Disclaimer. Don't rely on these old notes in lieu of reading the literature, but they can jog your memory. As a grad student long ago, my peers and I collaborated to write and exchange summaries of political science research. I posted them to a wiki-style website. "Wikisum" is now dead but archived here. I cannot vouch for these notes' accuracy, nor can I say who wrote them.
Lijphart. 1995. Multiethnic democracy. In The Encyclopedia of Democracy, ed. Seymour Lipset, vol. III, pp 853-65.
Lijphart reviews the literature on (1) what ethnicity is; (2) ways of solving problems of multiple ethnicity (and further problems these solutions may cuase); (3) whether ethnicity is instrumental or primordial (or even whether constructivists are right); (4) modernization and ethnicity; (5) democratic means of moderating ethnic disputes; (6) criticisms of power sharing.
As with most of Lijphart's work, his arguments are motivated by his long-running advocacy of consociational/consensus democracy. See Lijphart's Patterns of Democracy.
Lijphart uses the broad definition of ethnicity, whereby not just language, but also religion, race, customs, etc. can be the basis of ethnic differences.
Elimination of ethnic problems can be done immorally (genocide or expulsion) or morally, but using problematic methods (partition or assimilation).
Primordialists argue that ethnicity is inherited and difficult to change; instrumentalists argue that it is fluid and often not even an issue until politicians decide to stress ethnic differences. Both can be criticized: primordialists for failing to recognize quick change in ethnic boundaries, and instrumentalists for thinking that politicians are completely free to build on ethnic divisions or not (e.g. some noticed difference must exist, and if they ignore it, they may not have any chance of winning). Constructivists take a bit of a middle ground.
Modernization theory argued that economic development would cause us to put "pre-modern" ethnic differences behind us, allowing for regional and global integration. However, later developments have reminded us that even "modern" countries have ethnic problems; they don't go away.
Of the four ways that are argued to moderate ethnic divisions, three are majoritarian (cross-cutting cleavages, vote pooling, and majority control) and the fourth is consociational (power sharing).
Power sharing has four characteristics: representation of all ethnic groups; autonomy in ethnically internal affairs (this may take the form of federalism if the groups are appropriately distributed); proportionality; and minority vetoes (in some form, such as concurrent majorities of both the population and the ethnicities, or an agreement about higher thresholds for certain legislation, etc.).
Criticisms of power sharing: it doesn't always work, it can lead to deadlock (but that may be better sometimes than efficiency), it can strengthen ethnic differences by focusing on them (and prevent gradual assimilation and formation of cross-cutting alliances), it is not democracy because agreements are made behind closed doors by elected ethnic elites (but don't elected elites always do that?), some other minor concerns.
Cross-cutting cleavages (from Lipset 1960; see Lipset and Rokkan 1967 for related work): If there are many cross-cutting cleavages (which can moderate your preferences on all of them), two-party majoritarianism may be best because it will force the parties to moderate (centripetal incentives) in order to capture a majority. Following this logic, federalism is only good if it cross-cuts ethnic boundaries (this disagrees with power-sharing's ideas about internal autonomy).
Criticisms: when ethnic loyalties are strong, they will counter the moderating tendencies, and people may not gather to two large parties. Even if there are two large parties, they will not alternate in power because ethnic loyalties will reduce the number of swing votes. This can lead to the undemocratic permanent exclusion of many groups.
Vote pooling (from Horowitz 1985): Horowitz disagrees that "seat pooling" (forming a majority coalition by pooling seats in parliament) will work. Vote pooling (e.g. alternative vote for presidents, or pooling votes [not seats] to get a majority coalition). Horowitz wants a presidential system with the alternative vote. Horowitz also favors having many heterogeneous federal units, with large groups divided into several such units. Also approval of a system in which presidents need not only a majority, but also at least 25% of the vote in at least two-thirds of federal units (Nigeria for a while).
Criticisms: underestimates the incentives for compromise during parliamentary coalition building. Forgets that candidates who appeal for more "second choice" votes from other ethnic groups [in alternative vote] will get fewer "first choice" votes from their own. Overestimates incentives for moderation in alternative vote--it's really no different from a majoritarian system with a runoff.
Control theory (Lustick 1979): I think Lijphart mischaraterizes "control" theory when he speaks of it as a majoritarian alternative to consociationalism. Lustick seems to be outlining not an ideal democratic way of managing ethnic relations, but a less democratic alternative that can arise when consociationalism does not. Granted, it may look democratic, or at least nonviolent, but that doesn't mean we want it. He seems to be arguing that we should understand better the opposite of consociationalism, namely, controlled subordination of smaller groups.
Criticisms: This often spells majoritary dictatorship. In fact, Control theory stands at the opposite end of a continuum compared to Power Sharing theory.
Research by the same authors
Research on similar subjects