Different conditions preceding this recession may help cushion effects on the housing market, compared to 2008, but protracted lockdowns may create a worst-case scenario, explains Martin Beck.
The housing market surfacing from the last few months of enforced hibernation looks very different to the one we saw at the start of the year.
For much of the period from 2016 to 2019, the market was characterised by low and stable levels of transactions and slowing price growth. But it had flickered into life in early-2020, with mortgage approvals hitting a six-year high in February and most indices reporting a pick-up in house price inflation.
However, housing market activity virtually ground to a halt after the lockdown was implemented in late-March. Just 9,273 mortgages were approved for house purchase in May, some 87% down on the February level.
Following the lifting of restrictions on viewing houses and other activities related to buying residential property in mid-May, an increase in approvals to 40,000 meant June marked the start of a recovery.
But while a relaxation of lockdown restrictions may remove some of the physical barriers to house buying that the pandemic has raised, financial impediments are likely to have a more enduring effect. Our modelling suggests household income growth, the level of unemployment, and housing affordability are all key determinants of the number of mortgage approvals, and all three of these drivers are likely to be unsupportive over the next few years.
Though we think the UK’s Coronavirus Job Retention Scheme will limit the increase in joblessness, we still expect unemployment to rise to around seven percent by the end of the year. As well as the direct impact this sharp increase in unemployment will have on household incomes, it’s also likely to weigh on the pace of wage growth.
Furthermore, despite the more modest house price growth of recent years, affordability metrics have become increasingly stretched, with the price-to-income ratio back at pre-financial crash levels and the average income multiple for loans being higher than ever.
The chancellor has attempted to kick-start housing activity via a temporary cut in Stamp Duty, which will last until March 2021. But we have long been sceptical about the efficacy of Stamp Duty holidays in boosting activity, and the temporary change is unlikely to materially affect the outlook.
What’s more, while very low interest rates will continue to be supportive, this will only mitigate some of the drag from the weak labour market backdrop. Our modelling suggests mortgage approvals and housing transactions are unlikely to return to pre-pandemic levels on a sustained basis until 2023. Lower activity will also weigh on gross lending flows, which we only expect to return to pre-pandemic levels over a similar timeframe.
The weak labour market backdrop will also weigh on prices. But at just over seven percent, we expect the fall in prices to be smaller than in the two most recent downturns. This is partly because we have not seen the types of house price booms that preceded the two previous crashes. But it also reflects our view that these crashes were caused by factors which ‘forced’ greater numbers of borrowers to sell – conditions which are not present this time around.
Lessons from recent history
The prolonged correction of the late 1980s and early 1990s came against a backdrop of high mortgage interest rates and unemployment. Mortgage interest rates of 15% saw debt servicing costs peak at 11% of household disposable income in 1990. But today’s backdrop is more benign, with interest rates set to average just over two percent over the coming years, limiting debt servicing costs to one to two percent of income. Widespread use of mortgage holidays should further limit the degree to which owners are forced to sell.
The sharp drop in prices during the financial crisis also followed an increase in debt servicing costs, but this was accompanied by a significant tightening in credit conditions. In 2020, the Q1 Bank of England Credit Conditions Survey suggested that mortgage providers planned to reduce the supply of secured credit in Q2, particularly for high loan-to-value products, and there has been some evidence of this happening.
But our credit model suggests that the scale of write-offs is likely to be much lower than in the financial crash – largely due to the much lower debt servicing burden. And with the Bank of England keen to ensure the flow of lending is maintained, we expect the degree of credit rationing to be far less than during the financial crash.
While our forecast projects a relatively benign outcome compared with previous cycles, there is still huge uncertainty around the evolution of the pandemic, which skews the risks toward a worse outcome. Three of the four alternative scenarios we have modelled in our global scenarios service would result in higher unemployment and lower household incomes than the baseline and, thus, bigger falls in house prices.
Our risk-weighted forecast implies a 15% peak-to-trough decline in prices. In a worst-case scenario, where protracted lockdowns deepen the global downturn and trigger a financial crisis, the financial stress experienced by households would be compounded by a severe tightening in credit conditions. Under these conditions, the fall in house prices could be close to 40%.
The coronavirus lockdown caused the housing market to grind to a halt over the spring. Mortgage approvals fell to only 9,273 in May, almost 90% below the February level and around a third of their trough during the financial crisis in 2008. Gross mortgage lending averaged only £14bn over April and May, compared to around £23bn per month at the start of the year.
However, activity has seen a rebound since restrictions on home viewings were relaxed in May. Mortgage approvals recovered to 40,010 in June, and the same month saw gross mortgage lending pick up slightly to £16bn.
We expect a smaller peak-to-trough fall in house prices than in the two previous property market cycles. This is partly because we have not seen the types of house price booms that preceded the two previous crashes.
But it also reflects our view that these crashes were caused by factors which ‘forced’ greater numbers of borrowers to sell, notably very high mortgage interest costs and a rapidly rising burden of interest payments relative to household incomes – conditions which are not present this time around.