CHENG Xiao-yu,WANG Jian-ying ,Kevin Z.CHEN,

1 Institute of Agricultural Economics and Development,Chinese Academy of Agricultural Sciences,Beijing 100081,P.R.China 2 School of Economics,Zhejiang University of Finance & Economics,Hangzhou 310018,P.R.China 3 China Academy for Rural Development,Zhejiang University,Hangzhou 310058,P.R.China 4 International Food Policy Research Institute,Washington,D.C.20005,USA

Abstract Decentralized methods for targeting poverty are widely adopted in developing countries to improve the performance of various poverty alleviation programs. A common challenge for implementing successful decentralized targeting is the existence of elite capture. China has recently implemented a nationwide decentralized poverty targeting program,the targeted poverty alleviation (TPA) policy,to achieve the national goal of eliminating absolute poverty by the end of 2020. As the largest decentralized poverty targeting program in the world,TPA’s successful implementation was believed to be threatened by elite capture in some earlier reports. Since 2015,a targeting correction mechanism,called “follow-up checks” policy,has been introduced. With the “follow-up checks” policy,the elites and other ineligible households who receive benefits under TPA were removed from the program. This paper investigates the elite capture phenomenon in TPA using village census data from a poverty-stricken county in 2017 -two years after implementing the “follow-up checks” policy. We find no evidence of elite capture in TPA. The elites are unlikely to become beneficiaries or receive more benefits than non-elites. Our results contradict earlier findings that reported elite capture in TPA. We argue that the reason is the accountability emphasized by the central government in the “follow-up checks” policy. Our findings imply that having proper accountability is critical for improving targeting performance by global antipoverty initiatives.

Keywords:elite capture,decentralized poverty targeting,targeted poverty alleviation,the “follow-up checks” policy,accountability

1.Introduction

Decentralized methods for targeted poverty reduction are widely adopted in developing countries. The central government usually delegates the task of selecting beneficiaries to the local government or nongovernment organizations (Galasso and Ravallion 2005; Groshet al.2008). The advantage of the decentralized method is that the local government usually has better local knowledge than the central government (Alderman 2002; Alataset al.2012),and they are more likely to be accountable to local people(Galasso and Ravallion 2005); hence,they help direct the resource flow to the poor. However,a common challenge of decentralization is the elite capture phenomenon (Dutta 2009),especially in poverty reduction programs,as these programs are often conducted in remote and highly unequal communities (Pan and Christiaensen 2012).

In 2013,China started implementing a nationwide decentralized targeting poverty program:targeted poverty alleviation (TPA). TPA aims to lift 83 million rural poor people out of poverty,which is defined as under the national absolute poverty line of 2 300 CNY per capita in 2010,by the end of 2020. The implementation of TPA heavily relies on a decentralized approach. The village often selects the primary beneficiaries list,and review is carried out by the township and county-level governments. As the largest decentralized poverty-targeting policy,TPA was reported to suffer from the elite capture phenomenon earlier in the program. Hu and Wang (2017) found severe elite capture in the early stage of TPA implementation. They used data from 2012 and 2013 and found that 25% of the beneficiary quotas were captured by local non-poor village cadres and the wealthiest households. As a result,the State Council Leading Group Office of Poverty Alleviation and Development (LGOPAD) initiated a “follow-up checks” policy in 2015 to correct the targeting errors. With this policy,ineligible households,such as those who owned a car/van or real estate in the city,were removed from the beneficiary list. In particular,the “follow-up checks” policy emphasized that households with any family member working as a village cadre should not be TPA beneficiaries. The policy finally removed 9.29 million ineligible people and added 8.74 million eligible poor people. The “follow-up checks” policy appears conducive to limiting elite capture in TPA. However,little empirical study is available on elite capture in TPA after the“follow-up checks” policy. This paper intends to fill this gap.

This paper contributes to the literature by providing new evidence on elite capture in TPA after the “follow-up checks”policy. First,we conducted an empirical analysis using the village census data from Puding County,Guizhou Province,in 2017. The total sample was 824 households from 17 natural villages.1One administrative village in China consists of several natural villages.Second,we explained how accountability reduces elite capture in TPA. Our village census data have two advantages. First,the data were collected in 2017. At that time,the “follow-up checks” policy had been implemented,which allowed us to examine the policy’s impact. Second,the census data include all villagers’information,including every elite and their relatives in these three administrative villages. This helps reduce the impact of missing information on certain people,such as elites,in the random sampling method widely used in the literature.

