Underemployment Equilibrium in the Keynesian Model
The general case
Within the IS–LM model the existence of underemployment equilibrium can
be attributed to the existence of ‘rigidities’ in the system, especially two key
prices, the money wage and the interest rate. We begin with that of the
‘Keynesian’ assumption of downward rigidity in money wages. This case can
be illustrated using the four-quadrant diagram of Figure 3.6. Quadrant (a)
depicts the standard IS–LM model. Quadrant (c) shows the short-run production
function where, with the capital stock and technology taken as given, the
level of output/income (Y) depends on the level of employment (L) – see
Chapter 2, section 2.3. Quadrant (d) depicts the labour market in which it is
assumed that the demand for/supply of labour is negatively/positively related
to real wages (W/P). Finally, quadrant (b) shows, via a 45° line, equality
between the two axes, both of which depict income. The inclusion of this
quadrant allows us to see more easily the implications of a particular equilibrium
level of income, established in the goods and money markets in quadrant
(a), for the level of employment shown in quadrant (d). In other words, in
what follows the reader should always start in quadrant (a) and move in an
anti-clockwise direction to trace the implications of the level of income
(determined by aggregate demand) in terms of the level of employment in
quadrant (d).
Suppose the economy is initially at point E0, that is, the intersection of LM0
and IS in quadrant (a). While both the goods and money markets are in
equilibrium, the income level of Y0 is below the full employment income
level YF. Reference to quadrant (d) reveals that with a fixed money wage (set
exogenously) and a price level consistent with equilibrium in the money
market (that is, the curve LM0), the resultant level of real wages (W/P)0 is
inconsistent with the labour market clearing. In other words there is no
guarantee that the demand-determined level of employment (L0) will be at
full employment (LF). The excess supply of labour has no effect on the
money wage, so that it is possible for the economy to remain at less than full
employment equilibrium with persistent unemployment. We now consider
what effect combining the IS–LM model with the classical assumption of
flexible prices and money wages has on the theoretical possibility of underemployment
equilibrium.
Again suppose the economy is initially at point E0, that is, the intersection
of IS and LM0 in quadrant (a). As before, while both the goods and money
markets are in equilibrium, the income level of Y0 is below the full employment
income level YF. Reference to quadrant (d) reveals that this implies that
the level of employment (L0) is below its full employment level (LF) with real
wages (W/P)0 above their market-clearing level (W/P)1. As long as prices and
money wages are perfectly flexible, the macroeconomy will, however, selfequilibrate
at full employment. At (W/P)0 the excess supply of labour results
in a fall in money wages (W), which reduces firms’ costs and causes a fall in
prices (P). The fall in prices increases the real value of the money supply,
causing the LM curve to shift downwards to the right. Excess real balances
are channelled into the bond market where bond prices are bid up and the rate
of interest is bid down. The resultant fall in the rate of interest in turn
stimulates investment expenditure, increasing the level of aggregate demand
and therefore output and employment. The ‘indirect’ effect of falling money
wages and prices which stimulates spending via the interest rate is referred to
as the ‘Keynes effect’. The increase in aggregate demand moderates the rate
of fall in prices so that as money wages fall at a faster rate than prices (an
unbalanced deflation), the real wage falls towards its (full employment) market-
clearing level, that is (W/P)1 in quadrant (d). Money wages and prices
will continue to be bid down and the LM curve will continue to shift downwards
to the right until full employment is restored and the excess supply of
labour is eliminated. This occurs at point E1, the intersection of LM1 and IS. It
is important to stress that it is the increase in aggregate demand, via the
Keynes effect, which ensures that the economy returns to full employment.
Within this general framework there are, however, two limiting or special
cases where, despite perfect money wage and price flexibility, the economy
will fail to self-equilibrate at full employment. The two special cases of (i)
the liquidity trap and (ii) that where investment expenditure is interest-inelastic
are illustrated in Figures 3.7 and 3.8, respectively.
