Most firms are born small, stay small, and innovate little (Decker et al. 2016, Foster et al. 2016). Only few firms achieve high growth and engage in innovation on the way to becoming large, successful businesses that contribute significantly to aggregate productivity and growth. Ultimately, we want to understand who these firms are and the factors that contributed to their emergence and eventual success. The recent literature suggests that a firm’s initial conditions, such as entrepreneurs’ past experience, and indicators of early innovation, such as initial patenting and trademarking, are precursors to subsequent high growth (Fazio et al. 2016, Brown et al. 2017, Choi 2017).
In a recent paper, we study both empirically and theoretically the role of venture capital (VC) funding – a key source of startup financing – in identifying promising startups and turning them into engines of growth (Akcigit et al. 2019). In particular, we examine the types of startups that get funded by venture capitalists, study the extent to which synergies between venture capitalists and startups and venture capitalist experience matter for firm growth and innovation, and evaluate how critical VC is for growth in the US economy.
Stylised facts on venture capital-backed firms
To motivate the model and its analysis, we first empirically investigate both the selection and treatment effects associated with VC involvement in a startup. The analysis is carried out by combining data on all employer businesses in the US from the Census Bureau’s Longitudinal Business Database (LBD), with data on patenting from the USPTO, and deal-level data on firms receiving VC funding from VentureXpert for the period 1980-2012. A key advantage of the data is that it enables us to track the evolution of employment and patenting for all employer businesses in the US, and, critically, differentiate between the experience of VC-funded firms and other firms in the economy. The data reveal several key facts.
First, venture capitalists pick startups near the time of their inception to fund and nurture. The age distribution of VC-backed startups at first VC funding is shown in Figure 1, where age is measured by the number of years since a startup first appears in the LBD as an employer business. The striking feature of Figure 1 is that 57% of VC-funded firms receive their initial funding within their first year as an employer business.
Among young firms in the economy, venture capitalists disproportionately fund those that show the most promise. Figure 2 plots the probability of ever receiving VC funding by employment growth quintiles, where the quintiles are based on growth during a startup’s first three years as an employer business. The probability of funding jumps nearly 200-fold as one moves from the lowest quintile to the highest. We find a similar pattern when focusing on firms’ patenting activities. The relationship between a startup’s probability of receiving its first VC funding within five years of its first patent application and the quality of its early patents is plotted in Figure 3. The quality of a patent is measured by the citations that it receives. The probability of receiving VC funding increases 5% as one moves from the lowest quintile to the highest. These patterns suggest that venture capitalists select firms that exhibit relatively high growth and high-quality innovation in their early years.
Figure 1 The distribution of firm age in the year of first VC funding
Figure 2 Early firm employment growth and the probability of receiving VC funding
Figure 3 Early patent quality and the probability of receiving VC funding
Next, consider the effects of venture capitalist involvement on subsequent firm outcomes. In Figure 4, we plot the distribution of employment growth measured ten years after initial VC-funding, conditional on survival. The figure shows that VC-funded firms have substantially higher mean and variance of growth than other firms in the economy. Yet, in this figure it is impossible to disentangle whether the higher growth is driven entirely by VC selection, or whether VC involvement helped turn these firms into engines of growth.
Figure 4 Distribution of employment growth rate (measured 10 years post VC funding)
To assess the magnitude of VC treatment effects, the selection of startups by venture capitalists must be taken into account. To control for selection based on observables, we match VC-funded firms with observationally similar non-funded firms along key dimensions, including year of initial funding, industry, state, age, and employment. Figure 5 plots the evolution of (ln) average employment for VC-funded firms and their matched counterparts over the period spanning three years prior to initial funding to 10 years afterwards. Among firms that patent, Figure 6 plots the evolution of (ln) citation-adjusted patent stock of VC-funded firms and their matched counterparts.
In both figures, the VC-funded and non-funded groups exhibit virtually identical trajectories before VC funding. However, subsequently VC-funded firms grow and innovate much more. Average employment increases by approximately 475% by the end of the horizon for VC-funded firms, whereas growth is much more modest for the control group (230%). Similarly, the average patent stock of VC-funded firms grows by about 1,100% over the 10-year horizon, as opposed to 440% for the control group. These results suggest that venture capitalists play an important role in the making of successful firms.
Figure 5 Evolution of average employment before and after initial VC funding
Figure 6 Evolution of average quality-adjusted patent stock before and after initial VC funding
To understand a potential source of these treatment effects, we examine the heterogeneous impact on firm outcomes of being funded by more experienced versus less experienced venture capitalists. To do so, we first divide venture capitalists into two groups. Venture capitalists in the top decile of the ‘total deals’ distribution are labelled as “high quality” (high experience), and the remaining venture capitalists are labelled as “low quality” (low experience). Then, VC-funded firms are separated into those funded by high-quality versus low-quality venture capitalists. Figure 7 plots the evolution of (ln) average employment of firms in each of these categories, and Figure 8 plots the evolution of their (ln) average quality-adjusted patent stock.
