Notes for panel discussion, "Ensuring Women's Land Access,"


Material Information

Notes for panel discussion, "Ensuring Women's Land Access,"
Physical Description:
Mixed Material
Katz, Elizabeth
Publication Date:


General Note:
Latin America Regional Workshop on Land Issues, May19-22, Pachuca, Hidalgo, Mexico.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
System ID:

This item is only available as the following downloads:

Full Text

Notes for panel discussion, "Ensuring Women's Land Access," Latin America
Regional Workshop on Land Issues, May 19-22, Pachuca, Hidalgo, Mexico

Elizabeth Katz*
St. Mary's College of California

Deere and Le6n's work (in this paper and elsewhere) has been extraordinarily influential
in putting gender issues high on the Latin American land policy agenda. Their careful
compilation and analysis of national-level legislation, policy, and data has squarely put
all of the important "macro"-level advances and concerns on the table. What I would
like to contribute here are some initial observations on several "micro" (household-level)
topics that hopefully help to flesh out both the determinants and impacts of female
property rights in rural Latin America. Specifically, I would like to draw on new
household survey data from Nicaragua and Honduras to address several of the points
made by Deere and Le6n in their concluding section:

(1) Deficiencies in gender-disaggregated national-level data on land ownership;
(2) The possible variation by social class in the gender gap in the distribution of land;
(3) The potential impact of gender differences in land ownership on rural poverty;
(4) The relationship between women's land ownership, control over land use decisions
and land-related income; and
(5) The lack of empirical testing of the potential positive bargaining power effects of
women's land ownership.

' Invaluable research assistance was provided by Juan Sebastian Chamorro of the Department of Applied
and Agricultural Economics at the University of Wisconsin Madison. The empirical analysis presented
here is preliminary in nature and should not be cited. The financial support of the U.S. Agency for
International Development's BASIS CRSP, with support from the LAC Bureau, for gender analysis of the
Honduras data is also gratefully acknowledged.

(1) Deficiencies in gender-disaggregated national-level data on land

As the authors mention in the paper, since Latin American agricultural censuses do not
generally collect information on the sex of the landowner, household surveys can be an
important alternative source of data although they are not always nationally
representative. Two recent household surveys in Nicaragua and Honduras took special
care to record the identity (including sex and relationship to the household head) of
each landowner in the household. It should be emphasized that this extremely valuable
information is practically costless to obtain it only requires one additional cell in which
the interviewer records the household identification code of the landowner, which can
then be linked to any other individual characteristics the researcher may be interested
in, such as age, education, and labor market status.

The Nicaragua data, which are national in scope, cover 2476 rural households
(corresponding to 3662 land parcels) from the year 2000. They suggest that 79% of all
documented parcels (n = 2382) were owned by men, 16% by women, and 4% were
jointly titled. Considered from a household perspective, in only ten percent of male-
headed households do women have some form of land rights (including individual and
joint titles)', while women own land in 73% of female-headed households2 (see Table

The Honduran data, collected during 2001, while not national in coverage and with a
significantly smaller sample size (n = 850 households and 2500 parcels, which is
reduced to 702 households and 2071 parcels once landless families and undocumented
parcels are removed from the analysis), offers an interesting comparative picture from a

1 In male-headed households with some female land rights, women hold on average approximately 43%
of the total household land. In this same group, about one-third have exclusively female land ownership,
another third have separate individual ownership by men and women, and the remainder have joint
male-female ownership.

neighboring Central American country. Here, 26% of households with owned and
documented land have some form of women's land rights. Breaking it down by the
gender of the household head, in 16% of male-headed households and 66% of female-
headed households, women have individual and/or joint claims to at least some portion
of the family farm (see Table 1). In the case of male-headed households, where
women have property rights in land, they own on average 42.5% of total household
land, and in 70% of cases are individual owners separate from their spouses. This
latter point suggests the importance of measuring ownership at the level of the parcel,
since there may exist significant intra-household/farm heterogeniety in property rights,
with different members having individual daims over different pieces of the "family
farm." Women in female-headed households own an average of 88% of total land,
mainly as the exclusive landowner in their household.