The limiting or special cases
In the liquidity trap case illustrated in Figure 3.7, the economy is initially at
point E0, the intersection of IS0 and LM0. Although both the goods and money
markets are in equilibrium, the income level of Y0 is below the full employment
income level YF. Reference to quadrant (d) reveals that this implies that
the level of employment (L0) is below its full employment level (LF) with real
wages (W/P)0 above their market-clearing level (W/P)1. At (W/P)0 the excess
supply of labour results in a fall in money wages (W), which reduces firms’
costs and causes a fall in prices. Although the fall in prices increases the real
value of the money supply (which shifts the LM curve outwards, from LM0 to
LM1), the increased real balances are entirely absorbed into idle or speculaThe
tive balances. In other words, in the liquidity trap where the demand for
money is perfectly elastic with respect to the rate of interest at r* (see also
Figure 3.1), the excess balances will not be channelled into the bond market
and this prevents a reduction in the rate of interest to r1 (at point E2) which
would be required to stimulate aggregate demand and restore full employment.
With no increase in aggregate demand to moderate the rate of fall in
prices, prices fall proportionately to the fall in money wages (a balanced
deflation) and real wages remain at (W/P)0, above their market-clearing level
(W/P)1. Aggregate demand is insufficient to achieve full employment and the
economy remains at less than full employment equilibrium with persistent
‘involuntary’ unemployment. Finally, as noted earlier, in section 3.3.3, in the
case of the liquidity trap monetary policy becomes impotent, while fiscal
policy becomes all-powerful, as a means of increasing aggregate demand and
therefore the level of output and employment.
In the interest-inelastic investment case illustrated in Figure 3.8, the economy
will also fail to self-equilibrate at full employment. As before, we assume the
economy is initially at point E0 (the intersection of IS0 and LM0) at an income
level (Y0) which is below its full employment level (YF). This implies that the
level of employment (L0) is below its full employment level, with real wages
(W/P)0 above their market-clearing level (W/P)2. The excess supply of labour
results in a fall in money wages and prices. Although the increase in real
balances (which shifts the LM curve from LM0 to LM1) through the Keynes
effect results in a reduction in the rate of interest, the fall in the rate of
interest is insufficient to restore full employment. Reference to Figure 3.8
reveals that, with investment expenditure being so interest-inelastic, full employment
equilibrium could only be restored through the Keynes effect with
a negative rate of interest at r1. In theory the economy would come to rest at
E1 (with a zero rate of interest), a point of underemployment equilibrium (Y1)
with persistent involuntary unemployment.
At this stage it would be useful to highlight the essential points of the above
analysis. In summary, reductions in money wages and prices will fail to restore
full employment unless they succeed in increasing aggregate demand via the
Keynes effect. In the liquidity trap and interest-inelastic investment cases,
aggregate demand is insufficient to achieve full employment and persistent
involuntary unemployment will only be eliminated if the level of aggregate
demand is increased by expansionary fiscal policy. The effect of combining the
comparative-static IS–LM model with the classical assumption of flexible prices
and money wages is to imply that Keynes failed to provide a robust ‘general
theory’ of underemployment equilibrium and that the possibility of underemployment
equilibrium rests on two highly limiting/special cases.
The above equilibrium analysis, which owes much to the work of
Modigliani, implies, as we have seen, that it is possible for the economy to
come to rest with persistent (involuntary) unemployment due to ‘rigidities’ in
the system, that is, rigid money wages, the liquidity trap or the interestinelastic
investment case. In contrast, Patinkin (1956) has argued that
unemployment is a disequilibrium phenomenon and can prevail even when
money wages and prices are perfectly flexible. To illustrate the argument,
assume an initial position of full employment and suppose that there then
occurs a reduction in aggregate demand. This reduction will result in a period
of disequilibrium in which both prices and money wages will tend to fall.
Patinkin assumes that money wages and prices will fall at the same rate: a
balanced deflation. In consequence the fall in the level of employment is not
associated with a rise in real wages but with the fall in the level of aggregate
‘effective’ demand. In other words, firms would be forced off their demand
curves for labour. In terms of panel (d) of Figure 3.8, this would entail a
movement from point A to B. Nevertheless, in Patinkin’s view this disequilibrium
will not last indefinitely because, as money wages and prices fall, there
will be a ‘direct’ effect stimulating an increase in aggregate demand, via the
value of money balances, thereby restoring full employment, that is, a movement
back from point B to A. This particular version of the wealth effect on
spending is referred to as a ‘real balance’ effect (see Dimand, 2002b). More
generally, as we discuss in the next section, the introduction of the wealth or
Pigou effect on expenditure into the analysis ensures that, in theory, as long
as money wages and prices are flexible, even in the two special cases noted
above the macroeconomy will self-equilibrate at full employment. We now
turn to discuss the nature and role of the Pigou effect with respect to the
possibility of underemployment equilibrium in the Keynesian IS–LM model.
