Gender Gaps in the Federal Reserve System

Finance and Economics Discussion Series Federal Reserve Board, Washington, D.C. ISSN 1936-2854 (Print) ISSN 2767-3898 (Online) Gender Gaps in the Federal Reserve System Deepa D. Datta, Nitzan Tzur-Ilan 2024-092 Please cite this paper as: Datta, Deepa D., and Nitzan Tzur-Ilan (2024). “Gender Gaps in the Federal Reserve System,”FinanceandEconomicsDiscussionSeries2024-092. Washington: BoardofGovernors of the Federal Reserve System, https://doi.org/10.17016/FEDS.2024.092. NOTE: Staff working papers in the Finance and Economics Discussion Series (FEDS) are preliminary materials circulated to stimulate discussion and critical comment. The analysis and conclusions set forth are those of the authors and do not indicate concurrence by other members of the research staff or the Board of Governors. References in publications to the Finance and Economics Discussion Series (other than acknowledgement) should be cleared with the author(s) to protect the tentative character of these papers.

Gender Gaps in the Federal Reserve System∗ Deepa D. Datta† Nitzan Tzur-Ilan‡ November7,2024 Abstract Tobetterunderstandthestalledprogressofwomenineconomics,weconstructnewdata onwomen’srepresentationandresearchoutputinoneofthelargestpolicyinstitutions—the Federal Reserve System. We document a slight increase in women’s representation over the past 20 years, in line with academic trends. We also document a significant gender gap in research output, especially for years in which economists have greater domestic responsibilities,butnearlyabsentgendergapsinpolicyoutputandcareerprogression. This workcomplementsexistingresearchonwomeninacademia,allowingamorecomprehensive examinationofprogressintheeconomicsprofession. JELClassification: J16,A14,E58. Keywords: centralbanks,diversity,genderinequality,leakypipeline,researchoutput. ∗First version: June 30, 2023. We thank Julie Hotchkiss, Samuel Kruger, Yichen Su, and the conference participantsattheAFFECT(AcademicFemaleFinanceCommittee)NewIdeasSession,theFederalReserveBoard AppliedMicroSeminar,theWomeninCentralBankingConference,WomeninEconomics(FederalReserveBank ofDallas),theCommitteeontheStatusofWomenintheEconomicsProfession(CSWEP)sessionsattheEastern Economics Association Conference, the System Applied Micro conference, the Annual Meeting of the Southern Economic Association, and the 5th Conference on Diversity, Equity and Inclusion in Economics, Finance, and CentralBankingforhelpfulcommentsanddiscussion. WethankChristianZimmermannandhisteamforgivingus accesstodatafromRePEc(ResearchPapersinEconomics)onworkingpapersandpublicationsofFederalReserve System economists. We thank Andrew Raychawdhuri for excellent research assistance. The views expressed in thispaperaresolelythoseoftheauthorsanddonotnecessarilyreflecttheviewsoftheBoardofGovernorsofthe FederalReserveSystem,theFederalReserveBankofDallas,ortheFederalReserveSystem. †BoardofGovernorsoftheFederalReserveSystem,Email: deepa.d.datta@frb.gov. ‡FederalReserveBankofDallas,Email: nitzan.tzurilan@dal.frb.org.

1 Introduction Substantialrecentevidencesuggeststhatprogresstowardreachinggenderparityintheeconomics profession has stalled (Lundberg and Stearns, 2019; Sherman and Tookes, 2022). Women continue to face systematic barriers, including greater obstacles to publication success (Alexander et al., 2023; Hengel, 2022; Hengel and Moon, 2023; Card et al., 2020), lack of credit for equal work (Sarsons et al., 2021), and rampant harassment and discrimination (Wu, 2018). Much of the evidence on this stalled progress, however, focuses on the academic environment. To understand gender gaps in the economics profession, it is also necessary to explore women’s representationandoutputinpolicyinstitutions. This paper fills a gap in the literature by complementing existing work on academic institutions with evidence from policy institutions. The study of gender gaps in the policy institution environment is important for three key reasons. First, policy institutions can be large relative to academic ones, both in terms of the number of economists and their research output. The FederalReserveSystem,whichcomprisesthe12regionalFederalReserveBanksandtheBoard of Governors of the Federal Reserve System, is one of the largest policy institutions.1 The roughly 1,000 economists in the System at present are comparable to the number of all faculty in the top 20 economics departments (Chari, 2023). The System and other policy institutions are also ranked among top research departments. According to the January 2024 rankings by RePEc, 4 of the top 20 economic institutions are policy institutions, with the Federal Reserve Board, for example, ranked 15th.2 Consequently, policy institutions employ a large share of researcheconomistsandproducealargeshareofeconomicsresearch. Second, the policy institution setting allows us to investigate how women economists may advance in their careers when there exist success measures, such as those related to policy contributionsormanagement,thatareunrelatedtothepeer-reviewandtenurereviewprocesses, 1Weabbreviatetheseinstitutionsas“theSystem,”“theBanks,”and“theBoard,”respectively. 2MoreinformationcanbefoundatRePEc. (n.d.). “Top10%EconomicInstitutions,asofJuly2024.”IDEAS database,webpage,https://ideas.repec.org/top/top.inst.all.html. 1

which have been shown to exhibit some gender bias. Third, separate from being informative about the state of the economics profession, studying women’s representation and output in the Federal Reserve System is important because it is a public institution. It is essential to have diversity among the System’s researchers so that the institution may benefit from a diversity of perspectives,backgrounds,andexperiencesthatmightaffectoveralleconomicperformance.3 In this paper, we study how women’s representation has evolved in the Federal Reserve System and investigate whether there are gender gaps in research and policy output within the System. WefirstdocumentaslightincreaseinthefractionofwomeneconomistsintheSystem over the past 20 years, in line with trends in the economics profession. Second, we document a significant gender gap in research output but find no such gap in policy output among System economists. Our evidence on the complementarities and substitution between research and policy output supplements existing literature from the academic setting investigating gender differences in time allocation between research and activities such as teaching, mentoring, and service. Our results are based on our expansive data set on Federal Reserve System economists, whichwehavenewlyconstructedfromalargenumberofpublicsources. Wecollectnamesand datesofserviceofSystemeconomistsusingthecurrentandhistoricalpublicwebsitesofthe12 Banks and then use existing name dictionaries to gender tag these individuals. We supplement these data with the data constructed by Datta and Vigfusson (2024) on Board economists, also from public sources.4 We use these data to study female representation over time, across institutions, and across cohorts and compare results with analogous measures from academia. To measure research and policy output of System economists, we collect data from RePEc on the titles, authors, and publication years of all System working papers, research publications, 3Other papers that explore gender gaps in central banks also find underrepresentation of women (Azzimonti- Renzo,JarqueandWyckoff,2023). Forinstance,thereisnotasinglewomanonthe30-memberGeneralCouncil oftheEuropeanCentralBank(Hospido,LaevenandLamo,2022). 4Although this paper and Datta and Vigfusson (2024) overlap in their data used, the papers are addressing different topics, with this paper looking at System level paper output and career progression and Datta and Vigfusson(2024)studyingcoauthorrelationshipsattheFederalReserveBoardasameasureofinclusion. 2

and publicly-available policy reports. We use these data to study gender gaps in research and policy output both on a year-by-year basis and accumulated over each individual’s full tenure in the System. Using regression analysis, we also study the potential complementarities and substitutionamongvarioustypesofpolicyandresearchoutput. Using these data, we document that the female share of economists in the System increased only slightly between 2004 and 2021, from 20 to 22 percent. Overall, female representation in theSystemhastrendedinlinewiththatinacademia,showingsomeincreaseinrepresentationin theearly2000sandmorestagnationoflate(Chari,2023;ShermanandTookes,2022). Thatsaid, when comparing across the 12 individual Banks and the Board, we find substantial variation in both the levels and trends in representation. Lastly, although we do find fewer women at more seniorlevelsintheSystem,thedeclineisnotasstarkasinacademia.5 The relative success of women’s advancement within the Federal Reserve System may be relatedtothesuccessweobserveinwomen’srelativepolicyoutput. Wefindalarge,significant, and persistent gender gap in research output and a nearly absent gap in policy output. In our sampleofSystemeconomistsover2004to2021,womenproduce,onaverage,4.4fewerworking papers, 2.1 fewer peer-reviewed publications, and 1.9 fewer System research publications than men. Notably,thoughwomenalsoproduce1.0fewerpolicyreportsthanmen,thisgendergapis notstatisticallysignificant. Thoughthesegendergapsareestimatedoverthefullsample,similar averagegapsareobservablewhenoutputismeasuredonayear-by-yearbasis. We also investigate potential spillovers among various types of output, as well as patterns of specialization into research or policy. We find that individuals with greater research output also tend to have greater policy output. We also see positive spillovers from policy to research, though they seem a bit weaker on average. Overall, the results suggest complementarities between research and policy are strong, and we do not find much evidence for specialization. 5OurfindingsaregenerallyinlinewithAurioletal.(2022)andAzzimonti-Renzo,JarqueandWyckoff(2023). OurfindingofamoregradualdeclineinfemalerepresentationathigherlevelsofsenioritythanAzzimonti-Renzo, Jarque and Wyckoff (2023) can be attributed to our more expansive sample, which includes not just the Federal ReserveBanks,butalsotheFederalReserveBoard. 3

Amongindividualswhohavethesamelevelofpolicyreportoutput,wefindalargegendergapof 3.1workingpapersandasmallbutstatisticallysignificantgapof0.4peer-reviewedpublications. By contrast, among individuals who are equally productive in working paper output, we do not findasignificantgendergapinSystemresearchorpolicyoutput. Lastly, we explore patterns in production of research and policy reports over time, and, specifically, during the COVID-19 pandemic. Previous research has shown that pandemic disruptionsdisproportionatelyaffectedfemaleacademics(Myersetal.,2020;Kruger,Maturana and Nickerson, 2022). We find that policy and research production increased sharply for both menandwomenin2020and,perhapssurprisingly,thatfemaleeconomistsweremoreproductive duringthepandemicthantheirmalecolleagues,especiallyintermsofworkingpaperoutput. Comparedwiththeexistingliterature,wefindsimilargendergapsinresearchoutputamong System economists as have been found in academia (Lundberg and Stearns, 2019; Ceci et al., 2014;GintherandKahn,2004;Ductor,GoyalandPrummer,2023;ShermanandTookes,2022). Many of the previous findings in the literature that might explain this gap in research output also apply to both System and academic economists. For example, Lundberg and Stearns (2019) and Ceci et al. (2014) show that women in STEM (science, technology, engineering, and mathematics) fields, including economics, have fewer publications than men at equivalent stages of their career, though there appears to be no difference in hours worked. In economics, researchers have found that female-authored papers are evaluated more critically and may be held to tougher writing or other standards for publication (Krawczyk and Smyk, 2016; Card etal.,2020;Hengel,2022). Anotherobstacletoresearchoutputforwomenmaybetheirsmaller research networks (Ductor, Goyal and Prummer, 2023; Ductor and Prummer, 2024), which couldbeaconsequenceoframpantdiscriminationagainstwomenthathasbeendocumentedto beprevalentintheprofession(Wu,2018). Additionalfindingsintheliteraturemayalsohelpexplainwhywefindagapinresearchbut notpolicyoutput. First,wenotethatthoughweexpectthatalltheeconomistsinoursampleare salariedemployeeswhoareexpectedtowork40hoursperweek,itislikelythatmanyeconomists 4

