Since the Covid-19 shock, the exceptional growth in broad money, and more specifically in bank deposits from households and companies, has elicited numerous comments and analyses, some of which have been erroneous. In particular, some observers have associated it with a ‘wait-and-see’ attitude or with forced savings. This suggests that the creation of precautionary savings in the face of economic uncertainty, or savings “forced” by health protection measures, have contributed to the increase in the volume of bank deposits. Others have suggested that the temporary postponement of certain investment projects has driven the expansion of companies’ cash flow. Now that restrictions are gradually being lifted and consumer spending and investment are recovering, no one is predicting that there will be a destruction of the mass of deposits that have been accumulated. Whilst these analyses are not necessarily incorrect from a microeconomic point of view, they most certainly are when it comes to macroeconomics[1].
In reality, “prevented” consumption and additional savings for households as a result did not in and of themselves increase broad money in the economy. At most, they reduced its circulation and skewed its distribution between institutional sectors or, within sectors, between actors. Between households, the growth in deposits, which has been very strong in aggregate, was unevenly spread due to the loss of income or employment for some. Similarly, the cancellation by companies of certain expenditure or investment projects has limited the transfer of wealth between economic sectors. The varying degrees of the pandemic’s impact and the uneven take-up of cash flow support measures have also skewed the distribution of deposits between sectors. Difficulties in the circulation of money as a result of the health protection measures designed to control the Covid-19 pandemic are not, however, enough to explain, at the aggregate level, the rapid growth in deposits seen over the last year or more. Quite to the contrary, they hit economic growth and held back lending growth, the traditional channel of money creation. A recovery in consumer spending and investment is not in itself synonymous with the destruction of money as it will support economic activity and demand for credit (and thus money creation).
We propose here to dwell for a moment on the case of the USA. The first point of interest in our approach is that the identification of the sources of money creation and destruction, common practice in the eurozone, is not, as far as we are aware, widely used on the other side of the Atlantic.
In the eurozone, the underlying factors in money supply are identified in detail. The European Central Bank (ECB) publishes statistics on the main counterparts of broad money on a monthly basis (ECB, 2021).
Following this data provides information on the main sources of the creation or destruction of money. Analysis of the figures shows that the amplification of the ECB’s programme of purchasing sovereign debt and national government loan guarantee schemes between them explain the exceptional money creation that the eurozone has seen. Granted, the money created over the past year will gradually be destroyed as the central bank reduces its balance sheet, companies repay their government-guaranteed loans and when savers shift into term deposits that do not feed into broad money (see below) or invest in securities issued by the banking system. This partial destruction will result in a deceleration in the growth of money supply. However, continued growth in bank lending will sustain money creation.
In the USA, the concept of money supply is somewhat overlooked at present; most notably it is not included in the minutes of the Federal Reserve’s monetary policy committee. This was not always the case. The Great Depression of 1929 marked the Fed’s first efforts to monitor growth in broad money, encouraged, most notably, by economist Lauchlin Currie, who was one of the first to propose an empirical definition of money supply (Currie, 1935). The academic interest in monetary aggregates increased in the aftermath of the Second World War. In particular, a substantial body of literature was devoted to the velocity of money and the link between demand for money and inflation (Friedman, 1956 and 1960). Work on the definition and measurement of money supply also continued, with financial innovation and changes in the financial system bringing about the appearance of close substitutes for bank sight accounts in the 1970s and 1980s (Andersen, Bordo and Rockoff, 2003). From 1975, the Federal Reserve defined target ranges for growth in monetary aggregates and bank lending, and elevated growth in broad money to the rank of an intermediate goal of monetary policy. In practice, however, these targets were only rarely met and trends in money supply were not particularly decisive (Anderson and Kavajecz, 1994). Ultimately it was not until 1979 and the introduction of the ‘non-borrowed reserves’ management process[2] that the Federal Reserve, under the Chairmanship of Paul Volcker, took greater control over the growth of money supply. From 1982 onwards, however, broad money and lending aggregates no longer played a central role in the definition of US monetary policy. For various reasons (financial innovation, changes in payments technologies and practices, the increasing proportion of US currency held abroad), growth in money supply became harder to predict and to link to other macroeconomic measures (Bernanke, 2006). The Fed continues to supply information on the various components of broad money but the monitoring of its trends and the analysis of its counterparts are not included in publications.
