There is some blunt commentary over the quality and adequacy of the modelling of home loan risks in the Financial Stability Review published this week by the Reserve Bank of New Zealand.The context to this commentary is the introduction from the beginning of 2008 of the Basel 2 regime that bank regulators use to work out how much capital banks must hold.For all the bank bravado over how little Basel 2 would matter - on the grounds that it was a combination of their own estimates and ratings agency requirements that really dictated bank capital - it already appears that the new Basel formulas are fine by bank management, especially when the capital ratios turn out to be on the high side.In the case of New Zealand banking, the local regulator isn't impressed with the combined work of the nation's banks (which are, of course, also Australia's banks).One consequence is that banks are required to hold extra capital equal to 15 percent of the capital the bank's own models suggested was appropriate for credit risk from mortgage lending. This, the RBNZ said, was "to recognise that improvements are needed". The RBNZ wrote in the review, published on Wednesday, that during the process of accreditation of banks' approach to Basel 2 "we had concerns about the banks'housing risk models".Issues raised by the Reserve Bank included:• Estimates of    the long run probability of default varied across banks more than could reasonably be expected, given that New Zealand banks generally have a similar customer base.•  The bank PD estimates were generally low in relation to comparable estimates observed internationally, and low on the basis of the Reserve Bank's own modelling results. â€¢  Several PD models    were driven heavily by short-term indicators of distress, rather than long-term risk drivers such as debt servicing capacity. • Estimates of the loss given default - and which drive provisions - were not sufficiently calibrated    to economic downturn conditions and did not include loan-to-value ratios as a risk driver. The program of remedial work for the industry dictated by the RBNZ included:• That banks recalibrate estimates of the probability of default to better reflect a range of economic conditions that could reasonably be expected over the medium to long term. â€¢ Further reviews to determine long-term structural drivers of default risk.• Incorporation of more sensitive estimates of the loss given default and that take into account loan to valuation ratios, and derived from RBNZ rather than a bank's own modellings.• Still more reviews of the sensitivity of LGD models to economic risk drivers. In particular, the RBNZ wants banks to ensure their own models are calibrated to economic downturn conditions that incorporate a fall in average house prices of 30 per cent. Given that New Zealand's four major banks are owned by Australia's four major banks and on the assumption that the subsidiaries shared a great deal of work on modelling with their owners, it seems fair to expect that similar issues may arise in relation to the modelling of