Loan Guarantee

Those of us who have been railing on the government's increasing push to make massive loan guarantees available to individual energy firms are not surprised to see the first major bankruptcy.  Solyndra went down with $535 million in federal guarantees for lots of reasons.  The marketplace is increasingly competitive.  Power prices have fallen due to recession and fracking-induced reductions in the cost of natural gas.  China subsidizes its solar production, artificially manipulates exchange rates, and there is a growing supply overhang as PV subsidies in Europe get cut.  New technologies look to be better than Solyndra's.  John Hudson in The Atlantic's blog provides a nice overview of the political hemming and hawing as the parties try to frame the issue to protect their interests. 

Though all of these reasons are plausible contributors to the demise of Solyndra, they are hardly surprise risks.  Operational problems, competitors, changes in market conditions:  all are baseline risks for any type of new investment that a venture capital firm might take on. 

The difference isn't the nature of the risks, but rather the nature of the investor -- the federal government, straying far from its domain of demonstrated competance.  This is the core issue I see in the Solyndra situation, but one that seems to be getting overlooked.  The Title 17 loan guarantee program, in all of its various forms including a swath of more recent even larger plans to have a federal "Clean Energy Bank", is poorly structured to achieve success.  It puts a handful of largely invisible bureaucrats and advisors in the position of making unprecedented wealth transfers to fund high risk private ventures.  That last sentence alone should have been enough to doom the program by both the Bush administration (that passed it) and the Obama administration (which has been approving most of the lending commitments).

Yet there are other program weaknesses as well, compounding the barriers to success.  The risk assessments are not subject to much disclosure or vetting on the outside, and the payments from funding recipients (in the form of credit subsidy assessments on borrowers) are widely recognized to be well below the market value of the credit guarantees.  Thus, the riskier the venture, the better the federal guarantees look to the firm, dramatically increasing the adverse selection risks of the resulting portfolio.  Yes, I know -- there is government review of the deals to alleviate those risks.  But this is done by people with only their goodwill to rest upon; they have no personal financial risk bound up with making a good selection, may not always have the appropriate technical expertise, and have limited time.  And because the financial benefits from getting one of these loans are so high, the government review team is arrayed against the best talent money can buy.  There may be conflicts of interest as well; it is really quite hard for the public to know at this point as many of the key decision makers and reviewers are not named.

The structure of this program has many red flags indicating a high risk of failure and corruption.  Other than scoring political points, it hardly matters that a solar firm is the first one to go down.  The Vogtle nuclear reactor loan guarantee is roughly 16x larger, and was done at terms based on market structure prior to Fukushima.  Anybody want to bet on the default risks for that one as the operating requirements for reactors get ratcheted up in response, and natural gas plants continue to undercut everybody's pricing? 

Back in 2007 DOE was taking comments on their notice or proposed rulemaking (NOPR) for Title 17.  I put in a number of comments on program structure, which, as far as I can tell, were entirely ignored by DOE.  The Department focused instead on the comments from the big investment banks that would be financing these deals.  Lehman Brothers and Merrill Lynch, neither of which exist any longer, comprised a two of the six financial institutions that banded together to submit joint comments on how the multi-billion dollar loan guarantees ought to be structured to work for the banks.  Citi, Goldman Sachs, and Morgan Stanley, all of which went through dramatic restructuring and public bailouts, made up most of the rest.  Their comments resulted in even more generous terms for borrowers, and weakened checks and balances for taxpayers.  Below are some excerpts from the comments I submitted.

The NOPR did not go far enough in outlining how DOE would ensure non-political allocation of resources and protection of taxpayer capital. Many of the ways these issues were addressed in the current version were mostly descriptive in nature, leaving too much guesswork about how actual implementation and institutional oversight will proceed.

...

A combination of wide latitude in determining project eligibility to participate in funding rounds with imprecision on how performance "improvement" will be measured create the conditions for skewed and politicized distribution of billions of dollars in guarantees. These conditions bode poorly for the long-term success of this program.

...

DOE acknowledges wide latitude in targeting loan guarantees and acknowledges they are under no obligation to run open contests across all energy sources authorized under Title XVII of the Energy Policy Act. The NOPR is mostly silent on the establishment of rigorous project comparison metrics and robust institutions that would ensure the billions in guarantees are effectively targeted.

Not all energy resources eligible under the Energy Policy Act of 2005 (EPACT) are necessarily eligible in particular -- or indeed any -- funding rounds. In fact, DOE's NOPR lists specific programmatic objectives, loan guarantee authority or available funds as possible criteria by which guarantees under Title XVII can be awarded. The NOPR includes the example (p.7) of the Administration's 2008 budget that proposes $4 billion in guarantees for centralized power, $4 billion for biofuels and other clean fuels, and $1 billion for new electric transmission or renewable energy power systems. This example is an early indication of the power that the legislative or executive branches will seek to exert in earmarking funding to favored interest groups. The political influence leveraged by these groups is greatly enhanced by DOE's view that Congressional appropriations are needed to support DOE's credit authority under Title XVII, and that Congress can structure these appropriations however it sees fit. While in theory the executive and legislative branches could choose to exert their influence based on the technical merit of the projects alone, assuming they will in fact do so would be both imprudent and naïve.

