ВУЗ: Не указан
Категория: Не указан
Дисциплина: Не указана
Добавлен: 11.04.2021
Просмотров: 738
Скачиваний: 1
for large numbers of linguistically distinct sub-tribes, so focusing on members of these tribes allows us to
examine which level of ethnic identity people associate with most strongly.
The most striking characteristic of the data is that roughly half of respondents from both groups identified
themselves first in terms of their sub-tribe, and half identified themselves first in terms of their tribe
(Table 6). This result challenges the conventional wisdom that ethnicity is experienced the same way by
all members of an ethnic group. Ethnic identity is not uni-dimensional; rather, both levels of ethnic
association are salient for large portions of the sample population.
However, while Kenyans in the sample are evenly split about which dimension of ethnic identity they
associate with most strongly, their identification with their tribe or sub-tribe is not random. There are
some clear determinants of the dimension of identity that respondents associate with most strongly. We
consider the same individual-level factors as in the earlier Afrobarometer analysis: gender, age, education
level, occupation, urban/rural location, and media exposure.
We estimate two slightly different models
(see Table 7): column 1 reports results from the full sample; column 2 introduces controls for
membership in the Kalenjin ethnic group and the Tugen sub-tribe. The main results are robust to both
specifications.
We find that education significantly reduces the likelihood of sub-tribal identification – the higher a
respondent’s level of educational attainment, the greater the likelihood that they identify themselves in
terms of their tribe rather than sub-tribe. This effect continues to increase at higher levels of education, as
seen in the coefficient estimate on the indicator variable for post-secondary education (-0.240, standard
error 0.071 – Table 7, column 1). Occupation also matters. Petty traders and shopkeepers are
significantly more likely to identify themselves in tribal terms, with a large point estimate (-0.152,
standard error 0.058). Respondents in the “white collar, teacher or government employee” and “blue
collar or miner” categories are also significantly more likely to identify themselves as Luhya or Kalenjin
than as members of one of these communities’ sub-tribes. The strength and consistency of these
occupational findings are remarkable given that we have already controlled for respondents’ levels of
education.
Consistent with these results is the finding that respondents in Chwele (the more rural survey location) are
significantly more likely to see themselves in terms of their sub-tribe than their tribe. The interpretation
of the point estimate on the Chwele dummy is that, all else equal, a respondent in Chwele is between 21
percent and 31 percent (depending on the model) more likely to identify him or herself in sub-tribal terms
than a respondent in Eldoret. Women, who may often be less likely to have opportunities to leave their
home area, also tend to associate more strongly with their sub-tribes, although this result is statistically
significant in only one of the two specifications. In contrast, listening to the radio and reading the
newspaper have no effect on the relative salience of tribal and sub-tribal identities in this sample.
The results are consistent with a single over-arching theory: that the scope of the social sphere in which a
person operates affects the dimension of ethnic identity with which the person associates most strongly.
People with low levels of education, enmeshed in highly localized networks, and interacting principally
with others from their narrow rural arena have a circumscribed social universe and tend to see themselves
in sub-tribal terms. People who are more cosmopolitan and, through trade or because their occupations
bring them in contact with a wider set of individuals, interact in a broader social sphere, see themselves in
terms of a more expansive social unit – their tribe. This finding offers strong confirmation for situational
approaches to ethnic identity, as articulated by early scholars such as Mitchell (1956), Epstein (1958),
Gluckman (1960), and Young (1965). It also corroborates Posner’s (2004a) thesis that the salience of a
social identity will depend on the size of the group it defines relative to the scope of the social arena in
which the group is located.
