Download Gender Inequality in COVID-19 Research: Analysis of Working Papers and Co-Authorship and more Study notes Economics in PDF only on Docsity!
Cambridge-INET Working Paper Series No: 2020/
Cambridge Working Papers in Economics: 2038
THE UNEQUAL EFFECTS OF COVID-19 ON ECONOMISTS'
RESEARCH PRODUCTIVITY
Noriko Amano-Patiño Elisa Faraglia
(University of Cambridge) (University of Cambridge)
Chryssi Giannitsarou Zeina Hasna
(University of Cambridge) (University of Cambridge)
The current lock-down measures are expected to disproportionately reduce women's labor productivity in the short run. This paper analyzes the effects of these measures on economists' research productivity. We explore the patterns of working papers publications using data from the NBER Working Papers Series, the CEPR Discussion Paper Series, the newly established research repository Covid Economics: Vetted and Real Time Papers and VoxEU columns. Our analysis suggests that although the relative number of female authors in non-pandemic related research has remained stable with respect to recent years (at around 20%), women constitute only 12% of total number of authors working on COVID-19 research. Moreover, we see that it is primarily senior economists who are contributing to this new area. Mid-career and junior economists record the biggest gap between non-COVID and COVID research, and the gender differences are particularly stark at the mid-career level. Mid-career female economists have not yet started working on this new research area: only 12 mid-career female authors have contributed to COVID-19 related research so far, out of a total of 647 distinct authors in our dataset of papers (NBER, CEPR and CEPR Covid Economics).
Keywords: COVID-19, Economics Research, Gender Inequality
Cambridge-INET Institute
Faculty of Economics
The Unequal Effects of Covid-19 on Economists’ Research
Productivity∗
Noriko Amano-Pati˜no†^ Elisa Faraglia‡^ Chryssi Giannitsarou§^ Zeina Hasna¶
May 7, 2020
THIS IS A VERY PRELIMINARY VERSION OF AN EVOLVING PROJECT
Please click here for the latest version
Abstract The current lock-down measures are expected to disproportionately reduce women’s labor productivity in the short run. This paper analyzes the effects of these measures on economists’ research productivity. We explore the patterns of working papers publications using data from the NBER Working Papers Series, the CEPR Discussion Paper Series, the newly established research repository Covid Economics: Vetted and Real Time Papers and VoxEU columns. Our analysis suggests that although the relative number of female authors in non-pandemic related research has remained stable with respect to recent years (at around 20%), women constitute only 12% of total number of authors working on COVID-19 research. Moreover, we see that it is primarily senior economists who are contributing to this new area. Mid-career and junior economists record the biggest gap between non-COVID and COVID research, and the gender differences are particularly stark at the mid-career level. Mid-career female economists have not yet started working on this new research area: only 12 mid-career female authors have contributed to COVID-19 related research so far, out of a total of 647 distinct authors in our dataset of papers (NBER, CEPR and CEPR Covid Economics).
∗We are grateful to CEPR for providing data on VoxEU columns, as well as data on submissions to Covid Economics: Vetted and Real-Time Papers †Faculty of Economics, University of Cambridge, noriko.amanopatino@econ.cam.ac.uk ‡Faculty of Economics, University of Cambridge, and CEPR, ef307@cam.ac.uk §Faculty of Economics, University of Cambridge, and CEPR, cg349@cam.ac.uk ¶Faculty of Economics, University of Cambridge, zh274@cam.ac.uk
cession caused by the current pandemic is likely to affect sectors that employ mostly women. Moreover, the still unequal division of labour inside households will determine a higher burden on working women due to home schooling and the increase in household duties. However, de- spite the gloomy short and medium-run forecasts, they envisage possible changes in culture in workplaces and in households that can mitigate the long run effects of the shock. Adams-Prassl et al. (2020) find evidence that, since lock-down measures started, women in the US, the UK and Germany are spending more time on active childcare and home schooling than men. The second strand of the literature this paper relates to, studies the different constraints and outcomes of female economists in academia over the last twenty years. Women in the economics profession are less productive due to the fact that referees hold women to higher standards and increase the length of the peer-review process (see Hengel (2017) and Card et al. (2020)). Female economists write fewer single author papers and prefer to maintain strong productive ties with a small circle of coauthors (Ductor et al. (2018)). Moreover, the same authors find that women at any seniority usually prefer to write with senior male colleagues. More in general, studies find that women are less likely to be promoted, even conditioning on productivity (Ginther and Kahn (2004)) and hiring tends to be biased towards men (Williams and Ceci (2015)). These and other constraints make female economists underrepresented in the profession reaching, for example, only 29% in the top 100 European economics departments (Auriol et al. (2020)). In the remainder of the paper, we describe in more detail our dataset as well as our findings, followed by some thoughts about possible explanations of this stark evidence.
