This piece originally appeared in American Affairs.
Since late 2021, the Biden administration has set aside more than $1.6 trillion in infrastructure spending. Almost three years later, it’s very hard to know where most of that money has gone.
The Biden administration’s infrastructure laws are built upon two distinct and competing philosophies of industrial policy. The first, demonstrated by the chips and Science Act, works mostly through grants: funding awards are given directly to specific manufacturers by the Department of Commerce, with explicit terms around domestic production, intellectual property sharing, and reinvestment of profits. This has its benefits: the spending is easy to track, is highly targeted, and carries little risk of filtering out to foreign adversaries. No money may be allocated to foreign entities of concern, and award recipients may not engage in any “significant transactions” related to the expansion of advanced semiconductor manufacturing in countries of concern for a ten‑year period.
But there are downsides to this approach. Direct government spending requires administrative capacity: the federal, state, or local agency responsible for carrying out the program must solicit applications, review them, and make decisions about whom to award money. It also requires organization on the part of private sector applicants, who often lack the resources and know-how to put together grant applications in the first place, or even to gain awareness of relevant programs’ existence. The result is that grants tend to roll out slowly. At the time of this writing, almost two years since the passage of chips, less than $1 billion of the $52 billion appropriated for semiconductor funding has been formally awarded. (Despite the widespread headlines about awards being made, most of the Biden administration’s announcements to date have not actually been awards but rather “non-binding preliminary memoranda of terms” meant to serve as negotiation instruments.) Similarly, the $5 billion National Electric Vehicle Infrastructure program, passed through the Bipartisan Infrastructure Law in 2021, has led to exactly eight charging stations being built as of February 2024, while hundreds of other years-old federal spending programs have yet to make a single award.
The Biden administration’s second philosophy of industrial policy is exemplified by the Inflation Reduction Act (IRA). The IRA takes a climate-first approach, prioritizing emissions reductions over China competition, with the goal of deploying as much clean energy infrastructure as quickly as possible. Accordingly, the law works through “as-of-right” tax credits: companies that meet certain eligibility requirements can claim the credits, incentivizing clean energy investment while creating relatively little bureaucratic friction. The upshot is that IRA funding has reached the clean energy industry much more quickly than chips funding has reached semiconductor firms.
But this approach has its drawbacks as well. Whereas grant programs lack speed, tax credits lack transparency. Because as-of-right tax credits are facilitated through the tax code, they are protected by Internal Revenue Service confidentiality. As a result, no one—not the public, nor watchdog groups, nor even most members of Congress—knows exactly how much we’re spending, when we’re spending it, and on whom we’re spending it. Simply put, the implementation of a several-hundred-billion‑dollar bill is almost entirely withheld from the public eye.
Addressing America’s genuine industrial policy needs will require rethinking the mechanisms we use to pursue our policy goals, as the growing reliance on tax credit programs leaves much to be desired. The use of tax credits has made the administration of U.S. industrial policy smoother, but it has also made it very difficult to know what the policy is accomplishing. That tradeoff has been accepted by many as a necessary evil, but very little thinking has been done regarding whether such a lack of transparency is actually required for a successful infrastructure program. There is ample opportunity to improve the structure of tax credits—and there are also many cases in which other funding mechanisms offer a better way forward. If this administration and its successors are to continue investing trillions in American industry, they will need to ask not only what to fund, but also how to fund it.
How the IRA Works
The details of how the IRA works help illustrate the problems with “industrial policy by tax credit.” Five of the twenty-two credit programs under the Inflation Reduction Act are for individual taxpayers, while the remaining seventeen are aimed at businesses and other entities. The latter category establishes new credits, such as the clean hydrogen credit, and extends several others, such as the clean energy production credits. All of these differ slightly in structure—some are for production, others are for investment, and several more are for research, vehicle purchases, and energy efficiency.
They have an important commonality, however: nearly all of the credits under the IRA are uncapped. That is, there is no limit on the amount of funding available for the federal government to dole out in credit value each year. It is for this reason that, while the original budget estimate for the IRA was in the realm of $369 billion, a recent report from the Congressional Budget Office projects that the real cost will be closer to $820 billion—more than double the original estimated price. The spending under the law is therefore not just hard to track; it’s also theoretically unlimited.
