My simple prior was that if you have an SME loan that is secured by, say, your home, and is full recourse to you individually, then, all other things being equal (eg, income serviceability, probability of default, etc), the cost of that loan should be the same as any other home loan. This also assumes that the SME loan is the first ranking security, and not subordinated, and that small businesses cost the same to service as home owners.
Taking the above as given, if there was a differential between the interest rate on SME loans extended on this basis and your bog standard home loan, it would constitute something of a puzzle. Anyhow, I happened to be trawling through the RBA’s always-very-interesting interest rate data the other day and decided to check it out. And boy was I surprised by what I found.
Consistent with the intuition above, small business loans secured with residential property and residential home loans were priced at almost exactly the same rate for almost 11 years prior to the GFC (see the first chart below). To be clear, this comparison begins in 1996 only because that is when the RBA data commences. Over this period the average spread, or difference, between a variable SME loan (residential secured) and a variable home loan was just 0.22% (in percentage point terms, to be precise). Yet by December 2008 this spread had grown at an extraordinary rate to be 6.9x higher (or 1.50%). And while it has contracted a little since then, today the spread between these two products is still 1.20%, or around 5.5x higher than it was in the 11 years before the GFC (see second chart).
This is truly bizarre given that the pre-GFC data suggests that the risks, and hence cost, of these two loan products are near-identical. In particular, there are one of two possible small business loan puzzles: either the pricing of SME loans during the 11 years preceding the crisis was all messed up, or, assuming it was not, something fundamental has happened in the period since. Taking the latter as the more likely candidate for the moment, there are a number of possible explanations:
(1) Small businesses got indiscriminately ‘repriced’ alongside all other business lending during the GFC. Business lending, including SME finance generally, has historically been much riskier than residential lending, and nearly accounted for the collapse of ANZ and Westpac in 1991. Business loans have higher default rates and higher loss ratios. But the historical data implies that a small business loan fully secured by a residential asset, and with all the same characteristics as a home loan, is not really a business loan (more on this point later);
(2) Competition that may have kept small business rates in line with other residential rates evaporated during the GFC and allowed the incumbents to extract much higher margins;
(3) Perhaps there has been some sort of change to the standard small business loan product. This, however, seems unlikely given the ordinarily conservative RBA would have noted any such changes in its data (liaison with industry also discounts this explanation);
(4) Could APRA have altered its treatment of small business loans during the crisis? Business loans typically attract a 100% risk-weighting from APRA, which is much higher than the 25-50% risk-weighting found on home loans (due to the commensurately higher risks associated with business lending). To the best of my knowledge, however, there have been no regulatory changes of late; and/or
(5) Small business loans were cross-subsidising residential mortgages due to the political sensitivities associated with repricing the latter. This would mean that part of the margin expansion evidenced in SME lending should have been borne by residential borrowers, but banks did not have the will to impose it.
After consulting with industry specialists, my best guess is that the answer is a combination of points (1), (2) and (5). If this is indeed true, the small business space is currently offering lenders extraordinary (relative) returns. One would expect to see new entrants compete these margins away over time once the system normalises.
There is another possibility, which is worthwhile dwelling on. We know empirically that SME lending is much riskier than traditional residential credit. This is, for instance, borne out in the losses realised by the major banks during the 1991 recession. What we do not know is how the ‘residential secured’ sub-component of the SME finance portfolio performed. But it is probably reasonable to suppose that the default rates in this sector are higher than the very low probabilities of default associated with traditional residential borrowers. On this basis, one would expect there to be a spread—and potentially a significant one—between the two forms of finance. It is also plausible that it is more costly to service SME loans relative to residential credits.
But if all of this were true, it would suggest that banks got the pricing of SME loans horribly wrong for the best part of 11 years. And so our puzzle remains unresolved. It sounds like something the RBA should sink its very sharp teeth into!
PS: I have noticed some bloggers picking up on my content, writing their on views on the subject, and then not referencing back to their source (or trying to find an alternative source). The standard convention online (or anywhere else, for that matter!) is to acknowledge your sources via an HT: [insert link].
