Underrepresented Outperformers: Female Legislators in the Chinese Congress

Abstract This paper presents the first systematic study of the political behaviour of female members of China's national legislature, the National People's Congress (NPC). Women held 23 per cent of seats in the 12th NPC, yet they sponsored 44 per cent of all legislative bill proposals and more than half of the bills relevant to women's interests. Women sponsored more bills (4.8 bills) than did men (3.1 bills). We propose that there are two mechanisms driving women's outperformance: (i) women are more collaborative than men, and (ii) female leadership encourages female participation. We analyse 2,366 bills and show that women are disproportionately more active than men in all issue areas and are particularly engaged with women's issues. Our findings demonstrate that underrepresented regime outsiders (women) can carve out a space to amplify their voices, outperform insiders and shape policy direction to a certain extent within an authoritarian legislature.

While autocracies are usually characterized as being worse than democracies in terms of rights provision and political pluralism, they perform as well or even better than the average developing democracy in the area of female political representation. 1 Electoral autocracies are especially likely to adopt gender quotas. 2 The most notable example is Rwanda, an increasingly authoritarian regime which elected the highest proportion of female parliamentarians in the world in 2003 and became the first country with a female majority in the national parliament. 3esearch on gender and politics in autocracies and hybrid regimes has focused on elites' incentives to increase female descriptive representation.Autocrats promote female representation to bolster regime legitimacy, gain women's support and loyalty, improve their international reputations and to marginalize religious extremists. 4ne key debate in this literature is whether descriptive representation of women brings about substantive changes.Extant evidence is mixed, and findings appear to be highly case specific.In Morocco, for example, a gender quota in the parliament selects women who have powerful male relatives and has not resulted in meaningful policy change. 5In Jordan, smaller tribes nominate women candidates to enhance their electoral chances after gender quota adoption, but elected women councillors are sidelined and have little real power within the parliament. 6In other cases, women are able to find ways to exert some influence.In Pakistan, women legislators who have extensive networks are able to wage viable campaigns in the parliament. 7In the Ugandan parliament, female MPs consistently bring up women's issues in their plenary speeches. 8e add to this literature by introducing the case of China.Since the late 1970s, women have consistently held slightly more than a fifth of the seats in China's National People's Congress (NPC), which formally installed a gender quota of 22 per cent in 2008. 9Focusing on the 12th NPC (2013-2018), we show that although they only occupy about one-fifth of the seats and are less likely to hold leadership positions, women legislators sponsor 44 per cent of all legislative bill proposals and more than half of the bill proposals relevant to women's rights and interests.
In the remainder of this article, we use the term "bills" to refer to legislative bill proposals ( yi'an 议案).Among those who sponsor bills, women on average sponsor more bills (4.8 bills) than do men (3.1 bills).Thus, within the NPC, women are disproportionately more active and more productive than men.We therefore refer to these female legislators as "underrepresented outperformers." We explore two explanations to understand female legislators' outperformance.First, consistent with findings in various contexts, women in the NPC are more collaborative and thus more likely to build or join coalitions than men. 10 Second, using qualitative data and bill sponsorship information, we show that female leadership has a significant and positive effect on encouraging the voice of other female colleagues.This finding echoes studies in other contexts that show that female leadership encourages the political participation of other female politicians and citizens. 11inally, we examine the proposed bills' content and study women's substantive policy interests.The gender and politics literature suggests that women's shared experiences and social positioning (for example, women's place within the gendered division of labour in the household) result in gendered policy preferences. 12Women tend to be more oriented towards caring for vulnerable groups and innately concerned with the poor, the elderly and children. 13Women politicians are more willing and able to represent the preferences of female constituents, regardless of party identification. 14urther, women's status, in the labour market and in society more generally, could affect women's policy prioritization. 15We bring these insights to the Chinese context and find that not only are female legislators more engaged with gendered issues such as women's rights, children and healthcare, they also actively participate in other issue areas such as economics and finance, foreign affairs and the environment.
