Key takeaways
- Congress is considering approaches to extending the existing exemption from the Medicare Drug Price Negotiation Program for small employers. In trying to spur innovation, we question whether small employers are the most effective way to target support for innovation. Based on extant evidence on where the economic vulnerabilities to innovative firms lie, we find that policy to support small biotech works best when it reaches young firms with no product revenue and promising science, at the point where they cannot fund the next experiment. A refundable research and development (R&D) credit, paid promptly, reaches those firms more directly than an exemption from Medicare drug price negotiation
- Emerging firms drive drug innovation. About two-thirds of the novel drugs launched worldwide recently started in emerging biopharma companies, up from less than half in the five years prior. That shows these firms matter. It does not show that all of them need help.
- “Small” doesn’t mean “financially strapped.” About one in five small biotechs hold less than $10 million in cash. In the U.S., an estimated 59% of small biotech firms had less than two years of cash left, while 20% hold at least $300 million. The Small Biotech Innovation Act counts a firm’s qualifying drugs for Medicare negotiation and measures its R&D spending against net revenue. Neither measure separates firms that need money from firms that already have it.
- A negotiation exemption helps these firms only indirectly. Its value to a firm with no revenue depends on several uncertain factors well in the future, such as drug approval, Medicare sales more than a decade later, and selection for negotiation.
- A better-targeted policy approach involves a capped, refundable R&D credit. It would go to independent young firms with little or no product revenue. Refunds could be paid quarterly, or as advances against an approved research budget, so the money arrives when it’s needed. Milestone grants can complement the credit when careful scientific selection matters, or when a firm needs funding before it can spend. And success can be measured by research progress, not firm survival. An acquisition can be a good outcome.
Introduction
Recent efforts to reduce prescription drug prices have renewed debate over how to sustain biotechnology innovation. One response has been to propose exemptions from the Medicare Drug Price Negotiation Program for selected manufacturers and products. These proposals raise two distinct questions: which firms or research programs warrant assistance, and what form of assistance is most likely to generate additional socially valuable research?
The Small Biotech Innovation Act, introduced as S. 1930 and H.R. 3731, illustrates this approach. The introduced Senate bill would create an exception beginning in 2029 for drugs by manufacturers with five or fewer qualifying single-source drugs that meet specified R&D investment thresholds. Those thresholds rise from 30% to 70% according to the number of qualifying drugs and use average net revenue over the previous three years to estimate R&D intensity.1
We argue that product counts and R&D intensity provide an incomplete basis for targeting support to financially constrained innovators. Counts of qualifying single-source drugs are not a suitable proxy for understanding which firms need financial support. Small firms differ substantially in financial circumstances and access to capital. Young firms without product revenue may face particular difficulty financing the research needed to demonstrate their prospects to investors. These firms could potentially hold valuable innovations that cannot be produced due to early-stage financial constraints. If that is the problem policy seeks to address, assistance should prioritize promising research during this vulnerable period.
A negotiation exemption can in theory increase expected commercial returns and may thereby attract investment before approval. Its effect on immediate financing needs, however, depends on a series of uncertain links. Targeting small firms for the exemption also risks paying a well-financed firm for innovation that they would have pursued absent the subsidy. In contrast, a targeted refundable R&D credit, paid promptly, offers a more direct way to support research during periods of financing scarcity. The credit’s effectiveness would depend on eligibility, payment timing, and the extent to which it finances research that would otherwise be delayed or abandoned. We first explain the economic rationale for assistance, then examine targeting, assess exemptions, and consider the design of a more direct alternative.
Why promising biotech research may be underfinanced
Drug development involves substantial uncertainty, and the termination of unsuccessful projects is part of an efficient innovation process. Financial distress alone therefore does not establish a case for public support. The relevant question is whether projects with expected social benefits exceeding their development costs are interrupted because private financing is inadequate.
