Does Stronger Land Tenure Security Incentivize Smallholder Climate-smart

Agriculture?Initial Evidence from Zambia’s Eastern

Province

Lauren Persha1, M. Mercedes Stickler2, Heather Huntington3

1Department of Geography, University of North Carolina at Chapel Hill2Land Tenure and Resource Management Office, USAID

3The Cloudburst Consulting Group, Inc

Motivation• Resurgence in promoting agroforestry in SSA: to meet food

security challenges and CC adaptation objectives for poor rural farmers

• Many small-scale studies show a range of benefits to smallholders from agroforestry:– Improved soil fertility & crop yields– Risk smoothing via crop diversification; increased or more reliable

income– Increased availability of wood products for HH use (e.g., fuelwood,

fodder)

• Yet, uptake in SSA has been persistently low– Cash, resource and inputs constraints

• e.g., Labor & credit availability; Farm size– Insufficient technical knowledge– Incompatible land management practices within villages

• e.g., Communal livestock browsing during past-harvest season– Broader cultural, demographic, institutional factors– Insecurity over rights to land

Motivation• To what extent does land tenure insecurity serve as a barrier to

agroforestry uptake, and wider implementation of other climate-smart agricultural practices?– Widely hypothesized, but few studies show a definitive link– Difficult to rigorously test at scale (challenges of piloting tenure interventions;

introducing experiments)– Empirical studies have substantial endogeneity challenges to overcome

• Widely varying results across existing empirical work

• Methods challenges also contribute to equivocal knowledge base– Different definitions & measures of tenure security– Different empirical strategies for analysis– Generally, small-N studies: few villages, couple hundred HHs– Household level rather than field level data– No clear consensus; little understanding of how this relationship might vary

across different country, socio-cultural, policy contexts

USAID’s Tenure and Global Climate Change (TGCC) Project• 2-year RCT in 5 chiefdoms of Chipata District, Eastern Province• Cross-cutting tenure and agroforestry interventions aimed at increasing the

adoption of CSA• Opportunity to test relative contributions of strengthening customary land

governance and agroforestry extension support on HH tenure security and CSA adoption

– IE focuses on identifying effects of village and household level interventions

• Tenure intervention focuses on:– Establishing Village Land Committees– Participatory mapping to aid land allocation– Support dissemination of land management rules agreed at chief level– Capacity building to headmen, indunas, VLCs, around land administration

processes, including land dispute resolution and administration of customary land certificates

– Facilitate implementation of customary land certificates (at chief’s consent)– HH-level paralegal extension support around land rights, certification and dispute

resolution

• Agroforestry intervention focuses on:– Extension support around establishment of 3 agroforestry species: Faidherbia albida (msangu); Gliricidia sepium; Cajanus cajan (Pigeon pea)

TGCC Project: A Randomized Control Trial of Land Tenure Strengthening and Agroforestry Extension Impacts on Smallholder Adoption of Climate-Smart Agriculture

Eligible Chiefdoms

TGCC Chiefdoms

ControlVillages

AgroforestryVillages

Land TenureVillages

Agroforestry +

Land TenureVillages

Control

Agroforestry

Land Tenure

Agroforestry and Land Tenure

Randomization Step

ChipataDistrict

Baseline Data Collection

• Baseline completed Aug 2014; Endline expected Aug 2017; True panel survey

• Qualitative & Quantitative:– 3,500 households surveyed across

294 villages in Chipata District– Headperson survey– Key informant interviews:

• agricultural extension officers; individuals involved in local land issues (e.g., Indunas, village elders)

– Focus groups with women, youth, land-constrained HHs (45 villages)

– Nested data collection across fields, HHs, villages

Baseline Analyses: Research Questions

1. How secure are the rights of Eastern Province smallholders over the land they use, under Zambia’s prevailing dual customary and state tenure system?

2. What are household level determinants of stronger tenure security?

3. What role does tenure security play in shaping household land use strategies and CSA investments (and particularly with respect to agroforestry uptake)?

4. Heterogeneous outcomes: How do these results differ for female-headed and poorer households?

To what extent does land tenure insecurity serve as a barrier to agroforestry uptake, and wider implementation of other climate-smart agricultural practices?