The Pigou effect
Pigou was one of the last great classical economists who spoke for the
classical school in the 1940s (for example, 1941, 1943, 1947) arguing that,
providing money wages and prices were flexible, the orthodox Keynesian
model would not come to rest at less than full employment equilibrium. The
Pigou effect (see, for example, Patinkin, 1948, for a classic discussion) concerns
the effect that falling prices have on increasing real wealth, which in
turn increases consumption expenditure. Suppose, as is the case in Figure
3.7, the economy is at underemployment equilibrium (Y0) in the liquidity trap
at point E0, the intersection of IS0 and LM0. As prices fall, not only will the
LM curve shift outwards to the right (from LM0 to LM1) as the real value of
the money supply increases, but the IS curve will also shift to the right, from
IS0 to IS1, as the resultant increase in real wealth increases consumption
expenditure. In theory the economy cannot settle at underemployment equilibrium
but will automatically adjust until full employment is achieved at
point E1, the intersection of IS1 and LM1. The reader should verify that, once
the Pigou or wealth effect on expenditure is incorporated into the analysis, in
the special interest-inelastic investment case illustrated in Figure 3.8 the
economy will automatically adjust to restore full employment, at point E2.
The importance of the Pigou effect at the theoretical level has been neatly
summarized by Johnson (1964, p. 239): ‘the Pigou effect finally disposes of
the Keynesian contention that underemployment equilibrium does not depend
on the assumption of wage rigidity. It does.’
Over the years a number of reservations have been put forward which
question whether, in practice, the Pigou or wealth effect will ensure a quick
return to full employment (see, for example, Tobin, 1980a). In what follows
we consider two of the main criticisms of the effect. First, dynamic considerations
may invalidate the Pigou effect as a rapid self-equilibrating
mechanism. For example, if individuals expect a further future fall in prices,
they may postpone consumption, causing unemployment to rise. At the same
time, if firms expect a recession to continue, they may postpone their investment
plans, again causing unemployment to rise. Furthermore, in a deep
recession bankruptcies are likely to increase, reducing expenditure still further
(see, for example, Fisher, 1933b). In terms of the diagrammatic analysis
we have been considering, falling prices may cause the IS curve to shift to the
left, driving the economy further away from full employment equilibrium. In
these circumstances expansionary fiscal policy would ensure a more rapid
return to full employment.
Second, we need to consider briefly the debate on which assets constitute
‘net’ wealth. Net wealth can be defined as total wealth less outstanding
liabilities. In the Keynesian model wealth can be held in money and bonds.
Consider first money, which is widely accepted as comprising currency plus
bank deposits. Outside money can be defined as currency, plus bank deposits
which are matched by banks’ holdings of cash reserves or reserves at the
central bank. Outside money may be considered as net wealth to the private
sector as there is no offsetting private sector liability. In contrast, inside
money can be defined as bank deposits which are created by lending to the
private sector. As these bank deposits are matched by a corresponding private
sector liability (bank loans), it can be argued that inside money cannot be
regarded as net wealth. It is worth noting that the argument that inside money
does not constitute net wealth has been challenged by, among others, Pesek
and Saving (1967) and Johnson (1969). While this is an interesting debate
within monetary economics, it goes beyond what is required for our purposes.
Suffice it to say that, if one accepts the argument that only outside
money unambiguously constitutes net wealth, the wealth effect of falling
prices on consumption expenditure is greatly diminished. Next, as noted
earlier, in section 3.3.3, there is debate over whether government bonds can
be regarded as net wealth. It could be argued that the private sector will
realize that, following a fall in prices, the increase in the real value of
government debt outstanding will necessitate future increases in taxes to
meet the increased value of interest payments on, and redemption of, government
bonds. If the rise in the present value of future tax liabilities exactly
offsets the increase in the real value of government debt outstanding there
would be no wealth-induced shift in the IS curve. Again, while this view is
not one that is universally accepted, it does nevertheless cast doubt on the
self-equilibrating properties of the economy via the Pigou effect. The empirical
evidence for the strength of the Pigou effect shows it to be extremely
weak. For example, both Glahe (1973, pp. 213–14) for the USA and Morgan
(1978, pp. 55–7) for the UK found that the Pigou effect was not strong
enough to restore full employment in the interwar period, with actual price
level falls taking place alongside a decline in expenditure and output. Furthermore,
on reasonable assumptions, Stiglitz (1992) has shown that, if prices
were to fall by 10 per cent per year, then ceteris paribus ‘to increase consumption
by 25 per cent would take roughly 400 years’ and ‘it is hard to see
even under the most optimistic view, the quantitative significance of the real
balance effect for short-run macroeconomic analysis’. Given such doubts,
orthodox Keynesians prescribe expansionary fiscal policy to ensure a more
rapid return to full employment.
