Lecture 8: Constitutions
The Comparative Institutions Approach
The Emperor's tale.
Dictatorship has obvious problems. How do the problems of democracy
Economically literate defenders of democracy
typically focus on government's special ability to supply public goods. After analyzing how democracy actually
works, what can we say about this?
If people voted rationally and instrumentally, the
public goods argument makes a great deal of sense.
However, since voting is largely expressive and
not instrumental, public goods are supplied only by coincidence.
Ex: National defense may have expressive as well
as instrumental value.
But: Expressive voting is likely to lead to
government involvement in a wide array of "feel-good" issues that
have nothing to do with public goods.
Care of the old
Similarly, whenever voters hold systematically
mistaken beliefs, we are likely to see government wasting resources or being
Labor market regulation
All this adds up to an efficiency case for
limiting the scope of democracy in order to keep "political
pollution" under control.
Still, if dictatorship and democracy are our
only choices, democracy seems like the lesser evil.
Alternatives to Democracy and Dictatorship
Is democracy the only alternative to
dictatorship? Most countries that
we call “democracies” have many
undemocratic elements, including:
Division of powers
Limitations on the scope of democracy
Franchise restrictions – most notably the
restriction that only citizens get to vote!
Non-economists often insist, “This is a
republic, not a democracy.”
Strictly speaking, they’re right, even if unpopular restrictions
on democracy tend to disappear.
Alternate formulation: “There are many
different kinds of democracy.”
Whether or not you call them “democratic,”
couldn’t many of these restrictions mitigate democracy’s political
Ex: Supermajority rules (2/3 vote for price
controls), limitations on the scope of democracy (“Congress shall pass no
law infringing the freedom of production and trade”), and federalism
could all mitigate anti-market bias.
Ex: Giving two votes to college grads could
mitigate anti-foreign bias.
Ex: How about giving the CEA veto power over
Constitutional Reform and Endogenous Institutions
Many people in public choice – most
prominently, James Buchanan – advocate constitutional reforms to improve
upon the status quo.
Example: 2/3 rule for spending. Perhaps then, pork barrel projects will
fail, and only genuine public goods will receive funding.
Problem: It seems like any public choice problem
that afflicts day-to-day democracy will afflict constitutional choice as
In other words, constitutions are endogenous institutions. They are a product of the same forces
that generate other social outcomes, and no easier to change.
French/U.S. switch thought experiment.
Some proposed constitutional amendments have no
obvious impact on policy. These are
relatively easy to pass, but what's the point?
Other proposed constitutional amendments would
obviously affect policy. These are
hard to pass because the policies that currently exist tend to be popular.
Note: This doesn’t mean that direct democracy
always prevails, just that invoking “constitutional changes” as a
solution to problems with the status quo is probably wishful thinking.
Are Constitutional Politics Different?
Still, Buchanan and others insist that
constitutional politics are
different. Their central argument:
Constitutional politics operates behind a real
(not merely hypothetical) “veil of ignorance.” This leads people to selfishly favor socially efficient
Buchanan’s favorite example: Auto accident
liability rules. At the
constitutional level, no one knows whether he’ll be a plaintiff or
defendant, so we can get unanimous (or “virtually unanimous”)
support for efficient policies.
This whole argument rests on the discredited
SIVH, but it fails even on its own terms.
There may be some
constitutional rules where a veil of ignorance applies (though even
Buchanan’s favorite example overlooks lawyers). But most constitutional rules are about permanently locking in existing political
Consider a few examples from the U.S.
Constitution. Cui bono ex ante?
The purpose of the Senate is to permanently give
small states disproportionate influence.
The purpose of the slave trade provision is to
make sure that the slave trade remains legal until 1808.
The purpose of the three-fifths compromise is to
reduce the total influence of the South, but increase the influence of Southern
Or consider modern some Constitutional decisions. Cui bono ex ante?
The purpose of court rulings on religion is to
prevent the religious majority from doing what it wants to the secular
The purpose of court rulings on abortion is to
prevent states with pro-life majorities from restricting abortion.