Two economic problems can contribute to such underinvestment. First, research can produce knowledge that benefits other developers and is not fully captured by the originating firm. A discovery in one firm, sector, or country can trigger new avenues of research, inspire new research projects, or find new applications in other firms, sectors, or countries.2 In this case, the innovative firms themselves do not reap the full rewards of their R&D, and thus, the social benefits outweigh the private benefits. This creates a disincentive for firms to invest more in R&D, because they do not stand to gain from this additional investment. This results in an environment in which too little research is being done. In an ideal setting, the private returns to R&D would be equal to the social returns to R&D. An older Federal Reserve paper (but still relevant due to the lack of change in R&D incentives) puts forth that a conservative estimate of the externalities would be a social return of 30% and a private rate of return to capital of 7%––which would suggest that optimal R&D spending as a share of GDP is more than four times larger than actual spending.3 A 2026 Congressional Research Service report estimates similarly that social returns to R&D are typically two to four times that of private returns (with the highest estimates reaching 20 times).4 A young pre-revenue firm bears the full cost and risk of research and faces unique financial constraints while competitors and society capture some benefits through published knowledge, skilled workers, and follow-on innovation.
Second, in biotech specifically, outside investors may have difficulty evaluating a young firm’s scientific prospects. A developer may need to complete experiments or early clinical work before investors can judge their technology yet lack the resources to produce that evidence.5 Thus, it is often the case that investors cannot easily distinguish genuinely promising science from weak projects, and they may underfund even good firms.6 This leads investors to put their money into more established, less risky science rather than truly new ideas.7 As opposed to widely well-known ideas, small companies tend to have specialist knowledge that make it difficult for outside investors to evaluate. This means that it is challenging for a small firm to bring even a good idea to fruition. These two problems create a rationale for assistance, although their importance must be established empirically.
The timing of financial assistance matters. Funding that enables a firm to reach an informative milestone can reduce scientific uncertainty and improve its ability to obtain subsequent private financing. Howell provides evidence for such a mechanism, although its applicability to biotech firms requires attention to the industry’s longer development periods and scientific risks.8
The policy objective should be to maintain and advance valuable research. Acquisition or licensing can accomplish that objective even when a company ceases to operate independently. Conversely, continued operation does not necessarily indicate meaningful R&D progress. Policy assessment should therefore consider development milestones and the continuation of research programs alongside firm survival.
Targeting firms with promising research and financing needs
Emerging firms play a substantial role in the creation of new innovative drugs. A recent IQVIA report focuses on emerging biopharma companies, using thresholds of R&D spending less than $200 million or prescription sales less than $500 million a year to define these companies.9 For origination analyses, companies are classified at the time of the first patent filing or first clinical trial. Table 1 reports the share of drugs subsequently launched as novel active substances that were originated by emerging biopharmaceutical companies (EBPs) (IQVIA defines launches as the first commercial availability of a drug).
The table shows that in recent years, about two-thirds of new novel drugs launched were originated by emerging biomedical firms. These findings establish the importance of emerging companies as sources of innovation. They do not establish that all such firms lack financing or that a broad subsidy would induce substantial additional research.
Financial heterogeneity in this context is central to efficient targeting. The financial data assembled for this paper indicate that approximately 22% of small biotech firms held less than $10 million in cash, while 20% held more than $300 million.10 In a U.S. sample, 59% of small biotech firms11 had less than two years of cash runway, and 27% had at least three years.12 These data caution against treating small firms as uniformly exposed to financing shortfalls.
Being a young enterprise and the absence of product revenue are characteristics that may improve targeting. Such firms have limited internal resources and are often still developing the evidence needed to attract investors or partners. Research on high-technology firms finds that young firms’ R&D can be more sensitive to capital availability than that of mature firms.13 A German survey cited in the underlying evidence reports that 39% of start-ups less than five years old would expand innovation projects following a substantial cash injection.14 These findings support the relevance of financing, but neither establishes a biotech-specific five-year eligibility boundary.
Acquisition and licensing can provide resources or transfer research to a better-financed developer. In the analytic sample of 185 acquired small biotech firms summarized in Table 2 (see Appendix B for a more complete description of the data we assembled), the median age at acquisition was eight years, and the mean was 10 years. These observations indicate that acquisition often occurred after firms had financed several years of activity. Note that the sample includes only acquired firms.
A separate descriptive exercise follows 320 U.S. based biotech firms with less than 100 employees founded between 2016 and 2018 using PitchBook data (further described in Appendix B). As Table 3 shows, approximately 76% of the sample firms were operating privately or publicly in the most recent recorded data. Acquisition and closure each accounted for approximately 12%. These outcomes describe firm status roughly seven to ten years after founding; they do not identify when financing constraints arose, why firms closed, or whether their research continued elsewhere.
