Guest Post: A Closer Look at Industrial Policy

This guest post is by occasional contributor Ilya Podolyako, a third-year student at the Yale Law School and an executive editor of the Yale Journal on Regulation.

In my last post, I compared Obama’s plan for the automakers to that of the Chinese government. I concluded that the two shared goals, but that these goals fit poorly into the traditional American ethos of free enterprise. For better or worse, the administration will have to remedy this mismatch sometime soon, either by letting the automakers fail or by openly stating that jobs, national pride, and a green fleet justify a government-backed industry.

Some of the comments to that piece strongly favored the latter course of action. Accordingly, I think it is worthwhile to take a normative look at directed industrial policy generally. In the abstract sense, this system is the opposite of laissez-faire capitalism – the government owns enterprises and dictates both the nature and quantity of their output. This description naturally evokes images of the USSR, Cuba, or the People’s Republic of China (before 1992), all countries with “command” economies. The distinctive characteristic of these nations, however, was not the presence of state-owned enterprises (SOEs) per se, but the absence of a legal, noticeable private sector.

Countries like Sweden, South Korea (before 1988), and, indeed, the modern China have demonstrated that directed industrial policy is not an all-or-nothing game.

The state can own some firms while the private sector owns others; these groups can compete in the same industry or service different sectors of the economy. Examples of the middle-ground approach also abound in developed economies: in Norway, the government owns over 60% percent of Statoil, the country’s largest employer (now part of Statoil Hydro); in France, the government owns over 80% of the giant power company EDF; Germany, France, and Spain effectively control the management of the defense contractor EADS, despite the fact that 41% of the company’s stock is publicly held; until 2006, the Republic of Ireland held 85% of the shares in Aer Lingus, the country’s largest airline, a share that has since fallen to 25%; and Canada’s government owns dozens of so-called Crown Corporations. Of course, these statistics have lost much of their panache now that the US Treasury owns 80% of AIG, but the figures still show that (intentionally) state-owned industry is not altogether unusual.

That is not to say that SOEs are entirely benevolent. Over the years, economists have identified several key flaws. First, the absence of the profit motive, which often runs along with the guarantee of lifetime employment, removes incentives for employees to work hard. Second, the people involved in managing these conglomerates may retain their profit motive and act on it, even if the companies officially do not, leading to widespread corruption. Third, a company that decides what to produce based on official decree, not market demand information, will have an extremely difficult time determining the “right” amount of output. If said firm makes widgets, it will make either too many or too few widgets, creating waste in the first instance and shortages (assuming that the government regulates price) in the second.

A government committed to firing lazy workers, who could then fend for themselves in the remaining, robust private sector, would seem to be able to solve the first problem with SOEs. A robust criminal code that punishes corruption, when coupled with politically ambitious prosecutors, would probably mitigate the second. The third flaw, however, appears immutable, not merely because of societal momentum or even the laws of human nature, but because mathematically, pricing information appears to be the only way to measure efficient output.

Where do the above observations leave us? Conventional wisdom argues strongly against government control of industry if this control leads to an outcome different than what the market would otherwise dictate (passive taxpayer ownership poses a somewhat different set of problems). Yet the last 24 months have forced everyone, including old-guard, free-market purists in charge of the Fed to reexamine conventional wisdom. If the central bank can selectively lend to industries that appear to be failing in the open market, one would think that Treasury or the Department of Commerce could also selectively push certain industries to generate products that they believe would improve the national economy.

Such intervention could come in one of two forms. The government could try to dictate business strategy to firms that receive a substantial amount of bailout money, but as I explained before, this would amount to stacking an inefficient bureaucracy on top of a underperforming business. Alternatively, the President and Congress could fund brand-new entities that would lay high-speed rail, construct nuclear power plants, and build a modern national electric grid. These companies would be staffed with regular employees and managers who would have to answer directly to someone high up in the executive branch. They would also compete alongside their private sector peers, who could choose to provide variations on the basic product (like UPS or FedEx) or attempt to displace the basic product altogether (like private charter school operators).

