167
B. Results
The results are presented by hypothesis rather than by dependent variable, although references
to specific tenure-agricultural indicator effects are noted. The results which were found to be
significant are indicated in table 5.10. Insignificant or absent relationships are left blank.
Hypothesis 1: Matrilineal inheritance systems
The case of mode of acquisition is complex owing to its multiple possibilities. The significant
findings are from Southern province as only one minor finding was obtained in the case of Eastern
province. As opposed to households who acquired their land through matrilineal descent patterns,
households who acquired their land patrilineally were more likely to use credit and to make earth
structures. Those who acquired their land through the chief or headman were more likely to make
organic inputs, earthwork structures, or to plant fruit trees (than those who acquired matrilineally).
Lastly, those who purchased land or were allocated land from the state were more likely to own oxen,
use organic inputs, and to make earthwork structures. These results support the hypothesis in that
matrilineal land was found to be inferior to at least one other type of land acquisition for most of the
indicators. On the other hand, a superior mode of acquisition does not emerge from the data. The lone
result from Eastern province did not support the null hypothesis: the use of formal credit was greater
on farms with matrilineal land than on those who acquired from the chief or headman. 27
Hypothesis 2: Land alienation rights
The right to alienate land outside the family was significant in only a handful of cases. The
right to permanently transfer land was positively related to the use of organic inputs in Southern
province, but was negatively related to multipurpose tree planting in the same site and to fruit tree
planting in Eastern province. Given the insignificant and mixed nature of the results, it is difficult to
draw any strong conclusions about the effects of land rights. The negative relationships with respect
to tree planting suggests that tree investment behavior differs from other improvements. One
explanation for this is that there tends to be a noticeable input of tree planting project personnel into
the observed pattern of on-farm tree planting.
Hypothesis 3: Perception of landownership
In Eastern province, households who felt they owned their land were more likely to own oxen,
to make earthwork improvements, and to plant fruit trees. In Southern province the ownership variable
was positively linked to use of credit and, again, earthwork structures. The results are supportive of
the hypothesis. Because ownership is a perception of households, it may be that it is influenced by
certain peculiarities of households (i.e., it is endogenous). If so, there may be other important
variables which explain both the perceptions of tenure security and the agricultural indicators. To
attempt to elucidate these issues, a regression was made using ownership (household or not) as a
dependent variable with several household and farm characteristics as independent variables. A strong
relationship was found between household ownership and wealth indicators (including number of cattle
and farm size) raising the possibility that the wealth of a household may lead both to more individual
claims of tenure control and to more investments.
This may be explained by the fact that the latter households are more likely to come from outside the area.