International Differences in Decentralization and Productivity

12/01/2009
Featured in print Digest

Social capital -- as proxied by regional trust -- can improve aggregate productivity by facilitating greater decentralization of firms.

In The Organization of Firms across Countries (NBER Working Paper No. 15129), co-authors Nicholas Bloom, Raffaella Sadun, and John Van Reenen collect the first international data on the decentralization of investment, hiring, production, and sales decisions from corporate headquarters to local plants. They survey managers in almost 4,000 medium-sized manufacturing firms in the United States, Europe, and Asia. They find that social capital -- as proxied by regional trust of other people in society -- can improve aggregate productivity by facilitating greater decentralization of firms. Trust appears to facilitate delegation, with higher trust between CEOs and middle-managers leading to more decentralized decision making.

For several reasons, countries in which firms decentralize gain economically. First, in those countries it is easier for more efficient firms to grow. Decentralization is essential for the creation of large firms, because CEOs are otherwise constrained in the number of decisions they can make directly -- it is therefore critical for enabling productive firms to grow large and to take market share from unproductive ones. Because trust is strongly linked with more decentralization, it in turn affects productivity. The absence of trust, in developing countries like Brazil, China, and India, explains why productive firms do not grow large and drive out unproductive firms from the market place.

Second, economies with low trust tend to specialize in industries where decentralization is less important. If some industries require high levels of decentralization in the organization, for example, complex electronics, then these industries will tend to locate in countries with high levels of trust. These may also be particularly productive industries. By contrast, industries that limited decentralization, like basic woodwork industries, will tend to develop in lower trust countries.

Third, greater trust appears to encourage both cross-country trade and cross-country investment. Multinational firms have a greater need to decentralize to foreign affiliates because of the local managers' better information, but they will be reluctant to do so when they do not trust the local management.

Finally, decentralization seems to directly influence within-firm productivity. One reason is that in highly centralized firms, many decisions do not get made because CEOs do not have the time to make them, and other managers do not have the authority to do so. For example, production management meetings happen less frequently in centralized firms because the CEO needs to attend but is too time-constrained to participate.

In addition to regional trust, three other factors stand out in explaining decentralization: rule of law, less hierarchical religion, and product market competition. The authors find that countries with better protection of the rule of law are characterized by more decentralized plants, arguably because efficient law enforcement reduces the probability that local managers will abuse their authority. Plants in locations with a high share of hierarchical religions are significantly more centralized, which is potentially attributable to a lower taste for decision making authority by local managers. Decentralization is also positively correlated with product market competition, possibly because of the greater importance of local knowledge in competitive environments, or because of disciplining effects that reduce the need for central monitoring. Together, social capital, rule of law, religious structure, and product market competition account for four-fifths of the cross-country variation in decentralization.

-- Claire Brunel