What is risk?

Understanding and managing risk is a critical part of being a successful investor.  But what is risk?  We would look at it from two different angles:

risk as uncertainty

Many academics and the more quantitatively-oriented investors view risk as volatility of returns, often expressed by measures such as the the standard deviation of returns.

This approach to measuring risk has certain advantages:  with a relatively modest amount of return history one can make some statistical inferences about risk.  In addition, a full range of statistical procedures can be used to make predictions about risk.

This approach also has certain disadvantages.  For example, the measurement of risk depends in part on the measurement intervals.  For the same stock, risk measured using daily returns often looks very different than risk measured using monthly returns.  Also, statistical risk measurements usually are backwards looking; but the investor typically cares about the risk embedded in future returns.

Using the standard deviation of returns can lead to some big surprises.  A common practice is to calculate the standard deviation of past returns and assume that the standard of deviation of future returns will be similar. Often this works but sometimes it doesn’t in a really bad way, with the future being far more volatile than the past would suggest.

A final problem with variance or standard deviation is that upside volatility is considered as undesirable as downside volatility.  Actual investors may care only about downside volatility.

risk as losing money

Some professional investors and a large share of the public define risk as the possibility of losing money.  This sounds simple, but in practice it is can be difficult to incorporate into the management of a portfolio.

For example, how does one reasonably estimate the potential for losing money if stocks have gone up for several years in a row?  Lacking data that is both recent and meaningful, the risk assessment becomes almost entirely subjective.

Another problem is whether the losses will be temporary in nature or more permanent.  If losses are likely temporary, then should they be disregarded?  How does one decide whether a loss will be temporary?

What do you think?


Do Consumers Make Terrible Stock Analysts?

Early in my career, I read Peter Lynch’s One up On Wall Street How to Use What You Already Know to Make Money on Wall Street.  Lynch made the case that one can pick stocks on the basis of what one knows personally about a company.  If you like a company’s products, maybe its stock is a good investment.

It’s a strategy that has its adherents, but I can see some challenges standing in the way of an amateur investor’s success with it.

First, the strategy would tend to orient most people’s stock selections to United States-domiciled consumer-oriented stocks.  As consumers we interact with a wider range of companies than what we might encounter at work.  Consumer stocks are not necessarily always the best investment opportunity.  In addition, one would have an orientation towards domestic companies.  Many more stocks trade publicly outside of the United States than trade within the United States.

If the investor focuses too much on U.S. consumer stocks, the investor could lose in two ways:  (a) he may miss out on opportunities that are not U.S. consumer stocks, and (b) he likely will have a portfolio that has less-than-ideal diversification.

Second, whether one likes or dislikes a company’s product has at best a tenuous relationship to the future performance of the stock.  Other factors like valuation, quality and stock price momentum probably say more about where a stock may go.

Third, a high regard for a company’s products may create a counterproductive emotional attachment to an investment in the company’s stock.  In our opinion, a disciplined and unemotional approach to investments carries a better chance of success.  Emotions tend to get in the way of such an approach, or at least make such an approach more difficult to sustain.

Physical Science or Social Science?

My good friends at AlphaArchitect recently published a guest post on how engineers can transition to a career in investment management. If one is an engineer and wants to make the career transition, the article seems like a good place to start for ideas on how to do it.

I would like to add to the discussion an underappreciated question that engineers would not normally consider, specifically is investing a physical science or a social science?  I raise this question as a liberal arts major who has done quantitative investing for a long time, and who worked with engineers who had transitioned to investment management.

Engineers (as well as all people with rigorous STEM training) have a certain way of looking at the world.  If we want bridges that do not collapse, we want civil engineers who have that certain way of thinking. Create theoretical models and test against scaled down real-world models. Scale things up while comparing real-world results with the theoretical models.  Take lots of measurements and constantly test to make sure we understand every aspect of bridge design so that the bridge does not fall down.

The education of engineers is designed to get them to think rigorously about the physical realities of whatever specialty they have chosen.  I don’t think anyone could graduate with a degree in engineering without having fully internalized the engineering mindset.  Which is a good thing: We don’t want fuzzy-thinking people designing our bridges.