work more than 40 hours in total. If economists prioritize meeting institutional requirements forpolicyoutput,andspendtheremainderoftheirtimeonresearch,thenindividualswhowork longer hours may spend more time on research. Previous literature has found that women may have greater domestic responsibilities and therefore spend less time working (Lundberg and Stearns, 2019). Consequently, it is possible that women may be working fewer hours overall, and therefore also spending fewer hours on research than men. We explore output per year across different age ranges and show that most of the gender output gap materializes within the 31-to-49 age group, which corresponds to the years with peak probability of having a young childforbothgenders. Wedonotfindanysignificantgendergapinresearchoutputperyearfor researchersage30orbeloworaboveage50. Relatedly, studies of faculty in science have also found a gender discrepancy in time use within working hours, with women spending more time on teaching, service, and other nonresearch academic activities such as mentoring students (Xie and Shauman, 2003; Misra et al., 2011). Similarly,Babcocketal.(2017)showthatfemalefacultyaremorelikelytovolunteerfor tasksthatareunlikelytocontributetochancesforpromotion. Thesepatternsmaybeexplained bydifferencesinaspirationsorexpectationsaboutresearchsuccessorthereturnstoinvestment inresearch(Azmat,CuñatandHenry,2024;ShastryandShurchkov,2024;McFall,Parolinand Zafar, 2024). Women System economists may similarly be devoting more time to mentoring, service,andpolicyworkthantheirmalecolleaguesandrelativelylesstimetoresearch. Though women seem to face fewer obstacles to advancement in the System than in academia, these dynamicsintimeallocationmaycontributetothegendergapweobserveinresearchoutput. Theremainderofthepaperisorganizedasfollows: Section2describesthedataandsample construction. Section 3 discusses the framework and empirical methodology used in the paper, andSection4presentstheempiricalresults. Section5concludes. 5

2 Data 2.1 Data collection and construction Our data set focuses on the 12 regional Federal Reserve Banks and the Federal Reserve Board. These13institutionsareaffiliatedthroughcommonpurposeandparticipationinthemissionof the Federal Reserve System, and collaboration on both policy and research projects is common across the System. Focusing on these institutions allows us to have unique and diverse data on 1,291economistsbetween2004and2021,whichisalargenumberrelativetoevenacollectionof thetopeconomicdepartments(Chari,2023). Althoughtheseinstitutionsoperateindependently withrespecttohiring,itiscommontoobserveeconomistswhomoveacrossinstitutionswithin the System, given that experience at either the Board or one of the Banks would be relevant to hiringinterestsatanother.6 WehandcollecteddataonthesetofSystemPh.D.economists,basedontheindividualpublic websitesofthe12BanksandtheBoard.7 ForhistoricaldataonSystemeconomists,weusethe WaybackMachine,whicharchiveswebpages,oftenmultipletimesamonth.8 UsingPython,we scrape the names of all System economists listed each year from 2004 to 2021. We use stringmatching techniques to aggregate records for individuals who may have changed their names or moved across System institutions over time. Given the appearance and disappearance of individualsfromthewebsitesovertime,thescrapingalsoallowsustotrackeconomists’starting years and departure years, giving us individuals’ length and timing of service. We supplement thesedatawiththedatacollectedbyDattaandVigfusson(2024)onBoardeconomists,whichis alsosourcedfromWaybackMachinearchivesoftheBoardwebsitefrom2004to2021. Formost 6Wefindthataround10percentofresearcheconomistsinoursampleworkedatmorethanoneFederalReserve institutionoverthecourseoftheircareer. 7For example, the Board website listing research economists is https://www.federalreserve.gov/econres/researchers.htm, and the Federal Reserve Bank of Dallas website listingeconomistsishttps://www.dallasfed.org/Home/research/economists.aspx. 8WecollectdatafromInternetArchive. (n.d.). WaybackMachine,https://web.archive.org/,thoughthearchive is missing some Banks in some years: Federal Reserve Banks of Richmond (2004), Cleveland (2004), Chicago (2004-06),andKansasCity(2009). 6

of our empirical work, we focus on economists who are in the System for at least three years. Thisfilteringhelpsusexcludevisitingscholars,whoaresometimeslistedonthesewebsitesbut tendtohaveshorterperiodsofaffiliationandadifferentsetofprofessionalresponsibilitiesfrom permanentstaff. Overall,weareleftwith1,291economistsintheSystem,including1,002men and289womeneconomists,orabouta22percentfemaleshare.9 The gender tags we use for this study are also the result of our data construction effort. We take as given the gender tags of Board economists from Datta and Vigfusson (2024) and then generate the gender tags for Bank economists following a similar procedure. We first use data sets with names of economists and gender tags generated by Chari and Goldsmith-Pinkham (2017) and Hengel (2022). Next, as in Chari and Goldsmith-Pinkham (2017), we gender tag individuals using the Tang et al. (2011) gender dictionary. If a name is associated with one genderatleast95percentofthetimeinthisdictionary,weapplythatgendertagtoallindividuals inourdatasetwiththatname. Finally,wedowebsearchestoidentifyindividualswhosegenders remainunidentifiedbythesesources. Ofthe1,291individualsinourdataset,456areidentified bytheChariandGoldsmith-Pinkham(2017)data,274areidentifiedbytheHengel(2022)data, 473aremanuallyidentified,and88areidentifiedbythesupplementarydatasets.10 To construct our data set on research and policy output from the Federal Reserve System, we use the website Fed in Print, which catalogs all series published by the Federal Reserve BanksandtheBoard.11 TheFedinPrintwebsitecatalogspublicationsfromtheSystemresearch departments, including research-related technical publications such as Staff Studies and major articlesandtestimonyaslistedintheFederalReserveBulletin. Usingthiscatalog,weconstruct alistofallseriesthathaveatleast70individualpublicationentriessincetheyear2000. Taking some signal from Fed in Print descriptors for each series’ publication type, we then classify each series as either a policy report series (such as regional economic reports or economic 9AppendixTableA.1presentsadditionaldetail,includingcountsofeconomistsbyinstitution. 10WealsoverifiedthegendertaggingwithNamsorgendertagging;seeNamsor. (n.d.). “Recognizethetypeof apropernoun,”https://namsor.app/features/name-type-recognition/. 11FedinPrintisavailableathttps://fedinprint.org/. 7

commentary) or a research publication series. Notably, we classify as “policy” reports even articles which are not official policy documents within the Federal Reserve System. Rather, we seethesearticlesasreflectinganalysisthatislikelyrelevanttoorinspiredbypolicydiscussions, even when they represent individual staff views. Additionally, we note that these policy reports capture only a small fraction of the policy work conducted within the Federal Reserve System, as a majority of policy documents and discussions remain internal. After using Fed in Print to identify these series, we collect from RePEc the individual publication titles, authors, and publication years.12 For the series published by the Board, we use data on working papers collectedbyDattaandVigfusson(2024)andsupplementwiththeadditionalserieslistedbyFed inPrint. Notably,thelistofresearchpublicationseriesavailablefromFedinPrintandRePEcincludes both working paper series and journals published by the Banks. To determine which series are “final”Systemresearchpublicationsandwhichareworkingpapers,weusedataprovidedtousby RePEcthatlinksindividualworkingpaperstotheirappearanceinpeer-reviewedpublications.13 Wetakethe“republicationrate”foreachfullseriesasasignalofwhetherthatseriesisintended as a working paper series or a final publication series. In particular, we consider a series to be a working paper series if, according to RePEc data, more than 10 percent of the series is published again in another periodical series. Conversely, we consider a series to be a System research publication if less than 10 percent of the series is republished. Using all these data, we construct for each economist in our data set the career total as well as annual count of each of our four types of publications: working papers, peer-reviewed publications, System research publications,andpolicyreports. AppendixTableA.2presentsadditionaldetail. Overall, we include 17 working paper series, covering about 10,000 working papers from across the System, around 3,000 of which were also published in a peer-reviewed journal. 12LogEc. (2024). “AccessStatisticsforParticipatingRePEcServices,”https://logec.RePEc.org/. 13WethankChristianZimmermannandhisteamforgivingusaccesstothesedata. Notealsothatourmeasure ofpeer-reviewedpublicationsbySystemeconomistsislimitedtothosepublicationsthatfirstappearedasaSystem workingpaper. 8

Appendix Table A.3 presents all the working paper series and associated publication counts from the Banks and the Board. Appendix Table A.4 lists by institution the policy reports and Systemresearchseriesweincludeinourdata. Amongtheworkingpaperseries,aboutone-third ofpapersarefromtheBoard’stwoworkingpaperseries,theFinanceandEconomicsDiscussion SeriesandtheInternationalFinanceDiscussionPapersseries. Wealsohave12Systemresearch series, including nearly 3,500 observations, and 34 policy reports series, including about 6,000 observations. 2.2 Representation of women in the Federal Reserve System Using the data we constructed, we document that the female share of economists in the Federal Reserve System increased only slightly in recent decades, from around 20 percent early in the sample to around 22 percent most recently. That said, as shown in panel A of Figure 1, System institutionshaveroughlydoubledinsizeoveroursampleperiod. Consequently,evenastagnant female share still has resulted in greater numbers of women in each department, which may be helpful for mentorship and collaboration. Trends in female representation across the System are in line with trends in academia. Although the female share of economics faculty in higher education increased in earlier decades, it has been nearly stagnant over the past two decades, andwomenremainunderrepresented,especiallyinhigherfacultyranks.14 Alsoasinacademia,progressinfemalerepresentationisunevenacrossinstitutionsandover time. Panel B of Figure 1 shows substantial variation among the Federal Reserve Banks and the Board. For example, the share of female economists in Boston declined sharply from 45 to 27 percent between 2004 and 2021, and we observe a sharp increase in the share of female economists in Kansas City from 9 to 30 percent over the same period. Some variation across institutions is to be expected given their smaller sizes. Individual Banks have around 20 to 30 economists each in 2004, rising to around 30 to 70 economists in 2021. When we include 14Chari(2023)shows,usingCSWEPAnnualSurveys,thatduringthis2004-21period,thefemaleshareoftenuretrack faculty members increased from 15 to 23 percent. For the top 20 schools, the female share of tenure-track facultymembersincreasedonlyslightly,from13to17percent. 9

