What determines productivity?

Chad Syverson 25 June 2010



Productivity – the efficiency with which firms transform inputs into outputs – is the elixir of economic success. Nations that enjoy rising productivity experience sustainable growth that simplifies a broad swath of economic and social problems. The same is true at the corporate level. Naturally then, productivity is the focus of a great deal of government and corporate policy. But what determines productivity?

A full, detailed and widely accepted answer to this critical question still eludes the economic profession. Productivity, after all, is a residual – the variation in output not explained by observable inputs. It is therefore something like a measure of our ignorance.

During the past 25 years, economists have gained increased access to broad and detailed data on firms’ production activities. This has sparked a broad research effort to identify the determinants of productivity. In other words, to “put faces on” that residual.

When Bartelsman and Doms (2000) reviewed the state of knowledge in this area a decade ago, one of the most prominent and substantive findings were the enormous and persistent differences in measured productivity across producers, even within narrowly defined industries.

That some producers obtained twice as much output (or more) from the same measured inputs shaped research agendas over the past decade in a number of fields, including macroeconomics, industrial organisation, labour, organisational economics, and trade. Recent research efforts have been directed at trying to understand why these productivity differences exist and are so ubiquitous. What are their sources, and what supports them in equilibrium?

This new research, which I review in Syverson (2010), has offered explanations that can broadly be classified into two sets.

The new findings offer two categories of explanations for inter-firm productivity differences.

  • One includes factors that operate primarily within businesses, be it at the firm, plant, or even production line level. These are potentially under the control of management or other economic actors inside the firm.
  • The second set contains elements external to the firm. The impact of these “environmental” factors might not always be direct, but they can affect producers’ willingness and ability to harness factors in the first set. They may also influence the amount of productivity dispersion that is sustainable in equilibrium.

Within-business determinants

Research has pointed to several within-business productivity drivers, which of each is, in essence, an input that is not measured or mis-measured in the standard data sets.

  • Managerial practice/talent

Managers are conductors of an input orchestra, coordinating the application of labour, capital, and intermediate inputs. Just as a poor conductor can lead to a cacophony rather than a symphony, poor management can lead to discordant production operations.

  • Higher-quality general labour and capital inputs

Quality differences in standard inputs like capital and non-managerial labour (coming from capital-embodied technology in the former case and human capital in the latter, for example) will lead to measured productivity differences if standard input measures don’t fully reflect quality differences.

  • Information technology and research and development

It is arguably the general purpose technology of our time, and as such has great potential for broad productivity effects. R&D reflects investment in a knowledge stock that can raise productivity.

  • Learning-by-doing

The very act of operating can increase productivity, as experience allows producers to identify opportunities for process improvements.

  • Product innovation

While innovations in product quality may not necessarily raise the physical quantity of output a firm obtains from a given set of inputs, they can enhance productivity on a quality-adjusted output basis.

  • Firm structure decisions

The organisational structure of the firm’s production units – the vertical and horizontal linkages between the industries they operate in, their relative sizes, etc. – can affect the productivity levels of the firm’s component business units.

‘Environmental’ determinants

The second set of productivity factors, so-called environmental determinants – affect productivity in two ways. Either they incentivise individual producers to become more efficient, or they foster Darwinian selection that shifts economic activity toward more efficient producers.

  • Productivity spillovers

These across-firm externalities are often discussed in the context of classic agglomeration mechanisms like thick-input-market effects and knowledge transfers. Note that the latter, in particular, does not need to be tied to any single geographic market.

  • Competition

Pressures from threatening or actual competitors – whether from other producers in the same market or foreign competitors operating through trade channels – affect productivity levels within an industry. Competition fosters efficiency-based selection as lower-cost producers take market share from their less efficient competitors. Competition also raises the productivity bar that new producers must meet to successfully enter the market. More directly, heightened competition can induce firms to make productivity-raising efforts that they may otherwise not.

