This article is only available to Macro Hive subscribers. Sign-up to receive world-class macro analysis with a daily curated newsletter, podcast, original content from award-winning researchers, cross market strategy, equity insights, trade ideas, crypto flow frameworks, academic paper summaries, explanation and analysis of market-moving events, community investor chat room, and more.
- The precedent of the 1980s and 1990s suggests there is a slower ‘last mile’ phase of the disinflation process.
- We could be about to enter this last mile – that is, disinflation could be about to slow.
- This time, however, the last mile could still be faster than during the 1990s because of lower inflation inertia and a more efficient economy, even though unemployment is lower.
- Under plausible assumptions, disinflation could slow to about 70bp/year in 2024, compared with a disinflation pace of 1.3ppt/year in 2022-23.
- The Fed expects disinflation to slow this year, which will make it more difficult to ascertain whether disinflation is continuing and makes a March cut unlikely – against markets pricing a 20% risk.
- It will likely take a wait-and-see response to any rise in MoM inflation prints.
Disinflation’s Last Mile
In this note, I look at inflation from the top down and bottom up. Starting with top down, I compare the 1980s and 1990s disinflation with the current one.
The post-oil-shock disinflation went through two phases: a fast phase in 1981-83 and a slower phase during 1989-97 (Chart 1). Between these two phases, inflation recovered in 1986-88, which likely reflected the impact of the 1985 Plaza Accord and dollar weakening on inflation expectations. During 1985-87, the dollar depreciated almost 40% on a trade-weighted basis.
The two phases of disinflation, first fast, then slow, likely inform central bank views that ‘the last mile’ of disinflation could be the hardest, mainly due to base effects and services price stickiness. This raises two questions: when will the last mile start, and how fast will it be?
Faster Last Mile Than 1990s
The fast disinflation phase has been slower this time than after the oil shocks of the 1970s. During 1981-83, core PCE fell by about 1.8 ppt/year. This compares with a slowdown of about 1.3 ppt/year during 2022-23. The slower pace this time likely reflects the lower starting point and smaller base effects: core PCE peaked at 5.5% in early 2022, against a peak of 10.0% in late 1980.
During 1989-1998, the post-oil shocks’ last mile, core PCE slowed by about 40 bp/year. Overall, I expect disinflation to be faster now.
First, inflation has less inertia than in the 1980s, which reflects:
- This time, the inflation surge started after over two decades of low inflation. Consequently, market institutions were not prepared to deal with high inflation. For instance, in the labour market, wage indexation was rare, wage bargaining infrequent and union density low. This repressed wage growth: real wages initially fell and have barely recovered to their pre-pandemic levels.
- The policy response was faster this time. By contrast, core PCE remained above 2% during 1966-80, but the Fed decisively tightened monetary policy only in 1980.
- These factors meant long-term inflation expectations have remained stable, unlike in the 1970s (Chart 2).
Second, the US economy is more open and efficient now than in the 1970s (Chart 3).
Against these favourable factors, unemployment is lower now than during the 1990s (Chart 4). I do not think this will be a major factor preventing disinflation. First, the US economy is not overheating, as shown by the disinflation so far and the decline in the current account deficit. These reflect the positive supply shocks benefitting the US economy, mainly lower energy prices, higher labour supply and supply-chain improvements.
Second, disinflation is essentially about lowering inflation expectations. With enough credibility, a central bank can do it without triggering a recession – a feat the Fed seems to have pulled off. In 2019, core PCE was 1.5% despite unemployment below 4%.
Third, the 1990s unemployment spike reflected the 1990-91 recession. That was not triggered by the Fed to keep disinflation on track. Rather, the recession was caused by factors such as the oil price shock, S&L crisis, and tighter fiscal policy. The Fed actually cut rates from mid-1989.
These factors suggest the last mile could be faster now than during the 1990s, for instance about 70 bp/year.
In addition, historically core inflation has tended to be either high or below 2%. This is consistent with the view that inflation dynamics are driven by self-reinforcing regimes. Once inflation is in a low regime, it tends to stay there. This creates a risk that once inflation nears 2%, it could get pulled down to the low levels prevailing before the pandemic.
The biggest risk to this view could be an oil price shock. A spike in oil prices could rekindle inflation expectations and reverse the disinflation progress. The correlation between energy and core inflation is weakening but remains more elevated than before the pandemic (Chart 5). However, Viresh expects stable to somewhat rising prices this year, rather than a price spike.
And conversely, a marked decline in oil prices could accelerate disinflation and prove the last mile doubters right.
Last Mile May Have Already Started
In 1989, the last disinflation mile started when inflation had fallen to 4.7%, 5.1ppt below the 9.8% peak – or two thirds of the way back to 2%. Core PCE at 2.9% is currently about three quarter of the way back to 2% from a 5.6% peak.
Perhaps a stronger reason to expect the last mile to have started is the recent pickup in wage inflation. This reflects the lagged transmission of strong growth to wages. It need not lead to faster inflation because it has been accompanied by increased productivity. But it suggests disinflation could slow ahead.
I now turn to a bottom-up inflation scenario. Core PCE slowing to about 2.2% in 2024 is consistent with recent inflation trends. A simple scenario shows how (Chart 6; Table 1).
My key assumptions are:
Goods prices stay flat until end-2024 (Table 1). Contrary to my expectations, goods prices fell last year. This suggests businesses have circumvented residual global supply-chain disruptions or that global suppliers, especially China, are facing weak demand.
Going forward, I do not expect global supply chains to worsen much. In addition, the US has limited exposure to Red Sea traffic disruptions since most of its trade is with Canada and Mexico or trans-Pacific or trans-Atlantic.
My assumptions are consistent with recent trends in import prices and PCE (Chart 7).
Housing Inflation: in line with the new tenant index showing a downward trend, I expect housing inflation to slow through 2024.
Core services ex-housing (super core): I expect MoM to pick up in the first quarter due to a limited passthrough from higher wage growth, and to slow over the reminder of the year.
A pickup in MoM core inflation prints in the first quarter relative to 2023 H2 seems likely as the US enters the last disinflation mile. It would not however signal the end of disinflation. In addition, because inflation expectations are adjusting down, continued low unemployment and high growth would not signal the end of disinflation.
The Fed expects slower disinflation this year and therefore is likely to respond to any higher Q1 PCE prints with a wait-and-see attitude. At the same time, it will want to see evidence that disinflation is continuing before cutting and is therefore unlikely to start cutting in March.
Dominique Dwor-Frecaut is a macro strategist based in Southern California. She has worked on EM and DMs at hedge funds, on the sell side, the NY Fed , the IMF and the World Bank. She publishes the blog Macro Sis that discusses the drivers of macro returns.