(2) The possible variation by social class in the gender gap in the
distribution of land

The data from Nicaragua and Honduras allow for an examination of hypothesis that
households occupying different places along the rural socio-economic spectrum may
also have different intra-household distributions of property rights by gender.

Quintile Analysis. We initially approached this question by dividing the sample into land
and income quintiles, and then examining two measures of gender-specific land rights -
a dummy variable set equal to one if the woman has either individual or joint title, and
the percentage of total farm land owned by a women either individually or jointly --
across the quintiles. The results are presented in Table 2.

For both male- and female-headed households in Nicaragua, there appears to be some
evidence of a positive correlation between income/wealth and female property rights in

2 Of that 73%, women on average control 93% of household land, the vast majority as exclusive land
owners in the household.

land, although the patterns are not linear, and are likely to be sensitive to the way in
which both "class" and land rights are measured. For female-headed households, 84%
of women in the highest income group have some form of land rights, compared to
76% of women in the lowest income group; and women in the largest land quintile own
an average 63% of their total acreage compared to only 50% for the land poorest
group. Women have formal land rights in approximately 7% of the poorest male-
headed households and in 12-13% of the wealthiest, although their share of land only
increases from about 4% to 6% at either end of the income spectrum.

The Honduran data are more difficult to interpret. For female-headed households,
there is no clear trend in the relationship between either measure of wealth and
women's land rights. For male-headed households, while a significantly higher
percentage of women (25%) in the highest land quintile have land rights compared to
those in the land poorest households (10%), households sorted by income instead of
land show no such pattern.

Multivariate analysis. A multivariate exploration of the household-level determinants of
female land rights yields somewhat more interesting results. For Nicaragua, a probit
model of women's individual or joint property in land suggests that while farm size is
not a statistically significant correlate, a woman's age and level of education, along with
female household headship, are strongly positively associated with the likelihood that
she will have some claim on land. A similar (Tobit) model of the percentage of land
owned by women has almost identical results.3

These results (which are still preliminary) suggest that women's access to land is
mediated more through claims based on age (eg., inheritance), human capital (eg.,
labor market earnings) and headship status as opposed to membership in a household
with a certain level of claim on land resources. Indeed, as Deere and Leon have

3 Results are too preliminary to present in a separate table. Similarly specified models are also being
tested with the Honduras data.

found for the rest of the region, inheritance is significantly more important means of
land acquisition for women than for men: in Nicaragua, 37% of female-owned plots but
only 22% of male-owned plots were acquired through inheritance; the comparable
figures for Honduras are 40% for women and 19% for men. Somewhat surprisingly,
many women also use the land market to purchase their holdings: 47% of the land
owned by women in Nicaragua, and 57% of female-owned land in Honduras, was the
result of a sales transaction (see Table 3). While the data do not contain information
on the financing of past land purchases, it seems reasonable to believe that many of
these purchases would have been made out of savings generated by labor market
earnings (including rural "microenterprise"). Older women are therefore in a better
position to acquire land both from parental and spousal inheritance, as well as from
accumulated earnings from other income-generating activities.

(3) The potential impact of gender differences in land ownership on rural

While the analysis above considered the determinants of women's land rights, what
about the effects of women's land rights on poverty? A simplified (and preliminary)
multivariate model of total household income suggests that, controlling for assets and
some regional and household characteristics, as well as for unobservable factors that
may affect both women's land rights and income,4 households in which women have
either individual or joint rights in land generate significantly higher incomes than
households where women have no formal claims to land.5

Assuming that these results holds up to more careful testing, what mechanisms could
explain the positive association between female property rights and rural incomes? As
Agarwal (1994) and Deere and Le6n (2001) suggest (in somewhat different language),

4 A probit estimate of women's land rights (discussed above) was used to generate estimated values,
which were then used in place of observed values in the income model.
5The results are too preliminary to be included as a table.

it may be useful to divide the effects into two: property rights as economic access, and
property rights as social access. Economic access refers to the idea that the legal claim
to land allows the owner to control its use as a key input into agricultural production; as
a source of income from rental or sale; and as collateral for credit which can be used
for either consumption or investment purposes. Thus we could expect a positive
correlation between women's rights to land, and women's income generated either by
direct exploitation of the land resource (agricultural and livestock activities), and/or by
renting or lending out the land, and perhaps investing the proceeds in a non-
agricultural venture. We would also expect to see women with land rights making
greater use of the rural capital market. Social access refers to the ways in which
property rights may serve to empower women in their negotiations with other
household members, and with the community and society at large. "Bargaining power"
effects on household resource allocation could be expected to influence not only the
ways in which income is generated, but also the ways it is divided among different
investment and consumption expenditures. The following two sections address each of
these mechanisms economic and social access in turn.

