Policy Research Working Paper 9520 Investment Impacts of Gendered Land Rights in Customary Tenure Systems Substantive and Methodological Insights from Malawi Klaus Deininger Fang Xia Talip Kilic Heather Moylan Development Economics Development Research Group January 2021 Policy Research Working Paper 9520 Abstract Compared with the vast literature on the investment and shows that (i) having long-term land rights of bequest and productivity effects of land rights formalization, little atten- sale has a significant impact on investment and cash crop tion has been paid to the impact of variation in individuals’ adoption; (ii) women’s land rights of bequest and sale, joint tenure security under customary tenure regimes. This is a with local institutional arrangements, can amplify the mag- serious gap not only because most of Africa’s rural land is nitude of such effects; and (iii) the effects found here can be held under informal arrangements, but also because gradual obscured by measurement error associated with traditional erosion of long-term rights by women and migrants is often approaches to survey data collection on land ownership an indication of traditional systems coming under stress. and rights. Using a unique survey experiment in Malawi, the analysis This paper is a product of the Development Research Group, Development Economics. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at kdeininger@worldbank.org; xia.fang.fx@gmail.com; tkilic@worldbank.org, and hmoylan@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Investment Impacts of Gendered Land Rights in Customary Tenure Systems: Substantive and Methodological Insights from Malawi Klaus Deininger†, Fang Xia‡, Talip Kilic# and Heather Moylan* 1 JEL Codes: C83, J16, O13, Q15. Keywords: Land rights, Respondent Selection, Gender, Agricultural investment, Malawi 1 † Lead Economist, Development Research Group, World Bank, Washington, DC, kdeininger@worldbank.org. ‡Associate Professor, Research Institute for Global Value Chains, University of International Business and Economics, Beijing, China, xia.fang.fx@gmail.com. #Senior Economist, Development Data Group, World Bank, Washington, DC, tkilic@worldbank.org.*Survey Specialist, Development Data Group, World Bank, Rome, Italy, hmoylan@worldbank.org. The views presented are those of the authors and do not necessarily represent those of the World Bank, its Executive Directors or the member countries they represent. 1. Introduction While there is a large literature on the investment and productivity effects of land rights formalization, the impact of individuals’ tenure security under traditional regimes has received little attention. This is a key knowledge gap because most of Africa’s rural land is managed under customary tenure and long-term rights by vulnerable groups such as women, migrants or herders may be the first to be attenuated in view of increasing land scarcity. To help fill this gap, we use unique survey data from Malawi to show that individuals’ long-term rights of bequest and sale significantly affect investment and cash crop adoption; that the gender of right holders matters, together with local institutional arrangements; and that effects may be obscured by measurement error associated with traditional data collection methods. We draw implications for policy and research in Malawi and beyond. How individuals can access land or exercise their land rights has far-reaching implications for their well-being and ability to take advantage of economic opportunities. While formally documented rights are the norm in developed countries, across the developing world -especially in rural Africa- formalization remains out of reach for the majority of the population. Instead, land is almost exclusively accessed through customary arrangements that are administered by traditional authorities. Individuals’ land rights in such systems depend on societal status (Honig 2017) and relationships with local leaders (Goldstein and Udry 2008). The literature has long highlighted that, as long as land is relatively abundant, customary systems’ ability to flexibly adjust to changes in external conditions allows use of land as a social safety net (Andolfatto 2002) and for mutual insurance (Beck et al. 2019). Where risk is high and other mechanisms for insurance are absent, this is a key advantage (Bruce and Migot-Adholla 1994) that may well lead land owners to prefer traditional arrangements to more formal alternatives (Atwood 1990). Yet, once land becomes scarcer, conflicts that are difficult for traditional authorities to resolve in a predictable way may exacerbate insecurity (Eck 2014) and leaders may act in their own, rather than the group’s, best interest (Greiner 2017). This can lead to a divergence between private and social benefits so that land may be privatized even where this reduces overall social welfare (Leeson and Harris 2018) or a shift in land rights within the household in ways that may disadvantage women. Growing land scarcity, possibly exacerbated by shifts in land demand due to exogenous shocks such as the 2007/08 food and financial crisis or the COVID-19 pandemic may thus gradually erode the rights of traditionally disadvantaged groups such as women, migrants, and pastoralists. These changes may affect not only equity but also efficiency, by impacting investment incentives and the ability to transfer land to better users. 2 Accurate measurement of the land rights held by individuals in customary systems as well as their impact on investment and productivity is thus essential for policy. This paper focuses on the case of Malawi, a poor and predominantly agricultural country with widely varying institutional arrangements, to provide substantive and methodological insights on the link between individuals’ land rights and investment. To do so, we rely on two nationally representative surveys, the Malawi Integrated Household Panel Survey (IHPS) and the Fourth Integrated Household Survey (IHS4), that were conducted in parallel in 2016/17 by the Malawi National Statistical Office. Both surveys elicited parcel-level gender-disaggregated data on reported, documented, and economic ownership and on rights to bequest, sell, rent out, use as collateral, and invest in it to make improvements. The data allow us to (i) rely on within-household variation in rights for identification and (ii) to use application of organic manure and planting of cash crops as indicators for investment and diversification, to explore the extent to which such rights affect economic outcomes. While the two surveys’ instruments are virtually identical, they adopted fundamentally different approaches to respondent selection that allow us to explore impacts of data collection methodology on substantive results. The IHPS aimed to conduct private interviews with each adult household member to elicit self- reported information regarding their personal ownership and rights for each parcel, a higher cost approach that is likely to result in higher data quality. Conversely, the IHS4 followed the traditional approach to survey data collection using proxy respondents, i.e. interview the self-identified most knowledgeable household member for each parcel to identify owners and right holders within the household. By comparing results from analysis of both data sets, we are able to draw rigorous inferences on the impact of data collection methods on survey responses and measurement of land rights. At a descriptive level, the standard practice of obtaining data on individuals’ land rights from proxy respondents -even if they profess to be knowledgeable- rather than directly from (co-)owners may under- (over)estimate the extent to which women (men) have (co-)ownership rights while exhibiting the opposite bias in terms of the type of rights female or male (co-)owners can exercise. Regression analysis using the IHPS data as -reported by all individual parcel owners suggests that gender- specific (long-term) rights to transfer land via bequest or sale (but not (short-term) rights to make decisions on investment or production) affect investment and diversification into cash crops. Right holders’ gender matters on its own and in interaction with local norms: Female bequest and, in non-matrilineal inheritance regimes, sale rights affect investment in soil fertility (proxied by the use of organic manure) whereas male bequest rights increase the level of cash crop adoption. These results cannot be replicated using the data from the IHS4 which allowed for proxy reporting on parcel ownership and rights. This finding materializes despite the IHS4 sample being significantly larger and 3 random respondent selection in both cases. One explanation could be that information on (informal) land rights is not fully shared within the household as observed for other asset types (Ashraf 2009). However, the fact that attempts at replication fail also for the subset of parcels for which survey respondents claimed to be (co)-owner may suggest that interviewees respond strategically -either by wrongly claiming rights they do not possess or by willingly failing to mention other co-owners. Irrespectively of the reason -which could be explored by follow-up research- this suggests that measurement error due to information on land rights being provided by proxy respondents rather than owners directly could be one reason for insignificant or contradictory