Normally, this setup would raise concerns about “crowding out” private-sector capital or innovation, but this theory applies equally to indirect government spending, not just direct industrial participation. True, private firms may be reluctant to compete with organizations backed by “infinite” treasury pockets willing to absorb a loss, but this fear cant get much worse than it is now. Moreover, Congress could fix this problem by specifying early on that these new enterprises will have to shut down after posting 3 years’ worth of consecutive losses.

Here’s the crux: current strategies for economic recovery focus on stabilizing financial institutions and melting credit markets, but they fail to provide an internal engine of growth for the American economy. Theoretically, this growth should continue to come from private-sector innovation, but the private sector is in deep and worsening trouble.   According to the latest Fed report, consumers continue to be saddled with roughly as much revolving (non-mortgage, non-auto) debt as they were in Q2 of 2008. This debt load would have to decrease by 17% to get us back to 2004 levels, when an average household owed around $6500 on their credit cards. At the current rate of revolving debt change, the adjustment would take at least two years, but this calculation does not account for the decrease in individuals’ personal income, which is falling at around 1.5-3% a year. A sharp decline in domestic production means that the US will also continue to be a net importer. These characteristics are part of the reason why the country got to where it is in the first place. Without a new strategy, it is unlikely that the ingenuity of a few individual entrepreneurs will be sufficient to pull us out of this depression.

Of course, the government shouldn’t pick its areas of economic involvement arbitrarily. A smart policy would be to attack problems that the private sector has traditionally been unable to address (producing cars doesn’t fall into this category). The new enterprises should first engage projects that suffer from free-rider and coordination issues which only a central authority can overcome – public works and infrastructure. No one wants to build the first mile of a road to a store if the guy at the other end can get its benefits for free. Next, the government should invest heavily in pharmaceutical, engineering, and biotech research, both at universities and newly invigorated national labs. Since anyone can build on fresh advances, basic science carries many of the same unpleasant investment characteristics as physical public goods.

The above programs aren’t that far from the status quo. Congress already chartered Fannie and Freddie; the Pentagon already subsidizes an enormous defense sector. It is time, however, to try and change federal involvement in industry from covert efforts to promote opaque policy goals to thought-out measures designed to help firms and workers prosper. In the worst-case scenario, we privatize our way back to the present.

B y Ilya Podolyako

22 thoughts on “Guest Post: A Closer Look at Industrial Policy

  1. I’m not sure how this idea of prospective government-directed industry is different from the massive corporatist social engineering we’ve often had in this country.

    To give two examples, it was the federal government’s railroad policy and BuRec dam-building and irrigation infrastructure for the benefit of nascent agribiz which subsidized the development of the West (one of the most vicious of American myths is that of the self-reliant pioneer West).

    And, the government chose to structure modern society around the car and suburbia, instead of trying to revitalize mass-transit oriented cities.

    I could add more – financialization, globalization, consumerism, the exponential debt economy.

    (Then we have government programs like private health insurance or the ethanol “industry” which are purely parasitic and rent-seeking, which add literally nothing but complexity, cost, and in the case of monocrop ethanol mass hunger for the poor.)

    All of these were policies of choice, engineered collaboratively by big government and big corporations. There was nothing inevitable or natural about any of them.

    That is not to say that SOEs are entirely benevolent. Over the years, economists have identified several key flaws. First, the absence of the profit motive, which often runs along with the guarantee of lifetime employment, removes incentives for employees to work hard. Second, the people involved in managing these conglomerates may retain their profit motive and act on it, even if the companies officially do not, leading to widespread corruption. Third, a company that decides what to produce based on official decree, not market demand information, will have an extremely difficult time determining the “right” amount of output. If said firm makes widgets, it will make either too many or too few widgets, creating waste in the first instance and shortages (assuming that the government regulates price) in the second.

    It seems the first two are rampant in the private sector as well. It’s evident that even “real” industries like the auto industry (let alone completely fake ones like bank holding companies) are filled with time-servers who do no quality work, while the “work” that is done at a place like AIG or these big banks often is nothing more than sociopathic personal profiteering at the expense of everything else including even the company.

    As for the third, people should read Dmitri Orlov on how the vaunted inefficiencies of the Soviet system were very helpful to the people when the system collapsed. Everything was so ponderous that inertia kept food, jobs, goods flowing for a long time, while everyone could stay in his apartment.