14
Copyright Afrobarometer
Table 7: Sub-Tribal versus Tribal Ethnic Identification in Chwele and Eldoret, Kenya
Dependent variable:
Named sub-tribe (rather than tribe)
first when asked for ethnic identity
(1) (2)
Female
0.050
(0.033)
0.068
**
(0.033)
Age (years)
-0.0017
(0.0016)
-0.0008
(0.0016)
No formal education
-0.092
(0.107)
-0.076
(0.109)
Completed primary education
-0.079
*
(0.045)
-0.077
*
(0.046)
Some secondary education
-0.114
**
(0.053)
-0.096
*
(0.055)
Completed secondary education
-0.163
***
(0.051)
-0.178
***
(0.051)
At least some post-secondary education
-0.240
***
(0.071)
-0.236
***
(0.072)
Occupation: Farming or fishing
0.038
(0.039)
0.022
(0.040)
Occupation: White collar, teacher, or government employee
-0.152
*
(0.080)
-0.135
(0.081)
Occupation: Blue collar or miner
-0.131
*
(0.069)
-0.079
(0.072)
Occupation: Student
-0.149
(0.144)
-0.124
(0.146)
Occupation: Business, shop keeper, or petty trader
-0.152
***
(0.058)
-0.126
**
(0.059)
Occupation: Other (e.g., unemployed, housewife, don’t know)
-0.110
(0.070)
-0.103
(0.072)
Rural data collection site (Chwele)
0.21
***
(0.03)
0.31
***
(0.04)
Listens to radio daily
-0.040
(0.039)
-0.020
(0.040)
Reads the newspaper at least weekly
0.040
(0.038)
0.055
(0.038)
Kalenjin ethnic group
0.21
***
(0.04)
Tugen sub-tribe (Kalenjin)
-0.11
(0.10)
Observations (respondents)
1186
1186
Notes: The sample includes all Kalenjin and Luhya respondents. The survey question that generated the dependent
variable was phrased: “What is your tribe or sub-tribe?”. Probit estimation, with marginal coefficient estimates (at
mean values for the explanatory variables). Huber robust standard errors in parentheses. Significantly different
than zero at 90 percent (*), 95 percent (**), 99 percent (***) confidence. The omitted education category is “Some
primary education”. Enumerator fixed effects are included in all specifications.
Note that these findings say nothing about the salience of ethnicity per se. What the results show is that
ethnicity is expressed in different ways – that is, through identification with different dimensions of social
identity – depending on the degree to which respondents are enmeshed in the broader social networks to
which urbanization, education, and working in non-traditional occupations provide exposure. Whereas, in
the analysis of the Afrobarometer data, being more “modern” affected the salience of ethnic
identifications, in the analysis of the Kenya data, being more “modern” defines a situation (characterized
15
Copyright Afrobarometer
by exposure to a broader universe of interacting partners) in which some kinds of ethnic identities become
more salient than others. However, a person’s education, urban-rural location, and occupation are just
three of the many factors that define their “situation.” If his or her situation changes – for example, if the
person travels to a neighboring country or is introduced to a person of a different race – then so too will
the dimension of ethnic identity that the person finds most salient, and this will be true irrespective of
how “modern” the person happens to be.
Political Sources of Tribal and Sub-Tribal Identification
A key finding from the first part of this paper was the close link between exposure to political competition
and the salience of ethnic identifications. The timing of the Kenya survey, which was undertaken almost
immediately after Kenya’s watershed December 2002 national election, allows us to test the effects of
politics on a slightly different outcome: the kinds of ethnic identities that people use to describe who they
are.
The 2002 general election was the first election in over twenty years not contested by President Daniel
arap Moi, and it resulted in the transfer of power from the Kenya African National Union (KANU), which
had controlled the government since independence, to a grouping of opposition parties united under the
banner of the National Rainbow Coalition (NARC), which was led by former vice president Mwai
Kibaki. The significance of the transfer of power from Moi to Kibaki lay not just in the shift of authority
from KANU to the opposition but in the fact that it was perceived to transfer power from one ethnic
group to another. The political landscape under Moi (and, before him, Jomo Kenyatta) was characterized
by tribal cronyism. Members of Moi’s Kalenjin tribe reaped significant political patronage during his
twenty-five year reign, and Moi’s home region enjoyed disproportionate economic development and
investment (Barkan and Chege 1989). Moi’s Tugen sub-tribe was widely considered to have enjoyed the
most favoritism. While the Kalenjin enjoyed a privileged status under the Moi regime, other groups were
marginalized, including the Luhya, a tribe usually in the opposition, and the Kikuyu, who, as the former
ruling group under Kenyatta, were subjected to particularly concerted repression. The 2002 election saw
the Kalenjin lose the presidency to a Kikuyu (Kibaki) and the vice presidency to a Luhya (Michael Kijana
Wamalwa).
