Newsbite article

Bank of England using computer models to predict Buy-to-Let housing market effect

It’s no exaggeration to say that buy-to-let has fallen out of favour with both politicians and regulators.
 
Warnings from the Financial Policy Committee (FPC) about the risk buy-to-let landlords could pose to the wider housing market manifested in major taxation changes last year designed to make buy-to-let less attractive.
 
First to be announced were changes to the mortgage relief that landlords can claim on their mortgage, followed by a higher rate Stamp Duty charge of 3% for additional property purchases.
 
The Stamp Duty measure was announced in November 2015 and kicked in 1 April 2016 giving buyers a 4-month window to bring forward any purchases to avoid the tax.  Research by the Office of Budget Responsibility estimates that thousands of Brits did exactly that. Some 60,000 transactions were brought forward in that period, generating a net tax loss of £300 million as canny buyers evaded the 3% surcharge to buy an additional property.
 
From Q2 onwards there have been reports of landlords switching corporate to buy-to-lets – specialist brokerage Mortgages for Business has reported that for Q3 2016, 63% of its approvals used a limited company structure, up from just 21% pre- July 2015.  And at a recent conference the FCA stated that it had seen “a massive correction” in the buy-to-let market from the end of Q1, with it arguing that buy-to-let was no longer the growth area it once was. 
 
The FPC and Bank of England have continued to round on buy-to-let from a regulatory standpoint this year. In addition to new data requirements, there are new underwriting standards for buy-to-let lending, in particular harsher requirements for landlords with four or more properties.
 
Trade groups representing landlords campaigned hard to get the Government to overturn the decision in the run up to the Autumn Statement, which ultimately proved unsuccessful.
 
At the last FPC meeting in September buy-to-let was once again earmarked as an area of the market to watch given the “pro-cyclical behaviour of buy-to-let investors” amplifying housing market movements.
 
Lenders and trade bodies continue to point out that the buy-to-let sector has a pretty solid track record. However, some of the most damning Bank of England research on the buy-to-let sector does not rely on past historical data of loan performance.  Instead it’s based on computer algorithms called Agent Based Modelling (ABM) that are used to predict the future behaviour of participants in a system.
 
The concept behind ABM has technically been around since the 1940s but only started to be applied in a significant fashion following the development in computing from the 1990s onwards.
 
ABMs are computerised models that replicate the actions and interactions of “autonomous agents” with the aim of working out what effect they have on the overall system.  As Bank of England chief economist Andy Haldane explained in a recent speech, the technique has already been applied to all manner of sectors, from military planning to biology, the physical sciences and ecology.
Following the financial crash, and the general abandonment of the idea that markets have a natural equilibrium, ABMs have started to be applied to financial markets as well.
The Bank of England recently published the results of its own ABM study on the UK housing market.
 
Unsurprisingly, given the continued negative approach to the sector, buy-to-let doesn’t come out well in the study.
 
But interestingly the research does note that you get two different results where landlords are motivated by either capital appreciation or rental income.
 
  • Where landlords are motivated by capital appreciation, there is “pro-cyclical feedback”. In other words, landlords buy when prices are rising and sell when prices fall, exaggerating peaks and troughs.
  • By contrast, where landlords are motivated by rental income, the study found it can actually have a stabilising effect on overall housing markets. This is because as more buy-to-let landlords pile into the market as house prices rise, increased supply reduces rents. This in turn leads to more people choosing to rent rather than buy, with the lower rents even encouraging homeowners to switch to renting.
     
As the Bank of England states in its report, ABMs do have some drawbacks. However, on the opposing side it argues that a model of the housing market can be useful to policy-makers, with the Loan to Income (LTI) cap that was recently introduced into the market cited as an example.
 
With the Bank of England already looking to test other parts of the mortgage market using ABMs, computer based models could become a key part of how regulators weigh up the relative strengths and weaknesses of the UK’s market. 
 
In the meantime, we remain of the view that a healthy housing market has within it a range of different tenures with the private rental sector retaining an important place within the mix.  With around a third of this sector currently funded by buy to let, it is an area that we and others are watching carefully.