[on ROE] Peak ROE levels are rarely sustainable, especially if that peak ROE is achieved with minimal leverage (debt). ROE is another one of those misused "magic numbers" that seemingly allows people to reduce a company's performance to a single number. No single financial ratio will ever give a complete picture. Plus, it's just a snapshot, and doesn't consider those all-important trends.
I've seen many different ROE studies, with results ranging from "holy grail" to "useless." As with any statistical study, slight changes to assumptions, definitions, or timeframes can change the results ENORMOUSLY. Even the definition of ROE isn't standardized, so a person can't just compare the ROE of one company to another between data services, without looking behind the numbers.
The numerator of ROE can and is taken from several different points on the income statement, and even the denominator sometimes contains adjustments. Each variation of ROE is defended by its user as being the most accurate. I'd argue no single definition is best for all cases.
One of the most thorough and careful ROE studies I saw used operating earnings plus tax, but before interest (most published ROE numbers use net income because it is easy to get), and ran a complicated correlation analysis of ROE versus stock price on individual issues. It was very clear on the conditions where correlation was high and why this was so. It also made a good argument why broad market ROE studies were flawed, or at least prone to misinterpretation.
Despite the seemingly simple definition of ROE, it actually breaks down into as many as 5 operational components. Analyzing these individually gives clues as to why the company's ROE is what it is. ROE trends, in a broader operational sense, are also linked to business cycles and product trends. In an emerging market there are (at the beginning) many high and increasing ROE's among the competitors. As time passes, some of the competitors are consumed, margins decrease to stable (lower) levels, and the market funnels into a single (or much fewer than initially, anyway) dominant player. Market sectors follow this pattern over and over. Unrelated sectors have different cycle lengths and peaks.
In summary, it's perfectly logical that, for a given company, ROE using net income rises to a peak as the company is in its highest growth period, then declines an amount in proportion to the company's position in the resultant 'pecking order' that shakes out of the market/product dynamics (in other words, the 'leader' declines the least and holds at a high level, while the many others in the sector decline further).
dh |