In each of these cases, it is obvious ex ante
who will benefit and who will lose.
The point is to reassure the winners of today that they will continue to
get their way even if political conditions change to their disadvantage.
The most original and thoughtful suggestion for
constitutional change in decades, if not centuries: Robin Hanson’s
Background: Empirically, prediction markets (a.k.a.
betting markets) are the best way to estimate the future. They are the turbo-powered version of
“Put up or shut up.”
question that eventually has a verifiable answer can be turned into a financial
instrument – and its market price will efficiently aggregate all available
This includes contingent instruments such as…
A stock price conditional on firing the CEO.
The unemployment rate conditional on Obama
Terrorism deaths conditional on invading Iraq.
Tax rates in 2020 conditional on TARP.
Do betting markets give perverse incentives to
make bad things happen? We rarely
worry about this for traditional financial instruments, and in any case there
is a simple solution: Register the bettors, and/or cap the bets.
9/11 Commission found no evidence that anyone
used prior knowledge of the attacks to profit.
Do betting markets encourage manipulation to
mislead people using the market price to make decisions? No.
Manipulation just provokes arbitrage. See the Hanson-Oprea experiment.
Robin’s innovation: A constitution could require decisions to be based on betting
Moderate example: A corporate charter could
include a “fire the CEO” provision that says that if the value of
the firms’ stock conditional on firing the CEO ever exceeds its value
conditional on retaining the CEO, the CEO gets fired.
Robin advocates turning this approach into the
basis for a whole system of government.
He calls it “futarchy.”
Slogan: “Vote on values, but bet on beliefs.” This means that:
The political process defines an objective
function, such as “maximize GDP” or “maximize GDP plus the
market value of leisure” or “maximize GDP per capita times life
expectancy.” As a short-hand,
Robin calls the maximand GDP+.
If betting markets say that a policy has a
higher expected value of GDP+ than the status quo, the constitution requires the adoption of that
Ex: Someone proposes TARP. Betting markets on GDP+|TARP and
GDP+|~TARP go online. If and when the
value of the former exceeds the latter, TARP gets adopted. Critics can then set up betting markets
about the expected effect of abolishing TARP.
Most criticisms of futarchy argue that betting
market odds are not to be trusted.
These criticisms are weak, and often ignore extensive, specific
counter-evidence and simple fixes.
Will the marginal suicide bomber change his mind
if he add $100 to his will?
My main concern is with the definition of
GDP+. A key weakness of central
planning was that managers were given maximands that sounded good at the time,
but gave perverse incentives.
The Paradox of Public Good Provision
Isn't monitoring the government to act in socially beneficial ways
itself a public good?
The Paradox: If citizens can voluntarily produce the public good of
monitoring government, why can't they solve other public goods problems without
government? If they can't
voluntarily solve this problem, what reason is there to expect government to
Cowen and Kavka (2003) offer several solutions to this paradox:
The tying hypothesis. This
is very similar to the stationary bandit model: As long as leaders have a
reasonably long time horizon, socially beneficial policies are a private good for government leaders.
Altruism and noise hypothesis.
Governments can leverage a small initial altruistic donation. “The establishment of a government…
may require a smaller number of acts of altruistic support than
non-monopolistic mechanisms for producing public goods would require. Unlike a
competitive firm, once a monopolistic government obtains a certain amount of
initial support, it can use its monopoly power to induce continued support in
Bootstrap hypothesis: Support for government is really a coordination
game, not a Prisoners’ Dilemma, because if most people support a state
that provides public goods, it is selfishly optimal for you to do so as well.
From discussion with Tyler: Unless anarchy is better than government,
government must on net produce public
Are the Functions of the Night-Watchman
State Really Public Goods?
OK, so what about anarchy?
If you remain dissatisfied with democracy,
reject constitutional reform (even futarchy!) as a solution, and abhor
dictatorship, anarchy is all that’s left.