findings on the effect of informal land rights in the literature (Kang et al. 2020). The evidence provided here contributes to several strands of the literature. First, we add to the literature on empowerment and productivity impacts of female asset ownership and land rights. Earlier studies show the effect of individual asset ownership on women’s agency, the relevance of land as a key asset in developing countries (Meinzen-Dick et al. 2019), and the importance of female inheritance rights at group level (Dillon and Voena 2018). We add to this by showing that, in contrast to short term management rights -the fluidity of which is a key advantage of customary tenure- long-term rights to bequeath or sell at individual level affect investment. As nature and strength of such rights may change in the evolution of customary systems transcending the dichotomy of statutory vs. customary rights and focusing on individuals’ ability to exercise such rights will be important. Second, though earlier studies noted that males’ and females’ perception of their rights to a plot may differ from each other (Twyman et al. 2015), to the best of our knowledge, ours is the first study to show that measurement error inherent in traditional ways of measuring gender differences in land rights may affect the conclusions derived from such data. The analysis reinforces the importance of survey research to develop improved methods of microdata collection (see Carletto and Gourlay (2019) for a review) that could lead to revision of relationships long accepted as ‘stylized facts’ of development (Christiaensen 2017). Third, evidence of considerable variation in informal rights across localities highlights that customary institutions are neither static nor monolithic and that study of the incentive structures shaping behavior of actors in such systems is likely to be a fruitful area for research. Leaders’ behavior and associated socio-economic outcomes are affected by a host of factors including (i) traditional authorities’ legal status (Henn 2020) and ability to interact with statutory institutions, in particular local government (de Kadt and Larreguy 2018); (ii) leaders’ local presence and access to private information (Casey et al. 2019); and (iii) mechanisms to hold traditional leaders and local authorities to account (Baldwin 2018). It will be important to identify how, in view of exogenous changes affecting customary tenure systems, individuals’ rights could be strengthened through adoption of governance arrangements and institutional mechanisms that provide 4 incentives for effective land management in customary settings such as Mexico (de Janvry et al. 2015; Zepeda 2000) or Colombia (Vélez et al. 2020). Finally, our study contributes to the large literature on land rights regularization by showing that the size of estimated effects from greater tenure security is comparable to or in excess of those often attributed to land right formalization (Fenske 2011; Lawry et al. 2016). This suggests that, even where establishment of functioning and viable land registries is a long-term goal, the traditional model of individual titling may be neither the only nor the most expeditious and effective way to secure rights. Legal and regulatory reforms to improve women’s bequest rights, as implemented for example in Kenya (Harari 2019), might yield productivity benefits comparable in size to those from land titling more quickly and at much lower cost. Further study to explore the impact of such reforms compared to land titling in rural areas -and the scope for enhancing it e.g. via awareness raising and legal support- is needed. The rest of the paper is structured as follows. Section 2 discusses evidence regarding the effect of female rights to land and provides an overview of the salient characteristics of Malawi’s land sector. Section 3 discusses data sources and descriptive statistics and introduces the empirical strategy. Section 4 discusses the results and section 5 five concludes with implications for data collection and future research. 2. Context and approach A large body of evidence suggests that formally documenting land rights can improve investment incentives and operation of factor markets and female empowerment. Much less is, however, known about the nature of land rights and their impact in settings where formal documentation does not exist, despite the global prevalence of such systems. We highlight the methodological challenges associated with obtaining reliable information on land rights in such circumstances and document them specifically in Malawi, a country where low levels of crop diversification despite high land pressure; the coexistence of matrilineal and patrilineal inheritance regimes; and a recently adopted land law further increase the relevance of such study. 2.1 Conceptual framework The literature has long emphasized that formally documenting land rights can have multiple benefits. By reducing expropriation risk (Besley and Ghatak 2010) and making it easier to identify owners or boundaries (Ali et al. 2014; Deininger et al. 2008), it creates incentives for land investments; allows transactions with unrelated third parties; and frees up resources that would otherwise have been spent on protecting claims (Goldstein et al. 2018). Providing reliable information on land ownership and contracts via public registries allows long-term transfers to improve the efficiency of resource allocation (Restuccia and Santaeulalia- Llopis 2015); encourage movement of labor between agriculture and non-agriculture (Chen 2017); and 5 facilitate migration (de Janvry et al. 2015) in the context of economic transformation. Active land markets can furthermore permit the use of land as collateral for credit (Deininger and Goyal 2012). In many cases, efforts to document individuals’ land rights proved particularly advantageous for women. This is largely because, contrary to in traditional settings where women’s land rights are defined through their informal connections (Goldstein and Udry 2008) and relationships (Jayachandran 2015), formal documents make it easier to formally enforce claims. Consistent with the evidence on the impact of changes in female asset ownership, 2 registering women as land owners has been shown to reduce the need for female guarding labor (Field 2007; Goldstein et al. 2018); increase investment especially by female-headed households (Ali et al. 2014) and women’s labor force participation (Newman et al. 2015); and welfare of future generations (Menon et al. 2014). Such impacts were found even if certificates are not transferrable, as in Madagascar (Widman 2014) or if land sales are legally prohibited, as in Ethiopia (Deininger et al. 2008; Kumar and Quisumbing 2015; Melesse et al. 2018; Muchomba 2017) where investment and empowerment effects have been amplified by family law reform (Kumar and Quisumbing 2015). Yet, although recent technological developments reduced the costs of formalization by orders of magnitude, running national land registries requires resources, organizational capabilities, and a suitable regulatory framework. The widespread absence of these factors in Africa implies that most land rights in rural areas are administered informally by traditional leaders under customary arrangements (Boone 2019). Such systems prohibit land transfers to unrelated outsiders and guarantee land access for lineage members (Baland and Francois 2005) to prevent destitution (Andolfatto 2002). Appreciation of their risk-sharing benefits and cost advantages has traditionally led experts to view customary systems as the most viable option in many settings (Bruce and Migot-Adholla 1994) and prompted many countries to recognize them in law, though less frequently in practice (Alden Wily 2018). There is evidence that population growth (Greiner 2017) and demand for land by outsiders (Lentz 2010) can set in motion processes of individualization that may lead to an unmediated co-existence of modern and traditional systems. This co-existence may in turn increase contestability and insecurity (Eck 2014) and reduce or undermine customary systems’ risk sharing advantages (Delpierre et al. 2019), with negative consequences especially for weaker groups such as migrants and women (Bambio and Bouayad Agha 2018). A first sign of the erosion of such groups’ longer-term rights can be a surge in inheritance-related land disputes that may disproportionately involve women (Deininger and Castagnini 2006). As inheritance rights affect incentives for adopting productivity-enhancing practices (Dillon and Voena 2018), this could 2 Such interventions have been shown to decrease consumption of male-favored goods in the household (Wang 2014) and to increase girls’ survival rates (Qian 2008), their anthropometric status (Duflo 2003), and their level of schooling (Deininger et al. 2013; Luke and Munshi 2011). 6 reduce productivity and female well-being more broadly. 