A qualitative perspective on volatility as “Risk”
In part 1 of this white paper, we examined some of the quantitative problems with blindly accepting standard deviation as the quintessential definition of risk. There are also qualitative analyses which refute volatility as the sole embodiment of risk.
For example, imagine a retiree owns a bond ladder – a portfolio of individual bonds designed to generate a structured stream of interest payments for retirement income. The price of an individual bond can fluctuate for fundamental reasons; for example, an increasing chance of bankruptcy should decrease the price of a corporate bond. Let’s ignore fundamental worries and acknowledge a bond’s market value may go up and down for non-fundamental reasons as well. For example, an increase in interest rates will lower prices of a bond. Similarly, the market premium changes for bonds which pay interest on a particular schedule; for example, the market demand may be higher for bonds which pay interest more often than semi-annually. Like any other security, a bond simply may not reflect its intrinsic value due to simple market inefficiency or neglect. For these reasons, a bond’s market price moves – it experiences price volatility – but the payments due to bondholders may not be suspect and, as a result, the bondholder simply may not care. In this case, the retiree can watch the market price move for an individual bond, but there isn’t any material effect to the retiree. In other words, investors appreciate volatility risk differently depending on their goals.
The bond ladder example brings up another qualitative issue. We have been considering an individual investment’s risk and a portfolio’s risk as interchangeable since they are measured by standard deviation. Within the bond ladder example, the retiree does not absorb the risk. However, the same retirement may have a separate sleeve of return generating investments which are fairly evaluated against volatility risk.
Furthermore, we will see that different investment products and strategies may exist to protect a portfolio against different types or risk but may themselves exhibit high standard deviations. For example, a traditional investment product may try to generate positive risk adjusted returns relative to its benchmark, but a real return income fund may accept higher volatility in the pursuit of generating returns that exceed inflation. Including individual strategies that have high volatility might not be the worst action to a portfolio; asset optimization studies demonstrate that investments with high standard deviations can still reduce the total standard deviation of an efficient portfolio through the magic of low correlation and the subadditive features of portfolio construction. Besides, a savvy investor may be well served by a combination of investments that hedge against different types of risk. In other words, optimizing the disparate elements of your portfolio is often more complex than simply picking the most attractive volatility/return investment within each asset class.
Your primary measure of risk depends on your primary goal
So, what is risk, really? Like others before us, we suggest that risk is the possibility of not meeting your goals. To expand on this idea, we will consider how the goals of different entities influences the risks they face.
The most common goal for individual investors is to fund retirement. For them, risk may be best described by the possibility of insufficient returns to fund retirement during sensitive time periods. A 25 year old will only a small amount saved and they’ll have the benefit of time to compensate for any market downturns. By comparison, a 65 year old retiree is in a more precarious position. After a lifetime of saving, retirees typically have the highest net worth on their first day of their retirement, so they have highest sensitivity to unfortunately timed market corrections. Investors planning for retirement have a finite time horizon, their life expectancy, and a limited amount of time to recoup losses caused by market pullbacks. The lost decade of flat market returns in the 2000’s has, for example, impacted many retires in the baby boomer demographic. Individual investors may be better served by modeling techniques that are optimized to reduce a portfolio’s “Conditional Value at Risk” instead of simple standard deviation, particularly when they are nearing retirement.
Foundations, Endowments and other Charitable Organizations
It is said there are two basic drains to investment returns: taxes and inflation. Foundations and endowments, like individual investors, are return seekers. However, foundations and endowments have radically different sensitivity to both taxes and inflation. First, these eleemosynary have fewer incentives to pursue tax sensitive securities (like Municipal bonds) because their income is often tax exempt. However, to achieve the beneficial tax status, they must divest a substantial percentage (e.g. 5%) every year. Eleemosynaries have a minimum spending target to hit and they must invest more aggressively to reach that target. In the event of a market downturn, an individual retiree can adjust their spending habits downward; a foundation does not have that flexibility. Finally, a frugal retiree may have saved a substantial amount for their retirement during their career and require a nominal return; a foundation, existing in perpetuity, must always strive for a substantial return goal.
