1. IntroductIon¹

Clear and well-enforced property rights are essential for agricultural productivity and economic development. Increased tenure security incentivizes longterm investment (^{Demsetz, 1967}; ^{Alchian and Demsetz, 1973}; ^{Besley, 1995})² and can improve the allocation of production factors like land and labour.³ As economies grow, households with non-farming advantages often shift to emerging non-agricultural sectors (^{Duarte and Restuccia, 2010}; ^{Gollin et al., 2002}; ^{Do and Iyer, 2008}).⁴ However, in developing countries with weak property rights, high transaction costs and fear of appropriation hinder land trade and discourage investment, leading to inefficient labour and capital allocation and lower productivity. This issue significantly affects economic development and structural transformation.⁵

Why would we expect land property rights to affect households’ labour supply? ^{Besley and Ghatak (2010)} explain how in a context of weak property rights, labour effort can be used to maintain ownership of your property. We can think of tenure insecurity as a random tax on production, which discourages labour effort. However, if labour effort reduces the risk of expropriation, farmers have an incentive to work on the land -possibly more than the optimal level- to maintain claims over their property.

Empirical evidence has found that tenure security of residential homes can affect household’s allocation of labour. In the context of Peru, ^{Field (2007)} shows that granting title deeds for urban squatters increased their labour outside their home, and allowed families to substitute child labour for adult labour. She argues that the certification increased tenure security, which eliminated the need to spend human resources to maintain ownership of their home. ^{De Moura et al. (2014)} obtained similar results in the case of Brazil, where a property certification program increased the hours worked per week of adult members in the household. These findings align with the argument that greater tenure security reduces the need for household members to defend their property, freeing up time and energy to engage in productive labour outside the home (^{Galiani and Schargrodsky, 2010}). In both cases, the formalization of property rights alleviated the pressure on families to safeguard their homes, thereby allowing for a more efficient allocation of labor resources (^{Besley and Ghatak, 2010}).

Tenure security for productive assets like agricultural land has distinct implications from residential property. In developing countries, land ownership is often closely tied to usage. A Mexican land reform legally separated land rights from use, issuing household certificates and legalizing intra-community land transfers, which led to a substantial increase in migration as owners no longer feared expropriation when living elsewhere (^{De Janvry et al., 2015}; ^{Valsecchi, 2014}). Improved property rights also boost investments and promote land transfers (^{Besley, 1995}; ^{Goldstein and Udry, 2008}; ^{Markussen and Tarp, 2014}; ^{Do and Iyer, 2008}; ^{Kimura et al., 2011}; ^{Chen et al., 2021}; ^{Chari et al., 2021}). In Vietnam, tenure security increased plot-level investments in irrigation and soil conservation (^{Bellemare et al., 2020}) and, after the 1993 land reform, improved labor and land market efficiencies, allowing poorer farmers better access to land as wealthier households shifted to non-agricultural sectors (^{Deininger and Jin, 2008}).

In 1993, the Vietnamese government passed a law where it granted individual usufruct rights to rural households. The duration of these rights was set at 20 years for annual crops, and 50 years for perennial crops. When the rights of annual plots were set to expire, the Vietnamese government passed the 2013 Land Law in which, among other things, extended the usage rights period of annual crops from 20 to 50 years. This unanticipated change in policy improved all landowners’ property rights, but it was particularly beneficial for annual crop growers, whose rights were set to expire soon. This paper explores the effect of land tenure security caused by this policy change on farmers’ household labour supply. The main hypothesis is that the 2013 Land Law led to a reduction in labour supply as households felt more secure in their land tenure, allowing them to allocate labour more efficiently. Additionally, the complementary hypothesis examines the differences in labour supply impacts by gender and age, considering how these factors influence household dynamics. The underlying mechanism explores the roles of capital and land holdings, assessing how changes in tenure security affect labour allocation decisions across various household members.

Using panel data on over 2 000 rural Vietnamese farming households from 2008 to 2016, we find that the 2013 Land Law, which increased tenure security, led to a reduction in agricultural labor among annual crop growers, particularly female adults. These results are robust across specifications and align with the parallel trends assumption. However, we find no evidence that the law boosted off-farm participation. Two explanations are proposed: first, enhanced land rights may encourage households to adopt a more capital-intensive input structure, reducing labour demand in farming; second, greater tenure security may facilitate land market participation, leading to smaller operational scales and less labour. Findings indicate increased capital intensity and decreased landholdings among annual crop growers post-2013.

