Prof. Bryan Caplan

Econ 410


Week 8: Division of Powers, Special Interests, and Rent-Seeking

I.             Principal-Agent Problems, Multiple Agents, and Overlapping Principals

A.           We have already thought about simple principal-agent problems, where there is one principal (a shareholder or voter) who tries to set up good incentives for a single agent (a manager or politician).

B.           In many contexts, the problem is more complicated because there are multiple agents and overlapping principals.  In other words, a lot of different principals with potentially conflicting interests are holding the same agent responsible, and multiple agents share credit or blame when something goes right or wrong.

C.           The most natural intuition is probably that agents' performance falls because it is hard for principals to decide who to blame.  And if there are a lot of blameless agents, imposing probability multipliers on punishment may be a bad idea too.

D.           But there are a variety of contrary intuitions:

1.            Avoiding concentration of power

2.            Increasing stability

3.            Efficiency of specialization

E.           Many political systems, including that of the U.S., appear to act on one or more of these contrary intuitions by setting up a division of powers.

II.            Division of Powers and Minimum Winning Coalitions

A.           In the standard median voter model, whoever wins the election imposes its will.

B.           But as mentioned previously, in the U.S. and many other systems, political outcomes are a compromise between several different "winners."  How do these compromises work?

C.           One popular answer appeals to the notion of the "minimum winning coalition."  (MWC)  A minimum winning coalition is the smallest group large enough to impose its decision.

D.           Intuition: Why include more people in your winning coalition than you need?  That means dividing up the same pie between more people.

III.          Bicameralism and MWCs

A.           To simplify the analysis, imagine that you need a simple majority of two legislative houses to pass a bill (nothing more).  The MWC is 50%+1 of the lower house and 50%+1 of the upper house.

B.           So what happens?  Following Cooter, imagine putting median lower and upper ($L, $U) house preferences for spending on a line, and assume that $L<$U.  Similarly, put $L0 and $U0 to indicate the spending levels that leave the median lower and upper house legislators indifferent to spending $0.

C.           Cooter labels the region between $L and $U as the Pareto set.  Movement within this set makes one bargainer better off and the other worse off.  Moving from outside the Pareto set to inside, in contrast, makes all bargainers better off.

D.           Similarly, he labels the region between $L and $L0 as the bargain set.  This is the part of the Pareto set that  both sides prefer to $0. 

E.           Given the political rules, we know that policy must lie in the bargain set.  Theory can't narrow it down further, but it's a pretty good start.

F.            What does moving from one to two houses do?  It (weakly) tends to reduce the total level of spending because spending will never rise above the lower of $L0 and $U0.

G.           But this is not the only effect.  It could conceivably increase spending by bargaining the lower house to accept something between $L and $L0.  With only the lower house, you would get $L exactly.

H.           These effects are likely to matter more if the lower and upper houses represent constituencies with very different preferences.

I.             Why might this be so?  Cooter points out that since House members are elected by districts, it is mathematically conceivable that 25% of the population controls a majority of House seats.  (Just "pack" supporters of your opponents into as few districts as possible!)  To do the same for the Senate, you would have to pack states rather than districts. 

1.            This is the intuition behind "gerrymandering."

IV.          Presidential Veto the MWCs

A.           First imagine a situation with one legislative house and a president with absolute veto power.  ($E<$L) What happens?

1.            It looks just like the previous situation.

B.           Next, imagine adding a president with absolute veto power to a system with two houses. 

C.           Adding the president (with $E and $E0) tends to expand the Pareto set but shrink the bargain set.

D.           Last, imagine a situation with one house and a President with a 2/3 over-rideable veto.  This tends to expand the bargain set, because the President is no longer absolutely necessary for legislation to go through. 

E.           Just add the preferences of the 33rd percentile of the legislature to the diagram.  If these are closer to the median legislator than the President, they can strike a bargain without him.

V.           The Role of the Judiciary

A.           People frequently think of the judiciary as mechanically "interpreting" laws.  But judges' political views frequently seem to affect their decisions.