3 In fact, reforms to equalize women’s inheritance rights in Kenya -albeit not applied equally everywhere (Linkow, 2019)- had multiple positive impacts on women’s livelihoods (Harari 2019). While the above implies that forestalling gradual erosion of women’s land rights could be an important policy issue in customary tenure systems, gathering data to measure women’s rights and tenure security is likely to require considerable effort. Contrary to formal systems where the type of right associated with each document is defined in law and holders’ names are written down, household surveys conducted in countries dominated by customary tenure systems ought to elicit information on whether the rights to bequeath, sell, rent out, use as collateral, make improvements are held individually or jointly by different individuals for each parcel (Doss et al., 2019). While there is little doubt that asking all adult individuals in private about their personal ownership of and rights to assets is the most desirable and recommended approach, the cost of doing so led most large-scale surveys to obtain this information through proxy reporting and specifically through the use of most knowledgeable household members. Little is known about the magnitude of errors potentially associated with this practice. 2.2 Nature and relevance of land rights in Malawi Malawi is one of Africa’s poorest and most densely populated countries with an economy that is heavily dependent on agriculture. Historically, land allocation in Malawi has been dualistic. Land for cultivation of commercial crops was provided to estates under freehold or leasehold (Deininger and Xia 2017) while smallholders were left to produce food crops, mainly maize, under customary tenure. Although restrictions on small farmers’ participation in cash crop cultivation were eliminated in the 1990s, smallholders’ limited resources (Ricker-Gilbert et al. 2014) and vulnerability to risk, particularly in the face of climate-related shocks (Sesmero et al. 2018), constrained their ability to diversify into other crops or adopt measures to manage soil fertility (Krah et al. 2019). Agricultural production patterns are thus characterized by dominance of maize, a staple with limited scope for value addition and employment generation. Despite unequal coverage, fertilizer subsidies may further have crowded out more sustainable practices (Khataza et al. 2017) such as investment in organic manure (Holden and Lunduka 2012). Growing land scarcity has also triggered land disputes (van Donge 1999), often centered around inheritance, and land disputes are often associated with higher levels of tenure insecurity and quantitatively large productivity losses (Deininger et al. 2019), especially for women. 4 3 While access to inherited land increased women’s level of self-employment and earnings in Ethiopia (Kumar and Quisumbing 2012), data from Demographic and Health Surveys for 15 African countries show that there are only two countries (Rwanda and Senegal) where widows and their children inherit most of a deceased’s assets whereas in more than half of the countries (Benin, the Democratic Republic of Congo, Guinea, Mali, Namibia, Niger, Nigeria, Rwanda, Senegal, Sierra Leone, Tanzania, Uganda, Zambia and Zimbabwe), widows report no inheritance of assets whatsoever (Peterman 2012). 4 With 22% of farmers fearing potential land loss, annual losses due to tenure insecurity are estimated at US$14 million (Deininger et al. 2019). 7 An important characteristic of Malawi’s land tenure system is the co-existence of matrilineal and patrilineal inheritance systems. As the ability to decide on long-term transfer of land rights has been shown to affect investment decisions (Deininger and Jin 2006), individuals’ ability to influence decisions on intergenerational land transfers is likely to affect their willingness to make non-observable investments. Place and Otsuka (2001), Lunduka (2009), Lovo (2016) and Bhaumik et al. (2016) indeed argue that in matrilineal and matrilocal systems, the discrepancy between females’ land ownership and males’ social position and superior access to markets for inputs and outputs impedes land-related investments. To the extent that they give individuals control over assets, inheritance regimes may also affect spouses’ outside options. 5 Women have been documented to be more vulnerable to temperature shocks in patrilineal systems, possibly due to lower levels of tenure security (Asfaw and Maggio 2018). Moreover, higher household consumption in matrilineal as compared to patrilineal systems has been attributed to differences in women’s tenure security across inheritance regimes (Telalagic 2014). With the objective of integrating customary and statutory rights in innovative ways, Malawi passed a new land law in 2016. The law aims to demarcate territories controlled by individual traditional leaders as a pre- condition for documented rights to individuals. 6 In addition, it introduced sweeping changes of inheritance practices that have been highly controversial with scholars who argue that such change could disenfranchise women on a massive scale (Peters 2010). 3. Data, descriptive statistics and empirical approach The descriptive analysis of our data shows that (co)-owners often do not have the full bundle of rights and that women lag men in all rights. Comparing descriptive statistics between surveys relying either on self- reporting (IHPS) or on responses by a proxy (IHS4) suggests that similar figures in the aggregate conceal important discrepancies between surveys. Compared to the data provided by individuals directly, proxy respondents (i) overstate levels of male (co-)ownership, (ii) understate levels of female (co-)ownership, and (iii) for parcels that are reported as (co-)owned, introduce an upward bias regarding the rights held by women and conversely, a downward bias regarding the rights held by men. The latter finding is consistent with informational imperfections as well as strategic responses. Econometric analysis to ascertain the net effects investment and land use is thus needed and we discuss the empirical framework to be used. 5 While historically, matrilineal inheritance regimes are related to the physical force required for food provision (BenYishay et al. 2017), there is some evidence of short-term changes in Malawi which are particularly relevant as divorce rates are high (Cherchye et al. 2016). 6 The law aims to demarcate traditional land management areas (TLMAs) and establish customary land committees (CLCs) with at least 50% female participation, normally at Group Village Headman level. 8 3.1 Data and descriptive statistics Our data are from two nationally-representative surveys in Malawi, namely the Fourth Integrated Household Survey 2016/17 (IHS4), a cross-sectional survey of 12,480 households, and the Integrated Household Panel Survey 2016 (IHPS), a longitudinal survey of 2,508 households that had been followed since 2010. 7 The IHS4 asked the most knowledgeable household member to provide information on household members’ ownership of and rights to agricultural parcels and other assets. 8 On the other hand, the IHPS aimed to conduct private interviews with each adult household member on his/her personal ownership of and rights to the residential and each agricultural parcel, based on a common roster of parcels that is used across all private interviews in each household. In contrast to the IHS4, the IHPS questions on land rights for a parcel were asked only if the respondent identified him or herself as a (co)-owner. 9 The mean values for key variables at the household- and parcel-level from the analysis sample of parcels that were reported to be owned are provided in Table 1 for the IHPS (column 1) and the IHS4, split into the total sample (column 2) and the sample of respondents claiming to be owner or co-owner of the parcel in question (column 3). The average household has about 5 members, with a 44-year old head. Eighty percent of heads of households are Christian; approximately 30 percent are female; 70 percent can read and write Chichewa; and have an average of 5.5 years of education. The mean area for agricultural parcels is 0.4 hectare and the incidence of organic fertilizer application and cash crop cultivation is 23 percent and 6 percent, respectively. The residential parcels are smaller and dwellings on them have been occupied for an average of 7 years. Thirty-nine percent of household residential dwellings have improved roofs, walls, or floors. 10 As only 2.3 percent (in IHPS) or 1.6 percent (in IHS4) of households report to have formal documentation to agricultural land, or 6.5 percent or 3.5 percent, respectively, to residential land, we focus on individual (co)-owners’ rights at the parcel level irrespective of whether or not formal documents exist. 11 Information on rights for agricultural parcels in Table 2 highlights several interesting observations. First, at the household-level, (co)-owners often can only exercise a subset of the rights generally associated with full ownership and the incidence of short-term rights is higher than that of long-term ones. The IHPS data 7 Data, questionnaires and basic information documents for the IHS4 2016/17 and the IHPS 2016 can be accessed here and here. Both surveys were implemented with technical and financial assistance from the World Bank Living Standards Measurement Study Integrated Surveys on Agriculture (LSMS-ISA) using the Surveys Solutions Computer-Assisted Personal Interviewing (CAPI) platform. Implementation of individual interviews as part of the IHPS 2016 received technical and financial support from the World Bank LSMS Plus (LSMS+) initiative, aiming to improve the availability and quality of individual-disaggregated survey data on asset ownership, work and employment and entrepreneurship in IDA countries. 