Author: cgmallin

“Financing flows to the EA economy”

The impact of ECB policy seen through 12-month cumulative flows

The key chart

12-month cumulative financing flows (EUR bn) presented in stylised consolidated balance sheet format (Source: ECB; CMMP)

The key message

Monetary developments in the euro area (EA) highlight the elevated risks of ECB policy errors and their potential, negative impact on financing flows to the EA economy.

Headline YoY growth numbers for broad money and its key component (narrow money) and counterpart (private sector credit) highlight the speed at which EA money and credit cycles are rolling over.

Broad money (M3) fell -0.4% YoY, the first annual decline since May 2010. Narrow money (M1) fell -9.2% YoY, driven by a 10.5% YoY decline in overnight deposits. Growth in private sector credit slowed to 1.6% YoY, the slowest annual rate of growth since May 2016. The warning signs are there…

Within broad money, arbitrage continues in favour of the highest remunerated deposits – depositors are actively seeking higher returns – but this is insufficient to compensate for outflows from overnight deposits. The very slow/limited pass through from higher policy rates to the cost of overnight deposits has been one of the unique features of the current hiking cycle.

Financing flows to the private sector, largely in the form of loans, remain positive in absolute terms. They are slowing very sharply on a cumulative 12-month basis, however (see next post for details).

In short, cumulative financing flows to the EA economy were -€96bn in the 12-months to July 2023, compared with flows of €1,574bn, €1,014bn and €92bn in the 12-months to July 2020, July 2021 and July 2022 respectively.

The risks of significant policy errors are rising with negative implications for financing flows to the EA economy. How will the “data-dependent” ECB respond in September?

The impact of ECB policy on financing flows to the EA economy

Headline YoY growth numbers for broad money and its key component (narrow money) and counterpart (private sector credit) highlight the speed at which EA money and credit cycles are rolling over (see chart below).

Growth rates (% YoY) in M3, M1 and PS credit (Source: ECB; CMMP)

Broad money (M3) fell -0.4% YoY, the first annual decline since May 2010. The outstanding stock of money (€15,957bn) has fallen -1.6% from its September 2022 peak (€16,214bn). Narrow money, the key component of broad money, fell -9.2% YoY, driven by a 10.5% YoY decline in overnight deposits. Growth in private sector credit, the key counterpart to broad money, slowed to 1.6% YoY, its slowest annual rate of growth since May 2016.

Recall that monetary aggregates are derived from the consolidated balance sheet of MFIs. The key components are found on the liabilities side of the balance sheet – narrow money (M1) which comprises currency in circulation, other short-term deposits (M2-M1) and marketable instruments (M3-M2). Note that longer-term liabilities are not part of M3 as they are regarded more as portfolio instrument than as a means of carrying out transactions. The key chart above presents 12-month cumulative flows in the form of a stylised consolidated balance sheet.

Growth rates (% YoY) in M1 and M2-M1 (Source: ECB; CMMP)

Within broad money, arbitrage continues in favour of the highest remunerated deposits but this is insufficient to compensate for outflows from overnight deposits. The annual growth rate in other short-term deposits (M2-M1) was 24% YoY in June and July 2023, the highest rate of growth since the start of the EMU. In contrast, the -9.2% YoY decline in narrow money was the sharpest contraction since the start of EMU (see chart above).

Monthly flows (EUR bn) in EA monetary aggregates (Source: ECB; CMMP)

The EA banking system has seen eleven consecutive months of outflows in overnight deposits (see chart above). The outflow of narrow money totalled -€1,072bn (see key chart above), overshadowing the positive inflows of €852bn and €153bn into other short-term deposits (M2-M1) and marketable securities (M3-M2) and, outside broad money, the €262bn inflow into longer-term financial liabilities (mainly debt securities issued by banks).

Note that the very slow/limited pass through from higher policy rates to the cost of overnight deposits has been one of the unique features of the current hiking cycle.

Trend in 12-month cumulative monthly flows of loans to the private sector (Source: ECB; CMMP)

Financing flows to the private sector, largely in the form of loans, remain positive in absolute terms. They are slowing sharply on a cumulative 12-month basis, however (see chart above and next post for details).

12-month cumulative financing flows (EUR bn) presented in stylised consolidated balance sheet format (Source: ECB; CMMP)

In short, cumulative financing flows were -€96bn in the 12-months to July 2023, compared with flows of €1,576bn, €1,014bn and €92bn in the 12-months to July 2020, July 2021 and July 2022 respectively.

The risks of significant policy errors are rising. How will a data-dependent ECB respond in September?

Please note that the summary charts and comments above are abstracts from more detailed analysis that is available separately.

“When will US and UK consumers read the recession script?”

Consumer credit flows remain resilient YTD

The key chart

Quarterly consumer credit flows expressed as a multiple of pre-pandemic average flows (Source: FRED; BoE; CMMP)

The key message

In May 2023, I asked, “have US and UK consumers read the recession script?” I noted that, “despite rising borrowing costs, quarterly consumer credit flows in 1Q23 remained above pre-pandemic levels. Fast-forward another quarter and the question remains largely the same – when will US and UK consumers read the recession script?

Contrasting consumer credit dynamics suggests divergent growth outlooks for investors and different challenges for the Federal Reserve, the Bank of England and the ECB.

US consumer credit demand has moderated from its elevated 2022 levels, but remains above pre-pandemic average levels. UK consumer credit demand remains surprisingly resilient, and above pre-pandemic levels too. Chair Powell and Governor Bailey face similar challenges as consumers continue to borrow to support consumption despite rising borrowing costs.

The contrast with consumer credit dynamics in the euro area (EA) is sharp. EA consumer credit demand remains subdued and well below pre-pandemic levels. President Lagarde faces a very different balancing act between weaker growth/recession and lower inflation.

In short, the messages from the US, UK and EA money sectors remind us that the risks of policy errors remain elevated in all three regions but for contrasting reasons…

When will US and UK consumers read the recession script?

