Financial Crises and Political Crises

1
Financial Crises and Political Crises

• Roberto Chang
• Rutgers University
• September 2002

2
Introduction and Motivation

• Often, but not always, financial crises are
  associated with political collapse
• Current examples Argentina, Brazil
• Why are some crises and not others associated
  with political turmoil?
• Do political distortions always matter?
• What are the implications for policy?

3
Summary of the Paper

• A simple model of a country with a debt default
  problem
• Political distortions ? large debt and low
  reserves lead to default and political crises
• But the debt problem itself depends on the
  political outlook
• Market expectations may be self fulfilling and
  lead to default and to government collapse

4
Key Information Transmission

• The government makes default decision with better
  information than the public about the social cost
  of default
• To control the governments behavior, the
  representative agent can fire the government at
  the cost of a political crisis.
• For this to happen, there must be a political
  distortion, in that the policymakers objectives
  may differ from those of the public.

5
Main Results

• Some crises are only financial, while others
  are associated with political debacle.
• Which ones obtain depend on inherited debt and on
  the stock of reserves in intuitive ways
• There may be multiple equilibria financial
  fragility may lead to political collapse.

6

• If there are multiple equilibria, an
  international loan guarantee can select the best
  one at zero expected cost
• However, if a political crisis occurs with
  positive probability, such a guarantee must be
  made conditional on the political crisis

7
The Political Setting

• Small open economy with a representative agent
  and a policymaker
• One period and one good that costs one dollar in
  world market
• D amount due for repayment at end of period
• R lt D dollar reserves
• Debt repayment requires collecting a tax X D
  R, with deadweight loss ?(X).

8
Default Benefits and Costs

• The value of default is twofold the tax does not
  have to be paid and reserves are kept. Hence V
  R X ?(X)
• There is a direct cost of default ?, which takes
  values ?H and ?L with probabilities q and (1-q)
• So, if the representative agent chose whether to
  default after observing ?, the decision would be
  straightforward

9
Introducing a political dimension

• The default decision is made by the policymaker
  on behalf of the rep.agent
• The representative agent can dismiss the
  policymaker and overrule her decision at some
  cost ? this is a political crisis
• Only the policymaker observes ? freely.

10
Sequence of events

• Policymaker observes the realization of ? and
  proposes default or repayment
• Representative agent chooses whether to fire the
  policymaker or not
• If policymaker is retained, her proposal is
  implemented
• If not, representative agent learns ?, and
  chooses whether or not to default

11
Equilibria Without Political Distortions

• Focus on Perfect Bayesian Equilibria
• If the governments objectives coincide with
  those of the representative agent, the outcome is
  the same as if the latter chose policy directly,
  and the government is never fired
• Financial crises occur but political ones do not

12
Equilibria With Political Distortions

• Assume the policymaker does not only care about
  social welfare but may also suffer a personal
  cost if she proposes and implements a default
• The personal cost takes values 0 or ??, with
  probs. p and (1-p), and is private info.
• Interpretation career concerns or heterogeneity
  and delegation problems.

13

• Assume
• (1?) ?L ? ?H
• The previous outcome is ruled out, as the self
  interested government would repay the debt even
  if the social cost of default is low
• There are five kinds of equilibria, depending on V

14
PBE Type i Neither default nor political crises

• If V R X ?(X) ? ?L , the representative
  agent and both government types prefer to avoid
  default
• So the self interested government, in particular,
  does not gain from lying
• The representative agent cannot gain from
  dismissing the government

15
PBE Type ii Only financial crises

• The policymaker proposes to default only if she
  is benevolent and ? ?L .
• The representative agent chooses not to dismiss
  the policymaker
• Analysis if the policymaker proposes default,
  the rep. agent infers ? ?L with prob. one. So
  there is no gain from firing the government

16

• The inference is more difficult after a proposal
  not to default the cost of retaining the
  policymaker is known to be X ?(X), but the cost
  of firing her depends on the rep. agents beliefs
  conditional on the default proposal

17
From Bayes Rule,
18

• Hence dismissal costs ? for sure plus z(X ?(X))
  (1-z)(?L R)
• Accepting the policymakers proposal is optimal
  if
• V ? ?L ?/(1-z)
• Intuition for role of z a political crisis is
  only worth it if it leads to correcting a wrong
  outcome.

19

• So, PBE type ii exists if
• ?L lt V ? ?L ?/(1-z)
• Only financial crises occur
• But political distortions do matter nevertheless
  there is too little default in this equilibrium.

20
PBE Type iii Optimal Default with Political
Crises

• If V gt ?L ?/(1-z), there may be a PBE in which
  the policymakers strategy is the same as in PBE
  type ii, but she is fired unless she proposes to
  default.
• Here, if the social cost of default is high, the
  benevolent policymaker proposes repayment and is
  sacked, after which the debt is repaid anyway.

