…for debt vulnerabilities to return to GFC levels?
The key chart
Share of UK HHs with mortgage COLA-DSRs above 70% (Source: BoE Financial Stability Review, July 2023)
The key message
How high would the UK base rate have to increase for HH debt vulnerabilities to return to GFC levels? To around 9% according to the Bank of England today, up from a broader, estimated range of 5-8% indicated a year ago…
The Bank of England (BoE) held a press conference regarding its July 2023 Financial Stability Review (FSR) this morning (12 July 2023). As always, the question and answer session was particularly important not least in the request for clarification on rising HH debt vulnerabilities.
Recall that a year ago, Sir Jon Cunliffe, the Deputy Governor for Financial Stability, was asked for clarification on this topic – how far would interest rates have to rise for HH debt vulnerabilities to reach the GFC peak levels?
In response, he indicated that rates would have to rise by between 200-500bp above the existing market expectations. These were 3% at the time. In other words, the base rate would need to return to between 5% and 8% for debt vulnerabilities to return to the GFC levels. Today’s UK base rate is 5%, at the bottom end of this range.
In the latest FSR, the Bank of England returned to this topic. The Bank stated that, “The proportion of HHs with high mortgage cost-of-living-adjusted debt-servicing ratios (COLA-DSRs) is expected to continue to increase [from 2% in 1Q23] to around 2.3%…by the end of the year. But it would stay below the recent peak reached in 2007 of 3.4%.” (see key chart)
Importantly, the report then notes that, “To reach that peak level by the end of 2024, it would require mortgage rates to be around three percentage points higher relative to current expectations [c.6%], other things being equal.”
In short, the BoE has narrowed the range and increased the level of the base rate that would lead to a return of GFC levels of HH debt vulnerabilities – to 9%.
The main reason, according to Sir Jon Cunliffe, is the level of support provided by banks in terms of extending mortgage terms, changing mortgage terms and supporting moves to interest-only mortgage. In other words, more options and support from banks has given slightly more “breathing room” before debt vulnerabilities return to their recent 2007 peak.
A year ago, I felt that the BoE was verging on complacency in terms of rising HH debt vulnerabilities. Ninety days later the tone of the message changed and higher risks acknowledged. Today, we have a narrow and clearer “risk target” but the body language was less-than-convincing, in my opinion.
Trends and breakdown of UK FIRE-based (red and pink) and COCO-based (blue) lending (£bn)(Source: BoE; CMMP)
After two decades of fuelling FIRE-based lending, is it time to ask, “what is the purpose of UK banking?”
Bank lending falls into two categories: lending that supports capital gains largely through higher asset prices (FIRE-based); and lending that supports production and income formation, i.e. productive enterprise (COCO-based). The former includes mortgage or real estate lending and lending to NBFIs. The latter includes corporate lending and consumer credit.
Trends in the breakdown of UK sterling lending since 2002 (% total) (Source: BoE; CMMP)
Twenty years ago, less productive FIRE-based lending accounted for 67% of M4L, the Bank of England’s headline credit series. Today it accounts for 78%. The contribution of mortgages, the largest component of FIRE-based lending, has risen from 47% to 53% over the period. This means that nearly 80 pence in every pound lent by UK MFIs finances transactions in pre-existing assets (real estate) or in financial assets. Note that mortgages are the only component of M4L to register a record high at the end of 2022.
In contrast, only just over 20 pence in every pound lent in the UK finances productive enterprise. The contribution of lending to corporates, the largest component of productive COCO-based lending has fallen from 20% to 17% over the past two decades. Note that this lending supports sales revenues, wages, profits and economic expansion. Lending that increases debt in the economy BUT critically also increases the income required to finance it. The outstanding stock of sterling loans to UK corporates at the end of 2022 (£445bn) was £61bn or 12% lower than its peak (£516bn) recorded in August 2008.
Lending in any economy involves a balance between these different forms. A key point here is that the shift from COCO-based lending to FIRE-based lending as seen in the UK reflects different borrower motivations and different levels of risks to financial stability. This has negative implications for leverage, growth, financial stability and income inequality.
Classifying lending according to its productive purpose tells us what the purpose of UK banking is today – largely to support capital gains rather than production and (direct) income formation.
The obvious follow-on question is, “what should it be?”
Please note that the summary comments and charts above are abstracts from more detailed research that is available separately.