Quarterly US consumer credit flows (Source: FRED; CMMP)

US consumer credit demand is moderating sharply but remains above pre-pandemic levels. Quarterly flows of consumer credit have fallen from $87bn in 4Q22 to $52bn in 1Q23 and $50bn in 2Q23, but remain above their pre-pandemic level of €45bn (see chart above).

Monthly US consumer credit flows (Source: FRED; CMMP)

The monthly flow of US consumer credit was more volatile during the 2Q23 (see chart above). Flows ranged from $22.3bn (1.5x pre-pandemic average) in April 2023 to only $9.5bn (0.6x pre-pandemic average) in May 2023 and then $17.8bn (1.2x pre-pandemic flow) in June 2023.

Quarterly UK consumer credit flows (Source: BoE; CMMP)

UK consumer credit demand remains surprisingly resilient. Quarterly flows increased from £3.1bn in 4Q22 to £4.5bn in 1Q23 and £4.3bn in 2Q23. These flows were 0.9x, 1.2x and 1.2x the pre-pandemic average quarterly flow of £3.6bn (see chart above).

Monthly UK consumer credit flows (Source: BoE; CMMP)

Monthly flows during the 2Q23 were £1.6bn, £1.1bn and £1.7bn in April, May and June 2023 respectively. These flows were 1.6x, 1.1x and 1.7x the pre-pandemic average flows respectively. (see chart above).

Quarterly EA consumer credit flows (Source: ECB; CMMP)

Euro area consumer credit demand remains consistently subdued, in sharp contrast to trends observed in both the US and the UK. Quarterly flows have fallen from €5.2bn in 4Q22 to €4.2bn in 1Q23 and €3.4bn in 2Q23, the lowest quarterly flow since 2Q21 (see chart above). Quarterly flows have failed to recover to their pre-pandemic level of €10.3bn.

Monthly EA consumer credit flows (Source: ECB; CMMP)

Monthly flows have also declined during the quarter from €2.0bn in April to €1.3bn in May to only €23m in June 2023 (see chart above). As noted in previous posts, the fact that demand for consumer credit remains very subdued in absolute terms and in relation to trends observed in the US and the UK makes the ECB’s tasks slightly easier (while elevating policy risks at the same time).

Conclusion

Contrasting consumer credit dynamics suggests divergent growth outlooks for investors and different challenges for the Federal Reserve, the Bank of England and the ECB.

Demand is moderating sharply in the US but remains resilient and above the levels seen pre-pandemic, at least so far. UK demand also remains surprisingly resilient and above pre-pandemic levels, at least for the time being (and as suggested by OBR forecasts). Chair Powell and Governor Bailey face similar challenges as consumers continue to borrow to support consumptions despite rising borrowing costs.

In contrast, the message from the money sector for EA growth is much weaker and presents President Lagarde with a very different balancing act between weaker growth/recession and lower inflation.

The risks of policy errors remain elevated in all three regions but for contrasting reasons…

Please note that the summary comments and charts above are abstracts from more detailed analysis that is available separately.

“Why is Richard Koo’s profile rising in China?”

Because the Chinese government knows there is a disease called “balance sheet recession”.

The key chart

Trends in “excess credit growth” in China since December 2012 (Source: BIS; CMMP)

The key message

“The Chinese government knows that there is a disease called balance sheet recession, and they should know how to handle it”

Richard Koo, quoted by Bloomberg on 10 July 2023

Richard Koo is the Chief Economist at the Nomura Research Institute who developed the concept of “balance sheet recessions”. These recessions occur when “a nationwide asset bubble financed by debt bursts” (Koo, 2008). According to a Bloomberg article this week, “Koo’s ideas are being taken seriously in China”.

In this post, I explain why this is the case. I revisit my January 2023 arguments and update the data points and charts from “The missing link in the China re-opening story?”. I begin with the same question that I posed at the start of the year:

What happens if rather than seeking to maximise profit/utility as traditional economics assumes, the Chinese private sector (PS) turns to minimising debt or maximising savings instead? What if China experiences a balance sheet recession?

What does the data tell us?

While China’s PS indebtedness has declined from its 3Q20 peak, it remains above Japan’s “peak-bubble” level. Affordability risks remain elevated in China too. The PS debt ratio is not only high in absolute terms, but it is also elevated in relation to its long-term average. Chinese debt dynamics have shifted from excess growth in corporate credit growth, through excess growth in household debt, to a “passive deleveraging” phase. The risk remains that this extends to full PS debt minimisation i.e., a balance sheet recession.

Following recent re-intermediation, the Chinese banking sector is relatively exposed to the risks associated with these dynamics. The bank sector debt ratio exceeds the level reached at the height of the Spanish private sector debt bubble, for example, and is currently the highest ratio among BIS reporting economies.

What are the implications?

China’s debt dynamics point to potential demand (debt minimisation) and supply side (bank sector debt) constraints to future consumption.

From a policy perspective, Koo argues that the Chinese government must ramp up spending to offset private sector deleveraging. According to Bloomberg, he recommends a fresh wave of fiscal stimulus, targeted towards the real estate sector.

From an investment perspective, these factors need to be included in the investment narrative. It was a mistake to ignore private sector debt dynamics at the start of the year. It is a mistake to ignore them now…

Why is Richard Koo’s profile rising in China?

Personal background

I met Richard Koo in Tokyo when I was a graduate trainee at Nomura Securities in 1986. I left Nomura in 1988 to build the top-rated Japanese equities team (for European clients) at Baring Securities. I subsequently moved on from Japanese equities in the early 1990s to develop a career in banks research and macro strategy. I continued to follow Koo’s work, read his books and applied his approaches to my own analysis of monetary and macro developments in other advanced and emerging economies. We share a common analytical foundation based on a sector-balances framework.

