How Destructive is Innovation?

By
Daniel Garcia-Macia, International Monetary Fund
Chang-Tai Hsieh, University of Chicago and NBER
Peter J. Klenow, Stanford University and NBER

Abstract

Entrants and incumbents can create new products and displace the products of competitors. Incumbents can also improve their existing products. How much of aggregate productivity growth occurs through each of these channels? Using data from the U.S. Longitudinal Business Database on all non-farm private businesses from 1976–1986 and 2003–2013, we arrive at three main conclusions: First, most growth appears to come from incumbents. We infer this from the modest employment share of entering firms (defined as those less than 5 years old). Second, most growth seems to occur through improvements of existing varieties rather than creation of brand new varieties. Third, own-product improvements by incumbents appear to be more important than creative destruction. We infer this because the distribution of job creation and destruction has thinner tails than implied by a model with a dominant role for creative destruction.

Conclusion

How much innovation takes the form of creative destruction versus firms improving their own products versus new varieties? How much innovation occurs through entrants versus incumbents? We try to infer the sources of innovation from the employment dynamics of U.S. firms in the non-farm private sector from 1976–1986 and 2003–2013. We conclude that creative destruction is vital for understanding job destruction and accounts for around one-fourth of growth. Own-product quality improvements by incumbents appear to be the biggest source of growth. Net variety growth contributes much less than quality improvements.

Our findings could be relevant for innovation policy because the sources of growth we identify have different business stealing effects versus knowledge spillovers. The importance of creative destruction ties into political economy theories in which incumbents block entry and hinder growth and development, such as Krusell and Rios-Rull (1996), Parente and Prescott (2002), and Acemoglu and Robinson (2012). And creative destruction underscores the employment dislocations that come along with some growth.

It would be interesting to extend our analysis to individual sectors, other time periods, and countries. Retail trade experienced a big-box revolution in the U.S. led by Wal-Mart’s expansion. Online retailing has made inroads at the expense of brick-and-mortar stores. In Chinese manufacturing private enterprises have entered and expanded while state-owned enterprises have closed (Hsieh and Klenow, 2009). In India, incumbents do not expand as much as in the U.S. (Hsieh and Klenow, 2014) and therefore contribute less to growth. Our accounting is silent on how the types of innovation interact. In Klette and Kortum (2004) more entrant creative destruction discourages R&D by incumbents.

Or, as stressed by Aghion et al. (2001), a greater threat of competition from entrants could stimulate incumbents to “escape from competition” by improving their own products. Creative destruction and own innovation could be strategic complements, rather than substitutes. Our conclusions are tentative in part because they are model-dependent.

We followed the literature in several ways that might not be innocuous for our inference. We assumed that spillovers are just as strong for incumbent innovations as for entrant innovations. Young firms might instead generate more knowledge spillovers than old firms do — Akcigit and Kerr (2015) provide evidence for this hypothesis from patent citations.

We assumed no frictions in employment growth or misallocation of labor across firms. In reality, the market share of young firms could be suppressed by adjustment costs, financing frictions, and uncertainty. On top of adjustment costs for capital and labor, firms may take awhile to build up a customer base, as in work by Gourio and Rudanko (2014) and Foster et al. (2016). Irreversibilities could combine with uncertainty about the firm’s quality to keep Young firms small, as in the Jovanovic (1982) model. We defined young firms as those younger than five years, but these dynamics could play out for longer. Meanwhile, markups could vary across varieties and firms. All of these would create a more complicated mapping from firm employment growth to firm innovation.

Leave a Reply

Your email address will not be published. Required fields are marked *

Scroll to top