You could, but those two are similar measures. Both are earning ratios. Earnings yield divides total
earnings by market cap. ROE divides total earnings by total equity as reported on the balance sheet.
They have a factor in common, but they are measuring very different things.
One, the ROE, attempts to measure how attractive the underlying business seems to be in terms of its economics.
A high return on equity (or return on assets), especially if not falling on trend, is a sign that a company has a business with pricing power.
If a competitor could reproduce the assets and compete with you, they would–they would get a high return on the money they put up.
But as they haven’t done so, there is probably something preventing them. An economic moat, in the parlance.
Their clients appear to believe that the company’s product or service has no close substitute: definitely a good thing for shareholders.
But that business could be either expensive or cheap at the moment.
The earnings yield is an attempt to decide which it is.
As the two metrics are so different, it’s possible to have any combination:
good cheap business, bad cheap business, good expensive business, bad expensive business.
By looking at both, you can try to find the stuff that’s both good and cheap.
One metric is no substitute for the other.
An interesting corollary: if two firms have comparable earnings yields, the one with the high ROE is definitely the one to go for.
As you note, this is tied in with price-to-book. The better firm will have fewer assets needed to
produce the same earnings, so if they’re at the same earnings yield it will have the higher P/B ratio.
This is the reason that looking for a low P/B firm as an attempt to get “cheap” stocks doesn’t work:
more often it is simply finding poor quality businesses that require the most assets to produce a dollar of earnings.
These are on average the businesses most exposed to head-to-head competition and lacking in pricing power.
Plus, they generally need lots of new capital in order to expand, leading to debt or dilution.
The fact that low P/B doesn’t work as a way to outperform is the reason that most “value” indexes don’t work.
They are looking more for bad businesses than for companies offering a lot of value for the price.
Some look for low growth rates too. They should be called “dud” indexes, not “value” indexes.
If you’re going to skip one of the two tests, ROE works surprisingly well on its own.
Arguably you do better with a good business not checking to see if it’s expensive than you do with a cheap business not checking to see if it’s good.
But it’s good not to get sucked in by a firm with too much leverage. That can boost ROE but create dire levels of fragility.
When looking at individual firms I don’t look at ROE without first looking at the gearing, e.g. earnings to debt ratio.
Jim