The financial markets seem an obvious fit for engineers. Large amounts of data are available for the engineer to study.  Statistical techniques honed in undergraduate or graduate-level engineering programs can be applied to financial data to develop seemingly promising investment strategies.

However, the opportunity to apply engineering analytical techniques to financial data does not mean that the behavior of investments can be understood in the same way as the physical sciences.  Security prices do not behave with the predictability of cement and steel.  They impound the foibles of humanity.  Thus, as Isaac Newton said, “I can calculate the motion of heavenly bodies, but not the madness of people.”

As a thought exercise, the aspiring investment manager-engineer should consider what he will do when he suffers bad investment performance that equates to a Z-score of less than -5.  In my career, I have seen this happen at least twice but the probability should have been nearly nil.  It happened because investing is not a physical science.  One cannot underestimate the psychological impact of Z<-5.  This is the financial equivalent of the bridge collapsing.

The collapsing bridge triggers very powerful and unpleasant emotions. When besieged by such emotions, one’s ability to engage in open minded higher order thinking shrinks.  Most people revert to whatever mode of thinking comes naturally to them – in specific, how they have been trained to process information.  For the engineer turned investor, mental reversion means rechecking the calculations, re-running the backtests, more portfolio simulations.  But the vast majority of the data on which the engineer based the strategy has not changed – just the most recent data. With so little data having changed, the conclusions will not change, either. If one’s backtest used 15 years of history in order to achieve statistical significance and subsequently something has changed in the market, then it maybe another 10 to 15 years before there is enough data to detect that change.

Few clients are willing to wait that long.  So the engineer will be in an uncomfortable position:

  • stick with the engineering training, frantically re-running simulations and hoping that the bad performance is an anomaly, or
  • take a guess at the nature of the new reality (if, in fact, things have changed) and hope that the guess is correct.

Either way, it will not be comfortable!

14 ways to calculate value

Value is a stock selection strategy focused on finding stocks that trade for prices significantly below intrinsic value.

Unfortunately, “intrinsic value” is a matter of opinion and the opinions vary widely.  As a substitute, practitioners often compare the price of a stock to one or more of a company’s financial measures.  Here are fourteen ways to calculate value and some of the problems associated with each measurement:

