market fluctuations.   

Even a small difference in measurement period changes the inputs into a portfolio optimization program, because returns and risks do not scale up linearly. If returns were independent across time-that is, if a 12-month return were simply 12 times a one-month return-then you could use one set of monthly returns for each client. But such is not the case. An easy way to demonstrate the challenges of picking inputs to a portfolio optimization program is to look at the returns and risks of an index over different holding periods. For the S&P 500, the highest ratio of average return to standard deviation of returns occurs with a two month holding period (see Figure 1). The "correct" inputs to a portfolio optimization will depend on the client. How much cash does the client have to withstand market downturns, and what is the client's need for liquidity?

Figure 1
Ratio of Average Return to Standard Deviation of Returns for S&P 500 over Different Holding Periods

An adviser can use different holding periods to determine what risk-return parameters are most attractive to, and appropriate for, the client in determining the cash allocation. This method makes the cash balance an endogenous variable, where the percent of the portfolio to be held in cash is selected to coincide with the most attractive and appropriate risk-return characteristics of the equity and fixed-income portions of the portfolio.

Mistake No. 1: Confusing Static for Strategic Asset Allocation

Once an adviser determines the measurement period-which for an investor without additional sources of income determines cash holdings-the next step is to define the client's investable universe, which may change over time. A typical investor will want access to all types of securities. However, no investor is typical; thus, a client will want to hold a portfolio that is customized to his or her needs and preferences. A portfolio will consist of at least two types of risk: active risk and systematic risk. Systematic risk is risk that comes from a passive portfolio. Active risk comes from altering the composition of a portfolio. 

We believe that active risk is inseparable from systematic risk. If you measure the active risk of a manager by his or her tracking error, this tracking error will likely depend on whether markets are up or down. For example, the Wells Fargo Advantage Common Stock Fund has a daily tracking error of 0.36 percent in down markets (i.e., when the Russell 2500 had a negative daily return), but a daily tracking error of 0.27 percent in up markets. If active risk (which is risk added by an active manager) was separable from systematic risk (represented by broad exposure to a benchmark), then the down market and up market tracking error would be nearly identical.   

Most asset allocation models use broad market indices to determine the asset allocation, and then it is assumed that active managers only introduce active risk to the portfolio. However, this approach will likely result in an inadvertent addition of systematic risk to a portfolio compared to an optimization using active managers instead of passive indices. The adviser must evaluate active risk and systematic risk jointly and not separately. All sources of systematic risk and active risk need to be monitored and evaluated to make sure they are consistent with the client's risk preferences. 

For example, domestic equity exposure could be modeled by using three asset classes represented by the Russell Top 200, the Russell Mid Cap Index, and the Russell Small Cap Index. Using the historical risk-return numbers, as September 30, 2007, with monthly returns would have given the following descriptive statistics as inputs to an optimization tool:

Table 1
Average and Standard Deviation (annualized) of Monthly Returns for Indices, from April 1999 to September 2007 (Source: Zephyr Allocation Advisor)

Index Name	Average Monthly Return (annualized)	Standard Deviation of Monthly Returns (annualized)
Russell Top 200	13.23%	14.79%
Russell Midcap	15.57%	16.02%
Russell Small Cap Completeness	10.12%	19.09%

Table 2
Correlations of Monthly Returns for Indices, from April 1999 to September 2007 (Source: Zephyr Allocation Advisor)

	Russell Top 200	Russell Midcap	Russell Small Cap Completeness
Russell Top 200	1.00	0.8970	0.7758
Russell Midcap	0.8970	1.00	0.9473
Russell Small Cap Completeness	0.7758	0.9473	1.00

A portfolio with a target risk of 14.76 percent (the minimum risk portfolio, assuming non-negative investments in each index), would have been 12.4 percent in the Russell Top 200 Index and 87.6 percent in the Russell Midcap Index. The average return for this portfolio was 13.52 percent.

If, instead, the Wells Fargo Advantage Small Cap Opportunities Fund was used as a replacement for the Russell Small Cap Completeness Index, then the same target risk (14.76 percent) would have been achieved with a 37.4 percent allocation to the Russell Top 200, 30.8 percent allocation to the Russell Midcap, and 31.8 percent allocation to the Small Cap Opportunities Fund. The portfolio's average return would have also increased from 13.52 percent to 14.40 percent. This partially reflects the fact that active risk and systematic risk are not necessarily separable.

Strategic Allocations Need To Be Dynamic, Not Static

After the client's investable universe has been determined and an initial allocation made among passive investments, active managers, and any other strategies, the adviser needs to update the allocation periodically. The client's appropriate allocation-even amongst cash, equities, and bonds-will change with time. One common rule is that as an individual gets older, he or she should allocate more assets to bonds. The rationale is that investors early in their careers carry a lot of wealth in the form of human capital, or future income potential. Because work income typically occurs periodically and predictably, much like a bond, as investors approach retirement, their human capital begins to mature. To keep a constant total allocation to bonds (including bond-like human capital), equity, and cash, as investors age they should allocate more of their financial capital to bonds.