This paper contributes to multiple strands of literature on women in politics and beyond.First, it adds to the growing literature on female candidates and legislators. 16Similar to many other contexts, we find that female politicians in China are more productive and more collaborative than their male counterparts. 17Notably, this literature mostly investigates women's performance in electoral politics.We speak to this body of work by examining female politicians in a non-electoral context, where legislators and their parties do not face strong electoral pressure.Owing to the opaqueness of the legislator selection process in China, we are not able to study how legislators are selected and evaluated and so we are providing a largely descriptive account of their legislative behaviour.Nevertheless, we do find similarity in patterns of political behaviour between female legislators in China and in other contexts.
Our study also has implications for comparative politics beyond the study of gender.Many authoritarian regimes install institutions and promote norms that resemble those in liberal democracies.Although some studies have concluded that features such as gender quotas merely help autocrats to mimic the appearance of democracies but do not necessarily translate representation into political power for women, other studies investigate the circumstances under which a gender quota can address gender equality concerns in electoral authoritarian contexts. 18We join the latter group and show that female politicians, as strategic political outsiders, can exploit institutional features to amplify their voices and even to outperform regime insiders in an authoritarian context.

NPC as a Platform for Policymaking
The National People's Congress (NPC) is China's national legislature.Its members exercise state power by passing law and regulations as well as supervising the government and appointing and dismissing government officials.The full NPC meets once a year in the spring (usually March) for a period of about two weeks to deliberate and vote on bills as well as to approve budgets, work reports and leadership appointments. 19During the annual NPC plenary sessions, legislators are divided into small groups (usually in their delegation) to discuss the items on the agenda.
The Chinese Communist Party (CCP) monopolizes organized political power.NPC legislators are vetted by the CCP and are chosen each term (every five years) by legislators serving in the provincial-level congresses.The 12th National People's Congress (2013-2018) had a total of 2,987 legislators. 20Each province is allocated seats proportional to its population.The largest delegation is the People's Liberation Army (PLA) with 268 seats.The Electoral Law requires a specific number of seats to be reserved for women, ethnic minorities, various political parties, workers, and so forth.Overall, these legislators are not only "agents" of the state; they can also be "remonstrators" who bring group demands to the attention of decision makers. 21ny legislator can introduce and sponsor a legislative bill proposal (bill).Sponsoring a bill requires a significant amount of time and effort on the part of the sponsor or co-sponsors.The China Quarterly Sponsors, working with government agencies, scholars and other interested parties, hold public conferences, conduct research trips to different localities, collect first-hand data and participate in discussions with officials and representatives from the relevant fields. 22They also spend time and use their political capital to promote their proposed bills and to acquire the required 30 signatures from fellow legislators.Legislators who sign a proposed bill are referred to as "co-signers."A bill only requires one sponsor but sometimes multiple sponsors work across provinces to promote the same proposed bill, and these sponsors are referred to as "co-sponsors" in this article.
Bill proposals proceed through the NPC in three main stages.First, there is the agenda-setting phase, when the NPC and the State Council consider proposed bills drafted by legislators and other relevant agencies.Then comes the interagency consensus-building phase, when relevant ministries and agencies offer their opinions.The third stage is the leadership review phase, when the CCP leadership decides whether to move forward with promulgating the law. 23Bills that pass all three phases have already obtained endorsement from the ruling elites, so casting a "no" vote is usually not a viable option for legislators. 24Legislators are usually given little time to review the content of finalized bills and do not see the agenda until right before plenary sessions, although they may suggest points of improvement and alter the language and content of draft laws substantially. 25egislators can, however, choose to sponsor or sign certain bills on issues of their choosing as an "alternative voting process" to signal their policy preferences and advocate their policy agenda. 26ar from being ceremonial, legislatures in China provide a venue that allows some degree of interest representation, articulation and coalition building, both at the local level and the national level. 27Sponsoring a bill is a key way to articulate a policy position and push for policy changes, and is more than a performative act in the national legislature. 28Beijing does not dictate the process of the passing of a law or policy change, and individuals and agencies can exert a degree of influence and affect the outcome of legislative bills. 29Murray Scot Tanner describes two types of legislators, or what he calls "legislative entrepreneurs."The first kind patiently works to promote policy issues and legislative actions by raising leadership awareness of one specific issue, while a second type seizes on a forthcoming bill, which would be certain to pass, and tries to attach a rider amendment.Both types of legislators, working closely with relevant government agencies and ministries, try to influence law and policymaking. 30Even NPC "outsiders"for example, scholars, business owners and lower-level bureaucratsare sometimes invited to offer their input on drafts of bill proposals and to engage in lobbying. 