Finally it is interesting to quote Pigou (1947), who suggested that the
‘puzzles we have been considering … are academic exercises, of some slight
use perhaps for clarifying thought, but with very little chance of ever being
posed on the chequer board of actual life’.
The neoclassical synthesis
From the discussion of sections 3.4.1–3.4.3 it will be apparent that, if money
wages and prices are flexible, the Keynesian IS–LM model can in theory, via
the Pigou or wealth effect, automatically adjust to achieve full employment,
the main prediction of classical economics. In terms of pure analytical theory,
Pigou was said to have won the intellectual battle, establishing a triumph for
classical theory. Some writers (for example, Wilson, 1980; Presley, 1986;
Bridel, 1987) have suggested that Keynes anticipated the wealth effect but
rejected it on theoretical and practical grounds. Notwithstanding this neglected
point, Keynesians regarded themselves as having won the policy
debate in that the process of adjustment via the Pigou effect might be so slow
that interventionist policies (notably expansionary fiscal policy) would be
required to ensure a more rapid return to full employment. During the late
1950s and early 1960s a consensus view emerged, the so-called ‘neoclassical
synthesis’ (see Fletcher, 2002), in which the General Theory was seen as a
special case of a more general classical theory (that is, the case where
downward money wage rigidity prevents the classical automatic adjustment
to full employment), while the need was recognized for Keynesian interventionist
policies to ensure a more rapid return to full employment.
The general case
Within the IS–LM model the existence of underemployment equilibrium can
be attributed to the existence of ‘rigidities’ in the system, especially two key
prices, the money wage and the interest rate. We begin with that of the
‘Keynesian’ assumption of downward rigidity in money wages. This case can
be illustrated using the four-quadrant diagram of Figure 3.6. Quadrant (a)
depicts the standard IS–LM model. Quadrant (c) shows the short-run production
function where, with the capital stock and technology taken as given, the
level of output/income (Y) depends on the level of employment (L) – see
Chapter 2, section 2.3. Quadrant (d) depicts the labour market in which it is
assumed that the demand for/supply of labour is negatively/positively related
to real wages (W/P). Finally, quadrant (b) shows, via a 45° line, equality
between the two axes, both of which depict income. The inclusion of this
quadrant allows us to see more easily the implications of a particular equilibrium
level of income, established in the goods and money markets in quadrant
(a), for the level of employment shown in quadrant (d). In other words, in
what follows the reader should always start in quadrant (a) and move in an
anti-clockwise direction to trace the implications of the level of income
(determined by aggregate demand) in terms of the level of employment in
quadrant (d).
Suppose the economy is initially at point E0, that is, the intersection of LM0
and IS in quadrant (a). While both the goods and money markets are in
equilibrium, the income level of Y0 is below the full employment income
level YF. Reference to quadrant (d) reveals that with a fixed money wage (set
exogenously) and a price level consistent with equilibrium in the money
market (that is, the curve LM0), the resultant level of real wages (W/P)0 is
inconsistent with the labour market clearing. In other words there is no
guarantee that the demand-determined level of employment (L0) will be at
full employment (LF). The excess supply of labour has no effect on the
money wage, so that it is possible for the economy to remain at less than full
employment equilibrium with persistent unemployment. We now consider
what effect combining the IS–LM model with the classical assumption of
flexible prices and money wages has on the theoretical possibility of underemployment
equilibrium.