(4) The relationship between women's land ownership, control over land
use decisions and land-related income

While the Central American data do not include measures of land use decisions, for
Nicaragua we do have information on who within the household "administers" the
income generated from crops and livestock, as well as off-farm income and credit
disaggregated by gender for both countries (see Table 4).

Looking first at agricultural income, it is clear that there is a positive correlation
between women's land rights and their control over income generated from crop and
livestock production. In Nicaragua (the only country for which we have disaggregated
agricultural income), women in male-headed households who have some form of land
rights administer over half of crop income a much larger proportion than their share

of total farmland. Their counterparts who have no land rights only control 14% of crop
income themselves, although they report greater joint administration (29% as
compared to 8%). The differences are less marked for control over livestock income:
regardless of land rights, women administer less than 10% themselves and about 30%
jointly with their husbands.6

Turning to off-farm income, the land rights effect on small business earnings seems to
exist in Honduras but not Nicaragua: in Honduras, women in male-headed households
who have some land rights generate almost half of "microenterprise" income, compared
to less than 20% generated by landless women. In both countries, women with land
contribute a significantly greater proportion of wage and salary income to the
household.7 Also in both countries, women with land rights are more likely to have
received credit, although the figures are quite low for all women in male-headed
households (ranging from 3 10% of women receiving credit during the reference
year). Somewhat surprisingly, female household heads especially those with land
rights appear to be fairly active capital market participants: almost one-quarter of
Nicaraguan and one-third of Honduran female household heads with property rights in
land received credit during the reference year.

(5) Bargaining power effects of women's land ownership

The Central American data allow for empirical testing of one of the central hypotheses
of the economic literature on intra-household resource allocation: that individual assets
can skew household preferences in favor of household members with relatively strong
"fall back positions" or "threat points." While we are still in the very early stages of this

6 Women in female-headed households who have land rights, however, control about half of all livestock
' In all of these comparisons, it should be kept in mind that land rights are correlated with women's age
and education, which may also be contributing factors to their greater control over income. A simple
multivariate model of Nicaraguan women's share of farm income suggests that, controlling for age,
education, headship, and total household land and income, women with land rights administer 3% more

portion of the research, it seems worthwhile to share the goals of the work, especially
since to our knowledge analysis of this kind has not previously been carried out with
data from the Latin American region.

We plan to explore the impact of women's property rights on two indicators of
household and child welfare: food expenditures and children's education. Our
hypothesis is that, controlling for other relevant determinants, households in which
women hold land will allocate a greater proportion of their income towards food, and
will dedicate more resources to their children's education. Because many of the same
factors that influence these outcomes (such as women's education) are also
determinants of women's land rights, and because there may also be unobserved
individual or household characteristics that affect both sets of variables (female
property rights and intra-household resource allocation), appropriate statistical
techniques will be employed to control for this kind of "selection" bias.

In conclusion, we look forward to complementing Deere and Le6n's important work with
empirical microeconomic analysis that will hopefully shed greater light on the
household-level determinants and impacts of the changing gender distribution of land
rights in rural Latin America.

crop and livestock income than women with no land rights. The result is significant at the 10% level and
is likely to change as further modifications to the model are introduced.

Table 4: Income Sources and Receipt of Credit, by Female Land Rights and Household Headship

Male-headed Households Female-headed households Male-headed Households Female-headed households
Some female land rights No female land rights Some female land rights No female land rights Some female land rights No female land rights Some female land rights No female land rights

Farm Production


Total US

Total US

Off-farm Income

Off-farm business Income
Total US

Off-farm wage and salary income
Total US

Received credit in reference year
% 30%
Average amount ($)

% 9%
Average amount 100






























Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EZCNNZ11Y_R89GO1 INGEST_TIME 2013-08-27T23:57:51Z PACKAGE AA00016342_00001