The second point of interest in our approach is that with the eurozone having not so far experienced the process of withdrawing from a Quantitative Easing process, the US experience gives an opportunity to consider its effects. The trends seen during periods of reductions in central bank liquidity, tightening of monetary policy and the reduction in the size of the Fed’s balance sheet between 2015 and 2019 were admittedly in part specific to the details of the Fed’s intervention and the nature of the US financial system. However, they remind us that money does not simply evaporate as non-conventional monetary policy comes to an end. Although the purpose of this article is more to inform than to predict, we will consider the main factors likely to boost or hamper money creation in the USA over the next few quarters.
The European analysis of money creation applied to the American situation
We have transposed the methodology developed in the eurozone to US data in order to identify the sources of money creation and destruction in the USA. Completion of this tricky operation required a number of arbitrary assumptions, as US statistics are not designed with this use in mind. The findings require interpretation in the light of the particular features of the US financial system and the financing of the country’s economy.
Components of broad money
Money supply corresponds to all assets held by economic actors that can be instantaneously used as a means of payment or immediately converted into cash[3]. In the eurozone, the volume of money in circulation in the economy is measured by the M3 monetary aggregate[4]. In the USA, the monetary measure closest to the M3 aggregate in Europe is the M2 measure[5].
The structure of broad money is identical in the two economies (Table 1). Scriptural money or demand deposits (liabilities on bank balance sheets) make up the majority of money supply (around 85% of both M3 and M2), well ahead of fiduciary money (notes and coins on the liabilities side of the central banks’ balance sheets), which accounts for around 10%, and market instruments which only make up a marginal fragment of 5% or so. An element of bank deposits is not included in money supply (20%[6] on average in the two economies): this is the case for term deposits with maturities of more than two years in the eurozone and term deposits of more than USD100,000 in the USA.
The consolidated balance sheet of issuers of money
The European methodology consists of drawing up a consolidated balance sheet for monetary and financial institutions (MFIs), which, in the eurozone, means the Eurosystem, credit institutions and money market funds. This is also known as the “monetary balance sheet“. Under the definition of the M2 aggregate, the three sectors in the USA issuing money are: the Fed, depository institutions and retail money market funds[7]. However, we have excluded the contribution of retail money market funds due to the lack of historical depth of available data and the insufficiently detailed breakdown of their investment portfolios[8]. Our analysis therefore focuses on the consolidated balance sheet of US “MFIs”, excluding retail money market funds, and on the counterparts of the M2 aggregate reduced by the share of such funds (i.e., 94% of M2). Box 1 sets out our approach.
Counterparts of broad money
MFIS CONSOLIDATED BALANCE SHEETPreparing a consolidated balance sheet for MFIs allows the identification of the main counterparts of money supply: credit to the resident private sector (loans by MFIs and debt securities issued by the non-monetary private sector held by MFIs), net claims on the public sector (MFI financing of central government bodies net of their deposits with MFIs), net external assets (MFI credits with non-resident counterparties net of their debts to such counterparties) and longer-term financial liabilities (debt [excluding deposits included in money supply], capital and reserves).
The balance (difference between the volume of exposures identified here and money supply) constitutes “other counterparts” of money supply. Box 2 sets out the respective consolidated balance sheets for MFIs in the USA and the eurozone at 31 December 2020 (the latest date for which comparable data is available). From the MFIs’ viewpoint, the elements making up broad money represent a resource and are thus included on the liabilities of their consolidated balance sheet. An increase in broad money thus finds a counterpart either in an increase in items on the asset side of the consolidated MFI balance sheet, or a reduction in non-monetary resource on the liabilities side.
Sources of money creation and destruction and specific features of the US
The first lesson from our comparison is that over a long period, the counterparts of US broad money are similar in nature and proportion to those of the European M3 aggregate (Charts 1 and 2). We will now examine the main mechanisms in play and the specific features of the US economy.
Credit to the private sector
“Credit creates deposits“
In both economies – the USA and eurozone – and in normal conditions, lending to the private sector represents the main engine of growth in money supply (McLeay, Radia and Thomas, 2014). When a bank lends, it increases the volume of deposits at the same time. In other words, it creates money by crediting its client’s account. This ability to create money from nothing is in part tempered by prudential regulation and is influenced by monetary policy. Meanwhile, the repayment of a loan results in the destruction of money. On the scale of the economy as a whole, an increase in credit (when aggregate new production exceeds total repayments) is thus a source of money creation. The deposit balance can “shift” to the current account of a different bank (for example if the borrower buys a car from a client of a competing bank), but at the aggregate level the volume of deposits rises in line with growth in lending to the economy (save for the conversion of deposits into notes and coins or their “leakage” abroad).