economists from the entire sample period, the female share of economists by institution varies from8.3percentinPhiladelphiato34percentinBoston. Wenextdocumentthat,asinacademia,weobservegenerallygreaterfemalerepresentationat juniorlevelsofseniorityandadeclineinfemalerepresentationatmoreseniorlevels. However, this pattern is not as strong in the System as in academia. Additionally, though we see some improvement over time in female representation at each seniority level, progress is uneven. In our data, we have two measures of seniority: years since Ph.D. and tenure in the System. Notably, these two measures may differ more than in academia, as the System frequently hires researchers after they spend some time in academia. Consequently, the distribution of Ph.D. experienceandSystemexperiencemaydiffersomewhat,includingbygender. Focusing first on measuring seniority by years since Ph.D. attainment, panel A of Figure 2 shows how female representation has changed over time for each level of seniority. We find thatfemalerepresentationisgenerallyincreasingonlyatthemidcareerlevel(showninyellow). We see only the slightest of indications of increasing female representation for senior women over time (in red) and a decline in the share of female representation for more recent Ph.D. graduatesinthelastdecade(ingreen). Bycontrast,notwithstandingthemostrecentstagnation of earlier upward trends, female shares for assistant, associate, and full professors do seem to haveincreasedabout5to10percentagepointsfrom2004to2021(Chari,2023). Itmaybethat as efforts to increase women’s representation in academia have intensified in recent decades, the System faces increased competition for women job candidates, who tend to be more recent Ph.D. graduates. That said, we do not see a complementary increase in the share of women assistantprofessorsoverthesameperiod,callingintoquestionthispotentialexplanationforthe decliningfemaleshareamongthemostjuniorSystemeconomistsinrecentyears. For our second measure of seniority, Panel B of Figure 2 shows the number of economists grouped by tenure in the System in 2020. The female share of economists declines from 24 percent for those with up to 8 years of experience to around 21 percent for those with either 9 to 15 years of tenure or with 16 years or more of tenure. Though we see some evidence of 10

declining representation among more senior cohorts, the decline is not as strong as we observe in academia. For comparison, the shares of female assistant professors, associate professors, andfullprofessorsamongthetop20schoolswere22.7percent,21.2percent,and13.4percent, respectively(Chari,2023).15 3 Empirical framework Weexaminetherelationshipbetweengenderandpublicationsusingthefollowingregression: Y =α+β Female +β X +ζ +ε , (1) i 1 i 2 i i i where is one of five measures for economist : working papers, peer-reviewed publications, Y i i Systemresearchpublications,policyreports,or“totalpublications.”16 Eachmeasurerepresents the total count of that publication type each economist generates during our sample period of 2004 to 2021.17 Female is an indicator for a female economist, and ζ is an institution fixed i i effect. representsavectorofcontrolvariables,includingtenureintheSystemand,depending X i ontheoutcomevariableofinterest,someofthealternativepublicationmeasures. Forexample, whentheoutcomevariableofinterestisthecountofworkingpapers,weincludeSystemresearch andpolicyreportsascontrols. Using this regression, the coefficient on the indicator for a female economist provides an estimateofthegendergapinpaperproduction. Apositivecoefficientwouldindicatethatwomen produce more publications, on average, than men, while a negative coefficient would indicate the reverse. The tenure variable helps control for the fact that we may have different seniority distributions between men and women in our data and allows us to compare individuals with 15Amongalldepartmentswithdoctoralstudents,thesesharesare32.8percent,28.0percent,and15.5percent, respectively. 16Ourmeasureofpeer-reviewedpublicationscapturesonlypublishedversionsofworkingpapersalreadyincluded inourdataset,notpeer-reviewedpublicationsthatwerenotreleasedasSystemworkingpapers. Consequently,we excludepeer-reviewedpublicationsfromthistotaltoavoiddoublecounting. 17For economists who move across System institutions, we aggregate all their publications and associate their recordwiththeirlastobservedinstitution. 11

the same level of System tenure.18 The additional controls for alternative publications help us investigatetheinteractionamongthevarioustypesofoutput. Before turning to the regression results, we examine the distributions of each type of publication count by gender. Economists in the System produce, on average, 10 working papers, 4 peer-reviewed publications, 4 System research publications, and 5 policy reports.19 Figure 3 splits individuals into five bins for each type of output, ranging from economists with zero publications to those with at least 20 publications. We find that the distributions for each type of output are positively skewed, such that the reported means are strongly affected by the most productiveeconomists,andthemedianeconomistproducessomewhatloweroutput. Regarding the gender differences in distributions, we see in the left-most two bins of each panel that the share of female economists tends to be larger than the share of male economists. Thelargeshareofwomenwithzeropeer-reviewedorSystemresearchpublicationsisparticularly notable. By contrast, we see larger shares of male economists in the bins with higher paper counts, and we find a strikingly large share of men with more than 20 working papers. Using the gender-specific mean of each distribution to measure the gender gap, we find that women in our sample produce 4.4 fewer working papers, 2.1 fewer peer-reviewed publications, 1.9 fewer System research publications, and 1 less policy report than men. Notably, when we use the median measure, however, the gaps in working papers and peer-reviewed publications are roughlycutinhalf,thegapinSystemresearchdisappears,andthegapinpolicyreportsreverses suchthatthemedianwomanhasonemorepolicyreportthanthemedianmaninoursample. In the regression setting, we can go beyond these simple differences in means and medians and also investigate the interaction among types of outputs. Consider first the regression with working papers as the outcome variable and with policy reports and System research as controls. A variety of effects may influence the estimates of the coefficients on policy reports 18Menandwomenbothhaveabout10yearsoftenureintheSystemonaverage. Notethatourtenurevariable isbiaseddownward: Becauseourdatasetbeginsin2004, ourtenurevariableiscensoredtoonlycountyearsof tenuresince2004. 19AdditionaldetailisreportedinTablesA.1andA.2. 12

and System research. First, it may be that individual characteristics, such as hours worked or individual productivity will be most important. In this case, individuals who are more productive or who work more hours may produce more of all types of publications, and then we may see positive coefficients on the alternative publication count variables. It is also the case that potential complementarities between working papers and policy reports could result in a positive coefficient on policy reports, and analogously for System research. For example, a single research project may result in multiple types of outputs. However, there is necessarily somesubstitutionofhoursworkedondifferenttypesofoutput,andwewouldexpectthiseffectto contributenegativelytothecoefficientsonpolicyreportsandSystemresearchinthisregression. That is, more time spent on a policy report may result in less time available for research. This typeofspecialization,timeallocation,orsubstitutioneffectcouldresultinnegativecoefficients on the alternative output measures used as controls. To the extent that such specialization is correlated with gender, inclusion of controls would diminish our estimate of the gender gap in workingpaperproductionasrepresentedbytheestimatedcoefficientongender. Importantly,in thissetting,thecoefficientongenderrepresentsthegendergapamongindividualswhenholding othertypesofoutputconstant. Oursecondsetofanalysesfocusesonoutputperyearusingthefollowingregression: Y =α+β Female +β X +ζ +τ +ε , (2) it 1 i 2 it it t it where Y measures paper production for economist i in year t . We include institution (ζ ) and it it year (τ ) fixed effects.20 X represents controls for tenure in the System and, again, depending t it on the outcome variable of interest, some of the alternative publication measures. In this specification,thetenurevariablemeasurestenureintheSystemasofyear . Standarderrorsare t clusteredbyyear. 20ForeconomistswhomoveacrossSysteminstitutionsovertheircareer,theinstitutionalaffiliationwillchange overtime. 13

The interpretation of this regression is similar to the previous model but with a few key differences. First, whereas the previous regression measures the average gap between men and women over their entire careers in the System, this regression measures the average gender gap forasingleyear. Asweconsiderthecompetingeffectsofpolicyandresearchcomplementarity andsubstitutionofhoursworked,wemighthypothesizethathoursworkedarenecessarilymore constrainedwithinasingleyearandsothesubstitutioneffectmaydominate,leadingtonegative coefficientestimatesonthealternativeoutputmeasures. Ontheotherhand,itmaybethatthere aresomeyearsinwhichindividualshaveburstsofindividualproductivity,higherhoursworked, ormultiplerelatedpolicyandresearchoutputs,leadingtohigheroutputwithinagivenyearand resultinginpositivecoefficientestimatesonthesealternativeoutputmeasures. Figure 4 shows the raw data on publication output over time, as measured by the average numberofeachtypeofpublicationperyearbetween2004and2021formenandwomen. These unconditionalmeansillustratethegendergapinoutput,wherebywomenproducefewerworking papers and peer-reviewed publications on average. These gaps seem to diminish over time and eventually reverse for working papers in the final year of the sample. We see a similar gap in System research in the early period, which is greatly reduced in the latter half of the sample period. Lastly, we see a small gap in policy output that opens and closes repeatedly over the sampleperiod. Notably,coincidingwiththestartofthepandemic,asignificantjumpinthecountofworking papers and policy reports is evident for both men and women in 2020. Perhaps surprisingly, the spike for women is steeper than that for men. We do not see a step-up in peer-reviewed publicationsatthesametime,likelyduetothelaginpublicationtiming,thoughwemayexpect to see one as more data become available. We also observe a more modest uptick for System research in 2021, which may be consistent with System research taking more time to publish thanworkingpapers,butlesstimethanpeer-reviewedpublications. 14

Motivated by these observations, we include a related set of analyses using the following regression: Y =α+β Female +ζ +τ +Female ∗τ +ε . (3) it 1 i it t i t it Relative to equation 2, this model omits the controls for the alternative publication measures andaddsinteractiontermsbetweentheyearfixedeffectsandthefemaleindicatortomeasurethe gender-specificpatternsinpaperproductionduringthepandemic. Wediscussthepandemic-era findingsfurtherinthenextsection. 4 Results Panel A of Table 1 shows the coefficient estimates for equation 1, which we use to investigate therelationshipbetweengenderandthecountofpublicationseachSystemeconomistgenerates over the full sample period. Column 1 shows that female economists produce, on average, five fewer total publications than their male colleagues, even when controlling for tenure in the System. The following columns in panel A of Table 1 report coefficient estimates for the same regression,whenusingeachmeasureofoutputinturn,andwhilecontrollingforthealternative output measures. In column 2, we estimate that female economists produce 3.1 fewer working papers than men on average. These results suggest that the gap in working paper production accounts for most of the gender gap estimated in column 1. Additionally, this estimate is only slightly smaller than the raw difference between men and women of 4.4 working papers, as impliedbyFigure3.21 Turningtotheotherpublicationmeasures,incolumn3,wealsofindastatisticallysignificant gendergap. Thatis,evenwhenwecontrolforthenumberofworkingpapers(aswellasSystem research papers and policy reports), female economists produce, on average, 0.36 fewer peerreviewed publications. By contrast, columns 4 and 5 show that there are no gender gaps in System research and policy reports. Notably, though we see a gender gap of 1.9 in the raw 21The results are robust to dropping the different alternative measures of output, as shown in Appendix Table A.5. 15

dataforSystemresearchasreportedinFigure3,thisgaplargelydisappearsoncewecontrolfor System tenure and alternative measures of output. By contrast, our finding of no statistically significantgendergapforpolicyreportsisconsistentwiththerawdataaswell.22 Given our finding in Figure 3 that the distribution of output is highly right skewed, we next investigate measures of the gender gap in paper production when focused on highly productive economists, defined here as individuals in the 60th to 90th percentiles of working paper production (in panel B of Table 1) or policy reports (in panel C).23 Panel B shows that among the economistswhoproducethemostworkingpapers,wefindnogendergapinpolicyreports(column5)andaSystemresearchgendergapthatisappreciablylargebutnotstatisticallysignificant (column 4). However, we do still find a gap in peer-reviewed publications that is significant at the 10 percent level, again underscoring the gender differential in the likelihood of converting workingpaperstopublications(column3). Whereaswefindnopolicyoutputgendergapamongresearch-productiveeconomists,wedo find a research output gender gap among economists who are most policy-productive. That is, it seems that when averaging across the men and women who produce the most policy reports, wefindthatthewomentendtoproduceabout4fewerworkingpapers,onaverage,and2.5fewer peer-reviewed publications (columns 2 and 3 of panel C). The gender gap for System research among this group (column 4) is small and not statistically significant, however. Overall, these results indicate that research-productive men and women have similar policy output but that women produce fewer peer-reviewed publications. Meanwhile, among policy-productive men andwomen,weseealargegendergapinresearchoutput. Wenextturntoinvestigatingtherelationshipbetweengenderandpaperproductionperyear. On average, System economists produce one publication of some variety per year. That said, giventhevarioustypesofoutput,anindividualwhoproducesjustoneofanytypeofoutputina given year is in the top quartile of the distribution for that type of output. We estimate average 22InAppendixTablesA.6throughA.9,weshowthattheresultsinTable1arerobusttodroppingthemostand leastproductiveindividuals. 23AppendixTableA.10presentssummarystatisticsforthemostproductiveeconomistsintheSystem. 16

