heterodox economics

=economics

 

 

This is a summary of some disagreements I have with what media such as the New York Times, Wall Street Journal, or The Economist would consider "conventional" economics. It could be called "contemporary American neoliberal economics" so I'll abbreviate it as CANE.


 

 

humans are not (mainly) utility maximizers

 

An individual person acts like a collection of agents with different goals and time preferences, where each agent has varying influence depending on current circumstances. This leads to things like time inconsistency, but is still better than a single agent because the optimal optimization targets are unknowable. There's an ancient Japanese saying about "cognitive biases": バカっていうほうがバカなのよ!

 

Most of the time, most people act more as status-imitators than utility-optimizers.

 

A common type of advertisement is: show an attractive person, show them doing something high-status, and then show them using a product.

That's effective because many viewers tend to imitate high status people.

It's also common to pay large amounts of money to celebrities/athletes for endorsements of products.

 

The CANE view of such advertisements is basically the libertarian view: the ad-watching is transactional and so overall utility must be going up. I disagree; my view is that television has managed to hijack instincts for imitation of successful people, and with American households now watching 8+ hours of TV/Netflix/etc a day, this is a big problem.

 

It would be easy to say, "People just need to be deliberative utility maximizers like me instead." But I've known a lot of smart people who decided to do something unconventional with their life, and on average, their decisions turned out worse than just following conventional wisdom. Imagine if a 14-year-old kid was somehow unable to know what their peers were doing - would they say, "I need to spend years and money studying pointless and often wrong material in a high school, so I can get into a college, where I'll study material I'll never use except to get good grades in college, so I can get my resume past someone at the HR department of a company, and get a job"? In general, it's impossible to do that. Do you own some index funds? If so, you didn't decide that specific things were worth investing in; you have no idea what your money is doing on a real-world physical level. And that's OK, but you shouldn't pretend you're "maximizing utility" instead of choosing a group to imitate.

 

 

 

jobs have high value for employees

 

Empirically, people are unhappy when they lose their job, but employers are much more neutral about hiring and firing.

 

The CANE view of employment value is:

job_value_for_employee = pay - work

job_value_for_employer = work - pay

 

But if a worker is only paid enough to be neutral about having their job, they won't put any effort into keeping their job, which makes them worthless as an employee.

 

My view is:

job_value_for_employee_before_effort = pay - job_inherent_costs_before_effort

job_value_for_employer = employee_skill * log(job_value_for_employee) - management_time * manager_time_value

 

 

Laws restricting non-compete agreements are common.

 

The CANE view of those seems to be: "Government limits on employment contracts are harmful and only happen because voters are dumb."

 

My view is different. If employers want jobs to have high marginal value, then they'll do things that reduce the net value for employees of their jobs without reducing the marginal value. Non-compete agreements have that effect. Non-compete arguments make an employer better off, but their employees and competitors worse off, so overall, banning them is welfare-improving.

 

 

 

"sectors" are not adequate categories

 

Suppose a country builds too many houses due to mispredictions about how many people will want their own house. Then, few new houses are wanted for a while.

 

- This results in insufficient work for people who build houses, but not people who build large commercial buildings. That's the "construction sector".

- Wood is used for those houses, so wood demand goes down. That's the "agriculture and forestry sector".

- Concrete is used for those houses, so concrete demand goes down. That's the "mining" sector.

- Fewer mortgages are needed than expected, so financial stocks go down. That's the "finance" sector.

 

CANE economists look at all these "different sectors" going down together and say, "Aha, this is a depression hitting all these different areas, which means it's probably about money, not real reallocations."

 

 

 

making useful predictions is important

 

I had the following conversation with an economist:

me: As Peter Thiel has noted, CANE predicted capital flowing from America to China during a period of catch-up growth, but we've actually seen the opposite.

economist: That was the Solow model. Now there's the Romer model, which is consistent with trade numbers.

me: What useful predictions has the Romer model made?

economist: Predictions? We just make the models; other people can make predictions with them if they need to.

 

 

 

big companies are a bad sign

 

If companies exist because internal transaction costs are lower than market transaction costs, then huge companies are a sign that market transaction costs are high, which is bad.

 

If companies are huge because they make internal software tools, then that's bad. It's more efficient for tools to be provided independently, because then they can be shared between more users, and users get more choices for software. The same is true for physical infrastructure, like distribution warehouses, that's vertically integrated to prevent competitors from using it. This kind of integration is a form of monopoly.

 

If companies are huge because they negotiate special treatment from regulators, then that's bad.

 

If companies are huge because they have a profitable "core" that management uses to buy other companies despite those purchases lowering the company's stock value, then that's bad.

 

There are several reasons why big companies are a bad sign and you should be suspicious of them. And yet, the CANE view is that people are suspicious of huge corporations because people have a "corrupt emotional makeup". "How can people be ungrateful toward corporations? Corporations give us everything! Corporations do everything for us!"

Personally, I think you're the one with a problem, Bryan Caplan, but the entire field of macroeconomics shares some of the blame for that.

 

 

 

acquisitions can be bad for consumers

 

When EA buys a game company that people like, are those gamers right to be upset?

 

The CANE view of such aquirements is: if EA tries to extract more rents from consumers after buying game companies, the money spent will go to and/or encourage new game companies, so people will still be better off.

 

The fact is, that's not how it works. I've done some game development, and I've known people making new game companies, and empirically, such acquisitions have done nothing to help the creation of new game companies. The people making good games at new companies are the people who make games because they just want to make good games. Maybe the CANE view here would make sense if money was the only motivation of humanity, but I actually think it still wouldn't.

