Friday CPI Preview: Inflation Risks Shift, Will the Fed Hike in September?

TL;DR

· A former Fed economist has shifted her policy stance from "hold rates steady" to "hike," arguing the Fed should raise rates by 25 basis points in September, with cumulative hikes of 50 to 75 basis points by year-end.

· Current inflation data still supports holding steady, but disinflation progress is limited. July core PCE rose 0.2% month-over-month, an annualized rate of about 3%, still far from the 2% target.

· Sahm believes the recent three-month inflation slowdown may be influenced by seasonal factors and does not necessarily indicate a clear improvement in underlying inflation trends.

· If the Middle East conflict keeps gasoline and diesel prices elevated for a prolonged period, cost pressures may gradually spread from energy to transportation and other core goods and services.

· Trade frictions between the U.S. and Canada suggest the tariff hike cycle may not be over, and goods inflation could face renewed pressure.

· AI infrastructure investment may push up prices of key components like memory chips in the short term, with inflationary characteristics different from one-off energy or tariff shocks.

· A rate hike is not a complete reversal of Sahm's baseline inflation scenario, but rather a risk management move: using moderate tightening to hedge against the possibility of inflation staying above target over the next year.

Editor's note: The dilemma facing the Fed's September meeting is not just whether the latest CPI report runs hot or cold.

Recent inflation data has eased from earlier in the year, but the pace of decline remains slow. Meanwhile, the Middle East situation, trade frictions between the U.S. and Canada, and chip demand driven by AI infrastructure investment are creating new price pressures.

Economist Claudia Sahm, creator of the "Sahm rule," has therefore changed her previous stance of supporting a rate hold. She believes that based on current data alone, the Fed could still barely justify a wait-and-see approach, but monetary policy needs to address future risks, not just explain inflation that has already occurred. Persistently high energy prices could feed into core inflation, new tariffs could interrupt goods disinflation, and AI capital spending could create more durable demand pressure.

She still views inflation falling back as the baseline scenario and acknowledges that whether to hike is a close to 50-50 call. But without confidence that PCE inflation will return to 2% on its own within the next year or two, she argues the Fed should use a small rate hike as "insurance."

The following is a translated compilation of the original:

Why I shifted from "hold steady" to supporting a rate hike

Hike or hold? That is the question the Fed must answer at next week's meeting. What is certain is that Fed officials are already divided, and that division likely cannot be resolved by a single CPI report.

Recently, I adjusted my preferred policy choice from "hold rates steady" to "hike." The reason is that I can no longer be confident that PCE inflation will fall back to 2% within the next year or two without further rate increases.

More importantly, upside risks to inflation have been increasing since July. I believe the Fed could start with a 25-basis-point hike in September and cumulatively raise rates by 50 to 75 basis points by year-end to ensure inflation moves down in a timely and sustained manner.

This does not mean current inflation data has clearly deteriorated. On the contrary, recent data has been slightly positive. What really tilts the risk balance toward hiking is the unresolved Middle East situation, the trade war between the U.S. and Canada, and chip shortages driven by AI buildout.

If looking only at existing inflation data, the Fed could still choose to hold rates steady, but the justification is no longer sufficient.

As of July, both headline and core PCE inflation have fallen from their early-year highs, which can be seen as evidence that disinflation is continuing, but the improvement is limited. The 12-month inflation rate has declined slowly, and this week's PPI and CPI releases will be key inputs for the August PCE data. Markets widely expect further moderation, but the distance to the 2% target remains large.

I am cautious about the positive signals from the recent three-month inflation changes. Over the past few years, PCE inflation has shown some seasonality: higher early in the year, then gradually easing. July core PCE rose 0.2% month-over-month, an annualized rate of about 3%. That pace is lower than the first half but still clearly above the 2% target.

Therefore, recent data looks more like absorbing the abnormal early-year heat and returning to last year's already elevated pace, rather than proving a substantive improvement in underlying inflation trends.

The seasonal pattern of core PCE inflation—higher early in the year, easing mid-year—means the cooling signal from recent three-month data may be overstated

Temporary shocks are becoming more persistent

Given limited disinflation progress and a still-solid labor market, why has the Fed kept rates unchanged for so long? And why did I support that decision until just a few weeks ago?

The key lies in the type of shocks driving inflation.

Last year's sharp tariff increases pushed up goods prices. But once higher import costs are fully reflected in final prices, the incremental impact of tariffs on inflation typically fades. The same applies to Middle East energy supply disruptions: energy prices spike quickly, and once the conflict ends, prices may fall back and pull inflation lower.

Such shocks tend to cause one-time price level increases, not necessarily sustained inflation. If the Fed responds with rate hikes, it risks excessively weakening demand to suppress price pressures that would naturally fade.