The effect of securitisation of credit
Although the snapshot of the consolidated balance sheet of MFIs provides a reliable image of counterparts of broad money at a given moment, it does not take account of the many creative or destructive flows that cancel each other out. The equilibrium of US money supply, in particular, ignores substantial temporary monetary flows due to the intense use of securitisation of loans (Choulet and Quignon, 2021). Money supply is increased by credit originated by the banks (+) but reduced by the share that is securitised (-). In effect, the money created by adding a loan from a bank to its assets is destroyed when the bank in question sells the loan to a non-banking investor (elimination of the deposits used by the investor to acquire the loan).
MONETARY AGGREGATESIn the case of mortgage lending, monetary destruction occurs, more precisely, at the moment of the placement of Mortgage Backed Securities (MBS): the sale of loans to a mortgage refinancing agency (Government Sponsored Enterprise or GSE) results in a transfer of cash from the GSE’s current account with the Fed to that of the originating bank (increasing the bank’s reserves at the Fed); the subscription by a non-banking investor in an MBS issued by the GSE reduces the volume of deposits and results in a transfer of holdings at the central bank in the opposite direction, from the bank of the subscribing client to the GSE. However, when the investor is an MFI (for example the central bank as part of its quantitative easing programme), the deposit created at the time of the origination of the loan is preserved[9] as the counterpart, consisting of the loan, is simply transferred from one MFI to another and does not leave the consolidated MFI balance sheet.
Purchase of private sector securities
The purchase and sale of securities (debt securities, equities and non-money market fund shares) made by MFIs themselves are another significant source of money creation or destruction.
COUNTERPARTS’ CONTRIBUTIONS TO THE ANNUAL GROWTH RATE OF M2* IN THE UNITED STATESThe purchase or subscription by an MFI of a security issued by a non-monetary resident entity (a non-financial company for example) to another non-monetary resident entity results in an increase in the deposits held by the (resident) entity that sold or issued the security and thus contributes to the creation of money. This ‘monetisation’ phenomenon depends on the nature of the purchaser of or subscriber to the security. The purchase/subscription of the same security by a non-monetary entity (an insurance company for instance) results in the simple circulation of deposits between economic agents rather than a net creation of deposits (money).
Net claims on the Treasury
Purchase of public sector securities
COUNTERPARTS’ CONTRIBUTIONS TO THE ANNUAL GROWTH RATE OF M3 IN THE EUROZONENet credit to central government is a further significant source of money creation, particularly in a period of quantitative easing (QE). This virtually exclusively comes in the form of the acquisition of sovereign debt (with a view to increasing portfolios of liquid assets in the case of commercial banks, or reducing bond yields in the case of the central bank).
In practice, the government’s issuance of debt for the purpose of financing new spending results in several money creation/destruction movements.
Ultimately, the overall stock of bank deposits remains unchanged when the final subscriber to the securities is a non-banking resident entity and increases when it is an MFI. When the subscriber in a non-monetary agent, the debt raised by the Treasury, and its use to finance new public spending, results merely in the circulation of existing money. This is the case, for example, when funds move from the deposit account of an investment fund subscribing to a Treasury issue to the account of a non-financial company that carries out the renovation of a school, or to a household receiving direct payment as part of a stimulus programme[10].
However, when a commercial bank or the central bank purchases government securities from a non-monetary entity (e.g., from a pension fund), the value of the transaction is credited to the bank account of that entity, resulting in an increase in the volume of deposits. Money supply thus increases: the purchase of securities by the monetary institution (money creation) offsets the effect of the initial subscription to the securities by the pension fund (money destruction)[11], whilst the completion of the public spending results in additional deposits[12].
Liquidity circulating between agents, who are the final owners of the deposits created by QE, is not directly identifiable from monetary statistics. Moreover, second-order effects can attenuate or offset the direct effects of QE on broad money[13].
Expansion of Treasury holdings
The net money creation resulting from the purchase by MFIs of government debt securities from non-monetary resident entities is counteracted when the increase in the government deficit comes alongside an expansion in Treasury holdings. When the receipts from net issuance of government securities are not immediately used to finance additional spending, but are partly salted away in the Treasury’s accounts (with commercial banks or the central bank), the money created by the purchase of these securities by an MFI is ‘sterilised’. All other things being equal, there is no net creation of money unless net lending by MFIs to the Treasury increases. In 2020, the effect of QE on the growth in money supply was thus significantly moderated by an increase in the Treasury’s deposits at the Fed (these represented some 26% of the Fed’s liabilities at the end of July 2020, see below).