Empirical Strategy: Baseline Data

• Multi-level mixed effects models to identify determinants of:– Tenure security– Longer return, Intensive land investments– Shorter return, less intensive land investments– Agroforestry uptake specifically

• Tenure security measured at the field level, as expropriation risk: HH expectation of losing access to the field over (1) the next 1-3 years, and (2) beyond 3 years from now.

• Two tenure security measures:1. Index of HH perceived risk of expropriation for each field; responses across 6 different

groups:» Extended family; » Other HHs in village; » HHs in neighboring villages; » Village headperson; » Chief; » Elites from outside the village.

2. Number of seasons a HH felt they could let the field lie fallow without worrying about it being reallocated to another HH or for other use

DescriptiveStatistics

Variable Name Variable Description Mean SDDependent Variables

Tenure Security Index Mean score of 12 questions assess ing l i kel ihood of encroachment, confi scation or rea l location by 6 sources , over short (1-3 yrs ) and longer term (4+ yrs ), on a 6-point Likert sca le.

1.73 0.93

Fallow Length (0-2) 3-point interva l variable indicati ng how many seasons respondent could let a fi eld l ie fa l low without worrying about the fi eld being rea l located. 0= 0 seasons ; 1 = 1-5 seasons ; 2 = s ix or more seasons .

1.14 0.88

Long Return Investment (Y/N) 1 = indicates the presence of fencing, i rrigation or bas in planting on the fi eld. 0.11 0.31

Short Return Investment (# of investments) Number of short-term investments hh made on fi elds (e.g, planting ridges , zero ti l lage, herbicide, manure). 2.37 1.17

Agroforestry Uptake (Y/N) 1 = planti ng of agroforestry trees on a fi eld. 0.05 0.21

Independent Variables

Slope Interva l variable to assess the s teepness of a fi eld; ranging from fl at to s teep 1.41 0.53Soil Quality 4-point variable based on soi l texture, ranging from loamy to clay. 2.61 1.02Prior Land Dispute (Y/N) 1 = a land dispute has occurred on the fi eld. 0.11 0.32Length of landholding (yrs) Number of years that the household has owned or used the fi eld. 24.9 23.9Primary field use for subsistence (Y/N) 0 = crops grown on this fi eld are sold or bartered; 1 = crops grown on this fi eld are used for subs is tence. 0.64 0.48Field acquisition by inheritance (Y/N) 0 = fi eld was acquired through vi l lage a l location; 1 = HH inheri ted the fi eld through parents . 0.67 0.47Field replanted during prior season (Y/N) 1 = fi eld was replanted during previous agricul tura l season. 0.42 0.49Field distance to household (meters) Distance in meters from the household to the fi eld. 123.9 304.5Largest field used by household (Y/N) 1 fi eld i s the largest fi eld used by the household. 0.35 0.48

Relationship to chief (Y/N) 1 = the head of household or primary spouse i s related to the chief. 0.21 0.41Governance satisfaction index PCA measure of satisfaction with local land governance, based on 5 Qs that ask respondents to evaluate local leaders in terms of:

decision-making transparency, fairness in land allocation, accountability and commitment to protecting community land. 0.00 1.71

Total land farmed by household (ha) Tota l land area farmed by hh, in hectares . 2.03 1.88Household in poorest quartile (Y/N) 1 = household i s in lowest 25% of an asset-based weal th index run across a l l surveyed households . 0.24 0.43Highest education level of any hh member (yrs) Highest educati on level of any household member. 8.82 3.37Female headed (Y/N) 1 = female-headed household. 0.22 0.42Head residency time in vil lage (yrs) Number of years that the head of household has been a res ident of the vi l lage. 41.6 17.9Inheritance pattern 3-level indicator for whether hh i s defi ned by a matri l inea l , patri l inea l or mixed/unclear inheri tance regime. 1.84 0.53Labor availabil ity Number of adul ts in the household that are ava i lable for fi eld labor. 3.07 1.63Head hires labor out to other farms (Y/N) 1 = anyone in hh was involved in “ganyu” (hi red labor out to another fi eld) during past agricul tura l season. 0.43 0.50Head age (yrs) Age in years of the household head. 43.8 16.4Prior CSA extension exposure A count of the amount of CSA extens ion exposure received by hh in year prior to the survey. 2.09 2.64Prior agroforestry extension exposure (Y/N) 1 = the household received specifi c exposure to agroforestry extens ion. 0.31 0.46Land constrained household (Y/N) 1 = households uses < 1 hectare of land. 0.23 0.42

