Ch. 9: Credit constraints

[Aaron Swartz]

Farmers in the South could plant cotton or corn. After the Civil War,
the relative amount of cotton planted increased 50%, even as the
relative price for cotton went down. Estimates find the average farm
could have made 29% more had they planted cotton. Why didn't they? The
former slaves had to live on credit and credit was only extended to
people who planted cotton, because it was seen as safer. Not having
credit sucks.

One estimate is that two-thirds of Americans are credit-constrained.
An inheritance of $10,000 doubles the average British youth's
likelihood of starting their own business. (Additionally, a 10%
increase in UK house prices led to a 5% increase in startups.) And, of
course, the credit-constrained are typically the poor and young and
disadvantaged.

How can we model this? Imagine you're buying a machine that can be run
at various speeds. The faster you run it, the more it produces but the
more likely it is to break. (Why this machine is almost like a bet...)
If you're spending your own money on the machine, you'll try to
maximize the average amount it produces, so you'll run it at
half-speed. If you're borrowing money to buy the machine, the borrower
will charge you more in interest based on how fast you run the
machine, so again you'll run it at half-speed.

Now imagine that the lender can't tell you how fast to run the
machine. You run it at 3/4 speed, since you get the additional money
if it doesn't break and the lender eats the loss if it does. In this
case you're actually skimming money off the lender. Perhaps this could
be used as the basis for contingent renewal -- continuing to extend
loans as long as the machine doesn't break. This improves things, but
not completely: you now run the machine at 2/3.

Why would you invest your own money instead of taking out a loan? You
could be trying to prove to the lender that you really believe it's a
good investment, or you could be trying to lower this discrepancy
between lender and borrower (you won't run the machine as fast when
it's your own money on the line). As the wealth you risk increases,
the speed you run the machine at moves from 3/4 to 1/2.

In a perfectly competitive market, lenders will receive the risk free
interest rate from any investments they make. (That's kind of what it
means to be perfectly competitive.) Since the amount of wealth someone
can invest in their machine determines how fast they run it and thus
how much it makes, you need a minimum level of wealth for people to
lend to you. (If you don't have that much wealth invested, you'll run
the machine too fast and so no one will want to lend to you.)

In a noncompetitive market, lenders and borrowers can bargain anywhere
between the lender getting the risk-free rate of interest and all of
the profits. (This is the situation with "an urban pawn shop or
'payday lender'".) But in a perfectly competitive market, lenders will
offer the interest rate that makes them the risk-free rate of
interest, which means the wealthier you are, the lower your interest
rate is. Wealthier borrowers can also borrow more (because lenders
will require you put up more) and for lower-quality projects (ditto).

This can't be efficient -- rich people will get low-quality projects
funded, while poor people can't even get their high-quality projects
funded. Thus redistributing money from rich to poor is socially
optimal: it improves the overall quality of funded projects. But
here's the _really_ crazy thing: it can also be Pareto improving.
Normally, you'd think redistribution can't make everyone better off --
you have to take the money _from_ someone. But get a load of this
case:

Let's say you have a lot of money. You loan it out and you get the
risk-free rate of interest for it. Now imagine Big Government comes
and takes a chunk of it and gives it to a poor person to buy a
machine. It's now the poor person's own money at stake, so they
operate the machine at its optimum: half-speed. This makes them a
bunch of money, so they're happy. Then the government requires them to
pay the risk-free rate of interest back to you, so you're no worse-off
than before. Incredible: everybody wins! (Why don't you just lend the
poor person the machine for the same amount? Because the loan is
unenforceable; the magic comes from the government forcing the poor
person to pay you the risk-free rate of interest whether the machine
fails or not.)

The problem is that poor people don't like having to pay a bunch of
money to get a risky machine -- it's much less fun to risk your money
when you're poor than when you're rich. As a result, some poor people
may prefer to be employees (getting a guaranteed income) than own
their own machines (getting a higher average income, but with a
greater variance). [Then there's a bit I don't quite understand --
Chris? -- but it seems to argue that in most cases things will be such
that redistribution will be socially-improving, but not
Pareto-improving.]

In conclusion, estate taxes and wealth redistribution are a good thing
-- there is no efficiency-equity tradeoff. Taking a little bit of
money from the rich and giving it to the poor makes it so that the
poor can do new things, while not preventing the rich from doing
anything. The real trouble is that by moving from an economy where
some people are employees to one where everyone's an owner, you reduce
the overall level of risk-taking (since big companies will take more
risks than individual owner-operators).

It's worth thinking about policy responses to deal with these
problems. Options include microfinance, performance pay, and
insurance. [AS: Is there also room for state lending and submarket
rates of interest?] Our pareto-improving example showed that
government intervention can make everyone better off.

[Chris Mealy]

Shorter Bowles: economics in the real world is a game of musical chairs played with golden handcuffs. 

Chapter 10 concentrates on two things everybody cares about but economists never talk about: power and class.