2 The Data
In order to assess the different productivities of male and female authors we collect titles of papers and columns, names of the authors from four different sources. The first two are the National Bureau of Economic Research (NBER) Working Papers Series and the Centre of Economic Policy Research (CEPR) Discussion Paper Series. For each of these two, we cover the first four months of the year for the last six years (January to April from 2015 to 2020) to avoid seasonal research activity effects. Next, we use a novel dataset of all submissions to the newly established preprint of Covid Economics: Vetted and Real Time Papers, kindly provided by the CEPR, up until and including Issue 9. Last, we extract author seniority information from VoxEU columns in the first four months of the last two years (2019 and 2020).^1 In total, we analyse 2,361 papers with a total of 6,446 authors from NBER, 1,175 papers with 4,707 authors from CEPR, 129 papers with 240 authors from CEPR Covid Economics and 614 VoxEU columns with 1,611 authors. For more details, please refer to Table 1 in Appendix. To impute the gender of each author, we used the genderize.io API which assigns gender and a probability indicating the certainty of the assignment. All names that were assigned a gender with a probability less than 70% were checked and assigned manually. Additionally, we checked manually a random sample of 10% of the names with an assignment with probability above
(^1) We included VoxEU columns published until and including 27 April 2020.
70%, to make sure that their genders were correctly imputed. For the contributors of VoxEU columns, we assign seniority by defining five seniority cat- egories: Ph.D./postdocs, junior, mid-career, senior and others. Junior academics include As- sistant Professors and Lecturers, mid-career academics encompass Associate Professors, Se- nior Lecturers or Readers, and senior academics are composed of professors, full professors or Chaired positions. For non-academic researchers, We use standard metrics of seniority. The last category, “others”, is used for all authors who could not be easily classified.^2
3 Preliminary Findings
3.1 CEPR and NBER working papers
When we consider the CEPR and NBER working paper series, in the first four months of 2020, there are 798 distinct working papers, of which a large share appeared in March and April
- In contrast, the four-month average for 2015-2019 has been 606 papers, highlighting a substantial increase (of 31.3%) in research activity in recent months. At the same time, the proportion of women contributing to these papers in the four-month window has remained comparable to the last five-year average at approximately 20%. Figure 1 shows these results including the total number of working papers submitted, the total number of authors and the number of distinct authors. The first author measure, captures productivity by weighing authors’ output by the number of papers they have contributed to. The second measure gives equal weight to every author. We find that the fraction of women measured by either of the two measures is roughly the same (see Table 2 in the Appendix). Looking at the last four months, we identify the papers that have been submitted recently and classify them according to their topic: non COVID-19 and COVID-19 related research. The results are summarised in the last columns of Figure 1. Interestingly, when we compare the figures of the last five years with the first four months of 2020, the relative number of female authors in non-COVID research has remained stable with respect to recent years. In contrast, women constitute only 12% of total number of authors working on COVID-19 research. These facts suggest that while female researchers have managed to continue working on ongoing research and keep up with the recent growth of research activity, they have been less willing or less able to contribute to the budding literature on the economics of pandemics. This seems to be in contrast to the findings of the editors of some prominent economic journals such as the Review of Economic Studies and the Journal of European Economics Association. Both editors find that there has been a decrease in paper submissions by women.^3 We should keep in mind, however, that the economists that publish in the NBER and CEPR working paper series are
(^2) This group includes positions such as Member of the European Parliament, Member of the Advisory Council, independent researchers, economists at Banks such as Goldman Sachs and so on for which drawing parallels with academic seniority seemed difficult. (^3) Fuchs-Schundeln (2020) reports preliminary statistics from ReStud submissions in 2020 and Imran Rasul, editor of JEEA, tweeted on 24/04/2020 that ‘share of women submitters falls from 28 to 16%’
Figure 2: Breakdown of COVID research activity by source and gender
Note: Numbers in white are numbers of female authors. Source: VoxEU columns and authors’ calculations.