    Orlov wonders, when the American system collapses, where are its just-in-time distribution, Hobbesian employment, housing and health systems, and the completely shredded safety net going to leave us? We’re going to be a lot worse off than the Russians were.

    So maybe that third alleged bug is really a potential feature.

  2. Some specific responses:

    1) Efficient from the private perspective, and efficient from a social perspective, are two different things. Hence the problems with rent-seeking behavior and negative externalities. The faith-based economists who argued that rent-seeking never causes systemic issues seem to have had a come-to-Jesus moment lately.

    1) The “immutable” problem of choosing what/how much to produce (that government seems so unable to address) conjures up images of Stalin and his Cinnamon stockpile…

    “Apparently Chairman Stalin had ordered them to stockpile cinnamon after a shortage in the ’50s, and nobody ever countermanded the order, so they had quite a lot,”

    Funny, true. But this “immutable problem” has some easy solutions. The issue you are referencing is identical to the difference between managing trade through quotas vs. tariffs. For any given level of quotas, there’s a tariff level that would achieve it… but sometimes it’s easier to set a price than to set a quantity. (aka, when you know the social cost of an activitiy, set a tax equivalent to the social cost). Sometimes it’s easier to set a target quantity and let the (artificially created) market clear (aka, the sulfur-pollution market for coal).

    In other words, between “free-market” and “command-economy” is that practical place where most of the world resides… “managed markets”.

    Certainly Team Obama is aware of the range of economic instruments they have at their disposal, although they may be overly inclined to believe their toy models even when real-world problems break them. (For example, the failure of radio spectrum auctions to extract a fair price in non-liquid markets.) I think Team Obama has had a pretty quick education, and is getting real practical really fast – with the possible exception of a handful of cleptocratic individuals who should be smothered in honey and left out for the ants. (Disclaimer: Do not take that seriously – this in no way constitutes a threat, or suggestion of a threat, toward any elected or appointed official. Nor is it intended to incite violent action or to encourage the incitement of violent action.)

    3) Within a managed market, it is relatively easy to (mostly) align individual incentives with social incentives by maintaining <50% govt. ownership, holding non-voting stock, and making sure that executive compensation is directly linked to LONG TERM PERFORMANCE. (At the very least, you can do much better than the shareholders of GM have managed to do.)

    Sector level government choices regarding capital allocation (e.g. partial non-managing ownership of a firm) are several steps removed from “government management”.

    This is called having a real industrial policy (like we used to have, in the magically distant past under Eisenhower and the other Fairy Kings, before the Dark Times came).

    4) Finally, it is important to separate policies that work in a healthy market economy (govt.-backed loans, carbon taxes, subsidies to positive-externality activities, healthy regulation) from policies that work when the world is on the precipice of a deflationary debt-destroying depression (government pumping money/demand into the system, and trying to make sure that the demand it is creating will have some long-term social productivity).

    Team Obama – full of smart, geeky people – came in all ready to do the former. They had outgrown Keynesianism and Out-Thunk the Friedmanomicons. They were ready to start deploying “smart” policies, which would have worked 6 years ago.

    Then, just as they were grasping the helm of the ship, a hurricane blew up, and all those “smart” policies went overboard. Now they’re in crisis management, and crisis management does not always require smart and subtle. Sometimes it requires Big and Stupid.

  3. This is a software problem, not a hardware problem.

    American society, for whatever historical reason, believes there is a clear distinction between the public and private sectors. As Russ and StatsGuy point out, it isn’t true – Cigna and Lockheed exist in a government universe – but we behave that way.

    Suppose, as a minor thought experiment, we decided to pay the head of the FDIC $1mm/yr. Can you imagine the outrage? Well, she is supervising businesses worth billions whose managers each make many millions. We can contract out to PIMCO and let PIMCO pay its managers market wages, but we cannot internalize the function.

    If we went the SOE route, we the people would be completely unable to run the businesses in a “professional” manner. We would want to treat them as “government,” with all the symbols we hold dear. The fact that, say, Singapore is able to be a responsible financial shareholder in its companies would have no bearing on us; when the US taxpayer has equity, he makes his presence known.

    To some extent, this is why the government has taken such a weak stance with the financials – it is afraid of what it would be forced to do if it confronted the reality that we have put up the majority of the capital in the sector today.

