Political Decentralization and Public Goods Provision: Evidence from Kenya
Recent social science research has indicated that one of the fundamental determinants of long-run economic development is governance.
Sub-Saharan Africa (SSA) — with its history of centralized political power in the hands of corrupt, ‘big man’ presidents in the post-independence era — typically ranks lowest on cross-country rankings of good governance. One driver of SSA’s poor governance record is that, all too often, its leaders use the public coffers to systematically favor their co-ethnics to the detriment of the rest of their citizens.
Since the Third Wave of democratization in the 1990s, to rectify this, multilaterals and international development agencies have recommended that African policymakers decentralize political institutions to more local levels of government. Such reforms, advocates assert, would improve public service delivery by constraining the presidents’ ability to engage in ethnic favoritism when it comes to public goods provision. Indeed, in the last three decades over 80 percent of the governments in the Global South have heeded this advice and experimented with some form of political decentralization.
Despite the efforts of the World Bank, IMF, and other agencies to push such reforms, the actual evidence on the impacts of decentralization on the provision of public goods is decidedly murky, especially in SSA where data is often unreliable if not completely non-existent. If costly political decentralization reforms are going to be recommended by donor agencies and, in turn, implemented by cash-strapped SSA governments, it is important that they be backed by rigorous empirical evidence.
To that end, I examined the impacts of Kenya’s new constitution, adopted in a national referendum in 2010 and implemented across the country in March 2013. Enshrined in this constitution is the sweeping decentralization of political power away from the once-dominant presidency to 47 newly established subnational county governments. The World Bank called Kenya’s institutional reforms, “among the most rapid and ambitious devolution processes going on in the world.”
As is often the case with these types of reforms south of the Sahara, the motivation behind Kenya’s devolution of state structure stemmed from the country’s history of ‘imperial’ presidents that overwhelmingly favored their own ethnic groups with public investment. Over time, this led to heightened feelings of marginalization among ethnic groups not aligned with the sitting president — a sentiment that often boiled over into political violence (Kenya’s elections in 1992, 1997, and especially in 2007/08 all turned violent).
The 2010 Constitution proposed to solve these problems by decentralizing power. Under the 2010 Constitution, each of the 47 county governments is democratically elected and is responsible for administering various public functions (enumerated in Schedule 4) that were formerly provided by the central government at the discretion of the president (health services, agriculture, land administration, water and sanitation, infrastructure, among others). The constitution stipulates that the new county governments are to receive a minimum of 15 percent of national revenues each year to deliver these public goods locally. The theory goes that this constitutional change would mean that all citizens , regardless of ethnicity, would now get a slice of the national pie, and future elections would, in turn, be less likely to turn violent. But has the evidence supported this theory?
Using colonially imposed subnational borders within Kenya that ultimately became the basis for contemporary county borders delineated in the 2010 Constitution, I analyze how Kenya’s decentralization reforms have affected spending patterns for a key public good: roads. Road spending is significant because it is the largest single line item in Kenya’s annual development budget (about one-fifth of the 2016/17 budget), and Afrobarometer surveys indicate that improved infrastructure is one of the most prioritized devolved spending areas among Kenyan citizens.
The graph below illustrates how road expenditure has changed after the implementation of devolution in 2013. A value of 1 implies that a particular county received road spending exactly proportional to its population (values greater than/less than 1 imply road spending that is above/below the national per capita average).
Specifically, the graph shows that before devolution (pre-2013) counties ethnically aligned with the sitting president received road spending over 1.5 times their proportional share. Conversely, non-aligned counties remained consistently below 1 (the proportional share of road spending) pre-devolution. That is, ethnic favoritism in public goods provision was the status quo before political decentralization.
In contrast, after the implementation of devolution, aligned counties experience a drastic reduction in road expenditure which remains at or below the proportional share throughout the entire post-2013 period. At the same time, the road spending share in non-aligned counties stays consistently above 1 (and above the aligned share) every year after 2013. That is, after a significant portion of infrastructure spending was devolved to new county governments in 2013, ethnic favouritism seems to have dissipated.
In essence, the findings indicate that political decentralization in Kenya succeeded in bringing state institutions ‘closer to the people’ and away from the biased hands of the president. In so doing, these institutional reforms resulted in public goods provision (road spending) that is significantly more inclusive of all Kenya’s citizens — a key objective of the 2010 Constitution.