Recently I mentioned to someone, if only in passing, that I didn't believe that standard deviation was a rational source of identifying risk. Now, I want to be perfectly clear that I am well aware of the corresponding research that demonstrates its practicality. In fact, if one is solely concerned with future price fluctuations, then a historical measure of how widely an asset's price has varied over time is likely to be a noteworthy statistic. It is true that past performance does not indicate future results, but relative risk measures are often stickier than performance indicators. In this way, within the suggested price-only framework, standard deviation is a reasonable metric to identify.
Having this knowledge, the obvious next step is to go about defending why I disagree with the use of standard deviation as a measure of risk. This task is threefold: identifying my strategy, indicating the implication of such a strategy, and demonstrating the short-coming of ascertaining that fluctuating prices are risky in this context.
First and foremost, I am a dividend-growth investor. That is, I look to buy a collection of wide economic moat companies that have not only paid, but also increased their dividends over a long time period -- companies like Coca-Cola (KO), 3M (MMM), Emerson Electric (EMR), or Archer Daniels Midland (ADM). The idea is that these companies have both the propensity and ability to continue to increase their payouts in the future -- not only increase them, but increase them by a rate that far outpaces inflation. If a company happens to freeze, cut, or even moderately increase its dividend by a less than acceptable rate, then the company is further scrutinized to be considered for replacement. On the other hand, as long as my holdings continue to increase their payouts by an acceptable rate, while remaining together profitable and fundamental, then I will have no intention to sell even a single share.
The implication of such a strategy is that one is primarily concerned with a growing stream of income over time. As indicated, not just growing, but growing by a rate greater than inflation to increase your purchasing power over time. The end goal is to be able to live off the dividend proceeds. For example, let's say that one is generating $50,000 worth of today's purchasing power in 40 years' time, which incidentally, isn't that outrageous of a goal. If that amount of purchasing power is enough to cover your expenditures, you're essentially good-to-go with a dividend growth strategy. It is true that you would need to monitor your holdings to ensure that the businesses remain fundamental and that the dividends keep increasing by an acceptable rate. But beyond that, you really don't have to worry about the share prices, as you're not looking to liquidate any of your profitable endeavors. In the second year, perhaps those same holdings bring in $53,000, $56,100 in the third year, and so on.
In this way, you have no intention of selling shares or being concerned with short-term price movements. Moreover, just because one isn't necessarily concerned with share prices, that doesn't preclude them from capital appreciation. It just so happens that if the dividend continues to increase, the company's earnings must also increase over the long term. Higher earnings, given that you didn't pay too much in the first place, indicate a higher share price. Seen another way, it isn't very likely that we're going to see a 10%+ dividend yield on say, Kimberly-Clark (KMB) in the next 20 years. Yet, if KMB is able to grow its dividend payout by an average of 6% a year moving forward, this is exactly what the math dictates in order to have no capital appreciation whatsoever.
Now that we have developed a strategy context, I can better detail why standard deviation doesn't define dividend stock risk. In this regard, I plan on making three cases: theoretical, observational and a "true risk" case. Thus far, we have seen why one isn't necessarily concerned with stock price fluctuations of dividend-paying stocks, especially in the short term. It's not necessarily that the dividend-growth crowd doesn't care about stock price movements. It's simply that a growing dividend is a much more manageable thing to be concerned with. It's very easy to determine whether or not a certain payout has met your expectations. On the other hand, even if you are inherently correct about the relative intrinsic value of a security against the market's consensus, there's no way to ensure that you will be rewarded for your diligence. You're at the whim of a debatably rational market.
Let's apply this in a theoretical sense to using standard deviation as a measure of risk. Using such a measure, a higher standard deviation indicates a wider range of price swings and thus, a greater risk. In my opinion, if one is focused primarily on income, there is a useful purpose within this perceived risk. Imagine that everyone knew that say, Clorox (CLX) would increase in price by 4% a year for the foreseeable future, keeping all multiples the same. The standard deviation would be very low and therefore, the perceived risk would be low.
But what happens if CLX is currently trading slightly higher than you are willing to pay? That is, even with the 3.5% yield and 4% price premium, you can find better investing alternatives, even though you think that Clorox is a worthwhile company. Over time you would buy exactly zero shares of CLX. On the other hand, imagine that CLX has a high standard deviation such that the price swings widely by 10% intra-year. In this case, one might be able to buy shares of CLX at a price that they are more comfortable with, even if the perceived risk is higher. I would liken this ideology to having capital to deploy, but sitting on your hands during an up market. You're hoping for a rather immediate sharp decline in order to buy a greater portion of ownership. That is, you are hoping for the exact deviation that so many are trying to avoid.
