Rational Partisan Theory
As noted earlier, Hibbs (1977), in his partisan theory of macroeconomic
policy choices, argued that parties of the left will systematically select combinations
of unemployment and inflation which differ from those preferred
by right-wing parties. Following the rational expectations revolution, theoThe
rists questioned the ability of policy makers to influence real economic activity
using aggregate demand management policies. Alesina, in a series of
publications, has shown that the partisan theory of political business cycles
can survive in models incorporating rational expectations providing (i) voters
are uncertain about election outcomes, and (ii) uncontingent labour contracts
are signed for discrete periods and are not subject to renegotiation after the
election result is declared (see Alesina, 1987, 1988, 1989). In Alesina’s
model economic agents cannot enter into state-contingent nominal wage
contracts that provide insurance against electoral risk. Central to the rational
partisan theory is the idea that the political systems of many industrial democracies
are polarized. Alesina rejects the traditional view of politicians’
behaviour associated with Downs (1957) that vote-maximizing politicians in
a two-party system will generate a convergence of policies, as both parties
choose the policies favoured by the median voter (see also Minford and Peel,
1982). In the partisan theory politicians are ideological and adopt different
policies when in power. ‘There is no presumption that in a multi-party system
with self-interested politicians one should observe policy convergence’
(Alesina, 1989). In the case of the USA, empirical work has shown that
although the degree of polarization has varied throughout American history,
the Republican and Democratic parties have never fully converged (Alesina
and Rosenthal, 1995). This in part reflects the fact that a presidential candidate
has to appeal to the median voter in his/her own party in order to win the
nomination. ‘Since the platform adopted in the primary is a constraint on the
choice of platform for the presidential elections, even self-interested politicians
may have to choose polarized policies’ (Alesina, 1989). Alesina follows
Wittman (1977) and Hibbs (1977) and emphasizes the ideological preferences
of politicians who aim to please their supporters by implementing
policies which are likely to lead to a redistribution of income in their favour.
Hence the rational partisan theory shows how parties follow different macroeconomic
strategies because of their impact on the redistribution of income.
It is assumed that voters are well aware of these ideological differences
between the parties. In this framework macroeconomic policies create shortrun
aggregate disturbances because rational voters are uncertain about election
results. When Republicans or Conservatives are elected, economic agents are
confronted in the period immediately following the election with a deflationary
shock, that is, inflation is lower than expected. When Democratic and
Socialist governments are elected, the opposite occurs. There is a larger than
expected ‘inflation surprise’. It is the combination of election outcome uncertainty
combined with ideological differences between the two parties that
leads to aggregate instability in Alesina’s model. If a common macroeconomic
policy could be agreed between the two political parties aggregate
fluctuations would be reducen
The ‘eclectic’ macroeconomic framework adopted by Alesina is based on
Fischer’s (1977) well-known rational expectations model which includes a
labour market where nominal wage contracts are signed and extend for considerable
time periods (see Chapter 7). The ‘neutrality’ result or ‘policy
ineffectiveness proposition’ associated with Lucas (1972a) and Sargent and
Wallace (1975) depends on perfect (instantaneous) wage and price flexibility
as well as agents having rational expectations (see Chapter 5). Fischer (1977)
demonstrated that the crucial assumption for new classical results is instantaneous
market clearing. With nominal wage contracts, an element of price
stickiness is introduced into the model and in a non-market-clearing setting
policy effectiveness is restored with monetary policy having real effects on
aggregate output and employment. When ideological politicians use monetary
policy in such a setting a rational partisan business cycle is generated.
How does this come about? Consider equations (10.10)–(10.13) below.
yt Pt Wt yNt = β[ ˙ − ˙ ]+ , and β > 0 (10.10)
Ignoring capital accumulation, equation (10.10) shows that the rate of output
growth (yt) depends on the natural rate of growth of output (yNt) and positively
on the difference between the rate of inflation (P˙t ), and nominal wage
growth (W˙t ), that is, the path of real wages. (By Okun’s Law unemployment
and output are assumed to be inversely related and the natural rate of output
growth is that which is compatible with the natural rate of unemployment.)
