Magic Aristocrats

Tired of volatility and tracking error?

Applying a “Magic Formula like” strategy to long term dividend growers



This article explains why we designed the Magic Aristocrat formula and why we believe it will yield above average results with limited volatility on the long run. The Magic Aristocrat stock screeners are based on the observation that long term dividend growth and consistent high ROTCE are correlated and have one of the best predictive value for future financial performance of underlying companies.

The two main questions for a long term investor: the quality of an asset and its price

There are two key questions that a fundamental investor has to answer before allocating capital to an asset:

Question 1: What will the purchasing power of the asset be in the future and what are the financial performances of the asset (cash flows, earnings, ROE, dividends) going to be in the future ? Are the past financials indicative of future performance?

Question 2: Given the reasonable range of possible outcomes provided by Question 1), does the market price of the asset justify a buy or sell decision?

Smart indexing works, but it is very volatile

Systematic value investing strategies, also called “smart indexing” strategies, such as Joe Greenblatt’s Magic Formula or James O’Shaughnessy’s “What works on wall street” are well known and good attempts at answering question 1 and 2 mentioned above. Long term studies of those strategies indeed show that applying disciplined fundamental ranking methods to select an equally weighted portfolio of stocks tends to outperform the market on the long run. The studies published by these authors are of great value, but applying them in real life investing entails risks and can be a strain on an investor’s psychology and/or clients, because of the strong volatility and tracking error of the portfolios built with these investing styles.

Applying the Magic Formula or similar approaches, indeed, may be quite frustrating, or even impossible to do for professionals, since the companies that make the list because they seem cheap based on past metrics, often have serious fundamental problems which explain the low valuations.

As per our own experience, the companies returned by such screeners are often exposed to qualitative risk factors, which cannot be detected in the past fundamental financials reported by the companies. These risk factors can have different natures, such as commodity prices, management upheaval, technology or legal disruptions, increasing

competition, to name a few.

Screeners focusing on one past year’s fundamentals have limited predictive value

One of the major limitations of single or double criteria screeners is that they focus on a small number of fundamentals for the year before the screening. For instance, the Magic Formula only uses the past year’s earnings and ROTCE to select stocks. Yet, a past year’s earnings and ROTCE only convey a very limited amount of information about the long term profitability of a business: the past year may have been exceptionally good and may not re-iterate in the future.

As a result, the forward looking financial performance of the companies returned by these screeners often have little to do their past financial performance, making the screening process potentially irrelevant.

These issues can have a major impact on the short-term portfolio’s performance, such as a very strong tracking error and/or a high volatility.

Screening only makes sense if past fundamentals are indicative of future performance

This is the reason why we, at, are looking for screeners that are likely to find companies of which future financials are likely to remain close to their historical averages.

In other words, we are looking for good companies with a high probability of remaining good companies, and trading at a reasonable price.

In the following paragraph, we will explain why, in order to achieve such a goal, it might make sense to restrict stock screeners to a certain sub-set of the market: long-term dividend growers, and to apply the stock ranking philosophies to these subsets.

Long history of dividend growth and high ROE tend to be predictive of future good performance

Our analysis of long term databases, which is confirmed by the available literature, is that most companies which achieved high ROEs over a long period (more than 5 years), tend to maintain their financial performance, i.e. they tend to have superior ROEs in the following 5 years. The same is true for long-term dividend growers: most companies who achieved dividend growth for a long enough period (above 10 years) tend to keep growing their dividends in subsequent years.

Furthermore, a consistently high ROE is correlated to the ability of growing dividends on the long run for good reasons: being able to return ever increasing cash flows to shareholders requires a cash generating business model, disciplined capital allocation and a high moat. The ROE or ROTCE metric mostly captures those factors.

Coca-Cola (KO), or CH Robinson (CHRW), both long term dividend growers with high ROE/ROTCE metrics are paramount examples of such companies.

Using ranking screeners to find reasonably priced, high performing long term dividend growers

At, we propose to apply multi-factor ranking formulas, inspired by James O’Shaughnessy and Joel Greenblatt to a very specific subset of the stock market: long-term dividend growers with above average long term ROE/ROTCE.