Streetwise Professor

The Biden administration, courtesy of the delusionally titled Inflation Reduction Act, has made a huge spending commitment on alternative fuels, and in particular “clean” hydrogen, i.e., hydrogen not produced from fossil fuels (such as methane). Most of the “green” hydrogen stimulus involves supply-side subsidies (especially a $3/kg production tax credit, but also loans to be doled out by the administrative state). The Infrastructure Law sets aside funds for hydrogen electrolysis and hydrogen “hubs” (like that just announced for Germany). The administration is also attempting to make “the economic case for demand-side support,” such power purchase agreements (PPAs), contracts-for-differences (CFDs), advanced market commitments (made by whom?), and prizes (funded by whom?).

It’s hard to know where to begin in criticizing this mess. The biggest problem is that it attempts to address the climate issue (which I will take as a given, focusing on means not ends) by picking technologies. This almost never ends well. First, there is the knowledge problem–bureaucratic governments do not possess the information to make these technology choices. Second, there is the rent seeking/corruption problem–which exacerbates the knowledge problem, as interested parties exploit the ignorance of bureaucrats and funders, and their political connections, to induce investments based not on their economic virtues but instead on political influence.

There are also serious doubts about whether hydrogen qua hydrogen is the right alternative fuel given that it poses numerous problems and costs. The first is that using renewable energy to produce green hydrogen is extremely expensive. The second is that, well, hydrogen is highly explosive: I distinctly remember my 8th grade science teacher, Mr. Fisch, using electrolysis to fill a test tube with hydrogen, putting in a piece of chalk, then lighting a match to set off an explosion that sent the chalk flying across the room. You didn’t have Mr. Fisch as a teacher, but perhaps you’ve heard of the Hindenberg:

Explosiveness creates hazards, of course, and mitigation of them is expensive. Hydrogen is also extremely expensive to transport and store and requires a new and distinct transportation and storage system.

We are talking trillions of dollars to create “the hydrogen economy”–something even its boosters admit. Hell, they brag about it.

Hydrogen “carried” with carbon, in the form of ammonia or methanol, pose fewer problems (although ammonia in particular is nasty stuff). They are also costly, and it is clearly uncertain whether “green” forms of these hydrogen carriers are economical ways to reduce carbon emissions from fuels for transportation and power generation.

But the administration (and Europe too) have gone all in on hydrogen. Why? Maybe because their extreme antipathy towards carbon leads them to disdain fuels with any carbon in them.

Having chosen its technology, for better or more likely worse, now the administration is focused on how to force its adoption. The supply-side incentives are clear enough, so now there is a pivot to the demand-side, as expressed in the appallingly shoddy Council of Economic Advisors document linked above.

According to the CEA–and not just the CEA, as will be seen shortly–the problem is that “[r]eal or perceived risks around clean energy projects can raise the cost of accessing capital,  which could slow the rate at which projects like those in the hydrogen hubs program achieve commercialization..”

Well, I should hope so! That is, I should hope that risks are taken into account when allocating capital!

John Kerry flogged the risk issue on MSNBC (h/t Powerline):

“What’s preventing it is, to some degree, fear, uncertainty about the marketplace. People who manage very significant amounts of money have a fiduciary responsibility, an obligation to the people they manage it for not to lose the money, but to produce returns on that investment. Pension funds, many of them, are very careful about those investments in order to make certain they have the money to pay out to the pensioners who work for that money all their lives. So, there are tricky components of making sure that you have taken the risk away from these investments. And energy, which is what the climate crisis is all about, it’s about energy, it’s about how we fuel our homes, how we heat our homes, how we light our factories, how we drive and go from place to place.”

Damn those money managers for taking into account the risks and rewards of the money their investors entrust to them! Don’t they understand that John Effing Kerry knows what is right for humanity????? After all, he flies around the world in a private jet sharing his wisdom (and then dissembles about it before Congress).

I loved this part: “So, there are tricky components of making sure that you have taken the risk away from these investments.” Does John Kerry have a magic box into which he can make the risks disappear? Do tell!

