A New Direction in the Measurement of Household Debt Burdens
Another month, another, even gloomier, forecast from the Bank of England. Yesterday, as the Bank raised the base rate to 1.75%, its Monetary Policy Committee noted that:
“The United Kingdom is now projected to enter recession from the fourth quarter of this year. Real household post-tax income is projected to fall sharply in 2022 and 2023, while consumption growth turns negative.”
The Bank also released its forecasts for the oncoming recession, with Financial Times[i] analysis indicating that this will last for over a year, “match that of the early 1990s” and include “the biggest fall in household incomes for more than 60 years.”
However, it could be even worse if the Bank is underplaying the extent to which households are likely to default on mortgage and credit commitments. As is now widely accepted amongst economists, high debt levels are a “key determinant”[ii] of the intensity of recessions.
Problems of Measurement
Measuring the extent, and distribution of, household debt burdens accurately is therefore critical if we are to anticipate how this recession may unfold and take appropriate action to mitigate it.
However, the traditional measures used by the Bank and other policymakers – which are based on debt repayments relative to gross incomes – do not capture the true burden of debt when the cost-of-living is rising, and therefore the amount of disposable income from which to make repayments is falling.
We have made this point consistently over the past six years. In our report ‘Britain in the Red’, published by the TUC in 2016, we urged policymakers to adopt an alternative measure based on “total interest payments as a proportion of the household surplus”. And in 2020, we published a paper[iii], subsequently shared with HM Treasury, pointing out that:
“Measures of debt to income and debt servicing to income ratios ignore the impact of cost-of-living increases on household budgets. In the UK, the measures are becoming increasingly redundant as a means of assessing the financial health of households, and the thresholds used by the Bank of England and other policymakers appear to be set at far too high a level to provide any real guide to policymaking.”
A response from the Bank
It now appears the Bank has belatedly come to the same conclusion. The Financial Stability Report of July, reveals[iv]:
“Staff have developed a new measure of household debt affordability using the share of income available to repay debt, after adjusting for an estimate of essential spending and taxes. This enables better assessment of the combined impact from rising prices and interest rates.”
We very much welcome this development. We are, however, concerned that the precise method used by the Bank to calculate its “adjusted debt servicing ratio” - which it calculates separately for mortgage and consumer credit debts - continues to underplay the burdens of UK households and the potential financial stability risks involved.
What’s wrong with the new measures?
The Financial Stability Report provides limited detail of the new adjusted debt servicing ratios (DSRs). However, it tells us that its estimate of essential spending “includes utility and council tax bills, housing maintenance, food and non-alcoholic beverages, motor fuels, vehicle maintenance, public transport and communication.”
Whilst a reasonable start, this definition apparently ignores repayments on student loans – although these may be netted off in the calculation of taxes, as they are recovered through PAYE – and other key essentials, such as childcare, clothing and footwear, personal care items and prescriptions, and household appliances.
If absent from the Bank’s denominator, these are major omissions. For example, childcare costs alone average £138 per week - over £7,000 per year - for a part-time nursery place for a child under two[v].
We also have concerns about the ‘thresholds’ used by the Bank to determine when households are likely to be struggling with their debt repayments. These appear to have been set at unreasonably high levels. With respect to the ‘adjusted mortgage to debt servicing measure’ this is set at 70%, and for the ‘adjusted consumer credit to debt servicing measure’ the threshold is 80%.
The Financial Stability Report states that these have been estimated by taking “the thresholds at which households become much more likely to experience repayment difficulties for gross DSRs" and adjusting these to reflect the share of income spent on taxes and essentials.
But the report is silent as to the precise nature of the adjustment that has taken place.
What do the thresholds mean in practice?
In our opinion, the thresholds appear too high. Take, for example, a household with median gross income (in 2021) of £31,400. According to the Bank’s estimates[vi], these households spend around 50% of their gross incomes on taxes and ‘essentials’. This reduces their disposable income to £15,700 or £1,308 per month.
Assuming these households have a mortgage, then they would need to spend 70% of their disposable income on repayments to meet the bank’s threshold: equivalent to a repayment of £915 per month. And, if the household holds both mortgage and consumer credit debts, they will need to be paying £1,046 out on these before the bank would consider them to be in danger of default.
In the case of households with mortgages only, this would leave just £393 per month for ‘non-essential’ spending, or £90 per week to cover all the items listed above, which the Bank’s definition appears to ignore.
For households with both a mortgage and consumer credit debt, the amount left per week would be just £60.
Lower income households
The position is, unsurprisingly, worse for lower income households. The Bank estimates that those in the lowest income decile are currently paying out 90% of their disposable incomes on items it considers to be ‘essentials’. For the second-and third-income deciles, the Bank estimates 60%.