Net external assets
The net external counterpart (which results, when positive, in net entries of capital) plays little part in the (net) change in money supply in either the eurozone or the USA. In the case of the USA in particular, the snapshot provided by the consolidated MFI balance sheet obscures significant capital inflows and outflows; the former result from direct or portfolio investments by non-residents, the latter from the USA’s current account deficit (Choulet and Quignon, 2021). Where monetary outflows (to pay for imports for example) are not fully offset by capital inflows, ‘net external assets’ at monetary institutions (which includes most notably the central bank’s currency reserves) and broad money both shrink. In the USA, the approximate match-up between capital outflows and inflows (direct and portfolio investments) results if not in neutrality then at most in a modest effect on money supply from net external credit at US MFIs. Despite their size, inflows and outflows largely cancel out, with the result that the “external credit” counterpart plays a negligible role in trends in US broad money.
Longer-term financial liabilities
Subscription by non-monetary agents to term deposits not included in money supply (those with a maturity of more than two years in the eurozone or with unitary values of over USD100,000 in the USA), or to bonds and equities issued by banks, has in the past resulted in significant destruction of money (whilst still preserving the stock of bank deposits in the former case). However, this negative contribution has eased significantly over recent years (the very low interest rate environment has reduced the opportunity cost of holding highly liquid savings, receiving little or no interest, whilst QE has reduced the need for banks to issue debt on the market by improving the coverage of loans by deposits). In 2020, other than in a first quarter marked by extreme liquidity stress[14], the downward trend in these resources even contributed to a net creation of money. In the USA, the central bank maintains close relationships with certain non-monetary financial institutions, due to their central role in the operation of financial markets. At certain times, trends in their deposits at the Fed played a not insignificant role (reverse repos for money market funds[15], deposits from GSEs and clearing houses).
Money does not evaporate
COUNTERPARTS OF THE BROAD MONETARY AGGREGATE M2* SINCE THE 2008 FINANCIAL CRISISThe US experience of quantitative easing and then tightening (reducing the size of the Fed’s balance sheet) provides an opportunity to consider the sensitivity of money supply to the non-conventional monetary policies introduced since the financial crisis, and the factors likely to strengthen or attenuate these effects.
Counterparts of money supply since the financial crisis
In Table 2 we summarise trends in the counterparts of M2 since the financial crisis[16]. Columns can not be compared with each other as they cover periods of different lengths.
COUNTERPARTS OF M2* ON THE FED’S BALANCE SHEET**However, they do provide a picture, for each period of expansion, stabilisation and contraction of the Fed’s balance sheet, of the relative weight of each of the counterparts in the overall change in money supply. This breakdown can mask the effects, significant in some cases, of certain monetary policy measures or certain portfolio choices. We would therefore also refer to Charts 3, 4 and 5, which provide a more detailed, though less unified, view of the respective contributions.
COUNTERPARTS OF M2* ON BANKS’ BALANCE SHEETS** The three successive waves of quantitative easing between end-2008 and end-2014
Purchase of securities, a substitute for the traditional channel of money creation
Net financing of central government and the purchase of private securities (MBS) were the main vectors of money creation between 2008 and 2014 (green and light-grey hatched bars in Charts 4 and 5). Between September 2008 and January 2015 the Fed’s balance sheet grew by USD3,600 billion (and bank reserves at the Fed by USD2,700 billion).
As discussed above, through its purchases of Treasuries (forming the “net claims on central government” counterpart) and MBS (“net credit to the private sector”), the Fed created money by ‘monetising’ long-term debt securities.
Made to a large extent from non-monetary entities, these purchases made up for the drying up of the traditional channel of money creation during the major financial crisis of 2008: thus from the end of 2008, when bank lending to the economy contracted temporarily (repayments or write-offs of a part of mortgage lending, fall in new loan production - light grey bar in Chart 5), client deposits continued to grow rapidly[17]. In addition, from the end of 2013, the prospect of the introduction of the Basel Liquidity Coverage Ratio (LCR[18]) into US regulations was accompanied by an expansion of portfolios of Treasuries on bank balance sheets.
The negative effect of growth in net external debt
The dollar liquidity loaned by the Fed under swap deals with other central banks between late 2008 and early 2009 (“external assets”, dark grey bar in Chart 4) was, ultimately, used as a substitute for the discount window by US branches of foreign banks (facing the abrupt drying up of short-term dollar debt markets, hatched dark grey bar in Chart 5). Its contribution to growth in the money supply was therefore modest[19].