Of course he doesn’t. What he means, clearly, is that the government must somehow absorb the risks inherent in the technology that they have already decided upon–apparently without analyzing those risks fully or carefully, or wondering whether maybe these damned investors might know something they don’t. (Of course they don’t wonder that! They are all knowing, right?)

At least the CEA attempts to put lipstick on the pig and raise some economic arguments to justify the need for demand-side support. There are market failures! Government never fails, but markets do, right?

In my experience the concept of market failure is most likely to be advanced when the market fails to do what someone thinks should be done, or wants to be done, based on their own vision. That is, when the market disagrees with someone, the market has failed! Especially when that someone is a member of what Thomas Sowell calls “The Anointed.”

The CEA basically cites to some theoretical possibilities. At the core of their argument is that learning by doing, including learning-by-doing that “spills over” among companies, can lead to inefficient investment. The CEA advances a couple of reasons.

One is a contracting failure. LBD–moving down the learning curve–reduces costs, meaning that prices are expected to fall. So, according the CEA, potential buyers are unwilling to enter into long term contracts for fear of agreeing to pay a price that will turn out to be too high: “if rapid declines in technology costs are expected, the willingness of private sector end-users to seek out such contracts with clean energy developers will be limited” (emphasis added). Without such contracts, hydrogen project developers can’t secure financing, so plants won’t get built, no learning takes place, and costs don’t fall. The Curly Equilibrium, in other words:

Really? If costs are expected to fall, market participants can enter contracts with de-escalator clauses, i.e., contractual prices that fall over time. Apparently the CEA only envisions contracts at a fixed price that extends through the life of the contract. But even then, given anticipated cost declines, the developer would be willing to sell at a price below the initial cost, basically, at the average cost expected over the life of the contract.

The CEA mentions the risks of of the magnitude of cost declines, but again, that should be a material consideration in any contracting and investment decision. Is the CEA arguing that the risk compensation demanded by borrowers will be excessive? They don’t say so explicitly, but that’s what you would need to argue that the prices in these contracts would be “too low” and thereby stymie investment.

I’d also note that indexed prices, widely used in a variety of commodity off-take agreements, eliminate the risk to buyers of locking in too high a price. They also address the asymmetric information problem that the CEA frets about. If the developer has better information about the likely trajectory of price declines, then yes, buyers looking at fixed price deals or deals with mechanical (non-market based) price de-escalators face a “winners’ curse” problem: the developer will agree to terms that overestimate his (better) forecast of future prices, and reject deals that underestimate.

I think in fact that the issue is that there is considerable uncertainty among all parties, developers and buyers alike, regarding what the future cost trajectory will look like. That is, there is a real risk here, and that risk should be taken into consideration when deciding whether hydrogen investments make sense. And market participants are far better at assessing the risks, and the pricing of those risks, than the government, which is clearly taking a “Damn the risks, full speed ahead!” Approach.

Sorry, but John Kerry et al don’t inspire confidence like Admiral Farragut at Mobile Bay.

One of the proposals under discussion is Contracts for Differences (“CFDs”) in which the government would (perhaps through a non-profit intermediary) provide a guaranteed revenue stream to a developer and absorb the price risk. To work, CFDs require indexing to some market price–and the market price for H2 hasn’t really been created. Further, they require some mechanism to set the guaranteed price, a non-trivial task given the very information asymmetries that the CEA worries about. The government-appointed third party (or the government for that matter) will certainly be the less informed party in any negotiations with developers, and will almost certainly overpay. (Not that they will mind–not their money!) Meaning that the asymmetric information problem the CEA frets about is present in spades in one of their preferred means of addressing it. Further, CFDs have already presented performance issues, with the sellers (those getting the guaranteed revenue stream) treating these contracts like options rather than forwards, and spurning their CFD commitments when market prices rise above the guaranteed price (as has happened with with generators in the UK when power prices spiked).

The CEA also invokes capital market imperfections also driven by asymmetric information that may impede financing if developers know more about the economics of projects than the financiers. This is a hoary old story that has been used to identify alleged market failures since time immemorial. So long ago, in fact, that when Stigler wrote “Imperfections in the Capital Market” (JPE) 56 years ago, he (in typical Stigler fashion) drolly started thus: “The adult economist, once the subject is called to his attention, will recall the frequency and variety of contexts in which he has encountered ‘imperfections-in-the-capital market.'” That is, “capital market imperfections” were an old joke decades ago.