The BoE introduces an improved measure of affordability
The key chart
Share of outstanding mortgages and consumer credit by income decile (Source: BoE; CMMP)
The key message
Rising financial inequality means that lower-income HHs have less flexibility to adjust their spending in response to rising prices and are less likely to have a cushion of savings to protect them. But what does this mean for debt vulnerability in the UK, given that lower-income HHs also hold a relatively small share of outstanding mortgages and consumer credit?
In its latest “Financial Stability Review” (5 July 2022), the Bank of England introduced a new measure of HH debt affordability that takes account of these factors to deliver an improved assessment of HH vulnerability to rising prices and higher interest rates. The key points are:
The share of HHs with high, adjusted debt service ratios (DSRs) i.e. those who are typically more likely to struggle with repayments, is currently in-line with historic averages and well below pre-GFC peaks
The BoE believes that this share is unlikely to rise substantially in 2022. Further fiscal measures will cushion serviceability this year and the shift towards fixed mortgages (80% outstanding stock) delays the pass through effect of higher rates
Looking forward, the BoE expects this share to increase above its historic average in 2023, but to remain “significantly below the peaks seen ahead of the GFC.”
[In response to questions, Sir Jon Cunliffe, the Deputy Governor for Financial Stability, indicated that rates would have to rise significantly (200-500bp) above current market expectations for the bank rate (3.0%) for the share to reach previous highs.]
The obvious risks to this positive assessment, acknowledged by the BOE, include weaker growth, higher unemployment, persistent inflation, higher rates etc.
In the context of the recent deterioration in the net lending position of the HH sector, the risk that HHs may increasing their borrowing in order to fund rising living costs also remains important.
Financial inequality and debt vulnerability
What is the impact of financial inequality on household (HH) debt vulnerability?
Financial inequality means that lower-income HHs have less flexibility to adjust their spending in response to rising prices and are less likely to have a cushion of savings to protect them.
Share of income spent on taxes and essentials by gross income decile (Source: BoE; CMMP)
The share of income spent on essentials and taxes varies considerably across the UK income distribution (see chart above). HHs in the lowest income decile, for example, spend 94% of their gross income on taxes and essentials. In contrast, HHs in the highest income decile spend only 47% of their gross income on taxes and essentials. This means that lower-income HHs have much less freedom to change their behaviour in response to rising inflation.
HH savings ratio (%) broken down by gross income decile (Source: BoE; CMMP)
The savings ratio of lower-income HHs is also much lower than the savings ratio of higher-income HHs (see chart above). HHs in the bottom three income deciles save less than 6% of their gross income. This contrasts with HHs in the top two income deciles who save more than 30% of their gross income.
Obviously, this means that lower-income HHs are much less likely to have a savings cushion than can protect against rising prices. ( Note also that the COVID-19 pandemic led to a further widening of the savings disparity in the UK, as noted in previous posts.)
Share of outstanding mortgages and consumer credit by income decile (Source: BoE; CMMP)
That said, lower income HHs are also likely to hold a smaller share of both outstanding mortgages and consumer credit (see chart above).
Share of outstanding mortgages by income decile (Source: BoE; CMMP)
The bottom three income deciles account for 1.6%, 1.2% and 2.3% of total outstanding mortgages respectively, a cumulative market share of just over 5% (see chart above). The top income decile accounts for 33% of total mortgages alone and the top three income deciles together account for more than two-thirds of total mortgages.
The breakdown of consumer credit follows a similar, if less extreme, pattern. The bottom three income deciles account for 1.5%, 3.0% and 4.0% respectively, a cumulative market share of just over 8% (see chart below). The top income decile accounts for 20% of total consumer credit alone and the top three income deciles accounts for more than half of total consumer credit.
Share of outstanding consumer credit by income decile (Source: BoE; CMMP)
In its latest “Financial Stability Review”, the Bank of England introduces a new measure of HH debt affordability that takes account of these factors to deliver an improved assessment of the HH vulnerability to rising prices and higher interest rates.
The BoE claims that, “the share of HHs with high cost of living adjusted DSRs on either their mortgage or consumer credit has remained significantly below the pre-GFC peaks over the past few years.”
Share of HHs with high adjusted DSR on mortgage debt (Source: BoE; CMMP)
In the case of mortgages, the BoE estimates that 1.7% of HHs had a high, adjusted DSR at the end of 1Q22, up from 1.4% in 1Q20. This is close to the historic average but below the pre-GFC peak of 2.8% (see chart above). The BoE is expecting this share to remain at around the current level for the rest of 2022. This assumes that government support measures will relieve the pressure of rising living costs and also reflects the fact that 80% of outstanding mortgages are fixed rate now versus 55% five years ago.