31lthough the CCP tightly controls the NPC's legislative agenda, bills that pass the vetting process are not equal.Some bills are deemed to be more important or urgent than others.High-level CCP leaders may intervene and promote certain preferred policies, and policy entrepreneurs, including the NPC legislators, lobby behind closed doors to promote their position.During this process, government agencies and legislators build coalitions to advance their policy agendas and strengthen their bills.Favoured bills are more likely to be included in a five-year legislative plan and the annual legislative work plans, both of which are issued by the NPC Standing Committee (NPCSC).These plans prioritize proposed laws for deliberation and eventual passage. 32 Deng and Liu 2017.23 Tanner 1995.24 Lü, Liu and Li 2020.25 Truex 2020.26 Lü, Liu and Li 2020.27 For the local level, see Hou 2019; Manion 2015; for the national level, see Tanner 1995; Truex 2016; Yu and Yang 2021.28 Deng and Liu 2017.29 Tanner 1995; Truex 2020.30 Tanner 1995.  31  Deng and Liu 2017; Tanner 1995.  32 Truex 2020; Yu and Yang 2021.We next show that while female legislators are a minority within the NPC and occupy few leadership positions, they outperform their male colleagues in legislative productivity.

Descriptive Results
Of the 2,987 legislators in the 12th NPC, fewer than one-quarter (23.4 per cent) are women.We collected extensive biographical information for all legislators from government websites and other trustworthy sources and extracted the list of bills and bill sponsorship information from the NPC official website. 33According to the data, 33.8 per cent (N=211) of the legislators who sponsored at least one bill during this NPC are women.Co-sponsorship is common: 17.8 per cent (N=533) of all legislators co-sponsored at least one bill.Among co-sponsors, 35.7 per cent are female legislators and 64.3 per cent male.Proportionally, female legislators are more active than their male counterparts in sponsoring bills and in co-sponsoring bills during this period.
Table 1 presents descriptive statistics of legislators by gender.Among the NPC legislators, 13 per cent are ethnic minorities; 73 per cent are CCP members; 37 per cent come from the government; 30 per cent come from the business sector, and the rest come from public organizations, 34 the PLA, and other sectors; and 57 per cent hold a graduate degree.Table 1 shows that female legislators are fundamentally different from the male pool: female legislators are on average less likely to be CCP members, more likely to come from an ethnic minority group and more likely to have less 33 We extracted bill sponsorship information from the NPCSC communiqués (Quanguo renmin daibiao dahui changwei hui gongbao), which are published on the NPC website at http://www.npc.gov.cn/wxzlhgb/ccc/new.shtml.Note that the website only publishes the gazettes of the current NPC.The NPC website does not publish a list of suggestions or policy proposals, and we therefore do not focus on these documents.34 These include nongovernmental organizations, social service organizations, public schools and universities and public hospitals.
education. 35Male legislators tend to come from the government or the PLA, while women are more likely to come from public organizations.Importantly, female legislators are more likely than male legislators to sponsor bills: 30.19 per cent of all female legislators sponsored at least one bill, compared to only 18.05 per cent of male legislators.It is worth noting that when legislators do not sponsor bills, they can still act as bill co-signers.Our data only provide the sponsors of bills but not the names of co-signers.Bill sponsors undertake research, draft bills and mobilize support by inviting co-signers; co-signers only need to add their signatures.Each bill needs one sponsor and at least 29 co-signers. 36Therefore, there are more co-signers than bill sponsors.Given the large number of bills proposed and the total number of signatures needed, we can further confirm that bill sponsorship and co-signing are major and meaningful political activities in the NPC. 37mong legislators who sponsored at least one bill, women on average sponsored 4.8 bills while men sponsored 3.1 bills.In comparison, in the 60th Mexican Chamber of Deputies, women on average sponsored 6.9 bills while men sponsored 5.3 bills. 38In the United States House of Representatives, the average legislator sponsored 18 bills per congress between 1984 and 2004, and women on average sponsored about three more bills than did men. 39e next study what factors explain legislative productivity.The baseline specification is the following fixed effects model: where Y i,p is one of the two outcome variables: (1) whether legislator i in province p sponsors any bill during the 12th NPC, or (2) the total number of bills sponsored by legislator i in province p; Female i indicates whether legislator i is a woman; δ p represents province fixed effects; and ε i,p is the error term.X i is a vector of time-invariant controls for legislator i, including political party membership, ethnic identity, occupation (government, business, public organization, military), educational attainment, and political experience (whether a legislator is a first-time legislator; how long a legislator has served in the NPC).Table 2 presents the baseline results.In the first three columns, the dependent variablebill sponsorshiptakes the value of 1 if a legislator has sponsored any bill, and 0 otherwise.In columns 4 to 6, the dependent variable is the total number of bills a legislator sponsored during the 12th NPC.Models 1 and 4 are bivariate regressions without any controls or fixed effects, and models 2, 3, 5 and 6 add time-invariant controls and provincial fixed effects.Among all legislators, female legislators are 8 per cent more likely than men to sponsor a bill, and women sponsored 0.75 more bills after adjusting for observables such as occupation (government, business or public organization job), party affiliation (CCP or otherwise), educational attainment and political experience (whether one is a first-time legislator, or number of years in legislative office).