First, I loved Bowles's one-paragraph history of capitalism (p. 335):

"What capitalism accomplished, and what accounts for much of its productive success, is that it allowed some individuals to innovate and take risks on a grand scale with a reasonable expectation of reaping the rewards of a successful project while bearing the costs of a failure. Inequalities in wealth combined with credit and other financial markets (aided by the introduction of limited liability) allowed a single individual or a small group to amass substantial resources under unified direction. Labor markets allowed these material resources to be put to use to employ vast numbers of workers, thereby reaping both technological and organizational increasing returns. The wealth (and the creditworthiness) of those directing these business projects made the risks of innovation tolerable. It also allowed them to offer a modicum of de facto insurance to those whom they employed, in the form of the wage contract. The result was that those with nothing to supply other than their labor could be mobilized into projects the risks of which they would never have been willing to assume personally. For the first time in history, competition among members of the economic elite depended on one’s success in introducing unprecedented ways of organizing production and sales, new technologies, and novel products. The success of these arrangements hinged critically on the relative security of possession associated with the rule of law, accomplished in large part by the increasingly powerful national states that grew in symbiosis with capitalist economic institutions."

Alright then, power. In a world of complete contracts there's no power. The chapter's epigraph from Paul Samuelson captures how strange that is: 

"Remember that in a perfectly competitive market it really doesn't matter who hires whom; so have labor hire capital." 

The efficiency of markets hinges on complete contracts. Even when economists acknowledge the lack of complete contracts they'll often try to salvage the assumption of efficiency. They'll assume that by means of competition (like natural selection or something) efficient institutional arrangements automatically emerge. (This reminds me of adaptationist/Dawkins vs pluralist/Gould debate). Bowles challenges the assumption of efficiency:

"The widespread (often implicit) use of the efficient design assumption is explained in part by the fact that as an empirical matter it is virtually impossible to determine whether a particular institutional setup IS efficient. Few attempts to do so exist, and even fewer enjoy the endorsement of a consensus of scholars in the relevant fields." (p. 337)  

Market efficiency is too good to check!  Bowles thinks we can do better:

"if the structure of contracts and other institutions are not the result of some hidden algorithm that implements efficient solutions to allocational problems, what analytical tools can we deploy to explain empirically observed institutions and their evolution?" (p. 337)

This brings us to the seldom-noted flip side of Coasean bargaining: with transaction costs the distribution of property rights will have an effect on efficiency. This is where power comes in:

"The combined effect of incomplete contracts and conflicts of interest is that the determination of outcomes depends on who exercises power in the transaction. Power is generally exercised by those who hold the residual rights of control, meaning the right to determine what is not specified contractually." (p. 333)

Bowles cites Robert Dahl's (apparently famous) definition of power:

"A has power over B to the extent that he can get B to do something that B would not otherwise do." (p. 344)

Bowles and Gintis have their own definition:

"For B to have power over A, it sufficient that, by imposing or threatening to impose sanctions on A, B is capable of affecting A’s actions in ways that advance B’s interests, while A lacks this capacity with respect to B" (p. 345)

This sounds a lot like the endogenous enforcement / enforcement rent from the last couple of chapters. Here it comes:

"those with power are transacting with agents who receive rents and hence are not indifferent between the current transaction and their next best alternative. This being the case, there must exist other identical agents who are quantity constrained, namely, the unemployed, those excluded from the loan market or restricted in the amount they can borrow, and sellers who fail to make a sale." (p. 347)

(This is the musical chairs/golden handcuffs part. The firm overpays you slightly so you'll want to keep your job. After three chapters it's still blowing my mind that most markets don't clear.)

Bowles defines "short-side power" and "the long-side of the market." Note that he applies power only to the short-side:

"Those holding power in these cases are those on the side of the market for which in equilibrium the number of desired transactions is least, or what is termed the short side of the market. Note that the short side may be either demand (the employer, the buyer) or supply (the lender). Those on the long side of the market will be of two types: those transacting the amount they wish, and those who are quantity constrained (either excluded or transacting less then they wish)." (p. 347)

Later Bowles has a terrific example of how this differs from the saying "money talks": in the old communist countries cars were sold below market-clearing prices. In that case the buyers (the cash money) were quantity-constrained. They were on the long-side of the market. (p. 348)

While he's on the topic Bowles discusses other kinds of power:

 "Purchasing power is just another word for one’s budget constraint (or wealth), and it does not concern the exercise of sanctions or indeed any strategic interaction at all." (p. 348)

"Market power" means the ability to set prices, but without the capacity to sanction it's not necessarily power at all. (p. 349)

"Bargaining power" is not the same as short-side power: "bargaining power refers to outcomes -- to how much advantage one may gain -- rather than to any particular means of attaining it" (p. 349) When an employer holds short-side power it's because they're offering an enforcement rent. The rent is going to the employee while the employer is producing at the level of the zero profit condition. The hired employee is capturing the gains of cooperation and in this case has the bargaining power even though the employer has the short-side power.

This is too long already and I'm only halfway through. In part two, the emergence of class!

-- Chris