Alesina uses output growth rather than the level of output since this is the
variable that empirical research uses to capture partisan effects on the economy.
If we assume that non-indexed nominal wage contracts lasting one period
(for example, two years) are drawn up with the objective of maintaining a
real wage consistent with natural output growth, we derive equation (10.11),
which shows that nominal wage growth is set equal to the current expected
rate of inflation (P˙t ) :
e
W˙ P˙ t t
= e (10.11)
Unlike the situation in the Nordhaus and Hibbs models, agents form their
expectations rationally. The rational expectations hypothesis is given in equation
(10.12):
P˙ E[P˙ | I ] t
e
= t t−1 (10.12)
Here E is the mathematical expectations operator and It–1 indicates the information
which has been accumulated by economic agents up to the end of
time period t–1. By combining (10.10) and (10.11) we arrive at equation
(10.13), which tells us that output growth will deviate from its natural rate if
there is an inflation surprise, that is, an unexpected change in monetary
policy:
yt Pt Pt y
e
Nt = β[ ˙ − ˙ ]+ (10.13)
In rational partisan models it is assumed that the incumbent politicians
have the ability to control the rate of inflation by monetary policy. Since party
preferences with respect to the extent of aversion to inflation differ, with the
right being assumed to be more averse than the left, an election which brings
about a change of government will lead to an inflation surprise, causing
output to deviate from its natural growth path. Agents have rational expectations
but are uncertain about forthcoming election results. Because agents
sign nominal wage contracts before the election result is known, the rate of
inflation forthcoming after the election can differ from the rational expectation
of inflation formed by wage negotiators in the pre-election period. Consider
the following possible sequence of events. Suppose the current administration
is a party of the left (Democrats in the USA, Labour in the UK).
Following Hibbs (1977), we can assume that left-wing governments have a
reputation for wanting to lower unemployment. Wage negotiators, if they
assume that the incumbents will win the election, will sign nominal wage
contracts which have built into them a high expected rate of inflation. Even if
the right-wing party looks like winning the election, risk-averse negotiators
will probably sign contracts involving inflation expectations higher than would
have been the case if agents knew for certain that the right would win. If a
Conservative (inflation-averse) government replaces the left-wing incumbents,
they will begin to tighten monetary policy in order to reduce inflation, thus
creating a surprise which has not been built into the wage contracts. As a
result, following an election victory by a Republican or Conservative party,
the Alesina rational partisan theory predicts a recession in output growth and
a rise in unemployment as inflation declines. The opposite sequence of events
would follow a change in government from right to left (Maloney et al.,
2003). In the period following an election, left-wing governments expand the
economy and reduce unemployment. Eventually, when inflationary expectations
adjust to the new situation, output growth returns to the natural rate but
the economy is locked into an equilibrium with high inflation. The election of
parties of the left will, according to Alesina’s model, be followed by a
cyclical pattern which is the opposite of that predicted by the Nordhaus
model. In both cases expectations are assumed to adjust to the actual rate of
inflation in the second half of a term of office, and from equation (10. 13) we
can see that output growth will settle back to its natural rate for both rightand
left-wing governments during this period of the electoral cycle. Since
there are no election surprises in the second half of an administration’s term
of office, real variables settle down to their natural rates. However, because
left-wing governments will be locked into a high-inflation equilibrium in
Alesina’s model, they may feel obliged to fight inflation before the next
election. President Carter found himself in such a situation in 1979–80.
We can now summarize the predictions of the rational partisan theory of
the business cycle:
A1 A change to a Conservative or Republican government will be followed
by a recession and rising unemployment. Once inflationary expectations
have been reduced, output growth returns to its natural rate. Inflation
is low as the next election approaches.
A2 A change to a Labour or Democratic government will be followed by
an acceleration of inflation as the economy expands more rapidly. Unemployment
will initially fall. Once inflationary expectations adjust,
output growth returns to its natural rate but inflation remains high.
Should a government of the left attempt to fight inflation in the run-up
to the next election, it will create a recession.
A3 The stronger the ideological convictions of the two parties, the greater
will be the disturbance to output and employment following a change of
policy regime after an election.
