Countries all over the world are suffering major disruptions from the Covid-19 pandemic. In particular, new and young firms may be disproportionately disrupted by the crisis. If these firms are also considered to be important for innovation, disruptions to their activities could drag down overall productivity growth beyond just the next year. Numerous studies (e.g. Haltiwanger et al. 2013, Moreira 2016, Sedláček and Sterk 2020) have found that young firms are an important source of job creation, on average and over the business cycle.
How does one quantify the contribution of new and young firms to innovation and growth? Many studies look at patenting or R&D spending to assess innovation output or input. While this approach has been extraordinarily fruitful, it is limited by the propensity to patent or report R&D. For one, patents may not capture innovation in non-manufacturing industries. Consider the contrast between Walmart and Ford. Walmart has about 1,000 patents, whereas Ford has 30,000 (Justia Patents 2020). However, Walmart is many times larger than Ford in terms of sales and employment, having helped to revolutionise the retail sector. Even within manufacturing, the propensity to patent a particular innovation differs across firms. For consumer products, Argente et al. (2020) find that half of product innovation comes from firms that do not patent, and that larger firms have a much higher propensity to patent a new product compared to smaller firms. Since new and young firms are likely to be small, measuring innovation using patents alone may understate their creativity.
Similar to patents, about 85% of reported R&D spending in the US comes from manufacturing and information sectors (National Science Foundation BRDIS 2016). But these sectors produce approximately 20% of aggregate value added. R&D is a less useful guide to innovative activity for ‘the other 80%’ of the economy. For example, R&D would not successfully capture the revolution in the services, retail, or wholesale sectors. Further, conditional on reporting positive R&D spending, R&D intensity (spending relative to sales) does not appear to increase with firm size. As a result, even though larger firms are much more likely to report R&D spending, this could reflect a higher propensity to report, rather than a greater contribution to innovation.
Another common approach is to use growth in firm ‘revenue productivity’ (revenue relative to inputs) to evaluate innovation within firms. This is often implemented using revenue per worker (i.e. labour productivity). The rationale is that the most innovative firms should enjoy higher revenue per worker because their innovations give them a bigger lead over their competitors, enabling them to charge higher markups. However, firm markups could stem in part from natural monopoly power or government restrictions on entry (rather than innovativeness). Moreover, stable firm revenue productivity can be consistent with firms innovating by adding more markets and products with similar price-cost markup. This might be the case, for example, with Apple adding iPhones and then iPads, or Walmart adding new stores. Figure 1 illustrates the case of Walmart, whose sales and employment grew tremendously as their stores spread across the country. In contrast, Walmart’s revenue per worker appears flat. Thus, employment looks like a better proxy than revenue per worker for Walmart’s innovativeness. Figure 2 provides the more recent example of Amazon’s ascent. Amazon’s sales soared while its revenue per worker rose much more modestly. These patterns can also be seen in other innovative firms such as Microsoft, Google, Facebook, Apple, and Starbucks.
Figure 1 Cumulative growth in employment and revenue productivity, Walmart
Figure 2 Cumulative growth in employment and revenue productivity, Amazon
Source: Authors calculations from Compustat data.
As the examples of Walmart and Amazon illustrate, the market share of a firm may be informative about its innovative activities. In our recent study (Klenow and Li 2020), we use market share to gauge the contributions of firms of various sizes and ages to US productivity growth in recent decades. Out study builds on Feenstra (1994), who infers quality and variety growth from a product’s market share. In our framework, all firms have the same revenue productivity for simplicity. A firm’s market share increases when it adds a product. Similarly, if firms improve their own products (e.g. successive car models), their market shares should rise relative to that of firms whose products do not improve.
As pointed out by Garcia et al. (2019), however, not all market share gains contribute the same amount to growth. If a firm gains market share by introducing a brand new variety or by improving its own product, it contributes more to growth than if it replaces an incumbent producer. The latter is known as ‘creative destruction’, examples of which include Walmart and Amazon replacing mom-and-pop retailers, and Apple iPhones grabbing market share from Blackberry and Nokia devices. Distinguishing creative destruction from ‘new varieties’ or ‘own innovation’ is important because firms arguably have too much incentive to steal business from their competitors, as opposed to producing genuinely novel products or improving their own products. New and young firms may be especially good at coming up with novel products, but may also have more incentive to overtake businesses than incumbents who already enjoy a big share of the market. This finding is known as the Arrow Replacement Effect (Arrow 1962).