  • Deregulation or proper regulation

Poorly regulated markets can create perverse incentives that reduce productivity. Deregulating or reformatting to smarter forms of regulation can reverse this.

  • Flexible input markets

Just as one can think of competition as flexibility in product markets, more flexible input markets can also raise productivity levels. Indeed, there are almost surely complementarities between product- and input-market flexibility. When consumers want to reallocate purchases across producers, firms need to be able to easily reallocate inputs to meet the new demand pattern.

Questions that remain unanswered

While we have made considerable progress explaining the sources of productivity differences across business, many pressing issues remain. I see the following as some of the most important questions that the productivity research agenda should address going forward.

  • Which productivity drivers matter most?

The relative quantitative importance of the factors listed above is still unclear. Of course, it is quite likely that quantitative impacts vary across industries or markets, so which factors matter most in what sectors, and what features of the sector determine this?

  • What is the importance of demand?

Productivity is typically thought of as a supply-side concept. However, because production microdata typically lacks producer-specific price information, within-industry price differences – often caused by differences in demand conditions – are embodied in output and productivity measures. High- (low-) price producers look more (less) technically efficient than they really are. Research has begun to separately model and measure technological and demand idiosyncrasies and to explore their roles in businesses’ growth and survival, but there is much left to learn.

  • How can government policies encourage productivity growth?

This question is particularly applicable to the environmental set of productivity drivers, which by their very nature they are the easiest to manipulate via government policy. We need to know more about what kinds of reforms are most effective for different types of markets or frictions, and their optimal size and timing.

  • How important is input misallocation in emerging economies?

Productivity differences explain much of the per capita income variation across countries. Recent work has been building a case that a substantial portion of these productivity gaps arise from poor allocation of inputs across production units in developing countries. But while we know some distortions exist, we haven’t really yet pinned down exactly what those distortions are.

  • What is the importance of higher variance in productivity?

Some new work suggests that the variance of productivity outcomes may be broadly increasing. It isn’t clear exactly what is driving this, or whether the forces of selection will stem this spread.

  • Can we predict innovation?

Innovation drives productivity growth. We need to know more about how features of output markets and technologies spur innovative activity, and whether product or process innovation will dominate in any particular setting.

  • What is the nature of intangible capital?

One way to explain the observed persistence of productivity differences across businesses is to interpret firms’ application of the first set of productivity drivers as investments in intangible capital stocks – the creation of business know-how that is embodied in the organisation. Understanding more about how such stocks are built and sustained (which is, alas, inherently difficult given their nature) would shed light on many productivity-related issues.

  • Is it management or managers?

Management practices seem to matter, but we don’t know if these practices alone are sufficient to raise productivity. Can any organisation that follows “the cookbook” become more efficient, or must such practices be implemented by managers with complementary skills?

It is clear that, as much as we’ve learned about the factors that create and support productivity differences, much is left to be done. Fortunately, I see no sign that our knowledge accumulation rate is slowing in this area.

The need for data

I will close with a request (plea?) to economists and policymakers for a renewed focus on data.

Virtually everything we now know stems from detailed production data accessible to researchers. But much of this data was originally collected by statistical agencies in order to construct aggregates, not to learn about productivity per se. The data’s ability to offer insights into productivity was in many ways a positive externality.

Now that we know the value of such information, more directed efforts to record currently unmeasured aspects of business’s production practices (such as data on managers and management practices, producer-level prices, input quality measures, intangible capital proxies, innovation spending, and so on) should be a priority.

Collecting more data obviously involves costs, but it seems clear that there is much to be gained in exchange.


Bartelsman, Eric and Mark Doms (2000), “Understanding Productivity: Lessons from Longitudinal Microdata”, Journal of Economic Literature, 38(3):569-594.

Syverson, Chad (2010), “What Determines Productivity?”, Journal of Economic Literature.



Topics:  Productivity and Innovation

Tags:  productivity, Management, firms

Professor of Economics at the University of Chicago Booth School of Business


CEPR Policy Research