B.           But judges' "interpretations" can be overruled by further legislation.  So can judges really affect legislative outcomes?

C.           Yes, as long as they keep their ruling within the Pareto set. 

D.           Imagine, for example, that two houses of Congress finally agree to spend $X dollars, which lies somewhere between $L and $U.  But a dispute arises and a court gets to rule on the "true meaning" of the bill.

E.           Even if the judiciary ruled for $U exactly, the lower house would be unable to get the upper house to pass another bill "clarifying" the meaning of the first bill.

F.            Example: Both Houses of Congress and the President sign an anti-discrimination bill that bans any consideration of race.  Could the Court reinterpret this measure to allow discrimination against whites?  Yes, as long as ONE of the three governing bodies actually prefers the court's version.

VI.          The Logic of Divided Government

A.           Overall, then, division of powers tends to raise the transactions costs of passing legislation; it leads to so-called gridlock.

B.           If the main principal-agent problem is that agents pass too much legislation, this makes sense.

C.           But division of powers arguably reduces the quality of the legislation that does pass, because it is harder to assign credit or blame.  If the deficit skyrockets, who should you vote out?

D.           Also interesting: With divided government, one party rarely controls everything.  Voters could reduce "gridlock" by voting for unified government, but they often do the opposite!  "Ticket-splitting" is fairly common. 

1.            Note further: President's party normally loses seats in subsequent elections.  Once a President is in office, voters appear to deliberately make gridlock stronger!

E.           As usual, the way to get around gridlock is log-rolling.  The various branches of government can strike bargains with each other, either within bills or across bills.

VII.        Special Interests and the Median Voter Model

A.           The Median Voter Model makes no room for "special interests."  Voters get what they want.

B.           But observers have frequently argued that "special interests" - small groups of activists "behind-the-scenes" - somehow foil the majority's wishes.

1.            Industry groups

2.            "Public interest" groups

3.            Ethnic groups

4.            The bureaucracy

C.           Definition: A special interest is any organized group with preferences different from the median voter's.  The goal of a special interest group is to pull policy in their preferred direction.

D.           The distinction between mean and median preferences suggests the possibility that this is actually efficient, though few observers see it that way.

E.           If interest groups disagree, moreover, they may directly cancel each other out. 

F.            Similarly, if there is only "so much money to go around," in the budget interest groups may indirectly cancel each other out.  More money for schools means less for the environment, or defense, or something else.

G.           Leaving all this aside, how exactly could special interests exert influence anyway?  Doesn't political competition prevent it?

VIII.       Campaign Contributions, Political Advertising, and Rational Ignorance

A.           The most popular answer is campaign contributions.  Special interests move policy away from voter preferences by making donations to politicians.

B.           Cash donations are everyone's favorite example, but in-kind contributions of labor, printing, and so on are also important.  (Labor unions do a lot of the latter).

C.           Politicians then use these donations to "buy votes."  Political advertising is the standard example here.

D.           Why would political advertising work?  The usual answer appeals to voters' rational ignorance.  The ads are aimed at uninformed voters who are easy to sway.

E.           While this story sounds good, does it really make sense?  Why do uninformed voters pay any attention to ads, if they are typically funded by subversive special interests?

F.            Compare this to buying a used car.  Just because the buyer is ignorant, does he have to idiotically believe every word the seller tells him?

IX.          Concentrated Benefits and Diffuse Costs

A.           A more general story that builds on voters' rational ignorance goes by the name of "concentrated benefits, diffuse costs."

B.           Simple idea: Special interests are well-informed because they have so much riding on the political outcome.  Regular voters aren't informed because they have so little riding on it.

C.           Thus, special interests find it much easier to pressure politicians than voters.  They can say "Vote against us, and we'll withdraw all of our support.  And if you support us, the voters will never know."

D.           Suppose, for example, that voters are selfish, and 90% are uninformed.  Then the 10% of informed voters could in essence tell politicians: "We'll vote for whichever candidate promises us the most, not the candidate who will do the best job overall."