Thereafter, under the combined effect of QE and the change in the calculation of the FDIC commission, repayment flows of cross-border intra-group loans from parent companies, which were more rapid at foreign banks (foreign parents of US branches) than US banks (with regard to their foreign subsidiaries and branches), served to increase commercial banks’ net external debt (“external liabilities”, grey area of Chart 6 - see Choulet 2015). The trend was prolonged by net inflows of intra-group financing that all resident banks enjoyed (particularly in 2011 and 2013). The increase in banks’ net external debt between 2008 and 2014 therefore contributed to a destruction of money (dark grey hatched bar in Chart 5).
This effect was partly offset by a reduction in longer-term resources
The Fed’s Reverse Repo Program (RRP) for money market funds between late 2013 and 2015 had a negative effect on the volume of deposits (see below - Table 2 and black bar in Chart?4). The first three episodes of QE nevertheless favoured an even more significant contraction in banks’ non-monetary liabilities: deposits not included in M2, secured debt (repo, Federal Home Loan Bank advances) and unsecured debt (Fed Funds - see Table 2 and the black bar in Chart?5). On a consolidated basis, the contribution of the “longer-term financial liabilities” counterpart for MFIs was thus positive between 2008 and 2014 (Table 2).
Periods of stabilisation and then reduction in the Fed’s balance sheet
Over these two phases, base money (bank reserves with the Fed and fiduciary money, recorded as a liability on the Fed’s balance sheet) contracted, whilst growth in M2 merely slowed.
Between October 2014 and October 2017, the Fed stabilised the size of its balance sheet, by reinvesting, in full, maturing debt in its securities portfolio[20]. All other things being equal, its purchases of government and private sector securities, which were nil when netted against redemptions, will thus have had no effect on money supply. Over this period, money creation stemmed from a resumption of bank lending and banks’ subscription to Treasuries and MBS (responding to the LCR constraint, light grey bars of Charts 3 and 5). Other counterparts had little impact on trends in M2 over this period[21]. From October 2017, the Fed undertook a programme to reduce the size of its balance sheet, by limiting the amount of reinvestment on maturity. Within the space of less than two years (from October 2017 to August 2019), its balance sheet was reduced by USD700 billion (destruction of USD750 billion in central bank reserves). This policy, which can be considered as the net disposal of securities, resulted in the destruction of deposits (green and light grey hatched bars of Chart 4).
NET EXTERNAL DEBT OF US RESIDENT BANKS SINCE 2010In addition, some non-banking entities acquired newly issued Treasuries and MBS, thus reducing their deposits. Growth in bank loans, repurchase arrangements for securities ‘stuck’ on the balance sheets of primary dealers (Choulet, 2019) and purchases of Treasuries and MBS by banks nevertheless offset these effects (light grey and light grey hatched bars of Chart 5).
The negative effect on M2 of the increase in longer-term financial liabilities at banks (subscription of term deposits against a background of rising interest rates from the end of 2015, black bar in Chart 5) was, amongst other things, largely attenuated by the reduction in the Fed’s non-monetary commitments (ending of the overnight renewal of repo arrangements with money market funds, black bar in Chart 4).
The latest phase of expansion in the Fed’s balance sheet and the injection of central bank liquidity
THE TREASURY’S ACCOUNT AT THE FED
BREAKDOWN OF TREASURIES BY HOLDER SECTORAgainst the background of the Covid-19 pandemic, purchases of securities by the Fed (and by banks) once again represented the main vector of money creation[22] (green and light grey hatched bars in Charts 4 and 5). Within the space of just eighteen months (August 2019 to end-December 2020), compared to six years in earlier QE phases, the Fed’s balance sheet grew by USD 3,600 billion (increase of USD1,600 billion in central bank reserves).
Although over this period the Treasury sharply increased its deposits with the Fed (up by USD1,400 billion between September 2019 and December 2020, Chart 7), thus sterilising part of the newly created money (green hatched bar in Chart 4), and although non-residents reduced their holdings of Treasuries, the resulting growth in deposits was strong, particularly as part of these purchases were, as in previous phases of QE, made by households[23] and US pension and investment funds[24] (Charts 8 and 9). In 2020, bank lending, the traditional engine of money creation (light grey bar in Chart?5), was also protected, temporarily, by clients drawing on lines of credit and then by the Treasury’s introduction of a programme of guaranteed loans to small and medium-sized enterprises (Paycheck Protection Program, PPP).