Another key feature of the IRA is its relatively few restrictions on foreign entities of concern. According to an analysis from the Wall Street Journal, as of early 2024, China-based companies were responsible for nearly a quarter of new U.S. domestic solar capacity announced since the law’s passage. This has led many Republican members of Congress to argue that the IRA, despite its ostensibly domestic focus, represents a handout to the Chinese Communist Party. Indeed, many IRA-funded companies will source their raw materials from Chinese manufacturers and their components from Chinese suppliers. Less appreciated, however, is the fact that when Chinese firms inevitably profit from these American tax dollars, it will be nearly impossible to prove it.
The IRA credits come at a time of increasingly aggressive Chinese industrial competition. In spite of a growing number of restrictions on federal funding for purchases of foreign goods, Chinese companies have made explicit, and often successful, efforts to access U.S. government grants. BYD USA, an American subsidiary of the Chinese electric vehicle manufacturer BYD, has successfully secured funding through the Bipartisan Infrastructure Law’s Clean School Bus program. The Chinese battery manufacturing behemoth CATL, meanwhile, has partnered with Ford to secure federal incentives for Ford’s Michigan-based battery plant.
Elsewhere, U.S. reliance on Chinese manufacturing has led to the rollback of the very provisions that were meant to increase American manufacturing independence. The $7.5 billion National Electric Vehicle Infrastructure (NEVI) program, for example, was passed alongside the Bipartisan Infrastructure Law’s “Build America, Buy America” requirement, which mandated that the majority of the cost of all EV charger components come from components manufactured in the United States. But within less than a year of the law’s passage, the EV charger manufacturing industry requested a waiver from the requirement, citing supply-chain constraints and limited domestic production capacity. The Department of Transportation obliged. Eight months later, Autel Energy, the U.S. subsidiary of the Chinese company Autel Intelligent Technology Corp., started production on its NEVI-compliant electric vehicle chargers at its plant in Greensboro, North Carolina.
Most of these examples are from grant programs, spending which is designed to be much harder for Chinese firms to access compared to IRA tax credits. So while Chinese companies have been able to secure U.S. federal funding in spite of rules around foreign entities of concern, most IRA credits don’t have those rules in the first place. In short, even in highly regulated, broadly transparent government spending programs, U.S. taxpayer money is going to Chinese companies. What must it look like, then, when it comes to lightly regulated, unaccountable tax credit schemes? Answering this question requires first understanding just how these tax credit programs are designed.
Transferability
There are two main tax credit structures: allocated credits and as-of-right credits. Allocated credits are made available through a competitive process. Typically, these credits are limited in quantity, and entities must apply to a government agency to receive them. Allocated credits are significantly less popular than as-of-right credits, as they involve many of the same administrative processes as grants—applications, awarding agencies, and all of the attendant bureaucratic friction. Agencies’ involvement in the allocation process, however, means that there is often a way to access details about allocated credit recipients. For example, the Low Income Housing Tax Credit program—perhaps the best-known allocated credit program—has a comprehensive database of its property and tenant data.
In contrast, as-of-right tax credit programs, which make up the majority of those in Biden’s infrastructure laws, are inherently obscured by the tax code. With the as-of-right structure, eligible companies can claim credits on their tax returns without needing to apply or be selected by a government agency. While this approach eliminates administrative hurdles, it also protects recipient information, as the credits are claimed through the confidential tax filing process. This lack of transparency has been a consistent feature of as-of-right credit programs, from the Energy Policy Act of 2005 to the American Recovery and Reinvestment Act of 2009.
But the Inflation Reduction Act goes a step further, adding an additional layer of complexity that sets it apart from the federal tax credit programs of the past. New in the IRA is a funding mechanism known as transferability, under which companies generating credits can sell them directly to unrelated third parties in exchange for cash, tax free.