From our perspective, we believe this can justify a greater emphasis on active management strategies as an investor ages, as well as bonds. Human capital-in the form of income-has option-like qualities. An option is a right, but not an obligation, to engage in an activity. A person can choose to change jobs or to change the retirement date. But an option has time decay, in which the value of the option declines rapidly as the option expiration date approaches. Human capital has a similar time-decay; as people age, it becomes more difficult to exercise the option of switching jobs or planning to extend or contract the work life. 

Active management strategies have option-like characteristics because the manager has the choice to invest a portion of the portfolio outside the passive index's investment universe. The manager also has the ability to change the allocation to different portfolio holdings or engage in active trading strategies. These options mimic an investor's options in changing jobs or extending/contracting employment. As investors approach retirement, not only should their allocation to bonds increase, but perhaps their allocation to active management strategies as well. In fact, an allocation to active strategies that introduce more active risk to a portfolio might well substitute for some of the bond allocation.

Mistake No. 2: Using Historical Instead of Prospective Estimates as Inputs into an Asset Allocation Optimization

A client's asset allocation should be dynamically determined, not just as a function of the client's age, but also as a function of the economic and market environment. Some investors choose a static allocation or stick to a disciplined rule for asset allocation, but we believe there is actionable information contained in economic and market data that can inform tactical shifts in allocations. 

Client-specific information should inform the long-run allocation, especially which securities are included in the investment universe. Conservative investors value predictable cash flows and may therefore prefer bonds, preferred stocks, and dividend-paying stocks. Aggressive investors tend to place a greater value on future capital appreciation than on predictable cash flows.

For the fixed-income sleeve of an asset allocation model, advisers typically focus on managing credit exposure (credit quality) and duration. However, advisers should also be sensitive to the yield curve and the risk curve. If the yield curve is currently flat, for example, it can be expected to eventually steepen. An upward-sloping yield curve is called "normal" because, normally, the longer the term to maturity, the higher the yield to maturity. The time it takes to return to normal can be very difficult to predict, though.

To illustrate how investors can use yield curve data to make tactical decisions, as of October 12, 2007, the Treasury yield curve had a peculiar bulge (see Figure 2); the one-month Treasury yield was 4.092 percent, the six-month Treasury yield was 4.301 percent, the three-year Treasury yield was 4.234 percent, and the 30-year Treasury yield was 4.902 percent.  

Figure 2
Treasury Yield Curve as of October 12, 2007 and October 12, 2008, for Comparison Purposes

The fact that the one-month and the 30-year Treasury were within one percentage point of each other suggested that future growth was likely to be anemic, as a flat yield curve is historically associated with a low-growth environment. An inverted yield curve, where short-term rates are higher than long-term rates, is commonly considered a predictor of an economic downturn. 

From September 2007 to October 2007, rates had risen by nearly 70 basis points on the short end of the curve, which suggested that an economic downturn was imminent.

The yield-curve quickly steepened, with three-month Treasury rates going from 3.433 percent in December 2007 to 1.618 percent in March 2008. This readjustment provided a tremendous opportunity for investors to tactically shift from equities into Treasuries.

There is also typically a hierarchy of yields among the different types of bonds: tax-free municipal yields have the lowest yield (due to the tax advantage and generally high quality of municipal debt), and high-yield bonds (aka junk bonds) are the highest yielding. When these yields compress, or the ranking of yields changes, there may be mis-pricings in the market, giving an adviser the chance to add value to a portfolio.

To illustrate, Figure 3 shows the ratio of the Barclays Capital U.S. High Yield-Corporate yield to the Barclays Capital Global U.S. Treasury yield, along with the difference in yields. We look at both the ratio of the yields and the differences, because the ratio tends to be more stable over time, while the difference tends to provide more information when the Treasury yields get very low. For example, the average ratio from January 1995 to October 2007 was 2.098, with a standard deviation of 0.662. The average difference was 4.935, but with a standard deviation of 2.009. The coefficient of variation of each of these suggests that the ratio is more stable than the difference. That is why we tend to prefer looking at the ratio of yields instead of the differences of yields. After October 2007, the Treasury index yield dropped from 4.23 percent to a low of 1.55 percent in December 2008. At such low levels, yield ratios tend to get distorted, and that is why we look at both ratios and differences for identifying mis-pricings.