21

• For this to be optimal for the benevolent
  policymaker, the cost of a political crisis
  cannot be too large relative to the cost of
  default
• V ? ?H - ?
• Proposing repayment when the cost of default is
  low is optimal for the self interested
  policymaker assuming ??L gt ?

22

• First instance of a political crisis in
  equilibrium
• Default occurs when socially desirable, but
  costly government dismissal occurs with positive
  probability
• The political distortion makes the good
  government unable to convince the population that
  repayment is the best option (Argentina?)

23
PBE Type iv Excessive default and political
crisis

• The benevolent government proposes default
  regardless of ?, which is accepted
• The self interested government proposes default
  and is fired. The rep. agent then defaults if the
  social cost if low.
• Here, if ? ?L , the self interested government
  chooses to be fired than to be associated with a
  default.

24
PBE v Total collapse

• If V gt ?H , it is known that default is socially
  optimal
• Again, however, the self interested government
  chooses to propose repayment and be fired, rather
  than to be associated with default
• Default occurs for sure, a political crisis with
  prob. (1-p).

25
Political Stage Summary
26

• The nature of equilibria depend on V and, hence,
  on D and R in natural ways
• Interestingly, while the probability of default
  increases with V, the probability of a political
  crisis is not monotonic
• This is because a political crisis has little
  value if the economic fundamentals are dismal

27
Liquidity Crises and Political Crises

• Now embed the model as the last, political
  stage of a two stage investment problem
• Suppose that, at the beginning of the period, the
  economy has an investment opportunity that costs
  I dollars then and returns R at the end of the
  period.

28

• Assume that the investment entails enough non
  pecuniary benefits
• The government must then borrow R from foreign
  risk neutral investors
• To do so, the government sells claims to D1
  dollars payable at the end of the period.
• If ? is the prob. of repayment, D1 I/?

29

• After the debt auction, the model is as before,
  with D D0 D1 , where D0 denotes prior debt.
• A rational expectations equilibrium is a pair (D,
  ?) such that D D0 I/? and, given D, ? is
  given by the PBE of the political stage

30
No default equilibria

• ? must be one
• Then D D0 I
• For this to be an equilibrium, D must be
  consistent with PBE Type i
• V D0 I ?(D0 I R) ? ?L
• Intuitively, this is more likely if initial debt
  is small and R is large

31
Equilibria with only financial crisis

• The political stage outcome is PBE Type ii
• ? must be 1 p(1-q)
• It must be that
• ?L lt V D0 I/1 p(1-q)
• ?(D0 I/ 1 p(1-q) R)
• ? ?L ?/(1-z)

32

• Bad fundamentals increase the likelihood of a
  financial crisis
• These equilibria may coexist with the no default
  equilibria
• These results are consistent with previous
  literature

33
Equilibria with Political Crises

• The political stage outcome is PBE Type iii
• Then ? must equal q, and equilibrium requires
• ?L ?/(1-z) lt
• V D0 I/q ?(D0 I/ q R)
• ? ?H - ?

34

• High D0 , large I, or low R now lead not only to
  default, but also to political crises
• More importantly, these equilibria may coexist
  with equilibria of the previous kind
• Lenders expectations may then be self fulfilling
  and result in political collapse

35
Welfare and Policy Implications

• Suppose that a no crisis equilibrium coexists
  with an equilibrium with only financial crisis.
• Then a loan guarantee provided by the
  international community eliminates the bad
  equilibrium
• And, in the no crisis equilibrium, the guarantee
  is never activated, so the policy has zero
  expected cost

36

• Now, suppose that there is an equilibrium with
  only financial crisis, and an equilibrium with
  both financial and political crisis
• Then, the loan guarantee may eliminate the last
  equilibrium
• But there is an expected loss to the guarantor,
  since the guarantee is invoked in a default,
  which occurs with positive probability.

37

• A superior alternative is a guarantee that the
  international community will pay the debt if
  there is default and a political crisis
• Such a limited guarantee eliminates the
  equilibrium with political crisis
• And it has zero cost to the international
  community, since it is not needed in the
  remaining equilibrium

38

• Policy design is, therefore, delicate
• Also, if equilibrium is unique, loan guarantees
  are not as helpful

39
Final Remarks

• Larger point by neglecting the possible
  connections between financial crises and
  political crises, the international community may
  be missing a golden opportunity to help debtor
  countries at minimal cost
• The model suggests that some reforms to deal with
  the information transmission problem
  (transparency) do contribute to both financial
  stability and political stability

40

• This paper shows that we must study the links
  between financial issues and political
  distortions
• Consistent with some empirical literature
• Many questions remain in the air (example role
  of elections)