Consider a real-world example: Ashtrom Renewable Energy, a renewable energy developer, is constructing a roughly 400-megawatt solar project in Texas called Tierra Bonita. Through this project, Ashtrom expects to generate around $300 million in production tax credits (PTCs) over ten years. In the past, Ashtrom would have needed $300 million in tax liability over that decade to fully reap the value of the credits. Alternatively, they could have partnered with tax equity investors to monetize the credits. But with transferability, Ashtrom can now simply sell the $300 million in PTCs to an unrelated company with sufficient tax liability to absorb the credits. In October 2023, Ashtrom announced that they had signed a tax credit transfer agreement to sell the Tierra Bonita PTCs to an unnamed institutional buyer, described only as a “highly rated U.S. institutional entity.” Ashtrom will get cash upfront, and the buyer will get to offset its taxes over ten years.
Proponents argue that the transferability mechanism will significantly accelerate clean energy deployment, and will offer a simpler, more cost-effective approach. By allowing developers to sell tax credits for cash, transferability greatly simplifies the monetization process. Historically, clean energy developers relied on complex and costly tax equity financing structures, forming partnerships with large banks and investors to trade tax credits for upfront investment. These partnerships involved hefty legal and accounting fees that could eat up a sizeable portion of the tax credit’s value. Now, however, developers can sell credits to any taxpayer without getting entangled in partnership structures and associated expenses. This reduces transaction costs and complexity, and theoretically expands the pool of potential investors, increasing competition among buyers and improving access to financing for smaller clean energy projects.
But at the same time, transferability adds yet another layer of obscurity to government spending, as demonstrated by the Ashtrom sale. There’s no way for the public to see who is buying credits, at what price, and from whom, unless voluntarily disclosed.
Here, too, there are opportunities for foreign competitors to benefit. The IRA’s New Clean Vehicle Credit comes with a foreign entity of concern provision that aims to limit the use of Chinese products. The Advanced Manufacturing Production Credit may be subject to a similar restriction in the final IRS rule. But these provisions cover only the first layer of eligibility, not the secondary market, meaning that while foreign entities and their subsidiaries may not directly qualify for credits, nothing prevents them from buying credits on the transfer market from a broker. A Chinese state-owned enterprise could acquire U.S. clean energy tax credits, and the public would never know.
There’s every reason to think that the secondary market will be huge. In the first six months following the release of the IRS guidance around transferability, $7–9 billion worth of tax credits changed hands, according to one report. With the IRA providing hundreds of billions of dollars in credits over the next decade, the market has substantial room to grow. If even a fraction of those credits are transferred, annual trading volumes could easily reach the tens of billions.
How often is the transferability provision being used? While there is no perfect analogue to the IRA’s federal transferability scheme, there is a long history of transferability at the state level in the form of film tax credits. With these credits, states offer incentives to film production companies in hopes that Hollywood will bring economic benefits to the state. Recently, the nonprofit Good Jobs First sent in public records requests to assess the rate at which film tax credits are being transferred under Illinois’s transferability program. It found that from 2014 to 2021, a whopping 98 percent of the $540 million in credits awarded were being transferred each year. Of those, 60 percent were purchased by just five companies: U.S. Bank, Walmart, Comcast, Verizon, and Bank of America.
This is not to say that film companies and clean energy firms will necessarily transfer credits at similar rates. But it does show that when transferability is an option, it gets used at scale. The little data available on IRA transfers seem to confirm this. According to the Department of the Treasury, as of mid-April, “more than 900 entities have requested approximately 59,000 [tax] registration numbers for projects or facilities located across all 50 states plus territories. Approximately 97% of these projects are pursuing transferability.”
It’s almost certain, then, that the secondary market will continue to metastasize. Tax equity investors appear to be syndicating portions of their credits, and dedicated brokers and insurers are emerging to facilitate trades. These middlemen are seeking to maximize their own returns, of course, and their cut ultimately reduces the amount of public money reaching actual projects. There’s also the simple fact that if tax credit recipients pursue transferability, then the credit is not worth its full dollar value to the seller. The upshot is that the IRA credits are being transferred at an average of about 93 cents per dollar of credit value—in other words, at a 7 percent loss in credit efficiency.
Some will argue that the “haircuts” from transfers are ultimately lower than they would be in a more conventional project financing scenario. Indeed, the cost of monetizing credits through the tax equity market can be as high as 15 percent of the credits’ value. But there’s also a more basic problem with transferability: the Inflation Reduction Act is a climate bill, and yet its benefits are ending up with companies that have nothing to do with the climate.