2.Policy background and literature review

2.1.China’s targeted poverty alleviation policy

To eliminate absolute poverty by the end of 2020,the central government has implemented the targeted poverty alleviation (TPA) policy nationwide since 2013. For TPA,accurately identifying poor people is a prerequisite. The government provided personalized assistance based on the identified causes of poverty to the poor. TPA is a special mechanism that accurately allocates poverty alleviation resources to the poor through institutional arrangements and policy support (Liet al.2017). TPA created archives for each beneficiary household,which is then referred to as “poverty registration”,and developed personalized measures to improve their living standards and development ability (Liet al.2017). The specific goal of TPA is to lift all rural poor people (82.49 million in 2013) out of absolute poverty according to the 2010 rural poverty line. The poor are guaranteed to have enough food and clothing and to have access to compulsory education,basic medical care,and safe housing under the so-called “two worry-free and three guarantees” framework. There are two phases of TPA implementation in rural China.

Phase 1:lnitial selection process for beneficiariesThe identification of TPA beneficiaries is a result of a combination of top-down and bottom-up processes in rural China. The total number of beneficiaries identified was 89.62 million poor people and 29.48 million poor households in 2014.LGOPAD mainly distributes the beneficiary quota through the top-down approach. First,LGOPAD set beneficiary quotas for each province and developed a method to distribute the quotas from province to county. The initial quota was calculated by the National Bureau of Statistics of China (NBSC) based on its survey data. Provincial governments distributed the quotas to each municipal government and then to each county government. NBSC provided the poverty incidence of each province,each city,and each county. The total rural poor population is equal to the number of rural populations multiplied by the poverty incidence. Second,the beneficiary quotas were distributed from the county to the villages. In this step,due to the lack of income information,it is difficult to calculate each village’s poverty incidence. Typically,LGOPAD allowed each village to independently determine the calculation method for poverty incidence.

The bottom-up approach is that the village mainly determined the preliminary beneficiary list in the village.A typical process includes five steps:poor households application,the applications are assessed through democratic voting or a participatory appraisal process,the preliminary beneficiary list is screened and published in the village,the township government appraises the list,and county TPA administrators conduct a final appraisal. Among these five steps,participatory appraisal in the second step plays a crucial role in deciding the potential beneficiaries(Wang and Guo 2015). Like China’s rural minimum living standard guarantee (Dibao) program,TPA’s participatory appraisal is carried out by a village committee,consisting of village chiefs,village’s Party leaders,no fewer than three county-level TPA administrators,and some village representatives (Han and Gao 2019). Therefore,the village officials; village chiefs; village’s Party leaders; village representatives,who are usually the natural village leaders;and other village cadres play a crucial role in determining the beneficiaries as they have better knowledge than the county administrators. However,since the village officials are given the power to decide the TPA beneficiary,they may manipulate the participatory appraisals to allocate TPA to themselves or their relatives.

Phase 2:The “follow-up checks” policyIn 2015,the National Audit Office (NAO) found relatively serious poverty identification errors in TPA through auditing. For example,3 119 ineligible beneficiaries were found in Mashan County,Guangxi Zhuang Autonomous Region,China. To improve the targeting performance,LGOPAD nationally implemented a policy to correct identification errors of poor households in rural areas,which was referred to as the “follow-up checks”policy,in August 2015.

The specific regulations and standards were promulgated by the provincial governments. Therefore,there are some differences among different provinces in the eligibility criteria. Two common eligibility criteria are (1) the per capita disposable income of poor households is lower than the national poverty line and (2) the “two worry-free and three guarantees” could not be reached for beneficiaries. There are some common ineligible criteria. Specifically,beneficiary households cannot (1) have any family member who is a village cadre,including the village chiefs,village’s Party secretary,or the village’s deputy Party secretary; (2) have any family member who has a formal job in the government;(3) own real estate in the town or city; or (4) own a car or a van. If the eligibility criteria conflict with ineligible criteria,this household is considered eligible. For example,if a village chief’s per capita income is lower than the poverty line,he/she can be a beneficiary,conditional on a rigorous review by the county-level government,and separate publishing in the village (Liet al.2017). Therefore,the policy appears conducive to restraining village officials’ behavior in selecting the beneficiary and eliminating elite capture. The policy finally removed 9.29 million mistargeted people and added 8.74 million eligible poor people. TPA identified a total of 55.75 million beneficiaries by the end of 2015 (Liet al.2017).