annualoutputforSystemeconomistsof0.6workingpapers,0.2peer-reviewedpublications,0.2 Systemresearchpublications,and0.3policyreports.24 Panel A of Table 2 reports estimation results for equation 2. Paper production is measured at the economist-year level, and our primary variable of interest is the indicator for a female economist. Column 1 in Table 2 shows that, on average, each year, female economists produce 0.11 fewer total publications. Columns 2 and 3 show that, on average, each year, and after we control for other types of output, female economists produce 0.10 fewer working papers and 0.06 fewer peer-reviewed publications, respectively. For context, we find that male economists produce0.64workingpapersperyearonaverage,andfemaleeconomistsproduce0.50peryear. Notably,theunconditionaldifferenceintheseratesofworkingpaperproductionis0.14,whereas ourregressionestimatesimplythatthemagnitudeofthegapisamodestlysmaller0.10oncewe control for tenure in the System as well as alternative measures of output. In columns 4 and 5, the estimated coefficient on the female indicator is near zero and not significant. We conclude that after controlling for other types of output, women produce a similar number of System researchpublicationsandpolicyreportsperyearasmen. Theseresultsaregenerallyconsistent withtheresultsfromTable1.25 A leading hypothesis for why female academics are less productive in research is that women tend to have more domestic responsibilities and therefore may work fewer hours (Kim andMoser,2021). Thisexplanationseemstobemostapplicableforwomenwithyoungchildren. Consequently, we next investigate output per year for various age groups. Given the limitations of our data, we follow Kruger, Maturana and Nickerson (2022) and estimate age by assuming Ph.D. completion at age 27. Next, we estimate the regression in equation 2 separately for each agegroup. Figure5plotsthecoefficientonthegendervariablefortheseregressionsandshows that the gaps for working papers and peer-reviewed publications (in panels A and C) are largest for the 31-to-49 age group, which corresponds to the peak probability of having a young child 24AppendixTableA.2reportssummarystatisticsforpapersperyear. 25In Appendix Tables A.11 through A.13, we show that the results in Table 2 are also robust to dropping the mostandleastproductiveindividuals. 17

for both genders (Kruger, Maturana and Nickerson, 2022).26 We conclude that the gender gap inthenumberofworkingpapersandpublicationsislargelymaterializingduringthisagerange. Wedonotfindanysignificantgenderdifferentialinoutputperyearforresearchersuptoage30 oraboveage50. AnotherpotentialexplanationforwhyfemaleeconomistsattheSystemproducelessresearch outputbutnotlesspolicyoutputcomparedwiththeirmalecolleaguesisthatfemaleeconomists may be choosing to specialize in policy or management instead of research. To investigate this possibility, we examine gender gaps in research and policy output across different levels of Systemtenure,thatis,thenumberofyearsthateacheconomisthasbeenemployedasaresearcher in the System. We hypothesize that economists with greater tenure will have more managerial responsibilities or will spend more time on policy work. Figure 6 illustrates the gender gap coefficient when studying subsamples of economists divided by tenure in the System. We find that the gender gap in research output as measured by working papers and peer-reviewed publications is fairly stable at first and then widens sharply among the most senior economists, who have more than 15 years of tenure in the system. By contrast, we find scant evidence of a gendergapinpolicyorSystemresearchoutput,atanyleveloftenure.27 In sum, our investigation of gender gaps by age and tenure suggests that women may be falling behind men in research output, particularly during the period when they are most likely to have greater family responsibilities. Gender bias in publication processes may also be a contributing factor. Further, senior women may be specializing in managerial or policy work more frequently than their male colleagues. Though the gender gap in research output per year 26FigureA.1showssimilarresultswhenweexplorethedifferenceinpaperproductionperyear,bytenuregroups. Notably,whenmeasuringsenioritybytenure,wedoseeagendergapinresearchproductionamongthemostjunior economists,whichisabsentwhenmeasuringsenioritybyPh.D.year. Thedifferenceinresultsmaybeattributable tothedifferenceinhiringpatternsbetweenmenandwomen: AtthetimeofhiringintotheSystem,wefindthatthe numberofyearssincePh.D.isoftenabithigherforwomenthanmen. 27Figure A.2 illustrates the gender gap coefficient when studying subsamples of economists divided by years since Ph.D. We find similar results to those presented in Figure 6. Of course, these results may be related to thepreviousresultsonpaperproductionbyyear. Thoughthegendergapinresearchoutputperyearnarrowsfor economistsaboveage50,thegendergapintotalnumberofpaperscouldremainlargerfortheseindividualsdueto thecumulativeeffectfrompreviousyears. 18

narrows for more senior women, the cumulative gap remains wider for more senior than more junior women. Relative to their male colleagues, then, it may be the case that women find it easier to succeed in policy or management roles after having a period of lower research output. Though such a pattern would be particularly harmful to career progression in academia, in the System, women may have a wider set of potential paths to career success. The availability of these alternative paths—for example, in management or policy—may be mitigating the “leaky pipeline”effectintheSystemrelativetoacademia. 4.1 COVID-19 pandemic disruptions Previous research has shown that COVID-19 pandemic disruptions disproportionately affected femaleacademics. Forexample,Myersetal.(2020),Barberetal.(2021),andKruger,Maturana and Nickerson (2022) show that during the pandemic, the gender gap in paper production widened, even as the overall rate of working paper production increased. To explore whether thesepatternsareobservableintheresearchandpolicyoutputofSystemeconomists,weuseour yearlyoutputdataandestimateequation3. Relativetotheregressionspecificationinequation2, we drop the controls for alternative outputs, so that we may focus on each output individually, andweaddinteractiontermsbetweentheyearfixedeffectsandthefemaleindicator. Thoughwe includeinteractiontermsforeachyearofthesampleinourestimation,wereportthecoefficients hereonlyfor2020and2021. InpanelBofTable2,thelargepositivecoefficientsontheyeardummyfor2020incolumns2 and 5 provide evidence of a strong uptick in working paper and policy report production in 2020. We then see full retracement for working papers in 2021 (given the negative coefficient in column 2), and partial retracement for policy reports. Surprisingly, we find that female economists in the System increased their production even more than their male colleagues did during the pandemic, as evidenced by the large positive coefficients on the interaction between the female indicator and the fixed effects for 2020 and especially for 2021 for both working 19

papers and policy reports. These results are consistent with the raw data on paper production overtimeinFigure4. Our results are at odds with existing results noting a more negative effect of the pandemic on women. Pandemic-period findings using survey data, including Myers et al. (2020), Barber et al. (2021), and Deryugina, Shurchkov and Stearns (2021), suggest that financial economists and academics more generally suffered a reduction in research production, with a larger effect among women. Similarly, Cui, Ding and Zhu (2022) find a decline in journal submissions by womenduringthepandemicyears. Thatsaid,ourresultsareconsistentwiththemorenuancedfindingbyKruger,Maturanaand Nickerson(2022)thatbothmenandwomenincreasedtheirresearchoutputduringthepandemic and that the most harmful effects of the pandemic can be observed for women in the 35-to-49 age group, who registered the smallest increase in paper production during this period and who are the most likely to have young children. Men in this age range experienced large gains in paper production, and younger and older women fared similarly to their male counterparts. Consequently, Kruger, Maturana and Nickerson (2022) find that the gender gap widens most sharplyforthe35-to49-year-oldgroup. Wedonotfindsimilareffects,asourdatainsteadshow gainsforwomeninallcohortsduringthepandemic.28 How can we reconcile our results with these existing studies? One potential explanation is related to the observation in Barber et al. (2021) that relative to men, women allocated significantly less time to research and significantly more time to teaching following the onset of the pandemic. It may be that, in addition to family dynamics, differential expectations or investment in teaching was a key driver of the pandemic effects on the gender gap in research production, and the absence of teaching demands in the Federal Reserve System contributed to the relatively better outcomes for women researchers. Kruger, Maturana and Nickerson (2022) also find negative effects for individuals with more coauthors whose productivity was most affected by thepandemic disruptions. Similarly, we knowfrom Ductor (2015) that coauthoring 28Resultsareavailablefromtheauthorsuponrequest. 20

networks have large spillovers to researcher productivity. That said, we know from Datta and Vigfusson (2024) that at least at the Federal Reserve Board, coauthoring networks are heavily concentrated within the Federal Reserve System. Consequently, it may be that the thick coauthoring networks, along with the observed increase in System research output, were strong enough to offset potentially negative effects on women during this period. That is, to the extent that women researchers in the System experienced fewer disruptions, there was also less amplification of these negative effects across research networks and indeed more positive amplification,leadingtobetteroutcomesforwomenintheSystemoverall. 4.2 Complementarity among outputs Havinginvestigatedthegendergapsinproduction,wenowturntoexamininghowtheinclusionof thealternativeoutputcontrolvariablesinTables1and2canshedlightonthecomplementarities andsubstitutionthatmaybepresentamongoutputsandtheirinteractionwithgenderdifferences in paper production. First, if the gender gap in research is driven by a pattern in which women devotemoretimetopolicyreportsandmendevotemoretimetoworkingpapers,thenwemight expect the estimated gender gap for working papers to be smaller when we add the control for the number of policy reports. That is, we may observe a larger gender gap in working papers on average, but a smaller one when comparing men and women who have the same number of policy reports. The results support this hypothesis to only a modest degree, as we generally observe a gender gap in working paper production even among individuals with the same level ofpolicyreportproduction. Eveniftosomedegree,womenspecializeinpolicyoutputandmen specializeinresearchoutput,itcannotfullyexplainthegapinworkingpaperproduction. Second, we can examine the coefficient estimates on the controls themselves. There are necessarily some tradeoffs in the allocation of hours worked on different types of output. Consequently, we might expect to see a negative coefficient on the working papers control variable when policy reports are the outcome variable, indicating that, on average, individuals with higher research output have lower policy output. Instead, we generally see positive significant 21

coefficients on the controls across the various regression estimations, indicating that the effect oftimeallocationacrossvariousoutputsmaynotbethatstrong. Rather,thecomplementarities among outputs and effects of higher individual productivity or greater hours worked are likely strongerthantheeffectoftradeoffsintimeallocation. Third,wenotethatthesecomplementaritiesareparticularlystrongforSystemresearch. For Systemresearch,weseeagendergapintherawdata,asindicatedbythedifferenceinthemeans of the distributions in Figure 3. In our regression setting, once we control for working papers, thisgapinSystemresearchpaperproductionlargelydisappears. Thisresultsuggeststhatclosing the gender gap in working paper production could also be beneficial in closing the gender gap inSystemresearchproduction. Fourth, it is particularly salient that the gender gap exists for peer-reviewed publications. Peer-reviewed publications are a key signal of research success and lead to potential outside options for System economists. We find in Figure 3 that female economists produce about two fewer peer-reviewed publications over the course of their careers. Furthermore, in panel A of Table 1, we find that this gender gap remains even when controlling for the number of working papers (see column 3), indicating that women are also less likely than men to convert working papers into publications. In panel B of Table 1, we also show this gender gap exists even among the most productive researchers. Previous research has found empirical evidence of various obstacles to women’s success in research, including higher standards for publications, longer time spent in peer review, and more rounds of revisions required for paper acceptance (Hengel, 2022; Alexander et al., 2023; Hengel and Moon, 2023; Krawczyk and Smyk, 2016; Cardetal.,2020).29 Thestatisticallysignificantgendergapforpeer-reviewedpublications,even after controlling for the count of working papers, represents a potentially concerning feature of thedatathatwarrantsfurtherunderstanding. 29Anotherstrandofresearchfindsthatwomentendtohavelesswell-developedcoauthorshipnetworks,leading toreducedspilloverstoresearch(Ductor,GoyalandPrummer,2023). 22