A4 Unlike the Hibbs model, the rational partisan theory predicts that differences
in unemployment and growth resulting from changes in government
will only be a temporary phenomenon
Empirical evidence for rational partisan cycles
A considerable amount of empirical work has been done in recent years to
test the rational partisan theory (see Alesina and Sachs, 1988; Alesina, 1989;
Alesina and Roubini, 1992; Alesina and Rosenthal, 1995). These studies have
found supporting evidence for temporary partisan effects on output and employment
and long-run partisan effects on the rate of inflation as predicted.
For the USA systematic differences have been found to occur in the first half,
but not the second half, of a large number of administrations. This evidence is
reported in Table 10.3. Although the rational partisan theory of Alesina and
the political business cycle model of Nordhaus (1975) give similar predictions
for Republican and Conservative administrations, the data in Table 10.3
do not show, on average, evidence of opportunistic behaviour. In line with the
predictions of the Alesina model, ‘Every Republican administration in the
post-World War II period except Reagan’s second one has started with a
recession. No recessions have occurred at the beginning of Democratic administrations’
(Alesina, 1995). The deep recessions in the US and UK
economies following the elections of President Reagan and Prime Minister
Thatcher conform very well to predictions A1 and A3 above. Prediction A2
fits very well the experience in France in the period 1981–3. During the early
part of the French Socialist administration of Mitterand, expansionary policies
were pursued initially, even though many other major economies were in
recession. Both President Mitterand and President Carter in the USA ended
their administrations trying to fight inflation, although it should be recognized
that the second OPEC oil-price shock complicates matters in the case
of the Carter administration.
Alesina concludes that the more recent rational versions of politico-economic
models of cycles have been much more successful empirically than the
earlier models of Nordhaus and Hibbs. In particular, ‘partisan effects’ appear
to be ‘quite strong’ while ‘opportunistic effects’ appear to be ‘small in magni544
Modern macroeconomics
tude’ and seem to affect only certain policy instruments, particularly fiscal
variables (see Alesina, 1995). It is also well documented that during the
Reagan and Thatcher administrations the income distribution consequences
of the micro and macro policies were ‘particularly partisan’. Inequality increased
under both administrations (see Alesina, 1989).
Criticisms of the rational partisan theory
There are a number of important weaknesses in the rational partisan theory.
First of all, if the cyclical effects are due to the signing of wage contracts
before an election, then one obvious solution is to delay the signing of
contracts until the election result is known. This solution is, of course, not as
applicable where the timing of elections is fixed endogenously. However, in
the USA wage contracts are staggered and overlapping, which means that at
least a significant proportion of wage contracts will inevitably go over the
election date. A second important criticism is that, in line with other models
which assume nominal wage rigidity, the Alesina model implies a countercyclical
real wage which is at odds with the stylized facts of the business
cycle. For theoretical purists a third criticism relates to the lack of firm
microeconomic foundations in such models to explain the mechanism of
nominal wage contracting. Alesina (1995) describes this as the ‘Achilles heel’
of the rational partisan theory. A fourth line of criticism originates from the
most recent generation of equilibrium business cycle theories. According to
real business cycle theorists, monetary policy cannot be used to produce real
effects on output and employment, although they agree that monetary growth
determines the rate of inflation. In real business cycle models aggregate
fluctuations are determined mainly by shocks to the production function, and
such shocks are endemic. The pre-electoral behaviour of politicians and postelectoral
monetary surprises are largely irrelevant. A benign monetary policy
would not bring to an end aggregate fluctuations (see Chapter 6). A fifth
criticism relates to hysteresis effects. If the natural rate properties of rational
partisan models do not hold due to persistence effects following an aggregate
demand disturbance, the political business cycle may be turned ‘upside down’
(see Gartner, 1996). A sixth criticism relates to the empirical evidence. In an
extensive survey Carmignani (2003) concludes that monetary policy is not
the source of political cycles in real variables (see also Drazen, 2000a, 2000b).
Finally, some theorists argue that partisan and opportunistic models are not
incompatible and a more complete model should incorporate both influences
(see Frey and Schneider, 1978a, 1978b; Schultz, 1995). It is to this latter
criticism that we now turn.
