INVESTMENT RISKS

OVERVIEW

There is a chance you do not earn the profit you predict, you may lose all your money, or you may end up needing more than expected. These risks come from many causes but apply to every investment you make, no matter how basic. In all cases investors demand compensation for assuming risk, or else everyone would invest only in ‘safe’ investments. That compensation is usually the expectation of higher returns. It can also be the reduction of another type of risk. It can also be the love of your family, when they need financial help.

You need not compute complicated equations or look up meaningless metrics, but you must explicitly recognize and heuristically evaluate all risks before you invest. Too often people simply ignore the possibility of bad outcomes that are obvious from the start. It is far better to assume the worst, rather than assume the best.

Money is always flowing between markets. The changing risk of one type of asset will effect the market prices (and risk premium) of other asset types. You must always evaluate the ‘relative’ attractiveness of an asset given its risk. An example of this effect was in 2008 when the economic risk of all businesses all around the world exploded in the ‘credit crunch’. The ‘flight to safety’ resulted in the price of US treasuries rising sharply, even though their own risk profile had not improved (and actually worsened).

Instead of working from an evaluation of risk toward determining the appropriate return, retail investors often reverse the process. They judge an investment by its expected return. They think large dividends and low P/E’s mean the investment is a good one. In fact, these metrics indicate the investment is NOT a good one; that there are some impending problems; that markets are demanding extra returns for assuming that risk.

The expected return is just one point on a probability distribution of returns. When you assume risk you expect a higher return, but that is not guaranteed. As risk increases the width of possible outcomes also increases. With increasing risk come the possiblity of outcomes that are worse than for safer investments. For each of the following types of risks, you should understand their cause and effect; how they are measured; and how they are hedged.

REPAYMENT RISK (default risk, capital risk)

Most investments involve giving your money to someone else to do something with. You always run the risk that they abscond with your money, or that they go bust and cannot repay it. The US government is considered to have the lowest repayment risk, assuming they have the power to tax their subjects. Foreign governments are much more likely to default. They may even earn political points at home for doing so.

Businesses with majority owners pose the largest risk of default when things go wrong. The insiders will know, well ahead of public investors, when to start worrying. There are many, quite legal ways for them to recover their own money, leaving the company shell to the public. Once a business officially goes broke, laws determine who gets paid first. The tax man and employees rank high on this list. Debt holders come next, then Preferred Share owners. Equity owners come last, with the most risk of getting nothing.

Credit bureau’s exist to measure the probability of default for specific debt securities. They cover bonds, debentures, preferred shares and income trust units. They usually have access to more information than the retail investor, so their ratings should be consulted. Be clear that they rate only the probability of default, not the risk of price volatility. A security may lose value temporarily because of market sentiment — sentiment that may not be correct in the longer term. As the ABCP crisis has shown, the rating agencies can be co-opted by the businesses whose products they rate, because they are paid by those same businesses. The rating of each security signals to the market what interest rate its issuer should pay. See this website for US bond spreads for each rating.

Regulators sometimes provide insurance to cover default risk. American (FDIC) and Canadian (CDIC) regulators insure cash in bank accounts (to a limit). Both countries have another plan to insure your securities held by brokerages. Note that this does not insure the value of those securities. It insures that the brokerage will not walk away with your ‘ownership certificates’.

Sometimes investors can buy insurance to protect their invested capital. E.g. real estate is frequently insured by owners. Its small cost is a good trade off against the value lost in a fire. ‘Counterparty risk’ refers to the risk that the party on the other side of the derivative trade (e.g. insurance) cannot pay the claims they contracted to cover. The insurance you buy, is only as good as the reliability of the insurer to pay claims. During the 2008 credit crisis, it turned out that institutions were buying insurance (credit default swaps) from companies that could not pay the claim.

Principal Protected Notes (PPN) are a product offered by banks and insurance companies that allow you to participate (to an extent) in the risky stock markets while your initial investment is guaranteed to be repaid in (say) five or ten years. You must weigh the cost of the insurance against the default risk. Say the PPN allows only a 50 percent participation in the market return.

