A tremendous credit boom took place in 2022 and it may not even be over. The combination of healthy banks, financially strong households, and attractive rates appears to have to led to a surge in bank lending. Banks and credit unions together created $1.5t in cash last year that likely has not yet fully filtered into economic activity. Recall, bank lending creates money out of thin air. Interestingly, higher interest rates have so far only shown very tentative signs of moderating the boom. This post reviews the credit boom of 2022, suggests it was due to the strong financial position of banks and households, and notes that it will be supportive of demand throughout the year.
Partying Like It’s 2006
The total size of the banking sector was little changed over 2022, but the static surface obscures a boom in lending of epic proportions. Banks changed the composition of their assets by replacing their cash and security holdings with loans to the real economy. Around $1.2t in loans were made in 2022, a level around three times higher than that of recent years. The same explosive growth is also seen in credit unions, which are functionally similar to small banks. Credit union loans outstanding grew $0.23t from 2021Q3 to 2022Q3 (Q4 data not available), a level also three times higher than recent years. Loan growth was strong across categories and appeared to persist despite rising rates.
Note that explosive loan growth is likely the key reason banks recently increased market borrowings, as deposit growth usually lags loan growth. Banks may be temporarily borrowing from Federal Home Loan Banks and even the discount window to keep up with loan growth.
Initial Conditions
Credit growth is influenced by interest rates, but also the willingness and capacity of banks to lend and borrowers to borrow. As noted before, the Fed’s post-GFC change to transmitting monetary policy by adjusting marginal asset returns rather than marginal funding costs encourages loan growth when rates are high. Higher rates mechanically raise net interest margins as funding costs are little changed. As shown by their resilience throughout March 2020, banks today are financially sound and capable of lending. This is a very different context than the post-GFC years when they were nursing significant losses.
The strong financial position of households suggests they have a high capacity to take on additional debt. Household net worth remains historically high and many have a favorable debt profile comprised largely of long term mortgages at very low rates. At the same time, the historically strong labor market continues to push wages higher. This is a combination that makes lenders comfortable to lend and borrowers confident to borrow. It may also make credit growth more resistant to higher interest rates.
Money, but not Free
Increasing the supply of money boosts economic activity by directly increasing the purchasing power of the public. The enormous fiscal stimulus in 2020 created a few trillion out of thin air and just gave it away to the public – predictably supercharging growth and inflation. Note that fiscal stimulus is very different from QE, which merely exchanges Treasuries for cash. QE changes the composition of liquid assets held by non-banks (fewer Treasuries, more cash), but not their purchasing power. In contrast, stimmy checks and forgivable loans are essentially free “helicopter money” that increase potential demand.
The huge credit growth in 2022 can be likened to the prior fiscal stimulus, with the exception that the money must one day be repaid. Borrowers have $1.5t more in purchasing power that they did not have before. The need to repay the money may affect their spending decisions and willingness to take on additional debt, but credit cycles can last for years. The housing led boom of the early 2000s continued for years even as rates rose. Financial conditions today have been loosening with rising equity pricing and declining yields, potentially extending the cycle. At the very least, the credit boom thus far should sustain demand for the coming months.
31 comments On Credit Boom
First, thank you for your commentaries Joseph, I find them quite interesting and informative. I’m wondering if you know Mike Ashton, aka @inflation_guy, but his point has been that as interest rates go higher, so will money velocity and therefore the opportunity for higher inflation is very realistic. This seems in accord with your view that demand will be supported more fully thus making for a more robust economic picture, maybe that mythical soft-landing.
Great article! Can someone please translate “adjusting marginal asset returns rather than marginal funding costs” to a 5 year old mind here? Also, an example of each would be very helpful. Thank you!!
You should click on the link in the sentence you are referencing.
Jake, what he is trying to say is that the Fed now conducts monetary policy by changing the interest it pays on excess reserves (which is an asset to commerical banks) instead of the fed funds market (a liabliity to a borrowing commerical bank in a world where excess reserves dont exist a la pre 2008)
That being said, i disagree with his point that lending has increased because rates have increased. By paying higher rates on excess reserves, commercial banks are earning even more on a risk-free asset. At the same time, the beta on their deposits is less than 1, so the marginal return on excess reserves exceeds the marginal cost of their liablities by doing nothing.
