Transcript PS curve shifts up - Lecture Notes

Lecture 3
• 3.1 Supply-side structures, policies and shocks
• 3.2 Price push factors (zp) on PS curve
[W/P=λ.f(μ,zp)]
• 3.3 Wage push factors (zw)on WS curve
[W/P=b(λ,E,zw)]
• 3.4 Unions, wage-setting and the ERU
• 3.5 Hysteresis – how actual U affects ERU
• 3.6 Conclusions from Chapter
• 3.7 Ball and Mankiw’s analysis of supply-side
factors shifting the natural rate of unemployment
(US 1960-90’s)
3.1
Supply-side structures, policies and shocks
• Supply-side policies refer to those that shift the
wage-setting (WS) and price-setting (PS) curves
– WS shifted by changes in unemployment benefits,
minimum wages, employment legislation, child-care
policy
– PS shifted by – changes in taxes or competition
policies
• Supply-side changes in WS and PS effect the level
of equilibrium output ye and thereby effect the
equilibrium rate of unemployment ERU (aka the
Nairu)
Effect of a supply-side shock (eg improved education
outcomes or reduces union bargaining power)
• In Fig 4.1 a downward shift in WS leads to a shift form A’ to Z’ i.e. employment
increases from Ee to E’e at an unchanged real wage
• Scenario 1 (if there are lags in the CB’s response to the shift in WS):
– then the CB will take some time to realise that there are disinflationary pressures in the
economy and the CB will only reduce the interest rate once the lower inflation rate
form 4% to 2% has revealed itself i.e. the move from A to B,
– in the labour market there is downward pressure on wages as more people are willing
to work at the going wage and at a given level of employment there is downward
pressure on wages (e.g. at B’)
– then the CB lowers the interest rate and stimulate AD moving economy from B to C
– Due to the lagged response the r overshoots from rs to r’ before moving to r’s)
• Scenario 2 (if CB correctly anticipates the shift in WS):
– then the CB will realise that there will be disinflation in the economy and will also
realise that equilibrium employment has shifted to E’e and equilibrium output to y’e
– The CB will immediately in the first period cut the interest rate form rs to r’s (with no
overshooting)
– The economy would move straight from A to Z and inflation will always remain on
target (πT)
– in the labour market there is no downward pressure on wages as the cut in the interest
rate means that there is increased demand for along with increased supply of labour
Comparing demand and supply shocks
• Note: Supply shocks differs from demand
shocks as follows:
• For a supply shock:
– to a new equilibrium output (y’e) level
– the MR schedule to point at which inflation target
(πT) intersects new equilibrium output (y’e)
• For a demand shock:
– The equilibrium output (y’e) level and the MR
schedule do NOT SHIFT
Time profile of inflation
• Scenario 1 (if there are lags in the CB’s
response to the shift in WS):
– the time profile of inflation is given by Fig 4.2
– inflation falls sharply and then returns to (πT)
• Scenario 2 (if CB correctly anticipates the shift
in WS):
– Inflation remains constant at (πT)
IS curve and MR curve (where CB
response is sluggish)
• IS CURVE
• Initially and temporarily the CB overshoots as it sets r to r’ as the sluggish
CB must re-inflate the system due to wages and prices falling and to
achieve πT (in the IS diagram 4.3 there is a cut in the interest rate from rs
to r’) (from A’ to C’)
• Overall (the move is from rs to r’s) (from A’ to Z’) - the stabilising interest
rate rs’ is lower in the new equilibrium because equilibrium output is
higher due to the supply-side shift (AS) and a lower real interest rate is
required to provide the appropriate level of AD (AD)
• MR CURVE
• MR shifts from MR to MR’ – comprising the new objective point due to the
fall in wages at (πT, y’e) (point Z) and the subjective preferences of the CB’s
loss function (giving the slope of the MR curve)
• Once the economy is on the new MR line MR’ the adjustment takes place
from C’ to Z’ (i.e. r remains below r’s but rises towards r’s as inflation rises
to πT)
3.2 Price push factors
• PS – price-push factors (zp) shift the PS curve
• And cause supply-side shifts in equilibrium output ye and
the equilibrium rate of unemployment
• PS is given by W/P=λ(1 – μ)
– PS shifts up with productivity λ gains (increasing equilibrium
output ye and employment), and
– PS shifts down with increases in mark-up/pricing power μ,
(decreasing equilibrium output ye and employment)
• If we include the price push effects of taxes (called the
‘tax wedge’) then:
– PS is given by W/P = λ(1 – μ) / (1+td)(1 + tv) or
– W/P=λ.f(μ,zp), where:
• PS shifts up with increased productivity λ
• PS shifts down with increased pricing power μ, and
• PS shifts down with increased price-push factors (zp) such as direct
taxes td and indirect taxes tv
Note on the tax wedge
• Relevant to WORKER utility:
– w = W/Pc is the real consumption wage, where
– W = the post-tax money wage, and
– P or Pc = the consumer price index (inclusive of indirect taxes) (ie Pc = P(1+tv))
• Relevant to FIRM utility:
– w=W/P
– W is the full cost to company real product wage (inclusive of taxes and nonwage labour costs) , and
– P the price that the firm gets for its product (excluding indirect taxes)
• The difference between the real consumption wage and the real product
wage is the Tax Wedge
• The tax wedge shows up as a price-push factor (zp) and any rise in taxes or
the tax wedge pushes the PS curve down
• Any increase in direct or indirect taxes increases the tax wedge, eg raises
VAT, raise tax on employees salaries, raises the cost of employment, etc.