My January 2023 message

In January 2023, I questioned the so-called “China re-opening” narrative. My scepticism centred on the level of Chinese private sector debt, the growth in HH debt and the affordability of private sector debt.

I asked, “What happens, for example, if rather than seeking to maximise profit/utility as traditional economics assumes, the Chinese private sector turns to minimising debt or maximising savings instead? What if China experiences a balance sheet recession?”

The July 2023 update

Trends in Japanese, Spanish and Chinese PS debt ratios since 1980 (Source: BIS; CMMP)

China’s private sector indebtedness exceeds the “peak-bubble” level seen in Japan in 4Q94. The private sector debt ratio was 220% GDP at the end of 4Q22, below its recent 225% peak but still above Japan’s peak debt ratio of 214% GDP. Note the similarity in debt ratio trends in Japan (debt bubble), Spain (debt bubble) and in China (see chart above). History rhymes…

Trend in the Chinese PS debt service ratio since 2000 (Source: BIS; CMMP)

Private sector affordability risks remain elevated too. The private sector debt service ratio (PS DSR) is not only elevated in absolute terms (20.6%) but it is also elevated in relation to its long-term average (15.7%). The PS DSR peaked in 3Q20 (see chart above) and has trended between 20-21% since then.

Trends in stock of HH and NFC debt (RMB tr) and PS debt ratio (% GDP, RHS) (Source: BIS; CMMP)

The Chinese private sector has already entered a period of passive deleveraging (see chart above). In other words, while to outstanding stock of debt continues to increase, it has been growing at a slower pace than nominal GDP since 3Q20.

Chinese debt dynamics have shifted from excess growth in corporate debt to excess credit growth in household debt and then to passive deleveraging. The risks remain that these trends extend to an extended period of private sector debt minimisation – i.e., a balance sheet recession.

Bank credit as a percentage of total PS credit (%) since 2007 (Source: BIS; CMMP)

With recent re-intermediation (see chart above), the Chinese banking sector is relatively exposed to the risks associated with current debt dynamics. The bank sector debt ratio (see chart below) exceeds the level reached at the height of the Spanish private sector debt bubble, for example, and is currently the highest ratio among BIS reporting economies (excluding Hong Kong).

Trends in Japanese, Spanish and Chinese bank sector debt ratios (% GDP) (Source: BIS; CMMP)

Conclusion

China’s debt dynamics point to potential demand (debt minimisation) and supply side (bank sector debt) constraints to future consumption.

From a policy perspective, Koo argues that the Chinese government must ramp up spending to offset private sector deleveraging. According to Bloomberg, he recommends a fresh wave of fiscal stimulus, targeted towards the real estate sector.

From an investment perspective, these factors need to be included in the investment narrative. It was a mistake to ignore private sector debt dynamics at the start of the year. It is a mistake to ignore them now…

Please note that the summary comments and charts above are extracts from more detailed analysis that is available separately.

“How high would the UK base rate have to go…”

…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.

“Is there such a thing as the EA mortgage market?”

Yes, but it’s complicated…

The key chart

Trends in total EA mortgages (EUR bn, LHS) and annual growth (% YoY, RHS) (Source: ECB; CMMP)

The key message

The answer to the question, “Is there such a thing as the EA mortgage market?” seems obvious. Of course there is.

We know its size (€5,228bn), its structure (biased towards fixed rate lending) and its importance to banks (40% of total lending). We also know the current cost of borrowing (3.44%) and the speed with which higher policy rates have passed through to this cost (147bp so far). We can monitor the rate of growth in mortgages (3.0% YoY in nominal terms, -3.7% in real terms) and in monthly flows (slowing sharply).

Not so fast…!

The complication here is that these aggregate data points mask very important variations at the national level. These include:

Size (€5,288bn): five national markets dominate (“the big five”). Germany and France account for 56% of total EA mortgages alone and for 85% together with the Netherlands, Spain and Italy.

Structure of new mortgages (over 75% fixed-rate): varies from over 90% variable-rate in Finland, Lithuania, Estonia and Latvia to over 90% fixed-rate in Slovenia, Slovakia, France, Belgium and Ireland. Among the big five, relatively high exposures to fixed rate lending in France, Germany and the Netherlands, relatively low exposures in Italy and Spain.

Exposure to mortgage lending (40% of total lending): ranges from 50% in Malta and Slovakia to 25% or less in Luxembourg and Greece. Among the big five, above average exposures in the Netherlands, Germany and Spain, below average exposures in France, and more noticeably in Italy.

Cost of borrowing (3.44%, April 2023): ranges from 5.32%, 5.27% and 4.99% in Latvia, Lithuania and Estonia respectively to 2.24% and 2.61% in Malta and France respectively. Among the big five above average costs in all markets with the exception of France. Note that French banks have (1) relatively high exposure to fixed rate mortgages and, (2) unlike most other EA lenders, are constrained by the Banque de France on the amount they can charge borrowers.

Transmission mechanism of higher policy rates (147bp, so far): most rapid in Lithuania, Latvia, Estonia, Portugal, and in Italy and Spain among big five – all markets with above average exposure to variable-rate mortgages. Weakest in Malta, Ireland, Greece, and among the big five in France, Germany and the Netherlands. With the exception of Malta, these markets all have relatively high exposures to fixed-rate lending.

Growth (slowed to 3.0% YoY in April 2023, the slowest rate since May 2018): Skewed heavily towards German and French growth dynamics (1.1ppt and 1.0ppt of total 3.0% respectively). Large variations in nominal growth rates from 9.8% and 9.5% in Lithuania and Estonia respectively to -4.0% in Greece and -1.9% in Spain and Ireland. Among the big five, above average growth in France, Germany and the Netherlands, but below average growth in Italy and Spain. In real terms, mortgage growth peaked at 5.0% YoY in December 2020, eight months before the peak in nominal growth. It turned negative in February 2022 and has been negative ever since (-3.7% YoY, April 2023). Only Belgium and Malta are experiencing positive mortgage growth in real terms.

