Arnold Kling on Specialization and Trade
What this covers
Arnold Kling and Russ Roberts examine why mainstream macroeconomics fundamentally misrepresents how economies function. Kling contends that macro models reduce the entire economy to a single productive unit—a "factory" with homogeneous labor and capital—when reality is a constantly shifting network of specialization and trade across millions of firms and workers. This abstraction obscures the actual coordination problem: when disruptions occur, recovery depends not on aggregate spending or stimulus, but on entrepreneurs discovering new sustainable patterns of work. The conversation moves through Hayek's Knowledge Problem, the gap between economic theory and political implementation, and why markets often outperform government at error-correction despite their visible messiness.
The discussion ranges across how subsidize demand while restricting supply—as in owner-occupied housing, health care, and education—raises prices without expanding output, leaving consumers worse off. Kling argues that credentialization has ballooned from 5 percent of jobs in 1945 to roughly 30 percent today, blocking labor mobility. Roberts and Kling dissect G. A. Cohen's Camping Trip Parable, showing why organizing a small group of friends differs fundamentally from coordinating 330 million strangers through the price system. They address climate policy's intractable problem: nobody can identify an optimal Pigovian Tax, yet recommendations get mangled by Public Choice forces once they enter the political process. The conversation closes on technology's potential to route around regulatory failure in education and medicine, though without resolving whether disruption will truly break the existing framework or merely adjust within it.
Kling argues that mainstream macroeconomics misleads by modeling the economy as a single 'GDP factory' driven by aggregate spending, when the real economy is a constantly shifting web of specialization and trade whose disruptions and recoveries are microeconomic coordination problems, not problems solvable by fiscal or monetary gas pedals.
- Macro 'Model' reduces the economy to one factory with homogeneous labor and capital, hiding the coordination problem of specialization and trade
- Recovery from recession depends on entrepreneurs finding new sustainable patterns of specialization, not on stimulus
- The political process systematically subsidizes demand and restricts supply, so economists' optimal-policy recommendations get distorted into mere price increases
Macroeconomic modeling treats the economy as a single factory, losing sight of specialization and trade.
- Treating the economy as a machine operated with a gas pedal—where spending is the pedal—is the wrong model of coordination; organizing an economy is not a matter of hitting the gas or operating a few equations, but of people reconfiguring into new forms of specialization in response to changes in technology, tastes, and culture.
“thinking of the economy as some sort of machine that you can operate with a gas pedal. And this gas pedal is spending. Again, that's just not the right way to think of the problem of organizing or coordinating an economy.”
- Economists live by 'The Model'—writing equations where Y is total output, L is total labor, K is the capital stock—and in doing so reduce the entire economy to a single factory, having lost track of how far that abstraction differs from reality, much as accounting ledgers differ from a business: if accounts were the business, every business plan would succeed and all tycoons would be accountants.
“people write down a set of equations, of, say, 'Let Y be total output. Let L be the total amount of labor. Let K be the stock of capital.' And I'm sorry, but when you are doing that, you are reducing the whole economy to a single factory in your model.”
- Adam Smith founded economics on specialization, but economists have lost their way over the last 50 years by getting away from that focus, and refocusing on specialization and trade dispels widespread misconceptions held by economists, non-economists, socialists, and libertarians alike.
“Adam Smith, who is considered the founder of economics, started out, of course, talking about specialization. And I think that economists have a bit lost their way by getting away from that in the last 50 years.”
- Macroeconomics commits the 'GDP factory' misconception: it models the whole economy as a single business producing one good from homogeneous labor and capital, so unemployment looks like people not buying enough from that business; but this misapplies single-firm logic to an economy that is constantly splitting up and reconfiguring work into new patterns of specialization and trade.
“one of them is what I call the GDP factory [Gross Domestic Product]--thinking of the economy as producing just one type of good out of, you know, homogeneous labor, homogeneous capital, as if it were one factory, one business”
- Kling recalls discomfort from his first days at MIT, experiencing graduate macro training as 'mathematical brainwashing' that lacked the critical thinking and plurality of viewpoints of his undergraduate education, and he never bought into the hubris that building mathematical models was the right way to do economics—believing that endless modeling would teach very little about the real world.
“I can remember discomfort pretty much from Day One at MIT... it was really kind of a mathematical brainwashing.”
