Americans are spending a lot more at banks in ways they can’t see

Posted on November 18, 2016

Since the start of 2013, American consumers have boosted their spending on “financial services furnished without payment at commercial banks” by about 57 per cent. It’s been decades since this category of consumer spending has grown so rapidly:

US pce commercial banking imputed

(The source is table 2.4.5U of the National Income and Product Accounts.)

The recent boom is staggering when compared to the behaviour of total personal consumption. Since January 2013, the only category of spending to have grown faster is truck leasing — and truck leasing is only about one-sixth as big ($19 billion versus $121 billion as of September).

So what’s going on?

Start by understanding what the Bureau of Economic Analysis is actually measuring. By definition, these are services people don’t explicitly pay for, so economists have to impute their value using other data. The methodology is described in painstaking detail here, but the summary below should give you the gist.

Banks are engaged in two basic business models: they extend credit on the asset side and they provide money-like savings vehicles on the liability side. Depositors effectively pay banks the difference between a risk-free reference rate and the lower amount of interest they actually earn, while borrowers pay the difference between the interest rates on their debts and a “fair” market rate. Total consumer spending on “commercial banking services furnished without payment” is therefore the interest depositors forego plus the extra interest charged to borrowers.

You’ll notice there are lots of variables in there.

The risk-free reference rate is a blend of yields on Treasury debt and agency debt — excluding agency MBS — that’s smoothed according to a special formula. This reference rate is always between the average rate on deposits and the average rate on loans. Some of this is due to the spreads banks charge borrowers and take from depositors, and some of this can be explained by the fact that the average maturity of the debt used to calculate the reference rate is somewhere between the maturity of deposits and loans.

Imagine a depositor has $10,000 in a checking account earning no interest in a world where the risk-free reference rate is 10 per cent. This hapless creditor is effectively paying $1,000 each year to his bank for the benefit of whatever services are offered. The BEA uses this basic approach to measure the consumption spending by depositors by adding up the amount of household deposits and comparing the interest paid by banks against what would be fair.

The calculation for borrowers is trickier, because you can’t just compare the cost of a loan against the reference rate. After all, a mortgage or a credit card has a higher risk of loss than a US Treasury note. Lending money to people who don’t pay all of it back isn’t a sustainable business model unless you charge sufficient interest.

The BEA reasonably decomposes the actual interest rate paid by borrowers into 1) the risk-free reference rate 2) compensation for expected losses based on smoothed data on charge-offs, and 3) an additional spread to cover the costs of marketing and servicing and profits. Only that last component represents actual consumer spending.

Putting it all together, there are four variables that affect the BEA’s estimate of “consumer spending on commercial banking services furnished without payment”: the volume of household deposits, the stock of debt owed by households to banks, the reference rate, and the expected default margin. All four of these have moved significantly since the start of 2013.

According to the BEA — we asked them — rapid growth in household deposits and moderate growth in consumer debt would have boosted total nominal consumption of commercial banking services by about 24 per cent since the start of 2013 even if the reference rate and the default margin had stayed fixed. (Mortgage debt has been essentially flat.)

Put another way, about two-fifths of the total growth in imputed spending in the past few years can be explained by organic growth in bank balance sheets.

Meanwhile, the BEA tells us their risk-free reference rate has increased from about 0.9 per cent then to about 1.5 per cent now. Many deposit rates haven’t gone up at all, much less by 60 basis points, so the impact on the implied payments by short-term bank creditors would be worth tens of billions of dollars. That effect could have been partly muted by the impact higher reference rates might have had on the imputed payments by borrowers, especially since mortgage rates have been basically the same at the beginning of 2013 as they were back in September. However, we were told the change in the reference rate has been a big part of the expalantion.

Finally, the expected default margin has plunged because charge-off rates on bad debts are much lower now than they were just a few years ago. This is particularly noticeable in the case of single-family residential mortgages, where the charge-off rate went from 0.89 per cent at the start of 2013 to 0.11 per cent by the middle of 2016. (The recovery in house prices probably helped.) Over the same period, the charge-off rate on consumer loans — autos, credit cards, etc — fell by 51 basis points, from 2.33 per cent to 1.82 per cent:

Charge off rates

These charge-off rates aren’t the same as the default margin the BEA uses, because they aren’t smoothed, but they are the main inputs to their formulas and can help explain a big chunk of the boost in reported spending.

Where things get really interesting is when you convert these nominal figures into real data to get a sense of the inflation rate in these banking services. Chapter 5 of the BEA’s handbook to the National Income and Product Accounts explains how this is calculated:

The annual estimates of real PCE for commercial bank services are derived using a BLS banking output index that is based on volume measures for the deposit, loan, and trust functions of commercial banks…For U.S.-owned banks, the BLS deposit index consists of a demand deposit component, based on the number of checks processed and the number of electronic transactions; a time deposit component, based on estimated deposits and withdrawals; and an ATM component, based on ATM and point-of-sale volume.

The BLS loan index is based on the number of real estate, consumer, and commercial loans outstanding and on the volume of credit card transactions.

The chart below (note the y-axis) compares the change in amount of money indirectly spent by consumers at banks and the real value of the services they are getting:

US pce commercial banking imputed real vs nominal

Real and nominal spending tracked each other reasonably well between 1990 and 2008, implying the price of commercial banking services hadn’t moved much. There was a lot of growth in the real services provided to customers shortly following the crisis — think of ATM withdrawals, the growth of electronic payments, etc — and then stagnation.

Nominal spending disconnected from real in those years, but caught up by 2013. Since then, almost all the growth in nominal spending has come from inflation. The average inflation rate of commercial banking services furnished without payment since the start of 2013 — about 12 per cent per year — has been far faster than any other category of consumer spending. In a distant second place is “film and photographic supplies”, with an average annual inflation rate of around 7 per cent. Overall personal consumption inflation has averaged just 1 per cent over the same period.

The gap between the broad inflation rate and the change in the price of commercial banking services furnished without payment is unprecedented:

US imputed commerical bank services inflation vs pce inflation 45 months

(We looked at 45-month changes because that is what you get if you start in January 2013.)

One way of interpreting this: banks have made a lot of money off of consumers in the past few years by failing to adjust the interest rates they pay by changes in market conditions and by failing to adjust the interest rates they charge by changes in default risk. The net effect has been a boom in implied spending with almost no increase in the value of the service provided. Whether this can continue is anyone’s guess.

Related links:
Why does the BEA think American media are in a golden age? — FT Alphaville

Copyright The Financial Times Limited 2016. All rights reserved. You may share using our article tools. Please don’t cut articles from FT.com and redistribute by email or post to the web.

You must be logged in to post a comment Login