8 A parcel (referred to as ‘garden’ in Malawi) is a contiguous piece of land under a specific type of tenure that can comprise multiple plots, defined by the types of crop grown. 9 The basic information document for the survey (https://microdata.worldbank.org/index.php/catalog/2939/download/47216) and Kilic et al. (2020) give more information on survey organization and implementation modalities for individual data collection under the IHPS. 10 We define improved walls to be made of burnt bricks and/or concrete (59% of sample), improved roofs to be made of iron sheets, clay tiles, and/or concrete (18% of sample), and improved floors to be made of smooth cement, wood, and/or tiles (41% of sample) and compute an ‘improvement index’ by adding dummies for each of the indicators and dividing by 3. 11 Evidence on formal rights suggests that, for agricultural land, individuals with formal rights are more likely to have the right to bequeath and sell, an outcome driven by men’s rights as there is no significant difference for women. Similarly, men with formal documents are more likely to have bequest and sale rights but there is no difference in the aggregate. 9 in panel A suggest that members of (co)-owners’ households can make cultivation decisions for 87 percent of parcels, followed by rights to invest (84 percent), to transfer/mortgage (75 percent), to bequeath (73 percent), and to sell (66 percent). 12 The same general pattern is observed for IHS4 data (panel B) that point to a similar share of parcels that can be sold (66%) and a slightly higher one (78%) that can be transferred via bequest. Second, compared to individually reported information (from the IHPS), data provided by proxy- respondents (IHS4) overestimate the extent of (co)-ownership for men and underestimate it for women. The share of agricultural parcels reported to be (co)-owned by females in the IHPS is 4 percentage points lower than the comparable figure from the IHS4. In contrast, the share of agricultural parcels reported to be (co)-owned by males in the IHPS is 9 percentage points higher than the comparable figure from the IHS4. Finally, conditional on female (male) (co)-ownership, proxy reporting in the IHS4 seems to overstate the specific rights that can be exercised by female owners while understating the rights that can be exercised by male owners. For agricultural parcels that are (co)-owned by females, the IHS4 data indicate a much lower incidence of bequest rights (60 percent) vis-a-vis the IHPS (69 percent). This pattern is less pronounced for the right to sell. The opposite is true for parcels that are (co)-owned by males. The IHPS average parcel-level incidence of male rights to bequest and sell are 78 percent and 70 percent, respectively. The comparable estimates based on the IHS4 stand at 74 percent and 64 percent, respectively. 3.2 Empirical approach We use the parcel-level incidence of organic fertilizer application and cultivation of cash crops (tobacco, cotton, sunflower, sugar cane, and pepper) in the current season as proxies for land investment. The first of these, organic manure application, is widely used in the literature to quantify land-attached investment (Fenske 2011), possibly together with fallowing to restore soil fertility (Goldstein and Udry 2008). The reason is that it affects productivity well beyond the current production period (Jacoby et al. 2002) but cannot be directly observed and is thus non-contractible (Jacoby and Mansuri 2008). Although manure use is our preferred indicator, its short-term productivity impact in a setting dominated by subsistence cultivation of maize may be limited. We thus use adoption of cash crops as a second indicator that has been reported as key to moving out of subsistence and increasing household income (Bhaumik et al. 2016). Cash crops are also a key contributor to the gender differences in agricultural productivity (Kilic et al., 2015). Perennials and cash crops that provide benefits in the long term require considerable upfront investment that is unlikely to be undertaken without secure land rights (Place and Otsuka 2001). 12 The share of parcels that can be bequeathed is, with 82%, highest in communities in which the dominant inheritance regime is patrilineal, followed by matrilineal (71%) and mixed (69%) inheritance regimes. 10 After aggregating plot-level survey data on manure application and cash crop cultivation to the parcel level, 13 we estimate: = + + + (1) where is defined above; controls for unobserved household fixed effects such as risk-bearing capacity that do not vary across parcels; is a vector of parcel-specific indicators of land rights for those who report or are reported as land owners; is a vector of parcel controls including size, soil type, slope, and location in a swamp/wetland; 14 and is an error term. As public programs including input subsidies (Asfaw et al. 2017) are administered by any one of Malawi’s 28 districts, we cluster errors at the district- level and use the wild cluster bootstrap (Roodman et al. 2018) to account for the small number of clusters. The coefficient of interest is β, the vector of coefficients capturing the effect of different types of land rights on land-attached agricultural investment. To allow for heterogeneity of effects by gender, we denote women’s and men’s land rights by and , respectively. is an indicator variable for women’s land rights, equaling one if at least one woman in household j reports (or, in the case of IHS4 data are reported as) having land rights of a certain type to parcel i. Similarly, is the equivalent indicator variable for men’s land rights that equals one if at least one man in household j has rights of a certain nature to parcel i. We modify (1) as follows = + 1 + 2 + + (2) where 1 and 2 are estimated effects of land rights being held by women or men, respectively, and 1 + 2 denotes the estimated effect of rights being held jointly by male and female household members. As the nature of rights and ability to enforce them can vary by inheritance regime, we let be an indicator variable for household j living in a community with a patrilineal/mixed inheritance regime and estimate = + 1 + 2 + 3 ∗ + 4 ∗ + + (3) where 1 and 2 indicate estimated effects of women’s and men’s rights in matrilineal communities while 3 and 4 indicate if women’s or men’s rights affect outcomes differently in non-matrilineal communities. 13 Formally, we index plots by s, parcels by i and households by j and compute organic fertilizer application or plating of cash crops on parcel i of ∑ =1 × 1 ℎ household j as = ∑ , = � =1 0 ℎ where is an indicator variable for organic fertilizer application or planting of cash crops by household j on plot s of parcel i, and is the plot size in hectares. 14 Parcel size is in logs, soil types considered are sand, sandy/clay or clay; inclination can be flat, slightly sloped, moderately sloped, or hilly. These parcel characteristics are also aggregated from plot information. They are also weighted by plot size. 11 4. Results The analysis of the IHPS data suggests that (long-term) rights to bequeath and sell -but not for decision- making- affect land-related investment. Women’s right to bequeath in matrilineal systems or to sell in non- matrilineal ones affects organic manure use while men’s bequest rights affect take-up of cash-crop production. Using the full IHS4 data or the subset of self-reported (co-)owners yields completely different results that are often counterintuitive, suggesting that beyond measurement error due to respondents’ lack of knowledge, reliance on proxy respondents may also encourage strategic behavior. 4.1 Evidence from self-reported IHPS data The results from household fixed effects regressions for application of organic manure and planting of cash crops are presented in Table 3. The regressions presented in columns 1 and 3 do not control for parcel-level characteristics, while the opposite is true regarding the regressions presented in columns 2 and 4. The rights to make decisions or invest are consistently estimated to have insignificant effects regardless of the choice of outcome variable. By contrast, transfer rights are highly significant, similar to what was found in Ethiopia (Deininger and Jin 2006): the right to bequeath or sell land is estimated to significantly increase the likelihood of long-term investment via organic manure application. With 0.12 and 0.08 percentage points, the estimated effects are large; increasing bequest or sale rights by half a standard deviation (0.2) each would be expected to lead to an increase in organic manure application by 10 percent or 6 percent from the mean value of 0.23. The estimated effects are weaker for cash crop production where bequest rights are marginally significant at the 10% level if parcel characteristics are controlled for. An estimate of 0.06 points suggests that increasing bequest rights by half a standard deviation would increase cash crop production by 17 percent. Exploring differences by right holders’ gender points towards marked differences between manure use and cash crop production (Table 4). Female bequest rights emerge as key determinants of long-term investment. The estimated coefficient of 0.16 is highly significant and large: an increase of women’s bequest rights by half a standard deviation (0.2) each is estimated to increase manure application by 13 percent. This is consistent with evidence of inheritance rights as key determinants of long-term investment in the form of applying organic manure in Zambia (Dillon and Voena 2018) using a completely different specification. By contrast, the adoption of cash crops is estimated to be enhanced by males’ bequest rights with an increase of male bequest rights by half a standard deviation (0.2) predicted to increase cash crop production by 61 percent. This is consistent with the notion that males have better access to markets for output and other factors especially capital (Bhaumik et al. 2016) but that such market access will translate into cash crop adoption only if they have secure long-term rights to the land that benefits from such investment. 