  1. Price/Earnings (P/E):  one divides the price of the stock by the earnings per share.  In general, a lower but positive P/E indicates potential undervaluation.  P/E has lots of problems.  For many companies, the denominator is volatile; when earnings are low P/E is high and thus the stock presumably would be unattractive.  But what if the earnings decline is temporary?  Then P/E is misleading.
  2. Price/Operating Earnings (P/OE):  this approach excludes from earnings the effect of non-operating items such as extraordinary charges.  In excluding such items, one hopes to dampen the volatility of earnings and thus create a more stable and easily interpreted P/E ratio.  The problem here is that some managements will “kitchen sink” the earnings – they take all the bad news they have been accumulating for years and write it all off at once as a “non-recurring” item.  Should investors actually ignore such costs?  Maybe, maybe not.  If not, P/OE is misleading.
  3. Price/Cash Flow (P/CF):  same calculation as P/E, except one adds to the denominator depreciation and amortization per share.  Because the denominator is less volatile, P/CF is less volatile and hence easier to interpret.  However, it is biased because capital intensive industries have large depreciation and amortization charges.  Those companies may look cheap, but they must reinvest large amounts of cash to maintain their productive capacity.
  4. Price/Sales (P/S):  This ratio compares price to sales per share.  Compared to earnings and cash flow, sales appear relatively stable.  A low P/S ratio indicates potential undervaluation.  However, some of the cheapest companies by the P/S ratio may be unprofitable, and no company can remain unprofitable indefinitely.
  5. Price/Free Cash Flow (P/FCF):  This ratio compares price to free cash flow per share.  Free cash flow is cash flow (as defined above) less capital expenditures and dividends.  This is the cash that the operating business “throws off” and is not otherwise committed.  A low ratio is potentially attractive.  Problems with P/FCF include the fact that many companies do not have positive free cash flow.  How does one analyze them?  Also, management must decide what to do with the free cash flow.  Sometimes a management team will make ill-advised investments that harm the shareholders’ interests.
  6. Price/Book (P/B):  Book value is the difference of assets and liabilities, which are found on a company’s balance sheet.  A low P/B indicates potential undervaluation.  If liabilities exceed assets, this ratio becomes difficult to interpret.  Also, the intrinsic value of a company may depend more on how much income the company can produce than on its book value.  But book value may not account for such income or potential to produce income.
  7. Price/Tangible Book (P/TB):  A company’s assets may include intangibles such as good customer relations.  Tangible Book Value eliminates such items from a company’s assets before calculating book value.  P/TB is viewed as a more conservative measure of value than P/B.  Like P/B it also ignores income (actual or potential) which probably has more meaning as a proxy for intrinsic value.
  8. Price/Net Current Asset Value (P/NCAV):  NCAV assumes a very dire situation:  all but the company’s current assets are worthless; from that residual one subtracts all of the company’s liabilities.  A low and positive P/NCAV is attractive.  Very few companies have positive NCAVs, so under normal conditions this ratio applies only to a small number of companies.
  9. Dividend Yield:  By this approach, a higher yielding stock offers better value than a low- or zero-yield stock.  Dividend yield ignores the sustainability of the dividend.  Some stocks have very high yields because the current rate of dividends is unlikely to continue in the future.
  10. Shareholder Yield:  This valuation method recognizes that management can use cash to benefit shareholders in three ways:  dividends, debt repayments and share repurchases.  The numerator is the sum of dividends, net debt repayments and net share repurchases.  The denominator is the price times the number of shares.  A higher shareholder yield indicates potential undervaluation.
  11. Enterprise Value to Earnings Before Interest and Taxes (EV/EBIT): Enterprise value adds the value of a company’s equity and debt, less cash. EBIT is a measure of income.  A low EV/EBIT ratio indicates potential undervaluation.  If EBIT is temporarily low or negative, the EV/EBIT ratio becomes misleading.
  12. Enterprise Value to Earnings Before Interest, Taxes, Depreciation and Amortization (EV/EBITDA):  This ratio adjusts EV/EBIT in the same way that P/CF adjusts P/E.  It also suffers from the same flaw, in that capital intensive industries may appear undervalued when, in fact, they are not.
  13. Enterprise Value to Sales (EV/S):  This is analogous to P/S, but uses Enterprise Value as the denominator.  It suffers some of the same issues as P/S.
  14. Price to Discounted Cash Flows (P/DCF):  With this approach, one forecasts out the future cash flows from an investment in a stock, including dividends and the price at which one expects to sell the investment at some point in the future.  It is very difficult to know what the price of the stock will be.  Therefore, it is very difficult to make practical use of this ratio.

Clearly, each measure of value has its weaknesses.  Practitioners sometimes overcome the problems by using several different ratios as a “composite” value indicator.

Stocks and Closets

We like to keep track of what’s going on in the world of psychology. Mastering the technical details of investing isn’t too difficult.  Gaining mastery over one’s emotions is an entirely different matter.  The field of psychology can provide valuable insights in this regard.

Psychology Today published an article, Declutter NOW! How Prospect Theory Clutters Up Our Closets, which makes the case that we overvalue what we own; we don’t want to give it up even when it is irrational not to do so:

In one study, people randomly won either a nice pen or a new coffee mug. Both gifts are approximately equal in terms of dollar value. However, after winning your prize, you are then permitted to trade it for the other option, if you want.  Surprisingly, very few people traded. Once you get your gift, you don’t want to give it up.

The professors of finance have looked at this and drawn somewhat different conclusions.  See Grinblatt & Han’s “Prospect Theory, Mental Accounting and Momentum.”

