在高速工业化期间市场化与行政式农用土地再分配

__ _ _ _ _as VAo POLICY RESEARCH WORKING PAPER 2203 Market versus Property rights in China are moving in two different Administrative Reallocation directions. In some villages, of Agricultural Land private rights are secure and to some degree marketable; in a Period of Rapid in other villages, individual rights are increasingly Industrialization restricted and subject to more regulation and reallocation. Administrative reallocation Michael R. Carter tends to promote more equal access to land, but the price paid for the social insurance of land tenure may be forgone investment. The World Bank Development Research Group Rural Development U October 1999  Poi cY RESEARCH WORKING PAPER 2203 Summary findings Under commulnal farm production, there was ifttle households on the basis of economic and dermographic incentive to work hard: the communal system changes among households. In other villages, houLseholds guaranteed a livelihood, and there were few private gains were granted much greater imminunity against from additional efforts. The reform that introduced the redistribution of any sort. household responsibility system in China in the early Similarly, villages differed in the degree to which 1980s sharpened individual work incentives by assigning individual households could trade land among specific plots and the rights to residual income to themselves. Some villages did not regulate the practice, individual households. and others required village approval or prohibited land However, the household responsibility system left renral relationships. unresolved questions about the reallocation of land over Carter and Yao use simulated mnaximum likelihood time - questions that have become increasingly miethods to estinmate hybrid panel models of the important (for both efficiency and equity) with the rapid determinants of both marker-based and administrative growth of the nonfarm economy. reallocation of land. They also use them to estimate the Carter and Yao use household and village data to show insecurity-induced investment costs of market-based that the initially egalitarian distribution of land is reallocation of land. becoming more dispersed over time. They find that administrative reallocation responds to In wlhat has become a hybrid property rights system, in the increasing inequality but nonmarket reallocations some areas local village leaders (the cadre) were come at a significant cost in forgone investment. empower-ed to periodically redistribute land between This paper - a product of Rural Development, Development Research Group-is part of a larger effort in the group to study the determinants and impact of property f ights systems and land tenure regimes in the process of development. Copies of the paper are available free from the World Bank, 1818 H St-eet N_TW, Washington, DC 20433. Please contact Maria Fernandez, room MC3 -542, telephone 202 -473-3766, fax 202-522-115 I, Internet address mfernandez2 (@?worldbank.org. Policy Research Working Papers are also posted on the Web at http://-xsvw.worldbank.org/html/'dec/Publications/ Workpapers/home.html. The authors mnay be contacted at carterC aae.wisc.edu or yyaoCaccer.pku.edu.cn. October 1999. (35 pages) The Policy Research Working Paper Series disseminates the find:ngs of ivork in progress to enconrlge the exchange of ideas abont development issues. An obiectzr'e of tbe series is to get the findings otis qnickly, even iftbe presentations are less than foully polished. The papers carry the names of the authors anzd should be cited accordingly. The findings, interpretatiois, and coziclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the vieei of the hWo; Id Bank, .ts Executive Directors, or the countries they represent. Produced by the Policy Research Dissemination Center  MARKET VERSUS ADMINISTRATIVE REALLOCATION OF AGRICULTURAL LAND IN A PERIOD OF RAPID INDUSTRIALIZATION* Michael R. Carter Department of Agricultural and Applied Economics University of Wisconsin-Madison Madison, WI 53706 USA and Yang Yao The China Center for Economic Research Beijing University Beijing, China 100871 * The authors thank the World Bank and the Ford Foundation for their support of this research. We would also like to thank Klaus Deininger and Pedro Olinto for their extensive comments and help, and Weiping Chen for her research assistance. All views expressed in this report are the authors' responsibility and should not be attributed to the funding agencies.   MARKET VERSUS ADMINISTRATIVE REALLOCATION OF AGRICULTURAL LAND IN A PERIOD OF RAPID INDUSTRIALIZATION Section 1 Introduction The legalization and subsequent widespread adoption of the "Household Responsibility System" (HRS) in the early 1980s restored the peasant household as the primary production unit in Chinese agriculture. By assigning specific plots and residual income rights to individual households, the HRS sharpened individual incentives for the provision of careful labor and management. Many authors (e.g., Lin, 1992 and MacMillan, Whalley, and Zhu, 1989) have attributed the subsequent rapid growth of agricultural output to the fact that the HRS broke the syndrome of the iron rice bowl in which little labor effort was provided on the grounds that the prior communal production system guaranteed a livelihood and private gains from additional efforts were sparse.' While decisively individualizing residual income rights, the HRS more generally resulted in a hybrid property rights system in which use, transfer and reallocation rights were variously distributed between the individual land users and their villages that remained the formal custodians of land that legally remained the property of the state. As Liu et al. (1998) show, in some areas, local village leaders (the cadre) were empowered to periodically redistribute land between households based on economic and demographic changes among households, while in other villages, households were granted much greater immunity against redistribution of any sort. Similarly, villages differed in the degree to which individual households could trade land among themselves, with some villages not regulating the practice at all, whereas others required village approval or i There has, however, been some dispute about the severity of the labor incentive problems under the communal system and the extent the rapid agricultural of the 1980's was due to the HRS or the price reforms that shortly followed (e.g., see Puttennan, 1991).  