Governance greatly affects how well metropolitan areas function. How do we know? Even if differences in human capital levels, economic structure and agglomeration benefits are taken into account, large differences in productivity levels between a country’s metro areas remain. Governance arrangements – or the lack thereof – can explain an important part of these differences.
Metropolitan areas are in general more productive than smaller urban agglomerations and rural areas. Partly, this is due to higher human capital levels. The larger a metropolitan area, the higher the average education and talent of its residents, which is reflected in higher productivity levels. Another part arises through agglomeration benefits, positive externalities associated with metropolitan size. In line with the literature review by Combes et al. (2011), OECD estimates suggest that agglomeration benefits are responsible for an increase in residents’ productivity of between 2% and 5% if the population of a city doubles (OECD 2015a).
So why does metropolitan governance matter? Large urban agglomerations are characterised by manifold spatial connections and interdependencies that are often not reflected in the way they are governed. In most OECD countries, municipal borders are based on historical locations of towns and villages. Put differently, these administrative structures cannot fully cope with the challenges connected to economic and social realities in large urban agglomerations.
The OECD Metropolitan database defines ‘functional urban areas’ across the OECD on the basis of a common method that relies on settlement patterns and commuting flows rather than administrative borders (OECD 2012). Not a single one of the 275 OECD metropolitan areas – functional urban areas with populations in excess of 500,000 – is governed by a single local government. The metropolitan area of Paris, which consists of 1,375 municipalities, might be an extreme case, but more than 200 metro areas contain more than 10 local governments and over 60 contain more than 100 within their boundaries. Figure 1 shows the fragmentation of the metropolitan areas of Berlin and Madrid. The urban cores of the metro areas, defined as the contiguously built-up surface area and depicted in dark blue, are surrounded by a large number of administratively independent smaller municipalities that are closely connected to the urban core through commuting.
Figure 1. Municipalities within the metropolitan areas of Berlin and Madrid
Note: Dark blue: urban core / Light blue: functionally integrated commuting zones.
A large number of municipalities in metropolitan areas can complicate policy coordination among local governments. A potential solution to this coordination problem could be the amalgamation of municipalities within a metropolitan area. Many countries have successfully reduced administrative fragmentation, but rarely are these policies focused on creating administrative cohesion in large metro areas. An alternative to the amalgamation of municipalities is the creation of an organisation dedicated to the coordination of policies in metropolitan areas – a metropolitan governance body. The OECD Metropolitan Governance Survey has collected the first systematic overview of such metropolitan authorities across the OECD. The findings are described by Ahrend et al. (2014) and in the new report Governing the City (OECD 2015b).
Metropolitan governance across the OECD
Metropolitan governance bodies are common across the OECD, but their competences vary. Approximately two thirds of the 275 OECD metropolitan areas have some form of metropolitan authority. This average conceals a great deal of variation. The characteristics of the existing metropolitan authorities differ not only across countries, but also within them. Nevertheless, three fields of work emerge as clear priorities for most metropolitan authorities; regional (economic) development, transport, and spatial/land-use planning. The predominance of these three policy fields is not surprising. They are often mentioned by practitioners as the areas where the need for cooperation among municipalities is greatest.
Figure 2. Fields of work of metropolitan authorities
Most common is voluntary cooperation among municipalities, resulting in structures that do not have any regulatory powers. Without formal powers, such authorities gain their influence through the provision and exchange of information and voluntary commitment of member municipalities to implement their decisions. In some cases, they are also tasked by municipalities to provide selected services. In contrast, only about one fifth of all existing metropolitan authorities have formal regulatory or legislative powers to co-ordinate policies across a metropolitan area, potentially even overriding the wishes of some individual municipalities.