As noted earlier, Hibbs (1977), in his partisan theory of macroeconomic
policy choices, argued that parties of the left will systematically select combinations
of unemployment and inflation which differ from those preferred
by right-wing parties. Following the rational expectations revolution, theoThe
rists questioned the ability of policy makers to influence real economic activity
using aggregate demand management policies. Alesina, in a series of
publications, has shown that the partisan theory of political business cycles
can survive in models incorporating rational expectations providing (i) voters
are uncertain about election outcomes, and (ii) uncontingent labour contracts
are signed for discrete periods and are not subject to renegotiation after the
election result is declared (see Alesina, 1987, 1988, 1989). In Alesina’s
model economic agents cannot enter into state-contingent nominal wage
contracts that provide insurance against electoral risk. Central to the rational
partisan theory is the idea that the political systems of many industrial democracies
are polarized. Alesina rejects the traditional view of politicians’
behaviour associated with Downs (1957) that vote-maximizing politicians in
a two-party system will generate a convergence of policies, as both parties
choose the policies favoured by the median voter (see also Minford and Peel,
1982). In the partisan theory politicians are ideological and adopt different
policies when in power. ‘There is no presumption that in a multi-party system
with self-interested politicians one should observe policy convergence’
(Alesina, 1989). In the case of the USA, empirical work has shown that
although the degree of polarization has varied throughout American history,
the Republican and Democratic parties have never fully converged (Alesina
and Rosenthal, 1995). This in part reflects the fact that a presidential candidate
has to appeal to the median voter in his/her own party in order to win the
nomination. ‘Since the platform adopted in the primary is a constraint on the
choice of platform for the presidential elections, even self-interested politicians
may have to choose polarized policies’ (Alesina, 1989). Alesina follows
Wittman (1977) and Hibbs (1977) and emphasizes the ideological preferences
of politicians who aim to please their supporters by implementing
policies which are likely to lead to a redistribution of income in their favour.
Hence the rational partisan theory shows how parties follow different macroeconomic
strategies because of their impact on the redistribution of income.
It is assumed that voters are well aware of these ideological differences
between the parties. In this framework macroeconomic policies create shortrun
aggregate disturbances because rational voters are uncertain about election
results. When Republicans or Conservatives are elected, economic agents are
confronted in the period immediately following the election with a deflationary
shock, that is, inflation is lower than expected. When Democratic and
Socialist governments are elected, the opposite occurs. There is a larger than
expected ‘inflation surprise’. It is the combination of election outcome uncertainty
combined with ideological differences between the two parties that
leads to aggregate instability in Alesina’s model. If a common macroeconomic
policy could be agreed between the two political parties aggregate
fluctuations would be reducen
The ‘eclectic’ macroeconomic framework adopted by Alesina is based on
Fischer’s (1977) well-known rational expectations model which includes a
labour market where nominal wage contracts are signed and extend for considerable
time periods (see Chapter 7). The ‘neutrality’ result or ‘policy
ineffectiveness proposition’ associated with Lucas (1972a) and Sargent and
Wallace (1975) depends on perfect (instantaneous) wage and price flexibility
as well as agents having rational expectations (see Chapter 5). Fischer (1977)
demonstrated that the crucial assumption for new classical results is instantaneous
market clearing. With nominal wage contracts, an element of price
stickiness is introduced into the model and in a non-market-clearing setting
policy effectiveness is restored with monetary policy having real effects on
aggregate output and employment. When ideological politicians use monetary
policy in such a setting a rational partisan business cycle is generated.
How does this come about? Consider equations (10.10)–(10.13) below.
yt Pt Wt yNt = β[ ˙ − ˙ ]+ , and β > 0 (10.10)
Ignoring capital accumulation, equation (10.10) shows that the rate of output
growth (yt) depends on the natural rate of growth of output (yNt) and positively
on the difference between the rate of inflation (P˙t ), and nominal wage
growth (W˙t ), that is, the path of real wages. (By Okun’s Law unemployment
and output are assumed to be inversely related and the natural rate of output
growth is that which is compatible with the natural rate of unemployment.)