  4. SOE’s have typically failed. Let’s learn from the experience in India. After independence in 1947, India embarked on five year plans. India decided to adopt hybrid model for industrialization. The grand plan was to have both privately owned enterprises and state owned enterprises. The notion was that the SOEs will hold private enterprises in check by bring products to market at lower price than profit mongering private enterprises. Everything looked good on paper.

    In reality, the SOEs have been bastion of inefficiency and attracted incompetent, lazy bureaucrats who lacked professional management expertise and skills. Besides, the SOEs did not have incentive to make profit and hence did not care about the cost structure either. In India these SOEs exist even today. They employ hundreds of thousands of employees (for example. Indian Railways). In essence, the SOEs are welfare programs in disguise because nobody ever gets fired from these organizations for poor performance.

    The SOEs also turned out to be big time bonanza for the private enterprises. Because the cost structure was higher at SOEs, the price of products set by SOEs is always higher than what private enterprise would set. So, the margins enjoyed by private enterprise for the same class of products (for example, steel, fertilizer, petrochemicals, pharmaceuticals etc) is always much higher than the SOEs.

    Another lesson learned in India was that it is extremely difficult to privatize the SOEs. India has not been able to privatize Air India, Life Insurance Corporation and many other SOEs. On the other hand, the private enterprise has taken off and is thriving. Thanks to SOEs that they have set the benchmark for poor quality and high prices.

    The U.S. government can bail out the auto industry short term but finally the auto industry has to survive on its own.

  5. Some canonical research findings, then a simple inference:

    Stanford economist Paul Romer, who originated New Growth Theory, which updates growth economics for the information age:

    “Perhaps the most important ideas of all are…ideas about how to support the production and transmission of other ideas…North Americans invented the modern research university…As national markets for talent and education merge into unified global markets,
    opportunities for important policy innovation will surely emerge…There are…safe predictions. First, the country that takes the lead in the twenty-first century will be the one that implements an innovation that more effectively supports the production of new
    ideas in the private sector.”

    From a 2008 book co-authored by Clayton Christensen, a Harvard Business School professor who originated the canonical Model of Disruptive Innovation:

    “In the last three decades, increasing numbers of cognitive psychologists and neuroscientists…have produced a multitude of schemes to explain…that people learn differently from one another.

    …Students need customized pathways and paces to learn.”

    “We state above that in the first phase of the disruption of the instructional system the software will likely be complicated and expensive to build. The reasons for this can be traced to the use of
    the existing commercial system when marketing the system, as noted previously, as well as to the relative immaturity of Web 2.0 software. Within a few more years, however, two factors that were absent in stage 1 that are critical to the emergence of stage 2 will have fallen into place. The first will be platforms that facilitate the creation of user-generated content. The second will be the
    emergence of a user network, whose analogues in other industries would be eBay [i.e., an online market is a type of user network]…

    …The data suggest that by 2019, about 50 percent of high school courses will be delivered online.

    …80 percent of courses taken in 2024 will have been taught online.”

    The simple inference: the federal government should establish a venture capital (VC) firm to invest in companies (e.g., start-ups) that (will) provide:

    (1) an online market for customized education
    (2) complementary offerings (e.g., student loans, online markets that provide new and improved ways to showcase and earn money from expertise)

    The VC firm can be similar in form to In-Q-Tel, a VC firm that is funded solely by the CIA.

    Lots of related details are available at


  6. I would think of industrial policy as some kind of a middle ground in between laissez-faire and command-and-control. It seems like plenty of countries have as an explicit and significant government policy to support the development of high-value industries. Whether that means providing the necessary infrastructure (power, water, communications, suitable lands, etc.), or direct economic subsidies, I would think that there’s are plenty of very active things that could be done in that effort short of government ownership. Why is that not considered a possibility?

  7. Name any major software/technology/computer company today, and it can be traced back to government programs in the 60’s and 70’s. It’s doubtful Microsoft, Intel and Google would be the giants they are today if government-sponsored scientists (with the DARPA/ARPA agencies) didn’t have the freedom to muck around for a decade or two.

    The government is the ultimate angel investor. They are the only entity that can absorb years and years of losses or zero return on investment. But over decades what has been the rate of return on those investments? Best money the government has ever spent.