Recent financing trends provide further context. The venture capital (VC) and seed-round series assembled here and shown in Figures 1 and 2 indicate fewer transactions after the 2021 to 2022 period, with larger average VC deals (measured in millions of dollars) and a less consistent pattern in seed-round deal size. Specifically, Figure 1 highlights that since 2021, the number of early-stage VC deals has been declining. The decline since 2023 has been over 50% and was accompanied by a large increase in the median deal size. Figure 2 shows that the number of seed-round deals had a notable decline since 2021, but since 2023 the trend has been flatter. The median size of seed-round deals displayed a large spike in 2024 but has otherwise been relatively flat since 2021. We also examine alternative specification tests with larger sample sizes and see very similar patterns in all our tables and figures, which indicates the reliability of these trends (as seen in Appendix C). Changes in deal counts and sizes may indicate a shift in financing availability, but they cannot establish how many promising firms were denied capital.
The volatile nature of investment in biopharmaceutical R&D, combined with the inability of outside investors to accurately assess what science is promising, signals that particular corporate structures are being favored in ways that are not optimal. Whereas later in the developmental cycle, the market is effective at sorting bad and good ideas, in the earlier stages there is considerably more uncertainty, which leads to inefficient targeting of investments. This marks the need for more intentional targeting of support.
Taken together, the evidence supports using youth of firms and limited product revenue as screening criteria rather than treating either as proof of financial need. Eligibility could also reflect available cash relative to planned research spending and the cost of reaching a defined milestone. Scientific credibility remains essential: financial need alone cannot distinguish a promising but constrained developer from one whose prospects have deteriorated. A five-year age limit could simplify administration, but its advantages must be weighed against excluding older firms pursuing early development and including well-financed young firms.
How negotiation exemptions address financing constraints
The strongest case for a negotiation exemption is that expected future profits influence financing today. An increase in expected commercial returns could raise equity valuations, licensing payments, or acquisition offers before a drug reaches the market. An exemption could also increase resources available for R&D at a manufacturer that already sells an affected drug. The magnitude of any effect would depend on several uncertain links: eventual approval, Medicare sales, whether the drug would actually be selected for negotiation, and whether eligibility survives licensing or acquisition.
The question is how much additional research these channels would induce among firms facing financing constraints. For a pre-revenue developer, the current financing benefit depends on eventual approval, future Medicare sales (often 10 to 15 years in the future), selection for negotiation, and eligibility for the exemption when it would matter. Investors must also anticipate the effects of later transactions. Under the introduced Senate bill, an acquisition after 2029 by a manufacturer that fails the specified definition would end the acquired manufacturer’s drug eligibility beginning in the following plan year. This provision limits one route through which an exemption might otherwise increase an acquisition offer.
Direct benefits would accrue to qualifying drugs that would otherwise face negotiation, including commercially successful products whose development would have proceeded without the exemption.
An exemption may be defensible as a means of increasing expected returns to some selected drug development. The evidence presented here raises questions regarding the likelihood that such a policy would be an efficient means of resolving an immediate financing gap. That narrower objective calls for an instrument whose eligibility and payment timing are more closely tied to the research requiring support. Young, pre-revenue, or early post-revenue biotech firms bear the brunt of capital-driven risk aversion due to a stark lack of internal cash flows and an acute reliance on external equity or venture capital. Because novel drugs face lower likelihoods of regulatory approval and significant commercialization expenses, these financially constrained entities frequently depend on out-licensing agreements to fund their pipeline through partnerships with large pharmaceutical companies.15 Those types of deals have become less frequent over time, as shown in Figures 1 and 2.
Designing more direct assistance
In considering how to use simple instruments to promote R&D in financially vulnerable pharmaceutical companies, we consider one alternative that represents an important component of an R&D subsidy scheme. Currently, the U.S. has an R&D tax credit, but it is not refundable. That means that young pre-revenue research-based pharmaceutical firms are unlikely to benefit from such credits as a means of addressing immediate cash flow issues.16 A refundable R&D credit could aid firms without taxable profits and concentrate support during early development.17 Unlike an exemption whose value depends on a future commercial outcome, a credit tied to qualifying research expenditure would have a more immediate and observable connection to development activity. We would focus eligibility on independent young firms with little or no product revenue, while allowing development stage and financing needs to inform the final rules.
Refundability alone is insufficient. A credit that is paid long after expenses are incurred may arrive too late for a firm unable to finance the next experiment or clinical milestone. Quarterly refunds could reduce that delay. For firms that cannot first incur the expenditure, advances against an approved research budget could be considered, with reconciliation to verified spending. Advance payments would require controls against overpayment and misuse. These choices should be treated as central elements of a detailed proposal.
