by Ben McMillan, Portfolio Manager, Ramsey Quantitative Systems Inc
Conventional wisdom has long held that dividends are a key tenet of investing. Benjamin Graham, himself, was unwavering in his view that a company's record of paying a dividend was perhaps the best signal of "quality."1 As investors have increasingly begun to favor simplicity and quality, dividends have emerged as a simple solution to both. We take a closer look at the practice of using dividends as the only investing criteria, particularly among small-cap companies, and risks and implications that arise as a result.
Arguments in favor of a dividend-centric approach to investing generally center around two key rationales for why dividends are an important indicator of company "quality":
- The company's ability to pay a dividend signals a stable source of cash flows.
- The company's willingness to pay a dividend signals a management team committed to increasing shareholder value.
It is certainly a sensible notion that, all things equal, the presence of a dividend should be evidence of a stable business model with management that is disciplined and efficient in the deployment of cash - and confident in their ability to continue doing so. However, it is also not difficult to understand why a company's lack of willingness to pay a dividend is not necessarily at odds with management's desire to increase shareholder value - particularly when a company's size and industry are taken into consideration.
There is no doubt that dividends are an important source of stocks' total return over time. But this is truer for larger companies than smaller ones. Since 1990, dividends have contributed approximately 20% of the total return for the S&P (2.30% out of 11.34% per year) vs. only 13% (1.52% out of 11.63% per year) for the Russell 20002:
While small-cap stocks have returned slightly more than large-cap stocks over this period, much less of that return has come from dividends. Additionally, the percentage of small-cap stocks that pay a dividend has consistently been about half that of larger ones:
It can easily be seen that dividends structurally discriminate against small companies. As a result, one would expect a dividend filter to result in two (potentially very) different groups of stocks. This clearly presents its own problems for investors using dividends as a selection criteria across market-caps as it's ignoring nearly half of all small companies from the start. The question then (particularly for proponents of dividends as a proxy for "quality") is whether or not these small-cap companies are being filtered out for the right reason. As a next step, we explore how dividends might be more accurately interpreted within the context of the company lifecycle. To conceptualize this, we show a simplified example of the stages of Small business growth (based on the work from Churchill & Lewis):
This illustration is useful in that it presents a framework for thinking about the key points of similarity as small businesses become large ones. Intuitively, it's easy to recognize that companies will be facing very different choices depending on which stage of growth they're experiencing. Among the most important of those decisions will be how best to deploy cash. The choice to pay a dividend (or buyback shares3) is a more obvious course of action for established companies with stable cash flows and limited ways to invest cash in a productive manner. Conversely, companies that are earlier in their lifecycle (and thus smaller) presumably have a greater number of opportunities with which to invest their cash. Additionally, these companies are also likely facing limited (and costly) access to capital. As a result, for smaller companies, the choice to pay a dividend carries with it a much greater opportunity cost.
This is not to say that dividends aren't a signal of quality decision-making within smaller companies or that a lack of a dividend is a signal of poor quality in larger ones. The point is simply that the justification for paying a dividend can vary greatly across different segments (like market-cap and industry) and it is not realistic to expect dividends as the sole selection criteria to reliably extract the same effect across these segments.
Not all dividends are created equal
To further this example, let's imagine a hypothetical company in the Information Technology industry that successfully goes through all the stages of business growth shown above.
Survival stage: This would likely be a pre-IPO "start-up". At this stage the founders are subsisting on either VC or bootstrapped funding and the goal is simply to grow (or gain) revenues before the funding runs out.
Early Growth: At this stage the company has raised money through an initial public offering and would have demonstrated success in finding a market for its products and securing some key accounts. This stage is likely seeing relatively high spending on R&D on both new and existing products and management is focused on growing revenues.
Rapid Growth: At this stage, our hypothetical company's R&D efforts have paid off and it has found a large (or rapidly growing) market for its product(s). A catalyst may have helped such as shifts in macro technology trends, key strategic alliances or intellectual property breakthroughs. The company is experiencing rapidly accelerating earnings. Its growth in return on equity (ROE) is high and scaling to meet demand is management's primary focus. At this stage, management is no longer in survival mode and is facing many new choices on how to deploy cash such as possibly acquiring other companies or increasing branding. The length of time in this stage can vary but it will no longer be a small-cap company if it successfully graduates out.