What does this mean?

The EA mortgage market is as an aggregation of heterogeneous, national markets that differ greatly in terms of size, structure, importance, cost, transmission mechanism and growth rates.

The challenge for bankers, investors and analysts alike is to understand these differences and their implications. The far greater challenge for the ECB is to incorporate them all in the design of a “one-size-fits-all” monetary policy. The context, in part, for this week’s ECB press conference on Thursday 15 June 2023.

Is there such a thing as the EA mortgage market?

Market size

Trends in the outstanding stock of EA mortgages (EUR bn) (Source: ECB; CMMP)

The outstanding stock of mortgages across the EA was €5,228bn at the end of April 2023 (see chart above).

Five national markets (the “big five”) dominate in terms of size and account for 85% of the outstanding stock collectively (see chart below) – Germany (€1,574bn, 30% share), France (€1,333bn, 26% share), the Netherlands (€555bn, 11% share), Spain (€505bn, 10% share) and Italy (€426bn, 8% share).

National mortgage markets ranked by size (EUR bn, LHS) and cumulative market share (%, RHS) (Source: ECB; CMMP)

Mortgage types

Twenty year trends in share of variable rate loans in total new mortages (%) (Source: ECB; CMMP)

At the aggregate level, just over three quarters of new mortgages are fixed-rate mortgages, up from 13% in March 2022 (see chart above). The structure varies, however, from over 90% variable rate mortgages in Finland, Lithuania, Estonia and Latvia to less than 10% variable rate mortgages in Slovenia, Slovakia, France, Belgium and Ireland (see chart below).

National mortgage markets ranked by exposure to variable rate lending (% new loans) (Source: ECB; CMMP)

Among the big five markets, the share of variable rate mortgages in new loans ranges from 41% and 39% in Italy and Spain to 20% in the Netherlands, 16% in Germany and only 4% in France. Note also that, in aggregate, the exposure to variable rate mortgages in the EA is currently higher than in the UK (18%).

Exposure to mortgage lending

National mortgage markets ranked by exposure to mortgages (% total loans) (Source: ECB; CMMP)

Mortgages account for 40% of total lending to EA residents at the aggregate level. This exposure ranges from 50% of total loans in Malta and Slovakia to 23% and 25% in Luxembourg and Greece respectively. Among the big five, banks in the Netherlands (48%), Germany (43%) and Spain (41%) have above average exposures to mortgage lending, while banks in France (39%) and, more noticeably, Italy (28%) have lower-than-average exposures.

Composite cost of borrowing for house purchase

National mortgage markets ranked by cost of borrowing for house purchase (%) (Source: ECB; CMMP)

In nominal terms, the CCOB for house purchases ranges from 5.32%, 5.27% and 4.99% in Latvia, Lithuania and Estonia respectively to 2.24% and 2.61% in Malta and France respectively (see chart above). Among the big five markets, the CCOB is above average in Italy (4.15%), Germany (3.89%), the Netherlands (3.62%) and Spain and only below average in France (2.61%).

Note that French banks have (1) relatively high exposure to fixed-rate mortgages and (2), unlike most other EA lenders, are constrained by a limit, set by the Banque de France, on the amount that they can charge for mortgages. In short, they have a lower sensitivity to the positive benefits of rising interest rates.

Note also that the CCOB of borrowing for house purchases remains below the current rates of inflation (HICP) in all of the EA economies except Luxembourg, Cyprus and Belgium.

Pass through of higher policy rates

National mortgage markets ranked by pass through (bp) of higher ECB policy rates (Source: ECB; CMMP)

The composite cost (CCOB) for new loans to EA HHs for house purchase has increased by 147bp since June 2022 to 3.44% in April 2023.

The pass through from policy tightening has been greatest (in nominal terms) in Lithuania (315bp), Latvia (284bp), Estonia (274bp) and Portugal (251bp) and in Italy (197bp) and Spain (179bp) among the big five markets.

The pass though has been weakest in Malta (11bp), Ireland (74bp), Greece (84bp) and in France (126bp), Germany (132bp) and the Netherlands (141bp) among the big five markets.

Change in CCOB since tightening (bp) plotted against current CCOB (% April 2023) (Source: ECB; CMMP)

Growth in mortgage lending

Trend in annual growth rate (% YoY, nominal) of lending to the private sector (Source: ECB; CMMP)

The annual growth rate in EA mortgages slowed to 3.0% YoY in April 2023, down 2.8ppt from the August 2021 peak of 5.8%. Growth has slowed 2.4ppt since tightening began in 2022. April 2023’s growth rate is the slowest rate of growth recorded since May 2018.

Contribution (ppt) of big five and “others” to growth in EA mortgages (% YoY) (Source: ECB, CMMP)

Germany and France have been the main contributors to aggregate growth since 2015. In April 2023, Germany and France contributed 1.1ppt and 1.0ppt to the total YoY growth of 3.0% alone. Netherlands and Italy contributed 0.4ppt and 0.2ppt respectively. The most obvious contrast between the post-2015 recovery in EA mortgage demand and the pre-GFC period is the lack of contribution/negative contribution from Spain for large parts of post-GFC period, reflecting the bursting of the Spanish real estate bubble.

National mortgage markets ranked by nominal growth rate (% YoY) (Source: ECB; CMMP)

Large variations exists in the nominal YoY growth rates at the country level. In April 2023, these ranged from 9.8% and 9.5% in Lithuania and Estonia respectively to -4.0% in Greece and -1.9% in both Spain and Ireland. Among the big five markets, growth was above average in France (4.1%), Germany (3.8%), and the Netherlands (3.6%) but below average in Italy (2.7%) and Spain (-1.9%).

Trends in mortgage growth rates expressed in nominal and real terms (Source: ECB; CMMP)

Inflation also complicates the analysis of EA mortgage dynamic. In real terms, mortgage growth peaked at 5.0% YoY in December 2020, eight months before the peak in nominal growth. It turned negative in February 2022 and has been negative since then. In April 2023, mortgage lending fell -3.7% YoY in real terms.