The credit should have a defined expenditure base and limits on support. Eligible costs could be tied to specified research and development activities, with clear treatment of contract research and related-party spending. A firm-level cap would limit large transfers to already well-financed recipients. Rules should account for other public support for the same expenditure and aggregate affiliated entities to discourage artificial restructuring. The subsidy rate and cap should be calibrated to plausible milestone costs and evaluated for their effect on additional research; the evidence assembled here does not establish particular amounts.
There is a tradeoff between targeting precision and administrative simplicity. An automatic credit based on age, revenue, and eligible expenditures would be relatively straightforward but would assist some firms that could finance the work privately. Requiring independent scientific assessment and a documented milestone financing gap could improve selection, while increasing review costs, delay, and uncertainty. A financing screen should also avoid favoring firms simply because they have the least cash or discouraging them from raising private capital.
One option would combine a capped refundable credit for a broadly defined early-stage group with competitive milestone grants for projects requiring more intensive scientific review or financing before expenditures occur. Grants are a serious alternative where project selection is central to the policy. The appropriate balance depends on whether the objective is to lower early research costs across eligible firms or to sustain a smaller set of promising projects facing identifiable funding gaps.
Evaluation should measure whether assistance changes development decisions. Relevant outcomes include completion of prespecified milestones, subsequent private financing, and continuation of research through licensing or acquisition. Comparisons around eligibility thresholds or phased implementation could help distinguish the effect of assistance from recipients’ underlying prospects. Firm survival and patent counts alone would be inadequate measures of success.
Conclusion
Small biotech firms are important sources of new drugs, but size and product counts are imprecise indicators of financing need. The significant heterogeneity with respect to financial circumstance in the cohort of small biotech firms suggests an important risk of much spending going to well-financed or already profitable enterprises that are well positioned to obtain private financing for their R&D. The case for targeted assistance rests on identifying research with substantial expected social value that private financing would otherwise fail to advance. Young firms without product revenue are a plausible starting point for eligibility, supplemented by measures of development stage, scientific credibility, and resources needed to reach an informative milestone. In our reading of the extant evidence, these firms would be the best bet for use of public funds to promote new innovative prescription drugs.
Negotiation exemptions may offer some improvements in financing by increasing expected returns, but their effect on immediate funding needs is indirect and highly uncertain. A promptly paid refundable R&D credit offers a closer connection between public assistance and current research expenditure. Where support depends on detailed scientific selection or must precede expenditure, milestone grants may be better suited. Policy design should be judged by its ability to advance valuable research that would otherwise be interrupted, relative to its public cost.
In summary, we view the targeting of small biotech firms for an exemption from the Inflation Reduction Act’s prescription drug negotiation program as making use of the wrong target and the wrong mechanism for efficiently promoting R&D in emerging financially vulnerable biotech enterprises. We believe the evidence indicates that focusing on tax policy through a targeted refundable tax credit is a better bet.
Appendix
Appendix A: Exploratory patent comparison
For 16 firms recorded as exiting through acquisition or going out of business in 2025, acquired firms held an average of 10.4 active patents, and distressed firms held 10.9. The small, selected sample and absence of measures of patent quality significantly limit interpretation. Patent counts do not establish comparable scientific platforms or identify whether financing caused closure.
The data for this table were collected using the baseline filters detailed in Appendix B. Additionally, these data were restricted to firms that faced an exit in 2025 and were founded between 2016 and 2018. Only Merger and Acquisition (M&A) control transactions and firms going out of business were included in this sample. This subset is thus not representative of all firm exits in the industry but rather demonstrates a select sample. The table shows similar levels of patenting between acquired firms and those exiting due to financial distress. This suggests that the two groups had similar levels of patents as platforms for the development of new products.
Appendix B: Data sources and description
The following criteria apply to the tables and figures in the paper, with the exception of Table 1 (data are from IQVIA). Therefore, Tables 2 and 3 and Figures 1 and 2 are created from these base criteria.
We have attempted to create a universe of biopharmaceutical firms that truly encapsulates firms that are innovating and discovering new drugs. We are not interested in the biopharmaceutical firms that are only manufacturing, distributing, or packaging drugs. We want to isolate research and development with new ideas. We are focused on the U.S. and on small firms. The following are the criteria used as the baseline in the creation of our datasets.