3.3 Co-authorship patterns
For a richer analysis of economists’ research productivity, we analyze co-authorship patterns of NBER and CEPR working papers between 2015 and 2020. We notice that women tend be more present in co-authored papers than in solo-authored ones, confirming the analysis of Ductor et al. (2018). Out of 433 solo-authored papers submitted in the first four months of the past 6 years, only 12.5% were by female authors. However, this number increases to 19.4% for papers with 2 authors, 20.3% for papers with 3 authors, 21.6% for papers with 4 authors, 20.8% for papers with 5 authors, and 26.5% for papers with more than 5 authors. In order to better understand the micro-underpinnings of the recent increase in economists’ research productivity, we focus on the first four months of 2020 and look into the co-authorship patterns of COVID-19 related and unrelated papers. Figure 3 shows stark differences in female shares across co-authored groups between the two research topics. Female economists con- tributed to 13.6% of the solo-authored non-COVID related papers, but had no contribution to COVID-related solo-authored papers. More interestingly, for every network size of co-authors, the contribution by females is higher in non-COVID papers than in COVID-related papers. Figure 3 also shows that papers written by two, three and four authors constituted the lion’s share of the research disseminated in 2020 (85% of non-COVID papers and 81% of COVID- related papers). While female economists constituted 20.9% of the 618 non-COVID related papers in these 3 categories, they constituted only 10.9% of the 59 COVID-related papers coauthored by two, three and four economists.
Figure 3: Co-authorship Patterns in NBER and CEPR WP in 2020
Source: NBER, CEPR, and authors’ calculations.
3.4 Seniority analysis
It is interesting also to explore another dimension of the data available: the seniority of the authors. Considering only the recent COVID-19 research in our dataset, we find that it is primarily senior economists who are contributing to the discussion of this research and policy area. Meanwhile, mid-career and junior economists record the biggest gap between non-COVID and COVID research, and additionally the gender differences are particularly striking at the mid-career level. Mid-career female economists have not yet started working on this new research area: only 12 mid-career female authors have contributed to COVID-19 related research so far, out of a total of 647 distinct authors in our dataset of papers (NBER, CEPR and CEPR Covid Economics). As a comparison, we analyze the seniority of the contributors of the VoxEU columns. VoxEU columns showcase new research or provide summaries and opinions about ongoing research. The authors of the columns do not have to be the authors of new research papers on the topic. This analysis could give us a proxy on how engaged economists are in new research discussions. Figure 4 shows seniority breakdowns for contributors to VoxEU columns in the first four months of 2019, as well as contributors to VoxEU columns unrelated to the pandemic, for the first four months of 2020. These numbers highlight even more the absence of junior, mid-career and female researchers from COVID-19 discussion.