Political experience is positively correlated with bill sponsorship and productivity.Years in office is positively and significantly correlated with sponsoring more bills, and being a first-time legislator is negatively correlated with sponsorship and total number of bills, as expected.Being a CCP member is negatively correlated with being a sponsor and with the total number of bills.Being an ethnic minority legislator appears to be negatively correlated with sponsorship and productivity, although these effects do not reach statistical significance.In the online appendix, we show the results of robustness checks where we include newly added legislators and exclude legislators who are government and PLA officials, and results remain robust.Overall, we find a strong and positive correlation between legislators' gender and their productivity.Women legislators are more likely to sponsor a bill and, on average, sponsor a higher number of bills.
We next analyse the importance of bills by studying whether a proposed bill was later included in the NPCSC's five-year legislative plan or its annual legislative work plan, and whether there is any gender difference in bills included in these plans.In the first years of each new term of the NPC, the NPCSC issues a five-year legislative plan designating certain legislative areas or potential bills as priority issues that should be deliberated or require further research and investigation.The 12th NPC legislative plan (first issued in 2013, adjusted in 2015) included 102 priority bills.Besides the fiveyear plan, the NPCSC also issues an annual legislative work plan, which details and prioritizes bills that should be the focus of that particular year.Most bills included in the five-year plan appear in one of the annual plans.
Are bills sponsored by female legislators equally important as bills sponsored by their male counterparts?That is, are bills sponsored by women equally likely to be included in these legislative plans?Among all bills sponsored by individual legislators, 4.36 per cent were later included in the 12th five-year plan issued in 2015.Among female-sponsored bills, 3.94 per cent made it on to the five-year plan, while 4.70 per cent of male-sponsored bills were subsequently included in  3), the dependent variable is whether a legislator has sponsored a bill; in models ( 4)-( 6), the dependent variable is the total number of bills each legislator has sponsored.Robust standard errors are in parentheses.*** p < 0.01, ** p < 0.05, * p < 0.1 the five-year plan.The difference is statistically indistinguishable from 0 (top row, Table 3).A higher number of bills were included in one of the five annual legislative plans during the 12th NPC.Among female-sponsored bills, 10.24 per cent ended up in one of the five annual legislative work plans, while 10.85 per cent of male-sponsored bills were later included in an annual plan.
The difference is again statistically indistinguishable from 0 (bottom row, Table 3).Although we do not believe bill passage is a good measure for the importance of bills, we still test whether bills sponsored by women are more or less likely to be passed within the same term or by 2021. 40Results are reported in Table A.6: bills sponsored by women are slightly more likely to be passed faster than are bills sponsored by men.
In sum, women sponsor more bills not because they write and sponsor less-important bills or marginal bills.In line with scholars who use the inclusion of bills in legislative plans as a measure for the importance of bills, we find that bills sponsored by women are considered as important as bills sponsored by men. 41y are Female Legislators More Productive?We next explore potential mechanisms that could explain the outperformance of female legislators.The first mechanism is "collaboration."Using co-sponsorship data and qualitative evidence, we show that female legislators are more collaborative than their male counterparts.The second mechanism is a female leadership effect.Using qualitative evidence and bill sponsorship data, we show that female leadership has a positive effect on encouraging other women legislators in the same provincial delegation to be more active.