There is a dynamic adjustment of the beneficiary list. If the beneficiaries have been lifted out of poverty,they will be removed from the list. And households who fall into poverty and are not the beneficiaries will be added to the list.

2.2.Literature review

Elite capture refers to the resources allocated for the masses but received by a few usually politically and/or economically powerful groups (Dutta 2009). It is a typical feature of the decentralized program because it specifically occurs in lower-level governments (Bardhan and Mookherjee 2006a).More broadly,elite capture is defined as some who have great access to power and to influence the allocation of public resources or funds (Dutta 2009). The presence of elite capture leads to an ineffective allocation of resources.Public services tend to be over provided and undercharged to the local elite while charging the poor in a disproportionate share (Bardhan and Mookherjee 2006a),which is unfair and inefficient.

There are two types of elite capture discussed in existing studies,elite capture and the elite’s relatives capture. In essence,the elite’s relatives capture also reflects the abuse of power by local elites to benefit themselves rather than distributing resources according to the purpose of the project,which is also a kind of elite capture. In rural China,especially in poor and remote areas,village cadres are the elite. They are the carriers of state power at the village level,have the power to allocate public resources,and play an important role in many decentralized programs (Xing 2015). Therefore,in this paper,we define elite capture as,households with a family member as local village cadre or households with relatives as local village cadres,becoming beneficiaries or securing TPA benefits.

There is mixed empirical evidence in previous studies about elite capture in decentralized poverty programs. Pan and Christiaensen (2012) found that village officials captured 60% of voucher benefits in the input voucher program in Kilimanjaro. Platteau (2004) showed that there was very serious elite capture in a community-driven development project. On the other hand,Bardhan and Mookherjee(2006b) showed relatively little evidence of elite capture in intravillage allocations of private goods in West Bengal.Alataset al.(2019) found little evidence of elite capture among village elites and their relatives in aid programs in Indonesia. Moreover,Alataset al.(2019) also showed that welfare losses were minimal,even with elite capture.

Much of the literature suggests that elite capture is context-dependent (Bardhan and Mookherjee 2002; Fritzen 2007). Many pieces of evidence have shown that there is elite capture in the Dibao program (Li and Walker 2018;Han and Gao 2019; He and Zhu 2019),in public resource allocation (Xing 2014),and in microcredit projects (Wenet al.2016) in China. Hu and Wang (2017) showed that 25% of the beneficiary quota was assigned to the local non-poor village cadres and the wealthiest households in the early stage of TPA. However,there have been no quantitative studies examining elite capture in TPA after the “follow-up checks” correction policy.

A few papers discussed reasons for elite capture in decentralized targeting programs. These included information asymmetries and collusion (Platteau and Gaspart 2003; Dutta 2009) as well as the local political power structure (Bardhan and Mookherjee 2006b). Information asymmetry was found to be positively associated with elite capture in China’s context (Li and Walker 2018; He and Zhu 2019). Li and Walker (2018) found this effect in the early stage of Dibao. Because many community members did not have the relevant information,political elites directly put Dibao funds into their wallets. After the community members learned the relevant information and the government issued relevant regulations,the political elite capture phenomenon decreased significantly. However,elite capture still exists in Dibao. It was found that the collusion between the elite and other groups has captured the benefiting farmers’policy and excluded vulnerable groups in China (Li and Cao 2012).