12 Interacting men’s and women’s rights with the prevalent inheritance regime leaves estimates of impacts of female bequest rights on organic manure use virtually unchanged (see Table 5), while pointing towards no significant difference between matrilineal and patrilineal systems. Significant coefficients on women’s right to sell land in non-matrilineal systems suggest, however, that in such a regime, women’s ability to sell land further increases the likelihood of long-term investment. With a point estimate of 0.24, the estimated effects of increasing sales rights are meaningful economically. By comparison, the significance of estimated effects of male bequest rights on cash crop adoption weaken slightly in this specification. 4.2 Comparing results from self-reporting to indirect reporting As the size of the IHS4 sample exceeds that of the IHPS by an order of magnitude, it should, in the absence of measurement error, result in more precise coefficient estimates of the impact of land right variables included in both surveys. Inability to replicate results from the IHPS using the full or restricted IHS4 samples in most cases points towards the presence of not only measurement error but also strategic response bias. Tables 6 through 8 report the results from estimating equations (2) and (3) with the IHS4 data, with a focus on the bequest and sale rights. Columns 2 and 5 of each table include the results based on the full IHS4 sample; columns 3 and 6 include the results based on the sample of IHS4 respondents who report to be (co-)owners; and columns 1 and 4 include the results based on the IHPS sample for comparison purposes. While both the IHPS and the full IHS4 samples point towards weakly significant positive effects of bequest rights on use of organic manure, the estimated coefficient is much smaller for the full IHS4 sample and loses significance if only owner-respondents are considered (Table 6). The estimated effects of land rights on cash crop production are insignificant throughout. The loss of significance is consistent with the notion that relying on one respondent per household only increases measurement error that is not eliminated for those claiming to be a (co-)owner of a parcel. Allowing for heterogeneous effects by gender (Table 7) accentuates the differences between the two surveys: compared to point estimate of 0.16 for the impact of female bequest rights on manure use, estimates for both IHS4 samples are insignificantly different from zero for bequest and sale rights. For planting of cash crops, where the IHPS sample suggests a positive and significant coefficient on male bequest rights with an estimate of 0.20, the results from the full and restricted IHS4 samples imply insignificant effects of male rights and negative and significant effects of female bequest rights. Finally, interacting gender-specific rights with the inheritance regime indicators (Table 8) produces results that are similarly inconsistent with those obtained from the IHPS sample. While the estimated coefficients on women’s or men’s bequest rights for manure application and cash crops are insignificant in matrilineal communities, a negative and marginally significant effect of male bequest rights on cash crop adoption 13 emerges from the restricted sample. Together, the IHS4-based estimates suggest that bequest rights have either no effect or negative effect on investment, indicating systematic misreporting of bequest rights. 5. Conclusion and policy implications We use the example of Malawi to show that, even in customary settings where formal documentation is absent, the nature and extent to which women and men have rights to land vary widely. Variation in bequest and to some extent sale but not in other rights systematically affects land-based investment. Yet, standard household surveys may fail to capture such variation due to measurement error arising from reliance on proxy respondents rather than interviewing land owners themselves. We conclude by drawing out implications from our findings for research and policy. An important area for follow-up research is to assess the extent to which the findings presented here can be generalized beyond Malawi. Doing so would allow exploring if a failure to either include information on bequest rights or to collect such information directly from (co-)owners can explain the lack of clear results on the impact of individual land rights in the literature. Also, if results similar to what was presented here are found in other countries, getting information on the right to bequeath or sell land from concerned individuals at least for a subsample of national household surveys will be important. Beyond identifying the impacts of specific land rights on socio-economic outcomes, this would provide a basis for understanding of changes in individuals’ rights in response to exogenous factors. Our findings are relevant for Malawi in two respects. First, as most of our data were collected before the 2016 Land Act was assented to by the President in January 2017 and awareness of the specific provisions of the Act remains low, subsequent IHPS (or IHS) waves can help ascertain if passage of the Land Act, in particular its far-reaching modifications of inheritance patterns, affected individuals’ land rights and investment in agricultural and residential land. Second, defining customary estates to implement the Land Act will require adjudication of rights. Our analysis suggests that strategic motives can bias responses even in household surveys where stakes are low. To prevent such activities systematically disempowering weaker groups, information used to adjudicate rights in potential demarcation exercises will thus have to be carefully validated and cross-checked. Beyond Malawi, the fact that investment impacts associated with bequest and transfer rights under informal regimes are of a magnitude that is comparable to what is often ascribed to formal title in the literature implies that greater attention ought to be devoted to designing and evaluating such schemes, including changes in laws or regulations for inheritance. This does not negate the importance of formal registries to facilitate impersonal transactions and use of land as collateral but recognizes that, as centralized registries 14 for rural land alone cannot be sustained, 15 establishing property registries is a long-term agenda that has to start in urban areas. Given the adverse and potentially irreversible equity and productivity impacts from gradual erosion of individuals’ customary land rights, more agile means to measure such rights and protect them if needed may be required. 15 Analysis shows that, with costs that exceed any imaginable benefits as well as available resources by orders of magnitude, provisions regarding customary rights in Uganda’s 1998 Land Act are unlikely to be ever implementable (Hunt 2004). This has neither prevented donors from trying nor neighboring countries from passing similarly ambitious pieces of legislation. Ali et al. (2019) show that even in Rwanda, the only African country with a functioning national registry, sustaining rural registries will require efficient operation with a key role for IT; active engagement with the private sector (e.g. by using mobile phone providers as a first point of contact, and cross-subsidization from urban registry operations. 15 Table 1: Descriptive statistics across surveys and samples IHPS IHS4 IHS4 restricted Panel A: Household level information for owners of agricultural land Household size 5.010 4.457 4.409 Head’s age (years) 44.382 44.824 44.971 Head female† 0.264 0.311 0.342 Head reads & writes Chichewa† 0.685 0.672 0.661 Head’s education (years) 5.560 5.298 5.230 Somebody reads & writes Chichewa† 0.911 0.874 0.868 Highest educ.by any member (years) 7.881 7.332 7.244 Christian† 0.785 0.827 0.829 Household asset and dwelling wealth index 0.247 0.184 0.183 Household agricultural asset access index 0.208 0.195 0.195 Area of agricultural land (ha) 0.647 0.569 0.577 No. of parcels 1.537 1.399 1.420 Uses org. manure† 0.297 0.257 0.263 Cultivates cash crop† 0.097 0.073 0.075 Has formal document to agric. land† 0.023 0.016 0.016 No. of households 1,308 8,203 7,129 Panel B: Parcel level information for owned agricultural land Parcel size (ha) 0.421 0.407 0.410 Organic manure applied† 0.233 0.214 0.216 Cash crop planted† 0.070 0.055 0.055 Soil is sandy 0.197 0.188 0.189 Soil is between sandy and clay 0.527 0.566 0.570 Soil is clay 0.276 0.245 0.241 Slope is flat 0.502 0.56 0.562 Slope is small 0.370 0.334 0.331 Slope is moderate 0.097 0.070 0.070 Slope is steep 0.031 0.036 0.037 Parcel in swamp/wetland 0.131 0.124 0.124 No. of parcels 2,011 11,472 9,668 Panel C: Household level information for owned dwelling & residential land Property size (ha) 0.055 0.058 0.059 Years dwelling has been occupied 7.451 7.245 7.185 Improved dwelling† 0.397 0.393 0.385 Has formal document to non-agric. land† 0.065 0.035 0.030 No. of households 1,354 7,099 6,239 Source: Own computation from 2015/16 Malawi IHPS and IHS4. Note: ‘IHS4 restricted’ refers to the subsample of respondents reporting to be owners or co-owners of the parcel or dwelling. † Underlying variable is dichotomous variable. 