- Location: U.S.
- Employees: Max 100
- Industry: Drug Discovery (within Pharmaceuticals and Biotechnology, within Healthcare)
- Search Primary Industry Only
- Verticals: Life Sciences
- Verticals are investment themes or niche segments that cut across traditional industries.
- Keywords: PitchBook uses Boolean operators, and so we can use OR and AND, but also NOT to get rid of firms that may be out of our scope. Firms must match at least one term in the inclusion list, and none of the exclusion list to fit our criteria.
- OR list:
- novel therapeutic
- novel drug*
- new therapeutic*
- drug development
- therapeutic development
- drug discovery
- small molecule*
- targeted therap*
- precision medicine
- next-generation therap*
- first-in-class
- best-in-class
- innovative therapeutic*
- novel modality*
- new modality*
- NOT list:
- generics
- generic drug*
- biosimilar*
- contract manufacturing
- drug delivery
- CDMO
- CMO
- contract development
- API manufacturing
- active pharmaceutical ingredient*
- formulation
- packaging
- compounding
- OR list:
Data for Table 2 and Figures 1 and 2
In addition to the base criteria above, the data used to create Table 2 and Figures 1 and 2 are conditional on having a completed deal from January 1, 2016, to current day. Due to minimal missing data from PitchBook, the analytic samples differ slightly from total observations. For Table 2, M&A deals were isolated and analyzed (187 total observations, analytic sample is 185), for Figure 1, early-stage VC deals were isolated and analyzed (961 total observations, analytic sample is 870), and for Figure 2, seed-round deals were isolated and analyzed (504 total observations, analytic sample is 445)––all originating from the same greater dataset (total of 5,784 deals of all types).
Data for Table 3
In addition to the base criteria above, the data used to create Table 3 are restricted to firms founded between 2016 and 2018 (N=320 firms). The ownership status data are based on PitchBook’s most recently recorded available data.
Appendix C: Alternative criteria
An alternative sample set is comprised of less stringent criteria. Using PitchBook filters, this dataset would still be restricted to U.S. based firms with less than 100 employees, whose primary industry is drug discovery. However, these firms would also not need to match one of the keywords, and the firms matching the NOT keywords would not be excluded. This sample comes out to 11,804 total deals in the past 10 years as opposed to 5,784 total deals in the past ten years in our baseline dataset used above. While this broader dataset is clearly much bigger, it is likely less precise in that we may not be as closely targeting the firms with new innovations.
Either way, our data demonstrate similar patterns, emphasizing the reliability of these trends. A key difference is that the spike in median deal size we observed for seed-round deals in our earlier figure is largely flattened out. Below are the same figures and tables as introduced earlier in our paper but created from the broader set of criteria. The graph depicting early-stage VC deals is made up of 1,568 observations, and the graph depicting seed-round deals is made up of 926 observations. Table C1 detailing age at acquisition is made up of 410 observations, and Table C2 depicting current recorded ownership status is made up of 641 observations. Clearly, these alternative figures and tables have significantly larger sample sizes. However, the same conclusions can be drawn from either set of figures, as evidently, the figures and tables closely resemble the previously included figures and tables.
Observing both tables, it is clear that the results stay extremely consistent across both datasets, emphasizing the stability of our results.
Additional alternative criteria
Another set of alternative criteria further expands our baseline dataset. This broadest dataset replaces the Boolean operation AND with OR in our original criteria. This means that a firm could either have drug discovery as its primary industry OR it could match one of the keywords on the list, while still excluding all the words we are not interested in. This large dataset is comprised of 15,553 total deals of all types in the past 10 years (as opposed to 5,784 or 11,804). Despite the significantly larger sample size, which may make our universe of firms even less precise, we continue to observe the same trends in our figures and tables, as can be seen below.
Using this broadest set of criteria, the graph depicting early-stage VC deals is made up of 1,908 observations, and the graph depicting seed-round deals is made up of 1,244 observations. Table C3 detailing age at acquisition is made up of 559 observations, and Table C4 depicting current recorded ownership status is made up of 836 observations. Clearly, these alternative figures and tables have significantly larger sample sizes. Similar to the figures created by all drug discovery firms, the spike in 2024 has lessened. However, the same conclusions can be drawn from either set of figures, as evidently, the figures and tables closely resemble the previously included figures and tables, emphasizing the consistency of these patterns.