There is a surprisingly large economic
literature on the possibility of “free-market anarchism” or
“anarcho-capitalism,” beginning in 1849 with Gustave de
Molinari’s article, “The Production of Security” in the Journal
des Economistes. See Stringham’s Anarchy
and the Law for a broad survey.
Our analytical procedure: Start with the
functions of the night-watchman state (NWS) – dispute resolution, rule
formation, and enforcement – and work backwards.
Specifically: What aspects, if any, of these
functions are really public goods?
Dispute Resolution as a Private Good
Two people have a contract dispute. The night-watchman state takes it upon
itself to adjudicate. For disputes
it labels “criminal,” the NWS goes further by handling the prosecution.
Where is the public good?
Why couldn’t the parties (or their insurers) simply put an
arbitration clause in their contracts, which the night-watchman state
enforces? And why does the NWS have
to handle the prosecution itself?
Hard case for arbitration: complete
strangers. Night-watchman state
could still minimize its role by compelling defendant to choose from an
approved set of arbitrators.
Hard case for private prosecution: indigent
Rule Formation as a Private Good
The NWS also takes it upon itself to pass civil
and criminal laws.
Where is the public good?
Private arbitration firms could (and to some extent already do) ex ante
woo customers by offering the most efficient set of rules.
Efficient rules give parties incentives to
maximize joint wealth of signatories, factoring in expected cost of disputes.
Landes and Posner argue that lack of
intellectual property rights in precedents leaves little incentive for rule
In practice, though, public courts often defer
to expert judgments of arbitrators.
Intellectual property rights in rules could be strengthened, and
non-patent incentives are often effective.
Enforcement as a Private Good
The NWS also tries to monopolize enforcement.
Where is the public good?
Once an arbitrator makes a decision, why can’t it be enforced by
ostracism, bonding, or private security guards?
An unconventional solution to the problem of
indigence and crime: indentured servitude and private prisons.
NWS could again mechanically defer to
Moderate versus Radical Privatization
To a surprising extent, markets already share
the supposedly exclusive turf of the NWS.
Arbitrators’ codes of rules
It would be easy to drastically expand the role
of the market in these areas, while leaving the state as the final
authority. Let us call this moderate privatization.
But this is hardly an alternative to government,
because the state remains – and decides how far the alternatives are
allowed to go.
What about radical
privatization - getting rid of the last remnants of the NWS state?
What would this even look like? Standard picture:
Individuals subscribe to services of a defense
Defense firms arbitrate their disputes in
Private courts issue monetary judgments.
Reliable people unable to pay their debts have
their earnings attached.
Others, especially criminals, become indentured
servants and work off their debt in private prisons that bid for inmates.
Main Objections to Radical Privatization
Obviously, free-market anarchism is not a
popular proposal. What are the main
"Externalities of defense services.”
Reply: These have more to do with current policy
than the nature of the product. If police only help paying customers, if
judges charge for adjudication, if victims who prosecute win restitution, where
is the externality?
Turnaround: Government defense has lots of
externalities. Bureaucrats who make the world safer get paid the same as
those who don't. Oftentimes "crusaders" become very popular by
causing the crime they claim to be fighting (e.g. Prohibition).
"It would lead to violent chaos."
Reply: Why? It is cheaper to negotiate than to
fight, especially since police companies repeatedly interact with each other.
Also, police company employees, unlike conscripts, have to be paid more for a
Turnaround: Existence of government leads to
wars, which are far more serious than police agency shoot-outs because
governments control the resources of the whole society.
One strong agency would take over and become the
new government." (Alternate version: Agencies would merge until they had a
Reply: This would only be possible if there were
a large MES relative to the demand for defense services. In the current U.S., there are
already about 10,000 security guard companies, so there is little reason to
fear the dominance of one strong agency.
Turnaround: A much bigger risk from governments,
since their MES is much larger than for mere police agencies.
“Police agencies would build up demand by
defending their clients to the death."
Reply: This would create a severe "adverse
selection" problem, just like the one that insurance companies face. If you announce that you will protect
your clients to the death, you encourage high-risk, lawless people to hire you,
raising your costs enormously.