Here’s another one, George! Based on long experience, George was a skeptic. Based on even longer experience, I am too, in this case in particular.

And let’s look at the empirical record. Learning by doing is a ubiquitous phenomenon. Dynamically declining costs in industries with potential information asymmetries abound. Yet industries have developed and thrived nonetheless.

Some examples.

I recently finished a piece describing extensive learning-by-doing in the shale industry, including evidence of learning spillovers and dynamic cost reductions. Yet, the shale sector has not faced problems getting capital or expanding rapidly. Hell, if anything, a common criticism is that shale drillers have obtained too much capital and drilled too much, not that they are starved for capital and drilled too little.

Does the CEA (or John Kerry!) believe the shale sector in the US is too small?

Insofar as spillovers is concerned, the fact that the costs of firm A decline when firm B produces more output is a necessary, but not a sufficient condition for an externality. One plausible outcome in oil (as identified in a paper on LBD in conventional drilling by Kellogg in the QJE) is that service firms are the ones that do the learning, and capture and internalize it.

LBD is well-documented for computer chips, which have seen relentless cost and price declines over the years. Yet computer chip factories have been built, and companies especially in the US and Asia have attracted the capital necessary to build these very expensive facilities and build new chip lines nonetheless. (In this industry too, there have been chronic complaints about overcapacity, rather than undercapacity. I am not commenting on the validity of those complaints, just noting that their existence contradicts the notion that dynamic scale economies and price declines due to LBD starve an industry of capital.)

The LNG industry has many of the characteristics that the CEA attributes to hydrogen. Yet this industry has expanded apace for well over 50 years now.

I viewed a presentation by DOE people today in which LNG was raised several times, and as an example not to be followed. DOE advisor Leslie Biddle (ex-Goldman) mentioned LNG several times (“I keep going back to the LNG analogy”), and in a negative way. LNG took 30 years to move to a traded market, dontcha know. And we don’t have that time! We need to create such a market in a year! (DOE’s Undersecretary for Infrastructure David Crane was more generous, giving us all of 5 years.) (Crane was also hyping the idea of hydrogen for everything, including home heating–apparently oblivious to the fact that even Net Zero fanatical Britain has just recently determined that H2 is too dangerous to heat homes.)

In the context of the discussion of a grand government plan to transform the energy system, I couldn’t help but think of Gosplan, or Stalin’s race to industrialization (e.g., the Magnitogorsk Steel Factory). We will inevitably–inevitably–meet the “Dizzy With Success” phase in hydrogen, mark my words.

I note that LNG production grew substantially before it became a traded market, which actually undercuts Biddle’s argument. Even though there was not a liquid traded market for LNG in the first decades of its growth and development, long term contracts, usually using crude (no pun intended) indexing features (like tying prices to Brent), contracts were agreed to, financing was obtained on the backs of these contracts, and liquefaction plants were built.

Oil refining faced many of the conditions that worries the CEA about hydrogen. Kerosene was a radical product early on, with a lot of uncertainty about market adoption. But Rockefeller dramatically expanded output and reduced costs: the cost of kerosene by 2/3rds in 10 years (1870-1880), in large part due to extensive learning and research on all aspects of the value chain. Standard Oil’s supposedly predatory acquisitions of were actually ways by which SO’s knowledge could be combined with physical assets to improve their efficiency.

The co-evolution of gasoline refining and the adoption of the automobile represents another example of investment and falling prices in a new market in a capital intensive industry.

I note that the early refining examples occurred when capital markets were far less developed than is currently the case. I further note that large energy firms (IOCs and NOCs like Aramco in particular) can potentially finance hydrogen (and other alternative energy projects) with cash flows generated by their legacy fossil fuel investments: this would largely eliminate any asymmetric information problem between developer and financier (because the developer is the financier) and developer and customer (because the developer could finance without securing a long term price commitment).

Another example. Electricity generation. Beginning with its inception in the early-1880s, electricity generation was highly technologically dynamic, with substantially declining costs. Yet in a few short years most urban areas in the US were electrified, with numerous private companies competing with government utilities. This was another industry in which overbuilding, rather than under-building, was widely discussed. The movement to price regulation occurred well after the industry developed, and was a reaction to intense price competition: regulation effectively cartelized electricity generation.

