Many scholars (for instance) have noted a trend around the world of greater decentralization, at least on certain dimensions. Many non-federal, unitary states have tried to devolve some spending and decision-making authority on local or regional governments. Virtually every democratic government nowadays at least feigns some interest in decentralization.
Yet what strikes me is how little decentralization there has been, especially in the developing world. Some developing democracies that are sometimes described (or describe themselves) as “federal” or “semi-federal” include Mexico, India, Indonesia, Brazil, Argentina, Venezuela (before it went authoritarian some time in the 2000s), South Africa, Malaysia, Pakistan, Iraq, Nepal, and Nigeria. Yet none of these countries, other than Mexico, affords its constituent state or regional governments autonomy commensurate with that found in federal and semi-federal “Western liberal democracies” like Spain, Canada, the U.S., Switzerland, Belgium, Germany, Austria, Australia, and Italy. For instance, in Brazil, states do not have exclusive powers, and the federal government may overrule any state law with its own legislation. In India, the federal government may suspend state governments from operating at all and impose “President’s Rule.” Of all developing democracies, only India, Mexico, and Brazil routinely allow subcentral governments to raise significant revenue through autonomous taxation policies. (I count 9 Western democracies with such fiscal autonomy.)
Some of these developing countries are both huge and ethnically and regionally diverse, India and Indonesia most notably. One might think that these governments would have even more reason to decentralize than would the governments of comparatively homogeneous Western democracies. Therefore, the relative lack of decentralization in developing countries remains a puzzle.
One explanation might be the smaller talent pool in developing countries. Decentralization might not be feasible because uneducated or politically unsophisticated local officials require close supervision from a small cadre of Western-educated central administrators. While this explanation might have some weight in very poor democracies like Mali (before the recent coup), it likely does not apply to the majority of the cases just mentioned. If the talent pool in developing democracies were desperately shallow, then small developing democracies should have little state capacity plus all the adverse sequelae political scientists typically attribute to state weakness. Yet many small democracies in the developing world have performed fairly well: Costa Rica, Jamaica, Trinidad, Botswana, Mauritius, and Namibia, not to mention Slovenia and the Baltic republics in central and eastern Europe. There is no obvious positive relationship between country size and economic or political performance in the developing world.
Furthermore, many of the cases just mentioned do boast significant decentralization along some dimensions. For instance, India and Indonesia lack a unified internal market, allowing local and state or provincial governments to impose trade barriers on products from other regions. This is an economically perverse form of decentralization and one that has been nearly stamped out in the West, apart from certain discriminatory government procurement regulations. In addition, many developing democracies feature significant decentralization of expenditures: local and regional governments control significant budgets, but those budgets are funded by central grants, and most policy authority lies with the center. This set of policy choices is also likely economically perverse, as “vertical fiscal imbalance,” whereby subcentral governments depend heavily on grants or mandatory revenues from the center, tends to encourage fiscal irresponsibility. In Argentina in the 1980s and 1990s, provincial governments established their own banks, which were forced to lend money to those governments, leading to repeated fiscal crisis.
Another explanation might be that there is something about the Western liberal tradition of political philosophy that encourages decentralization. Many developing democracies fit within the category of “illiberal democracies,” where majorities use their political power to trample the rights of minorities. Sri Lanka might be just such a country, where the Sinhalese majority has repeatedly refused to countenance significant autonomy for the Tamil minority, and the central government fought a brutal civil war against Tamil rebels, complete with vast numbers of civilian killings and other human rights violations.
There may well be something to this explanation, but there are also hazards. As Vito Tanzi noted (PDF), demand for decentralization rises with size of government. A nightwatchman state can afford to be centralized because no one really cares about who controls it. Developing countries have bigger governments than Western democracies, not in the government spending as a share of GDP sense, but in the sense that the distribution of resources in such societies is more elastic with respect to the distribution of political power. So demand for decentralization should be higher there. True, the constraint might instead be supply: the views of political leadership in such societies. But then why the “perverse” decentralization in some countries?
To examine the extent and form of decentralization in developing democracies, I have, with the help of University at Buffalo Ph.D. student Govinda Bhattarai, developed a new dataset of regional self-rule in consolidated democracies worldwide. The coding scheme extends that introduced by Liesbet Hooghe, Gary Marks, and Arjan Schakel for Western democracies and various postsocialist European countries. Without going into details here, I will simply note that we coded the scope of policy powers of subcentral governments, the scope of taxation powers of subcentral governments, the local electoral accountability of subcentral officials, and the ability of the central government to veto subcentral laws.
Using those indicators, I then construct two higher-level, multiplicative indices of economic self-rule and political self-rule. Economic self-rule takes into account political self-rule as well as the tax autonomy of subcentral governments. Economic self-rule ranges from 0 (none) to 48 (maximum). Political self-rule ranges from 0 (none) to 16 (maximum).
The scatter plot below shows regional self-rule on the economic (Y axis) and political (X axis) dimensions in 2006, the latest year for which data on regional self-rule in the Hooghe, Marks, and Schakel dataset are available (our data go to 2010, however). Each observation in this plot is a type of region: either a particular region with its own autonomy statute (like Aaland in Finland or Scotland in the UK), or a type of regional government with the same autonomy arrangement (like states in the U.S. or in India).
Look at how few regions outside the developed West score highly on economic self-rule! Within the developed West, indicated by the blue “+” markers, the cluster of observations in the upper right, reaching the maximum on both economic and political self-rule, correspond to the provinces of Canada, the Yukon Territory in Canada, and the cantons of Switzerland. Also maxing out on economic self-rule are the Northwest Territory of Canada, Nunavut in Canada, Grønland in Denmark, Færøerne in Denmark, Euskadi in Spain, Navarra in Spain, the states of the United States, Alaska & Hawaii of the United States (considered separately because they were not states for the full length of the dataset), Aceh in Indonesia, West Bank & Gaza under Israel, the states of Mexico, and Gagauzia in Moldova. Only the last four are not in the developed West. After those, you have to drop all the way down to ESR=24,PSR=12 to find another region outside the Western liberal democracies with significant economic self-rule (the Autonomous Region of Muslim Mindanao in the Philippines).
So not only are developing democracies more centralized than Western liberal democracies (except on common market policies and access to the central bank, economically perverse forms of decentralization), but they are particularly centralized on tax policy. Very few developing democracies allow regional governments to establish their own fiscal resources. This form of centralization is also economically perverse, because it prevents federal systems from achieving the economic advantages of decentralization, such as Tiebout sorting and hard budget constraints. So why do developing democracies maintain such perverse forms of decentralized institutions? My answer to the question is found in a new paper I’m working on, and I’ll blog about it a little next week.