Excess savings are back, maybe

Receive free Global Economy updates

We’ll send you a myFT Daily Digest email rounding up the latest Global Economy news every morning.

The role of excess savings is one of the biggest economic stories of the pandemic (and post-pandemic) era.

There’s little dispute that widespread lockdowns led households to save far more than they might otherwise have done — what’s unclear is exactly how much they saved, how much is left, and what that means for the economy.

Our mainFT colleague Soumaya Keynes took a look at the different trends in a column last week, concluding:

It also seems unlikely that there is much pent-up demand yet to be unleashed. If the grandparents wanted to go on a cruise, they probably would have gone by now. Perhaps the stash might be used to cushion spending in case a recession strikes. But the risk of a surge is ebbing.

We broadly agree, although given our passion for the acute narrative instability of recent national statistics releases, Alphaville wanted to quickly dwell on last week’s updated US household savings figures.

Updated Bureau of Economic Analysis national accounts figures had the interesting statistical effect of suggesting US excess savings rates may be higher than previously thought — by lowering the bar.

To recap quickly, excess savings have typically been figured as the difference been the amount of savings since the pandemic began, versus the projected level of saving before the pandemic.

Research released by the San Francisco Fed in August put the total amount accumulated above trend at $2.1tn, with $1.9tn of that drawn down since savings fell below trend last year:

Authors Hamza Abdelrahman and Luiz E. Oliveira reckoned this left about $190bn of excess savings — suggesting further fuel for inflation or growth was very limited. (This was a fairly pessimistic estimate: JPMorgan had it at around $400bn, while Goldman calculated in July that there was $1.3tn left. Fitch Ratings put it at $900bn in June, but said it was dropping by $73bn a month.)

Whatever the reality is, it clearly matters, because the point at which these savings dry up might be the point at which consumers capitulate on spending and the Federal Reserve decides to call it a day on tightening.

Friday’s data may have changed the picture somewhat. The Bureau of Economic Analysis’s new figures suggest monthly saving had been lower than previously thought for a long time before the pandemic began. As a result, the projected level of saving is lower, and the actual course looks even more elevated. Here are the monthly figures, as blocks with pretty lines:

Or, as a spread:

It’s somewhat counter-intuitive to read this as bullish for the US consumer, as it indicates they’ve been saving less than thought for a long time. But remember that because excess savings is about reality measured against an extrapolated past trend, a sufficiently big historic undershoot can flatter recent figures (even if those recent figures also fell).

The theoretical effects are pretty substantial. JPMorgan, responding the new data said:

The combination of a higher recent saving rate and a lower pre-pandemic benchmark for saving suggests that there is a lot more “excess saving” left over for households than we had seen in the data before the revisions. It now appears that the leftover excess saving stood at around $1.2tr in 2Q, notably higher than the $0.4tr figure estimated based on pre-revision data (these estimates are derived by extrapolating pre-pandemic trend growth rates forward starting in 2020).

An extra $800bn, down the back of the sofa.

Clearly, excess savings are an extremely approximate measure. The SF Fed projected expected savings from 48 months ahead of when the US’s recession began in 2020. Others use substantially different measures — a calculation by TS Lombard’s Steven Blitz based on cumulative savings from 2012—2019 is particularly eye-catching:

We tried to recreate the SF Fed’s figures as best we could (ie very crudely). The US’s Covid-19 recession is considered to have lasted from February to April 2020, so we projected from the 48 months up to January 2020.

Based on that, here’s (roughly) how the first chart we shared above would look with the new numbers:

Crunching these (crude) numbers, we come up with a figure of $690bn of savings being left, based on an accumulation of about $2.6tn and a drawdown of about $1.9bn — ie nearly four times the SF Fed’s estimate, and a radically different amount of pent-up purchasing power, all because the old run-rate changed. Or, in charts, this:

What’s the tl;dr? Well,

1. national statistics are tricky

2. measuring anything relative to a trend is tricky

3. being part of the FOMC is presumably also quite tricky because you have to be worrying about (1) and (2)

4. ¯\_(ツ)_/¯