The fact that the survey was administered within weeks of the election and that the sample we treat is
comprised entirely of two communities whose political fortunes had changed so dramatically almost
guarantees that respondents’ answers were influenced by the political events of the moment. In addition,
the two survey sites were in areas that were particularly likely to have been affected by the election
campaign and aftermath. Eldoret was one of the primary beneficiaries under Moi’s presidency, enjoying
among other things a new international airport and university (Moi University). Eldoret was also the
scene of land grabs and tribal clashes in the early 1990s permitted and perhaps directly supported by Moi
(Kenya Human Rights Commission 1998). Eldoret’s Rift Valley Province was one of only two provinces
(out of a total of eight) where the KANU presidential candidate received a majority of votes in 2002.
Chwele sits in a more marginalized area, although it still contains a significant Kalenjin population that, at
the very least, was presumed to have benefited from the Moi presidency. Both sites also have large
Luhya populations.
In discussions with the authors before and immediately after the election, many Kenyans expressed the
belief that the Kalenjin had unfairly benefited from being members of Moi’s tribe and that, with a new
government, they would no longer enjoy such benefits. There was even speculation that the Kalenjin
might face a backlash for the perceived advantages they had enjoyed under the old regime. Indeed,
during the campaign season, Kalenjin leaders were accused of warning their constituents that they would
only be protected by the ruling KANU party (
Daily Nation
4 November 2002). And soon after the
election, 14 Members of Parliament from the Kalenjin-dominated Rift Valley Province threatened that
16
Copyright Afrobarometer
their region would secede from Kenya in response to the discrimination it was facing at the hands of the
new NARC government (
Daily Nation
24 February 2003).
To the extent that Kalenjins in the aftermath of the 2002 election felt that their Kalenjin identity would
associate them with the old regime and thus put them at risk of either retribution or exclusion from future
patronage flows, we might expect Kalenjin respondents to answer questions about their ethnic
background by claiming that they were not, in fact, Kalenjin – that is, by passing – or by identifying
themselves in sub-tribal rather than tribal terms. If so, it would provide empirical evidence for the
fundamentally political origins of identity choice.
Of course, it is difficult to study these dynamic political effects using a single cross-sectional survey in
only two sites. In the absence of longitudinal data on individuals, or at least repeated cross-sectional
surveys in the same study site, our strategy is to compare the frequency of sub-tribal identification among
Kalenjins and Luhyas. Since both groups have traditionally had strong sub-tribal associations, there is no
reason to expect Kalenjin respondents to be
a priori
more likely to identify themselves in sub-tribal terms
than their otherwise identical Luhya counterparts, so a difference in the response patterns between the two
communities can plausibly be interpreted as a product of ethnic redefinition triggered by the changed
political landscape.
Although we have no way of detecting Kalenjins passing as members of non-Kalenjin groups, we do find
suggestive evidence that, conditional on a range of individual characteristics, Kalenjins are more likely
than Luhyas to identify themselves in sub-tribal than tribal terms. The coefficient estimate on the
Kalenjin indicator variable (Table 7, column 2) indicates that Kalenjins are 21 percent more likely to
identify themselves in terms of their sub-tribal affiliations than are Luhyas (standard error 0.04). This
finding is consistent with the interpretation that respondents from either one group or both are altering the
way they identify themselves in response to the new political environment: Kalenjins are retreating from
their tribal identifications as Kalenjin to their sub-tribal identities as Nandi, Sabaot, Keiyo, and so forth as
a means of distancing themselves from their association with the old KANU regime, while Luhyas are
embracing their tribal identity as a means of signaling their relationship with the new ruling cohort.
Two percent of our respondents in Eldoret and Chwele identified themselves as Tugen, former President
Moi’s sub-tribe. In keeping with the political interpretation of our results, the negative coefficient on the
Tugen indicator implies that self-described Tugen were
less
likely to describe themselves in terms of sub-
tribe than other Kalenjin groups. Although the Tugen indicator is not statistically significant, the large
point estimate (-0.11) is suggestive that the Tugen are retreating into a broader Kalenjin ethnic identity to
distance themselves from their association with the former president.
Informal interviews undertaken
by the authors at the time in Western Kenya confirmed that many people believed that the Tugen had
uniquely (and unfairly) benefited under the Moi regime.