Two methods were discussed in the previous literature to avoid elite capture,namely,good project design and accountability. A good project design can minimize the role of the local elite (World Bank 2004),empower the poor(Platteau 2004),and formulate relevant eligible criteria,therefore reducing the likelihood of elite capture. However,there is a trade-off between minimizing the role of the elite and using local better knowledge. Accountability can restrict local officials’ behavior,thereby avoiding elite capture(Bardhan and Yang 2004). Accountability is the ability to ensure that officials in government or public institutions have an obligation to explain and justify their decision(Bovens 2007). Bardhanet al.(2008) believed that political competition can improve local accountability and reduce elite capture. Galasso and Ravallion (2005) believed that the community is less accountable to the poor when the local officials do not have to share the program objective. Besleyet al.(2002) provided evidence from India that election accountability makes the central government ensure equal distribution of public goods,preventing elite capture.

3.Data source and description

In this paper,we used a census-type rural household dataset. The dataset came from the “Public Policy and Rural Poverty” project,jointly conducted by the International Food Policy Research Institute (IFPRI),the Chinese Academy of Agricultural Sciences,and Guizhou University,China in January 2018. The project was initially implemented in 2004. In the first step,the project selected Guizhou Province,one of China’s most prominent poor provinces.In the second step,Puding County,followed by a town within it,were randomly selected from the national povertystricken counties in Guizhou Province. In the third step,three administrative villages with high,middle,and low per capita income were selected through a stratified random sampling method. We then conducted a census in these three administrative villages. The total sample consisted of 17 natural villages and 824 households.

Puding County,in the western part of Guizhou Province,is one of the 592 state-supported poverty-stricken counties in China. Puding is in the “Dian-Gui-Qian Stony Desertification Area”,one of the 14 poor areas with special difficulties. In 2017,the GDP of Puding was 11.30 billion CNY,and the per capita GDP was 28 797 CNY,accounting for 48.27% of the national average. And the per capita disposable income of rural residents was 8 337 CNY,accounting for 62.07% of the national average. In 2015,the poverty population in Puding reached 69 100,and the poverty incidence was 15.1%(Guizhou Provincial Bureau of Statistics of China 2017),which was close to that of Guizhou Province (14.7%) (NBSC 2017) and rural western China (10.0%) (NBSC 2019).

Table 1 presents the characteristics of the selected three administrative villages. We calculated the poverty incidence and poverty gap with China’s rural poverty line.We measured the poverty gap with the Foster-Greer-Thorbecke (FGT) index (Fosteret al.1984). The poverty line in rural China is 2 300 CNY per capita per year at a constant price in 2011,which equals 2 952 CNY per capita per year after CPI adjustment in 2017 (NBSC 2018). Village 1 is a poverty-stricken village and largely a minority village.Minorities account for 83% of the total population,mainly are Miao (71%) and Buyi (12%). In 2017,the poverty headcount ratio of village 1 was as high as 20.3%. A total of 9.38% of households lived below 2 952 CNY per capita per year in village 2. Village 3 is in the county’s suburban area and had the lowest headcount ratio of poverty,only 4.24%. Village 1 had the deepest poverty level,with a 4.95%poverty gap,which was higher than those of villages 2 and 3. We define households registered as poor households in the government as archived poor households (APH).Among the 824 households,there are 84 APHs and 734 non-archived households.

Table 2 presents descriptive statistics for household characteristics in APHs,and non-archived households. The head of APHs was more likely to be older and to have lower education and poorer health. The average age of the head of APHs was older than those of the other households. The average schooling years of APHs head was 3.77,which was almost one year less than other households. A total of 21.69% of APHs head reported bad health status,and 4.82%of them reported very bad health status. The APHs had the characteristics of more migrant workers,fewer odd-job workers,more minorities,fewer assets,and smaller houses than the other households. The APHs had more land than the nonarchived household. Because there were some suburban households in the sample,27.91% of households had no arable land.

We define the elite as the household with family members who are local village cadres or the household with relatives who are local village cadres. We found that the elites were more likely to be the beneficiaries. The local village cadres accounted for 4.76% among the beneficiaries,and their relatives accounted for 14.29%. Both were higher than the nonarchived households. The average benefit of APH was 2 665.64 CNY.

4.Results

4.1.Are the beneficiary quotas captured by the elites?