National mortgage markets ranked by real growth rate (% YoY) (Source: ECB; CMMP)

Conclusion

The EA mortgage market is as an aggregation of heterogeneous, national markets that differ greatly in terms of size, structure, importance, cost, transmission mechanism and growth rates. The challenge for bankers, investors and analysts alike is to understand these differences and their implications. The far greater challenge for the ECB is to incorporate them all in the design of a “one-size-fits-all” monetary policy. The context, in part, for this week’s ECB press conference on Thursday 15 June 2023.

Please note that the summary comments and charts above are abstracts from more detailed analysis that is available separately.

“Pschitt II…”

The sound of deflating UK and EA mortgage markets is getting even louder

The key chart

Monthly UK and EA mortgage flows (3m MVA) expressed as a multiple of pre-pandemic average flows (Source: BoE; CMMP)

The key message

The sound of deflating UK and EA mortgage markets increased further at the start of 2Q23, including the return of net mortgage repayments in the UK. More bad news and negative headlines followed in UK newspapers at the start of this week…

“Banks turn the screw with a weekend of worse mortgage rates”

The Times, 5 June 2023

UK monthly mortgage flow dynamics (LHS) and annual growth rate (RHS) (Source: BoE; CMMP)

According to the latest BoE statistics, borrowing of mortgage debt by UK individuals declined from net zero in March 2023 to net repayments of £1.4bn in April 2023. This represents the lowest level since the £1.8bn net repayments recorded in July 2021.

The BoE noted that, “If the period since the onset of the COVID-19 pandemic is excluded, net borrowing of mortgage debt was at its lowest level on record” (i.e. since April 1993). The 3m MVA of monthly flows fell to -£203m in April from £903m in March. The 3m MVA of monthly flows has been below pre-pandemic levels since November 2022.

The annual growth rate fell to 2.3% YoY, the slowest rate since September 2015. With approvals also falling from 51,500 in March to 48,700 in April, further weakness in monthly flows and annual growth rates are likely.

EA monthly mortgage flow dynamics (LHS) and annual growth rate (RHS) (Source: ECB; CMMP)

Monthly EA mortgage flows also slowed sharply in April 2023 according to the latest ECB statistics. The flow fell to €1.7bn in April from €7.5bn in March and €5.1bn in February. The 3m MVA average of mortgage flows fell to €4.7bn in April from €4.9bn in March, only 0.38x the pre-pandemic average flow. The annual growth rate fell to 3.0% YoY, the slowest rate since April 2018.

Weaknesses in mortgage flows matter for two important reasons:

First, mortgage demand typically displays a co-incident relationship with real GDP.

Second, and in this context, the message from the UK and EA money sectors is clear – central banks continue to tighten policy as the risks to the economic outlook intensify.

Please note that the summary comments and charts above are abstracts from more detailed analysis that is available separately.

“If you wanted to create panic about EA banks….”

…you might focus exclusively on narrow money.

The key chart

Recent trends in monthly flows (EUR bn) and YoY growth rates in EA monetary aggregates (Source: ECB; CMMP)

The key message

If you wanted to “create panic” about euro area (EA) banks you could focus exclusively on negative YoY growth rates and monthly outflows in narrow money (M1), and then develop a narrative about money destruction and an impending credit crisis. After all, the EA is experiencing the fastest contraction in M1 since the creation of the Economic and Monetary Union in 1999, and banks have experienced eight consecutive months of negative flows in narrow money too.

Unfortunately, there are a number of problems associated with this “panic narrative”. It ignores simple concepts such as opportunity cost and portfolio rebalancing and mispresents the causal links in money creation.

  • The opportunity cost problem: unprecedented tightening by the ECB has led to a rapid increase in the opportunity cost of holding ON deposits, in contrast to most of the past decade. It has triggered a partial reallocation of ON deposits to other ST deposits. M2-M1 (other ST deposits) increased 21% YoY in April 2023 – also the fastest rate since the creation of the EMU.
  • The portfolio rebalancing problem: The phasing out of net asset purchases and TLTROs has incentivised the issuance of bank bonds (up €170bn) since September 2022. This has led to portfolio rebalancing away from deposits (down €200bn over the same period) to longer term liabilities that do not form part of monetary aggregates (by definition).
  • The causal link problem: the principal way in which bank deposits are created is through commercial banks making loans. Banks are not simply intermediaries that take in deposits and then lend them out (“loanable funds theory”). Instead, banks create money. Growth in private sector credit peaked in September 2022 (7.0% YoY) and has slowed to 3.3% in April 2023, however. This reflects the relatively rapid pass through from higher policy rates to the cost of borrowing, weaker loan demand and tighter credit standards. Slower credit growth implies slower deposit growth (ceteris paribus).

In short, recent EA monetary dynamics are unprecedented in some respects. They are not a cause for panic over EA banks, however. They reflect instead a combination of an increase in the opportunity cost of holding money, portfolio rebalancing, weaker credit demand and tighter credit standards.

This is not to suggest that the message from EA banks is a positive one. On the contrary, it remains one of weaker economic activity and a challenging policy context in which the ECB is expected to continue tightening as economic stresses mount.

An important message, but a very different one to the “bank panic narratives” seen elsewhere…

If you wanted to create panic about EA banks

Focus on narrow money

Growth rates (% YoY) in EA narrow money since 1999 (Source: ECB; CMMP)

Narrow money (M1) in the EA is contracting at the fastest rate since the creation of the Economic and Monetary Union (EMU) in 1999 (see chart above). The annual growth rate in M1 turned negative in January 2023 (-0.8%) and has decelerated each month since then: -2.7% YoY in February 2023; -4.2% YoY in March 2023 and -5.2% YoY in April 2023.