E.           Standard examples:

1.            Tariffs

2.            Subsidies

3.            Regulation

F.            This could work.  But critics doubt that organizing such a "grand conspiracy" would succeed.  Few voters are 100% uninformed; wouldn't information leak out?

X.           Bureaus and Budget Maximization

A.           Though it gets little Constitutional mention, many people have identified "the bureaucracy" as the fourth branch of government.

B.           Leading bureaucrats are not elected, but are rather appointed by elected officials.

C.           While most people focus on private "special interests," couldn't bureaucrats themselves be seen as a special interest?

D.           How might bureaucrats want to pull policy away from the median voter's choice?  One early public choice model suggested that bureaucrats are budget-maximizers.  Since bureau chiefs aren't elected and can't earn dividends, they engage in "empire building" instead.  In other words, bureaucrats are a "special interest" like any other, and are always trying to get more funding, staff, etc.

E.           Imagine graphing the net social benefits of an agency as a function of its size.  Then while the median voter might prefer to maximize net social benefits, a bureau chief would prefer to be to the right of that maximum.

F.            What prevents bureaus from expanding infinitely?  Presumably, if their conduct gets bad enough, elected officials can hem them in.  But the pressure for expansion is always there.

G.           The difficulty of actually removing bureau heads seems consistent with this sort of story.  But why is it hard to remove bureau heads in the first place?

XI.          Lobbying in Equilibrium: Rent-Seeking and Rent Dissipation

A.           Gordon Tullock's deep insight: lobbying (also known as "rent-seeking") is a competitive industry like any other.  If lobbying earns a 10% rate of return, and the standard rate is 5%, this will induce "new entry" into the lobbying "business."

B.           If the government awards monopoly privileges, for example, then firms will hire lawyers, political insiders, and so on to try to win the privilege. 

C.           Firms will keep entering this "arms race" until the net profits of the privilege are zero.  This happens when the total costs of lobbying equal the total value of the monopoly privilege!  This is known as "full rent dissipation."

D.           The government could award monopoly privileges by taking bids (or bribes) rather than listening to lobbyists.  But then, Tullock pointed out, this intensifies political competition; if people can get rich in politics, they will pay more to win a seat.

E.           This even works in a dictatorship or monarchy; if the dictator can get rich by awarding monopoly privileges, this strengthens the incentives of "upstarts" to try to seize the throne, stage a coup, etc.

F.            The big point: Creating opportunities to get rich by rent-seeking might literally benefit no one because it is so socially wasteful.  The rate of return on lobbying is driven down to the normal competitive level, leaving fewer resources for production.

G.           Fun fact: Some international studies have found that economic growth decreases as the lawyer:engineer ratio rises.

XII.        Application: Import-Substitution Industrialization (ISI)

A.           After WWII, many less-developed countries turned bitterly against free trade, arguing that developed countries were preventing them from industrializing.

B.           They turned instead to ISI policies - high tariffs, import quotas, industrial subsidies, etc.  The idea was to start producing their own industrial goods, substituting away from foreign imports.

C.           ISI policies were a disaster on many levels.  One of the biggest was the spread of "corruption," lobbying, etc.  In short, rent-seeking.

D.           Simple idea: Once the government was handing out political favors in mass quantities, domestic firms switched much of their energy from production to politics.  It became more important to have friends in government than to actually produce a low-cost, high-quality product.

E.           Sophisticated theoretical rationales for government involvement, in contrast, had little real-world influence.  These served as "window-dressing," but political "muscle" was the driving force behind policy.

F.            The result: Frequently, the new domestic monopolies were still unable to cover their costs.  The next step was to subsidize them to keep them afloat.  To do so, high taxes and other burdens were imposed on the shrinking set of productive industries.

G.           The key point: The system of political favoritism had its own internal logic.  Once a large fraction of wealth depended on government approval, there was a massive expansion in the "industry" of influencing government policy.