Again suppose the economy is initially at point E0, that is, the intersection
of IS and LM0 in quadrant (a). As before, while both the goods and money
markets are in equilibrium, the income level of Y0 is below the full employment
income level YF. Reference to quadrant (d) reveals that this implies that
the level of employment (L0) is below its full employment level (LF) with real
wages (W/P)0 above their market-clearing level (W/P)1. As long as prices and
money wages are perfectly flexible, the macroeconomy will, however, selfequilibrate
at full employment. At (W/P)0 the excess supply of labour results
in a fall in money wages (W), which reduces firms’ costs and causes a fall in
prices (P). The fall in prices increases the real value of the money supply,
causing the LM curve to shift downwards to the right. Excess real balances
are channelled into the bond market where bond prices are bid up and the rate
of interest is bid down. The resultant fall in the rate of interest in turn
stimulates investment expenditure, increasing the level of aggregate demand
and therefore output and employment. The ‘indirect’ effect of falling money
wages and prices which stimulates spending via the interest rate is referred to
as the ‘Keynes effect’. The increase in aggregate demand moderates the rate
of fall in prices so that as money wages fall at a faster rate than prices (an
unbalanced deflation), the real wage falls towards its (full employment) market-
clearing level, that is (W/P)1 in quadrant (d). Money wages and prices
will continue to be bid down and the LM curve will continue to shift downwards
to the right until full employment is restored and the excess supply of
labour is eliminated. This occurs at point E1, the intersection of LM1 and IS. It
is important to stress that it is the increase in aggregate demand, via the
Keynes effect, which ensures that the economy returns to full employment.
Within this general framework there are, however, two limiting or special
cases where, despite perfect money wage and price flexibility, the economy
will fail to self-equilibrate at full employment. The two special cases of (i)
the liquidity trap and (ii) that where investment expenditure is interest-inelastic
are illustrated in Figures 3.7 and 3.8, respectively.
The limiting or special cases
In the liquidity trap case illustrated in Figure 3.7, the economy is initially at
point E0, the intersection of IS0 and LM0. Although both the goods and money
markets are in equilibrium, the income level of Y0 is below the full employment
income level YF. Reference to quadrant (d) reveals that this implies that
the level of employment (L0) is below its full employment level (LF) with real
wages (W/P)0 above their market-clearing level (W/P)1. At (W/P)0 the excess
supply of labour results in a fall in money wages (W), which reduces firms’
costs and causes a fall in prices. Although the fall in prices increases the real
value of the money supply (which shifts the LM curve outwards, from LM0 to
LM1), the increased real balances are entirely absorbed into idle or speculaThe
tive balances. In other words, in the liquidity trap where the demand for
money is perfectly elastic with respect to the rate of interest at r* (see also
Figure 3.1), the excess balances will not be channelled into the bond market
and this prevents a reduction in the rate of interest to r1 (at point E2) which
would be required to stimulate aggregate demand and restore full employment.
With no increase in aggregate demand to moderate the rate of fall in
prices, prices fall proportionately to the fall in money wages (a balanced
deflation) and real wages remain at (W/P)0, above their market-clearing level
(W/P)1. Aggregate demand is insufficient to achieve full employment and the
economy remains at less than full employment equilibrium with persistent
‘involuntary’ unemployment. Finally, as noted earlier, in section 3.3.3, in the
case of the liquidity trap monetary policy becomes impotent, while fiscal
policy becomes all-powerful, as a means of increasing aggregate demand and
therefore the level of output and employment.
In the interest-inelastic investment case illustrated in Figure 3.8, the economy
will also fail to self-equilibrate at full employment. As before, we assume the
economy is initially at point E0 (the intersection of IS0 and LM0) at an income
level (Y0) which is below its full employment level (YF). This implies that the
level of employment (L0) is below its full employment level, with real wages
(W/P)0 above their market-clearing level (W/P)2. The excess supply of labour
results in a fall in money wages and prices. Although the increase in real
balances (which shifts the LM curve from LM0 to LM1) through the Keynes
effect results in a reduction in the rate of interest, the fall in the rate of
interest is insufficient to restore full employment. Reference to Figure 3.8
reveals that, with investment expenditure being so interest-inelastic, full employment
equilibrium could only be restored through the Keynes effect with
a negative rate of interest at r1. In theory the economy would come to rest at
E1 (with a zero rate of interest), a point of underemployment equilibrium (Y1)
with persistent involuntary unemployment.