Monthly flows (EUR bn) in narrow money since June 2022 (Source: ECB; CMMP)

EA banks have also experienced eight consecutive monthly outflows of narrow money since September 2022, largely due to the outflow of overnight deposits. The outflow in April 2023 was €-75bn, compared with €-135bn in March 2023, and €-140bn in February 2023 (see chart above).

From here, it is tempting to create a “panic narrative” for EA banks. Tempting, but wrong…

Problems with the panic narrative – “opportunity cost”

Growth rate in M3 (% YoY) and contribution from ON deposits and other components (ppt) (Source: ECB; CMMP)

The first problem with the panic narrative is that it ignores the fundamental economic concept of opportunity cost.

Note that for most of the past decade, narrow money has been the main driver of EA broad money growth. The chart above illustrates the growth rate in broad money (M3) and the contributions to growth made by ON deposits (the light blue columns) and all other M3 components (the maroon columns).

The exceptionally large contribution and accumulation of ON deposits in the EA (and elsewhere) over the period reflects (1) the extended period of low interest rates and, more recently, (2) the COVID-19 pandemic. Unorthodox monetary policy reduced the opportunity cost of holding ON deposits dramatically and the pandemic resulted in a sharp rise in both forced and precautionary savings.

Policy tightening by the ECB since June 2022 has been notable for both its scale and pace. The pass through from higher policy rates to the cost of overnight deposits has been very slow/limited and has lagged the pass through to the cost of other forms of ST deposits, however.

In short, the opportunity cost of holding ON deposits has risen rapidly since the start of policy tightening.

Spread between rates on other HH ST deposits and HH ON deposits (ppt) (Source: ECB; CMMP)

The chart above illustrates the spread (or opportunity cost) between HH deposits redeemable at notice of up to three months and deposits with an agreed maturity of up to two years – the components of M2-M1 – versus HH ON deposits over the past 20 years.

The opportunity costs of holding notice deposits, and terms deposits with maturities of up to one year and between one and two years hit lows of 32bp, 13bp and 18bp in December 2021, June 2021 and March 2021 respectively. Since then they have risen to 106bp, 194bp and 194bp respectively, levels not seen since early 2013.

Spread between rates on other NFC ST deposits and NFC ON deposits (ppt) (Source: ECB; CMMP)

The chart above illustrates the spread (opportunity cost) between NFC time deposits – again M2-M1- and NFC ON deposits over the past twenty years. In this case the opportunity costs of holding term deposits hit lows of -33bp in November 2021 for maturities up to one year and 4bp for maturities between one and two years. Since, then they have risen to 215bp and 243bp respectively, levels last seen during the GFC.

Growth rates (% YoY) in M1 and M2-M1 since 1999 (Source: ECB; CMMP)

ECB policy has triggered a reallocation of funds from overnight deposits to other ST deposits. The growth rate in M2-M1 rose to 21% YoY in April. This is, in turn, the fastest rate of growth since the creation of the EMU.

Monthly flows (EUR bn) on EA monetary aggregates (Source: ECB; CMMP)

Inflows into other ST liabilities, M2-M1 and, to a lesser extent, M3-M2 have been important but still insufficient to compensate fully for the outflows in overnight deposits (see chart above). Hence, monthly flows of M3 have been negative in six of the past seven months.

Problems with the panic narrative – “portfolio rebalancing”

The second problem with the panic narrative is that it ignores portfolio rebalancing to other financial instruments. Note that money supply is derived from the banks’ ST liabilities and does not include longer-term liabilities since they are not close substitutes for money.

According to the ECB, bank bond issuance has increased by almost €170bn since September 2022. The terms and conditions of TLRTO II were recalibrated at this point resulting in sizeable repayments of funds borrowed under the programme and an increase in (more expensive) bond issuance. Bond issuance was close to total c€200bn decrease in the bank deposits over the period.

In short, M1 dynamics reflect both a substitution of ON deposits with time deposits (opportunity cost) and shifts to bank bonds (portfolio rebalancing) and, to a lesser extent money market fund shares. Funding costs may be rising but there is no evidence of liquidity-driven panic.

Problems with the panic narrative – “causal links in money creation”

The final problem with the panic narrative is that it misrepresents the causal links in money creation. Contrary to what is often taught (“loanable funds theory”), banks are not intermediaries that take in deposits first and then lend them out. Instead banks create money.

“Whenever a bank makes a loan, it simultaneously creates a matching deposit in the borrower’s bank account, thereby creating new money”

(Bank of England, 2014).

Growth rates in (adjusted) private sector credit (% YoY) since 2003 (Source: ECB; CMMP)

The principal way in which bank deposits are created is through commercial banks making loans. Growth in private sector lending peaked in September 2022 at 7% YoY, however (see graph above). In April 2023, growth has slowed to 3.3% YoY from 3.9% YoY in March 2023 and 4.3% YoY in February 2023. The moderation in bank lending reflects the relatively rapid pass through from policy rates to the cost of borrowing, weaker demand and tighter credit standards.

Growth rates in M3, M1 and private sector credit (% YoY) (Source: ECB; CMMP)

Conclusion – focus on the wider message not the panic narrative

In short, recent EA monetary dynamics are unprecedented in some areas. They are not a cause for panic over EA banks, however. They reflect instead a combination of an increase in the opportunity cost of holding money, portfolio rebalancing, weaker credit demand and tighter credit standards.

This is not to suggest that the message from EA banks is a positive one. On the contrary, it remains one of weaker economic activity and a challenging policy context in which the ECB is expected to continue tightening as economic stresses mount. An important message, but a very different one to the “bank panic narratives” seen elsewhere…

Please note that the summary comments and charts above are abstracts from more detailed analysis that is available separately.

“If you wanted to create panic over US public debt – II”

Be careful how you present your data and what you include.

The key chart

Long term trends in US GDP, public debt and private debt plotted on log scales (Source: FED; CMMP)

The key message

Further thoughts for those “seeking to create panic” over US public debt – (1) do not use log scales when presenting data graphically, and (2) do not include private sector debt in your analysis.

