Has Decentralization Worked?

10/04/2012 09:49 am ET | Updated Dec 04, 2012

Think back to the 1990s. The Soviet Union had just disintegrated. The American economy was on a roll. Technology and the Internet were starting to connect people. And democracy was spreading fast, especially among developing countries that never had it before. The headlines were about fresh new presidents and voters speaking up freely. Behind the headlines, a subtler political change was taking place: power was beginning to shift from central governments to states and municipalities. Vital public services for the common citizen -- like education, health, water, electricity and many others -- were becoming the responsibility of governors and mayors. With this responsibility came money, some through transfers from the national budget, some in new local taxes, and some from lenders eager to gain new clients. These were the times of "decentralization," the catchy idea that closer proximity -- literally -- between those who govern and those who are governed is always a good thing. No one could ever know people's preferences better than their local authorities, right? And if local leaders failed to deliver, it would be easier for you to hold them accountable -- after all, they are more likely to be your neighbors. This would surely improve service, reduce corruption and save money. Case closed. So countries decentralized fast and furious.

Twenty years later, it is fair to ask: Has decentralization worked? Well, the answer is a bit anticlimactic. When it was done well, decentralization did work. Success only came with smart design and careful implementation. You see, decentralization was -- and still is -- a high-risk, high-reward reform. A lot can go wrong with it. Local bureaucracies may not have the capacity to manage a school system or a power distribution network. They may not have the "scale" to keep costs down -- you can negotiate better prices for, say, trash collection trucks if you buy them by the thousands for a country rather than by the dozen for a county. Small-town politicians may be easier to lobby -- or to bribe. Labor disputes, obsolete equipment and irresponsible pension promises are just some of the problems that usually plague the public services that get decentralized -- federal governments are only too happy to see someone else blamed. Left to fend for themselves, remote poor areas may become even poorer, while big cities close to ports grow bigger and richer -- this is when geography begins to matter and regional resentment begins to fester. And then there is the "bail-out" issue: What should the central government do if a local government goes bankrupt? Can it watch and do nothing as a province's children go without school and its hospitals go without power? South America is living proof that, when that happens, "the federation" has no choice but to step in and pay up -- in effect, making everyone in the nation pay up, too.

With that much risk at play, it is not surprising that few countries can claim success in decentralization. In fact, there is no evidence that when the government is more decentralized, the economy grows faster or is more stable -- there is not enough data to tell one way or the other. Nor is it clear that more power to local governments automatically translates into less poverty. What we have is a growing inventory of experiences from around the world that show how specific public services improved -- sometimes a lot -- when local authorities began to run them. For example, giving Swiss cantons control over education raised student test scores. In Canada, infant mortality fell faster when provinces were responsible for it. The same happened amongst Spain's poorest regions. Bolivian municipalities managed to invest more in water and sanitation where it was most needed, something that translated into healthier local populations. Enrollment in Ethiopia's primary schools shot up when woredas -- a type of territorial division somewhere between a neighborhood and a municipality -- were put in charge. The list goes on.

But you can build a similar list with public services that deteriorated when they were decentralized. So, the real question is not whether decentralization has worked or not, but whether there is something in common among the cases where it has. There is. First, decentralization makes innovation easier. It seems that governors and mayors manage to experiment with new ways to deliver old services. Take the case of schools in Bogota, Colombia. The city wanted to improve its education system. It hired some of the best and most exclusive private schools to run 25 of its own establishments in low-income areas. Students in these "concession" schools -- where the power of teacher unions was limited -- did much better on standardized tests than their peers in the rest of the public system. (Many actually scored higher than their peers in the parent private school!) It would have been politically and practically impossible for the central government to try something like that in the country as a whole.

Second, technology helps decentralization. In the past two decades, computers, cell-phones and the web have made it easier to control and to provide public services at the local level. Training of municipal officials can now be done online. They have access to the same information as big bureaucracies in the capital city. They can learn from each other with a click of a mouse. New gadgets like transponders, remote meters and barcodes have made it simpler to charge for highways, water and licenses. And how about Twitter, Facebook and Youtube to embarrass the mayor if he is caught red-handed? Or the local electricity company if it fails to restore power fast enough? Or the city's sheriff if he keeps pulling over drivers always of the same skin-color? Instant accountability at the tip of the neighborhood's fingers.

And third, debt has to be kept low. Over time, local governments developed their own sources of income. Many began to receive large transfers from their countries' sales of oil, gas and minerals. This made them more creditworthy. Bankers started to offer them loans. Governors and mayors that avoided -- or were not allowed -- to go on borrowing binges did better than those who did. Why? Because paying off hefty debts meant less funding for schools, hospitals and roads. The alternative was, of course, to beg for money from the federal government in exchange for political favors -- a very messy alternative. Not surprisingly, central governments have tried to ban or at least control "sub-national" borrowing, with mixed results. (If you are looking for a good way to regulate this, check what Mexico has been doing.)

So, if decentralization needs such careful fine-tuning to work well, why do it at all? Because it's what people want. We like to choose our local leaders and to have a voice in the services we use day in and day out. There is no going back on that. The genie of decentralization is out of the bottle already. By now, in the average developing country, states and municipalities are in charge of a fifth of all public expenditures. That proportion is above a third in places like Argentina, India, Russia and South Africa. It will continue to grow. That's why the next time there are local elections where you live, be sure to vote.