At this stage it would be useful to highlight the essential points of the above
analysis. In summary, reductions in money wages and prices will fail to restore
full employment unless they succeed in increasing aggregate demand via the
Keynes effect. In the liquidity trap and interest-inelastic investment cases,
aggregate demand is insufficient to achieve full employment and persistent
involuntary unemployment will only be eliminated if the level of aggregate
demand is increased by expansionary fiscal policy. The effect of combining the
comparative-static IS–LM model with the classical assumption of flexible prices
and money wages is to imply that Keynes failed to provide a robust ‘general
theory’ of underemployment equilibrium and that the possibility of underemployment
equilibrium rests on two highly limiting/special cases.
The above equilibrium analysis, which owes much to the work of
Modigliani, implies, as we have seen, that it is possible for the economy to
come to rest with persistent (involuntary) unemployment due to ‘rigidities’ in
the system, that is, rigid money wages, the liquidity trap or the interestinelastic
investment case. In contrast, Patinkin (1956) has argued that
unemployment is a disequilibrium phenomenon and can prevail even when
money wages and prices are perfectly flexible. To illustrate the argument,
assume an initial position of full employment and suppose that there then
occurs a reduction in aggregate demand. This reduction will result in a period
of disequilibrium in which both prices and money wages will tend to fall.
Patinkin assumes that money wages and prices will fall at the same rate: a
balanced deflation. In consequence the fall in the level of employment is not
associated with a rise in real wages but with the fall in the level of aggregate
‘effective’ demand. In other words, firms would be forced off their demand
curves for labour. In terms of panel (d) of Figure 3.8, this would entail a
movement from point A to B. Nevertheless, in Patinkin’s view this disequilibrium
will not last indefinitely because, as money wages and prices fall, there
will be a ‘direct’ effect stimulating an increase in aggregate demand, via the
value of money balances, thereby restoring full employment, that is, a movement
back from point B to A. This particular version of the wealth effect on
spending is referred to as a ‘real balance’ effect (see Dimand, 2002b). More
generally, as we discuss in the next section, the introduction of the wealth or
Pigou effect on expenditure into the analysis ensures that, in theory, as long
as money wages and prices are flexible, even in the two special cases noted
above the macroeconomy will self-equilibrate at full employment. We now
turn to discuss the nature and role of the Pigou effect with respect to the
possibility of underemployment equilibrium in the Keynesian IS–LM model.
The Pigou effect
Pigou was one of the last great classical economists who spoke for the
classical school in the 1940s (for example, 1941, 1943, 1947) arguing that,
providing money wages and prices were flexible, the orthodox Keynesian
model would not come to rest at less than full employment equilibrium. The
Pigou effect (see, for example, Patinkin, 1948, for a classic discussion) concerns
the effect that falling prices have on increasing real wealth, which in
turn increases consumption expenditure. Suppose, as is the case in Figure
3.7, the economy is at underemployment equilibrium (Y0) in the liquidity trap
at point E0, the intersection of IS0 and LM0. As prices fall, not only will the
LM curve shift outwards to the right (from LM0 to LM1) as the real value of
the money supply increases, but the IS curve will also shift to the right, from
IS0 to IS1, as the resultant increase in real wealth increases consumption
expenditure. In theory the economy cannot settle at underemployment equilibrium
but will automatically adjust until full employment is achieved at
point E1, the intersection of IS1 and LM1. The reader should verify that, once
the Pigou or wealth effect on expenditure is incorporated into the analysis, in
the special interest-inelastic investment case illustrated in Figure 3.8 the
economy will automatically adjust to restore full employment, at point E2.
The importance of the Pigou effect at the theoretical level has been neatly
summarized by Johnson (1964, p. 239): ‘the Pigou effect finally disposes of
the Keynesian contention that underemployment equilibrium does not depend
on the assumption of wage rigidity. It does.’
Over the years a number of reservations have been put forward which
question whether, in practice, the Pigou or wealth effect will ensure a quick
return to full employment (see, for example, Tobin, 1980a). In what follows
we consider two of the main criticisms of the effect. First, dynamic considerations
may invalidate the Pigou effect as a rapid self-equilibrating
mechanism. For example, if individuals expect a further future fall in prices,
they may postpone consumption, causing unemployment to rise. At the same
time, if firms expect a recession to continue, they may postpone their investment
plans, again causing unemployment to rise. Furthermore, in a deep
recession bankruptcies are likely to increase, reducing expenditure still further
(see, for example, Fisher, 1933b). In terms of the diagrammatic analysis
we have been considering, falling prices may cause the IS curve to shift to the
left, driving the economy further away from full employment equilibrium. In
these circumstances expansionary fiscal policy would ensure a more rapid
return to full employment.