By contrast, we see little evidence of a gender gap for policy reports in both the raw distributions and the regression setting. One potential explanation for this finding is that these obstacles to publishing and research output are less relevant for policy output (and System research as well), or further, that such obstacles do not exist for policy output. For example, policy reports may not apply higher standards for women or may not take longer for peer review. Coauthorship networks also may play a lesser role for policy reports, which are more likely to result from policy work that is more centrally assigned. Another hypothesis is that Systememployeesarerequiredtohaveaminimumlevelofpolicyoutputandthenmayusetheir remainingworkinghourstoproduceresearchoutput. Underthismodel,itmaybethatmenhave more time for research, for instance, if (1) men take fewer hours than women to produce the required policy output, (2) men spend less time on service tasks, or (3) men work longer hours on average. Under this hypothesis, the lack of a gender gap for policy output is in some sense an institutional requirement, and the gap in research output remains a potentially concerning featureofthedata. 5 Conclusion ThispapermeasuresgendergapsinrepresentationandresearchandpolicyoutputamongFederal Reserve System economists. We document only a slight uptick in the female share of System economistsoverthepast20years,from20to22percent,whichisroughlyinlinewithtrendsin the economics profession. We also find that women produce fewer total publications over their career and on a per-year basis. The majority of this gender gap is accounted for by working papers, as the gender gap in System research papers is much smaller and the gap in policy reportsisnearlyabsent. Theseresultssuggestthatdifferenttimeallocationsbetweenpolicyand research may drive differences in research output. A successful career as an economist in the Federal Reserve System typically requires contributions to both research and policy analysis, andsogendergapsinresearchoutputareimportanttounderstand. 23

Though there are necessarily some tradeoffs in time allocation across outputs, our results suggest that there are strong complementarities across all types of paper production. Even on a year-to-year basis for a given individual, higher output of one type is correlated with higheroutputofothertypesinthesameyear. Theseresultssuggestpotentiallystrongspillovers betweencontemporaneouspolicyandresearchoutput,orlife-cycleeffectswherebyhoursworked or individual productivity is particularly high in some years. More generally, we find that the effects of individual output, higher working hours, or spillovers among various measures of output must be stronger than the negative spillovers across output types that must result from thelimitedtimethatindividualshavetoallocateacrossvarioustypesofoutput. What can we learn from these results? A leading hypothesis for why female academics are less productive is due to the gender discrepancy in time use even within working hours, with women spending more time on teaching, service, and other non-research academic activities. This result may also apply to Federal Reserve System researchers. Additionally, in the System setting, investment in research production could also be seen as a risky strategy for investing in outsideemploymentoptionsandcouldbelesspreferredbywomenrelativetothesaferstrategyof investmentinpolicyoutputthatmayhavemorecertainpayoffsintheSystemsetting. Consistent withcomplementaryfindingsintheliteraturesuggestingthatwomenfacegreaterobstaclesinthe peer-reviewandpublicationprocess,wefindthatevenaftercontrollingforthenumberofworking papers,thereremains agendergapinthecount ofpeer-reviewedpublications. Consequently, it mayalsobethatwomen’sinvestmentinresearchisriskierthanmen’sinvestmentinresearch. AnotherpotentialexplanationforthegendergapsinresearchoutputamongSystemeconomists isthatwomenhavemoreintensedomesticresponsibilitiesandthereforemayworkfewerhours; indeed, the evidence from most STEM fields is that publications by single, childless females are not significantly different from publications by single childless men. Although we do not haveinformationonthefamilystatusoftheeconomistsintheSystem,wedofindlargergender gaps in research output among economists in the age range in which individuals have a higher probabilityofhavingyoungkids. 24

Our investigation of gender and output patterns among central bank economists can be informative for the state of the economics profession. Potential barriers to research output and career progression in the System may echo similar obstacles for women in the economics profession at large, including in academia. Focusing on economists in the System provides a largenumberofeconomiststostudyandthusmayprovideinsightforthebroaderprofession. Beyondprovidinginformationabouttheeconomicsprofession,ourevidenceshowinglagging representation by and contributions from women in the Federal Reserve System may be of concern. The role of the Federal Reserve is to promote the health of the U.S. economy and the stability of the U.S. financial system. Research economists in the Federal Reserve System are responsible for providing economic analysis and recommendations to policymakers on a wide rangeoftopics,includingmonetary,regulatory,andsupervisorypolicies. Itisimportanttohave diversityamongtheSystem’sresearcherstobringinavarietyofperspectives,backgrounds,and experiencesthatmightaffectoveralleconomicperformance. 25

References Alexander, Diane, Olga Gorelkina, Erin Hengel, and Richard Tol. 2023. “Gender and the TimeCostofPeerReview.”TinbergenInstitute,TinbergenInstituteDiscussionPapers2023- 044/V. Auriol,Emmanuelle,GuidoFriebel,AlisaWeinberger,andSaschaWilhelm.2022.“UnderrepresentationofWomenintheEconomicsProfessionMorePronouncedintheUnitedStates Compared to Heterogeneous Europe.” Proceedings of the National Academy of Sciences, 119(16). Azmat, Ghazala, Vicente Cuñat, and Emeric Henry. 2024. “Gender Promotion Gaps and CareerAspirations.”ManagementScience. Azzimonti-Renzo, Marina, Arantxa Jarque, and Acacia Wyckoff. 2023. “How Are Women RepresentedinEconomicResearchattheFed?” EconomicBrief,23(40). Babcock, Linda, Maria P. Recalde, Lise Vesterlund, and Laurie Weingart. 2017. “Gender Differences in Accepting and Receiving Requests for Tasks with Low Promotability.” AmericanEconomicReview,107(3):714–47. Barber, Brad M., Wei Jiang, Adair Morse, Manju Puri, Heather Tookes, and Ingrid M. Werner. 2021. “What Explains Differences in Finance Research Productivity during the Pandemic?” JournalofFinance,76(4):1655–97. Card,David,StefanoDellaVigna,PatriciaFunk,andNagoreIriberri.2020.“AreReferees and Editors in Economics Gender Neutral?” Quarterly Journal of Economics, 135(1): 269– 327. Ceci, Stephen J., Donna K. Ginther, Shulamit Kahn, and Wendy M. Williams. 2014. “WomeninAcademicScience: AChangingLandscape.”PsychologicalScienceinthePublic Interest,15(3):75–141. Chari, Anusha. 2023. “Report of the Committee on the Status of Women in the Economics Profession.”Vol.113,815–39. Chari, Anusha, and Paul Goldsmith-Pinkham. 2017. “Gender Representation in Economics across Topics and Time: Evidence from the NBER Summer Institute.” National Bureau of EconomicResearchWorkingPaper23953. Cui, Ruomeng, Hao Ding, and Feng Zhu. 2022. “Gender Inequality in Research Productivity during the COVID-19 Pandemic.” Manufacturing & Service Operations Management, 24(2):707–26. Datta, Deepa D., and Robert J. Vigfusson. 2024. “Measuring Inclusion: Gender and CoauthorshipattheFederalReserveBoard.”Workingpaper. 26

Deryugina, Tatyana, Olga Shurchkov, and Jenna Stearns. 2021. “COVID-19 Disruptions DisproportionatelyAffectFemaleAcademics.”Vol.111,164–68. Ductor,Lorenzo.2015.“DoesCo-authorshipLeadtoHigherAcademicProductivity?” Oxford BulletinofEconomicsandStatistics,77(3):385–407. Ductor,Lorenzo,andAnjaPrummer.2024.“GenderHomophily,Collaboration,andOutput.” JournalofEconomicBehavior&Organization,221:477–92. Ductor, Lorenzo, Sanjeev Goyal, and Anja Prummer. 2023. “Gender and Collaboration.” ReviewofEconomicsandStatistics,105(6):1366–78. Ginther,DonnaK.,andShulamitKahn.2004.“WomeninEconomics: MovingUporFalling OfftheAcademicCareerLadder?” JournalofEconomicPerspectives,18(3):193–214. Hengel, Erin. 2022. “Publishing while Female: Are Women Held to Higher Standards? EvidencefromPeerReview.”EconomicJournal,132(648):2951–91. Hengel,Erin,andEunyoungMoon.2023.“GenderandEqualityatTopEconomicsJournals.” WorkingPaper. Hospido,Laura,LucLaeven,andAnaLamo.2022.“TheGenderPromotionGap: Evidence fromCentralBanking.”ReviewofEconomicsandStatistics,104(5):981–96. Kim, Scott Daewon, and Petra Moser. 2021. “Women in Science. Lessons from the Baby Boom.”NationalBureauofEconomicResearchWorkingPaper29436. Krawczyk, Michał, and Magdalena Smyk. 2016. “Author’s Gender Affects Rating of Academic Articles: Evidence from an Incentivized, Deception-Free Laboratory Experiment.” EuropeanEconomicReview,90:326–35. Kruger, Samuel, Gonzalo Maturana, and Jordan Nickerson. 2022. “How Has COVID-19 Impacted Research Production in Economics and Finance?” Review of Financial Studies, 36(8):3348–81. Lundberg, Shelly, and Jenna Stearns. 2019. “Women in Economics: Stalled Progress.” JournalofEconomicPerspectives,33(1):3–22. McFall, Brooke Helppie, Eric D. Parolin, and Basit Zafar. 2024. “Career Expectations and Outcomes: Evidence (on Gender Gaps) from the Economics Job Market.” National Bureau ofEconomicResearchWorkingPaper32446. Misra,Joya,JenniferHickesLundquist,ElissaHolmes,andStephanieAgiomavritis.2011. “TheIvoryCeilingofServiceWork.”Academe,97(1):22–26. Myers, Kyle R., Wei Yang Tham, Yian Yin, Nina Cohodes, Jerry G. Thursby, Marie C. Thursby, Peter Schiffer, Joseph T. Walsh, Karim R. Lakhani, and Dashun Wang. 2020. “Unequal Effects of the COVID-19 Pandemic on Scientists.” Nature Human Behaviour, 4(9):880–83. 27

Sarsons, Heather, Klarita Gërxhani, Ernesto Reuben, and Arthur Schram. 2021. “Gender DifferencesinRecognitionforGroupWork.”JournalofPoliticalEconomy,129(1):101–47. Shastry, Gauri Kartini, and Olga Shurchkov. 2024. “Reject or Revise: Gender Differences inPersistenceandPublishinginEconomics.”EconomicInquiry,62(3):933–56. Sherman, Mila Getmansky, and Heather Tookes. 2022. “Female Representation in the AcademicFinanceProfession.”JournalofFinance,77(1):317–65. Tang, Cong, Keith Ross, Nitesh Saxena, and Ruichuan Chen. 2011. “What’s in a Name: A StudyofNames,GenderInference,andGenderBehaviorinFacebook.”InDatabaseSystems for Advanced Applications. , ed. Jianliang Xu, Ge Yu, Shuigeng Zhou and Rainer Unland, 344–56.SpringerBerlinHeidelberg. Wu, Alice H. 2018. “Gendered Language on the Economics Job Market Rumors Forum.” In AEAPapersandProceedings.Vol.108,,ed.WilliamR.JohnsonandKellyMarkel,175–79. Xie,Yu,andKimberleeAShauman.2003.WomeninScience.HarvardUniversityPress. 28

FIGURE1 FemaleRepresentationacrossInstitutions PanelA.FederalReserveBoardandReserveBanks PanelB.FemaleShare,byInstitution Notes: PanelAshowsthenumberofmaleandfemaleeconomistsattheFederalReserveBoardandReserveBanks between2004and2021. PanelBshowstheshareoffemaleeconomistsbetween2004and2021ateachofthe12 FederalReserveBanksandtheBoard. AfewBankswithnotabletrendsarehighlighted.