16 Table 2: Descriptive statistics on rights for agricultural land Agricultural Land All Matrilineal Mixed Patrilineal Panel A: IHPS Household level Parcel size (ha) 0.421 0.437 0.346 0.393 Member has right to make decision 0.870 0.861 0.923 0.908 Member has right to invest 0.840 0.834 0.812 0.913 Member has right to transfer/mortgage 0.749 0.734 0.751 0.821 Member has right to bequeath 0.728 0.712 0.685 0.817 Member has right to sell 0.657 0.658 0.569 0.670 Parcel level Parcel (co)-owned by female(s) 0.746 0.750 0.801 0.711 Females can decide on crop cultivation 0.847 0.834 0.924 0.884 Females have right to invest 0.758 0.745 0.738 0.845 Females can transfer/mortgage 0.626 0.614 0.579 0.645 Females can bequeath 0.604 0.592 0.510 0.652 Females can sell 0.530 0.539 0.393 0.484 Parcel (co)-owned by men 0.440 0.421 0.486 0.560 Males can decide on crop cultivation 0.880 0.875 0.864 0.926 Males can invest 0.889 0.893 0.807 0.943 Males can transfer/mortgage 0.794 0.803 0.761 0.820 Males can bequeath 0.782 0.769 0.739 0.910 Males can sell 0.698 0.715 0.602 0.705 No. of parcels 2,207 1,617 181 218 Panel B: IHS4 Household level Parcel size (ha) 0.407 0.409 0.396 0.403 Member has right to bequeath 0.757 0.747 0.705 0.838 Member has right to sell 0.642 0.645 0.642 0.641 Parcel level Parcel (co)-owned by women 0.704 0.715 0.637 0.683 Females can bequeath 0.691 0.706 0.588 0.672 Females can sell 0.551 0.588 0.519 0.398 Parcel (co)-owned by males 0.529 0.468 0.651 0.751 Males can bequeath 0.736 0.702 0.708 0.850 Males can sell 0.639 0.629 0.667 0.661 No. of parcels 13,011 9,745 1,051 2,110 Source: Own computation from 2015/16 Malawi IHPS and IHS4. 17 Table 3: Effects of land rights on agricultural investment (IHPS) Organic manure applied Cash crop planted Has right to decide on crop 0.029 0.030 0.035 0.036 cultivation (β11) [0.631] [0.558] [0.325] [0.284] Has right to invest (β12) -0.091 -0.083 -0.037 -0.034 [0.299] [0.333] [0.647] [0.686] Has right to bequeath land (β13) 0.118** 0.121** 0.059 0.059* [0.019] [0.021] [0.135] [0.095] Have right to sell land (β14) 0.078*** 0.068** -0.018 -0.023 [0.006] [0.029] [0.791] [0.745] Parcel characteristics No Yes No Yes No. of observations 2,011 2,011 2,011 2,011 R-squared 0.775 0.777 0.576 0.579 Mean of dependent variable 0.233 0.233 0.070 0.070 Sd. of dependent variable (0.416) (0.416) (0.238) (0.238) Mean of right to bequeath land 0.724 0.724 0.724 0.724 Sd. of right to bequeath land (0.447) (0.447) (0.447) (0.447) Mean of right to sell land 0.656 0.656 0.656 0.656 Sd. of right to sell land (0.475) (0.475) (0.475) (0.475) Notes: IHPS data are used and the sample includes owned parcels only. Household fixed effects and parcel size are controlled for throughout. Parcel characteristics include soil type (sandy, between sandy and clay, or clay), land slope (flat, slight slope, moderate slope, or hilly), and whether land is in swamp/wetland. Robust standard errors are clustered by district. Figures in square brackets are p-values from wild cluster bootstrap (Roodman, 2018) with Rademacher weights and 1000 replications. *** significant at 1%; significant at 5%; * significant at 10%. 18 Table 4: Effects of gender-differentiated land alienation rights on agricultural investment (IHPS) Organic manure Cash crop planted applied Women have the right to make decisions on crop cultivation (β11) 0.034 0.037 -0.004 -0.001 [0.271] [0.206] [0.920] [0.991] Women have the right to invest (β12) -0.014 -0.012 -0.018 -0.015 [0.854] [0.855] [0.797] [0.836] Women have the right to bequeath (β13) 0.158*** 0.164*** 0.046 0.049 [0.001] [0.000] [0.393] [0.394] Women have the right to sell (β14) 0.052 0.051 -0.047 -0.047 [0.328] [0.352] [0.653] [0.616] Men have the right to make decisions on crop cultivation (β21) 0.060 0.062 0.050 0.046 [0.552] [0.567] [0.502] [0.524] Men have the right to invest (β22) -0.096 -0.084 0.052 0.064 [0.430] [0.501] [0.498] [0.399] Men have the right to bequeath (β23) 0.021 0.012 0.179** 0.177** [0.858] [0.916] [0.015] [0.014] Men have the right to sell (β24) 0.053 0.033 0.026 0.019 [0.564] [0.704] [0.690] [0.739] Parcel characteristics No Yes No Yes Observations 2,011 2,011 2,011 2,011 R-squared 0.777 0.780 0.585 0.588 Mean of dependent variable 0.233 0.233 0.070 0.070 Sd. of dependent variable (0.416) (0.416) (0.238) (0.238) Mean of women’s right to bequeath 0.454 0.454 0.454 0.454 Sd. of women’s right to bequeath (0.498) (0.498) (0.498) (0.498) Mean of women’s right to sell 0.402 0.402 0.402 0.402 Sd. of women’s right to sell (0.490) (0.490) (0.490) (0.490) Mean of men’s right to bequeath 0.334 0.334 0.334 0.334 Sd. of men’s right to bequeath (0.472) (0.472) (0.472) (0.472) Mean of men’s right to sell 0.299 0.299 0.299 0.299 Sd. of men’s right to sell (0.458) (0.458) (0.458) (0.458) Notes: IHPS data are used and the sample includes owned parcels only. Indicator variables for women’ ownership or men’s ownership are controlled for. Household fixed effects and parcel size are controlled for throughout. Parcel characteristics include soil type (sandy, between sandy and clay, or clay), land slope (flat, slight slope, moderate slope, or hilly), and whether land is in swamp/wetland. Robust standard errors are clustered by district. Figures in square brackets are p-values from wild cluster bootstrap (Roodman, 2018) with Rademacher weights and 1000 replications. *** significant at 1%; significant at 5%; * significant at 10%. 19 Table 5: Effects of land rights on agricultural investment by gender and inheritance regime (IHPS) Organic manure applied Cash crop planted Women have right to make decision (β11) 0.033 0.038 -0.000 0.003 [0.409] [0.324] [0.989] [0.926] Women have right to invest (β12) 0.052 0.059 -0.038 -0.033 [0.340] [0.209] [0.690] [0.743] Women have right to bequeath land (β13) 0.152** 0.160*** 0.072 0.074 [0.014] [0.009] [0.219] [0.255] Women have right to sell land (β14) 0.005 0.005 -0.072 -0.071 [0.940] [0.936] [0.526] [0.556] Men have right to make decision (β21) 0.017 0.019 -0.004 -0.011 [0.858] [0.870] [0.962] [0.898] Men have right to invest (β22) -0.036 -0.036 0.095 0.105 [0.754] [0.750] [0.397] [0.347] Men have right to bequeath land (β23) -0.003 0.011 0.158* 0.163* [0.992] [0.956] [0.095] [0.096] Men have right to sell land (β24) 0.064 0.040 0.048 0.046 [0.611] [0.756] [0.440] [0.517] Women have right to make decision -0.019 -0.042 -0.203 -0.218 *Non-matrilineal community (β31) [0.686] [0.367] [0.131] [0.127] Women have right to invest -0.270 -0.289 0.086 0.080 *Non-matrilineal community (β32) [0.246] [0.102] [0.455] [0.481] Women have right to bequeath land -0.024 -0.031 -0.114 -0.115 *Non-matrilineal community (β33) [0.799] [0.766] [0.365] [0.370] Women have right to sell land 0.244** 0.237** 0.117 0.115 *Non-matrilineal community (β34) [0.016] [0.030] [0.404] [0.429] Men have right to make decision 0.095 0.103 0.186 0.195 *Non-matrilineal community (β41) [0.664] [0.651] [0.271] [0.285] Men have right to invest 0.139 0.180 -0.162 -0.159 *Non-matrilineal community (β42) [0.291] [0.220] [0.251] [0.265] Men have right to bequeath land -0.030 -0.095 -0.119 -0.139 *Non-matrilineal community (β43) [0.901] [0.669] [0.569] [0.497] Men have right to sell land -0.076 -0.066 0.051 0.048 *Non-matrilineal community (β44) [0.699] [0.743] [0.851] [0.836] Parcel characteristics No Yes No Yes Observations 1,932 1,932 1,932 1,932 R-squared 0.778 0.781 0.587 0.589 Mean of dependent variable 0.233 0.233 0.070 0.070 Means of independent variables: Female bequest in matrilineal com. 0.444 0.444 0.444 0.444 Female sale right in matrilineal com. 0.407 0.407 0.407 0.407 Male bequest right in matrilineal com. 0.324 0.324 0.324 0.324 Male sale right in matrilineal com. 0.299 0.299 0.299 0.299 Female bequest right in non-matrilineal community 0.456 0.456 0.456 0.456 Female sale right in non-matrilineal community 0.347 0.347 0.347 0.347 Male bequest right in non-matrilineal community 0.407 0.407 0.407 0.407 Male sale right in non-matrilineal community 0.333 0.333 0.333 0.333 Notes: IHPS data are used and the sample includes owned parcels only. Indicator variables for women’ ownership or men’s ownership and their interactions with non-matrilineal community are controlled for. Household fixed effects and parcel size are controlled for throughout. Parcel characteristics include soil type (sandy, between sandy and clay, or clay), land slope (flat, slight slope, moderate slope, or hilly), and whether land is in swamp/wetland. Robust standard errors are clustered by district. Figures in square brackets are p-values from wild cluster bootstrap (Roodman, 2018) with Rademacher weights and 1000 replications. *** significant at 1%; significant at 5%; * significant at 10%. 