-
Acknowledgements and disclosures
The authors thank Loren Adler and Kristi Martin for their thoughtful comments on an earlier draft, Cameron Crider for fact-checking, and Rasa Siniakovas for editorial and publishing support.
-
Footnotes
- U.S. Congress. Senate. Small Biotech Innovation Act. S. 1930. 119th Congress. Introduced June 3, 2025. Section 2. https://www.govinfo.gov/content/pkg/BILLS-119s1930is/html/BILLS-119s1930is.htm.
- Pecoraro, Brandon, et al. 2026. “Estimating the Present Value of R&D Tax Benefits in the United States.” NBER Working Paper 35208. Cambridge, MA: National Bureau of Economic Research.
- Jones, Charles I., and John C. Williams. 1997. “Measuring the Social Return to R&D.” Finance and Economics Discussion Series 1997-12. Washington, DC: Board of Governors of the Federal Reserve System.
- Gravelle, Jane G., and Mark P. Keightley. 2026. The Federal Research and Development (R&D) Tax Credit. CRS Report R48848. Washington, DC: Congressional Research Service.
- Lerner, Josh. 1999. “The Government as Venture Capitalist: The Long‐Run Impact of the SBIR Program.” Journal of Business 72 (3): 285–318. https://doi.org/10.1086/209616.
- Alvarez-Garrido, Elisa, and Gary Dushnitsky. 2016. “Are Entrepreneurial Venture’s Innovation Rates Sensitive to Investor Complementary Assets? Comparing Biotech Ventures Backed by Corporate and Independent VCs.” Strategic Management Journal 37: 819–834. https://doi.org/10.1002/smj.2359.
- Nanda, Ramana, and Matthew Rhodes-Kropf. 2017. “Financing Risk and Innovation.” Management Science 63 (4): 901–18. Models how uncertainty about obtaining subsequent funding shifts investors toward less innovative ventures.
- Howell, Sabrina T. 2017. “Financing Innovation: Evidence from R&D Grants.” American Economic Review 107 (4): 1136–1164.
- IQVIA Institute for Human Data Science. 2025. Expanding Options for Emerging Biopharma in the U.S.: A Decade of Change. October 29, 2025. https://www.iqvia.com/insights/the-iqvia-institute/reports-and-publications/reports/expanding-options-for-emerging-biopharma-in-the-us-a-decade-of-change.
- Stifel Investment Banking. 2025. “Biopharma Market Update.” Healthcare Newsletter, April 2025. https://www.stifel.com/Newsletters/InvestmentBanking/BAL/Marketing/Healthcare/Biopharma_TimOpler/2025/BiopharmaMarketUpdate_TO_040725.html.
- Note that these firms contain all the hallmarks of a small biotech, such as seeking support from big pharma, being in the queue to go public, and having previously been dependent on VC investment.
- EY. 2025. EY Biotech Beyond Borders Report 2025: Focus on Fundamentals. https://studylib.net/doc/28795584/ey-beyond-borders-2025-report-final?utm.
- Brown, James R., Steven M. Fazzari, and Bruce C. Petersen. 2009. “Financing Innovation and Growth: Cash Flow, External Equity, and the 1990s R&D Boom.” Journal of Finance, 64 (1): 151–85. https://doi.org/10.1111/j.1540-6261.2008.01431.x.
- Czarnitzki, Dirk, and Marek Giebel. 2025. “Financial Constraints for R&D and Innovation: New Evidence from a Survey Experiment.” Applied Economics 57 (56): 9495–951.
- Congressional Budget Office. 2021. Research and Development in the Pharmaceutical Industry. Washington, DC: Congressional Budget Office. https://www.cbo.gov/publication/57126.
- Pecoraro, Brandon, et al. 2026. “Estimating the Present Value of R&D Tax Benefits in the United States.” NBER Working Paper 35208. Cambridge, MA: National Bureau of Economic Research.
- Agrawal, Ajay, Carlos Rosell, and Timothy Simcoe. 2020. “Tax Credits and Small Firm R&D Spending.” American Economic Journal: Economic Policy 12 (2): 1–21. https://doi.org/10.1257/pol.20140467.
The Brookings Institution is committed to quality, independence, and impact.
We are supported by a diverse array of funders. In line with our values and policies, each Brookings publication represents the sole views of its author(s).