Politics, Time Inconsistency, Credibility and Reputation
Following the incorporation of the rational expectations hypothesis into macroeconomic
models, the theoretical literature on economic policy has been
dominated by the game-theoretic approach. Policy makers are seen to be
engaged in a complicated dynamic game with private sector economic agents
(who are also voters). This literature was stimulated by the seminal paper of
Kydland and Prescott (1977), who raised the general problem of time inconsistency
of policy (see Chapter 5). According to Kydland and Prescott,
governments that are free from rules (pre-commitment) so that they can use
discretionary policies will be unable to persuade rational agents that they will
stick to low-inflation policies. Agents know that if they lower their inflation
expectations the government will have an incentive to cheat and, by creating
an inflation surprise, increase employment temporarily. However, because
rational agents are aware of the policy makers’ incentives, the time-consistent
policy involves an inflationary bias. If a government has discretion, lowinflation
declarations are time-inconsistent and are not credible. Therefore a
credible policy announcement can be defined as one which is time-consistent.
Solutions to the time-inconsistency problem include contractual arrangements,
delegation of decisions and institutional and legal constraints (see
Drazen, 2000a).
As we have already seen, in industrial democracies subject to regular
democratic elections, politicians have an incentive to deviate from optimal
Table 10.4 Alternative politico-economic models
Politico-economic model Main assumptions Predictions
All governments, both left and right, behave the same.
Output will increase and unemployment will fall before an
election. Inflation will accelerate as the election approaches
but will peak and be observed by voters after the election.
Governments of different ideological persuasions will have
different macroeconomic objectives with respect to
inflation and unemployment. Governments of the right will
tend to have persistently higher unemployment and less
inflation than governments of the left.
All governments behave the same.
Monetary growth and fiscal expansions before elections.
Left-of-centre governments produce an inflation bias
compared to right-of-centre governments. Output is above
(below) the natural rate at the beginning of a left- (right-)
of-centre government. The partisan effect of monetary
policy on real variables is temporary.
If incumbent’s popularity is in deficit (below some critical
level) prior to an election, the incumbent switches from
partisan to opportunistic behaviour.
Incumbents with a popularity surplus behave ideologically.
Non-rational
opportunistic business
cycle theory (Nordhaus,
1975)
Strong partisan theory
(Hibbs, 1977)
Rational opportunistic
theory (Rogoff and
Sibert, 1988)
Rational partisan theory
(Alesina, 1987)
Weak partisan theory.
Opportunistic–partisan
synthesis. (Frey and
Schneider, 1978a)
Expectations-augmented Phillips curve. Politicians only
care about re-election. Agents are myopic and have nonrational
expectations.
Exploitable Phillips curve trade-off. Policy makers and
voters are ideological and non-rational. Left-of-centre
parties have a strong aversion to unemployment relative to
inflation. Right-of-centre parties have a strong aversion to
inflation relative to unemployment.
Short-run Phillips curve trade-off. Agents have rational
expectations, but imperfect information.
Voters elect the party that they expect to perform best.
Politicians care only about re-election.
Short-run Phillips curve trade-off. Agents have rational
expectations, but election outcome uncertain.
Left-of-centre parties have a strong aversion to
unemployment relative to inflation. Right-of-centre parties
have a strong aversion to inflation relative to
unemployment.
Politicians alternate between partisan and opportunistic
behaviour. Actual behaviour of politicians depends on their
‘popularity surplus’ in the polls.
Source: Adapted from Edwards (1994).
policies and create an inflation surprise. In the Nordhaus model there is an
incentive to expand the economy before an election in order to gain votes.
This is possible in the Nordhaus model because the policy makers never lose
credibility due to the assumption of non-rational economic agents and myopic
voters. In models with rational expectations and forward-looking voters,
policy makers are under pressure to establish their credibility and reputation
(see Blackburn and Christensen, 1989). By reputation in this context economists
are referring to ‘the actions that policymakers are expected to take’
(Drazen, 2000a). In a game-theoretic context the reputation of a player will
depend on the way they have played and reacted to events in the past.
Rational agents will only believe politicians who make ex ante policy announcements
which are also optimal to implement ex post. However, rational
agents have imperfect information about the real motives of politicians as
opposed to their pre-election promises. Therefore private sector agents will
need to analyse carefully the various signals that politicians give out. In this
scenario it may be difficult for voters to distinguish ‘hard-nosed’ (inflationaverse)
from ‘wet’ (inflation-prone) politicians since the latter will always
have an incentive to masquerade as ‘hard-nosed’ (see Backus and Driffill,
1985).