For these households to meet the Bank’s thresholds, they would have to be left with just 2% and 8% of their disposable incomes to spend on 'non-essentials', after taking account of debt repayments.
Is it realistic to expect that lower-income households won't default just because they have, say, between 10% and 15% of their disposable incomes left?
Downplaying the potential impact?
Finally, the Bank also appears to downplay the financial stability risks of increased default rates amongst low to middle income households, arguing that:
“These households are likely to hold a smaller share of total outstanding consumer credit and mortgage debt.”
Whilst this is certainly true, the levels of debt that households in the lower half of the income distribution hold is still, in our view, cumulatively significant. Taken together, these households hold approximately[vii] 12% of all mortgage debt and a fifth of consumer credit debt.
In the 1990’s recession, a mortgage repossession crisis was sparked by only around 2% of mortgages entering prolonged arrears of 6 months or more[viii]. And if the cost-of-living crisis and coming recession start to impact on those households only slightly further up the income distribution, the likelihood of systemic problems increases. For example, those in the sixth income decile account for 7% of mortgage debt, and over 10% of consumer credit liabilities.
What difference could better measurement make?
The Bank’s strategy for dealing with the cost-of-living crisis is heavily reliant on raising interest rates. This has the effect of taking money out of mortgage holders’ pockets – immediately for those on variable rates, but deferred for those on fixed rate deals. However, it may also provide all households with some relief by stopping Sterling’s further depreciation. If this doesn't reduce, then it may at least hold steady, the price of imported goods.
In this sense the strategy is reminiscent of Nigel Lawson's focus on interest rate rises as a means of defending the exchange rate prior to the 1990's recession itself. The problem then, however, was that the level of rate rises needed was such that highly leveraged mortgage borrowers could no longer afford their repayments.
Reminiscent, certainly, but not exactly the same. This time the interest rate rises are more modest, and action in recent years to restrict mortgage lending relative to income will also mitigate the risks of default to some degree.
However, even with its strategy of, admittedly more modest, interest rate rises in place, the Bank is forecasting that inflation will hit 13% by the end of the year, and that we will enter a recession that will last for five quarters. It is then looking for increased unemployment, and cutbacks in household consumption to start to pull inflation down over the medium term. In short, it sees a lengthy period of falling living standards as necessary in the battle against inflation. Were it confident that higher interest rates would do the trick without triggering mass mortgage payment problems, it would be going harder and faster with them.
As Ross Walker, economist at NatWest Markets, put it in the Financial Times yesterday:
“The immediate inflation outlook is now so dire that the Monetary Policy Committee feels it has not option but to engineer a more severe economic downturn”.
However, in making this decision, the Bank may be underplaying the risks for financial stability. It is certainly not impossible that inflation and recession combine over the next year to trigger large-scale default.
If its measures predicted a significant growth in mortgage arrears, and the potential collapse of financial institutions under the weight of rising defaults and falling house prices – as happened in 2008 - the Bank would now likely be looking for alternatives to rate rises. For example, it might be working harder to hold down prices in the first place, such as by providing interest free liquidity for key service providers (such as local authorities, housing associations, and genuinely stretched energy providers) in return for their commitments to cap Council Tax, rents, and energy prices on a temporary basis.
Ultimately, decisions about which direction to take are made best when the compass is accurate. We therefore call on the Bank to make public the full methodology involved in the calculation of its “adjusted debt servicing ratios”, and its thresholds.
[i] ‘Bank of England serves up a shock with its intensely gloomy outlook’, Financial Times, 4th August 2022
[ii] See, for example, Lombardi et al (2017), ‘The real effects of household debt in the short and long run’, Monetary and Economic Department, Bank for International Settlements.
[iii] The paper was published in the German Quarterly Journal of Economic Research, ‘Vierteljahrsheftezur Wirtschaftsforschung’ available at https://elibrary.duncker-humblot.com/article/28624/unsustainable-household-debt-problems-of-measurement
[iv] Para 1.4.1 Financial Stability Report July 2022, Bank of England - https://www.bankofengland.co.uk/financial-stability-report/2022/july-2022
[v] See ‘Childcare costs rise by 4% over the last year as providers struggle to remain sustainable during the pandemic’, Coram Family and Childcare, 8th March 2021Available at:https://www.familyandchildcaretrust.org/childcare-2021-press-release
[vi] As set out in Chart 1.2 of the Financial Stability Report
[vii] Again, based on Chart 1.2 of the Financial Stability Report
[viii] See, for example, Aron & Meullbauer (2010), ‘Modelling and forecasting UK mortgage arrears and possessions’, Department for Communities and Local Government. Available from https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/16077/1643676.pdf