Lending is actually slowing down now because the credit risk premiums bank need to charge above and beyond the risk-free rate is increasing to make it worth their while.
Also, lets remember that loan growth figures are stated in nominal terms, not real terms. Nominal GDP was off the charts in 2022, but driven by price increases / inflation. Lots of borrowing to beef up capacity / supply. But that is going away soon.
But his overarching point that lending has created more M2 is valid.
So in 2022, banks create $1.2T of money, and the Fed shrinks its balance sheet by roughly $400B. M2 is down roughly the same amount, $400B. I would have thought M2 would be up $800B. What is the other factor I’m missing here?
During that time, some private debt was paid off. Paying off debt destroys money.
The difference is largely due to the RRP absorbing an additional $700B over that period, plus growth in the Treasury Government Account of around $200B.
Thanks for the response. That math makes sense though I still don’t entirely understand the reverse repo accounting. If the majority of money in the reverse repo is from money market funds, isn’t that counted in M2?
The bizarre thing about the money stock is that the FED doesn’t know a credit from a debit.
See: https://www.jstor.org/stable/2976184
The O/N RRP turns inside money into outside money (like the TGA). Contrary to the FED’s spurious accounting, re: “the bond underlying the repo transaction is still recorded on the Fed balance sheet”, O/N RRPs are contractionary.
That, of course, is an accounting error according to the Federal Reserve Bank of Chicago’s “Modern Money Mechanics”. “If the buyer of a reverse repo or a security sold by the Fed is a nonbank (which 90% of RRPs are), and pays for the purchase using its bank account, the money supply is directly affected”.
As far as I know ‘Loans and Leases in Bank Credit, All Commercial Banks’ is not a leading indicator as ‘Net Percentage of Domestic Respondents Tightening Standards for C&I Loans’ has been. This leading indicator shows that lending standards have become much more restrictive and have preceded recessions in the next two years, counting from the low.
Joseph, your graph says the job losses stemming from Covid have been addressed. Yet this graph shows that only 60% of USA displaced workers have returned to work: https://d15g8hc4183yn4.cloudfront.net/wp-content/uploads/2023/01/02160813/j8rpfhz8nm9a1.jpeg
Also, why not be more sober in evaluation of US household wealth (and even this graph looks overly optimistic to me): https://thedailyshot.com/wp-content/uploads/US-Real-HH-NetWorth2212130536.png
Regarding your second point, here is an alternative viewpoint of the household balance sheet and household debt. When someone takes on a large mortgage, they will likely repay that using a portion of income earnt from employment. Using finance 101 principles, a high-paying job (cash flow is high) in a stable profession (risk is low) would result in a high present value stream of income. This does not get recorded as an asset but influences the size of household debt. So if incomes have increased and job security is higher then you would expect a typical household balance sheet to look more bearish.
Is this perhaps the explanation for the Atlanta Fed Now numbers being so high?
The contraction of M2 is already having its impact on weaker economy, weak commodities, lower inflation, peak in the rate of 2-year Treasury note.
M2 vs. CPI since peak
7115.2 ,,,,, 296.311… Jun
7133.5 ,,,,, 296.276… Jul
7151.1 ,,,,, 296.171… Aug
7375.7 ,,,,, 296.808… Sep
7308.0 ,,,,, 298.012… Oct
7280.8 ,,,,, 297.711… Nov
7288.5 ,,,,, 296.797… Dec
Nice one, thx! What about data from 2020? Why is that not in the chart?
Why would banks need to tap the discount window to keep up with loan growth/demand?
Thanks – an education, as always.
How do you reconcile this with the recent collapse in mortgage applications for purchase, and even moreso for refi?
Hi Jeff. Thank you for your work. What do you mean by deposits lagging loans? Aren’t loans supposed to create the deposits in the first place?