Alesina (1987) has shown that the prediction of the median voter theorem,
that in a two-party system there will be policy convergence, is time-inconsistent.
In the period before an election both parties find it in their interest to
announce convergent policies on the assumption that this will appeal to the
median voter. Ideological issues take back stage in order to maximize reelection
prospects. However, because there is no mechanism for holding an
elected government to its promises, these announced convergent policies
must be time-inconsistent. After the election the influence of partisan considerations
will predominate as the elected politicians re-optimize and follow a
programme which best fits their ideological stance. Thus the time-consistent
equilibrium involves no policy convergence and the two parties ‘follow their
most preferred policy’ (Alesina and Tabellini, 1988). This inevitably creates
too much volatility in policy making which, in turn, causes politically induced
business cycles.
It follows from the above analysis that only those pre-election announcements
and promises which are consistent with a party’s ideology should be
taken seriously by voters. Once elected, politicians will tend to follow a more
partisan strategy. This may prove to be a particular problem for parties of the
left that declare themselves to be ‘tough on inflation’. Given the underlying
assumptions of the Hibbs (1977) and Alesina (1987) models that politicians
on the left give a high priority to reducing unemployment, voters may well
look with suspicion at pronouncements claiming an aversion to inflation.
Inflation-prone parties have an incentive to masquerade as ‘hard-nosed’ on
this objective. Rational voters are likely to interpret such signalling as the
‘dissembling actions of an impostor’ (Blackburn, 1992). An implication of
this is that ‘a low inflation policy announced by a government concerned with
unemployment would not be credible; indeed if expected inflation was low,
this government would create an inflation surprise in order to reduce unemployment’
(Alesina, 1989).
These issues were very pertinent in the UK in the run-up to the 1997
election. The ‘New’ Labour Party, led by Tony Blair, declared that it intended
to be ‘tough on inflation’ and that it also aimed to achieve much lower
unemployment. In terms of Alesina’s model, the statement on inflation is
clearly time-inconsistent. However, to give credibility to its anti-inflation
rhetoric, on winning the 1997 election, new Labour immediately granted
operational independence to the Bank of England (see Snowdon, 1997).
Following the incorporation of the rational expectations hypothesis into macroeconomic
models, the theoretical literature on economic policy has been
dominated by the game-theoretic approach. Policy makers are seen to be
engaged in a complicated dynamic game with private sector economic agents
(who are also voters). This literature was stimulated by the seminal paper of
Kydland and Prescott (1977), who raised the general problem of time inconsistency
of policy (see Chapter 5). According to Kydland and Prescott,
governments that are free from rules (pre-commitment) so that they can use
discretionary policies will be unable to persuade rational agents that they will
stick to low-inflation policies. Agents know that if they lower their inflation
expectations the government will have an incentive to cheat and, by creating
an inflation surprise, increase employment temporarily. However, because
rational agents are aware of the policy makers’ incentives, the time-consistent
policy involves an inflationary bias. If a government has discretion, lowinflation
declarations are time-inconsistent and are not credible. Therefore a
credible policy announcement can be defined as one which is time-consistent.
Solutions to the time-inconsistency problem include contractual arrangements,
delegation of decisions and institutional and legal constraints (see
Drazen, 2000a).
As we have already seen, in industrial democracies subject to regular
democratic elections, politicians have an incentive to deviate from optimal
Table 10.4 Alternative politico-economic models
Politico-economic model Main assumptions Predictions
All governments, both left and right, behave the same.
Output will increase and unemployment will fall before an
election. Inflation will accelerate as the election approaches
but will peak and be observed by voters after the election.
Governments of different ideological persuasions will have
different macroeconomic objectives with respect to
inflation and unemployment. Governments of the right will
tend to have persistently higher unemployment and less
inflation than governments of the left.
All governments behave the same.
Monetary growth and fiscal expansions before elections.
Left-of-centre governments produce an inflation bias
compared to right-of-centre governments. Output is above
(below) the natural rate at the beginning of a left- (right-)
of-centre government. The partisan effect of monetary
policy on real variables is temporary.
If incumbent’s popularity is in deficit (below some critical
level) prior to an election, the incumbent switches from
partisan to opportunistic behaviour.
Incumbents with a popularity surplus behave ideologically.