The paper contributes to the literature of property rights and labour supply in three ways: first, it uses a difference-in-difference approach to examine the distortions of tenure insecurity in labour allocation, and it sheds light on the intra-household allocation of labour by examining labour supply by age and gender. Second, it explores the possibility of a shift in the structure of input shares by suggesting improvements in tenure security can increase capital intensity in agricultural production. Finally, results can be useful for policy makers as they show farmers devote resources to maintain their usufruct rights. Arguably, removing the need to renew your rights as owner every given time period can have significant benefits for landowners.

The paper is structured as follows: section 2 describes the context of land tenancy in Vietnam. Section 3 presents the empirical strategy. Section 4 describes the data used in the analysis and presents some descriptive statistics. Section 5 discusses the empirical findings and we conclude in section 6.

2. Background: land reform In Vietnam

In the 1980s the Vietnamese government enacted a series of reforms to transform its agricultural sector from a collective production system to one based on household initiative. As part of these efforts, the Doi Moi reforms of 1988 sought to increase rural household’s land tenure security and began a process of privatization and decentralization of output and input markets. However, many restrictions on land usage and transfers remained until the enactment of the 1993 Land Law. This law contemplated the issuance of farm-level Land Use Rights Certificates (LURC), and the duration of those rights was set at 20 years for plots with annual crops and 50 years for plots with perennial crops. Moreover, the Land Law of 1993 allowed households to transfer, lease, mortgage and bequest the usage rights of their property.⁶

The formalization of private usage rights paved the way for the development of an active land market and the transformation of the rural economy. The land reform of 1993 increased agricultural investment and efficiency, it also granted better access to land for the poor and allowed rural households to pursue opportunities outside of agriculture (^{Pingali and Xuan, 1992}; ^{Smith, 1997}; ^{Do and Iyer, 2008}; ^{Deininger and Jin, 2008}; ^{Khai et al., 2013}).

The Land Law of 1993 has had two revisions since it was issued. The first was in 2003 to include two names in the LURC where land was jointly owned (for example, in the case of husband and wife) and allowed for certification to be conducted at the plot level, instead of the farm-level as the 1993 law stipulated.⁷ The second was in 2013 when, among other things, the usage rights of all land -annual and perennial- was extended to 50 years. Because plots devoted to annual crops were originally given 20 years of usufruct rights, we expect this change in the law to increase tenure security for annual growers while perennial crop growers must be relatively unaffected by the law since they had at least 30 more years of usufruct rights.

Under Vietnamese laws, any agricultural land plot assigned to a household before or in 1993 began its usufruct rights in October 1993, while land acquired later began its rights upon acquisition (^{Ravallion and Van De Walle, 2004} and ²⁰⁰⁸). When usufruct rights are transferred (via sales or rentals), their duration is not reset. For instance, a plot acquired in 1993 and sold in 2010 would leave the new owner only three years of rights. Rights can be renewed if there has been no illegal use; however, Article 38 of the 2003 Land Law allows recovery of land used inefficiently or for the wrong purposes, requiring land return upon expired rights without extension.

This paper posits that usufruct right duration effectively characterizes land tenure security, as legal protections apply only within the rights period. Annual crop growers, particularly those with land from 1993 (representing 75% of such growers), faced significant tenure insecurity as expiration neared.⁸ The 2013 Land Law unexpectedly extended rights for annual crop plots from 20 to 50 years, improving security more for these farmers than for perennial growers, whose rights were still valid. This differential impact is used to analyse the effects of tenure security on household labour supply.

3. Empirical Specification

As discussed in section 2, the 1993 Land Law set the duration of usage rights for annual land at 20 years and perennial land at 50 years. In 2013, when annual land usage rights were set to expire, the Vietnamese government passed the 2013 Land Law which, among other things, extended the usage rights period of annual crops from 20 to 50 years. This unanticipated change in policy improved all landowners’ property rights, but it was particularly beneficial for annual crop growers.⁹ To investigate the effect of this policy change on household’s labour supply, we compare the difference in household’s labour decisions for annual crop growers with perennial crop growers.¹⁰ The following linear household fixed-effect specification is estimated:

