March CPI & Retail Sales

March core inflation was weaker than thought, but trimmed mean measures remain firm. Real retail sales were solid in March due to auto and bar/restaurant sales.

This post reviews trends from the most notable data releases this and last week: the March CPI and retail sales reports.

Key takeaways:

• Core CPI inflation was weaker than expected in March but trimmed mean CPI measures nonetheless continue to suggest that in terms of underlying PCE inflation, trends remain sticky above the Fed’s inflation target.

• Compared to underlying inflation trends in the euro area and Japan, U.S. underlying inflation continues to run well above the Fed’s inflation target. In particular the ECB thus has more scope to turn its policy stance more accommodative in response to any economic fallout of forthcoming U.S. tariff hikes, which could temper further dollar weakening in the near term.

• Inflation-adjusted retail sales growth accelerated over the month for both core goods and bar/restaurants but contracted for core goods excl. motor vehicles in March. Households goods spending has been dominated by expectations regarding forthcoming tariff hikes, especially for durable goods such as cars.

• The recent data suggest the Fed will remain on hold for a while given underlying inflation trends.

March CPI: A Mixed Bag

The topline numbers of the March CPI report were below consensus expectations: headline CPI fell 0.1% (forecast: +0.1%) in March after rising 0.2% in February, while core CPI rose +0.1% m/m (three-digits: +0.057 vs. +0.227 in February), missing the +0.2% forecast.

To gauge underlying CPI inflation, one can refer to the Cleveland Fed's trimmed mean CPI measures, which exclude extreme volatility by either taking the median (price change at the 50th percentile) or a 16% trimmed mean (weighted average after cutting the top and bottom 8% of price changes). These trimmed mean CPI measures provide a more nuanced picture than the topline numbers: Median CPI inflation accelerated from +0.29% m/m to +0.34% in March, while the 16% Trimmed Mean CPI eased to +0.21% from +0.27%. For both metrics (i.e., underlying inflation), Q1’s average monthly increase was roughly in line with Q4 2024’s—around +0.3%. Monthly core CPI inflation averaged +0.2% in Q1. So, trimmed mean inflation remains stickier than core CPI suggests.

As usual, a major factor in CPI dynamics was the OER component, which firmed from +0.3% m/m in February to +0.4%. Year/year OER rates remain stable above pre-COVID levels in the Midwest, Northeast, and South, with only the West showing below pre-COVID levels (chart above). It remains to be seen if a nationwide return to pre-COVID trends is in store. Also note that while housing services drive ~40% of core CPI, they account for under 20% of core PCE—the Fed’s preferred price gauge. Hence, disinflation must be broader based, and the trimmed mean metrics suggest this may not be the case.

Both Median and 16% Trimmed Mean CPI inflation exceed the Fed’s 2% target over a six-month horizon (at 3% and 3.2% in core PCE terms, respectively). The 6-month average was steady for the Median but eased for the Trimmed Mean (chart above).

When using the strong correlation between the CPI and PCE trimmed mean inflation series, statistical nowcasts of Median and Trimmed Mean PCE inflation rates for March (due later this month) show underlying PCE inflation likely stayed elevated within the 2.6%-3% range (diamonds in the above chart).

These nowcasts underscore that inflation trends remain well above the Fed's 2% target. Given signals from this CPI report, a solid March jobs report, and expected core inflation acceleration from tariff hikes, the Fed will likely stay on hold most of the year.

It is worthwhile to compare the U.S. underlying inflation trends with those in the two other main advanced economy blocks: the euro area and Japan. The chart above contrasts year/year changes in the U.S. Median and 16% Trimmed Mean CPI with those for the euro area’s Median and 15% Trimmed Mean HICP, measured as deviations from 2% core inflation using each central bank’s preferred index. It’s clear that since early 2024 underlying inflation trends in the euro have been running essentially at the ECB’s 2% HICP inflation target, in sharp contrast to the U.S. where it remained well above the Fed’s 2% PCE inflation target.

The next chart above reports on a similar U.S. — Japan comparison using the Bank of Japan’s Median and 20% Trimmed Mean CPI. Here we notice that while Japanese underlying inflation had been easing back to and somewhat below the Bank of Japan’s 2% CPI inflation target throughout the first half, these trends have reversed since mid-2024. Currently, Japanese underlying inflation sits at least at target and likely above it, with further upward momentum expected.

It’s clear the Fed is facing a big dilemma in that the forthcoming U.S. tariff hikes are stagflationary in nature for the U.S. economy, with inflation trends at the same time remaining sticky above its 2% PCE inflation target. IMO this will severely hamper the Fed’s response to any potential economic slowdown and instead will likely remain on hold for a while. Fed chair Powell’s most recent remarks underscored this, highlighting the challenge of balancing the Fed’s dual mandate.