Much like the person who only reluctantly parts with clutter in the closet, investors instinctively hang on to losing stocks.  They may tell themselves that they will sell when they get back to even.  Why?  Does the market care what they paid for it?  Of course not!  But they hang on to their belief that the stock should be trading for more than the current price.  If enough investors do this, a significant supply of shares can be held off the market; shares which likely become future supply as investor’s one-by-one give up on getting all of their money back.  Thus, investors become risk takers when they are sitting on a losing position.

When it comes to gains, the dynamic appears to be a little different.  The instinct for many investors is to take profits.  Thus, as a stock advances it must fight through waves of selling by profit takers.  Such selling tends to restrain unnaturally the price advance in the near term, perhaps allowing for more price rises in the future.  Investors become risk averse when they have a profit.

The Behavioral Finance Conundrum

Behavioral Finance has become a popular concept in the investment world. The idea is that humans are not entirely rational thinkers.  Since they are not always rational, from time-to-time they will make irrational decisions with respect to stock prices.  This creates an opportunity to make abnormal profits in the stocks market – one need only identify other people’s irrational stock price decisions and take the opposite action.

Behavioral Finance enthusiasts then sift through reams of data looking for ways to exploit the irrational decisions of others.  They find a few:  value and momentum, among others.  They construct portfolios around these investment criteria and hope for the best.

Pondering from beneath furrowed brows, the doubters point out that just because something worked in the past doesn’t mean it will work in the future …

But the behavioral finance enthusiasts have a retort:  These profitable stock market anomalies are rooted in the biased architecture of the brain. So long as the brain doesn’t change, the anomalies will not change.  And since evolution plays out over millenia, they argue that these phenomena will persist for at least a while.

We believe in behavioral finance, but we don’t believe it will release its fruit so easily.  How can we access the anomalous profits if we, too, have the same basic brain architecture that allows the anomaly to exist?

Notwithstanding the commendable efforts of the behavioralists, it seems that the market will adapt in ways that make matters as frustrating as possible for the largest majority of investors.  So frustrating that the strategies of the behavioralists will be difficult for most people to follow.

For example, recent research has suggested that newly discovered stock market anomalies usually perform much worse after discover than before.  This phenomenon would seem to be part of the market’s adaptive process.  The lower returns and higher risk raise the frustration level for the behavioralists.  Rising frustration makes it harder to stay the course.  Few investors reap the rewards.




Reflexivity and Feedback

My good friends at Alpha Architect have published an article, “Reflexivity and the Feedback Effect in Financial Markets.”  Quoting George Soros, who coined the term “Reflexivity,” some key excerpts from the article are,

Changes in prices can influence the fundamentals of stocks. Stated another way, the way we perceive fundamentals and how this affects price can cause fundamentals themselves to change …

Usually some error in the act of valuation is involved. The most common error is a failure to recognize that a so-called fundamental value is not really independent of the act of valuation. That was the case in the conglomerate boom, where per-share earnings growth could be manufactured by acquisitions, and also in the international lending boom where the lending activities of the banks helped im­prove the debt ratios that banks used to guide them in their lending activity.

It seems to me that, for stocks, reflexivity would work to the extent that changes in valuation affect management’s perception of the cost of capital.

For some management teams, the cost of capital has little to do with the actual valuation of their stock.  They will estimate the cost of equity capital by attempting to guess what the shareholders want in terms of a long-term rate of return and risk.  The actual behavior of the stock in a given year (or years) is not as important.  Thus, if the stock doubles, such management teams may not adjust their equity cost of capital all that much.  In this case, valuation will not feedback into fundamentals.

On the other hand, less cautious management teams seem to view their stock as just another opportunity for arbitrage.  Thus, if the stock valuation doubles, they are inclined to think that the cost of equity capital has declined by 50%.   Such a large reduction in cost will lead to greater investment spending – on capital equipment, on acquisitions and/or – if stock options are involved – by hiring more people.

One could easily imagine a situation in which management essentially squanders the additional investment that the higher stock valuation encourages:

  • doing overpriced acquisitions with stock, on the theory that the acquirer’s stock is even more overvalued; only to suffer merger integration problems later on.
  • aggressively expanding the company’s physical infrastructure in excess of underlying demand, resulting in pricing pressures, capacity utilization problems and deteriorating profit margins; all of which result in deteriorating return on capital.