Where Y_ht represents the number of days per capita in own-farm cropping activities of household h at time t, ANNUAL is a dummy variable that takes value 1 if household h had annual crops at time t, and 0 otherwise. Law13 is a dummy variable that takes value 1 if year is greater or equal to 2014, γ is a province-year fixed-effect, θ is a household fixed effect, μ is a vector of household characteristics that may vary over time (e.g. age of HH head, titles land, etc.) and ϵ is the error term. The household fixed-effects specification allows us to control for unobserved time-invariant characteristics that could be related to labour supply. The province-year fixed effects accounts for any macroeconomic shocks (e.g. off-farm labour demand, crop prices, etc.) that affect household’s labour allocation. We are mostly interested in coefficient β₃ because it will capture the effect of the Land Law of 2013 on household labour decisions of annual crop growers (our treatment group).¹¹

This methodology is particularly well-suited for this analysis as it allows for a robust evaluation of the causal impact of the 2013 Land Law on household labour supply among annual crop growers compared to perennial crop growers. By comparing the changes in labour supply before and after the policy change between these two groups, our model controls for both observed and unobserved factors that could influence labour allocation, thereby isolating the effect of the policy itself. This method assumes that, in the absence of the treatment, the trend in labour supply would have been the same for both groups, allowing for a more accurate estimation of the treatment effect. Additionally, the fixed effects in our specification further strengthen this approach by accounting for any time-invariant characteristics specific to households, such as preferences or production capabilities, and province-year fixed effects mitigate potential confounding factors related to external economic influences, such as changes in labour market conditions or crop prices. Thus, our framework provides a powerful tool for understanding how enhanced tenure security through the 2013 Land Law specifically influenced labour supply decisions in the agricultural sector.

^{Bellemare et al. (2020)} use this dataset to analyse the 2013 Land Law’s impact on plot-level investment in annual land. Their findings show that, before 2014, investment trends for annual and perennial plots were similar. After the Law’s enactment, however, strengthened ownership rights for annual land led to positive impacts on irrigation and soil conservation investments. Following Bellemare et al. (2020), we use a Difference-in-Differences approach to measure tenure security’s effect on labour supply. Our identification strategy thus assumes that household labour supply trends for annual and perennial crop growers would have remained similar without the tenure security improvements for annual plots introduced by the 2013 Land Law.

To verify this in Figure 1 we show the number of days per capita devoted to own-farm cropping activities of annual and perennial growers across time. We can see that labour supply to own-farm agriculture is very similar until 2012, before the new land law came into effect. This illustrates that before the change in the law both annual and perennial crop farmer have similar labour decisions before the Land Law of 2013, however these deviates afterwards. We corroborate the parallel trends assumption by estimating equation (1) using only pre-treatment data (i.e. before 2014) and using 2012 as the treatment period (see the appendix for more details on robustness checks).

Note: number of days households devoted to own-farm cropping activities divided by household size across time. Households with no cropping activities are excluded from the sample. There are 8 068 annual observations and 867 perennial observations in the sample. The horizontal line depicts the enactment of the 2014 Land Law.

Source: own elaboration.

Figure 1 Household days devoted to own-farm cropping activities 

Tenure insecurity may lead households to allocate “excessive” labour to farming to maintain usufruct rights. Improvements in tenure security could reduce this need, freeing up labour for other activities and lowering agricultural labour input (^{De Janvry et al., 2015}; ^{Do and Iyer, 2008}; ^{Li, 2020}). However, tenure security can also boost investment and facilitate land transactions (^{Macours et al., 2010}; ^{Bellemare, 2020}), potentially increasing returns to farming and encouraging more labor in agriculture. In Peru, ^{Nakasone (2011)} found that tenure security increased labor on own farms. Thus, the impact on household labour supply is ultimately empirical and depends on the dominant mechanism.

For the analysis, we exclude households not in agriculture or that grow exclusively on rented land, as ownership rights are the focus. Anticipated rights expiration may cause households to switch crops; however, only 4% of annual plots convert to perennials, making endogeneity concerns minimal. We also address potential bias from rice-growing restrictions, which affect 38% of annual and 9% of perennial crop households. ^{Markussen et al. (2011)} suggest such households work harder and receive more support; thus, we reestimate equation (1) for households without restricted plots.

