另类数据与货币政策

complement to traditional data in informing policy,not as a substitute. The central focus of monetary policy is stabili-zation of the business cycle, so there is a premiumon accurately and quickly assessing turning points.Nontraditional data can be particularly helpful inthese circumstances. That’s because governmentstatistics on key variables like unemployment, infla-tion, and economic growth are published weeks oreven months after the fact. The delay in releasingprivate company data is often substantially shorter,a few weeks or even days. proved useful to the Fed and other decision-mak-ers during the pandemic, which dramatically shiftedconsumer and business behavior. Private companydata on physical mobility was deployed to monitorthose shifts during social distancing, along withadministrative data on the number of COVID cases.Measures of supply-chain stress were also instru-mental in gauging inflationary pressures. In addi-tion to surveys of purchasing managers and ship-ping price indices, the stress on supply chains wasgauged with real-time data on shipping containermovements. To be sure, traditional data sources alsohelped fill gaps in policymakers’ understanding ofthe economy. The Census Bureau, a major sourceof traditional data, quickly stepped into the breach,launching short online surveys to gauge the pandem-ic’s impact on households and small businesses. The timeliness of alternative data sources wasparticularly useful at the start of the pandemic,which triggered a short, deep recession. A reviewby Fed staff noted that its internal weekly estimatesof employment, based on data from ADP, a largepayroll processor, showed large declines in lateMarch 2020. This was more than a month beforethe BLS published its own monthly employmentreport, which also showed large declines. The pandemic downturn was unusuallyfast-moving, but higher-frequency and timelierestimates of employment have broader applica-tions. For example, anytime the monthly BLS esti-mates of jobs shift down sharply, as occurred in2025, the weekly ADP estimates offer early insightinto whether the trend will persist or reverse. Inaddition, ADP estimates are highly relevant duringgovernment shutdowns resulting from a congres-sional impasse over the budget, when official dataisn’t available. Loss of precisionAlternative data can help maintain, and even improve, the quality and cost-effectiveness of tra-ditional statistics. Government agencies rely heav-ily on surveys of people and businesses, which aredesigned to be representative of the overall econ-omy. But these have drawbacks. For one, costshave increased over time as people and businessesbecome less willing to participate. For another, fall-ing participation rates reduce the precision of theresulting estimates. Inflation episodes The granularity of alternative data is anotheradvantage for Fed policymakers seeking to assessthe impact of changes in trade policy on consumerprice inflation. Theory and experience suggest thatan increase in import tariffs will cause a one-timerise in the level of prices, which only temporarilyincreases inflation. In that case, the Fed should “lookthrough” tariff-related inflation and not raise rates.But testing the hypothesis is challenging, becausekey statistics the Fed consults don’t identify prices ofgoods by country of origin. Instead, the analysis mustcompare the prices of broad categories of goods bytheir average share of imports in the past.This is where the granularity of alternative data This loss of precision can create uncertaintyabout inflation or employment dynamics and hin-der a timely, appropriate monetary policy response.Nontraditional data offers a potential solution. Forexample, the BLS now uses private company datainstead of surveys for several components of theconsumer price index, including prices of used cars,airline tickets, and wireless telephone contracts. offers a more direct path to monitoring tariff priceeffects. Alberto Cavallo, a professor at HarvardUniversity, and two collaborators are one such datasource. They have constructed daily price indicesusing online data from five major US retailers, whichinclude country of origin, tariff rates, and sellingprice for 350,000 goods. They find that the prices ofimported consumer goods have risen more quickly Thereisscopeforfurtheruseofprivatesectordata,thoughtheacquisitioncostandreliabilityofsuchdatapresent challenges. A private company could decidetostopsharingitsdataorsharplyraiseitsprice,whichcould threaten the continuity of the government sta-tistics. Careful testing at statistical agencies is alsonecessary to ensure that nontraditional sourcesimprove the precision of estimates rather than sub-stitute new sources of noise for old ones. Business formation outcomes for low earners. Unexpected easing, how-ever, decreased inequality.For all the advantages of nontraditional data Improving the accuracy of the initial estimates oftraditional data is another area where alternativeda