In the case of the euro area and Japan, the U.S. tariff hikes could be disinflationary demand shocks provided European and Japanese retaliation remains limited. Inflation trends in the euro area are near the ECB’s inflation target and policy rates are closing in on its own perceived neutral rate range of 1.75%-2.25%. The ECB thus has room for a more decisive shift in its policy stance towards accommodation, and there likely will be a hint at that in its March post-meeting statement. For the Bank of Japan, the story is more complex, as the recent firming of Japanese underlying inflation above target likely means that the BoJ would want to continue to be careful in moving its policy rate given the potentially disinflationary external risks.

The dollar has been losing value relative to the euro and the Japanese yen recently, as investors’ confidence in the safety of U.S. assets deteriorated, and this likely is part of a more medium-term trend. But over a more immediate horizon the above discussion of underlying inflation comparisons suggests the euro might become more range bound relative to the U.S. dollar in the near-term, muting any further dollar weakening vs. that currency.

March Retail Sales: Cautiously Resilient

Retail trade (i.e. retail sales pertaining to goods) grew 1.4% over the month in March, after it increased 0.4% month/month in February. As I always emphasize, it's crucial to remember to not take these figures at face value without looking under the hood:

• Retail sales measures spending on goods as well as bar/restaurants spending. Thus, it really mostly measures goods consumption which is a relatively small slice of the monthly consumption basket, as about 2/3 of U.S. consumption expenditures relates to services.

• Retail sales data does not correct for changes in prices, which for goods in particular can make a big impact: core goods prices in the March CPI report dropped 0.1% month/month. As such weaker retail sales growth, for example, could merely reflect more the pace of price increases while retail sales volumes were less downbeat.

• Deflating retail sales with CPI or core CPI overlooks the predominantly goods-focused nature of retail spending.

Dissecting retail trade data into subcomponents and aligning them with corresponding CPI subcomponents provides a more accurate assessment. Firstly, when it comes to retail trade data (that is, retail sales minus nominal bar and restaurant spending) I use the CPI Commodities, CPI Gasoline, CPI New & Used Vehicles and CPI Motor Vehicle Parts & Equipment indices to inflation adjust over retail trade as well as components related to sales at gasoline stations and motor vehicle dealers & parts. In case of bar and restaurant spending I deflate that component by means of the CPI Food Away from Home index.

Furthermore, I apply chain-weighting based on the Fisher index approach using current period and previous period prices and quantities of retail trade and the gas and motor vehicle components to parse out the impact of the latter two volatile components. This approach allows for time-varying weights, as prices and quantities change from period to period and consumers substitute between the different spending categories. It is the same methodology used by the BEA to compute real consumption and GDP.

Real core goods spending (inflation-adjusted retail sales excl. gas stations) went up ever so slightly in March at 0.7% on a 3-month annualized rate (AR) basis, after declining in January and February, as is evident from the chart above. Real core goods sales excluding motor vehicles grew at a 2.9% three-month AR pace, a slowdown from the 11.9% increase in February (orange line in the chart above). Finally, inflation-adjusted bar and restaurant spending grew 1.7% 3-month AR in inflation-adjusted terms in March, after declining in the preceding three months (purple line in the chart above).

The chart above indicates that while the three-month growth pace of inflation-adjusted core goods spending was tepid in March, on a monthly basis real goods sales excl. gas stations accelerated notably in March to +21.5% AR after increasing 3.7% in the preceding month (blue line). We have pretty much a mirror-image move for monthly rates of real core goods spending excl. motor vehicles (orange line), which had been accelerating for three consecutive months between December and February (+3.5%, +8% and +25.4% month/month AR respectively), but declined a notable -19.4% month/month AR.

In terms of inflation-adjusted bar and restaurant spending, this recovered significantly in March after a marked slowdown in the month/month pace since November (purple line in the above chart): +18.2% month/month AR in March vs -13.8% in February, +3.2% in January and -11% in December. This dynamic seems somewhat at odds with really downbeat readings from a range of consumer sentiment surveys since Q4 that reflected increased policy uncertainty owing in particular trade policy actions by the Trump 2.0 administration.

The above suggests that real goods spending was volatile throughout Q4 and Q1, which was mainly due to volatile inflation-adjusted motor vehicle sales. Comparing inflation-adjusted auto dealer sales to a measure of trade policy uncertainty in the chart above we notice that since September we’ve seen large swings in real auto sales line up with historically high levels of trade policy uncertainty. So, the prospect of forthcoming tariffs and their impact on future car (and other durable goods) prices have led to waves of elevated car purchases at the end of Q4 and in March in an attempt to get ahead of the expected price hikes, while at the same time households clamp down on non-durables spending.

For Q1 consumption spending in the forthcoming Q1 GDP report, the relevant metrics are real core goods excl. motor vehicle dealer and maintenance spending as well as real bar/restaurant spending. Both measures suggest a very modest but still positive pace for Q1 consumption growth.

While underlying trends in the March CPI report suggest sticky, above-target inflation trends, real spending remained solid in March as households ramp up durable goods spending ahead of forthcoming tariff hikes. Combined with a still solid labor market this means the Fed will remain cautious about further cuts (if any) in 2025.