In the absence of detailed information about products, our study uses the establishments owned by each firm as a proxy for its products. This captures Walmart moving into new geographic markets and car firms that devote each plant to devoting a different car model. Using plant entry and exit as a proxy for product entry and exit, we can separate out creative destruction (entry accompanied by exit) from new varieties (net entry). Further, when an incumbent improves its own products, this should show up as a rising market share of its surviving plant. Since plant data is available in the US for all sectors, our method can be applied equally well to firms inside and outside of manufacturing.
We use the Longitudinal Business Database from the US Census Bureau, which contains data on employment and payroll at all plants within all non-farm private employer firms in the US over 1982-2013. Strikingly, we find that new and young firms (aged 0 to 5 years) account for almost one-half of growth – three times their share of employment. Table 1 presents the contributions from new, young, mature, and old firms in the US. Even though new firms represent only 10% of all firms (and only 3% of employment) in a given year, they are responsible for almost 30% of US productivity growth (on average). Young firms (aged 1 to 5 years) contribute approximately one-fifth of growth, notably higher than their 13% share of employment. In contrast, mature and old firms (aged 6+) employ 83% of workers but contribute only one-half of overall growth.
Table 1 Contribution to US growth by firm age, 1982-2013
Source: Author’s calculations from the U.S. Census Bureau Longitudinal Business Database.
Notes: Average growth rate over the entire period is 1.66%. Firm age is the difference between the year of observation minus the birth year of the firm’s first plant. Young = age 1-5. Mature = age 6-10. Old = age 11+.
To understand why new and young firms contribute disproportionately to growth, it is useful to look at the percent of employment at new plants which can be traced to these firms. The last row of Table 1 shows that the new and young firms are responsible for over two-thirds of all employment at new plants. According to Hsieh and Klenow (2017), they account for one-half of all gross job creation. Finally, new firms contribute predominantly through new varieties, while young firms contribute mostly through own-innovation. As we mentioned previously, the market share gains associated with new varieties and own-innovation contribute more to growth than identical market share gains associated with creative destruction.
To recap, our growth accounting approach uses the dynamics of firm and plant market shares to infer the growth contributions of various firms. Compared to studies focusing on patents and R&D spending, we find a much bigger role for new and young firms (including those outside manufacturing) in accounting for productivity growth. Our findings illustrate the potential importance of policies to help new and young firms weather the Covid-19 crisis, in order to prevent longer term damage to overall innovation and growth.
Authors' note: The views expressed do not necessarily reflect the views of the Federal Reserve Bank of San Francisco, the Board of Governors or the Federal Reserve System.
Argente, D, S Baslandze, D Hanley, and S Moreira (2020), “Patents to Products: Product Innovation and Firm Dynamics”, working paper.
Arrow, K (1962), “Economic Welfare and the Allocation of Resources for Invention”, in Universities-National Bureau Committee for Economic Research (ed.) The Rate and Direction of Inventive Activity: Economic and Social Factors, Princeton, NJ: Princeton University Press. pp. 609-626.
BRDIS (2016), “Business R&D and Innovation Survey”, National Science Foundation.
Feenstra, R C (1994), “New product varieties and the measurement of international prices”, The American Economic Review 84(1): 157-177.
Garcia‐Macia, D, C T Hsieh and P J Klenow (2019), “How destructive is innovation?”, Econometrica 87(5): 1507-1541.
Haltiwanger, J, R S Jarmin and J Miranda (2013), “Who creates jobs? Small versus large versus young”, Review of Economics and Statistics 95(2): 347-361.
Hsieh, C T and P J Klenow (2017), “The Reallocation Myth”, Fostering a Dynamic Global Economy: Federal Reserve Bank of Kansas City Economic Policy Symposium.
Justia Patents (2020), “Patents by Company”.
Klenow, P J and H Li (2020), “Innovative Growth Accounting”, in NBER Macroeconomics Annual 2020, volume 35. Chicago, IL: University of Chicago Press.
Moreira, S (2016), “Firm Dynamics, Persistent Effects of Entry Conditions, and Business Cycles”, working paper.
Sedláček, P and V Sterk (2020), “Startups and employment following the COVID-19 pandemic: A calculator”, VoxEU.org, 25 April.