We examined the possibility of the elites to be beneficiaries of TPA using the following equation:

whereidenotes the household.Paricipationiis the TPA participation variable measured in householdi.Paricipationi=1 indicates that the household is an APH,Paricipationi=0 indicates that the household is not an APH.Eiis a variable of the elite measured in householdi.Ei=1 indicates that the householdihas family members who are village cadres or householdihas relatives who are village cadres. Village cadres usually include village chiefs; the village’s Party secretary; the village’s deputy Party secretary;village representatives,who are usually the natural village leaders; and other village cadres.βis the parameter of interest.Xiis a vector of household characteristics.εidenotes error term. The regression was estimated using a Probit model and included the village dummy variables.

The estimated results are shown in Table 3. We removed a few households with missing household head information and finally have 817 households. We controlled the wealth status of the household by a wealth index. The calculation method of the wealth index was from WFP (2009). As there are many types of durables owned by households,we chose durables with an ownership rate of 3 to 95% and used principal component analysis to calculate the wealth index.

In column (1),we began by examining elite capture in TPA. We found that there is no significant evidence showing that the elite is more likely to become beneficiaries than the non-elite. In column (2),we explored the heterogeneous effect of the different kinds of elites. We added the village cadres and their relatives in eq.(1) to replace the elite variable. We found that village cadres and the household with relatives as village cadres are no more likely to become APHs than the other.

As we expected,households with higher wealth levels are less likely to become APH. Hence,our results also show that there is no economic elite capture,which is found in the early stage in TPA targeting (Hu and Wang 2017).Households with more land were more likely to become APHs,which indicates that the more land a farmer owned,the more impoverished they were. This may be because the quality of farmland owned by farmers in poor areas is generally poor,and cannot increase their income (Chenget al.2019). The more odd-job workers in a family,the less likely that family was to become an APH. Households with more odd-job workers had a larger available labor force and higher income and hence were less likely to become an APH (Liet al.2019).

The selection process of the beneficiary in TPA is similar to that in Dibao reported in Han and Gao (2019). Previous studies have found substantial elite capture in Dibao targeting (Han and Gao 2019; He and Zhu 2019). However,we found that there is no elite capture in TPA targeting.One of the differences between Dibao and TPA is that the TPA implemented a correction policy,that is the “follow-up checks” policy. Moreover,TPA had a serious elite capture problem in the early stage. Hu and Wang (2017) found that the elite captured 25% of the beneficiary quota in 2012 and 2013. Building on their paper,our findings show that no elite capture existed in TPA targeting in a 2017 dataset,indicating that the “follow-up checks” policy effectively controlled the political elite’s behavior and thus reduced targeting errors.

Table 1 The characteristics of the three administrative villages

Table 2 Summary descriptive statistics of nonarchived households and archived poor households

4.2.Are benefits captured by the elites?

The elite capture may occur in the benefit transfer process(Alataset al.2019). In TPA,the transfer to the beneficiaries is often conducted by village officials,which creates the potential for the elite to capture the benefit. Personalized support measures are required for TPA (Liet al.2017) and are usually developed by the county government (Xing 2015). We summarized the support measures of our sample county and presented them in Table 4. There were many kinds of support measures and subsidies to APHs,including providing newborn livestock,job opportunities,and subsidies in 2017 (Table 4). We calculated the transfer value in the elite and non-elite groups,respectively,and tested the significance of the difference in Table 4. The results show that the elite have more job opportunities and more subsidies to construct livestock pens than non-elite. At the same time,there is no difference in total transfer value and most assistances’ transfer value between the elite and non-elite.

We turned to empirically test the elite capture effect for the benefits using the following equation:

whereidenotes the household.benefitiis the TPA’s transfer value variable measured for householdi. The main explanatory variables were the same as in eq.(1). We first used the OLS regression to estimate eq.(2). As many households did not receive TPA transfers,many zero valuesin the regression. We used a Tobit model to estimate eq.(2) as well. The estimation results are shown in Table 5.

Table 3 Estimation results on elite capture of the beneficiary quota in targeted poverty alleviation policy

We found that elites are no more likely to receive benefits than non-elite. Alataset al.(2019) provided evidence that elite capture occurred in the actual benefit transfer process through experimental data from Indonesia. Our results show that there is no evidence of elite capture in the benefit transfer process in TPA. In TPA,many resources are emphasized to be allocated only to APHs. Hence,this strategy avoids the mismatch between the final recipient and the official beneficiary list. Even if elites can be beneficiaries,they may not receive more benefits than others.2There may be endogenous problems caused by missing variables. For example,the elites may be the rich in the village,while usually only the poor can be APH. This means that the effect of elite capture can only be identified as conditional on the level of wealth of the household. Although we controls “wealth” and “land” in all regressions,this may still be insufficient. We further explored the instrumental variable (IV) estimation and failed to find a suitable IV. Therefore,readers still need to be cautious that there may exist other missing variables.