    Fast forward to today. While all the humbug is about the goverment owning yesterday’s businesses, what about the future? If the government is going to grow their way out of this problem, then they better get busy on building the businesses of tomorrow. The governments dividend is future tax reciepts on future private enterprises.

    Clearly, the way to jump-start real growth in this economy to solve the energy problem. When we are focused on big goals, unforseen side benefits happen that increase future productivity even more. When we went to the moon, we developed the microchip. If I had my chance to be the dictator of the U.S., I would tell the American people – “Look we’re pumping all our extra money into solving the energy problem. We’re not going to bailout car companys or banks. We’re going to leverage the greatest university system in the world and we’re going hard-core on solving the energy problem. Everybody will have to sacrifice for a few years but I promise you in 10 years we’ll be better off. If this doesn’t happen you can hang me.” err..or something like that.

  8. lots of good ideas here in these posts but they are isolated bits and pieces – the question is how do we bring it all together so that economic system, fiancial system, political system are permanently configured so as to contribute positively to the social system as a whole – that is what is called for – change as in a total revolution is what is required – nothing less will do

  9. Stats,
    You know what scares me to death? I actually understand what you mean….and that scares me.

    PS: I agree.

  10. BUT, there is a massive debt load left in it’s wake, which has future consequences. Worth it? Time will tell…

  11. I do not see why we should follow Indian Rail (bloated and low-quality as per prior poster) and not EADS (not bloated and very well competing with Boeing) as an example of SOE. For that matter, I don’t think there is a single railway worldwide that could match the inefficiencies and miserable performance of our supposedly independent Amtrak.

  12. There is some (gulp) Keynesian merit to having government spend during the down cycle and build the nationwide infrastructure that businesses can build on during the next recovery. Interstate highways, hydroelectric, etc. Even, if need be, to provide busy work to keep enough people fed, clothed, and not beating plowshares into swords.

    When it is thought that “[a] robust criminal code that punishes corruption, when coupled with politically ambitious prosecutors, would probably mitigate the second” –well, we already have this system, and look how far we are with the private sector which theoretically should be easier to root out.

    But this is the very thing we desperately need our government to do: ITS job, not ours.

    At this point though, it is (aptly for this post) probably all academic.

  13. One should not confuse State Owned Enterprise and State Managed Enterprise. These are two different concepts. Unfortunately the article completely identifies one with the other.

    Instead of always taking Stalin or India as comparable, the USA should be compared to a republic with a long and similar history that is fully comparable, namely France.

    France has had many SOEs, but is well aware that the government can’t manage enterprises and should not, a fortiori, micro manage demand. The article, unfortunately views both as unavoidable if SOEs are created.

    They are not unavoidable. SOEs in France are managed by top professionals, and get top results, and freely compete in the market. Renault, EDF, Areva, EADS are example. The later three are viewed as strategic, and so they are. They are nearly as important as the Defense Department itself. Without mastery of energy (EDF), nuclear (Areva) and aerospace (EADS), the country would be defenseless.

    Renault, though, has had to compete, for decades with the completely private Peugeot (second European car maker in volume), and many private car parts makers. In the end the state has a golden share, the rest is traded. By the way, the reverse can happen: state created enterprises such as Total (created in the 1930s) has now zero percent state ownership (it’s the world’s fourth largest oil giant).

    Taking India as an example is also unfortunate: its state sector is well known to not have performed well, and a lot of the recent economic ascendency in India comes from letting the free market be free.

    In the USA most large companies are just illusions of privacy and market. They tend to have both the defects of hyper private capitalism and hyper state capitalism.

    Patrice Ayme

  14. Let’s start with Mr. Podolyako’s crucially important practical statement “current strategies for economic recovery focus on stabilizing institutions and melting credit markets, but fail to provide an internal engine of growth for the American economy”. Amen!, and yet only the financial sector has been the main thrust of Simon Johnson’s valiant attempt to improve the Obama Administration’s recovery program – he seems to be saying “let’s fix the FINANCIAL SECTOR and all will be well”! But what about our INDUSTRIAL SECTOR, the “engine of growth” that has permitted mounting balance of trade deficits for more than 30 consecutive years while piling up over than $10 trillion of foreign debt.
    Theoretical discussions about whether an “industrial Policy” is good or bad are useful, but at the bottom line we must face the necessity of earning our way in a very competitive world. Both government and multi-national businesses must recognize this “fact of economic life” and find a way to deal with it by adopting effective industrial and trade policies, just as other nations do. There can be no more “free lunch” for America!
    Kenneth N. Davis, Jr.
    Former U.S. Assistant Secretary of Commerce/International

  15. Just as households need to distinguish between good debt and bad debt, so does the government. In the past 10 years, we squandered our low-interest rate credit on new spending that has had (so far) a negative rate of return.