We first examine the possibility that women are more collaborative than men.Prior studies have established that women's style of work in politics is different from men's style of work, and that the "female style" is more collaborative.In the US, women legislators are more likely to act as facilitators, whereas men tend to use their power to control the direction of debates. 42Women were credited for working across the aisle, "passing the budget, the transportation bill, the farm bill" and ultimately ending the government shutdown in 2013. 43In Argentina, women legislators collaborate more because they are marginalized and benefit from collaboration. 44re women legislators more productive because they collaborate more?Qualitative evidence suggests that there is indeed a collaboration ethos, especially among women within delegations, where deliberations and network building regularly take place during and between NPC sessions.In her memoir, Zhao Yongqiu 赵咏秋, a two-term NPC legislator who sponsored 14 bills during her tenure, reveals how she formed close relationships with other women within her delegation as, during the NPC sessions, female legislators would wake up early to get ready for the morning sessions 40 Bill passage is dependent upon factors such as how complicated the issue is, public attention and the CCP's priorities (Truex 2020).41 For instance, see Tanner 1995; Truex 2020; Yu and Yang 2021.42 Kathlene 1994.43 Lawless 2015.44 Barnes 2016.together and sometimes stay up late to discuss their agendas in one of the women's hotel rooms. 45he was especially "emboldened" by other outspoken female legislators when she attended her first NPC meetings and decided to speak up soon afterwards during discussion sessions. 46As the NPC sessions usually overlap with International Women's Day (8th March), the secretariat of delegations usually organize activities for female legislators. 47These women-only activities provide opportunities for women to bond in an informal setting.Hu Pingping 胡平平, a female legislator who served for three terms in the NPC, also mentions in her memoir that female legislators would usually seek countersignatures and support from other women first before reaching out to male colleagues. 48o systematically test the collaboration effect, we study all bills in the 12th NPC and examine how legislators collaborated with each other.Because of an informal norm that the NPC discourages explicit cross-province co-sponsorship, legislators from different provinces sponsor bills with identical or very similar titles and submit them simultaneously as a way to collaborate.These similar titles are not mere coincidences: this type of interprovincial collaboration has been well documented and verified in interviews with legislators. 49The co-sponsors of highly similar bills collaborate behind the scenes to collectively push their agendas. 50In terms of timing, most proposed bills pass within five years. 51It is thus reasonable to study coalition building among legislators who serve in the same five-year term.To note, this kind of interprovincial collaboration is different from the above-mentioned intra-provincial collaboration, for which we do not have data.
Among all 2,315 non-delegation bills in the 12th NPC, 519 are unique bills, meaning that these bills share no identical or similar title with any other bills. 52The other 1,769 bills all share the same or a highly similar title with at least one other bill.If two bills refer to the same law and the same article (if available), we identify them as "identical bills" or "co-sponsored bills."The majority of co-sponsored bills have five or fewer co-sponsors, but several have 20 or more co-sponsors.We next focus on bills sponsored by individual legislators.About 13.7 per cent of the co-sponsored bills are sponsored by an all-female team, 19.4 per cent by an all-male team, and the rest by a mixed-gender team.Table A.4 shows the gender breakdown of these co-sponsored bill teams.Importantly, more than 80.6 per cent of all co-sponsored bills have at least one female legislator as a co-sponsor.Women collaborate with both women and men at a disproportionately high rate.
We then run the same set of regression analyses as in equation ( 1) but use the following dependent variables: (a) whether a legislator has co-sponsored as least one bill, and (b) the total number of co-sponsored bills that a legislator has co-sponsored.Table 4 reports the results.Model 1 includes all legislators, while models 2 and 3 only include those who have sponsored at least one bill (active legislators).We find that women are 5.8 per cent (model 2) to 8.8 per cent (model 1) more likely than men to collaborate with colleagues on average.Among all legislators who have sponsored at least one bill, women on average have 1.4 more bills that are co-sponsored with colleagues (model 3).These results, taken together, further demonstrate that female legislators are particularly collaborative when it comes to co-sponsoring bills across delegations.