In my previous post, I illustrated how using linear scales (absolute levels) was better for creating panic than using log scales (rate of change). Recall that long time-series data that has low initial values and a constant growth rate (eg US public debt) lends itself very well to creating so-called “panic charts” (see chart below).

“Panic charts” in action (Source: CMMP)

The reason not to use log scales is simple. They emphasise instead, how the rate of growth has (1) been relatively constant over time, and (2) slower now than in early parts of the period.

Excluding private debt is a second key tactic. Since current macro thinking typically ignores private debt altogether while seeing public debt as a problem, this is relatively straightforward and uncontroversial.

What are the risks of ignoring this advice?

First, readers may compare its rate of growth (the green line in the key chart above) with the rates of growth in public debt (blue line) and nominal GDP (maroon line). They may note the extended period (1948-2007) when the rate of growth in private debt far outstripped the rate of growth in both GDP and public debt (the private debt ratio rose from 48% GDP to 165% of GDP).

Second, and from this, readers may explore the relationship between public and private debt growth rates (buying Amazon’s last copy of Wynne Godley’s, “Monetary Economics” in the process). They may note that during phase 1 (4Q1940 – 4Q1981) the public debt ratio fell from 86% GDP to 32% GDP. In other words, the net wealth of the non-government sector parked in a savings account marked “treasuries” grew at a much slower rate than GDP. They may realise that in response, the private sector borrowed more. A lot more. The private debt ratio increased from 48% GDP to 99% GDP.

From here, attention may shift naturally to the two periods of relatively high rates of growth in public debt – Phase 2 (3Q81 – Reaganomics – 3Q93) and the post–GFC Phase 4 (2Q07 – 1Q13). In phase 2, the public debt ratio rose from 32% GDP to 70% GDP. Private debt continued to grow but at a much slower place (to 118% GDP). In phase 4, the expansion of public debt reflected the government’s response to the GFC and the subsequent deleveraging of the private sector (the private debt ratio fell from 165% GDP to 149% GDP).

There are two risks here. First, attention may be drawn to the minority-held view (so far) that private debt typically causes crises while public debt typically limits their impact. Second, it may also be drawn to the fact that the US led the advanced world in the structural shift away from relatively high risk private debt to lower risk public debt in the post-GFC period.

To repeat and extent the message from the previous post – it is always worth considering how different parties chose to present their data to support their arguments and also to consider what they chose to include, or more importantly, exclude….

“If you wanted to create panic over US public debt – I”

How would you present your data?

The key chart

50 years of US public sector debt dynamics – plotted differently! (Source: FED;CMMP)

The key message

If you wanted to create panic over US public sector debt there are a number of tactics that you could use:

  • ignore the fact that US debt is a liability for one economic sector and an asset for another;
  • make no comparison with GDP;
  • ignore inflation etc.

Alternatively, you could simply present your data using linear scales – an easier and almost guaranteed method to instil panic!

1 unit of XZY growing at 10% p.a. over 100 years plotted on a linear scale (Source: CMMP)

Why is this a smart tactic for creating panic?

In the case of any long time series data that has low initial values and a constant growth, the use of a linear scale will always create a chart that is flat for a long period at the start, then becomes steeper and steeper before becoming virtually vertical (see chart above). Readers’ eyes will naturally focus on the steepening of the graph, thereby leading to the conclusion that some form of critical limit has been reached

BINGO – “PANIC ALERT!”

In contrast, the wrong tactic for creating panic would be to present data using a log scale (the maroon line in the key chart). Rather than illustrating the absolute level of US debt (the blue area) this would show its rate of change.

As can be seen, this is (1) fairly constant over time (US debt roughly doubles every ten years in nominal terms) and (2) is slower now than in early parts of the period shown. Presented this way (see chart below), the same underlying data is much less likely to create panic…

1 unit of XZY growing at 10% p.a. over 100 years plotted on a log scale (Source: CMMP)

For the rest of us, it is worth considering how different parties chose to present the same data to support their respective arguments, especially as concerns (false or otherwise) over the US debt ceiling intensify.

Remember, a chart of US public debt using a linear scale is highly likely to create panic, even if the rate of change of debt is constant.

“Why can’t the US be more like Denmark?”

A tempting question to ask, but also the wrong one!

The key chart

Twenty-year trends in debt ratios (% GDP) broken down by type (Source: BIS; CMMP)

The key message

Denmark and the US are the only two, advanced economies that maintain absolute limits on the level of government debt. Interestingly, their public debt ratios (as opposed to debt levels) were essentially the same twenty years ago (54-55% GDP). Today, however, Denmark has the third lowest public debt ratio in the world (30% GDP), while the US has the seventh highest (103% GDP).

Unsurprisingly perhaps, while we hear little “political noise” about the Danish debt ceiling (or debt ratio), we are subjected to an overdose from the US. This leads some to ask, “Why can’t the US be more like Denmark”.

A tempting question to ask, but also the wrong one…

To understand why, we need to adopt a systemic view of the Danish and US financial systems that incorporates both public and private sector debt and reflects the reality of modern money creation.

Put simply, money creation relies on actions that add to bank deposits. Bank lending and government DEFICITS (despite what is taught in many textbooks) qualify, but in different ways. The repayment of bank loans and government SURPLUSES have the opposite effect – they destroy money.

In this context, a better question to ask is, “How do the processes of money creation in Denmark and the US differ and why does this matter?”

Total debt ratios have increased in both economies over the past two decades (see key chart above). In Denmark from 223% GDP to 246% GDP. In the US from 198% GDP to 256% GDP. The processes have been very different, however.

Denmark has effectively substituted government deficits and public debt (an asset of the non-government sector) with more private debt (a liability of the non-government sector) over the past two decades. In contrast, the US has substituted private debt – mainly household debt – with more public debt. In the process, the ratio of the stock of private debt (largely ignored by policy makers and economists) to public debt has risen from 3.2x to 7.2x in Denmark over the period. In the US, it has fallen from 2.6x to 1.5x.