While firms backed by high- and low-quality venture capitalists are similar prior to VC involvement, the average employment and average patent stock of startups funded by high-quality venture capitalists is higher after VC involvement, and the gap between the two groups widens over the 10-year horizon. By the end of the horizon, average employment grows by about 400% in the high-quality group, versus 320% in the low-quality group. Similarly, by the end of the horizon, the average patent stock grows nearly 50-fold for the high-quality group, and only 19-fold for the low-quality group. We confirm that startups funded by high-quality VCs have better employment outcomes through a regression analysis that controls for both startup characteristics and initial funding infusion. These findings suggest that factors beyond funding, such as expertise and management quality associated with high-quality venture capitalists, matter for subsequent firm outcomes.
Figure 7 Evolution of average employment before and after initial VC funding, by quality of VC
Figure 8 Evolution of average quality-adjusted patent stock before and after initial VC funding, by quality of VC
Quantifying the impact of VC on aggregate growth
Motivated by the empirical findings, we develop a macroeconomic model to match the salient features of VC in the US data. In the model, startups are born, some with better ideas than others. All startups need financing to bring their ideas to market through research and development. The model features two types of financiers – banks and venture capitalists. While both financiers provide funding and expertise to startups, venture capitalists offer a higher level of expertise than banks. The framework also stresses the complementarity between a talented entrepreneur’s skill and that of a venture capitalist. Critically, not all talented entrepreneurs can find a venture capitalist to back their startup. Consequently, they turn to banks, which cannot provide the same nurturing as venture capitalists. For a talented entrepreneur, matching with a venture capitalist as opposed to a bank results in a higher probability of success, a greater level of funding for R&D, and a higher productivity.
The model is calibrated using key micro-level moments from the empirical analysis. It is then used to assess how synergies and assortative matching between entrepreneurs and venture capitalists, and the differential taxation of startups affect economic growth. We start by eliminating the synergies between VC funders and entrepreneurs by shutting down VC funding altogether (that is, all startups are funded by banks). Doing so lowers aggregate growth by 28%. This decline is driven by lower survival and innovation rates among talented entrepreneurs.
Next, we evaluate the importance of assortative matching – the fact that talented entrepreneurs are more likely to be funded by venture capitalists – by introducing purely random matching. This exercise leads to a 1% lower growth rate. We also consider the case of perfectly assortative matching – where VCs fund only talented entrepreneurs – and observe a 9% increase in growth. The higher growth in the latter case highlights the fact that there is room to increase the existing extent of assortative matching in the economy. Finally, we consider the impact of raising the tax rate on VC-funded startups. VC-funded startups are taxed when they are floated or sold at the capital gains rate (17%). If the VC-funded firms were instead taxed at the corporate income tax of 35%, growth would decline by 18%.
These quantitative exercises, along with our empirical analysis, highlight the important role VC plays in the aggregate economy. Venture capitalists disproportionately target the most promising young startups and play a critical role in turning these startups into engines of economic growth.
Authors’ note: Any opinions and conclusions expressed herein are those of the authors and do not necessarily represent the views of the US Census Bureau, Federal Reserve System, Board of Governors, or its staff. All results have been reviewed to ensure that no confidential information is disclosed.
Akcigit, U, E Dinlersoz, J Greenwood, and V Penciakova (2019), “Synergizing Ventures”, Federal Reserve Bank of Atlanta Working Paper no. 2019-17.
Brown, D, J Earle, M Jung Kim, and K Min Lee (2018), “High-Growth Entrepreneurship”, IZA Discussion Paper no. 11662.
Choi, J (2017), “Entrepreneurial Risk Taking, Young Firm Dynamics, and Aggregate Implications”, Unpublished paper, University of Maryland.
Fazio, C, J Guzman, F Murray, and S Stern (2016), “A New View of the Skew: A Quantitative Assessment of the Quality of American Entrepreneurship”, MIT Innovation Initiative Policy Brief Series.
Foster, L, C Grim, and N Zolas (2019), “A Portrait of U.S. Firms that Invest in R&D”, Economics of Innovation and New Technology.
Decker, R, J Haltiwanger, R Jarmin, and J Miranda (2016), “Where Has All the Skewness Gone? The Decline in High-Growth (Young) Firms in the U.S.”, European Economic Review, 86 (July).