Foreign Affairs has a very good piece on Affirmative Action. The summary:
“Across the globe, the lessons from affirmative action programs are clear: they can occasionally help in the economic sphere, produce mixed results in improving social cohesion, and are an unmitigated disaster when it comes to politics.”
Here are some excerpts:
Differences aside, however, affirmative action programs around the world have enjoyed similar successes—and fallen prey to many of the same failures. Most succeed in reducing economic inequality, although often less significantly than policymakers hope. They have a mixed record when it comes to improving social cohesion: affirmative action policies tend to underscore ethnic divisions rather than reduce them, although there is some evidence that racial unrest becomes less likely as economic inequality ebbs. And their most negative effects tend to be in the political sphere: in many countries, policies intended to assist the disadvantaged have been corrupted by political elites who manipulate the system to their own advantage and block any attempts at reform.
Balancing this less-than-impressive track record is the fact that affirmative action has often played an indispensable role in nation building in multiethnic societies. In the West, nation building is usually thought of as something that happens only in postcolonial societies or in places recovering from recent periods of intense intercommunal conflict. Today, in fact, thanks to immigration, shifting demographics, and changing beliefs about identity and ethnicity, prosperous Western countries also experience a kind of slow-motion nation building, as ideas about what it means to be “American,” or “British,” or “Dutch” evolve. As a result, the role of affirmative action in creating national identities in places such as Malaysia and India may be more relevant to places such as the United States than most Americans realize.
For the benefits of affirmative action to outweigh the harm, however, the programs must effectively target the poor. Programs that lift only a select few out of poverty or become corrupted by elites will neither reduce inequality nor bolster the nation. At best, such programs will quickly lose popular support; at worst, they will fray the social fabric they were meant to strengthen.
It has not been very successful in South Africa:
In South Africa and India, the economic effects of affirmative action have been more mixed. When apartheid ended in 1994, the South African government instituted a policy of “black economic empowerment,” which promoted the transfer of stakes in white-owned businesses to black investors. The government hoped that change at the top would trickle down to the bottom, as black-owned businesses would be more likely to hire and promote other black South Africans. In fact, the program succeeded primarily in creating a coterie of well-connected black entrepreneurs.
The problem seem to be primarily about privileged elites coopting the programs.
If the economic and social impacts of affirmative action have been mixed, its political impacts have been almost universally damaging. In a wide range of countries, political elites and special interests have consistently exploited affirmative action programs for their own gain.
And reform is politically damaging for the reformers:
Even when reforms are clearly necessary, politicians often have a vested interest in keeping flawed affirmative action programs the way they are, or even expanding them. The equation is simple: more benefits to more people equals more votes.
So how do you make them work? This is an important question because once you start them they are nearly impossible to stop.
The successes and failures of affirmative action in India, Malaysia, and South Africa offer important lessons for the United States, where government agencies and other institutions have often struggled to define, justify, and reform their programs. The major lesson is that affirmative action policies work best when they target the poor. The “creamy layer” problem undermines the very purpose of affirmative action and makes such programs politically unsustainable. Indeed, interethnic support for Malaysia’s affirmative action program reached its peak in the 1970s, when it primarily targeted poor and rural indigenous groups, and began to drop as the program became a tool of the political elite.
In designing affirmative action policies that cannot be exploited by the wealthy or the politically connected, transparency is key, both in the rules of eligibility and in the awarding of benefits.
…Instead of giving politicians discretionary control over the selection process, the government should award benefits by lottery or through a merit-based point system. Similarly, in the United States, policymakers should craft affirmative action policies with benefits that are awarded according to clear metrics.
…Above all, the success of affirmative action depends on preventing programs from outliving their economic and social efficacy.
Source: Does Affirmative Action Work?