FIGURE2 FemaleRepresentationacrossSeniorityLevels PanelA.FemaleRepresentationoverTime,byYearssincePh.D. PanelB.FemaleRepresentationin2020,byYearsofSystemTenure Notes: Using years since Ph.D. to measure seniority, panel A shows how the female share of Federal Reserve Systemeconomistsevolvesovertimeateachlevelofseniority. UsingyearsofSystemtenuretomeasureseniority, panelBshowsthenumberofmaleandfemaleSystemeconomistsandfemaleshareofSystemeconomicsateach levelofseniorityin2020.

FIGURE3 DistributionofPublicationCountbyGender PanelA.WorkingPapers PanelB.PolicyReports PanelC.Peer-ReviewedPublications PanelD.SystemResearchPublications Notes: This figure shows, for men and women economists, the distributions of the count of working papers (panel A), policy reports (panel B), peer-reviewed publications (panel C), and Federal Reserve System research publications (panel D). The sample is restricted to System economists with at least 3 years of service. The lists ofworkingpaper, Systemresearch, andpolicyreportseriesateachSysteminstitutioncanbefoundinAppendix TablesA.3andA.4.

FIGURE4 GenderandPaperProductionperYear,overTime PanelA.WorkingPapers PanelB.PolicyReports PanelC.Peer-ReviewedPublications PanelD.SystemResearchPublications Notes: This figure plots average annual paper production by men and women for working papers (panel A), policyreports(panelB),peer-reviewedpublications(panelC),andFederalReserveSystemresearchpublications (panelD).Thesamplecovers2004to2021andincludesallSystemeconomistswithatleast3yearsofservice. The verticallineineachpaneldenotestheonsetoftheCOVID-19pandemicperiodin2020.

FIGURE5 GenderGapsinPaperProductionperYear,byAge PanelA.WorkingPapers PanelB.PolicyReports PanelC.Peer-ReviewedPublications PanelD.SystemResearchPublications Notes: This figure illustrates gender gaps in average annual production of research working papers (panel A), policyreports(panelB),peer-reviewedpublications(panelC),andFederalReserveSystemresearchpublications (panelD),byage. WeestimatetheageofaneconomistbyassumingPh.D.completionatage27andaddingthe yearssincePh.D.attainmentthroughtherelevantyear. ThesampleisrestrictedtoSystemeconomistswithatleast 3yearsofservice. TableA.3listsresearchworkingpaperseriesateachFederalReserveinstitution. TableA.4lists Systemresearchandpolicyreportseries.

FIGURE6 GenderGapsinPaperProduction,byYearsofSystemTenure PanelA.WorkingPapers PanelB.PolicyReports PanelC.Peer-ReviewedPublications PanelD.SystemResearchPublications Notes: Thisfigureillustratesgendergapsinthetotalnumberofresearchworkingpapers(panelA),policyreports (panelB),peer-reviewedpublications(panelC),andFederalReserveSystemresearchpublications(panelD),by yearsofSystemtenure. ThesampleisrestrictedtoSystemeconomistswithatleast3yearsofservice. TableA.3 listsresearchworkingpaperseriesateachFederalReserveinstitution. TableA.4listsSystemresearchandpolicy reportseries.

TABLE1 GenderandPaperProduction 1 2 3 4 5 Total Working Peer-reviewed System Policy publications papers publications research reports PanelA.Alleconomists Female -5.00*** -3.09*** -0.36** -0.11 -0.10 (1.30) (0.52) (0.15) (0.39) (0.53) Tenure 1.77*** 0.68*** 0.04 0.15 0.36* (0.53) (0.09) (0.05) (0.11) (0.20) Policyreports 0.09 0.01 0.06 (0.07) (0.01) (0.05) Systemresearch 0.48*** -0.00 0.14 (0.11) (0.03) (0.16) Workingpapers 0.44*** 0.21*** 0.09** (0.02) (0.07) (0.03) Observations 1,291 1,291 1,291 1,291 1,291 R-squared 0.34 0.34 0.77 0.37 0.26 PanelB.Highlyproductiveworkingpaperauthors Female -0.57* -0.75 -0.01 (0.30) (0.52) (0.77) Tenure 0.17*** 0.18 0.53* (0.05) (0.10) (0.29) Observations 445 445 445 R-squared 0.15 0.38 0.18 PanelC.Highlyproductivepolicyreportauthors Female -3.98*** -2.51*** -0.25 (0.82) (0.29) (0.51) Tenure 0.69*** 0.41** 0.29 (0.19) (0.15) (0.26) Observations 258 258 258 R-squared 0.25 0.22 0.40 Notes: ThistablereportscoefficientsfromregressionsoftheformY =α+β Female +β X +ζ +ε,where i 1 i 2 i i i Y is one of five measures of publications for economist i: working papers, peer-reviewed publications, System i research, policy reports, or “total publications,” which we define as the sum of working papers, Federal Reserve System research, and policy reports. Each measure represents the total count of publications the economist has generatedoverthecourseoftheirSystemtenuresince2004. Female isanindicatorforafemaleeconomist,and i ζ represents institution fixed effects, which are included in all regressions. X represents the vector of control i i variables,whichincludethealternativepublicationsmeasuresforwhichcoefficientsarereportedinthetable. The sampleconsistsofalleconomistswithatleast3yearsoftenureintheSystemoverthe2004-21period. Thelistsof workingpapers,Systemresearch,andpolicyreportsseriesateachSysteminstitutionareinAppendixTablesA.3 andA.4. Allregressionsincludeinstitutionfixedeffects. PanelAincludesalltheeconomistsattheSystem,panel Bincludeseconomistswith8to23workingpapers,andpanelCincludeseconomistswith4to12policyreports (equivalent to the 75 to 90 percentile range for each working papers and 60 to 90 percentile for policy reports). PanelsBandCincludethesamecontrolsandfixedeffectsasinpanelA.***,**,and*indicatestatisticalsignificanceatthe1percent,5percent,and10percentlevels,respectively.

TABLE2 GenderandPaperProductionperYear 1 2 3 4 5 Total Working Peer-reviewed System Policy publications papers publications research reports PanelA.Outputperyear Female -0.11** -0.10** -0.06*** -0.02 0.03 (0.04) (0.04) (0.01) (0.01) (0.02) Tenure 0.01 -0.00 0.01** 0.00 0.01 (0.01) (0.01) (0.00) (0.00) (0.01) Policyreportsperyear 0.09*** 0.01 0.05** (0.01) (0.01) (0.02) Systemresearchperyear 0.16** 0.06* 0.11* (0.07) (0.03) (0.06) Workingpapersperyear 0.20*** 0.05** 0.06* (0.04) (0.02) (0.03) Observations 12,996 12,996 12,996 12,996 12,996 R-squared 0.13 0.06 0.17 0.12 0.19 PanelB.Focusonthepandemicperiod Female -0.21 -0.14 -0.13** -0.10 0.03 (0.19) (0.12) (0.05) (0.07) (0.10) Year2020 0.45 0.24 0.04 -0.13 0.34 (0.32) (0.14) (0.06) (0.08) (0.26) Year2021 0.10 -0.04 0.03 -0.06 0.20 (0.30) (0.11) (0.10) (0.06) (0.21) Female#2020 0.23 0.05 0.06 0.12 0.06 (0.19) (0.13) (0.04) (0.08) (0.10) Female#2021 0.50* 0.29** 0.03 0.10 0.11 (0.24) (0.12) (0.06) (0.08) (0.15) Tenure 0.01 -0.00 0.01 0.00 0.01 (0.01) (0.01) (0.00) (0.00) (0.01) Observations 12,996 12,996 12,996 12,996 12,996 R-squared 0.13 0.05 0.04 0.11 0.18 Notes: ForpanelA,thistablereportscoefficientsfromregressionsoftheformY =α+β Female +β X + it 1 i 2 it ζ +τ +ε , where Y is one of five measures of publications for economist i in year t: working papers, peeri t it it reviewedpublications,FederalReserveSystemresearch,policyreports,or“totalpublications,”whichwedefineas thesumofworkingpapers,Systemresearch,andpolicyreportsinagivenyear. Eachmeasurerepresentsthetotal countofpublicationsaneconomisthasproducedinagivenyear. Female isanindicatorforafemaleeconomist. i Allregressionsincludeinstitutionfixedeffects(ζ)andyearfixedeffects(τ )for2004to2021. X representsthe i t it vectorofcontrolvariables,whichincludesthealternativepublicationsmeasures. ForpanelB,theregressionisof theformY =α+β Female +ζ +τ +Female ∗τ +ε . RelativetotheregressioninpanelA,thismodel it 1 i i t i t it omits the controls for the alternative publication measures and adds an interaction term between each year fixed effect and the Female indicator. Though all years and year interactions are incorporated in the regression, only coefficientsofprimaryinterestarereported. Forbothpanels,thesamplecoversthe2004-21periodandincludesall economistswithatleast3yearsoftenureintheSystem. Thelistsofworkingpapers,Systemresearch,andpolicy reports series at each System institution are in Appendix Tables A.3 and A.4. ***, **, and * indicate statistical significanceatthe1,5,and10percentlevelsrespectively.

Online Appendix Gender and Productivity in the Federal Reserve System DeepaD.Datta NitzanTzur-Ilan

A Additional Figures and Tables FIGUREA.1 GenderGapsinPaperOutputperYear,bySystemTenure PanelA.WorkingPapers PanelB.PolicyReports PanelC.Peer-ReviewedPublications PanelD.SystemResearchPublications Notes: This figure illustrates gender gaps in average annual production of research working papers (panel A), policy reports (panel B), peer-reviewed publications (panel C), and System research publications (panel D), by yearsofSystemtenure. ThesampleisrestrictedtoFederalReserveSystemeconomistswithatleastthreeyearsof service. TableA.3listsresearchworkingpaperseriesateachFederalReserveinstitution. TableA.4listsSystem researchandpolicyreportseries.

