VoxEU Column Productivity and Innovation

The organisation of production and firm-level productivity

Reorganisation doesn’t always create a more efficient and effective firm. This column assesses the extent to which a firm’s physical productivity varies as a result of reorganisation. The results suggest significant variation. For policymakers, studying and understanding the internal organisational responses of firms to firm-specific and economy-wide shocks is essential to understanding the level and distribution of productivity in an economy.

Firms are extremely heterogeneous in their ability to transform a combination of inputs, like capital and labour, into the goods they decide to produce. Furthermore, the productivity of firms changes substantially over time (Asker et al. 2014). One can point to many potential sources of heterogeneity and changes in productivity. Some of these sources are exogenous to a firm’s actions and can be most simply modelled as idiosyncratic (like random innovations or production disruptions); others are the response to active organisational decisions by a firm’s owners and managers. These decisions can include responses to exogenous productivity disturbances, but also organisational responses to changes in demand and consumer tastes, factor prices, and other changes in the economic environment, like trade liberalisations and tax reforms.

Measuring the importance of endogenous organisational responses to productivity, demand and other shocks on the determination of a firm’s productivity is extremely important. Ultimately, the productivity of firms will determine the productivity of an economy and so these endogenous responses could be essential to understand the large observed total factor productivity differences across countries (Klenow and Rodriguez-Clare 1997). So how important are organisational decisions in determining a firm’s productivity?

Firm reorganisation and firm productivity

In a recent paper (Caliendo et al. 2015), we use detailed employer-employee matched data, as well as detailed firm production data, to measure the effect of a firm reorganisation on firm productivity. In doing so, we distinguish between different measures of productivity. An essential distinction is the one between revenue-based productivity and quantity-based productivity, namely, the ability of the firm to transform a set of inputs into revenue, compared to the ability of firms to transform a set of inputs into a number of units of a particular good. The first measure is commonly used to compare productivity across firms because a firm’s revenue is expressed in units of a common currency. However, a key complication is that it confounds the physical productivity of the firm with its pricing decisions. The quantity-based productivity of the firm refers only to its physical productivity, but it is expressed in units of the particular goods produced by the firm. Hence, it cannot be easily compared across firms unless one compares exceptionally homogenous products.

Consider a firm that experiences a surge in the demand of its product – say, because its product becomes fashionable. The increase in demand makes the firm want to increase production. To do so it chooses to hire more workers and perhaps change its organisational structure by hiring more upper-level managers (a division boss, an HR representative, perhaps an in-house accountant or lawyer). These organisational changes work against the decreasing returns faced by the firm when expanding production conditional on its current structure. By modifying its organisation, the firm can expand and increase its physical productivity. At the same time, the new organisational structure allows the firm to produce much more, and so the firm might optimally decide to move down its demand curve and charge a lower price. Hence, the response to a demand shock could result in an increase in quantity-based productivity and a decrease in revenue-based productivity.

In fact, the productivity response of any discrete decision of a firm can have similar productivity implications. Because discrete decisions (such as adding a new plant) lead to lumpy quantity and price responses, they can lead to differential quantity and revenue-based productivity responses. For the case of firm reorganisations that expand or contract the number of layers in a firm this is exactly what we find in Portugal for the period 1995-2005.

We measure reorganisations using changes in the occupational structure of firms as in Caliendo et al. (2012). As we show in Caliendo et al. (2015) for the case of France, in Portugal firms that reorganise change their labour force in ways that allow the firms to economise in the human capital (or knowledge) they employ. Namely, firms that expand and reorganise add management layers with well-trained experts, which allow them to routinise lower-level jobs for which they now hire less skilled and trained employees. The evidence that firms actively manage the knowledge characteristics of their employees to economise on labour costs is extremely robust and clear. The key feature is that firms that, say, face higher demand can reorganise to produce more and this allows them to achieve higher physical productivity. As in the theories developed in Garicano (2000), Garicano and Rossi-Hansberg (2006) and Caliendo and Rossi-Hansberg (2012), larger firms can use a better organisational technology due to their ability to concentrate knowledge in upper-level management. This allows them to economise on the knowledge, and therefore wage bill, of the lower level employees. This effect on the number of employees hired and on average wages in a layer is particularly clear in the data of both France and Portugal.

Using our data for Portugal, we measure the response of quantity and revenue-based productivity to a change in the number of layers of management of firms. We are careful to include industry, time, and, for many specifications, firm or sequence-of-layers-firm fixed effects. The results are quite stark. A reorganisation that increases by one the number of layers of management increases a firm’s hours of work by 25%, value added by slightly more than 3%, and quantity-based productivity by about 4%. That is, firms that add a layer are about 4% more productive in transforming inputs into units of output. In contrast, firms that reorganise and add a layer decrease revenue-based productivity by about 4%. That is, as a result of the reorganisation prices fall to an extent that converts the positive quantity-based productivity increase into a 4% decline. We show that these results are quite robust and hold under a variety of specifications. Furthermore, the results can be interpreted as causal since they are verified when we use an instrumentation strategy building upon lagged firm-specific variables.

Concluding remarks

In sum, a firm’s physical productivity varies significantly as a result of a firm’s reorganisation. On average, we show that among firms that add layers, the increase associated with reorganisations accounts essentially for all of the increase in the productivity experienced by these firms over time. Namely, for the average firm that adds layers all of the experienced increase in productivity comes from this source. Other exogenous and endogenous components might account for changes in particular firms, but do not matter on average, and therefore in the aggregate.

In our view, these results imply that studying and understanding the internal organisational responses of firms to firm-specific and economy-wide shocks is essential to understanding the level and distribution of productivity in an economy. An implication is that in countries where reorganisation is hard either because labour markets are very rigid, or because monitoring or delegating is costly, firms will not be able to reorganise and therefore will not be able to enhance their productivity and growth. This is consistent with the evidence in Hsieh and Klenow (2014) who show that firms in some developing countries like India and Mexico tend to grow less rapidly than in the US, over their lifecycle.


Asker, J, A Collard-Wexler and J De Loecker (2014), “Dynamic Inputs and Resource (Mis)Allocation”, Journal of Political Economy 122:5: 1013 – 1063.

Caliendo, L and E Rossi-Hansberg (2012), “The Impact of Trade on Organisation and Productivity”, Quarterly Journal of Economics 127:3: 1393-1467.

Caliendo, L, F Monte, and E Rossi-Hansberg (2012), “The Anatomy of French Production Hierarchies”, Journal of Political Economy 123:4: 809-852.

Caliendo, L, G  Mion, L Opromolla, and E Rossi-Hansberg (2015), “Productivity and Organisation in Portuguese Firms,” CEPR Discussion Papers 10993.

Garicano, L (2000), “Hierarchies and the Organisation of Knowledge in Production”, Journal of Political Economy 108:5: 874-904.

Garicano, L and E Rossi-Hansberg (2006), “Organisation and Inequality in a Knowledge Economy”, Quarterly Journal of Economics 121:4: 1383-1435.

Hsieh, C and P Klenow (2014), “The Life Cycle of Plants in India and Mexico”, Quarterly Journal of Economics 129:3: 1335-1384.

Klenow, P, and A Rodríguez-Clare (1997), “The Neoclassical Revival in Growth Economics: Has It Gone Too Far?”, in B S Bernanke and J J Rotemberg 73–114, Cambridge: MIT Press.

2,835 Reads