Decentralization is often expected to improve development outcomes by bringing public decisions closer to the people and the places they are intended to serve. Subnational governments may have better information about local needs, face stronger incentives to respond to local priorities, and be better positioned to tailor public investment to local circumstances.
But decentralizing authority also changes who has discretion over public resources—and the political incentives surrounding that discretion. A new study of one of Europe’s largest regional development programs provides unusually compelling historical evidence of why this matters. A recent article in the Journal of Development Economics examines what happened when Italy transferred important investment-allocation powers from a relatively autonomous central agency to newly created regional governments in the early 1970s.
Their findings offer an important insight for contemporary decentralization and place-based development policies: devolution works through institutions, and transferring discretion without appropriate accountability and institutional safeguards can change not only where public money goes, but also what that money achieves.
A natural experiment in the governance of regional development
The Cassa per il Mezzogiorno (the ‘Fund for Southern Italy’, or CasMez) was established in 1950 to address Italy’s persistent North–South development divide. It financed infrastructure, industrial development and other investments across Southern Italy.
Initially, allocation decisions were made by an autonomous central board largely composed of technical experts. That changed dramatically following the creation of Italy’s regional governments in 1970. A 1971 reform transferred control over CasMez projects to the regions, leaving the agency’s technicians largely in an advisory role. This institutional change provides the authors with something close to a natural experiment: the development program continued, but the governance of its resources changed substantially.
The authors combine project-level CasMez data with newly digitized political records from more than 5,600 archival files. Their principal analysis covers 374 Southern Italian municipalities with populations above 10,000 and compares funding before and after the reform according to whether municipal governments were politically aligned with their regional government. The resulting evidence suggests that the governance change had significant consequences.
Political alignment suddenly mattered
Before devolution, the authors find little evidence that political alignment with the national government systematically affected CasMez allocations. After regional governments gained control over allocation decisions, however, municipalities governed by the same political party as their regional government began receiving more funding than politically unaligned municipalities.
In the authors’ main specification, political alignment was associated with a 53.6 percent increase in the total amount of CasMez funding received, conditional on receiving funding. In the short run, aligned municipalities received both more projects and larger investments.
Over the longer term, however, the pattern increasingly took the form of more numerous, smaller projects, particularly public works.
This distinction is important. Rather than concentrating investment on projects offering the strongest economic returns, politicians may have had incentives to spread projects across constituencies and beneficiaries. In other words, decentralization did not simply relocate decision-making authority. It altered the political economy surrounding those decisions.
The evidence is particularly striking because political party affiliation by itself does not appear to explain the result. The partisan allocation effect emerged after regional governments acquired control over CasMez resources.
Local context shaped the risks of devolution
The effects were also not evenly distributed. Political alignment was particularly associated with additional projects in municipalities located in areas with higher historical corruption, lower levels of industrial employment, and larger electorates.
For the study’s authors, these findings are consistent with a “local capture” mechanism. Where administrative institutions and accountability mechanisms are weak, transferring broad discretion downward can make public resources more vulnerable to local political and rent-seeking pressures.
This is an important qualification to simplistic arguments either for or against decentralization. The relevant question is not merely whether a function should be centralized or decentralized. It is what kind of authority is being decentralized, to which institutions, under what accountability arrangements, and with what degree of discretion.
More spending, but weaker development results
Perhaps the study’s most consequential finding concerns what happened to the effectiveness of CasMez spending. Before the reform, greater CasMez investment was positively associated with subsequent local employment growth. After devolution, that relationship weakened considerably, with the deterioration concentrated especially among politically aligned municipalities.
The authors provide an illustrative calculation to convey the magnitude of the change. Before the reform, €1 million in CasMez funding was associated with approximately 14 additional jobs. After the reform, the corresponding estimate falls to fewer than one job. They are careful to emphasize that these figures are illustrative rather than causal cost-effectiveness estimates; changing economic circumstances and differences in empirical design mean that the numbers should not be interpreted too literally.
The broader empirical pattern is nonetheless noteworthy. Following devolution, politically aligned municipalities systematically attracted more resources, yet those additional resources were not accompanied by superior economic outcomes. Instead, the relationship between investment and employment growth deteriorated particularly strongly in those municipalities.
The lesson: decentralization needs institutional safeguards
It would be easy to read the Italian experience as an argument against decentralization. However, that is explicitly not the conclusion the authors draw.
Decentralization has genuine informational and accountability advantages. Subnational governments can understand local constraints better than distant central agencies and can adapt policies to different territorial circumstances. The problem arises when considerable expenditure discretion is transferred into an institutional environment where administrative capacity and accountability are weak and political incentives to influence allocation are strong.
The authors identify three conditions that proved particularly problematic in the Italian case: broad subnational discretion over allocation, weak accountability and administrative capacity, and strong electoral incentives surrounding the resources being allocated. That distinction has direct relevance for decentralization reforms today.
Decentralizing responsibility for regional development does not necessarily mean that every aspect of resource allocation should be discretionary. Formula-based allocations, transparent eligibility criteria, independent monitoring, performance-linked disbursement and strong procurement arrangements can constrain political manipulation while preserving substantial subnational authority over how development objectives are achieved.
The study also suggests that oversight should examine what governments spend money on, not simply how much they spend. The post-reform movement toward fragmented public works appears to have been one channel through which political incentives affected CasMez. The authors therefore suggest monitoring the sectoral and contractual composition of investment and considering rules that limit excessive fragmentation or connect funding to verifiable performance.
Getting the institutions of decentralization right
The experience of Southern Italy offers a broader lesson for multilevel governance. Most importantly, the success of decentralization depends neither on local autonomy alone nor on central oversight alone. It depends on how authority, discretion, incentives and accountability are combined across levels of government.
This lesson is especially relevant for place-based development programs, which deliberately channel substantial resources toward economically weaker territories. These are often precisely the places where administrative capacity and accountability institutions may be weakest. A uniform approach to devolution can therefore expose the most vulnerable regions to the greatest risks of political capture.
The study’s results thus reinforce an important principle for decentralization practitioners: moving authority closer to people does not automatically improve how that authority is used. Effective decentralization requires getting the institutional architecture around devolved power right. When those institutions are strong, local knowledge and responsiveness can become powerful development assets. When they are weak, the same discretion can instead amplify political incentives that pull resources away from their intended purpose.
Read the (open access) article in the Journal of Development Economics:
Tancredi Buscemi and Giulia Romani, “The Political Economy of Regional Development: Evidence from the Cassa per il Mezzogiorno,” Journal of Development Economics, Vol. 184 (2027), article 103900. The article was made available online in August 2026.