Regarding inflation, foundations and endowments have a higher time horizon than a typical retiree. A retiree at age 65 may need to prepare for 30 years of inflation, and the retiree’s need to hedge against inflation will grow smaller every year. Conversely, a foundation must account for a potentially limitless inflation given their infinite time horizon. Protecting the purchasing power of a foundation, given its high spending mandate, is usually the key goal.
Defined Contribution Plans
Defined Contribution plans have the burden of facilitating retirement options for their employees but they have limited control of investment selection. Given the limitations, Defined Contribution retirement committees should be sensitive to the unique challenges of their employees when selecting investments for the retirement plan lineup. For instance, an employer with a great deal of young employees may prefer an aggressive target date fund series which accounts for longevity risk (i.e. – not saving enough to adequately fund retirement). Conversely, an employer with a high proportion of employees nearing retirement may prefer a conservative target date fund series which is more sensitive to downside market risk.
Defined Benefit Plans
Imagine a Miracle investment that has equal chances of generating a 10% return or 60% return annually. The standard deviation of that investment would be high (+25%), but only a few real world investors would complain with these results.
This Miracle investment would be extremely attractive to a foundation, given their potentially infinite time horizon and high return goal. Conversely, this Miracle investment may have only limited benefit to a fully funded (+130% funded status), frozen, corporate pension plan. This pension’s goal is to satisfy their obligations to prior employees and unwind, not to exist in perpetuity. The pension would not be able to use the additional returns and the high volatility may actually impair their budgeting stability.
In fact, a Defined Benefit plan – whether it is set up for a Taft-Hartley Union, a corporation’s pension, or even the federal government’s Social Security scheme – exists to satisfy its liabilities. Unfunded liabilities, driven by interest rates and the legal requirements of the plan are key risks borne by these plans. These investors may be best served by a Liability Driven Investing approach, described in our October 2011 newsletter.
Alternative measures of Risk
A quick review of the previous page has suggested risks, other than traditional volatility, to consider such as taxes, inflation, longevity, interest rate risk, and so on.
The aforementioned Miracle Investment demonstrates another problem with standard deviation – not all volatility is equal. Most real world investors don’t mind if the investment returns are substantially positive, even if they are unpredictably volatile. Some market analysts advocate separating upside volatility (exceeding a target return) and downside volatility (achieving less than target return) and there are analytical measures which that reflect this attitude. For instance, the traditional measure of risk adjusted return within modern portfolio theory is the Sharpe Ratio - the excess return per unit of volatility. There is a variant to the Sharpe Ratio called the Sortino Ratio which measures excess return per unit of downside volatility. It is worth evaluating your portfolio in terms downside volatility risk instead of combined volatility risk.
Parsing upside risk and downside risk, although useful, still uses the quantitative framework established by classic definitions of standard volatility, but there are alternative types of risk which defy the classic definitions. For example, imagine you can purchase a long term fixed annuity with an insurance company. The insurance company promises you a 7% annual return, each and every year, so long as it is held for 10 years. An investor looks at a 10 year US Government bond with an anticipated yield of 2% and also notes that the bond has interest rate risk, which should result in some price volatility in the next decade. The investor reasons that the annuity is a much better deal since it exhibits moderate returns and no return volatility whatsoever. Is it fair to compare an annuity to traditional investments or to model it with an expected return of 7% and a standard deviation of 0%?
If an investment seems too good to be true, it almost certainly is. The traditional model does not accurately portray the annuity’s risk because the risk does not stem from standard deviation of the returns. To begin with, the annuity contract adds liquidity risk to the equation; recall that a 10 year treasury can be bought and sold at any time, but an annuity contract will come with stiff penalties for any early redemption. Liquidity risk demonstrated its importance during the 2008 financial crisis when several healthy businesses were essentially solvent and profitable, but unable to generate sufficient liquid assets to pay their short term debts in the cash-crunch.
Another risk borne by the annuity holder is counterparty risk: the risk that a financial institution will default on their contractual obligations. In the worst case scenario, your counterparty in a contract could be an outright fraud, like Bernie Madoff, that has no intention or ability to honor your agreement. On the other hand, unsound business decisions demonstrated weaknesses within the financial system and counterparty failures. Insurance companies, like AIG, came close to failure and may have become unable to honor their contracts without government bailouts. Money market funds required government insurance to maintain their ability to keep their net asset levels equal to the amounts invested (i.e. – not “breaking the buck”). Several businesses went bankrupt and changed their pension payment obligations to retirees. Each of these events represents an actual or potential change to the original terms of a contract, i.e. counterparty risk.