Reports from intermediaries suggest that the institutions that have historically been involved in tax credit deals are the same ones buying up IRA credits today: banks, insurance companies, and other major corporate taxpayers. This is not to mention the benefits to the many syndicators, tax advisers, and insurers that have emerged in the wake of the IRA’s passage, all of whom are skimming cents off the top of every dollar of credit value.
It is for this reason that finance executives such as J.P. Morgan CEO Jamie Dimon have suddenly taken a keen interest in niche areas of climate policy—for example, invoking eminent domain to build more renewable energy projects. J.P. Morgan and Bank of America are responsible for half of the country’s tax equity finance deals every year, and the introduction of transferability promises to bring even more value to their bottom line. In late 2023, renewable energy developer Arevon Energy announced that J.P. Morgan had committed to purchasing $191 million of investment and production tax credits for their solar-plus-storage project in California, celebrating the deal as “among the nation’s first transactions . . . that leverage the Inflation Reduction Act’s transferability provision.” As the secondary market for tax credits continues to grow, so too will the profits for the financial institutions that broker the deals or redeem the credits themselves.
Refundability
If it were a simple case of choosing between the conventional tax equity financing approach and the IRA’s transferability mechanism, the latter would offer certain advantages in terms of simplicity and cost-effectiveness. But this is a false dichotomy. Not only is there another option on the table; that option is already available for three of the credits under the Inflation Reduction Act.
The Advanced Manufacturing, Carbon Oxide Sequestration, and Clean Hydrogen tax credits all allow for a mechanism known as direct pay. For a five-year period, these credits are effectively refundable: the IRS will issue a direct cash payment to the entity claiming the credit equal to the credit’s face value. What’s more, entities exempt from income tax—nonprofits, state and local governments, rural electric cooperatives, and others—qualify for direct pay for twelve of the IRA credits, including the investment and production credits. In this way, direct pay allows organizations with little or no tax liability to receive the full value of the IRA credits.
Direct pay thus achieves the efficiency benefits of working through the tax code, while also enabling entities to receive the full value of the credit directly from the government as if it were a grant, cutting out the middlemen and haircuts involved in transferability. Moreover, direct pay has the advantage of far greater transparency compared to the opaque secondary markets that emerge with transferable credits.
So why not just use direct pay for all the IRA tax credits? After all, there is clearly demand for the refundable option: per a report from the Congressional Research Service, direct pay will account for an estimated 40–48 percent of the total costs of the three credits for which it is available over the next decade.
Skeptics have offered a few reasons. One common argument against expanding direct pay is that it transforms tax credits into something more akin to a spending program. Critics note that this shift could make the tax credits less politically palatable, as direct government spending is often scrutinized more heavily than tax reductions. But in budgetary terms, there is little difference between a refundable tax credit and a grant program. The reluctance stems more from optics than substantive distinctions.
Senate staffers also pointed out to me that direct pay increases the explicit cost of the credits for the government compared to transferability, where costs are more hidden. This, however, is actually a point in favor of direct pay from a transparency perspective, as it allows for a more honest accounting of the full fiscal impact of these subsidies. The government should be fully accountable for its expenditures regardless of whether they come on the front end through direct spending or on the back end through tax credits.
Ultimately, the preference for transferability over refundability seems more about political framing than sound policy considerations. Tax credits simply don’t feel the same as direct spending to many policymakers, even when they have the same economic impact. But given the substantial downsides of transferability in terms of transparency and potential for abuse, this preference ought to be reconsidered.
Social Policy and the Tax Code
The Inflation Reduction Act is far from the first instance of the federal government using the tax code as a tool to achieve policy objectives. The last several decades have seen Congress increasingly rely on tax credits to incentivize any number of desired economic and social outcomes—what the sociologist Joshua McCabe dubs “the fiscalization of social policy.”
The use of tax credits as a policy tool began in earnest in the post–World War II era and accelerated significantly in the 1970s. The Energy Tax Act of 1978, passed under President Jimmy Carter, created tax credits to encourage energy conservation and the development of alternative energy sources in response to the energy crisis of that decade. President Ronald Reagan later signed into law the Economic Recovery Tax Act of 1981, which included several business tax credits aimed at stimulating investment and economic growth.
