Minorities and Democratization David Epstein Sharyn OHalloran Columbia

Minorities and Democratization David Epstein & Sharyn O’Halloran Columbia University Bahar Leventoglu SUNY – Stonybrook

Project Background n n Interested in Racial Gerrymandering in the US n In the 1990’s, saw many Republican-Black Democrat coalitions in redistricting n Makes sense electorally: concentrating black voters in a few districts helps elect Republicans elsewhere n But not for policy – one side has to do better So do this with numbers n n Estimate plans that maximize substantive representation, and those that maximize descriptive rep. Are they the same, or is there a tradeoff?

Findings n In the 1970’s & 1980’s, maximize substantive representation by electing blacks to office n But starting in 1990’s, do better by creating districts of no more than 45% black n n n At that point, good chance of electing minority rep. But don’t lose too much influence elsewhere On the other hand, maximize # of minorities elected by concentrating up to 75% BVAP n n Not the same as maximizing substantive rep. So a Pareto Frontier has emerged over time

Lessons: AP Version n Districting can give minority voters: n n A lot of influence in a few districts, or Less influence in more districts At times one strategy is best, at times the other may be superior In the US, the answer changed due to: n n Decreased polarization in the electorate Increased polarization in Congress

Lessons: CP Version n n In a majoritarian system, minorities must form coalitions to influence policy There are two ways to do this: n n n In elections, via encompassing parties, or In the legislature, via vote trading Political institutions can favor one of these alternatives over the other n This holds true for all types of minorities

Segue into this paper… n n So, within majoritarian systems: institutions coalitions min. influence But of course, many systems are not majoritarian n n This is itself an object of choice and change Minorities don’t always do better w/democracy n n Voting to Violence Important issue now in many places

Questions – Reciprocal Effects n How does the presence of ethnic factionalization affect democratization? n n n Does it make it easier or harder to transition to democracy? Does it matter if the ruling party is from the majority or minority group? How does democratization affect minorities? n More discrimination, or less?

Model – Acemoglu & Robinson • Society with two classes: upper (U) and lower (L) • Group i is proportion li of the population, with l. L>l. U • Total wealth x in society, U starts with a x, a l. U U L

Model – Acemoglu & Robinson • Only issue is degree of redistributive taxes (t) • U, of course, wants t=0; L would tax rich & transfer • At first, only U members can vote, but can democratize U 0<t. L<1 L

Game Tree U: Democratize? Y N L: Set t. L U: Set t. U L: Revolt? Nature “Revolutionary Outcome” Success R Y N Not t. U Note: For A&R, revolutionary outcome destroys proportion Y of economy, and U gets 0 utility. For us, lose Y as before, but then move to democracy.

Model – ELO • To A&R model, we add a second ethnic group • Assume L>U, and 1>2 (so L 1 is largest group) • Groups decide independently whether to uprise. 1 2 U L

Revised Game Tree

Model – ELO • So now, two taxes are possible: economic and ethnic • Ethnic tax t 2 yields “discrimination rents” to majority 1 2 U te L Note: t 1 is always 0 in equilibrium. t 2

Model – ELO • In democracy, L 1 makes first proposal for {te, t 2} • If L 1 has a majority, proposal passes • Otherwise, form coalition: 1 U L 2

Model – ELO • In democracy, L 1 makes first proposal for {te, t 2} • If L 1 has a majority, proposal passes • Otherwise, form coalition: L 1 -U 1 (ethnic) 1 U L 2

Model – ELO • In democracy, L 1 makes first proposal for {te, t 2} • If L 1 has a majority, proposal passes • Otherwise, form coalition: L 1 -U 1 (ethnic) or L 1 -L 2 (class) 1 U L 2 Note: Reversion policy in case of disagreement is te=t 2=0.

Equilibrium n n In autocracy, U 1 sets {te=0, t 2=1} No way to credibly commit to any te>0 other than democratization n n Check equilibrium under democracy and see which U 1 prefers If L 1 > 50% of the population, then it sets {te=1, t 2=1} n So focus on L 1 < 50% case

Implications n n Less peaceful democratization the larger is l 2, the size of the minority Peaceful democratization occurs at intermediate levels of uprising cost Y Either both L 1 and L 2 revolt together, or L 2 alone revolts Probability of transition independent of size of the economy (x)

Extensions n Power Sharing in Autocracy n n n U 2 starts off in power n n U 1 can attract L 1 or U 2 as coalition partner Democracy can be worse for minorities Makes peaceful transitions less likely Two-period model with class mobility n n Induces upper-lower class sympathy Increases likelihood of transitions with antiminority coalition in democracy

Conclusions n Ethnic diversity reduces the probability of peaceful democratic transitions n n Lower class does well in: n n n Upper class loses discrimination rents Autocracy violence potential Democracy numbers Addresses question of when politics revolves around an ethnic dimension