I think Joseph’s choice of wording is confusing. “Deposit growth usually lags behind loan growth” can infer that deposit growth is temporarily below loan growth but will catch up eventually.
However, I think a better way to phrase it is “deposit growth is usually below loan growth” because banks will fund loans with bonds and equity, both of which when sourced from capital markets destroy deposits.
See Graph 2 from a Reserve Bank of Australia speech titled “Money – Born of Credit?”
Explaining this one step further, when a bank buys a security it is effectively making a loan to the issuer, creating a deposit in doing so, In the short-term banks fund new loans by terminating these security loans, so no net new bank credit is generated. Over time however, banks tend to rebuild their securities portfolios in proportion to their deposits.
I’ve not heard of this kind of mechanism and I’m struggling in following your example. I’d be interested in doing further reading on it if you have anything you can reference. I’ve got access to Joseph’s book if he covers it there.
I might be missing something here. How have commercial loans and leases gone by 1.2trillion? The H8 shows Commercials loans and leases increased from 2.47 trillion to 2.80 trillion. Thats approx 400bn. What am i missing?
Joseph, I humbly disagree. There is no difference between QE and fiscal stimulus. Anytime it is paid for by the Fed, it is fiscal stimulus.
Without Fed involvement if the government wants to spend money (covid, military, normal outlays, whatever) it can only do so by tax revenue or borrowing money. There is no free lunch, i.e. fiscal stimulus. It is at a cost: taxes, or higher interest rates pushing debt on the market.
Only the Fed can enable fiscal stimulus by buying the debt by printing money.
It makes no difference whether the Fed is buying the debt directly from the Treasury, or whether the debt has been out in the market place for a few days, months, years. It is still financing the treasury.
This has become one of the most confused financial issues in our times.
I agree 100%. I feel this topic often gets lost in technicalities and semantics (“it is not money printing – it is just an asset swap: bonds for bank reserves”).
The way I see it is this: the very moment the Treasury deficit-spends some money that it is not paid for via A) taxes or B) debt actually held by the private sector, then some level of money printing has taken place. Whether that printed money materializes in some obscure/virtual bank reserves held at the Fed is irrelevant, because the Treasury’s initial spendings into the economy was indeed very real. Furthermore, QE keeps interest rates artificially low, further enhancing the Treasury’s capacity to issue yet more debt. Finally, lets not forget that any interest coupons the Fed might harvest from the trillions of dollars worth of bonds held at its Balance Sheet get automatically reverted back to the Treasury, making their financing costs virtually free!!!
Lets simply ask ourselves the following question: if the Fed hadnt interfered, could the Treasury have financed itself as much and as cheaply as it has? The answer is a clear NO, and since the Fed has no revenue means other than the printing press, then we must conclude that the Treasury has directly benefited from the Fed printing. Period. All the rest is just noise, technicallities and semantics.
Ding ding ding.
Right, the 1951 Treasury-Federal Reserve Accord has been obverted.
I’m really having a hard time following the logic here. Has income for the average household gone up 3x ? I keep on reading here that the consumer is healthy and that is why they are willing to borrow more. However increasing loans to $1.2 trillion dollars without income that increases a proportional amount seems unsustainable. If the assumption is that balance sheets are strong because net worth is high, maybe that’s a lagging indicator and can change in 2023. Consumer cash flow unfinanced seems like a more reliable marker
Thank you Joseph, can you also comment on why the graph for “Personal Savings” vs. “Checkable Deposits and Currency Household Asset Level” at the FRED website are diverging? Is there no motivation for people to save their cash for some non-apparent reason?
It’s called dis-savings, a decrease in money demand.
I’m wondering to what extent the increase in commercial bank lending can be attributed to more educated consumers/ businesses. In early 2022 many were already warning of inflationary pressures and the Fed’s impending rate hikes. Perhaps the combination of higher net worths and knowledge that rates were going up led individuals and businesses to a one-time attempt to secure lower rates ahead of/ early during the Fed pivot. Perhaps we’ll see a precipitous fall in 2023.
Disclaimer: I have no idea wtf I’m talking about.