Reducing land fragmentation to curb cropland abandonment: Evidence from rural China

Reducing land fragmentation can theoretically curb cropland abandonment, thus ensuring food security. However, few studies have quantified this relationship using large-scale survey data at the household level. This study adopts a two-way fixed-effects (TWFE) model to examine effect of on abandonment nationally representative panel from China Rural Household Panel Survey (CRHPS). The set contains 15,138 households across 29 provinces in 2017 and 2019. We find that has significant positive with abandonment. mechanism analysis reveals is due high labor costs difficulties renting out fragmented land. heterogeneity indicates farmers poor human capital those living non-plain areas are higher risk abandoning their fragmentation. Furthermore, association between tends vary different rent-in scenarios. For instance, an increase number plots case not necessarily associated These findings conducive correcting underestimation role implications may be extended various countries. La réduction de la des terres peut théoriquement freiner l'abandon cultivées, assurant ainsi sécurité alimentaire. Cependant, peu d'études ont quantifié cette relation à l'aide données d'enquêtes grande échelle au niveau ménages. Cette étude adopte un modèle bidirectionnel effets fixes pour examiner l'effet sur cultivées représentatives national l'enquête par auprès ménages ruraux en Chine L'ensemble contient 15 138 dans et Nous constatons que une significative avec cultivées. L'analyse mécanismes révèle est aux coûts main-d'œuvre élevés difficultés location morcelées. l'hétérogénéité indique les agriculteurs ayant faible humain ceux ne vivant pas zones plaines courent risque plus élevé d'abandonner leurs raison terres. En outre, l'association entre tendance varier selon différents scénarios rente foncière. Par exemple, augmentation du nombre parcelles cas fermage n'est nécessairement associée Ces résultats sont propices corriger sous-estimation rôle peuvent être étendues divers pays. Cropland contributes increasing carbon storage (Kuemmerle al., 2011; Silver 2000), reducing soil erosion (Y. Liu 2012), restoring forest ecosystems (Chazdon, 2008). also impacts security (Queiroz 2014). When no longer used for farming, production area decreases, which could reduce overall supply (J. R. Campbell, 2015). global increasingly prominent context COVID-19 pandemic, frequent extreme weather events, ongoing conflict Russia Ukraine (Zheng 2023). Therefore, maximizing use maintain become many governments’ policy priorities. phenomenon occurs both developed developing Since beginning industrial revolution mid-19th century, European countries documented substantial (Mather, 2001). At 20th began occurring parts Americas Southeast Asian (Fox 2009; Rhemtulla 2009). After collapse socialism 1990s, former Soviet Union socialist Central Eastern Europe, such as Albania (Deininger 2012; Sikor 2009), Russia, Ukraine, Kazakhstan (Meyfroidt 2016), faced rate In 21st France (San Roman Sanz 2013), Portugal (Van Doorn & Bakker, 2007), Switzerland (Price 2015), Austria (Silber Wytrzens, 2006) severe. recent years, there been several countries, including (Deng 2019; Du et, W. Xu 2021; Yan 2016; Y. Zhang 2014), Chile (Díaz 2011), Argentina (Izquierdo Grau, According J. E. Campbell al. (2008), abandoned ranged approximately 3.85 million 4.72 km2 since century. complex multi-dimensional influenced by interaction environmental, economic, political, social determinants (Lambin Meyfroidt, 2010). Harsh natural conditions (e.g., steep topographies conditions), parlous economic state massive rural-urban migration low agricultural income), imperfect rental market failure distortion), unsatisfactory reform transfer rights limitation inheritance subdivision), among others, affect varying degrees Lieskovsky Lieskovska, MacDonald 2000; Prishchepov D. Zhou 2020; Specifically, critical driving force affecting (Q. 2018). little attention devoted exploring nexus well its mechanisms heterogeneous effects. To best our knowledge, only three quantitative correlation but reach conclusions, possibly difference sample size, cross-sectional nature difficult deal omitted variable biases. based 98 villages 2004, (2009) found evidence causes contrast, employing 1,796 farms 2005 Living Standards Survey, Deininger (2012) pointed 10% Like other witnessed From geographical perspective, Wang (2022) impact tenure 5479 observations Chinese Family Database 2017. They larger average plot less likely would abandoned. While most relevant paper study, they did employ latest fixed-effect models rule biases unobserved factors correlated Meanwhile, discuss how affects nor explore effects perspective endowment. while examined moderation rentals, further distinguish rent-out rent-in, although latter directly degree Given literature gap, we (CRHPS) 2019 alleviate potential bias possible estimation small samples. By comprehensively examining associations underlying mechanisms, deepen current understanding important factor changes plots, measurement core independent process when participate rental. More specifically, aims fill gap research answering following questions: (i) Does abandonment? (ii) If so, does (iii) Is consistent levels topographies? (iv) scenario, underestimated measured plots? applied economics, first address above questions context. Land hinders efficiency threatens China. context, governance Then, literature, analyze market, laying theoretical foundation empirical section. new It closely related scarcity irreversible population growth. today's intense discussion about dates back late 1970s or early 1980s, Responsibility System (HRS) was introduced China, profoundly influencing future pattern HRS prioritized fairness establishment. Before HRS, make own decisions production. Under collective system principle equal distribution production, farmers’ enthusiasm severely hindered, leading agriculture stagnation long time. exploration grassroots Anhui province, officially recognized government gradually promoted nationwide. Although system, essence, assigned rather than collectives. Farmers gained greater autonomy over signing leases allocation under characteristics. First, allocates farmland each village according size force, population, combination two. initially obtain years rights. Second, allocated lottery usually careful consideration distance fertility ensure fair starting point (Tan 2006). As result, occupy sizes, shapes, locations. Third, births, deaths, migrations, marriages), carry redistribution every guarantee process. increased extra piece collective. decreased asked give share assures farmer detains roughly same amount short run. advantages disadvantages apparent. On one hand, it improve output initial stage implementation. aggravated hindered modernization fragmentation, implemented measures. adopted consolidation, direct solution widely consolidation implies promoting continuous management through engineering, biological, measures region, objectives determined general plan use. Starting mid-1980s, local governments some eastern coastal begun implementing projects, followed later central 1999, Management Law stated should encouraged. 