simply prohibited land rental relationships. Dong (1994) has argued that the redistributive power of the cadre permits land to serve an important social safety net function, an argument ratified by Burgess (1997) who econometrically shows that the (presumably administratively maintained) egalitarian land distribution in China contributes to low levels of undernutrition. Kung (1997) presents evidence that peasants prefer this mixed to a fully privatized, redistribution-proof system. At the time of the HRS reform, most villages allocated land to households based on the workforce size, with adjustment sometimes made for the number of its dependents (Liu et al., 1998). Given this distributional mechanism, the initial post-HRS distribution of factor endowments-defined as the ratio of contract land to family labor-across households must have been fairly compressed. Moreover, with limited opportunities for off-farm jobs or for urban migration, there must have been relatively little scope for land and labor exchange between households. Even with limited exchange of these factors, and irrespective of the definition of tenure security and transfer rights, the intensity with which land was cultivated must have also been quite similar across households.2 While tenure insecurity may have dampened incentives for long term, fixed investment under the mixed HRS property rights regime (an issue to be investigated below), the productivity impacts of the secondary features of the HRS (e.g., limited transfer rights) must have initially been quite modest. However, the rapid growth and industrialization of the Chinese economy since 1980, with its concomitant withdrawal of labor from agriculture, would be expected to put new pressure on the fluidity and efficacy of the institutions that allocate rural land 2 Graynor and Putterman (199x) note that Chayanovian forces will tend to equalize factor proportions between households even when land is distributed on a need as opposed to a work capacity basis. 2  and labor. In this new environment, the absence of further administrative reallocation of endowments with no increase in private land or labor transactions, would imply an increasing dispersion in both the factor endowment and the factor intensity distributions. An increased dispersion in the latter distribution would signal a potentially costly (static) allocative inefficiency problem of under- (and over-) utilized land. Moreover, as Carter and Yao (1998) argue, in an environment of rapidly growing off-farm employment, imperfectly transferable rights may discourage investment by households that expect to exit or reduce their participation in agriculture in the near future, creating a sort of dynamic inefficiency problem as well. Punctuated by a slowdown in agricultural growth, these new challenges re-ignited debate over rural property rights and the efficacy of the HRS and its hybrid property rights regime. In theory, an allocatively more efficient economy could be achieved either by mechanisms that facilitate rental transactions between households in the face of an increasingly disperse endowment distribution; or, by mechanisms that reallocate distribution of land endowments such that no transactions are not required to equalize factor intensity production across production farms. Matching these theoretical possibilities, proposals for further reform in China range from those that look toward a more administrative mechanisms to adjust the endowment distribution, requiring a recentralization of property rights, to those that propose to rely on decentralized, market- mediated decisions and ask for the deepening of individual tenure security and transfer rights needed to make this mechanism work. While either market or administrative mechanisms could in theory resolve the new challenges posed by rapid industrialization, this paper turns to empirical data to 3  gauge how these mechanisms actually function in practice. Using panel data on a sample of 80 villages drawn from across rural China in 1988 and 1993, this paper tries to first clarify the extent of these new land tenure problems. Section 1 below first looks at the evolving distributions of endowments and factor proportions in production over the 1988 to 1993 time period. We find using a non-parametric Kolmogorov-Smirnov test that we cannot reject the hypothesis that these distributions have changed over time, and parametric maximum likelihood estimates show that the land-labor endowment distribution has indeed become more disperse over time, revealing the uneven impact of industrial employment growth and other demographic shifts. At the same time, we find that the that factor intensity distribution (in rice production) has, if anything, become slightly less disperse over this time period, indicating that either centralized or decentralized mechanisms are at least keeping pace with the challenge of demographic and industrial change. How then do these two competing mechanisms work and at what cost? Addressing this first issue, Section 3 explores how and how effectively administrative and market-mediated land reallocations respond to evolving opportunities to improve the efficiency or the equity of the rural economy. Simulated maximum likelihood methods are used to estimate a hybrid panel data tobit model of land reallocations and land rentals. Our strongest finding is that administrative reallocations seems to respond primarily to increasing dispersion in endowment ratios (i.e., to equity concerns) and hence can be seen to play a social insurance function. 