FIGUREA.2 GenderGapsinPaperOutput,byAge PanelA.WorkingPapers PanelB.PolicyReports PanelC.Peer-ReviewedPublications PanelD.SystemResearchPublications Notes: Thisfigureillustratesgendergapsinthetotalnumberofresearchworkingpapers(panelA),policyreports (panelB),peer-reviewedpublications(panelC),andFederalReserveSystemresearchpublications(panelD),by age. WeestimatetheageofaneconomistbyassumingPh.D.completionatage27andaddingtheyearssincePh.D. attainmentthrougheithertheendyearofoursample(2021)ortheirlastyearofSystememploymentifitisearlier than2021. ThesampleisrestrictedtoSystemeconomistswithatleast3yearsofservice. TableA.3listsresearch workingpaperseriesateachFederalReserveinstitution. TableA.4listsSystemresearchandpolicyreportseries.

TABLEA.1 SummaryStatistics: SystemEconomists Institution Numberof Shareof Yearsof Years Total Numberof Numberof Numberof Numberof economists women System since publications working peer-reviewed System policy tenure Ph.D. papers publications research reports publications Board 562 26.3 9.9 15.2 9.4 7.6 2.5 0.3 2.7 Atlanta 47 19.1 12.2 23.5 21.2 14.9 6.0 3.9 2.7 Boston 53 34.0 10.2 19.4 13.4 9.0 2.7 0.3 6.8 Chicago 51 25.5 12.2 20.2 24.1 13.7 5.1 4.1 8.3 Cleveland 80 18.8 8.0 18.5 13.7 8.8 3.0 5.4 1.2 Dallas 62 19.4 9.2 15.8 20.6 10.8 3.7 3.3 7.1 KansasCity 53 26.4 8.2 15.1 16.0 7.0 1.6 6.4 3.4 Minneapolis 35 22.9 9.4 23.7 16.1 17.3 7.1 0.5 1.4 NewYork 115 20.0 9.8 16.7 28.5 12.6 4.6 3.1 15.0 Philadelphia 36 8.3 10.5 18.9 24.1 18.6 6.0 5.2 1.9 Richmond 43 14.0 9.9 18.9 20.2 8.8 3.0 11.0 1.3 SanFrancisco 89 13.5 10.4 23.2 22.1 11.0 4.9 10.8 1.8 St.Louis 65 12.3 9.5 21.9 43.9 19.8 8.1 7.7 18.3 FullSystem 1,291 22.4 9.9 17.5 18.7 10.3 3.7 3.5 5.0 Women 289 9.8 16.9 13.0 6.8 1.9 2.0 4.2 Men 1,002 9.9 17.7 20.3 11.2 4.2 3.9 5.2 Notes: ThistableshowssummarystatisticsoneconomistsattheBoardofGovernorsandthe12FederalReserve Banks. Economists in the sample are those employed between 2004 and 2021 with at least 3 years of service. Column1presentscountsofeconomists,andcolumn2presentstheshareoffemalePh.D.economists. Column3 showstheaveragetotaltenureforeconomistsateachinstitution,definedasyearsofemploymentduringour2004- 21 sample period. Column 4 shows the average years since Ph.D. for economists at each institution, defined as yearssincePh.D.foreachindividualeconomistasoftheendyearofoursample(2021),orasoftheirlastyearof FederalReserveSystememploymentifitisearlierthan2021. Foreachinstitution,wereporttheaveragenumber ofpublicationspereconomist. Column5reportsaveragesfortotalpublicationspereconomist,whichisthesum oftheirworkingpapers,Systemresearchpublications,andpolicyreports. Remainingcolumnsreporttheaverage numberofpublicationsbytype, includingworkingpapers, peer-reviewedpublications, Systemresearchpublications, andpolicyreports. DataonpublicationsarecollectedfromRePEcandsupplementedbydatacollectedby DattaandVigfusson(2024). Thelistsofworkingpapers,Systemresearch,andpolicyreportsseriesateachSystem institutionareinAppendixTablesA.3andA.4.

TABLEA.2 SummaryStatistics: PapersperYear 75th Shareofobs. Mean Std. dev. percentile w. zeros Totalpublications 1.06 1.75 2 0.53 Totalpublications-Women 0.86 1.40 1 0.56 Totalpublications-Men 1.11 1.83 2 0.52 Workingpapers 0.61 1.11 1 0.65 Workingpapers-Women 0.50 0.90 1 0.68 Workingpapers-Men 0.64 1.17 1 0.64 Peer-reviewedpublications 0.22 0.61 0 0.84 Peerreviewedpublications-Women 0.14 0.44 0 0.89 Peer-reviewedpublications-Men 0.24 0.65 0 0.83 Systemresearchpublications 0.16 0.63 0 0.89 Systemresearchpublications-Women 0.09 0.36 0 0.93 Systemresearchpublications-Men 0.18 0.69 0 0.88 Policyreport 0.29 0.96 0 0.85 Policyreport-Women 0.27 0.82 0 0.84 Policyreport-Men 0.29 1.00 0 0.85 Notes: This table shows summary statistics for publications per year for all economists, split by gender. The sampleisrestrictedtoFederalReserveSystemeconomistswithatleast3yearsofservice.

TABLEA.3 WorkingPaperSeriesandPeer-ReviewedPublications,byFederalReserveInstitution FedinPrint Institution contenttype Series Workingpapers Publications Workingpaper InternationalFinanceDiscussionPapers 931 342 Board Workingpaper FinanceandEconomicsDiscussionSeries 2,054 665 Atlanta Workingpaper FRBAtlantaWorkingPaper 475 168 Boston Workingpaper WorkingPapers 327 85 Chicago Workingpaper WorkingPaperSeries 513 176 Workingpaper WorkingPapers 113 23 Cleveland Workingpaper WorkingPapers(OldSeries) 462 134 Workingpaper GlobalizationInstituteWorkingPapers 150 51 Dallas Workingpaper WorkingPapers 344 90 KansasCity Workingpaper ResearchWorkingPaper 340 80 Workingpaper WorkingPapers 181 52 Minneapolis Report StaffReport 351 148 NewYork Report StaffReports 924 285 Philadelphia Workingpaper WorkingPapers 674 187 Richmond Workingpaper WorkingPaper 293 79 SanFrancisco Workingpaper WorkingPaperSeries 605 255 St. Louis Workingpaper WorkingPapers 1,044 374 Notes: ThistableshowsthelistofresearchworkingpaperseriesandpublicationsbyFederalReserveBanks. We scrapealltheeconomists’researchworkingpaperstitles,authors,andpublicationyearforall12FederalReserve BanksfromtheRePEcwebsite. ThistablepresentsonlypapersdefinedasresearchpapersbytheFedinPrintwebsite. FortheFederalReserveBoardworkingpapers,weusethedatacollectedbyDattaandVigfusson(2024).

TABLEA.4 SystemResearchandPolicyReportsSeriesbyFederalReserveInstitution FedinPrint Numberof Institution contenttype Series Classifiedas papers Report ReportsandStudies Policy 37 Board Discussionpaper FEDSNotes Policy 454 Journalarticle FederalReserveBulletin Policy 221 Journalarticle EconSouth Policy 32 Atlanta Journalarticle EconomicReview Systemresearch 136 Discussionpaper PublicPolicyDiscussionPaper Policy 62 Conferencepaper ConferenceSeries;[Proceedings] Policy 39 Boston Journalarticle RegionalReview Policy 54 Journalarticle CommunitiesandBanking Policy 7 Journalarticle NewEnglandEconomicReview Policy 154 Newsletter ChicagoFedLetter Policy 430 Journalarticle EconomicPerspectives Systemresearch 244 Chicago Journalarticle Profitwise Policy 19 Conferencepaper Proceedings Policy 248 Conferencepaper Proceedings Policy 54 Cleveland Journalarticle EconomicCommentary Systemresearch 444 Journalarticle e-Perspectives Policy 5 Dallas Journalarticle SouthwestEconomy Policy 343 Journalarticle EconomicLetter Systemresearch 137 Conferencepaper Proceedings-EconomicPolicySymposium Policy 43 Journalarticle TEN Policy 33 KansasCity Journalarticle MainStreetEconomist Policy 64 Journalarticle EconomicReview Systemresearch 368 Journalarticle TheRegion Policy 19 Minneapolis Journalarticle Fedgazette Policy 113 Journalarticle CurrentIssuesinEconomicsandFinance Policy 162 NewYork Discussionpaper LibertyStreetEconomics Policy 864 Journalarticle EconomicPolicyReview Systemresearch 212 Discussionpaper ConsumerFinanceInstituteDiscussionPapers Policy 19 Philadelphia Journalarticle EconomicInsights Policy 53 Journalarticle BusinessReview Systemresearch 226 Briefing RichmondFedEconomicBrief Systemresearch 126 Richmond Journalarticle EconomicQuarterly Systemresearch 245 Journalarticle EconFocus Systemresearch 163 Conferencepaper Proceedings Policy 144 Journalarticle CommunityInvestments Policy 4 SanFrancisco Journalarticle CommunityDevelopmentInnovationReview Policy 83 Journalarticle FRBSFEconomicLetter Systemresearch 1,055 Newsletter PageOneEconomicsNewsletter Policy 25 Journalarticle Bridges Policy 33 Journalarticle MonetaryTrends Policy 105 Journalarticle BurgundyBooks Policy 36 St.Louis Journalarticle NationalEconomicTrends Policy 118 Journalarticle EconomicSynopses Policy 431 Journalarticle TheRegionalEconomist Policy 569 Journalarticle Review Systemresearch 527 Notes: WeexamineallseriesfromFederalReserveinstitutionscataloguedbytheFedinPrintwebsitethathave more than 70 papers published since 2000. We classify as working papers those series that have at least 10 percent of entries appearing in another peer-reviewed publication and report these series in Table A.3. The definition of Federal Reserve System research and policy reports series was taken from the Fed in Print website https://fedinprint.org/,whichisalsousedbyRePEcintheclassificationofresearchandpolicyreports.

TABLEA.5 GenderandPaperProduction 1 2 3 4 5 Total Working Peer-reviewed System Policy publications papers publications research reports PanelA.Genderandpaperproduction Female -5.00*** -3.57*** -1.92*** -0.89* -0.53 (1.30) (0.57) (0.27) (0.50) (0.55) Tenure 1.77*** 0.90*** 0.44*** 0.37* 0.49** (0.53) (0.17) (0.11) (0.20) (0.22) Observations 1,291 1,291 1,291 1,291 1,291 R-squared 0.34 0.26 0.22 0.29 0.24 PanelB.Genderandpaperproductionperyear Female -0.11** -0.10** -0.08*** -0.03 0.02 (0.04) (0.04) (0.02) (0.02) (0.02) Tenure 0.01 -0.00 0.01 0.00 0.01 (0.01) (0.01) (0.00) (0.00) (0.01) Observations 12,996 12,996 12,996 12,996 12,996 R-squared 0.13 0.04 0.03 0.11 0.17 Notes: ThistablereportscoefficientsfromregressionsoftheformY =α+β Female +β X +ζ +ε,where i 1 i 2 i i i Y is one of five measures of publications for economist i: working papers, peer-reviewed publications, Federal i Reserve System research, policy reports, or “total publications,” which we define as the sum of working papers, System research, and policy reports. Each measure represents the total count of publications the economist has generatedoverthecourseofhisorherSystemtenuresince2004. Female isanindicatorforafemaleeconomist, i andζ representsinstitutionfixedeffects,whichareincludedinallregressions. X representsthevectorofcontrol i i variable,whichincludetenureattheSystem. Thesampleconsistsofalleconomistswithatleast3yearsoftenure intheSystemoverthe2004-21period. Thelistsofworkingpapers,Systemresearch,andpolicyreportsseriesat eachSysteminstitutionareinAppendixTablesA.3andA.4. Allregressionsincludeinstitutionfixedeffects. ***, **,and*indicatestatisticalsignificanceatthe1percent,5percent,and10percentlevels,respectively.