Village holds land-related meetings (Y/N) 1 = the vi l lage held at least one land related meeti ngs in the past year. 0.67 0.47

Village N 293Household N 3437Field Obs N 9745

Level 1 - Field

Level 2 - Household

Level 3 - Village

Descriptive Statistics

Chiefdom 1 Chiefdom 2 Chiefdom 3 Chiefdom 4 Chiefdom 5Dependent Variables

Tenure Security Index 1.82 1.69 1.72 1.68 1.75Fallow Length (0-2) 1.19 1.16 1.14 1.07 1.14Long Return Investment (Y/N) 0.12 0.10 0.11 0.10 0.11Short Return Investment (Number of investments) 2.39 2.35 2.36 2.38 2.38Agroforestry Uptake (Y/N) 0.05 0.04 0.05 0.05 0.04

Independent Variables

Slope 1.45 1.42 1.49 1.34 1.38Soil Quality 2.73 2.61 2.69 2.46 2.58Prior Land Dispute (Y/N) 0.11 0.11 0.11 0.13 0.10Length of landholding (yrs) 24.67 25.83 23.71 25.12 24.41Primary field use for subsistence (Y/N) 0.64 0.63 0.64 0.64 0.62Field acquisition by inheritance (Y/N) 0.68 0.66 0.67 0.71 0.66Tenure security index 1.82 1.69 1.72 1.68 1.75Field replanted during prior season (Y/N) 0.44 0.44 0.45 0.36 0.38Field distance to household (meters) 88.0 135.7 137.9 127.3 126.0Largest field used by household (Y/N) 0.35 0.35 0.35 0.36 0.35

Relationship to chief (Y/N) 0.23 0.21 0.20 0.21 0.19Governance satisfaction index -0.05 0.03 -0.07 -0.05 0.11Total land farmed by household (ha) 2.17 2.02 2.15 1.72 2.08Household in poorest quartile (Y/N) 0.26 0.24 0.24 0.25 0.22Highest education level of any household member (yrs) 8.68 8.70 8.70 9.06 9.00Female headed (Y/N) 0.17 0.22 0.23 0.23 0.26Head residency time in vil lage (yrs) 38.22 43.45 40.58 42.29 42.59Inheritance pattern 1.97 1.81 1.75 1.85 1.80Labor availabil ity 3.11 3.07 2.96 3.12 3.09Head hires labor out to other farms (Y/N) 0.47 0.42 0.46 0.40 0.40Head age (yrs) 42.69 45.42 43.31 43.21 43.70Prior CSA extension exposure 2.31 2.10 2.01 1.96 2.09Prior agroforestry extension exposure (Y/N) 0.33 0.31 0.33 0.33 0.28Land constrained household (Y/N) 0.20 0.24 0.21 0.29 0.20

Village holds land-related meetings (Y/N) 0.69 0.66 0.69 0.67 0.66

Village Cluster N 59 85 56 49 45Household N 641 980 564 631 621Field Obs N 1824 2792 1598 1742 1789

Level 1 - Field

Level 2 - Household

Level 3 - Village

Results: Determinants ofTenure Security

Model 1 Model 2

Independent Variables Tenure Securitya Fallow Lengthb

Fixed Effects

Slope 0.079 (0.020)*** 0.895 (0.043)**Soil Quality 0.023 (0.011)** 1.060 (0.029)**Prior Land Dispute (Y/N) 0.428 (0.037)*** 0.791 (0.069)***Length of landholding (yrs) -0.003 (0.000)*** 1.01 (0.001)***Primary field use for subsistence (Y/N) -0.101 (0.023)*** 0.815 (0.047)***Field acquisition by inheritance (Y/N) 0.073 (0.026)*** 1.10 (0.068)

Relationship to chief (Y/N) 0.009 (0.030) 1.20 (0.101)**Governance satisfaction index 0.003 (0.007) 0.982 (0.019)Total land farmed by household (ha) 0.024 (0.006)*** 0.982 (0.017)Household in poorest quartile (Y/N) -0.016 (0.028) 0.889 (0.069)Highest education level of any household member (yrs) 0.000 (0.004) 0.997 (0.010)Female headed (Y/N) -0.002 (0.032) 0.957 (0.072)Head residency time in vil lage (yrs) 0.000 (0.001) 1.000 (0.002)Inheritance pattern 0.943 (0.058) patrilineal 0.045 (0.029) mixed inheritance / unclear 0.077 (0.052)