2000s, successively National Plan Development Consolidation (2001−2010), (2011−2015), (2016−2020), all require comprehensive effective cropland. supported rental, emphasizes integration subjects market-based elements. Doing transformed, switching original small-scale decentralized operations operations. promote transfer, issued Administrative Measures Transfer Contractual Rights 2005, encouraging lease others voluntarily. 2014, Opinions Guiding Orderly Use Develop Appropriate Scale Agricultural Operations, separate into ownership, contracting rights, so better serve farmers. enacted prohibition redistribution, required collectives not, principle, redistribute during contract period, probability divided. Moreover, period 30 years. newly revised Contract People's Republic 2019, another 75 Finally, exchange, emphasizing exchange far away homes neighbors private consultations achieve farming. internal breakthroughs arable helped situation abundant people limited land, resolved suggests need continuously future. urbanization growth, root cause needs improvement. scattering, pulverization, parcellization (Bentley, 1987), generally defined single farm divided numerous spatially separated belong least property subject (King Burton, 1982; McPherson, van Dijk, 2003). common challenge 2015; Kuemmerle Theoretically, increases input restrains rent-out, making easily (Figure 1). cost hired input. rational farmers, more abandon (Lu 2019). Especially condition being will greatly 2012). main reason means small, fragmented, scattered 2006); limits access machinery technologies development infrastructure like irrigation drainage, resulting (Kawasaki, 2010; 2017; Ma 2022). addition, locations, transportation these relatively invest time effort cultivating even hire insufficient (Huang 2008; Lu 2018; X. Wang, Yamauchi, 2020). sense, hiring thereby lessees lessors unable unwilling continue (FAO, 2004). Compared large contiguous generate investment, (Latruffe Piet, 2014; Veljanoska, avoid unless multiple connect whole doing inevitably involves transaction (Jin Deininger, investment marginal challenging (Cao Even if willing cultivate free, sometimes interested income (Zhou Zheng Qian, rural lower price, offer free cannot (H. Xue Consequently, reasonable choice especially plots. Hence, rent (at prices), aggravating line X i t ${X_{it}}$ vector control variables head's age quadratic term education level (D. 2019), dependency ratio, off-farm ratio certification, value, characteristics φ ${\varphi _t}\;$ time-fixed effect, controlling macroeconomic situation, policy, time-varying do change δ ${\delta _i}$ depicts household-fixed preferences, capabilities, household-specific time-invariant factors. ε ${\varepsilon _{it}}\;$ error term, α0, α 1 ${\rm{\;}}{\alpha _1}$ , 2 ${a_2}\;$ coefficients estimated. proposed controls unobservable level, observable mentioned previous studies, helping correct caused variables. third part analysis, investigate behavior groups endowment, topography, whether respectively. grouping estimation. compare significance magnitude draw conclusions CRHPS, Finance Database.2 2011, CRHPS collected two country, five rounds surveys conducted, namely, 2013, 2015, 2017, hence, analysis. collects (excluding Xinjiang Tibet) country employs stratified, three-stage sampling method proportional (Qian, stage, counties 10 layers per capita GDP, randomly selected stratification. second communities county 20 community village. comprised 11,085 samples 527 communities, 11,306 776 communities. total form unbalanced panel, excluding missing information Note attrition data. Table reports descriptive statistics Regarding dependent variable, 12% respondents exhibit behavior. 5.49, suggesting severe Concerning characteristics, household's head 54, proportion heads (12.4%) school diploma higher. mean elderly children 35.1%, average, 22.5% working-age members engaged employment. An endowed 2.98 mu capita, unit value 15,243 RMB mu. 55.9% reported having obtained certificates, insecurity still concern. household’ earns 5,000 10,000 RMB. There nearly eight types farming operators surveyed 21.4% behavior, 16.4% Whether householder went (0 = no; yes) range disposable annual (1 1,000 RMB, 1,000-3,000 3 3,000-5,000 4 5,000-1,0000 6 15,000-20,000 7 20,000-30,000 8 30,000-50,000 9 50,000-100,000RMB) Based data, Figures depict Figure shows that, southeastern experiencing development, Guangdong Fujian, significantly. spatial shown (2019) Labor-force Dynamics indicating representativeness southwestern provinces, Chongqing, Guizhou, Guangxi. A overlap observed rates inland provinces. suggest tested above, TWFE estimate results. column (1), (land fragmentation), fixed effects, included. (2), included, (3), province×time exclude interference provincial policies risks (Ward Shively, models, regression 1% coefficient 0.005, 0.5% additional plot. rationale smaller requiring output, ultimately With regards results variables, highly specifications. focus (2) example. Age square −0.023 (p < 0.01) 0.0003 0.01), respectively, revealing U-shaped householder's below certain threshold (approximately 38 old), likelihood experience. point, surprising, makes sense aging decline technical physical capacity. finding alignment (2014). negative determinant contrast Schneider Geoghegan (2006). plausible explanation capable managing activities; (Yan 2016). positive, explained fact availability activities. engagement employment resources sustain existing certification negatively significantly Qian (2022); positively negatively, practice economy perform robustness checks, presented 3. instead defining dummy measure (i.e., ratio) ordered categorical one, two, three, depending partially, completely refers area. alternative columns (1) significant.