20 Table 6: Estimated effects of type of land rights on agricultural investment across surveys Organic manure applied Cash crop planted IHPS IHS4 IHS4 rest. IHPS IHS4 IHS4 rest. Has right to bequeath land (β11) 0.105* 0.059* 0.048 0.054 -0.034 -0.044 [0.068] [0.051] [0.258] [0.178] [0.237] [0.299] Has right to sell land (β12) 0.050 -0.013 -0.011 -0.032 0.008 0.002 [0.164] [0.602] [0.723] [0.697] [0.725] [0.934] Observations 2,011 11,472 9,668 2,011 11,472 9,668 R-squared 0.777 0.817 0.834 0.578 0.618 0.634 Mean of dependent variable 0.233 0.214 0.216 0.070 0.055 0.055 Sd. of dependent variables: (0.416) (0.404) (0.406) (0.238) (0.217) (0.216) Mean of right to bequeath land 0.724 0.751 0.727 0.724 0.751 0.727 Sd. of right to bequeath land (0.447) (0.433) (0.445) (0.447) (0.433) (0.445) Mean of right to sell land 0.656 0.643 0.614 0.656 0.643 0.614 Sd. of right to sell land (0.475) (0.479) (0.487) (0.475) (0.479) (0.487) Notes: Regressions are for owned parcels. Household fixed effects, parcel attributes including size, soil type (sandy, between sandy and clay, clay), slope (flat, slight slope, moderate slope, hilly), and whether land is in swamp/wetland are controlled for throughout. Robust standard errors are clustered by district. Figures in square brackets are p-values from wild cluster bootstrap (Roodman, 2018) with Rademacher weights and 1000 replications. *** significant at 1%; significant at 5%; * significant at 10%. 21 Table 7: Compare effects by gender and number of right holders across surveys Organic manure applied Cash crop planted IHPS IHS4 IHS4 rest. IHPS IHS4 IHS4 rest. Women have the right to bequeath (β11) 0.165*** 0.025 0.027 0.045 -0.042* -0.088** [0.002] [0.320] [0.545] [0.535] [0.076] [0.034] Women have the right to sell (β12) 0.047 0.015 -0.008 -0.047 0.035* 0.009 [0.388] [0.684] [0.889] [0.594] [0.098] [0.760] Men have the right to bequeath (β21) -0.013 0.047 0.070 0.199*** -0.002 0.001 [0.916] [0.301] [0.301] [0.002] [0.969] [0.990] Men have the right to sell (β22) 0.019 -0.024 -0.010 0.027 -0.013 -0.003 [0.820] [0.449] [0.789] [0.659] [0.696] [0.951] Observations 2,011 11,472 9,668 2,011 11,472 9,668 R-squared 0.779 0.817 0.834 0.587 0.619 0.635 Mean of dependent variable 0.233 0.214 0.216 0.070 0.055 0.055 Sd. of dependent variable (0.416) (0.404) (0.406) (0.238) (0.217) (0.216) Mean of women’s right to bequeath 0.454 0.488 0.407 0.454 0.488 0.407 Sd. of women’s right to bequeath (0.498) (0.500) (0.491) (0.498) (0.500) (0.491) Mean of women’s right to sell 0.402 0.395 0.333 0.402 0.395 0.333 Sd. of women’s right to sell (0.490) (0.489) (0.471) (0.490) (0.489) (0.471) Mean of men’s right to bequeath 0.334 0.375 0.321 0.334 0.375 0.321 Sd. of men’s right to bequeath (0.472) (0.484) (0.467) (0.472) (0.484) (0.467) Mean of men’s right to sell 0.299 0.329 0.281 0.299 0.329 0.281 Sd. of men’s right to sell (0.458) (0.470) (0.449) (0.458) (0.470) (0.449) Notes: Regressions are for owned parcels. Household fixed effects, parcel attributes including size, soil type (sandy, between sandy and clay, clay), slope (flat, slight slope, moderate slope, hilly), and whether land is in swamp/wetland, and interactions between women’ or men’s ownership are controlled for throughout. Robust standard errors are clustered by district. Figures in square brackets are p-values from wild cluster bootstrap (Roodman, 2018) with Rademacher weights and 1000 replications. *** significant at 1%; significant at 5%; * significant at 10%. 22 Table 8: Effects by gender and inheritance regime across surveys Organic manure applied Cash crop planted IHPS IHS4 IHS4 rest. IHPS IHS4 IHS4 rest. Women have right to bequeath land (β11) 0.177** 0.047 0.016 0.062 -0.035 -0.070 [0.010] [0.118] [0.757] [0.413] [0.203] [0.159] Women have right to sell land (β12) 0.008 0.023 0.005 -0.073 0.046** 0.037 [0.896] [0.497] [0.911] [0.533] [0.048] [0.314] Men have right to bequeath land (β21) -0.022 0.038 0.036 0.194*** 0.012 0.051 [0.888] [0.464] [0.663] [0.008] [0.713] [0.232] Men have right to sell land (β22) 0.039 -0.003 -0.009 0.056 -0.026 -0.008 [0.753] [0.938] [0.882] [0.463] [0.279] [0.867] Women have right to bequeath land -0.062 -0.109 -0.035 -0.108 -0.016 -0.083 *Non-matrilineal community (β31) [0.583] [0.125] [0.629] [0.443] [0.749] [0.613] Women have right to sell land 0.200* 0.031 -0.010 0.143 -0.042 -0.117 *Non-matrilineal community (β32) [0.057] [0.698] [0.929] [0.357] [0.423] [0.139] Men have right to bequeath land 0.063 0.054 0.157 -0.050 -0.058 -0.243* *Non-matrilineal community (β41) [0.801] [0.393] [0.205] [0.637] [0.555] [0.081] Men have right to sell land -0.032 -0.094 -0.026 -0.014 0.072 0.077 *Non-matrilineal community (β42) [0.832] [0.231] [0.750] [0.950] [0.512] [0.464] Observations 1,932 11,201 9,439 1,932 11,201 9,439 R-squared 0.780 0.817 0.833 0.585 0.609 0.629 Mean of dependent variable 0.233 0.215 0.218 0.070 0.054 0.054 Means of independent variables: Female bequest in matrilineal com. 0.444 0.507 0.440 0.444 0.507 0.440 Female sale right in matrilineal com. 0.407 0.425 0.370 0.407 0.425 0.370 Male bequest right in matrilineal com. 0.324 0.324 0.284 0.324 0.324 0.284 Male sale right in matrilineal com. 0.299 0.291 0.255 0.299 0.291 0.255 Female bequest right in non-matrilineal community 0.456 0.423 0.293 0.456 0.423 0.293 Female sale right in non-matrilineal community 0.347 0.287 0.208 0.347 0.287 0.208 Male bequest right in non-matrilineal community 0.407 0.558 0.441 0.407 0.558 0.441 Male sale right in non-matrilineal community 0.333 0.465 0.365 0.333 0.465 0.365 Notes: Regressions are for owned parcels. Household fixed effects, parcel attributes including size, soil type (sandy, between sandy and clay, clay), slope (flat, slight slope, moderate slope, hilly), and whether land is in swamp/wetland, and interactions between women’ or men’s ownership with non-matrilineal community are controlled for throughout. Robust standard errors are clustered by district. Figures in square brackets are p-values from wild cluster bootstrap (Roodman, 2018) with Rademacher weights and 1000 replications. *** significant at 1%; significant at 5%; * significant at 10%. 23 References: Alden Wily, L. 2018. "Collective Land Ownership in the 21st Century: Overview of Global Trends." Land, 7(2), 68. Ali, D. A., K. Deininger and M. Goldstein. 2014. "Environmental and Gender Impacts of Land Tenure Regularization in Africa: Pilot Evidence from Rwanda." Journal of Development Economics, 110(0), 262-275. Ali, D. A., K. Deininger, G. Mahofa and R. Nyakulama. 2019. "Sustaining Land Registration Benefits by Addressing the Challenges of Reversion to Informality in Rwanda." Land Use Policy, 104317. Andolfatto, D. 2002. "A Theory of Inalienable Property Rights." Journal of Political Economy, 110(2), 382-393. Asfaw, S., A. Cattaneo, G. Pallante and A. Palma. 2017. "Improving the Efficiency Targeting of Malawi's Farm Input Subsidy Programme: Big Pain, Small Gain?" Food Policy, 73, 104-118. Asfaw, S. and G. Maggio. 2018. "Gender, Weather Shocks and Welfare: Evidence from Malawi." Journal of Development Studies, 54(2), 271-291. Ashraf, N. 2009. "Spousal Control and Intra-Household Decision Making: An Experimental Study in the Philippines." American Economic Review, 99(4), 1245-1277. Atwood, D. A. 1990. "Land Registration in Africa: The Impact on Agricultural Production." World Development, 18(5), 659-671. Baland, J. M. and P. Francois. 2005. "Commons as Insurance and the Welfare Impact of Privatization." Journal of Public Economics, 89(2-3), 211-231. Baldwin, K. 2018. "Elected Mps, Traditional Chiefs, and Local Public Goods: Evidence on the Role of Leaders in Co-Production from Rural Zambia." Comparative Political Studies, 52(12), 1925-1956. Bambio, Y. and S. Bouayad Agha. 2018. "Land Tenure Security and Investment: Does Strength of Land Right Really Matter in Rural Burkina Faso?" World Development, 111, 130-147. Beck, U., B. Bjerge and M. Fafchamps. 2019. "The Role of Social Ties in Factor Allocation." World Bank Economic Review, 33(3), 598-621. BenYishay, A., P. Grosjean and J. Vecci. 2017. "The Fish Is the Friend of Matriliny: Reef Density and Matrilineal Inheritance." Journal of Development Economics, 127, 234-249. Besley, T. and M. Ghatak. 2010. "Property Rights and Economic Development." in M. R. Rosenzweig and D. Rodrik (eds.), Handbook of Economic Development Vol 5, Elsevier, Oxford and Amsterdam. Bhaumik, S. K., R. Dimova and I. N. Gang. 2016. "Is Women's Ownership of Land a Panacea in Developing Countries? Evidence from Land-Owning Farm Households in Malawi." Journal of Development Studies, 52(2), 242-253. Boone, C. 2019. "Legal Empowerment of the Poor through Property Rights Reform: Tensions and Trade-Offs of Land Registration and Titling in Sub-Saharan Africa." Journal of Development Studies, 55(3), 384-400. Bruce, J. W. and S. E. Migot-Adholla. 1994. Searching for Land Tenure Security in Africa. Kendall/Hunt Publishers, Dubuque, IA. Carletto, C. and S. Gourlay. 2019. "A Thing of the Past? Household Surveys in a Rapidly Evolving (Agricultural) Data Landscape: Insights from the Lsms-Isa." Agricultural Economics, 50, 51-62. Casey, K., R. Glennerster, E. Miguel and M. Voors. 2019. "Skill Versus Voice in Local Development." Stanford University, Graduate School of Business, Research Papers. Chen, C. 2017. "Untitled Land, Occupational Choice, and Agricultural Productivity." American Economic Journal: Macroeconomics, 9(4), 91-121. Cherchye, L., B. De Rock, S. T. Walther and F. Vermeulen. 