• The cost of PPN’s insurance is the value of the stock returns you give up. If you think the stock market will return about 10%, then the insurance is costing you 5% each year.
• Contrast that against the probability that the stock market will have dropped after a

5 year period (historically 11 times over 80 years) or

10 year period (historically 2 times over 75 years).

See the graph on Sheet 8 of this Excel file .

Another way to analyze the cost is to compare

• the expected return from T-bills of the US government (approximating inflation) to
• the half of the S&P’s expected return (after tax for both choices).

In 30 of the 42 5-year periods, the T-bill return was higher than 1/2 S&P’s return. (Sheet 5 of the same Excel file).

PRICE VOLATILITY

When used by academics or finance industry people, the term ‘risk’ almost always refers to the volatility (or variability) of the investment’s price (of its rate of return really). This is the sense in which the term ‘risk-reward tradeoff’ was coined. E.g. The yearly return from a US bond (safe) will fall within 8.3 percentage points above and below the average 4.8% return, 2/3 of the time. The yearly return from a US stock (risky) will fall within 19.9 percentage points above and below the average 10.1% return, 2/3 of the time.

Notice that the width of possible returns includes both the upside as well as the downside. Both are considered equally bad. Standard deviation (SD) is its mathematical measure. The math is based on a presumption that the probability of returns follow a ‘normal curve’ (remember the Bell-curve for school grades?). US equity returns show that kind of history. But other countries’ equity returns have had outlier results that are more extreme than predicted by a ‘normal curve’. They are said to have a ‘fat tail’.

The risk-return trade-off gets turned on its head in bear markets. The expected monthly returns are 2 percentage points lower than expected in a bull market, while the standard deviation is 50 percent greater. Of course you would have to identify the bear market in real time to change your behavior. This difference in risk may well explain why active managers argue it is much easier to outperform passive indexing in bear markets. Bull markets are a steady plod up the wall-of-worry, but bear markets have sudden and violent drops.

The risk-return trade-off also gets turned on its head when measuring actual returns within the asset class of common stocks. This is called ‘the low-volatility anomaly’. Numerous studies have shown that stocks with lower price volatility have generated higher (not lower) returns for investors. There are a number of ideas why this happens.

• It could be a ‘lottery effect’ where people will over-pay for a big payoff. Those with small portfolios have been shown to evidence this attitude toward risk.
• It could be that most investors do not have access to leverage to get the high level of risk they are willing to accept, and so pay more for high-volatility stocks.
• Since we are talking about using the past volatility of a stock to predict its future return, the failure of the risk-return trade-off may be because volatility reverses. Investors chasing hot stocks, or uncertain earnings, create high idiosyncratic volatility that reverses to correct for over-pricing, or when uncertainty gets resolved.
• It may be that institutional money managers’ compensation schemes promote risk-taking with large bonuses for excess returns.
• It could be that the market is inefficient. Investors may interpret the statistics to think they will get higher returns when they buy riskier stocks, so they are willing to pay more to purchase them — putting the cart before the horse.
• The results might only be measuring a predicted math effect, if you think prices grow arithmetically (which they don’t, of course, because Betas change). When the market’s volatility is high enough, the cumulative return for a portfolio with a Beta greater than 1 will be less than the market portfolio’s.

Risky stocks are supposed to frighten you into using a larger discount rate — in other words you are supposed to pay less for them, pay a lower valuation. Your profit is determined by the price you pay. When you pay less your profit is higher. Risky stocks seduce the market into paying higher valuations because they have media coverage and become ‘market darling’. What happens is the exact opposite of what is supposed to happen.

There are even more problems with the concept of a risk-return trade-off. Between asset classes Beta seems to explain different returns — in the long run. But between 1980 and 2012 (more than 30 years), as interest rates fell from double-digit to miniscule, returns from low-risk treasury bonds have outperformed risky stocks. All along the way, stock investors told themselves interest rates cannot possibly go lower but they did. The risk-return trade off is only in effect at t=0. Expectations do not always pan out.

But high standard deviation is not necessarily bad if an investment’s average return is high — if you get paid for assuming the risk. The Sharpe ratio measures excess return per unit of risk (as measured by SD). Another metric that measures a risk-adjusted return is the Coefficient of Variation. It measures risk per unit of return — SD divided by the mean rate of return. The lower this number is, the less risk per unit of return.