After the first three phases of QE, the average ratio of loans to deposits at commercial banks had improved by 26 points in the space of six years, from 99% at end-October 2008 to 73% at end-October 2014, close to its record low. The latest round of QE brought about an improvement on a similar scale in just one year: the ratio dropped from 75% at the end of March 2020 to 61% a year later (Chart 10), a level not seen since the mid-1950s.
TRENDS IN COUNTERPARTS OF BROAD MONEYTowards more modest growth in broad money
BREAKDOWN OF AGENCY MBS BY HOLDER SECTORIn the first half of 2021, growth in money supply slowed significantly (Chart?11). Whilst over the medium term, the continuation of an accommodating monetary policy and the expected recovery in demand for credit will bolster M2 growth, in the shorter term, banks’ strategies of reducing balance sheets could slow this growth.
Continuation of the QE4 programme, even at a slower pace, will support money creation
A CLEAR IMPROVEMENT IN THE LOAN TO DEPOSIT RATIO OF US COMMERCIAL BANKS Over the first half of 2021, the Fed’s net claims on the Treasury, via the continuation of QE4 and the reduction in the Treasury’s account at the Fed (down by USD800 billion between end-2020 and early June 2021, Chart 7) have provided a powerful vector of money creation[25]. Their impact is likely to remain significant over the next few quarters. It is certain that drawing down of the Treasury’s deposits will slow[26]; to our knowledge, however, there is no factor that would warrant the rebuilding of deposits on the scale seen in 2020 (Cecchetti and Schoenholtz, 2020). Moreover, the Fed’s monetary policy committee is unlikely to shift from its accommodating stance for a number of quarters (until a return to full employment[27]). For the time being, analysts are suggesting an announcement in the second half of 2021, with tapering of net purchases starting between 9 and 12 months later. It is worth bearing in mind that three years elapsed after the end of QE3 before the Fed began to shrink its balance sheet (synonymous with money destruction).
The credit channel is temporarily weakened
In the first quarter of 2021, for the first time since the third quarter of 2011, bank loans to households and business contracted (down 2%, Chart 12). However, this contraction reflected a significant negative comparison effect: in March 2020, drawing on lines of credit and the introduction of PPP (Paycheck Protection Program) for small and medium-sized enterprises resulted in exceptional growth in lending to business, whilst the introduction of mortgage repayment forbearance artificially maintened households’ stock of debt (and reduced the likelihood of payment defaults). New production of PPP loans in 2021 (USD190 billion issued by banks in the first quarter, USD220 billion over the first five months of 2021) did not fully offset the elimination of guaranteed loans in 2020. Thus in the first quarter of 2021, 40% of outstanding PPP loans issued in 2020, for a total of USD210 billion, were eliminated (54% and USD280 billion respectively over the first five months). At the end of March 2021, only one third of households still benefited from mortgage payment forbearance (Haughwout, Lee, Scally and van der Klaauw, 2021). Lastly, nearly one third of the value of the ‘stimulus checks’ distributed to households was used to repay existing loans (see below). Banks surveyed by the Fed between 22 March and 2 April 2021[28] indicated that they had relaxed lending criteria on a large range of loans (commercial and industrial loans, mortgages, consumer loans) and had observed an increase in demand for credit, at least amongst consumers (mortgages, credit cards, car loans) in the first quarter. Over the next few quarters, growth in outstanding loans will remain closely linked with the outlook for the economy. In comparison with the aftermath of the financial crisis, the less damaged financial position of households (lower indebtedness rate, increased savings rate, greater real estate wealth[29], more modest increase in the unemployment rate) should help support lending growth. However, the possible emergence of credit risk could slow the process and require fresh allowances to provisions.
Monetary tightening would result in switching between deposits
In general terms, changes in money supply (M2 in the USA) and its principal counterpart (credit) are closely linked to movements in interest rates (Chart 13). An increase in interest rates is not only liable to hold back money creation (by slowing demand for credit) but, by increasing the opportunity cost of holding liquid savings, can also lead certain investors to switch to more lucrative deposits (Dreschler, Savov and Schnabel, 2017). The average interest rate paid on bank deposits is sensitive to changes in the effective Fed Funds rate (Chart 14). We can also see that the share of interest-bearing deposits in total deposits tends to rise (fall) in a period of monetary tightening (loosening). This said, the scale of the transmission of monetary policy to the interest rates paid on deposits, and the resulting arbitrage, varies according to the periods in question and the phase of the process (a reduction in Fed Funds rates is generally more widely passed on than an increase, Box 3). Moreover, at this stage it seems unlikely that the target range for Fed funds will be raised before 2023.
