I don't mean to pick on Donville, because they have a good track record and seem to have a lot of integrity, but I think that Home Capital is a great example of a company where consistently high ROE does not make an investment good.
What Home Capital does
In corporate presentations, Home Capital describes itself as "Canada's leading alternative financial institution" which underwrites "mortgages not meeting lending criteria at traditional financial institutions." In other words, it lends money to subprime and alt-a borrowers.
As Donville's newsletter makes clear, this has been a very profitable business, with ROEs above 20% every year since 1998:
For context, here is what Canadian house prices have done over the same period:
(Chart from World Housing Bubble)
As the chart shows, Canada is in the midst of a truly phenomenal housing bubble. The prudent thing to do in a market like this is to reduce risk by cutting back on lending, which results in lower ROEs in the near term. Home Capital has done the opposite: it has expanded its loan book as the market has gotten frothier, maintaining its ROE by lending against increasingly inflated assets.
Comparison with Golden West
Golden West Financial was a large savings and loan association in California. Like Home Capital, Golden West consistently earned high ROEs and was a pioneer in introducing new kinds of loans and serving borrowers that larger banks didn't serve very well.
Investors and competitors alike admired Golden West. According to Wikipedia, "In 1990 The New York Times called the company 'the Nation's Best-Run S.&L.'" I have an old Peter Lynch book in which he praises Golden West for its entrepreneurial flair and cost-control efforts, much like how Donville praises Home Capital.
Wachovia acquired Golden West in 2006. Two years later, Wachovia became one of the biggest casualties of the housing bust, with surging NPLs, and had to sell out to Wells Fargo in a distressed sale. Most of Wachovia's problem loans had originated at Golden West.
Golden West had lent aggressively during the housing bubble, and its loan book was so toxic that it nearly destroyed a much larger bank. If the company hadn't been acquired, it would have gone bust pretty quickly.
Reassuringly, Donville writes that "Home Capital typically only lends the first 70% of a home’s value. Therefore, if a housing correction were to ensue, the downside risk to Home Capital would be small because the first loss would be absorbed by other lenders or the home owner."
Unfortunately, Golden West's lending was pretty similar: "Golden West did not move into the business of making loans with loan-to-value (LTV) ratios of 90%, 100%, or more, which became an accepted practice in the early and mid-2000s. Golden West’s average LTV ratio remained at about 71%."
Golden West shows-- and I think Home Capital will soon show-- that earning consistently high returns isn't a good thing for lenders. To the extent that they can earn higher returns by operating more efficiently, that's great. But lending prudently often necessitates sacrificing ROE to maintain asset quality and avoid fatal risks.