“People would have no incentive to
Reply: Monetary restitution provides the
incentive; the ability to impose indentured servitude ensures that almost all
convicted persons will be solvent.
Turnaround: This is true today: the only
incentive of victims to cooperate with prosecutors is desire for revenge. E.g.
rape victims are often better off staying quiet.
“Criminals would be over-punished."
Reply: In many ways, profit-making prisons would
be more humane: there is an incentive to protect the safety of workers, to
separate workers by size and strength, and to provide useful job training.
Legal codes could incorporate prisoner protection as well.
Turnaround: Numerous non-violent offenders are
currently sentenced to harsh prison conditions.
Cowen, Anarchism, and Collusion
Tyler Cowen has a novel objection to free-market
anarchism: Defense is a network industry. Collusion is relatively easy to
orchestrate in such industries due to the threat of expulsion.
There might seem to be a lot of firms, but for
practical purposes there will only be one.
In such industries, some forms of collusion
(relating to product standardization) are actually beneficial; but this opens
the door to the traditional forms of inefficient collusion (like price-fixing).
According to Cowen, this leaves basically two
Case 1: Transactions costs are low enough to
make collusion work. Then both the
efficient and inefficient forms of collusion thrive. The public faces a profit-driven
monopoly defense firm, a great example of Hobbes’ Leviathan.
Transactions costs are too high to make collusion work. Then neither the efficient or the
inefficient forms of collusion appear.
The public faces chaos, a great example of Hobbes’ “state of
Stringham and I reply to Cowen in the RAE. Main argument: Cowen conflates
coordination problems with prisoners’ dilemmas.
Product standardization is a coordination
problem. Once firms expect certain
standards to prevail, it pays to conform.
Price-fixing, extortion, etc. are
prisoners’ dilemmas. If other
firms are changing high prices and mistreating customers, it pays to stay
honest and gain their business.
Historically, network industries do not seem to
have been especially able to achieve collusive outcomes, but they have solved a
lot of coordination problems:
After anti-trust: Credit cards
Before anti-trust: Clearinghouses
What about national defense? Isn’t that clearly a public
Answer: It depends. “National defense” is not a
public good for the world because if
no country had “national defense,” no country would need it!
Implication: Countries’ “national
defense” programs are often a public bad – and the losers typically
include their own citizens.
Simplest reply to the national defense
objection: Our country’s national defense is a public bad, and both we
and the rest of the world would be safer without it.
This argument became far more convincing after
the fall of the Soviet Union.
Still, how would a free-market anarchist society
defend itself against invasion by an aggressive state? Surely that would be a public good.
Voluntary charity. Members of the anarchist society could
volunteer both money and their labor to defend their society from attack.
A successful anarchist society would have many external sympathizers,
making it difficult to rationalize aggression. Look at how little the capitalist world
did to crush the Soviet Union in its infancy.
If this seems naïve, consider a Tullockian
challenge: “Give me one good reason not to invade Brazil.”
The Transition Problem
The biggest challenge to anarcho-capitalism: The
transition problem. How do we get
from here to there?
Most radical movements consider violent
revolution. But even if this were
feasible, what are the odds that violent revolution would create a freer society, much less a free society?
David Friedman’s quip: “Revolution is the hell of it.”
From a libertarian point of view, most
revolutions are a complete disaster.
Even the American Revolution is greatly overrated by libertarians
– government didn’t really shrink, and the rights of slaves and
Indians would have been better protected by the British.
Question: Are revolutions a good example of the
Other options are viable for marginal change,
but offer little hope of radical change in the foreseeable future:
Infiltrating the political system to take
advantage of slack
Coordinated movement to change the median voter
(e.g. the Free State Project)
What about just creating a new society?
“Create your own country” projects
end in failure or conquest by neighboring old country.
Paul Romer’s solution: Charter cities
Patri Friedman’s solution: Seasteading
My case against fatalism: Radical change is very
difficult to achieve, but what sensible person ever thought otherwise? To quote Tollison, “We’re
all part of the equilibrium.”
You can still make a marginal difference – and do good while doing