A number of smart FinSky types got in touch following our recent post on how UK households have become self-funding, meaning interest rates are having increasingly inter-household distributional impacts.
While that broad thesis remains intact, we made a couple of bold claims along the way. Firstly, we claimed that household deposits have grown to be around the same size as household debt. Second, we claimed that households had become — in aggregate — net recipients of interest.
While the first claim was uncontroversial and amazing, the second seemed too incredible to be, uh, credible? Sure, we had the charts to prove it — but it turns out that we’d made some assumptions that rendered the claim true only in a fairly narrow sense, and potentially only to national economic accounting fanboys/girls. This follow-up post explores the dank (in the traditional sense) niche of economic accounting that makes the statement possible, because dank niches are fun.
It’s easy to see why people had difficulty with the claim. If households have got £2.1tn of deposits and £2.1tn of debt, the idea that households are making a positive spread implies that someone is paying them a (net) spread. Overwhelmingly, households face banks. And many banks’ entire business model is making sure they get paid a spread, if only to make payroll on the six hundred thousand odd staff they employ.
Let’s consider potential explanations.
Perhaps banks have done a terrible job at extracting revenues from households? No. The Bank of England reckons that UK bank profitability has been strong and net interest margins consistently positive.
Perhaps some other sector of the economy is picking up the bill, and banks are just middlemen making a turn? If so, who would these others be?
Candidate number 1: swap counterparties. Given the quantity of mortgage borrowing done at historically low fixed rates (while deposits tend to receive variable), maybe households are reaping the rewards of having (accidentally or otherwise) gone short duration amid one of the worst bond bear markets of all time by initiating mortgage fixes, and it’s really banks’ swap counterparties that funding them? This sounds vaguely credible, and is not immediately easy to disprove.
Candidate number 2: taxpayers. Recall that the UK Government is spending around £37bn a year remunerating bank reserves. Those arguing against introducing unremunerated reserve requirements have done so on the basis that this would be a tax on banking (rather than on banks), sucking billions from some mix of depositors, lender and bank profits. Households are depositors and lenders. Maybe household net interest surplus is being funded by the taxpayer? We’re not sure this holds up to scrutiny.
Or perhaps we just screwed up our data feeds? Nope. Table 1.7.3 of the ONS’s Blue Book shows Households (S.14) Interest (D.41) Uses (HACY) are definitively lower than Resources (HAXV). Phew.
But did we really understand what was going on inside those ONS numbers? Several learnèd respondents asked about the choices we made in the treatment of FISIM.
Spoiler alert, we did not dwell on choices made around the treatment of FISIM. In fact, we didn’t consciously make choices about how to treat FISIM allocations before comparing household interest (uses and resources). This was a big mistake. Because FISIM choices matter.
Translation, please.
FISIM stands for financial intermediation services indirectly measured.
As the ECB 2023 primer on measuring household income explains:
FISIM represents the financial intermediation services provided by financial institutions that are not explicitly invoiced but are relevant for the measurement of the output of banks and thus may have an impact on GDP and national income.
In other words, it’s a measure of value attributed by statisticians to services banks and other financial intermediators provide to households in pounds and pence.
Digging through the methodology issued by the European System of Accounts (ESA 2010), we can see that FISIM is calculated by taking an estimate of the interbank refinancing rate and applying it first to household deposits, to arrive at a measure of utility that households derive from their savings.
We infer the logic behind this move is that household depositors — if they could be bothered — could choose to receive an interest rate equivalent to the interbank refinancing rate rather than the derisory rate offered by their bank.
The fact that households don’t opt to receive interbank refinancing rate on their deposits, and opt instead to receive their actual deposit rate, means that they must really love paying implicit bank charges. Right? And how much value do households receive from these implicit bank charges? Whatever the implicit bank charge happens to be. Obviously.
The ONS’s Blue Book of national economic accounts’ massive spreadsheet actually lays this all out — if you know where to look. At the macro level, Table 1.7.3 (allocation of primary income account, resources and uses) shows that household interest resources are greater than household interest uses. But 124 spreadsheet tabs later, Table 6.1.3 shows that the summary data used earlier is the sum of cash interest (aka interest) and FISIM adjustments (aka households’ love of banking). Here’s the quarterly chart for interest received, split into these two components:
As you can see, financial institutions are pocketing huge amounts of money UK household depositors really love their financial intermediators.
How about borrowers? They’re the other side of the same coin.
Just like the calculation around FISIM on deposits, FISIM on loans roughly assumes that the interest rate households could pay is the interbank refinancing rate. The amount they pay above and beyond this is the FISIM component. So while depositors are sort of assumed to pay FISIM out of the interest they could’ve received, borrowers pay interest with hard cash. FISIM is the portion of this hard cash that statisticians reckon they paid for financial services.
The eagle-eyed among you will see that FISIM on loans has been negative for the last six quarters. No — this isn’t a statistician’s guess as to the level of households’ newfound collective disdain towards their lenders. It looks to us like this is the ONS representing those lucky people who locked in fixed rate mortgage deals before interbank refinancing rates moved higher. And as fixed-rate deals roll off (and the yield curve steepens), we’d expect FISIM on loans to turn back positive.
Two questions remain. What does this do to net cash interest paid to/from the household sector? And is all this accounting as stupid as we’ve just made it sound?
Taking only the cash components of interest paid and interest received, we can no longer say that UK households are net recipients of interest. We can say that rising rates have improved the household sector’s net cash interest position rather than worsened it, for all the reasons we spoke about in the original post. And that — given the household sector has a comparable quantum of deposits and loans — interest rate changes are increasingly distributional within the sector.
Moving on to the question about whether all this accounting makes sense in human, despite our flippancy, we sort of think that adding a made-up number to household cash interest received … is not really as stupid as how we’ve portrayed it?
Think of it like this. You, a household, have £1,000 on deposit. Your bank pays you an interest rate of 2 per cent per annum despite the interbank refinancing rate being 5 per cent per annum. You also get a debit card, an account that can receive and transmit bank transfers, a network of cash points and a (shrinking) number of physical branches.
National economic accountants record that along with your £20 cash interest you also receive £30 worth of non-cash FISIM. Are your debit card, bank account, etc worth £30 per annum? Cheap at the price!
What about people who just want the interest and don’t want to pay for FISIM? They could move the entirety of their deposits into short-dated low-coupon gilts and most likely boost their post-tax cash interest substantially. But they’d probably find modern life more challenging.
Did we think about this before posting? No.
Does this mean that we’ve accidentally red-pilled ourselves by reading too many national accounting methodology documents? Probably.
https://www.ft.com/content/a4312e65-f1f1-4e20-bd29-9f38892bca58