Alesina uses output growth rather than the level of output since this is the
variable that empirical research uses to capture partisan effects on the economy.
If we assume that non-indexed nominal wage contracts lasting one period
(for example, two years) are drawn up with the objective of maintaining a
real wage consistent with natural output growth, we derive equation (10.11),
which shows that nominal wage growth is set equal to the current expected
rate of inflation (P˙t ) :
e
W˙ P˙ t t
= e (10.11)
Unlike the situation in the Nordhaus and Hibbs models, agents form their
expectations rationally. The rational expectations hypothesis is given in equation
(10.12):
P˙ E[P˙ | I ] t
e
= t t−1 (10.12)
Here E is the mathematical expectations operator and It–1 indicates the information
which has been accumulated by economic agents up to the end of
time period t–1. By combining (10.10) and (10.11) we arrive at equation
(10.13), which tells us that output growth will deviate from its natural rate if
there is an inflation surprise, that is, an unexpected change in monetary
policy:
yt Pt Pt y
e
Nt = β[ ˙ − ˙ ]+ (10.13)
In rational partisan models it is assumed that the incumbent politicians
have the ability to control the rate of inflation by monetary policy. Since party
preferences with respect to the extent of aversion to inflation differ, with the
right being assumed to be more averse than the left, an election which brings
about a change of government will lead to an inflation surprise, causing
output to deviate from its natural growth path. Agents have rational expectations
but are uncertain about forthcoming election results. Because agents
sign nominal wage contracts before the election result is known, the rate of
inflation forthcoming after the election can differ from the rational expectation
of inflation formed by wage negotiators in the pre-election period. Consider
the following possible sequence of events. Suppose the current administration
is a party of the left (Democrats in the USA, Labour in the UK).
Following Hibbs (1977), we can assume that left-wing governments have a
reputation for wanting to lower unemployment. Wage negotiators, if they
assume that the incumbents will win the election, will sign nominal wage
contracts which have built into them a high expected rate of inflation. Even if
the right-wing party looks like winning the election, risk-averse negotiators
will probably sign contracts involving inflation expectations higher than would
have been the case if agents knew for certain that the right would win. If a
Conservative (inflation-averse) government replaces the left-wing incumbents,
they will begin to tighten monetary policy in order to reduce inflation, thus
creating a surprise which has not been built into the wage contracts. As a
result, following an election victory by a Republican or Conservative party,
the Alesina rational partisan theory predicts a recession in output growth and
a rise in unemployment as inflation declines. The opposite sequence of events
would follow a change in government from right to left (Maloney et al.,
2003). In the period following an election, left-wing governments expand the
economy and reduce unemployment. Eventually, when inflationary expectations
adjust to the new situation, output growth returns to the natural rate but
the economy is locked into an equilibrium with high inflation. The election of
parties of the left will, according to Alesina’s model, be followed by a
cyclical pattern which is the opposite of that predicted by the Nordhaus
model. In both cases expectations are assumed to adjust to the actual rate of
inflation in the second half of a term of office, and from equation (10. 13) we
can see that output growth will settle back to its natural rate for both rightand
left-wing governments during this period of the electoral cycle. Since
there are no election surprises in the second half of an administration’s term
of office, real variables settle down to their natural rates. However, because
left-wing governments will be locked into a high-inflation equilibrium in
Alesina’s model, they may feel obliged to fight inflation before the next
election. President Carter found himself in such a situation in 1979–80.
We can now summarize the predictions of the rational partisan theory of
the business cycle:
A1 A change to a Conservative or Republican government will be followed
by a recession and rising unemployment. Once inflationary expectations
have been reduced, output growth returns to its natural rate. Inflation
is low as the next election approaches.
A2 A change to a Labour or Democratic government will be followed by
an acceleration of inflation as the economy expands more rapidly. Unemployment
will initially fall. Once inflationary expectations adjust,
output growth returns to its natural rate but inflation remains high.
Should a government of the left attempt to fight inflation in the run-up
to the next election, it will create a recession.
A3 The stronger the ideological convictions of the two parties, the greater
will be the disturbance to output and employment following a change of
policy regime after an election.
