Policy commentary

Loan Guarantees Reconsidered

Oct 3, 2011 | Joel Darmstadter, Joshua Linn

Loan guarantee programs—such as those run by the Department of Energy (DOE) to help bring innovative technologies to market and stimulate the economy—have received greater scrutiny after the recent bankruptcy filing by Solyndra, a DOE-backed company. Are loan guarantees the best option for promoting new technologies, green industry, and job growth?

Early in September, Solyndra, which manufactured a new type of solar photovoltaic module, declared bankruptcy and about 1,100 workers lost their jobs. The company had been awarded a $535 million loan guarantee from the Department of Energy’s (DOE) Loan Programs Office (LPO), and many politicians and industry advocates had previously touted Solyndra as a positive example of the “green economy.” DOE’s press release announcing the award in September 2009 predicted that Solyndra would “accelerate job creation and … introduce into large-scale commercial operation a new and highly innovative process for manufacturing a breakthrough design for photovoltaic panels.”

Solyndra’s collapse and the resulting cost to taxpayers have generated a flurry of media coverage. While some of that coverage delved into issues related to loan guarantees in general or Solyndra in particular, most of the media attention has focused rather narrowly on the loan approval process, largely ignoring the bigger policy questions. As the quote above reflects, loan guarantee programs such as DOE’s have rather broad aims: encouraging commercialization of brand new technology, but also promoting U.S. manufacturing of green technologies and stimulating the stagnant economy. Are loan guarantee programs a good way to achieve these objectives, and how can this approach be improved?

Three Loan Guarantee Programs in One Office

The basic workings of a loan guarantee are fairly simple. A private company borrows money for a project, and the government agrees to repay all or most of the loan if the project fails and the company has to default. By reducing the risk to the lender, the guarantee reduces the cost to the company of obtaining the loan. Effectively, the government is subsidizing the cost of the project.

DOE’s LPO oversees three programs that have overlapping objectives. The first program aims to encourage the commercialization of significantly new or improved technologies that reduce carbon dioxide (CO2) emissions. The second program aims to create jobs in the current weak economy and includes, in particular, projects that begin construction before the end of 2011. The first two programs include a broad range of technologies primarily aimed at reducing carbon emissions, such as alternative fuels, solar, and carbon sequestration. The third program specifically focuses on vehicle technologies. All told, the programs have guaranteed almost $40 billion in loans for 42 projects. The expected federal liability depends on the projects’ successes, and will almost surely be less than $40 billion.

Common Rationales for Loan Guarantees

Three primary arguments are commonly given in support of these multiple objectives:

Bridging the “valley of death.” Investment in new technologies tends to be highest in the earliest stages (think venture capital or public funding of basic research and development) and in the latest stages (private funding after commercialization); in between there is a “valley” in which new technology has trouble attracting private investment. For a variety of reasons, private investors may lack sufficient incentives to pursue technologies that have not yet been commercialized. For example, investors may not want to back a startup company because there is too much uncertainty about the quality of its product or because the time horizon required for payback is too long.

Providing immediate economic stimulus. Because of poor and, presumably, transitory economic conditions like the current economic downturn, many businesses may fail that would otherwise succeed in a stronger economy; loan guarantees would help support such businesses. Although this argument would apply to businesses in many sectors of the economy, supporting the green economy provides the additional bonus of improving the environment.

Supporting green industry in the long term. Many supporters of loan guarantees argue that either because of future carbon policy or depletion of exhaustible resources, the United States will have to transition to low-carbon technologies sometime in the future. U.S. citizens would be better off developing and manufacturing these technologies rather than importing them from China and elsewhere.

Some Important Economic Considerations of Loan Guarantees

Directly evaluating these arguments would require a good deal of space. Here, we make several points about the challenges of using loan guarantees to meet the stated objectives, particularly regarding the commercialization of new technology.

Risk. Investing in new technology is inherently risky—moving from the pilot project stage to commercialization is extremely difficult, and there is always uncertainty over finding sufficient customers. That doesn’t necessarily mean that the projects shouldn’t receive subsidies, but it does mean that loan guarantees for new technologies should be recognized as risky. Projects may fail because technologies turn out not to perform as expected, because demand drops unexpectedly, or for any number of reasons. If few of the projects fail, that may mean that the projects weren’t risky enough.

Learning-by-doing and market failures. Loan guarantees and other investment subsidies are often justified based on a learning-by-doing argument. The subtlety is that learning-by-doing does not necessarily justify a subsidy. For example, suppose an installer of solar panels learns as it installs more panels how to install them at lower cost. Anticipating this, the installer should reduce the price of its panels so that it can sell more and learn more—the installer should, in effect, invest in learning. But suppose there are learning spillovers, and another installer observes what the first installer does, copies it, and proceeds to compete with the first installer (we’ll assume that no patents are violated). In that case, the first installer does not profit from its own learning. Basic economic theory says that the first installer will invest less in learning than is socially optimal, and the government should offer subsidies. But if there are no spillovers and the first installer profits from its learning, subsidies are not justified.

Information. In a successful loan guarantee program, the government supports projects that private investors have decided not to back, but which nonetheless are beneficial to society. In other words, the government needs to figure out why the private market isn’t financing the project. Is it because the project is a bad idea? Or because the social benefits of the project outweigh the private benefits? For example, the solar installer may not have attracted private investors because of low consumer demand or because learning spillovers deprive the firm of a proprietary asset—that is, the value of information.

Designing the program to elicit information. The LPO addresses this information problem by conducting a detailed review of each proposed project. But, compared with other types of policies, this places a large burden on the government. Alternatively, an innovation or commercialization prize can be offered to a project that successfully meets certain criteria. Or, a production subsidy can be offered for each unit of output. In both cases, the subsidy can depend on the social value of the project, in which case the government no longer has to identify the best projects; private investors can consider the value of the subsidy and any additional profits they may earn from selling the resulting products—and then choose accordingly.

Is a Loan Guarantee Program the Best Policy for the Job?

One detail that sets apart loan guarantee programs from many other policies, such as the federal investment or production subsidies for renewable electricity generators, is that loan guarantees directly promote the commercialization of new technologies. Even so, it is not clear why loan guarantees are the best form of investment subsidy, much less the best policy for meeting the broader objectives of promoting new technology, green industry, and employment. It is particularly challenging for the government to acquire enough information to identify the projects that a) have the greatest social value and b) the private sector would not invest in otherwise. Identifying learning spillovers is very difficult, as is gathering sufficient information to evaluate a project’s risk and societal benefit. As noted above, other policies, such as production subsidies targeted at new technologies, may provide better incentives for commercialization.

We offer two concluding comments. First, loan guarantee programs should be designed to enable ex post evaluation of their effectiveness. The programs are advocated for correcting market failures related to the commercialization of new green technologies, and it should be possible to assess the extent to which the programs achieve that goal.

Second, some observers, in both political and business circles, have sought to explain the Solyndra debacle in terms of unfair price-cutting by Chinese firms. In fact, although Chinese trade and investment practices pose significant threats to competing U.S. technology firms, those practices are largely separable from problems associated with the purpose and shortcomings of federal loan guarantee programs, as discussed here. For policies such as China's subsidies, export controls, and domestic-content requirements, one had better seek redress through the World Trade Organization, the provisions of which, at least to some degree, China seems to be violating.