We should point out that in these analyses, we are unable to identify which one of the co-sponsors is the principal sponsor.We are also unable to determine whether women are actively building coalitions or are drawn into them, if we define co-sponsoring a bill across delegations as a form of coalition building.Coalition building can occur within a province, within a line ministry across provinces, or within an informal network.Our data only allow us to identify coalitions across provinces, because we can only observe the names of bill co-sponsors (usually across provinces) but not the names of co-signers (usually within province).
Women may collaborate more than men for various reasons.Given data availability, we can only show that women collaborate among themselves and across gender, but we do not have an in-depth understanding of the rationale behind these collaborations.We do observe that women legislators, in general, exhibit wider policy interests than men, measured by whether their bills belong to more than one committee (Table A .14).When a policy issue is outside of a legislator's expertise, that legislator is more likely to reach out to other legislators who do have the expertise.We propose that women's wider interests could be a plausible explanation for why women are more collaborative.In future research, we plan to obtain more finely grained data on bill sponsorship and signatures as well as expand the timeline to earlier years to investigate how these collaborations and coalitions form and evolve.Detailed case studies are also in order to understand how a bill is initiated and how collaborations are formed.
The second mechanism we test is whether female leadership amplifies the voices of other women.In other contexts, the presence of female leaders increases the participation of other female politicians in relevant debates. 53Having female leaders improves perceptions of female leader effectiveness, increases women's political knowledge and inspires other women to engage in political activity. 54ccording to Zhao Yongqiu, the NPC legislator mentioned above, it was particularly encouraging to her when a high-level female politician was involved in delegation work.She recalls feeling thrilled when a female minister, vice-premier Wu Yi 吴仪, worked closely with her provincial delegation for two weeks during the NPC sessions. 55Motivated by this experience, Zhao went on to become an active legislator who sponsored multiple bills during her tenure.
In the NPC, discussions usually occur within the provincial delegations instead of in congressional committees.Therefore, we investigate whether the gender of provincial delegation leaders affects female legislators' participation.Each delegation has one chair and multiple deputy chairs.The whole delegation elects its leaders, and the provincial Party secretary usually becomes the delegation chair. 56Across 35 delegations, only two delegations have a female delegation chair, and 15 delegations have at least one female delegation leader.There are only 19 female chairs and deputy chairs across all delegations.Women constitute 11.3 per cent (=19/168) of the delegation leadership, a small number even compared to the overall female representation in the NPC.Provincial delegation leaders change each year, so the number of female leaders varies across years.For instance, the Guizhou provincial delegation had female leaders between 2013 and 2015, but not in 2016 or 2017.
We account for these changes in delegation leadership when studying how female leadership affects political behaviour.To understand the effect of female leadership on legislative behaviour, we estimate: where Y i,p,t indicates whether (a) legislator i in province p has sponsored any bill in year t, or (b) the total number of bills that legislator i in province p has sponsored in year t; Female i indicates whether legislator i is a woman; FemaleLeader p,t indicates whether province p has any female leader in the delegation in year t; δ p and γ t represent province and year fixed effects; X i is a vector of timeinvariant controls for legislator i; and ε i,p,t the error term.The main coefficient of interest is β 3 , the interaction effect of Female i and FemaleLeader p,t .In these estimations, the unit of analysis is legislator-year.Table 5 presents the results of these analyses.Having a female leader in a provincial delegation in a given year makes it significantly more likely that women from that delegation will sponsor any bill that year, as indicated by the positive and significant interaction effect between Female i and FemaleLeader p,t in model 1.Having at least one female delegation leader is associated with a 2.7 percentage point increase in the likelihood of a female legislator sponsoring a bill in a given year, or a 31.9 per cent increase from the baseline.
We do not claim that provincial female leadership is randomly distributed after controlling for province fixed effects.Taken together, these findings provide only suggestive evidence of a positive female leadership effect on female legislators' political activities.Nevertheless, these results are consistent with our qualitative evidence that female legislators can be inspired and encouraged by female leaders they interact with and then go on to become more politically efficacious.
Our mechanism discussion does not rule out other possible explanations, including the possibility that women have to be of higher quality (for example, harder working, more highly qualified, etc.) 53 Blumenau 2021.54 See Schuler 2019; Beaman et al. 2009; Wolbrecht and Campbell 2007.55 Zhao 2012, 6, 20.56 See more on delegation composition at https://www.guancha.cn/WangWeiJia/2014_03_03_210207.shtml.Accessed 7 July 2020.