Another risk facing portfolios, particularly those dominated by fixed income instruments, is interest rate risk. Without going into detail, when interest rates go up, bond prices go down; our February 2012 article on the Federal Reserve, demonstrates how government monetary policy drives US interest rates. Globally, The European Central Bank and Federal Reserve have clearly had tremendous influence on the fixed income market in 2012, and that is expected to continue for the foreseeable future. Nobody can accurately separate central bank action (or inaction) from the fixed income market. Thus, the judgments of the Federal Open Market Committee determine if interest rate sensitive bonds (e.g. treasuries and longer duration bonds) will hold their value or not.
A related risk, inflation risk, is driven by the supply of money and the global demand for goods and services. Since inflation is a function of monetary policy and global aggregate actions, it is more complex to manage. The risk parallel to inflation, but sometimes ignored, is deflation. The global slowdown has particularly hit the European Union but the European Central Bank – chastened by the memory of hyperinflation within the Weimar Republic – had signaled resistance to injecting more money into the system. As a result, there was a very real possibility of a deflationary cycle, wherein prices fall, investors retain their cash, fewer goods and services are demanded, and prices fall even further in a never-ending spiral.
This list is by no means exhaustive. There are other types of risk to consider when designing a portfolio including concentration risk, style bias, beta driven risk (market risk), multi-manager risk (overlapping bets leading to unseen concentration or even offsetting their active positions), currency risk, leverage and so on.
One key risk: fiduciary responsibility
Institutional investors have an additional burden since they are trustees for the assets they control but do not own. As an example, imagine that you make investment decisions for a foundation. As part of your investment process, you had conducted an asset allocation study and determined a combination of investments that was designed to maximize the foundation’s long term investment return for an acceptable level of volatility. In other words, the foundation has selected an efficient portfolio.
In this white paper, we have promoted alternative measures of risk. Having absorbed the lessons of this whitepaper, you have decided to broaden your definition of risk beyond standard deviation; specifically, you increase your protections against long term inflation. However, by introduction new asset classes or different combinations, the new portfolio will move away from the previously established efficient frontier: those combinations of portfolios which maximize return for any given level of volatility.
To step away from the efficient frontier, at least two actions must be completed. First, you (or your investment consultant) should fine tune your optimization analysis. Imagine that the foundation has considered making a substantial allocation to precious metals as hedge against inflation. After extended analysis, you determine that a combination of TIPs, real estate, and diversified commodities (including metals) make an equally effective hedge against inflation, but add less volatility to your portfolio. Thoughtful analysis can minimize the distance from the updated portfolio and the established efficient frontier.
Second, the institution’s Investment Policy Statement should document the reasons for deviation from an efficient frontier. Document what steps are being taken, and why they were enacted. If, for example, the foundation adopts an inflation hedge, describe its purpose, investments and limits. For another example, if a pension fund moves to a Liability Driven Investing methodology, than the portfolio’s duration, allotment to fixed income and methodology should all be referenced in the IPS.
Accepting the limits of definable risk
Each investor, whether they are an individual or an institution, has unique goals. Since risk represents the possibility of not meeting a goal, it then follows risk is also unique for each investor. We have demonstrated how an appropriate investment for a foundation investor may be completely inappropriate for a pension plan, based upon their unique return goals and the risks they each face
On the other hand, having a common denominator, like standard deviation, establishes a baseline benchmark for any discussion of risk. Alternative types of risk are often hard to measure and compare. It is sometimes only possible to indirectly measure alternative types of risk; for example, the futures & options market for federal funds rates is used to infer central bank monetary action or intervention.
Challenging the definition of risk may remind investors of a quote from Winston Churchill, “Democracy is the worst form of government, except for all the others.” In a similar sense, standard deviation is not always the best measurement of risk for any investor, but it is a universally accepted and it is easily measured. Savvy consultants and investors may include standard deviation as a standard risk measure, so long as they understand its limitations and they are sensitive to the true risks which challenge their portfolios.