The risks of political influence are further compounded by the fact that funding under Title XVII will be in the form of loan guarantees, for which valuation and transparency are far more difficult to attain than with direct payments. In addition, recent mandates to publish legislative earmarks (albeit only partially effective thus far even for direct payments) would not seem to apply at all to Title XVII loan guarantee decisions.

By the time the larger lending proposals under CEDA started to surface in 2009, I was a bit more blunt in my concerns about corruption:

CEDA risk profile likely to be far more concentrated than conventional banks, increasing potential problems with systemic risks and corruption. The multi-billion dollar scale of centralized energy technologies suggest the size of the credit commitments for individual projects under CEDA will be far larger than undertaken in other areas of federal credit  guarantees. This opens the initiative to higher systemic risks and potential corruption.

It will be interesting to see what surfaces in the FBI review of documents.  It's a pity we can't learn about potential political influence on the nuclear deals at the same time, but Congressional sources tell me it is not about to happen.  At least the Solyndra failure can serve as a wake-up call on the incentive problems for these mega-lending initiatives, and help us avoid far worse financial blow-ups were we to have followed industry's lead in ramping up lending to nuclear or establishing a large scale "green energy bank" such as CEDA.

UPDATE, September 19th.  Michael Grunwald at Time Magazine has a nice overview of the problems with shining a light only into a small part of a big swamp.  Focusing on Louisana Senator David Vitter, Grunwald documents how Vitter's bill to increase scrutiny only of loan guarantees to renewable energy projects side-steps Vitter's own efforts to bring home the loan guarantee pork to his own constituents -- including for renewable energy projects.  We've got systemic problems here, and piece-meal solutions generally just shift the problem rather than solve it.

This is the first of what will hopefully be a series of posts looking at specific aspects of The American Power Act (APA) proposed by Senators John Kerry and Joe Lieberman. 

This one deals with the nuclear “loan guarantee retention fee” (section 1102), as an item I’d flagged an needing additional information in my initial review of  the Bill’s summary. 

More resolution on other nuclear subsidies from detailed bill language

Before delving in to that specific item, the detailed APA brought a few other nuclear-related subsidies into greater focus than could be discerned from the summary alone.  Clearly, Kerry and Lieberman have some broader strategy behind their strong desire to shift more and more of the investment risk for an ever larger number of new reactors from investors to taxpayers.  Maybe it is pure politics (MA has MIT, a strong backer of nuclear; CT is the headquarters of General Electric); maybe it is some downpayment on a future climate deal; maybe they really do believe that nuclear power should be the anointed winner of the post-carbon energy world, without even having to compete. 

Whatever it is, however, the risk-reward balance is worrying (with taxpayers on the hook for so much of the investment risk, K-L don’t even give us any of the upside), as are the potential impacts on inter-fuel competition.

• Nuclear delay insurance subsidy expansion is larger than was evident in the summary.  This is because the fixed number of contracts role to new facilities if no claims made.  Rather than "just" tripling the face value of the policies from $2 billion to $6 billion, and eligible reactors from 4 to 12, K-L also took allows all 12 available policies to be reused and reassigned to new projects if claims are not made on the inital plant.  This change greatly increases the likelihood of the government incurring losses under the program; and practically means that coverage will be available to any reactor that wants it during the period of delay insurance operation.

• Near-current deduction of investment tax credits by counting them as “progress payments.”  The bill summary made it clear that K-L thought production tax credits for new nukes weren’t enough, and that they wanted investment tax credits as well.  ITCs in general dramatically shift performance risks from plant owners to taxpayers relative to production tax credits (which can’t be earned until a plant is actually producing).  The progress payments allow more rapid deductions on partially complete investments. 
• Increased nuclear production tax credits (section 1124) – national cap is increased by 33%, from 6,000 MW to 8,000 MW.

A detailed look at nuclear loan guarantee retention fees

Including a retention fee is a de facto admission by K-L that federal loan guarantees for nukes will be among the cheapest money in the project.  I don't see this admission as a bad thing.  Quite to the contrary:  in any 12-step program, recognition is step 1.  Only by acknowledging the enormous value of the guarantees can K-L or others in Congress more directly assess their impact on the competitive dynamics of the energy industry.

For too long, the entire focus on the loan guarantees has been on whether or not they will default, not on the enormous distortions they will cause in energy investment patterns even if they don’t.  Look at the industry’s own cost models (which do not predict default and bankruptcy) and the loan guarantees top the list for most valuable “incentives”.  NEI also holds the guarantees at the top of their long and growing wish list of support.  This subsidy is what I refer to as the intermediation value of government loan guarantees – basically, the difference between what a true market transaction would price the risk at and the US Treasury’s cost of borrowing.

