RDP 9202: Some Tests of Competition in the Australian Housing Loan Market 5. Conclusions

We have estimated a conjectural variations model of bank lending for the owner-occupied housing market. We are able to decisively reject the hypothesis that this market is perfectly competitive, as well as the hypothesis that banks collude to form a cartel. We are unable to reject the hypothesis that the market for housing loans is a Cournot oligopoly, although this conclusion could be plausibly challenged on statistical grounds.

Of course, we do not view this paper as the final word on this subject. The imprecision of the parameter estimates on the interest rates indicate to us that more sophisticated supply functions ought to be estimated, as well as perhaps giving greater emphasis to dynamic specifications. Furthermore, banks may well compete in ways we have not modelled, e.g. by price competition, and this needs to be accounted for in a complete study of competition in this market.

Finally, we note that the degree of competition in any market is determined endogenously. If the market for housing loans is not competitive, the interesting issue for policy is to determine why this is the case. Future research needs to examine issues such as barriers to entry, transactions costs and consumer inertia as impediments to the existence of competitive markets.