The key point here it that private sector money creation faces unique supply (capital and regulation) and demand (ability of currency users to service debt) constraints. Despite more than a decade of HH sector deleveraging, for example, Denmark’s HH and NFC debt ratios remain above the levels where debt is considered a constraint on future growth. The result? Further private sector deleveraging, led by the HH sector.

Private sector models are also inherently less flexible, less stable and, in advanced economies, more prone to periods of crisis that require greater to levels of public debt to resolve/address.

While lending to the private sector is also the main source of money creation in the US, the government plays a much greater role than in Denmark. As a currency issuer, the US government does not face the same constraints as the private sector (although it does face other constraints). The US is also one of only three advanced economies where both HH and NFC debt ratios are below their respective threshold limits. In short, the structure of the US model provides more flexibility and more stability, as the response to the COVID-pandemic illustrated, and reduces the risk of crisis.

The US generates significantly more political noise about the management of government debt than Denmark, the only other country to maintain an absolute debt ceiling. This noise differential in not an accurate reflection of the relative strengths of the money creation processes in each economy. It is instead, a reflection of flawed macro thinking. Thinking that views public sector debt as a problem but largely ignores private debt and falsely equates the financial constraints of currency users and currency issuers.

The irony here is that a more accurate, systemic view of Danish and US money creations leads to a different question altogether – “Shouldn’t Denmark be more like the US?”

Why can’t the US be more like Denmark?

Twenty-year trends in public sector debt ratios (% GDP) (Source: BIS; CMMP)

Denmark and the US are the only two, advanced economies that maintain absolute limits on the level of government debt. Their public debt ratios (as opposed to debt levels) were essentially the same twenty years ago (54% and 55% GDP respectively, see graph above). Today, however, Denmark has the third lowest public debt ratio in the world (30% GDP, see graph below), while the US has the seventh highest (103% GDP).

BIS reporting nations ranked by 3Q22 public debt ratios (% GDP) (Source: BIS; CMMP)

Unsurprisingly, while we hear little political noise about the Danish debt ceiling (or debt ratio), we are subjected to an overdose of “political noise” from the US. This leads some to ask, “Why can’t the US be more like Denmark”. A tempting question to ask, but also the wrong one…

Money creation in Denmark and the US – a systemic view

To understand why, we need to adopt a systemic view of the Danish and US financial systems that incorporates both public and private sector debt and reflects the reality of modern money creation.

Money creation relies on actions that add to bank deposits, the main form of money today. Bank lending and government deficits (despite what is taught in many textbooks) qualify, but in different ways. The repayment of bank loans and government surpluses have the opposite effect – they destroy money.

In this context, a better question to ask is, “How do the processes of money creation in Denmark and the US differ and why does this matter?”

The changing nature of money creation – total debt broken down by type (% total debt) (Source: BIS; CMMP)

The graph above illustrates how Denmark has effectively substituted public debt (an asset of the non-government sector) with more private debt (a liability of the non-government sector) over the past two decades. In contrast, the US has substituted private debt – mainly household debt – with more public debt.

Twenty-year trends in private debt/public debt ratios (x) (Source: BIS; CMMP)

Note the relative scale of the stock of private sector debt (largely ignored by policy makers and economist) in relation to the stock of public sector debt over the period (see graph above). In Denmark, the ratio rose from 3.3x to almost 10x before the GFC, fell back and then rose again to 7.2x today. Again, in contrast, the US ratio has fallen from 2.6x to 1.5x over the period.

Twenty-year trends in Danish money creation (debt as % GDP) (Source: BIS; CMMP)

In other words, lending to the private sector is a far more important source of money creation in Denmark (see graph above) than in the US (see below). As an aside, the Danish government ran a budget surplus in 2022, destroying money in the process.

The key point here it that private sector money creation faces unique supply (capital and regulation) and demand (ability of currency users to service debt) constraints. Despite more than a decade of HH sector deleveraging, for example, Denmark’s HH and NFC debt ratios remain above the level where debt is considered a constraint on future growth (see graph below). For reference, the BIS considers that HH and NFC debt ratio of 85% GDP and 90% GDP represent threshold limits above which debt becomes a constraint on future growth.

HH debt ratios plotted against NFC debt ratios for selected BIS reporting nations – red lines represent BIS threshold limits (Source: BIS; CMMP)

The result? Further private sector deleveraging, led by the HH sector (see chart below). Note also that private sector models are also inherently less flexible, less stable and, in advanced economies, more prone to periods of crisis that require greater to levels of public debt to resolve/address.

Twenty-year trends in US and Danish HH and NFC debt ratios (% GDP) (Source: BIS; CMMP)

While lending to the private sector is also the main source of money creation in the US, the government plays a much greater role than in Denmark (see chart below). As a currency issuer, the US government does not face the same constraints as the private sector (although it does face other constraints). The US is also one of only three advanced economies where both HH and NFC debt ratios are below their respective threshold limits.

In short, the structure of the US model provides more flexibility and more stability, as the response to the COVID-pandemic illustrated, and reduces the risk of crisis.

Twenty-year trends in US money creation (debt as % GDP) (Source: BIS; CMMP)

Conclusion – shouldn’t Denmark be more like the US?

The US generates significantly more political noise about the management of government debt than Denmark, the only other country to maintain an absolute debt ceiling. This noise differential in not an accurate reflection of the relative strengths of the money creation processes in each economy. It is instead, a reflection of flawed macro thinking. Thinking that views public sector debt as a problem but largely ignores private debt and falsely equates the financial constraints of currency users and currency issuers.

The irony here is that a more accurate, systemic view of Danish and US money creation leads to a different question altogether – “Shouldn’t Denmark be more like the US?”

Please note that the summary comments and charts above are abstracts from more detailed analysis that is available separately.