Second, we need to consider briefly the debate on which assets constitute
‘net’ wealth. Net wealth can be defined as total wealth less outstanding
liabilities. In the Keynesian model wealth can be held in money and bonds.
Consider first money, which is widely accepted as comprising currency plus
bank deposits. Outside money can be defined as currency, plus bank deposits
which are matched by banks’ holdings of cash reserves or reserves at the
central bank. Outside money may be considered as net wealth to the private
sector as there is no offsetting private sector liability. In contrast, inside
money can be defined as bank deposits which are created by lending to the
private sector. As these bank deposits are matched by a corresponding private
sector liability (bank loans), it can be argued that inside money cannot be
regarded as net wealth. It is worth noting that the argument that inside money
does not constitute net wealth has been challenged by, among others, Pesek
and Saving (1967) and Johnson (1969). While this is an interesting debate
within monetary economics, it goes beyond what is required for our purposes.
Suffice it to say that, if one accepts the argument that only outside
money unambiguously constitutes net wealth, the wealth effect of falling
prices on consumption expenditure is greatly diminished. Next, as noted
earlier, in section 3.3.3, there is debate over whether government bonds can
be regarded as net wealth. It could be argued that the private sector will
realize that, following a fall in prices, the increase in the real value of
government debt outstanding will necessitate future increases in taxes to
meet the increased value of interest payments on, and redemption of, government
bonds. If the rise in the present value of future tax liabilities exactly
offsets the increase in the real value of government debt outstanding there
would be no wealth-induced shift in the IS curve. Again, while this view is
not one that is universally accepted, it does nevertheless cast doubt on the
self-equilibrating properties of the economy via the Pigou effect. The empirical
evidence for the strength of the Pigou effect shows it to be extremely
weak. For example, both Glahe (1973, pp. 213–14) for the USA and Morgan
(1978, pp. 55–7) for the UK found that the Pigou effect was not strong
enough to restore full employment in the interwar period, with actual price
level falls taking place alongside a decline in expenditure and output. Furthermore,
on reasonable assumptions, Stiglitz (1992) has shown that, if prices
were to fall by 10 per cent per year, then ceteris paribus ‘to increase consumption
by 25 per cent would take roughly 400 years’ and ‘it is hard to see
even under the most optimistic view, the quantitative significance of the real
balance effect for short-run macroeconomic analysis’. Given such doubts,
orthodox Keynesians prescribe expansionary fiscal policy to ensure a more
rapid return to full employment.
Finally it is interesting to quote Pigou (1947), who suggested that the
‘puzzles we have been considering … are academic exercises, of some slight
use perhaps for clarifying thought, but with very little chance of ever being
posed on the chequer board of actual life’.
The neoclassical synthesis
From the discussion of sections 3.4.1–3.4.3 it will be apparent that, if money
wages and prices are flexible, the Keynesian IS–LM model can in theory, via
the Pigou or wealth effect, automatically adjust to achieve full employment,
the main prediction of classical economics. In terms of pure analytical theory,
Pigou was said to have won the intellectual battle, establishing a triumph for
classical theory. Some writers (for example, Wilson, 1980; Presley, 1986;
Bridel, 1987) have suggested that Keynes anticipated the wealth effect but
rejected it on theoretical and practical grounds. Notwithstanding this neglected
point, Keynesians regarded themselves as having won the policy
debate in that the process of adjustment via the Pigou effect might be so slow
that interventionist policies (notably expansionary fiscal policy) would be
required to ensure a more rapid return to full employment. During the late
1950s and early 1960s a consensus view emerged, the so-called ‘neoclassical
synthesis’ (see Fletcher, 2002), in which the General Theory was seen as a
special case of a more general classical theory (that is, the case where
downward money wage rigidity prevents the classical automatic adjustment
to full employment), while the need was recognized for Keynesian interventionist
policies to ensure a more rapid return to full employment.
No comments:
Post a Comment