2016. "Where Did It Go Wrong? Marriage and Divorce in Malawi." University of Oxford, Department of Economics, Economics Series Working Papers: 786. Christiaensen, L. 2017. "Agriculture in Africa--Telling Myths from Facts: A Synthesis." Food Policy, 67, 1-11. de Janvry, A., K. Emerick, M. Gonzalez-Navarro and E. Sadoulet. 2015. "Delinking Land Rights from Land Use: Certification and Migration in Mexico." American Economic Review, 105(10), 3125-3149. de Kadt, D. and H. A. Larreguy. 2018. "Agents of the Regime? Traditional Leaders and Electoral Politics in South Africa." The Journal of Politics, 80(2), 382-399. Deininger, K. and R. Castagnini. 2006. "Incidence and Impact of Land Conflict in Uganda." Journal of Economic Behavior & Organization, 60(3), 321-345. Deininger, K. and S. Jin. 2006. "Tenure Security and Land-Related Investment: Evidence from Ethiopia." European Economic Review, 50(5), 1245-1277. Deininger, K., D. A. Ali, S. Holden and J. Zevenbergen. 2008. "Rural Land Certification in Ethiopia: Process, Initial Impact, and Implications for Other African Countries." World Development, 36(10), 1786-1812. Deininger, K. and A. Goyal. 2012. "Going Digital: Credict Effects of Land Registry Computerization in India." Journal of Development Economics, 99(2), 236-243. Deininger, K., A. Goyal and H. K. Nagarajan. 2013. "Women's Inheritance Rights and Intergenerational Transmission of Resources in India." Journal of Human Resources, 48(1), 114-141. Deininger, K. and F. Xia. 2017. "Assessing Impacts of Large Scale Land Transfers: Challenges and Opportunities in Malawi’s Estate Sector." Policy Research Working Paper, World Bank, Washington DC. Deininger, K., F. Xia and S. Holden. 2019. "Gendered Incidence and Impacts of Tenure Insecurity on Agricultural Performance in Malawi’s Customary Tenure System." The Journal of Development Studies, 55(4), 597-619. Delpierre, M., C. Guirkinger and J.-P. Platteau. 2019. "Risk as Impediment to Privatization? The Role of Collective Fields in Extended Agricultural Households." Economic Development and Cultural Change, 67(4), 863-905. 24 Dillon, B. and A. Voena. 2018. "Widows' Land Rights and Agricultural Investment." Journal of Development Economics, 135, 449-460. Eck, K. 2014. "The Law of the Land: Communal Conflict and Legal Authority." Journal of Peace Research, 51(4), 441-454. Fenske, J. 2011. "Land Tenure and Investment Incentives: Evidence from West Africa." Journal of Development Economics, 95(1), 137-156. Field, E. 2007. "Entitled to Work: Urban Property Rights and Labor Supply in Peru." Quarterly Journal of Economics, 122(4), 1561-1602. Goldstein, M. and C. Udry. 2008. "The Profits of Power: Land Rights and Agricultural Investment in Ghana." Journal of Political Economy, 116(6), 980-1022. Goldstein, M., Houngbedji, K., Kondylis, F., O'Sullivan, M., & Selod, H. (2018). "Formalization without certification? Experimental evidence on property rights and investment." Journal of Development Economics, 132, 57-74. Greiner, C. 2017. "Pastoralism and Land-Tenure Change in Kenya: The Failure of Customary Institutions." Development and Change, 48(1), 78-97. Harari, M. 2019. "Women’s Inheritance Rights and Bargaining Power: Evidence from Kenya." Economic Development and Cultural Change, 68(1), 189-238. Henn, S. J. 2020. "The State, Chiefs, and Development: Evidence from Africa." Working Paper, Boston University, Boston, MA. Holden, S. and R. Lunduka. 2012. "Do Fertilizer Subsidies Crowd out Organic Manures? The Case of Malawi." Agricultural Economics, 43(3), 303-314. Honig, L. 2017. "Selecting the State or Choosing the Chief? The Political Determinants of Smallholder Land Titling." World Development, 100, 94-107. Hunt, D. 2004. "Unintended Consequences of Land Rights Reform: The Case of the 1998 Uganda Land Act." Development Policy Review, 22(2), 173-191. Jacoby, H. G., G. Li and S. Rozelle. 2002. "Hazards of Expropriation: Tenure Insecurity and Investment in Rural China." American Economic Review, 92(5), 1420-1447. Jacoby, H. G. and G. Mansuri. 2008. "Land Tenancy and Non-Contractible Investment in Rural Pakistan." Review of Economic Studies, 75(3), 763-788. Jayachandran, S. 2015. "The Roots of Gender Inequality in Developing Countries." in K. J. Arrow and T. F. Bresnahan (eds.), Annual Review of Economics. Volume 7., Palo Alto, CA: Annual Reviews. Kang, M., B. Schwab and J. Yu. 2020. "Gender Differences in the Relationship between Land Ownership and Managerial Rights: Implications for Intrahousehold Farm Labor Allocation." World Development, 125, 104669. Khataza, R. R. B., A. Hailu, M. E. Kragt and G. J. Doole. 2017. "Estimating Shadow Price for Symbiotic Nitrogen and Technical Efficiency for Legume-Based Conservation Agriculture in Malawi." Australian Journal of Agricultural and Resource Economics, 61(3), 462-480. Krah, K., H. Michelson, E. Perge and R. Jindal. 2019. "Constraints to Adopting Soil Fertility Management Practices in Malawi: A Choice Experiment Approach." World Development, 124, 104651. Kumar, N. and A. Quisumbing. 2012. "Inheritance Practices and Gender Differences in Poverty and Well-Being in Rural Ethiopia." Development Policy Review, 30(5), 573-595. Kumar, N. and A. R. Quisumbing. 2015. "Policy Reform toward Gender Equality in Ethiopia: Little by Little the Egg Begins to Walk." World Development, 67, 406-423. Lawry, S., et al. 2016. "The Impact of Land Property Rights Interventions on Investment and Agricultural Productivity in Developing Countries: A Systematic Review." Journal of Development Effectiveness, 1-21. Leeson, P. T. and C. Harris. 2018. "Wealth-Destroying Private Property Rights." World Development, 107, 1-9. Lentz, C. 2010. "Land Inalienable? Historical and Current Debates on Land Transfers in Northern Ghana." Africa, 80(1), 56-80. Lovo, S. 2016. "Tenure Insecurity and Investment in Soil Conservation. Evidence from Malawi." World Development, 78, 219-229. Luke, N. and K. Munshi. 2011. "Women as Agents of Change: Female Income and Mobility in India." Journal of Development Economics, 94(1), 1-17. Lunduka, R. 2009. "Land Rental Markets, Investment and Productivity under Customary Land Tenure Systems in Malawi." PhD Thesis, Norwegian University of Life Sciences, Aas. Meinzen-Dick, R., A. Quisumbing, C. Doss and S. Theis. 2019. "Women's Land Rights as a Pathway to Poverty Reduction: Framework and Review of Available Evidence." Agricultural Systems, 172, 72-82. Melesse, M. B., A. Dabissa and E. Bulte. 2018. "Joint Land Certification Programmes and Women's Empowerment: Evidence from Ethiopia." Journal of Development Studies, 54(10), 1756-1774. Menon, N., Y. van der Meulen Rodgers and H. Nguyen. 2014. "Women's Land Rights and Children's Human Capital in Vietnam." World Development, 54, 18-31. Muchomba, F. M. 2017. "Women’s Land Tenure Security and Household Human Capital: Evidence from Ethiopia’s Land Certification." World Development, 98, 310-324. Newman, C., F. Tarp and K. van den Broeck. 2015. "Property Rights and Productivity: The Case of Joint Land Titling in Vietnam." 91(1), 91-105. Peterman, A. 2012. "Widowhood and Asset Inheritance in Sub-Saharan Africa: Empirical Evidence from 15 Countries." Development Policy Review, 30(5), 543-571. 25 Peters, P. E. 2010. ""Our Daughters Inherit Our Land, but Our Sons Use Their Wives' Fields": Matrilineal-Matrilocal Land Tenure and the New Land Policy in Malawi." Journal of Eastern African Studies, 4(1), 179-199. Place, F. and K. Otsuka. 2001. "Tenure, Agricultural Investment, and Productivity in the Customary Tenure Sector of Malawi." Economic Development and Cultural Change, 50(1), 77-99. Restuccia, D. and R. Santaeulalia-Llopis. 2015. "Land Misallocation and Productivity." University of Toronto, Department of Economics, Working Papers. Ricker-Gilbert, J., C. Jumbe and J. Chamberlin. 2014. "How Does Population Density Influence Agricultural Intensification and Productivity? Evidence from Malawi." Food Policy, 48, 114-128. Roodman, D., J. G. MacKinnon, M. O. Nielsen and M. D. Webb. 2018. "Fast and Wild: Bootstrap Inference in Stata Using Boottest." Queen's University, Department of Economics, Working Papers: 1406. Sesmero, J., J. Ricker-Gilbert and A. Cook. 2018. "How Do African Farm Households Respond to Changes in Current and Past Weather Patterns? A Structural Panel Data Analysis from Malawi." American Journal of Agricultural Economics, 100(1), 115- 144. Telalagic, S. 2014. "Kinship and Consumption: The Effect of Spouses' Outside Options on Household Productivity." University of Oxford, Department of Economics, Economics Series Working Papers: 720. Twyman, J., P. Useche and C. D. Deere. 2015. "Gendered Perceptions of Land Ownership and Agricultural Decision-Making in Ecuador: Who Are the Farm Managers?" Land Economics, 91(3), 479-500. van Donge, J. K. 1999. "Law and Order as a Development Issue: Land Conflicts and the Creation of Social Order in Southern Malawi." Journal of Development Studies, 36(2), 48-70. Vélez, M. A., et al. 2020. "Is Collective Titling Enough to Protect Forests? Evidence from Afro-Descendant Communities in the Colombian Pacific Region." World Development, 128, 104837. Wang, S.-Y. 2014. "Property Rights and Intra-Household Bargaining." Journal of Development Economics, 107, 192-201. Widman, M. 2014. "Land Tenure Insecurity and Formalizing Land Rights in Madagascar: A Gender Perspective on the Certification Program." Feminist Economics, 20(1), 130-154. Zepeda, G. 2000. "Transformación Agraria. Los Derechos De Propriedad En El Campo Mexicano Bajo El Nuevo Marco Institucional." CIOAC, Mexico. 26