The observational problem with standard deviation is plain to see. For illustrative purposes, I found the annual standard deviations for the S&P 500, Consolidated Edison (ED), Walgreen (WAG) and a hypothetical set of returns that followed the return pattern of 1% for three straight months followed by 20% every fourth month.
The above table is based on monthly percent change observations, which I annualized. In addition, we will get to the difference between the Consolidated Edison and Walgreen shortly. However, I want to focus on the seemingly high standard deviation of the hypothetical company. If one were to only look at this table and consequently agree that standard deviation is an acceptable measure of risk, then one would come to the conclusion that this set of returns had the most risk.
Yet look what happens when you manually view the returns: Consolidated Edison had 77 months of negative price returns, and Walgreen had 65 months of negative price returns. On the other hand, our hypothetical company had exactly zero months of negative returns. The underlying reason that the hypothetical company had a higher standard deviation is the outsized return that occurred every fourth month. This isn't risk that we are seeing, but simply a measure of dispersion. In relative terms, a positive return every period isn't so much risky as it is observable. Now there are corrective methods that can be used to mitigate this seeming anomaly -- the Sortino ratio comes to mind -- but I will still maintain that price fluctuations are not a risk to dividend stocks, especially if the payouts are growing.
Now the obvious criticism that results from disagreeing with a common known factor, in this case standard deviation as a measure of risk, goes something like this: "Oh yeah, if you don't like standard deviation, then what are you proposing, huh?" It's a fair point, and one that I would like to address through an example. Let's go back to the standard deviations of Consolidated Edison and Walgreen, 13.94% and 25.36%, respectively. Incidentally, the corresponding betas of these companies would also depict a similar story. (0.14 and 1.17, respectively on Yahoo Finance, although I would advocate doing the calculations yourself) Moving forward, it would appear that ED is less risky than WAG. Of course, you can pick whatever dividend growth stocks that you like, as the logic remains fundamentally the same. Given that WAG seemingly has more risk, one would expect more return to be provided over the long term. I am not disagreeing with this argument.
But let's think about the risk involved in the context of a dividend growth strategy -- a strategy where one is concerned with increasing their purchasing power over time. Consolidated Edison has increased its dividend for 38 straight years. Over the past decade, these increases have come in at an average annual rate of just under 1%. Walgreen has increased its dividend for 37 straight years, but by an average of almost 19% a year over the past decade. Your guess is as good as mine as to how quickly these respective payouts will increase in the future. But to continue with the example, let's say that ED continues to increase its payout by 1% a year for the foreseeable future (no indication to the contrary), while Walgreen comes down to a more manageable 7% annual rate. In the context of past price appreciation, it appears that ED is less risky than WAG. But in the context of future income generation, given an inflation rate above 1%, Consolidated Edison is guaranteeing not just a rising income, but the loss of purchasing power over time. On the other hand, Walgreen is guaranteeing an increase in purchasing power over time.
For that matter, any company that can outpace inflation in its dividend payouts over the long-term, be it Procter & Gamble (PG), AT&T (T) or Target (TGT), is increasing your dividend purchasing power, while Consolidated Edison is losing dividend purchasing power. Thus, the true risk with a dividend growth stock in this case is that dividend increases, which you are primary concerned with, do not passively increase your purchasing power over time. In a much wider sense, the true risk with dividend growth stocks can be extended to the probability that a collection of payouts are frozen or worse yet, cut altogether.
In addition, Warren Buffett detailed my illustrated points rather succinctly in his 2011 shareholder letter, found here. I'll let you review this rather informative piece at your leisure, but I would like to end with his prefacing point. It should be noted that beta and standard deviation are separate measures, but they are related in their relation to relative price movements.
The riskiness of an investment is not measured by beta (a Wall Street term encompassing volatility and often used in measuring risk) but rather by the probability -- the reasoned probability -- of that investment causing its owner a loss of purchasing-power over his contemplated holding period. Assets can fluctuate greatly in price and not be risky as long as they are reasonably certain to deliver increased purchasing power over their holding period. And as we will see, a non-fluctuating asset can be laden with risk.