5.Discussion

An interesting question is what part of the “follow-up checks” policy functioned to eliminate elite capture. The“follow-up checks” policy has three primary functions:clarifying ineligible criteria,speeding information,and establishing accountability (Yang 2017). Firstly,the policy clarified the eligible and ineligible criteria as we discussed in Section 2. And some counties developed a proxy means test method to identify the poor. For example,Weining County,Guizhou Province,has developed the“investigating four dimensions”,which includes living standards,agricultural production,demographics,and education. Secondly,the policy promoted the flow of TPA information in rural areas through extensive propaganda and emphasized the importance of public the beneficial list in the village (Yang 2017). Thirdly,the policy clarified the responsibilities of governments at all levels and established accountability (Yang 2017).

There are two types of accountability at the village level:accountability from higher levels of the government and election accountability. The former,which we defined as topdown accountability,plays an important role in elite capture in TPA. The election accountability is also called incumbent accountability,as village chiefs are democratically elected in rural China.

During the implementation of the “follow-up checks” policy,the central government realized that the lack of accountability leads to poor targeting. Therefore,top-down accountability had been introduced. Firstly,the clear responsibilities of the government at all levels have been listed. LGOPAD issued“Measures for the Implementation of the Responsibility System in Poverty Alleviation” in October 2016 (LGOPAD 2016). This regulation cleared and emphasized the responsibilities of all levels of government in TPA. The village cadres were responsible for doing the household surveys and convening democratic selection beneficiary meetings or participatory appraisal meetings. In practice,the village cadres were also responsible for allocating povertyalleviation resources. Secondly,the reward and punishment mechanism had been introduced. The central government issued the “Provincial Party Committee and Government’s Assessment in Poverty Alleviation Work” in February 2016.This regulation stipulated that LGOPAD annually evaluated the provincial implementation of TPA,including the targeting results. And the central government rewarded provinces with good performance while penalized provinces with poor performance.3For example,each of the eight best performing provinces was awarded 400 million CNY in 2017.Once the targeting error was found,the relevant responsible persons should be punished,including village cadres who did the household survey and submitted the preliminary beneficiary list (Yang 2017). Thirdly,the audit work conducted by NAO plays a role in monitoring.With the implementation of TPA,NAO has paid attention to audits of poverty reduction (Yan and Yu 2020). NAO issued audit announcements for poverty alleviation in 2015–2019.NAO focused on auditing poverty-stricken counties selected by sampling and reported targeting errors,including the relatives of village officials who captured TPA benefits.4For example,the NAO audited 158 poverty-stricken counties and reported targeting errors,one of them is the phenomenon of “relative”capture of the village cadres.Therefore,the audits effectively supervised the behavior of the village cadres. Also,some other measures had been introduced,such as third-party assessments and linking the responsibility of poverty alleviation with the promotion of government leadership.

Table 4 Assistance list and transfer value of targeted poverty alleviation policy 1)

We argue that accountability plays a critical role in eliminating elite capture. Firstly,Dibao and the early stage of TPA also conducted the clear ineligible criteria emphasized in the “follow-up checks” policy,but there existed the elite capture phenomenon. This means that these specific measures,clear ineligible criteria,and information publicity,are not the key to the success of the “follow-up checks”policy. Moreover,the implementation of these measures also relies on the village cadres. Therefore,how to ensure the village cadres’ execute correctly is the key to prevent elite capture. A strong accountability system can restrain the village cadres’ behavior,ensuring that they identify the poor and allocating the resources following governments’regulations and project design. As a result,top-down accountability plays a critical role in the “follow-up checks”policy to avoid elite capture.