SLOWDOWN IN THE MONETARY GROWTH SINCE MARCH 2021The Biden stimulus packages will encourage money creation
FIRST ANNUAL CONTRACTION IN OUTSTANDING BANK LOANS TO HOUSEHOLDS AND BUSINESSES SINCE 2011
THE MONETARY AGGREGATE M2 AND THE FED FUNDS RATEWith all other things being equal, the direct impact of the Biden administration’s stimulus plans on broad money and bank deposits will depend on how they are financed and allocated. The USD1,900 billion American Rescue Plan (Proutat, 2021a), approved in March 2021 and financed mainly from the Treasury’s deposits with the Fed and an increase in the deficit (part of which will be ‘absorbed’ by the Fed under QE and potentially by the banks), is likely to boost growth in M2 (increase in the “net claims on central government” counterpart). Increased appetite for issues of Treasuries amongst foreign investors will also contribute. However, second-order effects could affect the final impact. For instance, a survey conducted in January by the Federal Reserve of New York showed that US households that had received ‘stimulus checks’ under the stimulus packages approved in 2020 (the CARES Act and then the CRRSA Act) indicated that they had allocated – or planned to allocate – a substantial share of the money to the repayment of outstanding loans (34.5% of the initial payments in June 2020 and 37.4% for the second in January 2021 - Armentier, Goldman, Kosar and van der Klaauw, 2021)[30]. The planned allocation of sums distributed under the American Rescue Plan (cheques from the Treasury, additional unemployment benefits, tax credits) also suggest a ‘destruction’ of around one-third of the amount paid out to households (33.7% will be used to repay a debt, 24.7% will be consumed and 41.6% will be saved). The Biden administration, however, plans to finance its American Jobs Plan investment programme out of tax revenues over eight years (still under discussion in Congress, Proutat, 2021b). Thus its positive (direct) effect on M2 will in part be offset.
Over and above the details of their financing and the nature of the initial use of the funds paid, the additional economic growth and the knock-on effect on new loan production will mean that the Biden plans will favour M2 growth.
SENSITIVITY OF THE INTEREST RATE PAID ON DEPOSITS TO THE FED FUNDS RATEReduction in bank balance sheets liable to hold back growth in deposits
The money created by quantitative easing has also resulted in a marked expansion of bank balance sheets. This in turn has tightened regulatory capital constraints on the very largest banks (leverage ratios[31], systemic surcharge, total loss absorbing capacity[32]) and caused smaller banks to worry that their regulatory framework might be tightened up[33]. Various options have been cited by banks[34]: adjusting their dividend payout plans, issuing long-term debt (eligible for the Total Loss Absorbing Capacity calculation) and/or seeking balance sheet reduction solutions. These strategies would all affect the trend in deposits[35].
Strengthening Total Loss Absorbing Capacity (TLAC)
The ambitious payout plans announced by the Global Systemically Important Banks (G-SIBs) in early 2021, after a lean year for shareholders in 2020 (Chart 15), will help boost money creation (reduction in the “longer-term financial liabilities” counterpart).
US G-SIBS BUILT HIGHER CAPITAL BUFFERS SINCE 1 YEAR
SUPPLEMENTARY LEVERAGE RATIOS AS OF Q1 2021Cash paid out at the time of share repurchases or dividend payments by banks may be stashed away by shareholders in the form of bank deposits or could be reinvested in securities issued by other resident companies, whether financial or not[36]. However, the overall effect of these plans needs to be seen in context: 1/ they remain subject to the results of the 2021 CCAR stress tests; 2/ the sums paid out, although significant at the level of the banks, will probably be negligible relative to the contribution from other M2 counterparts[37]; 3/ leverage requirements for the main depository institution subsidiaries (Chart 16) and the increase in their systemicity scores (Chart 17) could force G-SIBs to reconsider these plans or take measures to shrink balance sheets (see below); 4/ G-SIBs could be required to rebuild their Total Loss Absorbing Capacity.