Standard Deviation also does not distinguish between upside and downside volatility. Investors have a problem ONLY with underperformance. The Sortino ratio is a modification of the Sharpe ratio that only penalizes an investment for downside deviation. Academics keep working. There is now a risk metric called ‘chi’ that measures the down-side-tail of a stock’s distribution curve vs. the down-side-tail of a benchmark. A low metric indicates lower probabilities of negative returns, and lower down-side risk, but also seems to predict higher nominal returns — the exact opposite of the high-risk, high-return doctrine.

A problem with the older metrics is the buried presumption. The context in which you hear them used is most often an attempt to dismiss someone’s exceptional portfolio returns. Oh sure, you outperformed the benchmark by 10%, but on a risk-adjusted basis you actually UNDERperformed. The presumption in that statement is that the additional volatility was unacceptable. But who says? The volatility may well have fallen within the owners’ tolerance.

The use of these risk-adjusted metrics is valid only if the investor changes his asset allocation according to the risk of his securities. If the investor has already decided on (say) 50:50 debt:equity, and will not change that AA no matter how risky his choice of securities within that equity allocation, then a metric that ‘normalizes’ returns serve no purpose. Most retail investors won’t change their AA. The use of these metrics assumes that AA is determined after the risks of the individual securities is considered — a higher allocation to equity if the securities are lower risk, and vice versa.

Volatility impacts the investor in five ways.

Investors react emotionally to swings in market prices. It is emotionally hard to NOT sell an investment that has lost value — you just cannot tolerate any deeper losses. The result is an emotional toll, as well as a possible loss of returns from selling at the bottom. Buying a basket of different stocks/bonds allows the volatility of one’s upswing to offset the other’s downswing. In total your portfolio suffers less volatility and is easier, emotionally, to maintain. How many different securities you need to get rid of the individual security’s risk is not agreed. 20 used to be the recommended number but it has increased. Some say you now need 50 securities because securities now move more closely in tandem. ( e.g. W. Bernstein ) Of course you will still be left with the volatility inherent in the asset class itself.

You can get rid of even more volatility by diversifying your portfolio’s asset classes — traditionally between stocks and bonds. Now real estate, commodities, etc are added to the mix of asset classes. Portfolio theory claims that you can add asset classes with lower expected returns without lowering the expected return of the portfolio, even as the addition lowers the portfolio’s volatility. This is all proven with dense math and backtesting. The Efficient Frontier is the best trade off you can get — the highest portfolio return for any given volatility. Many portfolios will lie below the line. But remember that this line is an average of historical returns. Individual decades have produced radically different looking charts. See also the warnings on the Backtesting page.

The measure of how closely a individual stock’s returns tracks that of the equities class is called the ‘beta coefficient’. The stock with a beta coefficient larger than one (or negative one) is riskier because its price swings wider than the asset class does. A beta between one and negative one means the volatility is less than the asset class. A negative beta coefficient means the asset’s price moves in the opposite direction to the market.

Since investors demand extra returns for assuming risk, they should incorporate beta into the discount rate used to value a specific company. Say 10-yr T-bonds yield 5% and you demand 9% from equity investments. The equity risk premium would be (9% — 5% =) 4%. If a stock’s beta is (say) 1.3, you can calculate a premium specific to the company equal to (4% * 1.3 =) 5.2%. The discount rate you would use equals the sum of the T-bond rate plus the specific equity risk premium (5% + 5.2% =) 10.2%.

You can find a stock’s beta on the Yahoo Finance site in the Key Statistics link, on the right under the heading Trading Info. You can also calculate Beta yourself using this spreadsheet’s longer tracking period .

One way to protect yourself from volatility is called ‘dollar cost averaging’. When buying into any position, you can divvy-up the cash into equal-dollar, multiple purchases over time. Each dollar will buy more shares when the price is low, and less when the price is high. Your resulting cost-per-unit will be higher if the market price marches upward, because you delayed the purchases, but lower if the price declines or is merely volatile around a flat line. The reality is that most people are enticed into the markets, or into a stock, after it has done well. In other words, they buy at the top (and sell at the bottom). Dollar-cost averaging will save you from this error.