How does accountability restrict the village cadres’behavior? Accountability is the ability to ensure that officials in government or public institutions have an obligation to explain and justify their decision (Bovens 2007). Exercising accountability usually involves the elements of monitoring and punishment. For example,the mass media can monitor election accountability and prompt government agencies to allocate resources more equitably (Besleyet al.2002).Due to the punishment for losing the leadership position,the incumbent will perform better,and political competition can improve accountability (Bardhanet al.2008). In short,accountability means that the government should be accountable for their decisions; otherwise,they should be punished for their misconduct.

Top-down accountability in the “follow-up checks” policy clarified the responsibilities of village cadres,set up a strict punishment,and conducted monitoring. On the one hand,accountability ensured that village cadres comply with the relevant regulations and requirements in targeting and resource allocation. On the other hand,accountabilityincreased the cost of village cadres capturing resources.Therefore,top-down accountability effectively constrains the behavior of village cadres and reduces the possibility of elite capture. The lack of top-down accountability explains why there was elite capture in Dibao and the early stage of TPA before the “follow-up checks” policy was enacted.

Table 5 Estimated results for elite capture of beneftis in targeted poverty alleviation policy

A related potential question is why the democratic election of a village leader fails to create accountability in the village.There are two reasons for this phenomenon:first,the local officials are often selected by the lineage within the village,which may lead to the elite and their relatives capturing the resources and excluding other groups (Mattingly 2016).Second,some village leaders were directly assigned by the township government because many people left the village to work in cities. Therefore,many village chiefs and village’s Party leaders have a very long term of service,which may undermine electoral accountability in the village,leading to elite capture. Hu and Wang (2017) found that the longer the service years of officials in poor villages,the more serious the phenomenon of elite capture.

6.Conclusion and implications

In this paper,we investigate the elite capture effect in the decentralized targeted poverty program in the context of TPA. In particular,we provide new evidence on the elite capture effect after the “follow-up checks” correction policy by using census data from three administrative villages in Puding County,Guizhou Province in 2017. We define elites as households with family members who are local village cadres or households with relatives who are local village cadres. There is little evidence showing that elites more likely became beneficiaries or captured benefits in TPA than non-elites after implementing the “follow-up checks” policy.This finding implies that the “follow-up checks” policy had a positive impact on the elimination of elite capture.

We argue that accountability has a vital role in eliminating elite capture through the “follow-up checks” policy. The role of “follow-up checks” policy is to clarify the criteria that made households ineligible or eligible and to reduce information asymmetry,but these were also introduced in Dibao and the early stage of TPA where the elite capture problem exists. Moreover,the implementation of these measures relies on village cadres. Therefore,how to restrict the behavior of village cadres is the key to eliminate elite capture. By establishing a clear responsibility,punishment,and supervision system from higher-level governments,the top-down accountability ensured the village cadres followed the program regulations and requirements and increased their capturing cost,restraining their behavior. As a result,top-down accountability led to success in avoiding elite capture.

However,we do not want to overstate the claims that the elite capture phenomenon has completely disappeared in TPA after the “follow-up checks” policy. The enforcement and supervision of accountability vary greatly in different counties and the study area that we focused on in the paper is only one poverty-stricken county. Therefore,we could not evaluate the effects of elite capture controlled by the “follow-up checks” policy of the overall TPA in rural China,which requires further researches with nationally representative data.

Our findings have important implications. Elite capture can be avoided in poverty targeting programs.Many decentralized projects restrict elite participation or marginalize elite roles through project design to prevent elite capture. For example,the “community-driven development” programs sponsored by the World Bank are often designed to restrict existing local leaders (Alataset al.2019). However,many projects need the local elite. The elite usually has better local knowledge than the project administrators,and their influence can help implement the project (Huet al.2019). Therefore,other effective ways to avoid elite capture should be explored. Accountability can play an important role in controlling the elite capture problem. By establishing a clear responsibility list,effective monitoring,and appropriate reward and punishment mechanisms,top-down accountability can ensure the local officials execute policies following higher-level government regulations or project administrators’ requirements,reducing the possibility of elite capture.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (71661147001 and 71603228),the China Postdoctoral Science Foundation (2018M630687 and 2018T110607),and the Zhejiang University-IFPRI Center for International Development Studies.

Declaration of competing interest

The authors declare that they have no conflict of interest.