Non-rational
opportunistic business
cycle theory (Nordhaus,
1975)
Strong partisan theory
(Hibbs, 1977)
Rational opportunistic
theory (Rogoff and
Sibert, 1988)
Rational partisan theory
(Alesina, 1987)
Weak partisan theory.
Opportunistic–partisan
synthesis. (Frey and
Schneider, 1978a)
Expectations-augmented Phillips curve. Politicians only
care about re-election. Agents are myopic and have nonrational
expectations.
Exploitable Phillips curve trade-off. Policy makers and
voters are ideological and non-rational. Left-of-centre
parties have a strong aversion to unemployment relative to
inflation. Right-of-centre parties have a strong aversion to
inflation relative to unemployment.
Short-run Phillips curve trade-off. Agents have rational
expectations, but imperfect information.
Voters elect the party that they expect to perform best.
Politicians care only about re-election.
Short-run Phillips curve trade-off. Agents have rational
expectations, but election outcome uncertain.
Left-of-centre parties have a strong aversion to
unemployment relative to inflation. Right-of-centre parties
have a strong aversion to inflation relative to
unemployment.
Politicians alternate between partisan and opportunistic
behaviour. Actual behaviour of politicians depends on their
‘popularity surplus’ in the polls.
Source: Adapted from Edwards (1994).
policies and create an inflation surprise. In the Nordhaus model there is an
incentive to expand the economy before an election in order to gain votes.
This is possible in the Nordhaus model because the policy makers never lose
credibility due to the assumption of non-rational economic agents and myopic
voters. In models with rational expectations and forward-looking voters,
policy makers are under pressure to establish their credibility and reputation
(see Blackburn and Christensen, 1989). By reputation in this context economists
are referring to ‘the actions that policymakers are expected to take’
(Drazen, 2000a). In a game-theoretic context the reputation of a player will
depend on the way they have played and reacted to events in the past.
Rational agents will only believe politicians who make ex ante policy announcements
which are also optimal to implement ex post. However, rational
agents have imperfect information about the real motives of politicians as
opposed to their pre-election promises. Therefore private sector agents will
need to analyse carefully the various signals that politicians give out. In this
scenario it may be difficult for voters to distinguish ‘hard-nosed’ (inflationaverse)
from ‘wet’ (inflation-prone) politicians since the latter will always
have an incentive to masquerade as ‘hard-nosed’ (see Backus and Driffill,
1985).
Alesina (1987) has shown that the prediction of the median voter theorem,
that in a two-party system there will be policy convergence, is time-inconsistent.
In the period before an election both parties find it in their interest to
announce convergent policies on the assumption that this will appeal to the
median voter. Ideological issues take back stage in order to maximize reelection
prospects. However, because there is no mechanism for holding an
elected government to its promises, these announced convergent policies
must be time-inconsistent. After the election the influence of partisan considerations
will predominate as the elected politicians re-optimize and follow a
programme which best fits their ideological stance. Thus the time-consistent
equilibrium involves no policy convergence and the two parties ‘follow their
most preferred policy’ (Alesina and Tabellini, 1988). This inevitably creates
too much volatility in policy making which, in turn, causes politically induced
business cycles.
It follows from the above analysis that only those pre-election announcements
and promises which are consistent with a party’s ideology should be
taken seriously by voters. Once elected, politicians will tend to follow a more
partisan strategy. This may prove to be a particular problem for parties of the
left that declare themselves to be ‘tough on inflation’. Given the underlying
assumptions of the Hibbs (1977) and Alesina (1987) models that politicians
on the left give a high priority to reducing unemployment, voters may well
look with suspicion at pronouncements claiming an aversion to inflation.
Inflation-prone parties have an incentive to masquerade as ‘hard-nosed’ on
this objective. Rational voters are likely to interpret such signalling as the
‘dissembling actions of an impostor’ (Blackburn, 1992). An implication of
this is that ‘a low inflation policy announced by a government concerned with
unemployment would not be credible; indeed if expected inflation was low,
this government would create an inflation surprise in order to reduce unemployment’
(Alesina, 1989).
These issues were very pertinent in the UK in the run-up to the 1997
election. The ‘New’ Labour Party, led by Tony Blair, declared that it intended
to be ‘tough on inflation’ and that it also aimed to achieve much lower
unemployment. In terms of Alesina’s model, the statement on inflation is
clearly time-inconsistent. However, to give credibility to its anti-inflation
rhetoric, on winning the 1997 election, new Labour immediately granted
operational independence to the Bank of England (see Snowdon, 1997).
No comments:
Post a Comment