The second reason to care about volatility is because many investors have a need to liquidate the investment at a specific time — e.g. to pay for college in 2 years. It is impossible to know in advance whether the security will be at its high price at that time, or at its low price. If cashed out at it’s low point, that loss gets crystallized, permanently damaging the portfolio’s value.

You can control this risk by choosing products with a certain cash flow. E.g. A bond ladder is set up so there are sufficient bonds maturing in each year to fund a retiree’s draws. The investor does not fret about any price fluctuation during the bond’s life because it is irrelevant. The redemption value at the time it will be needed is know in advance. That is all that matters.

The third reason to care about volatility is because may people invest to create a nestegg for retirement. Will I have enough? depends on the variable yearly returns your portfolio generates — which cannot be predicted. There is a common argument that volatility decreases as you expand the time horizon — that returns will even out or cancel out, over time.

The chart above is from Jeremy Siegel’s Stocks for the Long Run. It measures the variability of the AVERAGE return for the holding period. It shows that the longer the time span, the lower the variability — even for common stocks. This is very persuasive, but wrong. It measures the wrong attribute. a) Its measure is a math calculation in hind sight, but the investor must still suffer the anguish of price swings in real time. b) What really matters is the variability of the ending portfolio’s size — its cumulative return.

In fact time INCREASES the variability of the final portfolio’s value, as shown in the graph above. The spread of possible ending values gets larger and larger. Each different ending valuation represents the result of one distinct AVERAGE rate of return over the period. The ideas and graph here come from John Norstad. He admits that the distribution may not be quite so wide if the markets revert to the mean and show cyclicality of returns. (This is a pro-forma graph not actual historical results). The US experience has shown this, but it is not universal. When international stock indexes are added to the data, there is no reversion to the mean, or reduction of risk over longer periods of time (Philippe Jorion).

Notice how the graph has no results in negative territory. Its author presumed that once the principal is lost it is gone for good and that is the end of the story. But what if the investor uses leverage? What if the investor rebalances from debt securities regularly? Then the results would indeed flow into negative territory — giving an even wider final distribution.

But wait. Things get worse. (The following is from a paper in the Journal of Financial Planning Aug 2010, by Chittenden, Moon and Toles.) Norstad’s graph above reflects the resulting value of a lump-sum $$ amount invested at the beginning of the period. When a lump-sum investment is modeled, the sequence of returns experienced in each particular year does not matter. You can see why from the math discussed at Rates of Return. The following 20 year chart shows the value of $1 invested at the actual historical return for each year vs. its value if the path of returns was exactly reversed. Both values end up at the same place.

But in real life, savings are added each year to grow the portfolio — you need to model an annuity. (The flip side of this same issue is that withdrawals are staggered throughout retirement as the portfolio shrinks.) Now the sequence of yearly returns DOES make a difference.

The larger the volatility of yearly returns, the more important the sequence of returns becomes — the wider the range of possible final outcomes — even though all scenarios would report THE EXACT SAME AVERAGE ANNUAL RETURN for the period. The following chart shows the distribution of possible results from all the possible permutations of sequences of returns — 20 years of $12,000 savings (equaling $240,000 total) invested to earn an averaged 8.2% in stocks.

Each single point from Norstad’s distribution of final values, will itself be a distribution of possible values when savings are added gradually and the sequence of returns is unknown. The distribution of terminal values results from the volatility of yearly returns. The width of the curve is narrower for debt — wider for stocks. And a portfolio of mixed debt and equity is in-between.

The historical experience in the US markets falls within those expectations. The 20 year annuity terminal values range between $500,000 and $1,500,000 with an abnormal spike to $2,000,000. The 30 year annuities end up roughly the same. In both time frames the range of outcomes varies by a factor of three or four. The risks did not decrease with a longer time frame.

The less cyclical Canadian stock index has resulted in far less volatility in terminal values. The graph above comes from Sheets 24 and 25 of the data spreadsheet. Investors have no control at all over the specific sequence of returns the economy will produce in their lifetime. No matter how well they invest, their returns will be greatly determined by the overall market. Some generations do well, and others don’t.