The use of tax credits continued to expand in the following decades. The Energy Policy Act of 1992 under President George H. W. Bush created new credits for renewable energy. The Energy Policy Act of 2005 under President George W. Bush further extended credits for energy efficiency and investment. In the wake of the 2008–9 financial crisis, the American Recovery and Reinvestment Act established a variety of tax credits aimed at stimulating the economy, including firm-specific credits similar to those in the IRA. The Patient Protection and Affordable Care Act of 2010 created new tax credits to subsidize health insurance coverage, while the Consolidated Appropriations Act of 2021 under President Trump included extensions and expansions of several existing credits.
The IRA represents a culmination of these trends, with its extensive use of uncapped, transferable, and in some cases refundable tax credits to spur the development and deployment of clean energy technologies. Thus the problems with the IRA’s credit provisions—their lack of transparency, the potential for abuse in secondary markets, the efficiency questions around transferability—shouldn’t be seen as isolated concerns. “We now have the blueprint,” said Sunnova Energy CFO Robert Lane of the IRA’s transferability mechanism during a recent earnings call, “that we intend to keep leveraging.”
The tax code has seen its purpose continuously expanded over the last several decades. At some point, policymakers might take a step back and ask whether the tax mechanisms that have become popular in that time are actually the best levers for industrial policy.
What To Do
At a base level, if tax credits are to be the key funding mechanism in future infrastructure bills, then they ought to be restructured to better align efficiency and fiscal responsibility imperatives. To enhance transparency about the use of tax credits, Congress should require all recipients to submit a simple annual report to the Treasury providing basic information detailing the value of the credits they’ve received. This report should confirm the credits received, state their type and amount, briefly describe how they were used, and disclose any changes to foreign affiliations. Critically, this reporting requirement should be included in a legislative vehicle like the infrastructure bill itself to ensure it has the force of law. The reported information, excluding any sensitive business details, should then be made publicly available in a searchable database to enable public oversight.
Mandatory disclosure of foreign affiliations will be crucial for providing transparency about potential foreign influences. In relevant industrial policy programs, tax credit recipients should be required to disclose their foreign affiliations at two points: during the initial eligibility certification process and in the annual report for credit recipients. The disclosure should include information about ownership stakes held by foreign entities, strategic partnerships or joint ventures with foreign entities, and significant contracts or financial relationships with foreign entities. For credits that target at-risk industries critical to national security, policymakers should consider placing restrictions on foreign entities of concern. A bill to this end has already been introduced: the No Oppressive Giveaways of Taxpayers’ Income to Oppressive Nations (no gotion) Act. Named for the American subsidiary of the Chinese battery parts manufacturer, the bill would disqualify “any entity created or organized in, or controlled (in the aggregate) by, one or more countries of concern” from clean energy tax credits.
Policymakers should also establish rules around the transferability of credits to foreign-owned entities, particularly those with ties to China. This could involve requiring credit buyers to certify that they are not owned or controlled by foreign entities of concern, or placing outright prohibitions on the transfer of credits to such entities. This too involves a tradeoff between efficiency and national security considerations, so the nature of these restrictions should depend on the importance of the industry that the credit aims to support.
All entities that redeem transferred credits should be required to report essential information to the IRS, which should then make that information available in a public registry. This information should include the identity of the redeemer, the value of the credit being transferred, and the specific credit involved. This reporting requirement should be designed to minimize the compliance burden on market participants, with clear guidance and a simple, standardized process provided by the IRS.
As an alternative to transferability, expanding direct pay options for tax credits within bills like the IRA could significantly enhance their accessibility and impact. By allowing entities, particularly start-ups and small businesses in the energy sector, to receive the full value of the credits as direct payments from the government, this measure would enable them to benefit from the incentives even if they have limited tax liabilities, without the mission creep and transparency concerns associated with transferability.