    One of the best ways to help defend against a long term real-interest rate rise is to demonstrate to investors that the U.S. Government spending is focused on future growth, rather than transfer payments that fullfill little or negative economic purpose.

  16. Yes, innovation is key to competitive leadership, and disruptive innovation has shown great promise. I expect the U.S. to have even more challenge to its innovation leadership in the years ahead. Even so, we can and must adopt business strategies and government policies that recognize we cant go on borrowing abroad to sustain our high standard of living.We especially must discard the idea that it’s O.K. to transfer production of our most advanced technologies abroad, and thereby accelerate innovation in compteting nations (notably Japan and China). We don’t need to become “protectionist” against foreign-made goods, but we do need to safeguard our key technologies.
    Best, Ken Davis

  17. Ken,

    Thanks for your reply. For the U.S., one of the virtues of pursuing policies that leverage the insights of Christensen and Romer is that a leadership role in the global market for customized education can lead to much-increased exports, and correspondingly less borrowing. Better still, achieving said leadership role makes knowledge transfer a less problematic issue for the U.S., as any loss of competitive advantage anywhere feeds demand for education.

    Anything about the above that is particularly objectionable? :-)


  18. I’m getting that you have a big idea in customized education based on highly innovative software.If your initials are DT, I note that your Silicon Valley marketing company has excellent VC backing, so you know what it takes to interest them. Why not get some startup or early stage V.C.’s to back the first of your new breed of education companies? I’ve been raising capital for early stage tech companies for 20 years, and I don’t see why a new government-backed V.C. category is needed here. E-mail me at if you’d like a real dialog. Very best regards,
    Ken Davis

  19. An example of government directed and funded industry is the defense industry. The goverment directs what’s needed, private industry competes for the orders. The two industrial sectors in which the US is a world class competitor are aerospace and information technology, both of which have are and have always been heavily funded and directed by the military. Other sectors of the economy, notably the environmental business are the product of government regulation. To redirect the auto industry to build more fuel efficient cars the federal govt could use a combination of the approaches it has used with the military and environmental businesses, for instance requiring tighter standards and funding the development of electric cars. There doesn’t seem to me to be anything mysterious or unprecedented here.

  20. Russ: Thank you for pointing out how historically-inaccurate and over-simplified this free-market revisionism is.

    The example of the development of the West and of the auto industry are both powerful illustrations of the cooperation between state and private interests. Or at least an illustration of the difficulty of disentangling the two.

    And with unemployment numbers growing on top of already declining wages and the decreasing security of middle-class lifestyles, the tired argument about profit incentive just ads insult to injury. Thank you for pointing out the obvious: that the private sector can be outrageously wasteful, also.

  21. While searching for a write-up of the state of industrial-policy theory and/or best practices, I found this 2004 paper — Industrial Policy for
    the 21st Century — by Dani Rodrik, a Harvard economist:

    Click to access unidosep.pdf

    As it happens, this paper reads like an advertisement for the VC fund that I proposed above.

    Representative excerpts from the paper:

    From page 12:

    “Since self-discovery requires rents to be provided to entrepreneurs, one side of the policy has to take the form of a carrot. This can be a subsidy of some kind, trade protection, or the provision of venture

    From page 27:

    “Therefore, governments will generally need a facility to defray the costs of the early stages of the cost discovery process. The manner in which this would be done can be envisaged as a ‘contest’ whereby
    private-sector entrepreneurs would bid for public resources by bringing forth pre-investment proposals. The criteria for financing such studies would be that (i) they relate to substantially new activities; (ii) they have the potential to provide learning spillovers to others in the economy; and (iii) the private sector entities are willing to submit
    themselves to oversight and performance audits.”

Comments are closed.