The China Quarterly than men to make it to the legislature in the first place. 57If education is a good proxy for quality or competence, our data do not suggest that women legislators are of higher quality than men: male legislators on average have higher educational attainment.But, as Table 2 shows, educational attainment is not a strong predictor of legislative productivity.One study finds that Chinese government officials rate men and women similarly in terms of their "leadership quality" and "task implementation," and both can be seen as measures of competence or quality. 58In sum, we cannot rule out the hypothesis that quality differential explains the gender gap in productivity, but a better metric for legislator quality is required to test this hypothesis.Another possibility is that women have been marginalized in other, and arguably more central, positions of power in the state and in the CCP, and women's lower position can affect the nature of their legislative behaviour. 59Admittedly, legislators in positions of power might exhibit more caution and be more reluctant to initiate a bill than those who occupy a less powerful position: sponsoring a bill can be risky because the ruling elites may not have reached a consensus on certain policy issues. 60Tendencies to avoid unnecessary risk may limit the potential for powerful (and more likely male) legislators to publicly support a policy by sponsoring a relevant bill; they can find proxy fighters who may be willing to sponsor a bill on their behalf.This dynamic may have provided participatory opportunities for female legislators.

Legislative Bill Content
Next, we study the content of legislative bills and pay particular attention to female legislators' substantive policy interests.NPC bills are assigned to nine different congressional committees based on the issue of the bill. 61Among the committees that considered more than ten bills, female legislators sponsored more than 37 per cent of the bills.In the education, culture and health committee, women contributed almost half of all sponsored bills.In all committees, female legislators sponsored at least 25 per cent of all the bills, which is slightly higher than their representation in the NPC, which is 23.4 per cent.Female legislators are particularly active in a variety of issue areas including education, culture and health, environment and resources, and internal affairs.Overall, female legislators are well-balanced in their interests.
A different way to categorize these bills is to look at whether they are directly related to women's rights and interests.Literature on gender and politics suggests that female public officials care about different issues to their male colleagues, 62 and that their priorities differ across dimensions such as regions, political parties and social status. 63Following this literature, we classify a bill as a "women's issue" bill if the title includes one or some of these key words: women, girl, child/children, campus, school, minor, birth, birth control, marriage, divorce, domestic violence, or family.Using this rather liberal definition, we identify 141 bills related to women's rights and interests, of which about 58.87 per cent are sponsored by a female legislator.It is a relatively small percentage (6 per cent) of the total number of bills. 64Figure A.1 shows the breakdown of bills related to women's interests by the gender of their sponsor(s).Regression analysis shows that women are 7 per cent more likely than men to sponsor a bill related to women's issues (Table A.11).
Alternatively, we can define bills related to women's interests by identifying topics officially specified by the Women's Federation (WF) as important, including marriage, sexual harassment, female employment, women's land rights, among other issues. 65We label these bills as "WF-related" bills.There are 89 WF-related bills in the 12th NPC, 64 per cent of which are sponsored by female legislators.
These patterns are consistent with the existing literature on gender and politics that shows that women tend to have more divergent political preferences and priorities than men, and that women might be more interested in issues related to women's rights, public health, poverty alleviation and other topics. 66But, unlike women in democracies, female legislators in autocracies face real limitations, especially if their targeted issues are considered to be politically sensitive. 67 Figure A.2 shows the distribution of co-sponsored bills.

Table 1 .
Legislators' Descriptive Statistics, by Gender Notes: This table compares female and male legislators in the 12th People's Congress selected in late 2012.
35 Table 1 also shows that the NPC members are skewed in terms of socioeconomic class, education and other dimensions.36 Bills co-signed by 10 NPCSC members can also be added to the agenda.37 More than 2,000 bills were proposed during the 12th NPC.38 Kerevel and Atkeson 2013.39 Anzia and Berry 2011.

Table 3 .
Percentages of Bills Included in NPCSC Legislative Plans This table reports the percentages of bills later included in the NPCSC's five-year legislative plan or one of its annual legislative work plans for the 12th NPC between 2013 and 2017.The last column reports the p value from two-sided difference-in-means t tests. Note:

Table 4 .
Legislator Gender and Bill Collaboration