4. Data And Descriptive Statistics

Our data comes from the Vietnam Access to Resources Household Survey (VARHS), covering the period 2008-2016.¹² The VARHS is conducted every two years and is a representative sample of the rural population in Vietnam. We rely on the balanced panel sample which consists of 2 131 rural households that were surveyed in all the five waves collected in the period under analysis. The surveyed households are located in 476 communes across 12 different provinces. The dataset contains precise information on household demographics, economic activities, land use and agricultural production. Crucially for us, the survey reports the time spend in own-farm activities by each member of the household and detailed information on crop choice, input use and cultivated land at the plot-level.

Table 1 presents descriptive statistics of households growing annual and perennial crops, before and after the change in the land law. The number of households involved in the production of annual crops is significantly larger than those in perennial crops. Annual crop growers have on average larger households, lower incomes and more land per household member. Before the 2013 Land Law annual and perennial growers were equally likely to have their land titled, however after the change in the law perennial crop growers increased their certified land relatively more than annual farmers.¹³

For our identification strategy, it is crucial to consider that households may switch between growing annual and perennial crops. We document households exclusively cultivating either type: over two-thirds of annual crop farmers grow only annual crops, whereas only 21% of perennial crop farmers grow perennials exclusively. After the policy change, more households specialize in one crop type. Additionally, some plots face government restrictions on crop types. Approximately 38% of annual crop households and 9% of perennial crop households have restricted land in our sample, though this proportion declines following the land policy change for both crop types.

In Table 2 we explore the differences in household days devoted to agriculture (own-farm).¹⁴ In the first and second row we can see the average number of days per year devoted to cropping activities from all household members for the pre-reform period (years 2008, 2010 and 2012), for annual and perennial farmers. Before the 2013 Land Law, we notice that annual crop growers (households with at least one plot devoted to annual crops) spend more days in farming compared to perennial growers. However, after the policy change, annual crop growers reduced their average number of days in own-farm cropping activities, while perennial crop growers increased it.

Disaggregating labour supply by gender and age reveals notable patterns. In the pre-reform period, the difference in adult male labour days between annual and perennial crop farmers is small and not statistically significant. However, following the 2013 Land Law, male labour days increase significantly for perennial farmers and decrease for annual farmers. Among female household members, annual farmers previously worked 25 more days per year than perennial farmers, but this shifts post-reform, with women in annual-growing households reducing their farm labour while those in perennial-growing households increase it. For child labour, both groups reduce farm days, with the decrease being greater among annual crop growers.

There are two possible explanations for the reduction in own-farm labour supply. On the one hand it suggests that the “guard” effect of improved tenure security out-weights the “productivity” effect. The fact that women and children’s labour decreased after the Land Law suggests these households were devoting these labour resources to maintain usufruct right on their land.¹⁵ On the other hand, it could be that the increase in capital investments as found by ^{Bellemare et al. (2020)} leads to less labour resources being required to maintain productivity. We explore these channels in section 5.

5. Results

6. Conclusions

Weak enforcement of property rights distorts household labour allocation, forcing farmers to devote excessive resources to maintaining land ownership. This paper shows that Vietnamese farmers facing uncertainty about their usufruct rights experienced such constraints. The Land Law of 2013 increased tenure security for annual crop growers, eliminating the need for labour dedicated to guarding land.

Using detailed panel data from Vietnam, we find that increased tenure security reduced days spent on own-farm cropping activities, particularly for female adults, without increasing off-farm labour time, consistent with reduced guard labour. Two additional channels are explored: First, improved tenure security likely shifted input shares and increased capital intensity among annual crop growers, leading to reduced labour supply. Second, enhanced land rights enabled households to adjust their operational scale and reduced total landholdings, which also contributed to the decline in agricultural labour supply. Overall, this paper highlights the significance of clear, enforceable property rights for efficient labour allocation, demonstrating that farmers with enhanced tenure security reduced both their farming days and landholdings.

Our findings on Vietnam’s land tenure security can offer important lessons for Mexico and other Latin American countries, where insecure land rights continue to hamper agricultural productivity. In both regions, tenure insecurity limits investment in land and capital, leading to inefficient labour allocation. Similar to Vietnam, land reforms that formalize or extend property rights could encourage Mexican and Latin American farmers to invest in more capital-intensive practices, reduce reliance on manual labour, and improve overall productivity. For example, Mexico’s ejido system has often led to underinvestment due to unclear property rights. Reforms that enhance tenure security, similar to those in Vietnam, could encourage greater agricultural investment and labour reallocation (^{De Janvry et al., 2015}; ^{Galiani and Schargrodsky, 2010}).