Share of HHs with high adjusted DSR on consumer credit (Source: BoE; CMMP)
The story for consumer credit is much the same. The share of high, adjusted DSRs for consumer credit was 6.4% in 1Q22, up from 5.5% in 1Q20. This is also well below the pre-GFC peak of 9.5%. Again, the BOE is not expecting a major change here during the rest of 2022.
Looking slightly further ahead, the BoE believes that the shares of HHs with a high, adjusted DSR for both mortgages and consumer credit will increase in 2023, but “would remain significantly below the peaks seen ahead of the GFC.”
In response to questions, Sir Jon Cunliffe, the Deputy Governor for Financial Stability, indicated that rates would have to rise significantly (200-500bp) above current market expectations for the bank rate (3.0%) for the share to reach previous highs.
The obvious risks to this positive assessment, acknowledged by the BOE, include weaker growth, higher unemployment, persistent inflation, higher rates etc. In the context of the recent deterioration in the net lending position of the HH sector, the risk that HHs may increasing their borrowing in order to fund rising living costs also remains important.
Please note that the summary comments and charts above are abstracts from more detailed analysis that is available separately.
Trends in net lending/net borrowing, EURbn, rolling 4Q sum (Source: ECB)
The key message
Comments made by the Governor of the Banque de France in Paris last week (1) confirm that conventional macro thinking continues to (largely) ignore private debt while seeing public debt as a problem, and (2) suggests that reports of the death of out-dated fiscal rules in the euro area (EA) are premature.
What did he say? The Governor rejected arguments that (1) accommodative monetary policy was responsible for the rise in public debt, and (2) that “because of this high public debt, monetary policy is now unable to raise interest rates sufficiently to combat inflation”. He stressed that central bank independence was “notably designed to prevent any risk of fiscal domination.” The rest of the speech focused on why debt must remain a key issue and the future EA fiscal rules.
From a CMMP analysis perspective, there were three extraordinary features of the speech:
First, in discussing the exceptional (fiscal) response to exceptional circumstances, the Governor ignored the similarly exceptional disinvestment by the French private sector;
Second, and linked to this, he suggested that France “could keep the 3% deficit target, which is as a “useful anchor” and even the 60% debt target”;
Finally, he chose not to refer to the elevated risks associated with the level, growth or affordability of risks associated with French private sector debt, particularly in the corporate (NFC) sector..
Why does this matter? The Governor’s speech follows similar arguments presented earlier this year by the French state auditor. In both cases, the level of public sector debt was viewed as a problem but private sector debt was ignored, confirming a fundamental flaw in conventional macro thinking. The support for out-dated and arbitrarily determined fiscal rules also means that the risks of deficit reductions compounding further private sector deleveraging in the future remain.
Plus ça change, plus c’est la même chose…
Sonnez l’alarme II – the charts that matter
Trends in French public sector net borrowing (EURbn) since 1999 (Source: ECB)Trends in French and German government debt ratios (% GDP) versus fiscal target (Source: BIS) Trends in French private sector net lending (EURbn) since 1999 (Source: ECB)The view from a sector balances perspective (Source: ECB)Breakdown (% total) of French credit to the non-financial sector (Source: BIS)Rolling 3-year CAGR in NFC credit versus 3-year CAGR in nominal GDP (Source: BIS)French NFC sector debt ratio versus BIS maximim threshold limit (Source: BIS)French NFC debt service ratio versus LT average since 1999 (Source: BIS)EA private and public sector balances, EURbn (Source: ECB)
Please note that the summary comments and charts above are extracts from more detailed analysis that is available separately.
Where are the “real risks” in French debt dynamics?
The key chart
Trends in French debt ($bn) broken down by sector (Source: BIS; CMMP)
The key message
France’s state auditor, the Cour des Comptes, “sounded the alarm” about the impact of pandemic spending on France’s widening budget deficit and rising government debt levels last week. The auditor also raised concerns about potential risks to the cohesion of the EA. Are these concerns justified or do greater risks lie elsewhere within French debt dynamics?