 

 

 

high table stakes

 

SpaceX didn't exist until Elon Musk came and made a rocket company. It had lower launch costs than ULA before it started landing rockets, with very conventional rocket technology.

 

Elon Musk is a good businessman, sure, but there are other equally competent people. If conventional rocket technology was enough, why did it take Elon Musk to make SpaceX?

 

Because he was rich enough to have skin in the game, he'd succeeded in business before, and he started hiring people. Elon Musk has been able to easily pull in $billions of investment - he's a nucleation point in a financial supersaturated solution.

 

What happens when the central person doesn't put up those table stakes? Often, something like Theranos.

 

These days, the table stakes for a viable business are often over a billion dollars, and there's a shortage of people who can fill that role. That's why Elon has such a cult following.

(note: Personally, I'm not optimistic about Tesla Motors or The Boring Company.)

 

On the other hand, restaurants have "table" stakes within the means of relatively normal people, and so there's no shortage of new restaurants.

 

 

 

company management has power

 

When company A buys company B, the usual pattern is that A's stock goes down and B's stock goes up. Yet, company management pushes to buy companies, and resists its company being purchased. This implies that management has power and different goals from shareholders.

 

CANE theory often assumes near-perfect competition between companies, which are held back from monopoly formation or exploitation by government. If that was the case, companies would be deperately trying to merge to get monopoly power, with a lot of tension between them and government regulators. That's not what we see.

 

When near-perfect competition happens in some area, what we actually see is:

1) Conglomerates sell off their divisions in that area.

2) Those sold divisions and any independent companies in that area rapidly merge until there are 2 to 4 left. They stop before government gets involved.

3) Those companies are profitable because they've become an oligopoly with market power.

4) Once a company is profitable, management has rents it can extract, and being purchased would take away those rents. Internal resistance from management to being purchased, not government regulators, is usually what stops a wave of mergers.

 

 

 

opaque pricing is commom

 

What's the market price of 1000 tons of glycerol? You don't know?
It's a trick question - chemical companies don't know either, so there is no market price; chemical companies are paying significantly different prices.

 

What's the cost of some type of surgery at an American hospital?

Another trick question - surgery costs typically vary by more than 3x between different American hospitals.

 

What's the cost of tuition at an American university, after "financial aid" which is often a large % of the tuition? Once you're accepted, fill out some forms and they'll tell you what their overall price is.

 

CANE theory generally assumes that prices are common knowledge, but that's just not true for much of the economy.

 

 

 

property lines

 

When people buy a house in a city, they're paying more for the "location" than the house. "Location" means "what's in the surrounding area" and the closer something is to someone's property, the more they care about it.

So, most people are happy to accept limits on their property rights in exchange for limits on their neighbors' property rights. There are home owner assocations (HOAs) as well as laws about loud music at night and buildings that block neighbors' sunlight.

The CANE view of this is something like:

Instead of having laws about noise, the disagreement should be settled by one side paying the other to do what they want. Coase theorem.

I disagree, and not just because of "transaction costs".

But even without perverse incentives, there are reasons why normal people are skeptical of involving money in such negotiations, and economists are fools to assume that people are merely superstitious. A high-trust society relies on "stocks" of goodwill, and monetary negotations can be a mechanism for outflow of those stocks of goodwill, so they can contribute to collapse of a high-trust community.

 

 

 

cities consist mainly of people

 

A city consisting of empty buildings is basically worthless. People move to cities because they can buy and sell goods and services to other people there, people who have certain specialized skills.

 

Imagine a talented chef Joe who works for a restaurant with good but reasonably priced food. Joe rents a house from a landlord. The neighboring houses include an empty house owned by a rich guy living in China, and a house occupied by a professional panhandler and cellphone thief.

 

Now, suppose the property value goes up. That property value increase comes from the contributions of people like Joe, but his neighbors and landlord are the only ones who benefit from it. They're free-riding on people like Joe. As such, taxes and regulation on real estate investors holding empty houses are entirely reasonable.

 

 

 

some subsidies are strategic

 

first, see: industrial policies in production networks

 

That paper shows that subsidizing upstream industries can be logical when there are "market imperfections" from "financial and contracting frictions". But I think there's an even more significant argument here: management takes rents, and those managment rents are proportional to the amount of money being handled.

 

If a company SC makes shipping containers from steel, and government subsidizes steel but taxes shipping containers, that policy might appear to do nothing, but it's reducing the money handled by the management of SC, and if they can only extract a % of that money as rent, then efficiency might have been improved. This makes industrial policies subsidizing upstream products more appropriate in countries with high levels of management rent extraction.

 

 

 

some finance is harmful

 

In 2010, the state of Colorado tried to eliminate this feature of payday lending by mandating that they be offered in the form of installment loans. A legitimate question is why this contractual form should be mandated. The simple reason – very much in line with Gabaix and Laibson (2005) – is that unsophisticated borrowers cannot appreciate the convenience of installment loans. When they do not, lenders prefer conventional payday loans because they make customers borrow repeatedly, maximizing the fees they can charge.

 

The experience in Colorado is very positive. Three years after the reform, borrowers spent 44 percent less in interest than they had in 2009 under the conventional payday loan model, saving $41.9 million.

- Does Finance Benefit Society?

 

 

When it rains, interfering with high-frequency algorithmic trading, transaction costs go down.

When it rains or snows in the area between the two cities, the networks are disrupted because rain droplets and snowflakes block the microwave paths. With the networks temporarily down, information transmission falls back onto the fiber-optic cable – a more reliable, yet slower transmission medium – effectively eliminating the speed advantages of the fastest traders. We show that when this happens, adverse selection and trading costs decline.

 

 

 

 



back to index