- To the naked eye, when stimulus was hitting the gas pedal hardest employment performed poorly, and when policy switched to so-called austerity (easing off the gas) the economy and especially employment picked up, suggesting the gas pedal is meaningless—and the naked eye is at least as valid a way to read the data as torturing it through a model.
“with the naked eye, it looks like when we were hitting the gas pedal the hardest with stimulus, employment was performing poorly. And then when we switched to austerity... the economy picked up.”
Economic coordination emerges from decentralized trial-and-error by entrepreneurs, not central policy tinkering.
- Relative prices change all the time in response to market forces, but a sharp move like a halving of oil prices is meaningful only because most other prices are roughly stable; if the general price level were itself wildly uncertain, every price signal would be drowned in noise, so people naturally gravitate toward an expected, persistent price environment that the government cannot fine-tune.
“The reason that the price of oil is meaningful and we understand the impact of it is we have a sense that, at the moment, most other prices are roughly stable.”
- Money and prices are a consensual hallucination—a social convention in which yesterday's price anchors expectations of tomorrow's—so prices are far more stable than mechanical money-supply theories predict; people adapt their use of money and credit and can offset much of what the Fed does month-to-month, making the Fed just one bank among many whose money-supply role in coordination should be minimized.
“I have this phrase, that money and prices are a consensual hallucination. That is, it's more of a social convention that, okay, this was the price yesterday; that's what we expect it to be tomorrow.”
- Talking about what is 'optimal' wrongly treats technology and tastes as given and seeks a fixed social optimum, but since the optimum is unknown and constantly changing as people make discoveries and culture and tastes shift, the right goal is to design institutional processes that help these elements of change work out for the best, judged by institutional quality rather than optimality.
“when you are talking about institutional quality, you are saying, 'Well, nobody knows what the optimum is--and the optimum changes because people make new discoveries; culture changes, tastes change. So, what you want are institutional processes that you think help make these elements of change work out for the best.”
- Better economic outcomes arise from patterns of sustainable specialization and trade that come not from Federal Reserve tinkering or congressional deficit spending but from the creative, decentralized trial-and-error of thousands of entrepreneurs and individuals seeking the best use of their talents; thus the best government action to encourage new specialization is to rethink existing policies that restrict competition, discourage innovation, and retard mobility.
“Those patterns do not come about as a result of tinkering undertaken by the Federal Reserve, by deficit spending undertaken by Congress. It requires the creative decentralized trial and efforts of thousands of entrepreneurs”
- Rather than comparing an existing market configuration against an ideal based on a simple model (the Nirvana fallacy), one should compare the market's process of error-correction against the political process of error-correction; the market has the better trial-and-error process, as shown by private solutions to the Market for Lemons problem like CARFAX-style histories and CarMax guarantees that economists overlook when they assume only government can solve a discovered flaw.
“What we should be comparing is not the existing market configuration, with an ideal based on a simple model, but the market process of error-correction with the political process of error correction.”
Policy recommendations ignore how the political process systematically distorts them through special interests.
- Even a well-designed policy like a carbon tax is unlikely to survive the political process intact: rather than emerging as the recommended optimal tax, it comes out laden with special exceptions, subsidies, and favors, so an economist who recommends the pure policy knowing it will be distorted is being naive about how the sausage gets ground.
“what are the chances that the political process comes out with the tax that you recommend? More likely it comes out with all kinds of special exceptions and subsidies and favors. So, when you actually go through the sausage grinder you don't come up with anything close to what you like.”
- The political process makes systematic, not random, errors: it consistently subsidizes demand for a good while restricting its supply—as with owner-occupied housing—which contradicts public-goods/externality theory (which says produce more of positive-externality goods and less of negative-externality ones) and instead just raises prices, benefiting suppliers while consumers' subsidies are offset, because specialized producers have concentrated interests and diffuse consumers do not.
“What it does is it always seems to work in the direction of subsidizing demand for something and restricting supply for that same thing.”
- Recommending a Pigovian carbon tax assumes a knowable optimal level that nobody can actually identify, and the prescription is always one-directional—'I hate carbon, tax it more'—with no way to recognize when a tax has become too high, so it ignores all the adjustments the market is already making and overstates economists' ability to set the right number.
“Greg Mankiw is always talking about a Pigovian Tax, the Pigou Club: We need to tax carbon more. How does he know how much?”