The auditor is correct to highlight the impact of pandemic spending on the government’s net borrowing. This rose from 78bn in 3Q19 to €220bn in 3Q21. The level of government debt and the debt ratio are close to their record 1Q21 highs. The debt ratio has remained above the EA’s 60% GDP threshold for the past two decades and the divergence in the debt ratios of France and Germany in the post-GFC period represents a “potential risk to the cohesion of the EA”.
The first counter argument, based on national accounting principles, is that the correct level of government net borrowing is the one that balances the economy not the budget. The level of private sector net lending rose from €75bn in 3Q19 to €225bn in 3Q21, driven largely by HH net savings of €134bn. In other words, the net borrowing of the government was a necessary, timely and appropriate response to the scale of the private sector’s net lending/disinvestment.
The second counter argument is that that while the outstanding stock of French government debt may be the fourth highest in the world, France ranks lower in terms of government indebtedness. This argument will be more compelling to those who view debt sustainability (correctly) as a flow concept, but much less compelling to those who prefer the traditional stock-based approach.
From a risk and financial stability perspective, we are more concerned about France’s private sector debt dynamics, particularly in the NFC sector. France has relatively high exposure to the NFC sector, the fifth most indebted NFC sector globally. In spite of this, France has seen the third highest rate of excess NFC credit growth globally over the past three years. Affordability risks in the NFC sector are also elevated in absolute terms and in relation to historic trends. While the level of HH indebtedness in France is low in absolute terms, the risks associated with excess HH credit growth and affordability are elevated in this sector too.
The headlines resulting from the Cour des Comptes’ report support our wider hypothesis that conventional macro thinking is flawed to the extent that it typically ignores private debt while seeing government debt as a problem rather than as a solution.
Sonnez l’alarme?
France’s state auditor, the Cour des Comptes, “sounded the alarm” about the impact of pandemic spending on the widening budget deficit and level of government debt in its 2022 Annual Report published on 16 February 2022. The auditor also argued that the government should revise its deficit reduction plans after April’s presidential election, claiming that current plans risk fuelling divergences within the euro area, especially with more fiscally conservative countries such as Germany.
The supporting evidence
Trend in government net borrowing over the past twenty years (Source: ECB; CMMP)
The state auditor is correct to highlight the impact of pandemic spending on the budget deficit. As illustrated in the chart above, the government’s net borrowing rose from €78bn in 3Q19 to €220bn in 3Q21. The current level of net borrowing is also higher than the previous peak net borrowing of €143bn in the aftermath of the GFC. Viewed in isolation, this is a scary chart!
Trends in government debt and debt ratios over the past twenty years (Source: BIS; CMMP)
The level of government debt and the debt ratio are also close to their all-time highs (see chart above). The outstanding stock of government debt rose form €2,727bn in 2Q19 to €3,088bn in 2Q21, slightly below the peak 1Q21 level of €3,099bn. The government debt ratio (the blue line above) rose from 113% GDP in 2Q19 to 128% GDP in 2Q21. Again, the debt ratio also peaked at 134% GDP in 1Q21. Note that throughout the past two decades, France’s government debt ratio has exceeded the EA’s threshold of 60% GDP.
Trends in French and German government debt ratios (Source: BIS; CMMP)
The debt ratios of France and Germany have been on different trajectories for most of the post-GFC period. The German government debt ratio peaked at 86% GDP in 4Q12 and declined to a recent low of 64% GDP in 4Q19 (still above the EA’s threshold level). So again, the auditor is correct to highlight this as a “potential” source of risk to the cohesion of the EA.
The counter arguments
The first counter argument here is a simple one – the correct level of government net borrowing is the one that balances the economy not the budget. It is a basic principle of national accounting that the net borrowing of one economic sector (in this case the French government), must be equal to net lending of one or other economic sector(s) (see “Everyone has one”).
Private sector net lending versus public sector net borrowing (Source: ECB; CMMP)
The chart above plots the net lending of the private sector and the net borrowing of the public sector together. The level of private sector net lending, or disinvestment (the blue area), rose from €75bn in 3Q19 to €225bn in 3Q21, driven largely by HH net savings of €134bn and FI net savings of €125bn.
In other words, the increase in the government’s net borrowing position essentially matched the increase in the private sector’s net lending position. Rather than a source of alarm, the spending response of the French government was necessary, timely and appropriate.
The second counter-argument is that while, the outstanding stock of French government debt may be the fourth highest in the world, France is ranked lower in terms of government indebtedness (with a debt ratio similar to the UK).