Figure 3
Ratio of and Difference Between High Yield Index and Treasury Index Yields

The fact that corporate yields relative to Treasury yields spiked as the economic downturn intensified suggested that investors with a long-term outlook-perhaps without a need for liquidity for the next few years-could purchase debt that was being sold at fire-sale prices. Clearly, the risk was that any given issuer of corporate debt would default, rendering the investment worthless. It was difficult-if not impossible-to ascertain ahead of time which companies would survive and which would fail. This led to the "lemons problem," in which all securities from corporate issuers were lumped together as junk. The situation provided a good argument for a diversified-as opposed to a concentrated-approach to asset allocation. It could also be used as a justification for investing with fixed-income managers who conduct their own credit analysis instead of simply relying on credit rating agencies.

Another measure that can help inform whether an investor is being adequately compensated for risk is the "Days Shares Outstanding" (DSO) for various indices (see Figure 4 for a quarterly measure on the S&P 500). The DSO is a measure of the average holding period of market participants, calculated by comparing the dollar value of daily trading for components of an index to the overall market capitalization of the index. The DSO is also a measure of the liquidity, or the churning, of the markets. As the DSO decreases, it suggests that market participants are becoming more myopic and concerned about short-term events. An investor who has the willingness and ability to take a longer-term view can buy liquidity when it is cheap (low DSO) and sell when it is expensive (high DSO).

Figure 4
Number of Quarters for the Dollar Value of Trading to Equal the Market Capitalization of 500 Selected Stocks

Paying attention to various "fear" or "liquidity" measures like the yield ratios, yield differences, VIX (volatility index), and days shares outstanding, can help identify cyclical opportunities in both fixed-income and equity markets. The assumption is that these measures have long-term tendencies, and short-term deviations create tactical opportunities to reduce risk and add value to a portfolio. The danger is that a temporary deviation is actually a structural change in the markets and there is, in fact, no trading opportunity. This is where it is important to understand not only the mechanics of the markets, but the politics and economics of the time. If there is a regime change such that empirical tendencies are no longer relevant, being early to recognize the regime change can create an opportunity to add value through that awareness.

Mistake No. 3: Assuming Risk Management Is All About Size and Style While Ignoring Sectors, Countries, Liquidity, and Other Factors

It's not uncommon for asset allocation recommendations for the equity sleeve of a portfolio to focus on exposure to certain market capitalizations or market styles. Such an approach can ignore other important exposures. For example, as of September 30, 2007, value indices tended to be dominated by financial companies, while growth indices were dominated by information technology companies (see Figure 5 for sector comparisons of growth and value indices). If an asset allocation recommended a balanced exposure to value and growth, there was the risk that the so-called "balanced allocation" would have a sector imbalance: favoring financials and information technology stocks over other sectors. In addition to controlling exposure to size and style, a portfolio needs to have controlled exposure to sectors.

Figure 5
Sector Exposures of Style Indices as of September 30, 2007

Another common error is over-emphasizing the importance of countries compared to sectors. Up to the end of the 1990s, it was common-and reasonable-to think that country-specific risk was greater than sector-specific risk. In the current era of trade liberalization and capital market integration, however, the notion of country is of less significance to investing than in the past. Some of the notions of dividing the world into the United States and "them," are quaint, but not very helpful for asset allocation. True, the United States does retain its status as the bastion of democracy, but its role as the source of economic growth and investment opportunities has declined. The fate of U.S. businesses is not necessarily tied to the health of U.S. consumers either. In 2008, according to Standard & Poor's, nearly 50 percent of S&P 500 companies' revenues came from non-U.S. sources.

While country exposure is secondary to sector considerations, it does remain important because of currency risk. A U.S. investor buying stock in a foreign company is not only exposed to the risk of the company, but to the risk of the currency. This risk can be hedged by purchasing futures contracts or exchange-traded funds that move inversely with the currency, but it is an additional source of risk that a financial adviser will need to articulate to the investor and to manage.  

There are many other risk factors that should be measured and managed, such as political risk, which is especially relevant for international investing. This is one reason we prefer active managers for international investments, because managers have the ability to react quickly and-perhaps-preemptively to political developments (e.g., Venezuela nationalizing banks and oil companies). Even in the absence of such dramatic events, businesses operate in an ever-evolving political and regulatory environment. Active managers who can change portfolio compositions quickly are in a better position to respond to developments than a passive index in which index constituents are updated quarterly, yearly, or as needed by a committee.

Conclusion

Asset allocation is in a crisis. However, we have demonstrated ways that advisers can use asset allocation to reduce risk and add value to a portfolio throughout a crisis. The key is to not become complacent with asset allocation.

The basic sources of failure were:

1. confusing static for strategic asset allocation;
2. using historical instead of prospective estimates as inputs into an asset allocation optimization; or
3. focusing on style and size while ignoring sectors.

True asset allocation is dynamic, it is prospective, and it manages risks that extend beyond simple style and size exposure. These are simple things an adviser can do to reduce risk and add value to client portfolios.