Metropolitan governance reduces the cost of administrative fragmentation
Charles Tiebout (1956) famously argued that more administrative fragmentation – a larger number of local governments – is associated with a greater set of choices over public service provisions and their costs. Increased choice and competitive pressure among local governments improves the quality of local public services, which in turn may increase productivity in the municipalities and ultimately the metropolitan area. But Tiebout’s argument fails in respect of policies that require coherence across the whole metro area and generate externalities across administrative boundaries. For example, the planning of infrastructure provision is more complex if a large number of local governments have the power to veto individual projects. Which of the two forces – the positive impact from competition among local administrations or the need for co‑ordinated policies – prevails in determining a metropolitan area’s fortunes is ultimately an empirical question.
In recent research, we estimate the impact of administrative fragmentation and the presence of metropolitan authorities on productivity in five OECD countries (Ahrend et al 2014). Using observations for more than two million individuals from Germany, Mexico, Spain, the UK and the US, we estimate productivity differences across 430 functional urban areas. The estimates use wages as a proxy for individual productivity and account for the direct impact of individual characteristics, such as education, age, gender, occupation, and part-time work. Figure 3 plots the productivity differentials for the 430 functional urban areas against the number of local governments per capita, standardised for each country to have zero mean and unit variance. A clear negative association emerges – administrative fragmentation is associated with lower productivity. The effect is robust in multivariate regressions that take agglomeration benefits into account, and control for city aggregate skill level, industrial structure, and capital city or port city status.
Figure 3. The cost of administrative fragmentation
In our quantitative analysis, we find that the descriptive evidence understates the true penalty of fragmentation. Why? Metropolitan authorities have the potential to alleviate the cost of administrative fragmentation. Focussing on the 140 metropolitan areas in the aforementioned five countries and using the information collected in the OECD Metropolitan Governance Survey results in two striking findings. First, doubling the number of local governments within a metro area reduces productivity by 6%, thus possibly reducing the gains from agglomeration benefits. Second, the presence of a metropolitan governance body reduces this penalty, on average, by half. This shows how better policy coordination can have direct effects on productivity and hence GDP levels.1
The impact of better metropolitan governance is not limited to economic productivity. For example, metro areas without a metropolitan authority have experienced an increase in urban sprawl, as measured by the population density of the built-up area, whereas those with a metropolitan authority densified.2 Similarly, residents’ satisfaction with public transport in metro areas with a dedicated metropolitan transport authority is 14 percentage points higher than in those without, and air pollution (PM2.5) is significantly lower when a metropolitan authority exists.3
These findings are indicative that dedicated metropolitan authorities help to improve economic outcomes and the quality of life in metropolitan areas. They also correspond to the anecdotal experience of policymakers and the conclusions from a large number of case studies conducted by the OECD. Together with these insights, the new findings from the OECD Metropolitan Governance Survey make a strong case that well-designed metropolitan authorities are important for a country’s prospects, as they can improve the productivity and the quality of life of its metropolitan areas.
Ahrend, R, E Farchy, I Kaplanis and A C Lembcke (2014) “What makes cities more productive? Evidence on the role of urban governance from five OECD countries”, OECD Regional Development Working Papers, No. 2014/05, OECD Publishing, Paris.
Ahrend, R, C Gamper and A Schumann (2014) “The OECD metropolitan governance survey: A quantitative description of governance structures in large urban agglomerations”, OECD Regional Development Working Papers, No. 2014/04, OECD Publishing, Paris.
Combes P-P, G Duranton and L Gobillon (2011) “The identification of agglomeration economies”, Journal of Economic Geography, 11: 253-266.
OECD (2012) Redefining ‘urban’: A new way to measure metropolitan areas, OECD Publishing, Paris.
OECD (2015a) The metropolitan century: Understanding urbanisation and its consequences, OECD Publishing, Paris.
OECD (2015b) Governing the city, OECD Publishing, Paris.
Tiebout, C M (1956) “A pure theory of local expenditures”, Journal of Political Economy, 64(5): 416-424.
1 The existence of metropolitan authorities captures only one aspect of good governance in metropolitan areas, but not others such as stakeholder participation. Thus, the total impact of governance on economic performance is likely to be larger than this estimate.
2 This estimate is based on a regression that controls for country fixed-effects. It refers to the 2000 – 2006 period, the only period for which relevant data is available.
3 Controlling for population size and country fixed-effects.