A4 Unlike the Hibbs model, the rational partisan theory predicts that differences
in unemployment and growth resulting from changes in government
will only be a temporary phenomenon
Empirical evidence for rational partisan cycles
A considerable amount of empirical work has been done in recent years to
test the rational partisan theory (see Alesina and Sachs, 1988; Alesina, 1989;
Alesina and Roubini, 1992; Alesina and Rosenthal, 1995). These studies have
found supporting evidence for temporary partisan effects on output and employment
and long-run partisan effects on the rate of inflation as predicted.
For the USA systematic differences have been found to occur in the first half,
but not the second half, of a large number of administrations. This evidence is
reported in Table 10.3. Although the rational partisan theory of Alesina and
the political business cycle model of Nordhaus (1975) give similar predictions
for Republican and Conservative administrations, the data in Table 10.3
do not show, on average, evidence of opportunistic behaviour. In line with the
predictions of the Alesina model, ‘Every Republican administration in the
post-World War II period except Reagan’s second one has started with a
recession. No recessions have occurred at the beginning of Democratic administrations’
(Alesina, 1995). The deep recessions in the US and UK
economies following the elections of President Reagan and Prime Minister
Thatcher conform very well to predictions A1 and A3 above. Prediction A2
fits very well the experience in France in the period 1981–3. During the early
part of the French Socialist administration of Mitterand, expansionary policies
were pursued initially, even though many other major economies were in
recession. Both President Mitterand and President Carter in the USA ended
their administrations trying to fight inflation, although it should be recognized
that the second OPEC oil-price shock complicates matters in the case
of the Carter administration.
Alesina concludes that the more recent rational versions of politico-economic
models of cycles have been much more successful empirically than the
earlier models of Nordhaus and Hibbs. In particular, ‘partisan effects’ appear
to be ‘quite strong’ while ‘opportunistic effects’ appear to be ‘small in magni544
Modern macroeconomics
tude’ and seem to affect only certain policy instruments, particularly fiscal
variables (see Alesina, 1995). It is also well documented that during the
Reagan and Thatcher administrations the income distribution consequences
of the micro and macro policies were ‘particularly partisan’. Inequality increased
under both administrations (see Alesina, 1989).
Criticisms of the rational partisan theory
There are a number of important weaknesses in the rational partisan theory.
First of all, if the cyclical effects are due to the signing of wage contracts
before an election, then one obvious solution is to delay the signing of
contracts until the election result is known. This solution is, of course, not as
applicable where the timing of elections is fixed endogenously. However, in
the USA wage contracts are staggered and overlapping, which means that at
least a significant proportion of wage contracts will inevitably go over the
election date. A second important criticism is that, in line with other models
which assume nominal wage rigidity, the Alesina model implies a countercyclical
real wage which is at odds with the stylized facts of the business
cycle. For theoretical purists a third criticism relates to the lack of firm
microeconomic foundations in such models to explain the mechanism of
nominal wage contracting. Alesina (1995) describes this as the ‘Achilles heel’
of the rational partisan theory. A fourth line of criticism originates from the
most recent generation of equilibrium business cycle theories. According to
real business cycle theorists, monetary policy cannot be used to produce real
effects on output and employment, although they agree that monetary growth
determines the rate of inflation. In real business cycle models aggregate
fluctuations are determined mainly by shocks to the production function, and
such shocks are endemic. The pre-electoral behaviour of politicians and postelectoral
monetary surprises are largely irrelevant. A benign monetary policy
would not bring to an end aggregate fluctuations (see Chapter 6). A fifth
criticism relates to hysteresis effects. If the natural rate properties of rational
partisan models do not hold due to persistence effects following an aggregate
demand disturbance, the political business cycle may be turned ‘upside down’
(see Gartner, 1996). A sixth criticism relates to the empirical evidence. In an
extensive survey Carmignani (2003) concludes that monetary policy is not
the source of political cycles in real variables (see also Drazen, 2000a, 2000b).
Finally, some theorists argue that partisan and opportunistic models are not
incompatible and a more complete model should incorporate both influences
(see Frey and Schneider, 1978a, 1978b; Schultz, 1995). It is to this latter
criticism that we now turn.
No comments:
Post a Comment