Loan guarantees generate subsidies even without default in two main ways.  First, they reduce the cost of debt substantially.  The higher the market risk of your project, the larger the cost savings from replacing “debt investor” with “taxpayer” as your funding source.  Second, because title 17 of the Energy Policy Act now allows 100% of the debt to be federally guaranteed (up to 80% of the total project cost), nuclear investors get to use far more debt than any true merchant structure would have ever allowed.  Debt is cheaper than equity, so the weighted average cost of capital on the new reactor goes down further.

This creates some uncomfortable facts for K-L, however.  The first is that less risky projects – even if they also get federal guarantees – will end up relatively worse off vis-à-vis nuclear than without the guarantees.  If these other projects are also less politically powerful, they may end up far worse off – an unfortunate outcome if they also happen to be quicker, more scalable, less expensive, and lower risk to the taxpayer-investors than a new reactor.  The second issue is that if the feds are the cheapest source of funding, there’s no reason for the nuclear borrowers to pay it off any sooner than they have to (potentially 30 years out). 

The retention fee attempts to deal with this second issue.  It partially calls industry's bluff that they need financing only during the construction phase -- though even with the retention fee, low cost credit lasts well into the plant's operating life -- and possibly for the full duration of the loan.

Retention fee mechanics

Assuming a 4 year construction period (admittedly ambitious), you have the following schedule of retention fees (longer construction times would delay the year in which the retention fee kicks in) that escalate the cost of borrowing over time:

• Years 1-4: no retention fee due to construction period
• Years 5-9: no retention fee due to 5 year grace period
• Year 10: 0.5% fee added to whatever base rate the borrower is paying
• Year 11: 1.0% fee added (escalating 0.5%/year through year 14)
• Year 12: 1.5% fee added
• Year 13: 2.0% fee added
• Year 14: 2.5% fee added
• Year 15: Rate jumps to 5.0% above original note interest rate; this occurs no more than 10 years after operations begin, and stays there until the normal expiration of the note.

Will the retention fee accelerate repayment of government guaranteed debt?

For the borrower to want to pay off the federal debt earlier, the cost of that funding must rise higher than alternative sources of debt.  This is possible – if, for example, the reactor comes in on-time and below budget, and interest rates stay low, the jump to a 5% retention fee might trigger refinancing in year 15 instead of year 30. 

But there are also many scenarios where the refinancing is unlikely:

• Default.  The risk of default is highest during the construction phase.  With no fee until at least year 9, the majority of defaults will kick in prior to any impact from the retention fee.

• Financial stress.  If the reactor is under financial stress due to construction delays, cost over-runs, or an expectation that break-even power rates will not be achievable when the plant goes live, there will likely be no capacity to repay any of the debt lines early. If there is, the government debtor is likely to be more easily lobbyed for relaxed repayment terms than is a bank.  This would especially be true in a situation of systematic financial stress across the nuclear sector, as would occur following a reactor accident somewhere in the world; or from structural changes in electricity markets that alter the economic benefits of large scale, centralized power generation.

• Federal credit is still cheaper.  Even with the retention fee premium at 5%, it is quite possible that the alternative sources of credit will be more expensive.  This outcome could arise in two situations.  If the market situation is not great for nuclear, industry promises about take-out financing once the plant is built may not materialize inexpensively.  Second, replacing the Federal Financing Bank with a private lender, even if the initial interest rate seems lower, could well be far more expensive if the 80% debt structure of the project needs to be ramped down to 40 or 50% to meet private lender debt covenants. 

Unknowns

There are also a variety of unknowns in how the sketchy language on the retention fees would be deployed.  These include:

• Impact on existing $18.5 billion nuclear pool.  Though K-L want to amend the title 17 statutory language, it seems unlikely that they could retroactively institute this fee on guarantee applicants for the first $18.5 billion.  What is unknown is whether there would be any restrictions in modifying the terms of a program to which $122 billion in applications have already been submitted under the old rules.
  
• Retention fees versus credit subsidy and administrative fees.  While it is rational to assume that retention fees are in addition to, rather than in partial lieu of, credit subsidy and administrative fees, this is not explicit in the proposed language.  In the high stakes world of nuclear lobbying, nothing should be taken as a given.
  
• Cash cow for DOE?  Under title 17, the interest and principal goes directly back to Treasury.  Fees, on the other hand, appear to go to Treasury but remain available to DOE for subsequent use.  This is a very interesting potential wrinkle.  Consider that were this this interpretation is correct, the retention fees would create a return of up to $5 billion per year (5% of $100 billion in total authorizations) that would flow to DOE rather than to the general taxpayer.  If we further assume that DOE pools the money to support additional guarantees, this might create a revolving-fund like structure similar to what the CEDA legislation would have done, without the need to pass CEDA.  Any lawyers out there who can evaluate whether this type of an outcome would be possible?