Impact of LVR restrictions on mortgage portfolio resilience

This page contains information on the impact of LVR restrictions on mortgage portfolio resilience from the November 2017 Financial Stability Report.

The loan-to-value ratio (LVR) policy was first introduced in October 2013, with progressively tighter restrictions for investors introduced in November 2015 and October 2016. The primary goal of the LVR policy is to improve the resilience of the financial system to a significant housing market correction, given the Reserve Bank’s view that risks in this sector have been elevated.

There are a number of ways in which the LVR policy has improved the resilience of the financial system. This box focusses on how, by improving equity buffers of borrowers, the LVR policy reduces the likelihood that mortgage borrowers will default and reduces the magnitude of losses that banks would sustain in the event of default.

Prior to the policy being introduced, the share of banks’ mortgage portfolios with LVRs above 80 percent had steadily increased to 21 percent, posing a risk to financial stability. This reflected that around a third of new loans being originated had LVRs above 80 percent.1 As a result of the LVR policy, the share of outstanding mortgages with LVRs above 80 percent steadily declined to under 8 percent in September 2017 (figure A1). Tighter LVR rules applied to property investor lending have seen the share of outstanding mortgages at LVRs between 70 and 80 percent decline since late 2015.

Figure A1 Share of outstanding mortgages by LVR

Box A Figure A1 FSR November 2017

Source: RBNZ

Note: Vertical lines denote the dates of LVR policy changes.

If there was a major housing market correction or economic downturn, then this reduction in the share of lending at high LVRs is likely to mean that fewer housing loans would default, and overall bank losses would be lower. One way of quantifying this is to use data from recent stress tests to estimate how the change in banks’ portfolios would affect default and loss rates for a given downturn scenario. Based on the 2017 stress test scenario (see box B), we estimate that banks would experience around 10 percent lower default rates and around 20 percent lower credit loss rates than they would have if LVR restrictions had not been applied (figure A2).

Figure A2 Estimated loss rates and default rates (based on 2017 stress test scenario)

Box A Figure A2 FSR November 2017

Source: RBNZ

Aside from this direct impact on banking system resilience, there are two other key channels through which the LVR policy is likely to have improved bank resilience. First, the LVR policy has reduced demand in the housing market, which has been a contributing factor in slowing house price growth. This has left the housing market less exposed to a sharp correction in prices. Second, the improvement in households’ equity buffers means that fewer households will need to sell their house or cut back on consumption to meet debt obligations during a downturn. This limits the risk of significant feedback effects that could cause further deterioration in the housing market and broader economy – ultimately lessening the risk of severe credit losses for banks.

It is difficult to quantify exactly how large these indirect effects are. However, results from the recent bank stress tests suggest that credit loss rates on mortgage portfolios are highly sensitive to the magnitude of house price declines. For example, banks estimate that credit losses would rise by around a quarter if house prices declined by an additional 5 percent (from the scenario baseline of a 35 percent fall in house prices) – suggesting that these indirect effects of the LVR policy could potentially be large.

As noted in chapter 1, a prerequisite for easing LVR restrictions is that it would not undermine the resilience of the financial system. Our assessment is that the current restrictions are continuing to reduce the share of mortgages at high LVRs. The easing of the policy from 1 January will allow a slightly larger flow of high-LVR loans to be granted. Nevertheless, the share of bank portfolios with LVRs above 70 percent is still expected to trend down slightly, as the new policy settings will remain relatively restrictive compared to pre-LVR lending flows. In addition, if housing pressures continue to moderate, this will reduce the risk of a significant house price correction. As a result, the resilience of bank mortgage portfolios is not expected to diminish as a result of the announced policy easing.

1 Counterfactual modelling suggests that the share of outstanding mortgages at LVRs above 80 percent would have increased slightly had the share of new loans at LVRs above 80 percent remained at a third.