4  Section 4 then turns to see if this social insurance is indeed costly, using data on rice-producing households to explore whether or not the tenure insecurity implied by centralized powers of administrative reallocation significantly dampen investment incentives. Using panel data methods to control for the likely correlation between tenure regime and unobserved factors that influence returns to investment, the analysis in this section shows that tenure insecurity significantly reduces investment. Whether or not this price is one worth paying for this insurance is an issue that will require additional research, as discussed in the paper's final and concluding section. Section 2 Changes in the Factor Endowment and Factor Intensity Distributions, 1988 to 1993 A well-functioning economy that obtains maximal output from the resources that it utilizes will allocate inputs such that their marginal productivities are equal across different uses and production units. Were this not the case, total output could be increased by shifting inputs from lower to higher productivity locations. Under fairly general assumptions about the nature of technology, dispersion in the ratio of labor allocated to production per-unit area cultivated is a good indicator of the degree to which an agricultural economy has succeeded in equating factor productivities across farm units. While it is unlikely that any real world economy will perfectly equalize factor intensity across households,3 panel data from a sample of households that were surveyed in 1989 and again in 1994, permits us to explore whether the dispersion in factor proportions has increased over time. These same data also provide a window through 3 In measurement sense have trouble with unmeasured factor quality differentials that may make it efficient to have different factor proportions. Also expect some time of adjustment such that at any moment of time be some variation. 5  which to view how the endowment distribution (defined as the ratio of household contract land per-family member) has evolved under hybrid property rights and the pressure of rapid industrialization. The overall survey covered 800 households spread across 80 villages drawn from 4 provinces (Henan, Jiangxi, Jilin, and Zhejiang).4 Table 1 gives some descriptive indicators drawn from the approximately 400 surveyed households that primarily cultivate rice. As can be seen, as measured by the coefficient of variation, both the endowment and factor proportions distribution became more disperse for the sample as a whole, with the coefficients of variations for these two distributions rising from 74% to 100%, and 46% to 50%, respectively. Some of this increased variation or dispersion in these distributions may well result from differences between villages. For example, we might expect that more rapidly growing regions will experience an increase in wage rates and a matching fall in the labor intensity of rice production. In order to distinguish this source of increased variation from that occurs because of, say, faulty allocative institutions within the local village economy, Table 1 also presents an intra-village coefficients of variation for the endowment distribution defined as: nl, (2-1) it1 = iht,Th -i,)2 In,)If,, h=l 4 The data used in this study come from two comprehensive surveys administered in eight provinces of China in 1988 and 1993 for studies of land tenure evolution and its implications to agricultural productivity. Both surveys have a household and a village questionnaire. Villages and households were chosen from the Rural Survey Base maintained by the Rural Survey Team of the State Statistical Bureau of China. A household questionnaire was administered to ten randomly selected households in each village and asked questions ranging from land rights, land transactions and annual agricultural production to off- farm employment. The village questionnaire asked questions about land tenure arrangements at the village level both in the surveyed years and in the history. In order to eliminate extraneous noise, the analysis of investment and changing dispersion in factor proportions in production relies only on the subset of villages where rice is the predominant crop. 6  where -cit is the contract land-family labor ratio for household h in village i in year t; n,f is the number of sampled households in village i in year t; and, the mean endowment ratio for village i in year t isfit = Xt*h, /n,, . Letting t denote the ratio of land to labor in rice h=1 production, a similar expression defines the intra-village coefficient of variation in factor proportions, oy'. As can be seen, the intra-village coefficient of variation of factor endowments rose from 43% to 54%, whereas the intra-village the same measure for the distribution of factor proportions actually declined slightly (34% from to 32%). The Kolmogorov-Smimov statistics reported in Table 1 test the hypothesis that the 1988 and 1993 distributions are statistically different from another-i.e., they test whether or not the difference in the distributions signaled by the coefficients of variation are large enough that we can reject the hypothesis that the 1988 and 1993 samples were drawn from the same population distributions.5 Table 1 reports the value of the Kolmogorov-Smirnov test statistic and the figure in parenthesis gives the probability of randomly generating a test statistic of that size under the maintained hypotheis that the two distributions are identical. At a 5% significance level the data thus reject the hypotheses that the 1988 distributions are identical to the 1993 distributions for both endowments and factor proportions. However, we cannot reject the hypothesis that the intra-village factor proportions distribution was the same in 1988 as it was in 1993. There is, however, stronger evidence that intra-village distribution of factor endowments 5 The distribution free Kolmogorov-Smimov test compares the empirical cumulative density functions for two distributions, asking if the largest difference that occurs between the two functions is so large that it is unlikely to have happened based on random draws from identical distribution. 7  has changed as we can reject the hypothesis of identical distributions at the 11% significance level. While these tests confirm which distributions have changed in a significant way, they provide no information on the nature of the change nor its efficiency implications. In order to gain a more precise (and visual) idea about the nature of these evolving distributions, we assume that the factor endowments and factor proportions are distributed according to a gamma distribution: f (Tht) r(at t, I Pt, (2-2) f(h ) h I-(at" Pt') wheref(.) is the probability distribution function for the distribution of factor endowments, sit, and factor intensity, tit for household h in period t. The gamma distribution has been chosen for this analysis because it is a flexible distribution form that can take on a wide variety of shapes, from normal to skewed to exponential, depending on the value of the parameters of the distribution. In addition, in order to explore differences between the 1988 and 1993 distributions, we specify the two parameters of the distribution to be linear functions of time: (2-3) a, a0 + a,Dt where j= T,