14
Note that we cannot definitively rule out the possibility that baseline levels of sub-tribal identification might differ
between these two communities.
15
Given the small number of Tugen respondents in our sample, it would have been unlikely that we would find
significant results in any case – note the large standard errors on the Tugen coefficient estimate in Table 7, column
2. We do not look for similar patterns among the Luhya because the bulk of Luhyas surveyed were from the same
sub-tribe (Bukusu).
16
Of course, we would also expect the Tugen to be particularly likely to try to pass as non-Kalenjin, but this is not
something we can detect in our survey. We intend to repeat the survey in Eldoret and Chwele in early 2005 to create
a quasi-panel that will permit a more definitive investigation of how the 2002 election affected the salience of tribal
and sub-tribal identities among the Kalenjin and Luhya.
17
Copyright Afrobarometer
Conclusion
The findings of this study challenge two persistent conventional wisdoms about Africa: that Africans are
uniformly and uni-dimensionally ethnic, and that the salience of ethnicity is a product of the region’s low
levels of political and economic development. The study’s central result is that exposure to education,
non-traditional occupations, and political competition powerfully affects both whether or not people
identify themselves in ethnic terms and the particular ethnic identity they embrace when they do so.
Taken together the findings provide strong confirmation for modernization approaches to ethnicity, and
for theories that link identity choices with context and instrumentality. Beyond their relevance for these
academic literatures, the paper’s results also have important implications for policymakers and
researchers interested in ethnicity’s effects.
Economists and political scientists use the concept of ethnic salience to help explain everything from
economic growth to civil conflict and the effectiveness of foreign aid.
When they do so, they frequently
employ measures of ethnic diversity as indicators of ethnic salience, the nearly universal assumption
being that greater diversity implies greater ethnic salience. Perhaps surprisingly, then, we find that high
levels of country ethnic fractionalization actually
reduce
the likelihood that individuals will identify
themselves first and foremost in ethnic terms. The finding is sufficiently robust to call into question a
central assumption on which many studies are based.
We also find evidence that both the salience of ethnicity and the particular dimension of ethnic identity
that matters for individuals can change – not just over the course of years, but even over the course of a
few months, particularly at election time. This result, which is entirely consistent with situational
approaches to ethnicity, challenges empirical work that takes ethnic identities as static and historically
determined. Particularly for researchers undertaking survey work, it provides a caution that the timing of
data collection – particularly the proximity of the survey exercise to large-scale political events such as
national elections – can have significant effect on the answers respondents provide about their ethnic
identifications.
The strong relationship we find between the intensity of political and economic competition on the one
hand and the salience of ethnicity on the other also makes it clear that as African countries institute
democratic and market reforms it will become more urgent – not less – for African governments to
develop policies and institutional mechanisms that are capable of dealing with ethnic divisions. Kenya’s
recent political developments are informative. After the reintroduction of competitive multi-party politics
in the early 1990s, Kenya’s reform efforts have increasingly become mired in tribal politics, including
violent ethnic clashes that left hundreds dead. Policies and institutions such as those in place in
neighboring Tanzania – a country known for its efforts at nation-building through the promotion of
Swahili as a national language, public education, and institutional reforms, as described recently by
Miguel (2004) – might serve as a model for how Kenya, and other African countries, could dampen
destructive ethnic divisions. Tanzania has the lowest degree of ethnic identity salience in the Round 1
Afrobarometer sample, at just 3 percent.
Finally, our work brings new evidence to bear on the stubbornly persistent popular misconception that
ethnicity in Africa is an atavism that can be “solved” by political and economic development. Scholarly
consensus has long disputed this position, but the popular view remains firmly entrenched. Part of this
disconnect may lie in lingering racism, which leads some to uncritically accept representations of
Africans as backward and tribe-bound, and of Africa as a place where modern aspects of life somehow
fail to snuff out pre-modern social attachments. But another part of the answer may lie in the fact that
nearly all of the research that documents the association between modernization and deepening ethnic
identification is either anecdotal or based on analyses of single countries. Absent systematic, cross-
17
Refer to Easterly and Levine (1997) and Fearon and Laitin (2003) for just two of many such studies.
18
Copyright Afrobarometer