Village holds land-related meetings (Y/N) 0.011 (0.025) 0.904 (0.062)Chiefdom indicator 0.958 (0.029)Chiefdom 1 dummy 0.104 (0.045)**Chiefdom 2 dummy -0.051 (0.040)Chiefdom 3 dummy -0.038 (0.046)Chiefdom 4 dummy -0.041 (0.042)

Village Cluster N 293 293Field Obs N 9745 9745

Note: Robust standard errors reported in parentheses. a Higher index score indicates greater tenure insecurity. b Ordered logit with highest category indicating greatest tenure security. Estimates reported as odds ratios.* p < 0.10; ** p < 0.05; *** p < 0.01

Level 1 - Field

Level 2 - Household

Level 3 - Village

Results: Longer ReturnLand Investments

Model 3 Model 4 Model 5

Independent Variables

Long Return

Investmentb

Short Return

Investmentc Agroforestryb

Fixed Effects

Tenure security indexa 0.913 (0.044)* 1.005 (0.005) 0.975 (0.063)Field replanted during prior season (Y/N) 1.166 (0.142)Field distance to household (km) 1.000 (0.000) 1.000 (0.000)Largest field used by household (Y/N) 2.522 (0.225)*** 1.062 (0.010)*** 3.12 (0.434)***

Relationship to chief (Y/N)Governance satisfaction index 1.003 (0.023) 0.999 (0.002) 0.984 (0.032)Household in poorest quartile (Y/N) 0.854 (0.086) 0.999 (0.010) 0.987 (0.130)Highest education level of any household member (yrs) 1.049 (0.014)*** 1.003 (0.011)*** 1.054 (0.022)**Female headed (Y/N) 0.903 (0.989) 1.008 (0.011) 0.970 (0.147)Head residency time in vil lage (yrs) 1.006 (0.005) 1.000 (0.000) 0.990 (0.006)*Labor availabil ity 1.043 (0.028) 1.008 (0.003)*** 1.09 (0.037)**Head hires labor out to other farms (Y/N) 1.190 (0.107)** 0.991 (0.008) 1.05 (0.130)Head age (yrs) 0.992 (0.005) 0.999 (0.001) 1.02 (0.007)**Prior CSA extension exposure (Y/N) 1.011 (0.015) 0.996 (0.002)**Prior agroforestry extension exposure (Y/N) 0.975 (0.117)Land constrained household (Y/N) 0.838 (0.095) 0.956 (0.011)*** 0.715 (0.121)**

Village holds land-related meetings (Y/N) 0.908 (0.081) 1.002 (0.009) 1.004 (0.119)Chiefdom 1 dummy 1.164 (0.177) 1.006 (0.014) 1.194 (0.224)Chiefdom 2 dummy 0.884 (0.113) 0.986 (0.013) 1.046 (0.183)Chiefdom 3 dummy 0.966 (0.162) 1.010 (0.016) 1.215 (0.244)Chiefdom 4 dummy 0.833 (0.122) 0.983 (0.014) 1.164 (0.223) Constant 0.051 (0.013)*** 2.479 (0.057)*** 0.007 (0.003)***

Village Cluster N 293 293 293Field Obs N 9745 9745 9745

Note: Robust standard errors reported in parentheses; Estimates reported as odds ratios.a Higher index score indicates greater tenure insecurity.b Logit model.c Poisson model.* p < 0.10; ** p < 0.05; *** p < 0.01

Level 1 - Field

Level 2 - Household

Level 3 - Village

Findings: Tenure Security, Land Disputes, Land Documentation

Land expropriation, documentation and tenure security:• Relatively high tenure security across the surveyed chiefdoms• Land expropriation events are uncommon (experienced by < 2% HHs)• But, household concern over this is high (15-25% of fields surveyed). Two main

sources of concern: – dispossession by chiefs for investment purposes; and – boundary disputes with other HHs in village

• HHs commonly experience land-related disputes, despite fairly high overall tenure security: – 26% of households (N=707) experienced a land conflict on at least one field in

past 3 years;– prior disputes recorded on 10% of fields surveyed (N = 1007 fields)

• 91% of HHs would like to obtain documentation over customary land they use (current holders of documentation are very uncommon; < 1% of households)

