Political Agency (Endogenous Rents & Enforceability & Verifiability) Hard Flashcards
Political principal agent problem
Politicians (agents) exploit power by appropriating endogenous rents for themselves
(Against interests of principal - voters)
Are endogenous rents appropriated in MTV and PVT?
MVG: no, competed away as efficient equilibrium
PVM: inefficient equilibrium with positive endogenous rents (due to voters’ ideology, they would accept a bit of corruption if the candidate aligns with their ideology)
MVT: proving no rents extracted as efficient equilibrium algebra pg7
Voter i has income Yi and utility
Ui = Ci + ln(g)
(Utility from consumption and public good)
What is the individual and government BC
Individual BC
Ci = (1-t)Yi
Government BC
g = tybar
Find voter i’s optimal government spending (helps us find the level politicians will choose on next FC)
g*i = Ybar/Yi
Get there by subbing the individual BC, and the rearranged version of gov BC (t = g/Ybar) into the utility function
Then differentiating gets us g*i = Ybar/Yi , voter i’s preferred level of spending
So that was g* was for voter i. What will candaidate do tho
Propose the preferred g of the median voter
g*m = Ybar/Ym
As Ym increases, g*m falls (as median voter income increases, they want gov to spend less)
Now introduce endogenous rents to the model:
What does this look like in our model set up
Assume government revenue can also be used for rents (r) for politicians (baddd)
Include r in gov BS
g+r= tYbar
What do politicians also have to do now
Propose a 2 dimensional policy - propose level of spending (as before)
But also propose the rents r they want to extract
We assume they announce their rent-seeking, just to simplify. (Can think of it of announcing expenditure g, and tax level t)
Candidate if successful receives v = R + yr
R is salary
A) what does y represent?
B) Candidate A maximise their expected payoff; expression
y measures ease of appropriating resources
B)
Ea (Va) = pA (R +yrA)
Expected payoff = probability of winning (payoff)
So that was their expect payoff expression:
What does A’s expected payoff depend on
B) what is Pa value
Probability they win (Pa)
B) since we are looking at MVT, we know we get convergence in policies i.e ga=gb, thus probabilities of A and B are symmetrical equal chance of winning, Pa=Pb=1/2
So both A and B propose median voters preferred g (g*m) , but what is it?
After solving, we still get g~i = Ybar/Yi
Thus
g~m = g*m
Same case without rents as efficient equilibrium! No different when we introduce r to model
Important: So under MVT: gbarm=g*m, but what about r
B) reasoning
Both propose r=0 , NO ENDOGENOUS RENTS
B)
Since candidates are perfect subs! Probability of winning would change if deviate
Proves how MVT is efficient equilibrium where rents are compared away
Probabilistic voting model: inefficient equilibrium where positive rents can be extracted proof
We know policies converge, but what about when we add rents?
Candidate A maximises expected payoff; what do we get for r now?
Same working: Find A’s expected payoff:
Ea [Ua] = Pa (R+yra)
Differentiate with respect to rA to get:
0 = −Ψ (R+yrA) + 1/2y
Let Pa=1/2, and −Ψ comes from symmetric ideology DISTRIBUTIONS (does not mean symmetric ideology)
Rearrange to get r*a = 1/2Ψ - R/y >=0
i.e proof of positive rents!
Using equation r*a = 1/2Ψ - R/y >=0
When will rents be higher? (3)
If R (salary) is lower
If y (ease of appropriating resources) is higher (easy to steal)
If 1/Ψ (uncertainty about shock) is higher
Intuition why positive rents can be extracted
Candidates are no longer perfect substitutes for all voters, due to differences in the voters ideology preferences,
thus voter tolerate rent-seeking from candidates if their ideology is strong enough
Summary of MVT and PVM
Both predict policy convergence but different implications for endogenous rents
MVT - rents competed away r* = 0 since candidates are perfect subs (deviation will lose them the vote)
PVM - rents arise due to ideological preferences so voters tolerate rent seeking
So far we assume candidates stick to policy proposed.
Why may they not (2)
E.g Nick Clegg pledged not to rise tuition fee, but he tripled it.
May deviate from proposal to maximise rents
Changing costs may mean they are no longer able to fulfil proposed policy
Thus we need to model enforceability and verifiability….
Candidates make state-dependent proposals : depending on what
Pg 23
Cost of public good provision ø (high or low)
I.e If low cost (o low bar) then propose X, if high (O high bar) then propose Y
Scenario 1: Enforceability and verifiability:
What result do we get
Judiciary knows ø (cost of public good provision) and can hold politicians to state-specific policy (e.g if low cost, make sure they follow through with X)
Politicians propose policy to maximise probability of winning
This policy is optimal as otherwise probability of election goes to zero, so politicians get zero rents
So with enforceability and verifiability we get efficient equilibrium
Scenario 2: Enforceability (judges can hold them to promises) but no verifiability (we don’t know true cost)
Judiciary does not know true cost of public good provision, so cannnot make politician stick to state-specific policy
In election stage politicians propose policy to maximise probability of winning (same as first scenario)
However post-election can claim ø=øhigh bar (cost was high when it was actually low)
I.e can pocket different between high and low states
So inefficient equilibrium!
Scenario 3: What if no enforceability or verifiability
Judiciary cannot hold politicians to promises, nor know ø, meaning
Nothing to stop full exploitation - politicians can take everything, provide no public goods!
Empirical evidence on political principle agent: 2 key findings
Politicians behave differently depending on rents they can extract
Voters respond by punishing misbehavement
Does term length affect effort of legislators
Dal Bo and Rossi
Looked at 2 year vs 4 year terms
4 year term is better - more effort.
Why more effort? There are 3 counteracting effects
Accountability: work harder if prize is closer in reach. So longer term length decreases effort
Campaigning: re-election campaigns effort can crowd out effort in legislature (spend more time on campaign than actual legislative work): So longer term length increases effort (since less crowding out since are in office longer
Payback horizon; Larger rewards overtime, so longer term length increases effort
B)
Campaigning and payback horizon (2 positive effects) outweigh accountability (negative effect)
2nd finding: voters punish misbehaving
Ferraz and Finan
Voters were aware of mayors corruption before election
some were revealed corruption after election.
Found voters punish (reelection rates drop dramatically, esp for pre election audit (where voters were aware mayors were corrupt before)