continued working on non-COVID-19 research, but have been less willing to engage in this new research area. The few women who have started new research on the effects of the pandemic appear to prefer working in co-authorship groups more than their male counterparts and they collaborate with more senior colleagues on average. The current research and policy debate is more and more turning focus on the unequal impact of lock-down measures on the households and the general economy. Newspapers are inundated with articles describing the struggle of parents who have to juggle a career and a family in these non-ideal conditions.^5 In our analysis junior and mid-career female economists are more likely to be heavily involved in both professional and administrative (i.e. non-research related) duties, while also probably tending to families with young children during the pandemic lock-downs. However, the economic literature provides additional explanations in the inherent differences between the research approach of female and male academics. Female academics are more risk averse and may be less willing to start new projects in such a short notice, especially if the topic is inter-disciplinary.^6 All the above may preclude them from investing in new high-cost research as opposed to finishing off work-in-progress, for which most of the fixed cost had been incurred already before the pandemic. This is particularly noticeable from the absence of female authors in single authored COVID-related research in our dataset. Moreover, a recent volume published by VoxEU (Lundberg (2020)) offers a summary and explanations of known factors that contribute to the productivity gap between female and male economists. In particular, it is generally known that female academics tend to produce fewer papers than men. Hengel (2017) argues that female economists face a trade-off between quality and quantity due to higher standards demanded of female authors. This makes them spend more time on reviewing and polishing older research than on generating new ideas, in line with the current experience of research on the economics of COVID-19. Ductor et al. (2018) in the same research show that female authors prefer co-authorships to single-authored papers. Men and women have different co-authorship networks. On the one hand, men have more groups of co-authors and produce more single-authored papers. On the other hand, women prefer to maintain the same small number of co-authors that is persistent over time.
5 Conclusion and work-in-progress
The COVID-19 crisis has spurred a fast-growing new field in economic research. Female economists have not been as fast in reacting to the new research challenge as their male coun- terparts, and this seems to be due to a combination of unfortunate coincidences. The adverse effects of lock-downs on division of labour at home may have been detrimental to the research
(^5) For example, Minello (2020) offers a humorous but vivid commentary on the life of the female academic during lock-down. (^6) The behavioral literature shows that there is a wide range of experimental and survey evidence that supports the view that women are less willing to take risk, and they are also more averse to uncertainty, than men (e.g. Eckel and Grossman (2008)).
activity of all parent mid-career economists and especially women. In addition, women are in- herently more averse to risk and undoubtedly more cautious about how they approach research. Therefore, it seems that whenever they can set aside time to work on their research, it is largely devoted to completing well-developed work within their comfort zone. Going forward, we will continue to update and analyze these datasets along the lines pre- sented herein for the next months and to include also more years predating 2015. The evidence shown here could be only a temporary shock and in the next months we may see a convergence to previous years averages. Alternatively, this may be a more persistent shock that will cause long lasting effects on the productivity gap. Moreover, we are looking at the effect on different research fields sorting our papers by JEL codes and for the NBER working papers, by research themes. This will allow us to determine if some economic fields have been more affected than others. The current analysis gives a preliminary understanding of the effects of this unprecedented shock on the gender productivity gap in general. This evidence could provide the first input to design policies that will address the consequences of this unequal shock. In the literature review, we have provided a long list of all the evidence available showing possible obstacles to female academics. It is possible that these circumstances will tighten these constraints even more, and have long-lasting effects on female career progression. These effects need to prompt policies that address and reverse these outcomes.
Appendix
Table 1: Summary Statistics NBER CEPR CEPR Covid Economics VoxEU Columns Avg[2015-2019] 2020 Avg[2015-2019] 2020 Total Accepted Rejected 2019 2020 Number of papers/columns 380 461 274 403 129 55 74 217 397 Number of authors 1026 1318 713 1144 240 118 122 571 1040 %Female authors 20.5% 21.2% 19.8% 19.5% 10% 6.8% 13.1% 22.6% 21.4% Number of non-COVID papers/columns 380 407 274 379 217 244 Number of authors 1026 1157 713 1061 571 643 %Female authors 20.5% 21.8% 19.8% 20.6% 22.6% 23.8% Number of COVID papers/columns 0 54 0 24 0 153 Number of authors 0 161 0 83 0 397 %Female authors 0 17.4% 0 4.8% 0 17.6%
Table 2: Comparing Two Female Authorship Measures in CEPR/NBER Working Papers
%Female out of %Female out of total authors distinct authors 2015 19% 20% 2016 18% 19% 2017 20% 21% 2018 20% 22% 2019 23% 23% 2020 21% 22% 2020 non-COVID 21% 23% 2020 COVID 12% 15%