• Little evidence of strong marginalization around land issues for female-headed or poorer households

Land disputes and links with household tenure security over land:• Clear indication of dampening effect of land disputes on household perception of

security over land• Clear indication of concern over emergent challenges to land rights, such as

expropriation by chiefs for investments• Links to strong household-level interest in land documentation

Key Messages: Factors Associated With Tenure Insecurity

• Strong role of prior land disputes• Other key factors:

– Field acquisition by allocation rather than inheritance– Poor quality land: steeply sloped; poorer soils

• Little evidence of greater insecurity across traditionally vulnerable groups

• Little evidence of a moderating role for village-level governance factors• Role of elite networks: Some indication that HH ties to chief is

associated with lower insecurity, but not for headperson

Role of prior field-level disputes

02

04

06

08

01

00

% o

f Fie

lds

No prior disputes Prior dispute on this field

(within next 1-3 yrs)Households within the village will try to cross-over your boundaries

Impossible / would never happenHighly unlikelyUnsure / I don't know

LikelyVery likelyHappening right now

02

04

06

08

01

00

% o

f Fie

lds

Fields without prior disputes Fields with prior disputes

1 2 3 4 5 1 2 3 4 5

(within the next 1-3 yrs)Likelihood that other households within the village will try to cross over your boundaries

Impossible / would never happenHighly unlikelyUnsure / I don't know

LikelyVery likelyHappening right now

02

04

06

08

01

00

% o

f Fie

lds

1 2 3 4 5

(within the next 1-3 years)Likelihood that other households within the village will try to cross-over your boundaries

Impossible / would never happenHighly unlikelyUnsure / I don't know

LikelyVery likelyHappening right now

02

04

06

08

01

00

% o

f Fie

lds

1 2 3 4 5

(within next 1-3 yrs)Likelihood that the chief will give this field up for investment purposes

Impossible / would never happenHighly unlikelyUnsure / I don't know

LikelyVery likelyHappening right now

02

04

06

08

01

00

% o

f Fie

lds

Fields without prior disputes Fields with prior disputes

1 2 3 4 5 1 2 3 4 5

(within the next 1-3 yrs)Likelihood that the chief will give this field up for investment purposes

Impossible / would never happenHighly unlikelyUnsure / I don't know

LikelyVery likelyHappening right now

Findings: Agricultural Investments & Agroforestry

• Agroforestry uptake currently very low (11% of HHs (N=383); 5% of fields (N=404))

• Other upfront costly field investments also uncommon:• Planting basins: 10% of fields• Live fencing: 1% of fields• Drip irrigation: < 1% of fields

• Less costly CSA investments somewhat more common:• Zero tillage: 8% of fields• Manuring: 18% of fields• Fallowing: 7% of fields

• Ridging (85% of fields) and crop rotation (82% of fields) very common

Key Messages: Factors Associated With Agricultural Investment

• HHs strongly preference their largest field for such investments• Greater HH labor availability and education level have small but significant effects• Small but significant negative effect of tenure insecurity on HH likelihood to

undertake costly upfront land investments (but not agroforestry specifically)

• Significant factors for likelihood of agroforestry uptake:

• Field size (+)• HH labor availability (+)• Head age (+)• Education level of HH (+)• Land-constrained (-)

– Indication these HHs are significantly less likely to engage in agroforestry

• No evidence of a significant role for several factors often hypothesized in the literature:

• Quality of village-level land governance

• HH exposure to agricultural extension• Ties to elite networks• Wealth status• Female-headedness

Preliminary Overall Conclusions• Overall, relatively high tenure security across the surveyed chiefdoms• Collective understanding of cultural norms around customary land access,

allocation and inheritance appear to be strong overall and well-functioning, including mediating equitable access to land for traditionally vulnerable groups (also strongly supported by qualitative data)

• Traditional informal and customary norms over land access, allocation, inheritance appear to be strong

• Little evidence of strong marginalization around land issues for female-headed or poorer households.

However• Clear indication of dampening effect of land disputes on household perception of

security over land• Clear indication of concern over emergent challenges to land rights, such as

expropriation by chiefs for investments• Links to strong household-level interest in land documentation.

Role of stronger tenure security in promoting agroforestry?– Some indication of its effect on broader costly land investments– Currently very low uptake constrains the baseline analyses; – A need for more nuanced work (e.g., instrumenting as a next step)

Thank you.