TABLEA.6 GenderonProductivity-WorkingPapers 1 2 3 4 5 Baselineestimation WPs>0 WPs<60(99%) WPs<34(95%) WPs<24(90%) Female -3.09*** -3.22*** -2.60*** -1.66*** -1.26*** (0.52) (0.58) (0.42) (0.26) (0.23) Tenure 0.68*** 0.70*** 0.63*** 0.51*** 0.38*** (0.09) (0.12) (0.05) (0.04) (0.05) Policyreports 0.68*** 0.70*** 0.63*** 0.51*** 0.38*** (0.09) (0.12) (0.05) (0.04) (0.05) Systemresearch 0.48*** 0.46*** 0.42*** 0.25*** 0.18*** (0.11) (0.11) (0.08) (0.04) (0.04) InstitutionFE + + + + + Observations 1,291 1,115 1,278 1,218 1,151 R-squared 0.34 0.35 0.34 0.29 0.24 Notes: Thistablereportscoefficientsfromregressionsoftheformdescribedinequation1. Thesampleconsists ofalleconomistsattheFederalReserveSystem,withatleast3yearsoftenureintheSystem,andspans2004to 2021. Thistableexaminestherelationshipbetweengenderandproductivity,asmeasuredbythenumberofworkingpapers(WPs). TableA.3showsthelistofresearchworkingpaperswithineachoftheFederalReserveBanks. ***,**,and*indicatestatisticalsignificanceatthe1,5,and10percentlevels,respectively. FEisfixedeffects.

TABLEA.7 GenderandProductivity-Peer-ReviewedPublications 1 2 3 4 5 Baseline Publications>0 Publications<30(99%) Publications<16(95%) Publications<10(90%) Female -1.71*** -2.43*** -1.39*** -0.77*** -0.40*** (0.28) (0.47) (0.23) (0.16) (0.09) Tenure 0.33*** 0.33*** 0.30*** 0.22*** 0.15*** (0.08) (0.10) (0.04) (0.03) (0.02) Policyreports 0.05 0.05 0.03 0.01 0.01 (0.03) (0.04) (0.02) (0.02) (0.01) Systemresearch 0.21*** 0.22*** 0.16*** 0.06** 0.04** (0.06) (0.07) (0.04) (0.02) (0.02) InstitutionFE + + + + + Observations 1,291 799 1,278 1,224 1,153 R-squared 0.29 0.28 0.30 0.23 0.20 Notes: Thistablereportscoefficientsfromregressionsoftheformdescribedinequation1. Thesampleconsists ofalleconomistsattheFederalReserveSystem,withatleast3yearsoftenureintheSystem,andspans2004to 2021. Thistableexaminestherelationshipbetweengenderandproductivity,asmeasuredbythenumberofpeerreviewed publications. Table A.3 shows the list of research working papers within each of the Federal Reserve Banks. ***, **, and*indicatestatisticalsignificanceatthe1, 5, and10percentlevels, respectively. FEisfixed effects.

TABLEA.8 GenderandProductivity-SystemResearchPublications 1 2 3 4 5 Baseline Sys. research>0 Sys. research<42(99%) Sys. research<17(95%) Sys. research<8(90%) Female -0.11 -0.59 -0.14 -0.14 -0.15 (0.39) (0.85) (0.37) (0.09) (0.09) Tenure 0.15 0.31** 0.11 0.05 0.03 (0.11) (0.14) (0.08) (0.04) (0.02) Policyreports 0.06 0.05 0.07* 0.03 0.02* (0.05) (0.06) (0.03) (0.02) (0.01) Workingpapers 0.06 0.05 0.07* 0.03 0.02* (0.05) (0.06) (0.03) (0.02) (0.01) InstitutionFE + + + + + Observations 1,291 580 1,278 1,221 1,146 R-squared 1,291 580 1,278 1,221 1,146 Notes: Thistablereportscoefficientsfromregressionsoftheformdescribedinequation1. Thesampleconsistsof alleconomistsattheFederalReserveSystem,withatleast3yearsoftenureintheSystem,andspans2004to2021. This table examines the relationship between gender and productivity, as measured by the number of System researchpublications. TableA.3showsthelistofresearchworkingpaperswithineachoftheFederalReserveBanks. ***,**,and*indicatestatisticalsignificanceatthe1,5,and10percentlevels,respectively. FEisfixedeffects.

TABLEA.9 GenderandProductivity-PolicyReports 1 2 3 4 5 Baseline PolicyReports>0 PolicyReports<52(99%) Policy<24(95%) Policy<13(90%) Female -0.10 -0.50 0.58 0.67** 0.40* (0.53) (0.88) (0.35) (0.23) (0.20) Tenure -0.10 -0.50 0.58 0.67** 0.40* (0.53) (0.88) (0.35) (0.23) (0.20) Systemresearch 0.14 0.18 0.02 0.01 0.02* (0.16) (0.22) (0.05) (0.02) (0.01) Workingpapers 0.14 0.18 0.02 0.01 0.02* (0.16) (0.22) (0.05) (0.02) (0.01) InstitutionFE + + + + + Observations 1,291 826 1,278 1,226 1,152 R-squared 0.26 0.25 0.32 0.29 0.23 Notes: Thistablereportscoefficientsfromregressionsoftheformdescribedinequation1. Thesampleconsists ofalleconomistsattheFederalReserveSystem,withatleast3yearsoftenureintheSystem,andspans2004to 2021. Thistableexaminestherelationshipbetweengenderandproductivity,asmeasuredbythenumberofpolicy reports. TableA.3showsthelistofresearchworkingpaperswithineachoftheFederalReserveBanks. ***,**, and*indicatestatisticalsignificanceatthe1,5,and10percentlevels,respectively. FEisfixedeffects.

TABLEA.10 SummaryStatisticsforHighlyProductiveSystemEconomists Highlyproductive Highlyproductive Workingpaperauthors Policyreportauthors Female 0.2 0.3 Tenure 10.7 10.5 Workingpapers 13.2 12.0 Systemresearchpublications 3.8 3.5 Policyreports 5.7 5.6 Publications 4.4 4.3

TABLEA.11 GenderandProductivityoverTime-WorkingPapersperYear 1 2 3 4 5 Baseline WPsperyear>0 WPs<4(99%) WPs<3(95%) WPs<2(90%) Female -0.10** -0.13** -0.05 -0.03 -0.02 (0.04) (0.05) (0.04) (0.03) (0.02) Tenure -0.10** -0.13** -0.05 -0.03 -0.02 (0.04) (0.05) (0.04) (0.03) (0.02) Policyreports 0.09*** 0.02 0.07*** 0.05*** 0.04*** (0.01) (0.03) (0.01) (0.01) (0.01) Systemresearch 0.16** 0.08* 0.11** 0.09** 0.04** (0.07) (0.04) (0.05) (0.04) (0.02) InstitutionFE + + + + + YearFE + + + + + Observations 12,996 4,604 12,698 12,277 11,057 R-squared 0.06 0.08 0.05 0.04 0.02 Notes: Thistablereportscoefficientsfromregressionsoftheformdescribedinequation2. Thesampleconsistsof alleconomistsattheFederalReserveSystem,withatleast3yearsoftenureintheSystem,andspans2004to2021. Thistableexaminestherelationshipbetweengenderandproductivity,asmeasuredbythenumberofworkingpapers(WPs)peryear. TableA.3showsthelistofresearchworkingpaperswithineachoftheFederalReserveBanks. ***,**,and*indicatestatisticalsignificanceatthe1,5,and10percentlevels,respectively. FEisfixedeffects.

TABLEA.12 GenderandProductivityoverTime-Publicationsperyear 1 2 3 4 Baseline Publications>0 Publications<3(99%) Publications<2(95%) Female -0.08*** -0.12** -0.05*** -0.03*** (0.02) (0.05) (0.01) (0.01) Tenure 0.01 0.01 0.01** 0.00** (0.00) (0.01) (0.00) (0.00) Policyreports 0.03*** 0.01 0.03*** 0.02*** (0.01) (0.02) (0.01) (0.01) Systemresearch 0.10** 0.09** 0.06** 0.04** (0.04) (0.04) (0.03) (0.02) InstitutionFE + + + + YearFE + + + + Observations 12,996 2,015 12,830 12,398 R-squared 0.05 0.06 0.04 0.02 Notes: Thistablereportscoefficientsfromregressionsoftheformdescribedinequation2. Thesampleconsists ofalleconomistsattheFederalReserveSystem,withatleast3yearsoftenureintheSystem,andspans2004to 2021. Thistableexaminestherelationshipbetweengenderandproductivity,asmeasuredbythenumberofpublicationsperyear. TableA.3showsthelistofresearchworkingpaperswithineachoftheFederalReserveBanks. ***,**,and*indicatestatisticalsignificanceatthe1,5,and10percentlevels,respectively. FEisfixedeffects.

TABLEA.13 GenderandProductivityoverTime-SystemResearchandPolicyReportsperyear 1 2 3 4 5 6 Sys. research Sys. research>0 Policy(baseline) Policy>0 Policy<4(99%) Policy<2(95%) (baseline) Female -0.02 -0.26** 0.03 -0.06 0.04** 0.02** (0.01) (0.10) (0.02) (0.07) (0.01) (0.01) Tenure 0.00 -0.00 0.01 0.03 0.00 0.00 (0.00) (0.01) (0.01) (0.02) (0.00) (0.00) Policyreports 0.05** 0.02 (0.02) (0.01) Systemresearch 0.11* 0.14* 0.05* 0.02** (0.06) (0.08) (0.02) (0.01) Workingpapers 0.05** 0.01 0.06* -0.01 0.04* 0.02** (0.02) (0.02) (0.03) (0.08) (0.02) (0.01) InstitutionFE + + + + + + YearFE + + + + + + Constant 0.10** 1.50*** 0.14 1.45*** 0.15*** 0.07*** (0.03) (0.13) (0.09) (0.33) (0.02) (0.01) Observations 14,635 1,481 14,635 2,162 14,419 13,756 R-squared 0.12 0.10 0.18 0.21 0.14 0.06 Notes: Thistablereportscoefficientsfromregressionsoftheformdescribedinequation2. Thesampleconsists ofalleconomistsattheFederalReserveSystem,withatleast3yearsoftenureintheSystem,andspans2004to 2021. Thistableexaminestherelationshipbetweengenderandproductivity,asmeasuredbythenumberofpolicy reportsandSystemresearchpublicationsperyear. TableA.4showsthelistofresearchworkingpaperswithineach of the Federal Reserve Banks. ***, **, and * indicate statistical significance at the 1, 5, and 10 percent levels, respectively. FEisfixedeffects.