One more. Aircraft. LBD was first identified in the production of airframes. This phenomenon was first documented by Wright in 1936, and was subsequently observed in myriad other industries (e.g., Liberty Ship construction in WWII). LBD and the associated cost declines have continued in aircraft construction ever since. And aircraft have been built and aircraft manufacturers have been able to attract the capital to design and build new aircraft that benefit from these cost declines.

In the face of all these examples, the CEA and others making these market failure arguments should identify an industry that died aborning due to the alleged chicken-or-egg problem that makes demand side support of hydrogen investment necessary.

The CEA document has echoes of some rather common, but unpersuasive, arguments for government support of industry, such as the infant industry argument and the big push development literature. The latter has been demolished by practical experience: the list of its dismal failures is far too long. There are more than echoes of this discredited approach in the CEA document. It links to a paper that credulously recycles the old, bad, discredited theories.

What is amazing about the infant industry argument is how often it is invoked, and how little empirical evidence supports it. One of the few empirical papers, that of Krueger and Tuncer, rejects the argument in the case of Turkey.

A paper by Juhasz is often touted to support the theory. It shows that after the stimulus of the cotton spinning industry in France due to Napoleon’s Continental system, post-1815 the industry was competitive with the British, indicating that it had moved down the learning curve. Again, at most this identifies a necessary condition for protection–learning–but not a sufficient one. Even if LBD occurs, and even if there are spillovers, the cost of protection may exceed the benefits. A simple story demonstrates this. If the protected industry achieves cost parity with the first-mover (e.g., the UK in cotton), the protected firms merely displace firms in the first-mover country, leaving post-parity total costs unchanged. So in equilibrium, protection is costly but generates no benefits.

All in all, the CEA document reminds me of a rather conventional undergraduate econ paper, repeating textbook wisdom about externalities and market failures. It completely ignores the Coasean insight that market contracting methods are far more sophisticated than those in the textbooks, and that market participants have incentives to find clever ways to contract around what would be market failures if market transactions were limited to the forms considered in textbooks. It also ignores the historical record.

In other words, rather than writing off the difficulties of securing “bankable” contracts to secure funding for H2 developments to “market failures” or the excessive risk aversion of market participants, the government should step back and consider whether this alleged hesitation reflects a more sober and informed evaluation of risks than our betters in DC have undertaken.

I crack myself up sometimes.

In sum, the administration’s entire approach to hydrogen is utterly flawed. It attempts to pick technologies based on a pretense of knowledge it does not possess. It views flashing red lights warning of risks as signals to be suppressed rather than considered when making policy and investment choices. It engages in simplistic analyses of how real markets work, and how they have worked historically, to conclude that market failures requiring government intervention to fix abound in hydrogen.

All of these government failures could be eliminated by cutting the Gordion Knot, pricing carbon, and letting markets and private enterprise develop the technologies, products, contracting practices, and market mechanisms to trade off efficiently the benefits of reducing CO2 emissions. Decentralized mechanisms discover and utilize information, including information about new technologies, far more efficiently than governments. Decentralized mechanisms incentivize learning and innovation–including contracting and organizational innovations that can be instrumental in developing and adopting new technologies, products, and techniques.

In the case of hydrogen, pure or “contaminated” with carbon, priced carbon would address the problems that the CEA frets about, in particular the contracting problem. A carbon price would make it straightforward to index prices in contracts. A formula related to NG prices (because blue hydrogen is likely to drive the price of hydrogen at the margin, and because methane is likely to be the substitute at the margin for H2 in many applications) and the cost of carbon would send the appropriate signals and eliminate the need to fix prices in advance.

What the price of carbon should be and how it should be determined is a whole other question. But it would be far more productive, and not just in regards to hydrogen, to focus on that problem rather than leaving it to the John Kerrys of the world to pick technologies and then devise the coercive mechanisms necessary to force the adoption of those technologies.

Alas, we are on the latter path. And it will not take us to a good place. Probably figuratively, and perhaps literally, to the fate of the Hindenberg.