THE RISE IN G-SIB SCORES COULD TRANSLATE INTO A RISE IN SURCHARGES 
8 G-SIBS’ NON-OPERATING DEPOSITS INCREASED BY MORE THAN USD 380 BN BETWEEN Q4 2019 AND Q1 2021In 2020, the profits made and the ban on major banks buying back their own shares resulted in a temporary strengthening of the G-SIBs’ capital amounts and risk-weighted capital ratios. In the first quarter of 2021, the TLAC and Long-Term Debt (LTD) ratios, expressed as a proportion of average risk-weighted assets also remained comfortable relative to the minimum requirements[38]. However ratios expressed as a proportion of leverage exposure deteriorated for some banks, and appear close to the minimum required levels after taking account of the expiry on 31 March of the April 2020 rule (which allowed banks to deduct their reserves with the Fed and Treasuries portfolios from their leverage exposure, Chart 18).
LONG-TERM DEBT AS A % OF THE LEVERAGE EXPOSURE AS OF Q1 2021Penalising non-operating deposits
An easing of balance sheet constraints could also take the form of a penalty for non-operating deposits, which grew by 35% between December 2019 and March 2021 on the balance sheets of the 8 G-SIBs (Charts 19 and 20). Granted, taken alone, a penalty on deposits from institutional clients would not be enough to reduce the volume of deposits on a macroeconomic scale. Only if it were coupled with a re-activation of the Fed’s RRP would it drive this liquidity towards money market funds, and thence to the Fed, and thus contribute, as it did in 2014-2017[39], to a destruction of deposits (on at least a temporary basis).
JP MORGAN, BANK OF AMERICA AND CITIGROUP HELD AROUND USD 1000 BN OF NON-OPERATING DEPOSITS AS OF Q1 2021Although its main aim was to establish a floor for short-term market rates[40] by draining off excess liquidity from the repo and Fed Funds markets, the RRP[41] provides a form of sterilisation of the broad money. A reverse repo transaction can be considered as a secured loan: the Fed lodges securities held on its balance sheet with a money market fund which, in return, ‘lends’ it central bank money. In practice, these transactions result in a destruction of base money (reserves on the asset side of bank balance sheets) as they are conducted via bank balance sheets, but also a destruction of money supply (client deposits, in this case money market funds’ deposits, shown as liabilities on bank balance sheets). The central bank recognises a debt to the money market fund on its balance sheet and debits the same amount from the current account (reserves with the Fed) of the commercial bank acting as an intermediary in the transaction. This bank in turn debits the account of the money market fund which has accepted securities from the Fed. The ‘deposit’ made by the money market fund ‘with the Fed’ under the RRP transaction reduces the volume of cash that the fund invests in debt securities in the traditional fashion, or that it lends to other financial institutions on the repo markets or places on deposit with banks[42]. Stopping the renewal of these transactions (generally from one day to the next) results in the reinjection to the economy of deposits which had previously been sterilised. Thus these transactions only temporarily reduce broad money.
Although the current excess of liquidity, leverage constraints and the drying up of the market for short-term Treasuries[43] are also contributing, the rapid growth in liquidity lodged with the Fed under RRP since the end of March (even though the Fed is still offering only a zero interest rate) suggests that these transfers of deposits are already under way (Chart 21).
THE STERILISATION OF NON-OPERATING DEPOSITS MAY ALREADY BE IN PROGRESS****
Our exercise of reconstituting a consolidated balance sheet for American monetary and financial institutions shows that although the methods of financing the two economies are very different, the sources of money creation in the eurozone and the USA are quite similar. Since 2008, in particular, quantitative easing (QE) policies have become a key tool in crisis management and an important factor for creation of base money and broad money. In the aftermath of the 2008 financial crisis, the impact of the Fed’s programme of purchasing government debt on money growth was mitigated by the debt reduction efforts of the private sector and the increase in banks’ net debt to non-residents. In 2020, the effects of the exceptional measures taken by the US authorities to bolster the liquidity of companies and markets in response to the Covid-19 crisis have in their turn been partially mitigated by the expansion of the Treasury’s deposits at the Fed. The rapid growth in money supply that has resulted has nevertheless been unprecedented, and at a level liable to feed fears of inflation (Mandelman, 2021). Granted, strategies to shrink bank balance sheets, which are probably already under way, will continue to slow this growth. Even so, continued QE (albeit at a reduced pace), the expected economic recovery and the compression of long-term interest rates as a result of QE will all help boost money creation. A further area of research would be to look at the way in which monetary policy, changes in the regulatory and accounting framework, economic support measures and health protection measures have affected the distribution of deposits between sectors (households against non-financial companies) and how planned stimulus packages, adaptation strategies and the gradual normalisation of the health and economic situations could affect this distribution.