- Macroeconomics is not an experimental science but a set of observations that can be interpreted within any preferred framework, so statistical evidence is almost never decisive in macro (it changes minds only in microeconomics); fiscal stimulus appears to do nothing to the naked eye, yet a model like the Congressional Budget Office's shows it works only because it works in the model, with the data never actually informing the model.
“if you run it through a standard macroeconomic model like the Congressional Budget Office Model, it works. But it works because it works. It works because it works in the model. The data aren't really telling the model anything.”
Occupational licensing and health insurance tied to employment create harmful barriers to labor mobility and hiring.
- Occupational credentialization has expanded enormously, with roughly 5% of jobs requiring a license at the end of WWII compared to about 30% today, which is excessive and impedes labor mobility.
“somehow, like 5% of jobs at the end of WWII required a license and now it's something like 30%. That's excessive. And that impedes mobility.”
- Linking health insurance to employment locks people into jobs and acts as a large tax on employment, so that low-wage employment is now taxed higher than almost any other activity because of the high payroll tax plus the requirement that employers provide health insurance, creating a major barrier to experimenting with hiring; reducing payroll taxes (offset by lower government spending) would be Kling's single highest-priority reform.
“linking health insurance to employment. I think that does a couple of things. One is, it has a tendency to lock people into jobs. Another is it's a big tax on employment.”
- Health care and education exhibit the same subsidize-demand/restrict-supply pattern: demand is heavily subsidized (third-party payment, education subsidies) while supply is actively restricted through hospital Certificates of Need, state medical licensing that doesn't transfer across states, and accreditation controlled by industry incumbents so that hardly any new accredited college has appeared in decades—making doctors comfortable and hospitals profitable while raising prices.
“in the case of hospitals, they require this Certificate of Need. And that has restricted the supply of hospitals. And of course all sorts of medical licensing restrictions exist, especially at the state level.”
- Health insurance as currently structured is not really insurance but a pre-paid health plan, and recent reforms have, if anything, entrenched that mis-structuring further in the name of reform.
“health insurance as it's structured is just on the wrong track. It's not really insurance. It's a pre-paid health plan.”
Financial disruption of complex production networks explains why economic busts cluster together.
- Because every complex pattern of specialization and trade requires financing and multistage production must be financed, the financial industry sits naturally at the center of the economy's network effects, so a sudden collapse in finance (or another central industry like 1970s oil) reverberates into widespread disruption of other patterns and takes a long time to readjust—explaining why busts get clumped together.
“all patterns of specialization and trade that are complex involve financing... the financial industry is kind of at the center of a lot of these networks. And so you could imagine big disruptions in the financial industry reverberating into big disruptions in other patterns of specialization and trade.”
- Real economic activity involves multistage production—getting the plow, the tractor, planting the seed—and trade through financial intermediation rather than barter, as when a surgeon trades minutes of surgery for lawn-mowing service via money rather than doing it herself, so financing is integral to every complex pattern of specialization and trade.
“we know that a surgeon shouldn't mow her own lawn. She should in effect trade a few minutes of surgery for a few months of lawn-mowing service. But she doesn't do that by barter. She does that through financial intermediation.”
Market processes for error-correction outperform government solutions when compared fairly.
- Capitalism does have something ugly about it relative to Nirvana—unfair advantages of luck and birth—but the misconception shared by economists and non-economists is to imagine someone with enough information to dispense cosmic justice; in reality everyone is equally ignorant of most things and knowledgeable about a few, so letting the market process work is often a better solution than giving someone power whose knowledge cannot match it.
“There's just nobody who sits up there in the throne dispensing cosmic justice. We're all kind of equally ignorant about a lot of things”
- G.A. Cohen's camping-trip parable—where organizing chores by a market would seem crazy among friends—is valid for a camping trip but a bad metaphor for the whole economy, because the trip involves only 6-10 people doing simple, roughly equal tasks while relying on gear (tents, pots, clothing) whose production required millions of coordinated tasks the parable pretends away, and because in the real world people's talents, inclinations, and the desirability of tasks differ, so getting people to be sanitation workers or chicken farmers without pay requires heavy-handed coercion.
“The bad move is to say that that camping trip, then, is a valid metaphor for the economy as a whole. Because it's not.”