The opposite would seem to hold as well, for the stock that has fallen.  But the mechanism would be a little different.  Poor fundamentals weigh on the stock and eventually alarm the creditors, who force retrenchment. Assuming the company survives, retrenchment resuscitates profitability and return on capital, paving the way to a higher stock price in the future.

Book Review: Quantitative Momentum

Wesley R. Gray, PhD., and Jack R. Vogel, PhD., have done a commendable job in their recently published book, Quantitative Momentum A Practitioner’s Guide to Building a Momentum-Based Stock Selection System. Momentum is an investment style that calls for buying recent stock market winners, with the idea that such stocks will continue to win in the near future.

Dedicated amateur investors will find this book a welcome addition to their investment libraries. It provides an in-depth overview of momentum as an investment style while showing how to build a sophisticated momentum strategy from the ground up. Experienced practitioners will appreciate the authors’ intellectual rigor and the interesting features of their specific implementation of momentum.

The authors present their case in two parts. In the first part, “Understanding Momentum,” Gray and Vogel set forth their rationale for using momentum to select stocks. Contrasting two major philosophies for evaluating stocks (fundamental and technical), they make the case for using the two together – an approach that we ourselves support. Notwithstanding the success of passive strategies, the authors find a convincing case for active investment management provided investors can find the discipline to see through the inevitable (and potentially extended) periods of poor performance that active management may suffer. At the end of part one, Gray and Vogel show how momentum can be an attractive complement to a value-oriented strategy by smoothing out some of the volatility of value investing.

The second part explores techniques for enhancing the performance of a simple momentum strategy. The authors note that momentum works best if one evaluates returns over the past three to twelve months. The strategy does not work well on very short-term past returns (a month or less) or on long-term returns (more than a year). One can improve on its performance by focusing on consistency. The more consistent the upward movement of a stock, the stronger the potential future returns. In contrast, stocks that achieve their gains in a choppy manner are less likely to perform well. The authors also find a seasonal aspect to the returns of a momentum strategy, but have decided that this effect is best used to time the rebalancing of portfolios.

Practitioners with a thorough understanding of the momentum literature may fault the book for not exploring in depth the subtleties of combining momentum and value in light of recent work published by AQR; or for not exploring in detail Grinblatt & Han’s paper, “Prospect Theory, Mental Accounting and Momentum.” But such omissions are understandable in light of the intended audience and do not diminish the appeal of the book.

Eleven ways to calculate momentum

Momentum is a stock selection strategy of buying past winners with the hope that their winning performance will continue in the near-term.

Typically momentum investors will calculate the past return of each stock in a given universe and then sort the stocks in order of return.  By this approach, the stocks with the highest returns are considered the most attractive.  That the actual past returns of those stocks may be positive (in a bull market) or negative (in a bear market) is not the point.  Instead, the concern is which stocks have done the best regardless of return.

One would think that calculating momentum is not controversial.  Far from it.  Practitioners and academics have developed a surprising number of ways to calculate momentum:

  1. Point-to-point:  One compares the price today with the price (adjusted for dividends) at a preceding point in time.  The ratio minus one tells us the return.  Typically, the preceding point in time trails the current price by three to twelve months.
  2. Lagged point-to-point:  The idea that past winners continue to win in the near future does not appear to hold for return measurement periods of a month or less.  Instead, very short term returns tend to reverse so that the short-term losers tend to rebound.  To avoid commingling contrary signals, one can lag the calculation by up to a month so that the “current” price is the price from ago.
  3. Average Return:  A legitimate criticism of point-to-point measurement is that an outlier price many months ago can affect the momentum score of a stock today.  One can address this concern by calculating momentum as the average monthly return over the past three to twelve months.
  4. Multi-horizon return:  Another approach to reducing point-to-point outlier errors is to calculate momentum using the sum of returns over several different horizons – for example, 6, 9 and 12 months.
  5. Price to moving average:  One divides the current price by the average of the stock’s price over a certain period of time.  A higher ratio means higher momentum.
  6. Price to VWAP:  “VWAP” stands for “volume weighted average price.”  This is a variant of price to moving average.  Instead of taking a simple average of the price over a certain period of time, one weights each day of price history by the volume traded on that day.  A higher ratio of price to VWAP means higher momentum.
  7. Smoothed price history:  Instead of smoothing the past with a moving average or VWAP, one can smooth all of the prices – including the current price.  For example, one could convert the price history into a five- or ten- day moving average of the price and calculate the momentum score by comparing the most recent data point in the moving average to another data point further back in the history.
  8. Slope of the price history:  Taking average monthly return a step further, one can use regression analysis to fit a slope to a stock’s total return chart.  The independent variable is time and the dependent variable is price (adjusted for dividends and stock splits).  One estimates a slope using regression analysis; the slope becomes the momentum score.
  9. Slope of the log of the price history:  In extreme cases, the compounding of returns can distort the calculation of the slope.  Converting the price history to the log of the price history can solve that problem.
  10. Coefficient of determination adjusted slope:   Regression analysis will force a slope even when a stock’s price chart is choppy and no slope is obvious.  One can adjust for this phenomenon by multiplying the slope by the coefficient of determination.  Doing so will tend to remove such stocks from the tail ends of the momentum distribution.
  11. Basis simulation:  Based on a paper by Grinblatt and Han (“Prospect Theory, Mental Accounting and Momentum”), some practitioners attempt to calculate momentum by comparing the current price to the estimate cost basis of the average current holder of the stock by examing the stock’s price history, volume history and shares outstanding.

Do you know of any others?


Process over Goals

Gatis Roze has an excellent article on what has worked for him in a multidecade investing career.  The article may come up a little short for the left-brain (systematic/analytical) dominant thinkers, but I think it gets at some important concepts.

To paraphrase, the article takes the view that a focus on process has better potential to wring profits out of the market than a focus on one’s goals.

By process he means the design of an investment methodology. For Gatis Roze, disciplined following of a process leads to successful investment results.  In contrast, focusing on goals (profits) draws attention away from the process and reduces the likelihood of success.

We agree.  A well-designed process, followed with discipline over time, has the potential to achieve good results in the capital markets. Reasonable goals may be difficult to achieve if one has a poor investment methodology; or a good methodology but inconsistent application due to a lack of discipline.

An apt analogy may be found in sports.  I paraphrase a world-famous coach who I know personally.  He says that goals are counter productive.  He coaches his athletes to avoid thinking about them too much.

It is inevitable that people will fail to achieve at least a few of their goals on their journey to mastery.  The failures sap the will.  They may have to be revised down to a more “realistic” objective.  A few more failures, morale falls a bit more.

In contrast, the athletes who focus on process tend to perform better. Were the exercise drills at the required intensity?  Was the diet followed? Did people get enough sleep?  In this case, the coach and the athlete design the process – exercise, food and sleep – to achieve the goal.  Fantasizing about the goal does not lead to success.  Executing the process may achieve the goal assuming the athlete has the natural talent.

What football team excels by keeping its eyes focused on the score board?What team has gone from losing in the first half to winning in the second half by obsessing about the score?  What successful coach tells his players to neglect the fundamentals in favor of focusing on the result?  The team that focuses on mastery of strategy, conditioning and proper execution of plays probably has a decent chance of winning, assuming they have talented athletes.

The same goes for investing, but it so easy to get sidetracked!  All of us care intently about the goal – good returns – because, frankly, for most of us the stakes are very, very high.  It is hard to focus on the process when your future is on the scoreboard.At the same time, the behavior of the markets appear designed to draw the largest possible number of people away from whatever process they use.

At the same time, the behavior of the markets appear designed to draw the largest possible number of people away from whatever process they use. Even the best investment process will toss out unwelcome surprises far too often.  It’s the nature of the markets.

But whatever the stakes and whatever the market does to distract us, the process and the discipline (or lack thereof) ultimately determines what happens to our investments.