Wow: the October personal saving rate was the second lowest ever recorded (data goes back to 1959).
The two-month moving average was actually the lowest ever recorded and the three-month moving average was the third lowest.
The low and fall saving rate is supporting a huge divergence between stagnant real incomes and steadily rising consumption.
Relative to CBO's pre-pandemic forecast* real per capita:
Disposable personal income: -5%
Personal consumption expenditures: +2.5%
*CBO does not forecast these exact variables but has ones that are close (e.g., personal income instead of disposable personal income). I think my versions of their forecast are reasonably robust--you can see the levels and approximate CBO forecasts here.
The personal saving rate is typically about 7.5%.
During the pandemic it was above 10% for the 15 straights months--with the entire period accumulating $2.2T in excess savings.
It has been below average for 13 straight months, $800b in excess dissavings.
You can also think about this as the lagged impact of fiscal policy and pandemic-reduced consumption. The initial fiscal multipliers were relatively low as people saved money. But fiscal policy has long and variable lags and it is supporting consumption (and inflation) now.
This story is likely disproportionately about higher income households (but we can't be sure because BEA does not produce real-time distributional data).
One hint is that real compensation per capita, which matters more for middle-class households, is only slightly below trend.
Overall this is consistent with the excess saving story (the household budget constraint being expanded) & the pent up demand story (the marginal utility of spending being expanded as the pandemic eased).
Also of fiscal policy mattering--could debate if for better or worse.
Will we have a Wile E Coyote moment when consumers realize there is nothing below them and consumption plummets? Could happen--although these data plus more direct indicators of household balance sheets and financial distress suggest that moment could be 6-12 months away.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Core PCE inflation came in just as expected. It has been very tame for the last three months--but shouldn't think of them in isolation but as part of a noisy process where inflation was much higher before.
And in big inflation news, the CPI-based Ecumenical Underlying Inflation measure was exactly 2.0% in May, consistent with the Fed's target. This is the first time it has been there since I started this concept during the inflationary episode.
The ecumenical measure takes the median of 21 different measures: 7 different concepts (e.g., with and without housing) over 3, 6 and 12 months--all re-meaned to match the PCE inflation that the Fed targets.
In practice it is very similar to 6-month core CPI (re-meaned).
I didn't share the basic data earlier. Here is core CPI, came in well below expectations in May.
A boring jobs report, in a good way. 139K jobs added (140K private). Unemployment rate unchanged at 4.2%. Hours unchanged. Only notable deviations from steady state were participation down and unusual wage growth up.
Note, Federal employment continued to decline. But state and local added almost as much.
Strong jobs report. 177K jobs added. Unemployment rate steady at 4.2% but participation rate up and U-6 down. Hours steady. A slowdown in hourly wage growth.
Federal employment was down a bit but state and local more than made up for it. The trend in private jobs is basically the same as total.
Unemployment rate very slowly drifted up for the last year and a half.
Wednesday's Q1 GDP # will have a lot of economic noise, a lot of measurement noise, and could generate even more political noise.
A technical🧵on one aspect: what period does it reflect?
The answer is a combo of pre- and post- 1/20 because of the weirdness of quarterly averages
When I (and most people) look at things like CPI or jobs, we look at something like a three month average. That would be growth from Dec 2024 to Mar 2025. Which is also the (geometric) average of the growth rates in Jan, Feb and Mar. It tells you what happened in those 3 months.
But GDP is not reported monthly (fortunately, would be really volatile). So the numbers are the growth from the average of Oct/Nov/Dec to Jan/Feb/Mar. If there is weak growth in Nov or Dec that lowers part of Q4 but all of Q1 so lowers overall growth.