The fourth reason to care about volatility. Instead of looking at volatility in a negative light you can see it as presenting opportunity. Your investment returns come mostly from the difference between the purchase and sale price. Volatility presents opportunities to buy cheaply and sell when overpriced. Exactly what you want. Of course the glass-half-empty crowd will call this ‘price risk’ — the risk of paying too much.

Remember that volatility measures both the downside and the upside. You WANT upside volatility — you want capital gains. The pro-dividend crowd uses the lower volatility of dividend stocks as a selling point. Don’t be overly impressed. When there are no dividends, the investor relies on price appreciation for ALL his profits. He WANTS a stock price that moves more than the general index.

ECONOMIC RISK

Market economies tend toward a cyclical boom/bust, as they correct, then overshoot, then correct again. Investors must know where they are in the cycle. They need to listen to the economists and central bankers and keep track of interest rates and inflation. This is required — sorry. Understanding business cycles and how they work is a prerequisite to effective saving and investing.

Business cycle theory says that as the economy heats up, business profits increase and their stocks rise. The reverse happens in the bond market. As companies borrow money to expand capacity they compete for capital, pushing up interest rates. When rates increase, the value of pre-existing bonds falls.

Unfortunately this negative correlation between debt and equity does not always hold true. For example, during the 1990’s corporate earnings and stock prices increased with global affluence. But interest rates never increased because savings from the developing world flowed in to finance the US expansion — keeping rates low. This was called the Great Moderation.

Also, the expected negative correlation between debt and equity may not hold true in truly bad times. Bond values also fall when the business’s ability to repay the principal becomes questionable. In bad times when profits and stocks are falling, the debt may also fall in value due to insolvency worries. in spite of the government cutting interest rates.

Never-the-less, some assets gain value in a booming economy (equity). Others gain value in economic contractions (debt). Some assets rise in value with inflation (real estate supposedly, and commodities). Other assets lose value with unexpected inflation (debt, utilities). Investors can reduce the risk of being whipsawed by business cycles by holding different asset classes. This is called ‘asset allocation’.

The correlation between the returns of different asset classes is measured by the correlation coefficient for each pair. A coefficient of 1 (one) means that an x% increase in class A will happen at the same time as the same x% increase in class B. A coefficient of (negative one) means that class B will fall by the same x%. Most asset classes fall between these extremes. Asset allocation is still worthwhile, even when positively correlated, because different classes with have different returns in different years, thereby smoothing portfolio returns.

Advisors will have you think that the asset allocation decision will make or break your investing success. But as you can see from this list of risks, asset allocation is just one of many risks you have to manage. If you care to read this, here is the seminal academic paper on Portfolio Selection by Markowitz (1959)

‘Diversification’ is the term used for buying baskets of securities within each asset class. It reduces the risk of individual investments the same way ‘asset allocation’ reduces the risk of single asset classes. Your total return is a blend of each component’s return. It is easier to ignore the price swings of one investment if the portfolio, in total, has steady values.

• You allocate between asset classes to reduce the volatility of business cycles. In spite of cycling, the global trend is up. So you hold all asset classes for the long term, through the troughs.
• You diversify between countries for exposure to different types of economies. E.g. Russia is tied to oil, and Japan to manufacturing. View these series of 60 year graphs of stock returns from different countries. Also read this paper on comparative returns from country diversification. See also ‘Political Risk’ below.
• You diversify between industry sectors because they rotate in importance to the economy. E.g. Technology was hot in 1999 and Commodities were hot in the 2006.

The number of common stocks you should own also depends on their riskiness vs. your personal tolerance for price volatility. Diversification will only reduce the volatility of your portfolio’s returns down to the level of the total market’s own volatility, but your choice of risky assets may predispose you to additional price swings. Eg. If your experience with your own portfolio of small or mid-cap stocks is that commonly 3 at any time have tanked 33%, but you can only tolerate a 5% drop in the portfolio’s value (in addition to any over-all market drop), how many stocks should you own?

# of stocks that fall at once