Previous data largely supported this judgment, but future risks have changed.

Risk 1: Middle East conflict could transmit energy inflation to core prices

A rise in energy prices alone is not enough to justify a Fed rate hike. But if energy prices stay elevated for a long time and gradually feed into non-energy goods and services, monetary policy may need to respond.

Historical experience shows that after energy prices rise, airfares usually increase quickly; other core prices respond less and more slowly, with pass-through effects possibly peaking a year or more later. A 10% increase in gasoline prices adds roughly 0.2 percentage points to core inflation over the following year.

The pass-through of energy price increases to core inflation is slow, with effects possibly peaking a year or more later

This impact may seem limited, but the longer energy prices stay high, the more pronounced the cumulative push on inflation becomes.

When the Middle East conflict began, assuming a quick resolution as the baseline was reasonable, but that assumption is becoming harder to maintain. If gasoline prices remain at current levels, the impact on core inflation could extend into next year.

Diesel prices deserve particular attention. As a key cost in transportation and logistics, diesel increases affect a wide range of goods and services. The lack of progress in restoring passage through the Strait of Hormuz means core inflation still faces further upside risk.

Risk 2: Tariff hikes may not be over

The Fed's July meeting minutes showed most officials believe the impact of tariffs on inflation has largely peaked and will gradually fade.

This trend is already visible in the data. After last year's tariffs took effect, the three-month annualized increase in core goods prices rose rapidly, peaked early this year, then clearly eased in the summer.

After tariffs were implemented, U.S. core goods inflation rose notably, and though it has recently eased, further cooling depends on no additional tariff increases

But continued goods disinflation requires that tariff rates do not rise further.

Currently, the scale of Canadian imports subject to the additional 50% U.S. tariff is relatively limited, but Canada's retaliatory measures against U.S. goods could trigger further U.S. tariff increases. More importantly, this dispute shows the U.S. government is still using tariffs as a negotiating tool.

Therefore, compared with the last Fed meeting, the risk of tariffs re-igniting inflation has increased.

Risk 3: AI investment could also push up inflation in the short term

AI infrastructure buildout is another easily underestimated source of inflation. Nvidia's recent earnings show AI demand remains strong, with large cloud service providers' capital spending expected to exceed $1 trillion next year.

From a five- or ten-year perspective, AI could have deflationary effects after boosting productivity. But over the next year—the timeframe more relevant for monetary policy—the impact of AI investment on inflation is more likely to be upward.

Consumer and business investment price data already show memory chip prices rising. The share of such spending in the overall economy may be limited, but it still adds to upside inflation risk.

AI infrastructure investment is driving up demand for memory chips, and prices in related consumer and business investment categories are already showing signs of increase

The price pressure from AI buildout is fundamentally a demand-driven factor. For energy and tariff shocks, the Fed can choose to look through them because their effects may fade on their own; but the memory chip shortage is backed by sustained expanding investment demand, and the same logic may not apply.

A rate hike is insurance against future risks

Taking everything together, my baseline scenario remains that inflation continues to fall. However, upside risks to inflation are now substantial and spread across energy, goods, transportation, and tech investment.

From a risk management perspective, the Fed has reason to tighten monetary policy further. Raising rates during a disinflation phase may sound contradictory, but the policy goal is not just to get inflation to 2% eventually—it is also to ensure inflation declines credibly and sustainably over the next year.

A moderate rate hike now is like buying insurance against rising inflation risks.

Of course, holding rates steady also has reasonable grounds, especially if policymakers place more weight on current data than on forecasts and tail risks. If this week's inflation data improves markedly, or if the above risks ease, I could shift back to supporting a hold. Monetary policy judgments must be continuously adjusted as information changes.

Since 2024, I have participated in the "Shadow Economic Forecast Summary" organized by Duke University. In March this year, I still expected the Fed to cut rates within the year; by June, I shifted to supporting a hold; now, I believe two hikes may be needed this year, and the rate path over the next few years will also be higher.

This does not mean the Fed will definitely hike. For officials who previously thought holding steady was more appropriate, they would need to change their stance like I did to form a majority in favor of hiking. This remains a very close call.

The market needs more than just a decision

Whether to hike or hold at the September meeting will be a difficult decision. Current inflation data is still enough to support a wait-and-see approach, but the inflation outlook has deteriorated due to mounting upside risks. On that basis, I believe a slight increase in the federal funds rate is more appropriate.

Whatever the Fed ultimately decides, the market needs a clear explanation.

It is not terrible that the outcome is uncertain before the FOMC meeting; what is truly unacceptable is that after the press conference, the market still does not know why the Fed made its decision.

If most officials still believe inflation will return to 2% soon, they need to explain where that confidence comes from; if they no longer believe it, then the Fed should act to rebuild credibility that inflation can return to target.