Established: Finally, potentially decades later, this hypothetical company has established itself as a leader in its space. Its technology is ubiquitous, its revenues are in the billions and its stock has been a top performer for years. At this point, earnings are no longer growing at huge multiples but they are stable (and sizable). The company (and its stock) will be much more sensitive to the broader economy and maintaining market share is now the key concern. Management (which has successfully demonstrated an ability to strategically deploy cash in high-ROE activities over the years) has now decided that returning cash to shareholders in the form of a dividend is a prudent use of capital going forward. Management is confident in its ability to continue to pay (and more importantly increase) the dividend in future years.
While the example above was put forth as a hypothetical illustration of how a technology company might progress through different stages of growth, it actually accurately describes Intel Corp from its early beginnings in 1968 to its $6.8M IPO a couple years later to its decision to initiate a dividend in 1992 when Intel's market capitalization was $18B and it occupied #93 on the Fortune 500. It's also important to note that long-time CEO, Andy Grove, who oversaw many of the key strategic decisions over the years, including to institute a dividend, was also a day-1 employee.
The goal of this vignette was simply to underscore the point made previously: Dividends as the sole proxy for quality must be considered within the broader context of market cap and industry - otherwise, the practitioner risks conflating other factors that may bias the intended outcome at the expense of other, high-quality, exposures.
While this analysis isn't intended to be a detailed assessment of the SmallCap Technology space, it is worth a closer look in order to quantify some of the intuition above. Specifically, we start with a universe of about 2,100 small-cap stocks from 2005 through June of 20154. The sample is very evenly split among companies that pay a dividend (51%) vs. those that don't (49%). When we look at this breakdown by sector, however, a different picture emerges:
We see that the vast majority of small cap stocks in either Healthcare or Technology do not pay a dividend. Similarly, the opposite is true for those stocks in the Financials or Utilities industries. As a result, an investor relying solely on dividends as an indicator of "quality" among small-cap stocks would be heavily concentrated in Financials and Utility companies at the expense of Health Care and Technology.
A closer look at Technology
As a next step, we take a closer look at dividend vs. non-dividend paying stocks within the small-cap Technology sector (for the simple reason that it's generally a larger Index component than Healthcare). As of Q2, about 16% of small cap companies in our sample were technology companies (which is very similar to the Russell 2000). Of those technology companies, only a quarter currently pay a dividend:
While the proportion of small-cap technology companies paying a dividend has steadily increased over time, the vast majority still do not, again highlighting a large portion of the universe that would be ignored by the investor relying on dividends as the sole proxy for exposure to a "quality" factor. This calls into question what effect, exactly, is this filter expected to capture and is there a more appropriate selection criteria that can be applied to this group?
To explore this, we put forth another potential indicator of quality: Return on Equity (ROE)5. The figure below now shows what percentage of each of the groups shown above have had a positive Return on Equity over time:
On average, more than 70% of non-dividend paying small technology companies still had a positive ROE over this period. While the proportion of positive-ROE companies is higher among those that pay a dividend, the point to be shown is using dividends as a proxy for quality potentially misses a large portion of small cap technology companies that would be considered "quality" under another definition.
If we go one step further and look at the growth rate in ROE for these two groups we find that the non-dividend paying small cap technology companies are actually increasing their ROE at more than double the rate of those companies that chose to pay a dividend:
While conventional wisdom has long touted the benefits of dividend-paying stocks, there has been a seemingly growing trend of touting dividends as the "one-size-fits-all" solution to building portfolios of "quality" stocks. Such approaches are, at best, misguided. As we show, the tendency for companies to pay a dividend varies greatly according to their size and industry (and for good reason), so strategies that exclude stocks based solely on this criteria risk introducing large biases into the portfolio while ignoring high quality companies (for no good reason).
2 Figure compares average calendar year price vs. total returns since 1990
3 While the topic of share buybacks is not directly addressed in this analysis, it occupies a close place along with the decision to issue a dividend.
4 We construct a proprietary universe of stocks with price > $5, market cap between $200M and $2B over the period from 2005 to Jun 2015. Data is sourced from CapitalIQ
5 ROE is shown as a contrasting example. For a comprehensive analysis of different proxies for quality and their relationship with small-cap companies, the reader is referred to "Size Matters, if You Control Your Junk" by AQR