Top 10 BIS reporting economies ranked by total government debt (EURbn) (Source: BIS; CMMP)
The chart above ranks the top ten BIS reporting countries in terms of outstanding government debt. Total government debt was $3,670bn at the end of 2Q21, representing a 4% share of global government debt after the US (33%), Japan (14%) and the UK (5%). The chart below ranks the top ten BIS reporting countries in terms of the government debt ratio. In this case, France’s ranking drops to #8 globally and #7 in Europe after Greece, Italy, Portugal, Spain, Belgium and the UK.
Top 10 BIS reporting economies ranked by government debt ratio (% GDP) (Source: BIS; CMMP)
This second counter-argument will be less persuasive for those who view debt sustainability as a stock concept (the traditional approach). They will point to the fact that France’s government debt ratio is not only above the EA average, but it is also above the (largely arbitrarily chosen) 60% or 90% thresholds. CMMP analysis, which is centred on the sector balances framework, considers both fiscal space and debt sustainability as flow concepts and for reasons mentioned above (and possibly in future posts) is less concerned here.
What about private sector debt dynamics?
From a risk and financial stability perspective, we are more concerned about France’s private sector debt dynamics, particularly in the NFC sector.
Trends in the breakdown of French and EA debt (Source: BIS; CMMP)
France has a relatively high exposure to NFC debt. At the end of 2Q21, NFC debt accounted for 46% of total debt. This is down from 50% at the time of the GFC (see chart above) but remains above the aggregate shares of 32% and 39% for advanced and EA economies respectively. Why does this matter?
Top 10 BIS reporting economies ranked by NFC debt ratio (% GDP) (Source: BIS; CMMP)
France’s NFC sector is the fifth most indebted NFC sector among BIS economies. At the end of 2Q21, the NFC debt ratio was 170% GDP, after Luxembourg (322%), Hong Kong (304%), Sweden (179%) and Ireland (171%). The French NFC debt ratio is well above the 111% GDP and 98% GDP aggregate for all advanced and EA economies respectively and the 90% threshold level above which the BIS considers debt to be a drag on future growth.
In spite of the high level of NFC indebtedness, France has seen the third highest levels of excess NFC credit growth over the past three years (see chart above). The NFC sector’s RGF was 4.8% in 2Q21, after Switzerland (6.7%) and Japan (6.6%). The rate of excess NFC credit growth was well above the EA average of 1.8% and higher than the 3.0% average for all advanced economies. Risks are clearly elevated when excess rates of credit growth combine with high levels of indebtedness, as is the case here (for an explanation of the RGF framework see here.)
Trends in NFC debt service ratio (Source: BIS; CMMP)
The NFC debt service ratio (DSR) is also high in absolute terms and above its respective LT average despite the low absolute cost of NFC borrowing. As at the end of 2Q21, the DSR was 60%, 9ppt above its LT average of 51% (see chart above). France is one of four advanced economies where the DSRs are high in both absolute terms and in relation to LT averages, along with Sweden, Canada, and Norway (see chart below).
NFC debt service ratios and deviations from LT averages (Source: BIS; CMMP)
While the level of HH indebtedness in France is low in absolute and relative terms, the risks associated with excess HH credit growth and affordability are elevated in this sector too. At the end of 2Q21, the HH debt ratio was 67% GDP, slightly below the 4Q20 peak level of 68%. The HH debt ratio is higher than the 61% EA average but below the 77% advanced average and the BIS threshold level of 85% GDP. Nonetheless, France has the seen the highest rate of excess HH credit growth over the past three years among EA and other advanced economies. The HH’s debt service ratio, while low in absolute terms, is also above its LT average (see “Global debt dynamics – IV” for more details and charts).
Conclusion
The COVID-19 pandemic had a significant impact on the French government’s net borrowing and the level of government debt. The widening gap between France’s government debt ratio and those of the so-called “fiscally-conservative” economies is also a potential source of conflict with the EA. Viewed from a sector balances perspective, however, the government’s response was timely, necessary and appropriate. We are also more concerned about the risks associated with private sector debt dynamics, particularly in the highly indebted NFC sector.
More fundamentally, the headlines resulting from the Cour des Comptes’ report support our wider hypothesis that conventional macro thinking is flawed to the extent that it typically ignores private debt while seeing government debt as a problem rather than as a solution.
Please note that the summary comments and charts above are extracts from more detailed analysis that is available separately.
