Analysis suggests that in a global housing slump, the worst hit economies would include the Nordic countries, New Zealand, Australia and the UK.

Risks of a global housing slowdown appear to be growing. We recently looked at how a downturn could affect global growth through different channels, including wealth effects and reduced housing investment.

We use the Oxford Global Economic Model to estimate the potential impact of such a slowdown on individual economies. We identified high valuations and strong recent housing investment as two key factors that drive housing downturns. Among a sample of OECD countries, the most vulnerable, with high house price valuations and high recent rates of housing construction, look to be Sweden, Norway, New Zealand and Canada – followed by Australia, the UK and Denmark.

The least risky appear to be Portugal, Japan and Italy. Importantly, large economies such as the US and Germany also look relatively low risk. Housing upturns are more likely to end as their duration increases, consistent with the notion that housing booms represent departures from fundamentals that become increasingly hard to sustain. Over a 40-year period, the median length of housing upturns is 21 quarters and the 75th percentile is 32 quarters. On this basis, a number of economies in our sample look vulnerable: New Zealand, the US, the UK, Denmark and Korea have all enjoyed long upturns above the historical median – with the first two around the 75th percentile. For economies where prices are already falling, we also find that the average downturn lasts about four years.

So, with the exception of Italy, all the current downturns are relatively young. To model housing downturn risks, we first examine how well the cross-country pattern of (real) house price changes correlated with key risk factors during the last housing downturn between 2006 and 2013. Real house prices were clearly linked to peak house price valuations before the downturn began. The peak level of valuations explains around 35% of the cross-country pattern of real house price changes in 2006-2013. Economies with stretched valuations such as Spain, Ireland and Denmark, saw the biggest price declines peak-to-trough, while the smallest price falls were in low-valuation markets including Germany and Japan.

We have modelled two scenarios of differing severity. A pessimistic scenario simply takes the relationships between the risk factors (valuation, recent housing investment) and house prices and housing investment in the last downturn and applies these to the current levels of the risk factors for each economy.

In this scenario, house prices in our sample fall on average by 9% (with drops of up to 20% in some economies) and housing investment declines by 18% on average, relative to levels in our baseline forecast. The pessimistic scenario would be the second worst housing downturn in the last 40 years or so and is probably too severe: some factors that drove the downturn in 2006-2013, such as banking crises and acute credit crunches, look unlikely to be repeated. A more moderate scenario scales down the impacts on house prices and housing investment to a third of the levels in the pessimistic scenario, based on the fact that the 2006-2013 downturn was around three times as severe as the average of all downturns since 1981-1982.

So, on average, house prices in our sample fall around 3% and housing investment by 6%, relative to baseline. We run both scenarios through the Oxford Global Economic Model between Q3 2019 and Q4 2021. The results reflect the scale of the individual country shocks, model parameters such as the wealth effects embodied in consumption equations, the importance of consumer spending and housing investment in GDP and scope for monetary policy to adjust. For the Nordics, the UK, New Zealand and Australia, the level of GDP in 2021 is 0.5%-0.7% lower, relative to our baseline forecast in the moderate scenario and 1.5%-2.1% lower in the pessimistic scenario.

A relatively large impact in Korea reflects above-average initial housing investment levels and greater sensitivity to lower growth in key trading partners. Spain, Portugal, Japan and (importantly) the US suffer the lightest impact, with levels of GDP falling by just 0.2%-0.3% in the moderate scenario and 0.6%-0.8% (relative to baseline) in the pessimistic scenario. A key factor here is low initial levels of housing investment, in turn related to steep declines in house building in the last downturn.

Adam Slater is lead economist at Oxford Economics.

Originally published in Economia on 6 September 2019.