政策共识反映了对前景的不确定性

Restricted - External Investment SciencesRenate Marold(v)+1 212 526 2484renate.marold@barclays.comBCI, USEconomics ResearchMarc Giannoni(ii)+1 212 526 9373marc.giannoni@barclays.comBCI, USJonathan Millar(ii)+1 212 526 4876jonathan.millar@barclays.comBCI, US FIGURE 1. Participants showed elevated agreement regarding monetary policy, inflation and thelabor market, with muted agreement on the economic outlookNote: The "other" category includes fiscal policy, business conditions, credit conditions, financial market developments,real estate, and consumption.Source: Federal Reserve, Barclays ResearchThe agreement indicators point to the following:•Comparing the number of discussion items across topics, it appears that the focus of theMay minutesshiftednotably relative to recent norms. Discussion about miscellaneoustopics and the economic outlook were more numerous, with fewer items related to theoutlook for monetary policy and the labor market.We suspect that agreement indicatorsfrom the May minutes may be a less precise gauge of participants' level of agreement acrosstopics, as an increased share of discussion items did not receive an explicit quantifier.•FOMC participants may have turned even more aligned behind the "no rush" approach.Ourmonetary policy agreement indicator derived from the May minutes rose further, fromalready-elevated levels in the previous two meetings.•Although the minutes are now placing increased focus on the economic outlook,none of theeconomic outlook discussion statements garnered broad support, resulting in aneconomic outlook agreement indicator that is well below historical averages.•We think deterioration in the indicator of agreement about "other" topics was pulled downby a lengthier-than-normal discussion of topics pertaining to financial markets and businessactivity. These discussion items included reports and surveys indicating firms are planning topartially or fully pass ontariff-relatedcosts, and that they are pausing or delaying theircapital expenditure plans, all of which tended to garner only minority support.•Inflation agreement in the May minutes recovered from muted levels at the twomeetings earlier this year(January and March), with fewer discussion statements thatfound only narrow support. May "agreement" on inflation was mainly related to the risk thatinflation could prove to be more persistent than expected, whereas agreement in the Marchminutes was mostly about developments in near- and longer-term inflation expectations.We view these NLP results as broadly consistent with the interpretation of the May minutes inFederal Reserve Commentary | May FOMC minutes: FAIT no more, May 29, 2025, which inferredthat participants are in no rush to adjust policy rates.The US economics team expects one25bp rate cut this year, in December, in the face oftariff-ledslowing in economic growth, aslightly higher unemployment rate and rising inflation.They then expect the FOMC to cutrates three times in 2026, placing the policy rate at what it regards as a slightly restrictive level 2 2This assumption is reflected in the calculation of the agreement indicator, which weighs non-quantified discussionstatements at 80%, versus 70% and 100%, respectively, for statements that receive explicit medium support (eg, "manyparticipants") or broad support (eg, "almost all" or "all" participants).by year-end, but with the rate path viewed as unusually uncertain, amid elevated uncertaintyabout other policies.Focal pointsshiftedfrom the March minutes followingtariffdevelopmentsBased on the count of discussion items across topics, the focus in the May minutesshiftednotably relative to the March minutes and prior years' averages. There was a lengthierdiscussion on "other" topics and the economic outlook, and the discussion was lessheavily concentrated on the outlook for monetary policy and the labor market.Figure 2shows the number of discussion items in the minutes across topics, counting both quantifiedand unquantified discussion statements. The March minutes included 11 discussions of "other"topics, which increased to 24 in the May minutes. The uptick in economic-outlook-relateddiscussions stands out even more, with the number of discussion items rising from 15 to 19,which is double that of the typical minutes in 2023-24.FIGURE 2. The number of discussion items jumped for miscellaneous topics and the economicoutlook, while there was relatively little focus on the outlook for monetary policy and the labormarketSource: Federal Reserve, Barclays ResearchOur sense is that the agreement indicators from the May minutes may be a less precisegauge of agreement on topics, as more of the discussion items did not receive an explicitquantifieron the share of the committee that aligns itself with respective statements. Figure 3plots the percentage of sentences that did not contain a quantifier that helps us bucket thestatement as "broad agreement", "medium agreement" or "low agreement", even though wepresume that