In the absence of new legislation, there already exists one limited option for enhancing transparency: a provision of the tax code allowing certain congressional committees to request and obtain tax return information. As outlined in Section 6103 of the Internal Revenue Code, the chairs of the Senate Finance Committee, House Ways and Means Committee, and Joint Committee on Taxation can request and receive tax return information from the IRS. Though these committees cannot share taxpayer-specific information with the public (or, except in rare cases, with other members of Congress), they can request details on the use of tax credits by foreign-affiliated companies and release aggregate data to the public.
The goal in all of these policies is to create a framework that ensures transparency, accountability, and strategic alignment while still preserving the core administrative efficiency of tax credits. Striking the right balance will require ongoing refinement, but the above reforms would represent a promising and necessary start.
These are also reforms that could attract broad bipartisan support. Republicans and Democrats may differ on the specifics of industrial policy, but members of both parties have advocated for greater transparency and oversight in how public funds are deployed. Strengthening reporting requirements, limiting transferability to foreign entities, and expanding direct pay are all measures that could advance that shared goal.
Moving Beyond Tax Credits
Fixing tax credits’ transparency and accountability problems should be a near-term priority. But these problems should also prompt a broader reconsideration of how we do industrial policy in the United States. Simply put, there are many scenarios where grants—the main alternative to tax credits—would be a more appropriate mechanism for industrial policy spending.
Grants have been a staple of U.S. industrial policy for decades, from the Advanced Research Project Agency’s funding for energy R&D to the Department of Defense’s defense industrial base grant program. And while grants often involve more bureaucratic processes compared to tax credits, they offer several advantages.
First and foremost, grants provide greater up-front control and ability to target funds strategically. As one former Treasury official pointed out to me, when the government is trying to shape supply chains and make market allocation decisions based on nonmarket factors like national security, grants allow for more due diligence and selection of recipients aligned with those strategic goals. Policymakers can target specific industries, technologies, or geographic areas, and put in place guardrails to prevent funds from benefiting foreign competitors.
Grants also offer enhanced transparency and accountability compared to tax credits. While tax credit data is shrouded in IRS confidentiality, grant recipient information is typically public. This enables congressional oversight and public scrutiny to deter potential misuse of funds. In my conversations with Senate staffers, they emphasized that this transparency is a key reason why grants were preferred in the chips Act, where preventing funds from flowing to adversaries like China was a top priority.
Third, grant agreements provide an opportunity to set specific conditions and requirements for recipients. Agencies can specify the scope of work, set milestones and reporting requirements, and include clauses to suspend or reclaim funds if conditions aren’t met. This allows for much more tailoring and control compared to the relatively blunt instrument of tax credits.
For all of these reasons, grants are particularly well-suited for advancing national security priorities and shoring up strategic industries. On the other hand, there may be instances where the national security risks are so acute that the imperative for rapid action trumps other considerations. If a critical industry is on the brink of collapse, or a key capability is about to be lost, a two-year grant-making process may simply be too slow. Tax credits, for all their flaws, can inject resources more quickly than their more deliberative alternatives.
This raises another, more fundamental question: what actually constitutes a national security concern justifying industrial policy intervention in the first place? There’s a clear consensus that semiconductors are vital to both economic and national security, given their centrality to everything from consumer electronics to military systems. But what about less obvious cases?
The critical minerals supply chain, for instance, may not present the same acute risks as semiconductors, but it underpins a vast array of industries and technologies. Overreliance on foreign suppliers, particularly adversaries such as China, could create serious vulnerabilities down the line.
If Republicans are going to articulate a vision for industrial policy—one that is distinct from the “everything-bagel liberalism” offered up by the Left—they will need to be clear about what that policy is trying to achieve. Is the goal solely to shore up industries with direct and immediate bearing on national security? Is it to address market failures in areas with significant public goods aspects, such as R&D or environmental externalities? Or is it something broader—a sort of everything-bagel conservatism?
The scope of that vision will necessarily inform the choice of policy instruments. A more targeted approach focused squarely on acute security risks may prioritize grants for their control and precision. A more expansive agenda may rely more heavily on the scale and flexibility of tax credits. By fixing the flaws in our tax credit regime, we can take a step toward a more mature industrial policy. And by doing so, we can begin to build a foundation for the more far-reaching reforms that will be necessary in the years ahead.