Property rights? What property rights?


This page in:

Whom you ask matters for the answers you get. And this matters for what your data will tell you.  I’ve talked about this before (e.g. here), but today, let’s take a look at an interesting new paper by my colleagues Talip Kilic, Heather Moylan, and Gayatri Koolwal which unpacks how to get better information on property rights. 

The headline here is that asking adults individually gets you a significantly different quantity and configuration of property rights than asking the “best-informed” individual (which is the most common approach currently).  If you’re interested in the nitty gritty (and some surprising differences), read on.

The setting is Malawi, where Kilic and co. have the coincident timing (in 2016) of an LSMS+ supported national panel survey (IHPS) aimed at getting individual ownership of and rights to  land, as well as a national cross-sectional household survey (IHS4). 

Kilic and co. are focusing on 4 dimensions of rights on agricultural land:

·       Reported ownership: who is the owner of this parcel

·       Economic ownership: who would control the money from sale or rental

·       Right to sell

·       Right to bequeath

·       (a fifth dimension, documented ownership drops out of the analysis since so few people have documented ownership)

For the IHPS, a list of parcels owned by any household member was assembled as part of the household interview. Then, in private interviews (up to 4) each adult was asked separately about their personal ownership of and rights to each parcel. In the IHS4, the enumerator asked the self-identified most knowledgeable household member to identify owners and right holders for each parcel: aka the business as usual approach.

To start with, you get a very different picture of whether an individual owns land depending on method. For example, in the business as usual approach 80 percent of female household heads are reported (exclusive) owners, but only 63 percent of female heads own any land (exclusively) if you use the individual interview approach. Turning to economic ownership, it flips, with 37 percent of female heads as exclusive economic owners in the business as usual and 44 percent in the individual interview approach. Both the right to bequeath and right to sell follow the ownership with the business as usual (IHS4) approach giving higher rates for heads.

Part of the story on higher ownership from the business as usual approach comes from men. Among all men, the IHS4 gives us 27 percent having exclusive ownership, while in the IHPS when adults are interviewed separately, this drops to 18 percent. Interesting Malawi fact: both surveys show higher land ownership of just about every type for women versus men. The one exception to this is for non-household head women. Here, business as usual tells us that they are less likely to be exclusive owners than men overall, but the individual interview approach tells us they are more likely to be owners. Apparently the most knowledgeable person has a bit of a gender bias.  And the two approaches are telling us (statistically significant) different stories.

This is referring to exclusive ownership. In case you are wondering, this isn’t a case where one survey is picking up more joint ownership versus exclusive ownership – joint ownership shows no meaningful distinctions across the two surveys for all of the ownership variables.

So that’s what it looks like across households. One added bonus of the IHPS approach of asking all individual adults is that it lets us look at things within households. Recall that a roster of all parcels is created and then folks are asked (individually and privately) whether they own them (and with whom, if they report that there are also other owners).  For the most part husbands and wives agree – to the tune of 71 percent of the time on reported ownership. But 6 percent of the time the husband says it’s exclusively his and the wife says it’s joint, 5 percent of the time the husband says it’s a joint parcel, but she says it’s hers, and 2 percent of the time he says it’s his, she says it’s hers. But what’s really striking is that 8 percent of the time the husband says it’s joint and the wife says she has no rights, while 9 percent of the time she says its joint and he says he has no rights. So there are a bunch of parcels (17 percent) that wouldn’t register if you were just interviewing one of the two (despite their spouse thinking they did have some claim to the parcel). 

The results for economic ownership (controlling the proceeds from sale or rental) are starker – here spouses agree only 58 percent of the time, with larger cases of a man or woman denying any rights when the spouse says things are joint. Perhaps less surprising: there is much less disagreement on the right to sell (12 percent) and right to bequeath (14 percent).

Kilic and co. spend some time looking for correlates for the differences in reporting across the two surveys and then they zoom in on the joint-deny pattern in the individual interviews within households. Looking at the case where the husband says its joint, but the wife denies any reported ownership, this is less likely in female headed households, in larger households, when the wife is listed as a decision maker, and when the household has electricity. Flipping it around (wife says joint, husband denies), this is more likely when she suffers from a chronic illness and when the household has piped water (both significant at 10 percent).  So, not much clarity – although there is some suggestion that the woman is more likely to claim rights relative to the husband in households with more wealth (the water and the electricity).  But definitely more to unpack and understand here.

The bottom line is that to get accurate landholdings for the household, we likely need to interview more than the most knowledgeable person. And more than one person. On to the next question.



Markus Goldstein

Lead Economist, Africa Gender Innovation Lab and Chief Economists Office

Join the Conversation

Sarah Hughes
May 21, 2021

Very interesting results! I looked at a couple of the reports but could not find the average duration for the individual-interview component. Is a methodological report forthcoming? The paper includes a footnote that indicates administering the individual module adds one day per 16-household EA but since the team used CAPI, the average duration of the survey administration should be readily calculable, no?