• The increase in taxes shifts PS downwards, reducing equilibrium output
and employment
Change in taxes shift PS curve
• Any fall in taxes implies an upward shift in the PS
curve (as td and tv are in the denominator) of PS
equation
• i.e. the real wage is higher if the tax take is
smaller
• => WS and PS curves intersect at a higher level of
employment
• Conclusion: higher taxes (indirect and direct
taxes) mean a larger tax wedge which will result
in higher unemployment (lower real wage and
PS curve shifts downwards)
Price-push factors (including tax wedge)
• PS curve can be written compactly as: W/Pc =
λ.f(μ, zp)
• PS curve shifts up (employment up,
equilibrium output up and real wage up)
where:
– The is a rise in productivity (λ)
– Fall in the mark-up price μ (increased competition)
– There is a fall in the tax wedge (zp )
3.3 Wage-push factors
• WS equation can be written as: W/P =b (E,zW)
• zW are wage-push variables including institutional,
policy, structural and shock variables
• WS shifts down (increasing employment levels as zW
falls) where:
– There is a fall in employment benefits (or ratio of benefits
to average wage)
– Union protection falls or unions weaker (reducing gap
between WS and labour supply curve)
– Unions / employees exercise bargaining restraint e.g. a
wage accord (e.g. proposal for wage accord for high and
low income earners in South Africa’s New Growth Path 2010)
Wage-push factors and productivity increases
• There is debate as to how to include labour productivity
changes (λ) in the WS and PS equations
• First approach - λ is included in WS and PS symmetrically (ie
WS rise is equal to the PS rise as a result of λ) => no change
in ERU (this is supported by stylised fact that productivity
has been rising but the ERU has not changed), in this
approach wage aspirations rise inline with productivity
– Specified as W/P= λ b(E,Zw) (full symmetrical effect of λ is
assumed)
• Second approach – λ changes take some time to for the
change in productivity trend to influence wage-setting
behaviour (slow adjustment or lag so PS and WS do not
shift symmetrically) => a slow down in λ will lead to a rise
in ERU and a increase in λ will lead to a fall in ERU (See Ball
and Mankiw (2002) on productivity and the Nairu)
– Specified as W/P = b(λ ,E,Zw) (λ is inside b function to reflect
uncertainty of the adjustment)
Wage-push factors and training programmes
• Training may have two effects:
• First – training increases productivity of work force and
pushes PS upwards (and possibly WS curve up
symmetrically or lagged) increasing wages and
employment
• Second – education may weaken the bargaining power
of employees shifting the WS curve downwards
(increasing employment but decreasing wages)
• Both effects are theoretically possible and the result
will depend on the specific empirical circumstances
e.g. in SA would widening the skills pool assist in
reducing the premium on skills and thus narrow the
wage gap?
Wage-push factors and wage accords
• Wage accord or incomes policy agreements can
be entered into between labour, business and
government which have the effect of limiting
wage increases and lowering the WS curve
• See Fig 4.4 where it is assumed that wage
negotiations outcomes can results in setting
wages between two limits – the WS ceiling set by
employers (they will not go higher) and the WS
floor set by employees (they will not go lower)
• The gap between the ceiling and the floor is
termed the “zone of bargaining discretion”
Wage-push factors and wage accords (cont.)
• Question why don’t unions always insist on the WS ceiling?
Because of unions concern for long-run future of the
industry, maximum industrial action may jeopardise
investment plans in the industry (unions and employers are
in a relationship that is simultaneously antagonistic and codependent)
• Where there is a wage accord (or incomes policy) the WS
curve moves downwards in the zone of bargaining
discretion (often with govt offering policy measure
supported by the unions e.g. tax cuts / pension policy, etc.)
• Examples of successful wage accords are Germany (1990’s),
Australia (1980’s),Netherlands (1982) and Ireland (1987)
where there was wage restraint and increased employment
assisted by the productivity growth of the 1990’s
• In SA despite wage accord proposals – relations have been
relatively conflictual and antagonistic in the post 1994
period.
Notes on Australian Labour Accord in the 1980’s
• There was a 1983 accord between the Australian Labour Party
(ALP) and the Australian Council of Trade Unions (ACTU), which
served as a platform to bring the labour party to power and
secure long-term gains for the broader working class in Australia.
• In 1983 the Australian economy was suffering stagflation
(stagnant employment growth and inflation) and on the eve of
elections, which Bob Hawke’s Australian Labour Party eventually
won, the ALP and ACTU signed an accord which sought to secure
employment growth, put in place tax cuts for the poor as well as
increased social security, parental rights of workers and social
services aimed at benefiting poor families.
• Through a process of negotiation the ALP and ACTU arrived at a
common “realistic appraisal of the economic situation” and
“identified jointly how best long term gains could be made for
the working class”.
• The working class, which he defined as comprising three
components, that is, the organised working class, the
unorganised working class and the unemployed.
Australian Labour Accord (cont).
• The following gains were achieved:
– thirteen years of wages increases, although moderate, were better
than would have been secured without the accord,
– a record number of jobs were created and stagflation was ended,
– the working week was shortened and parental rights were secured,
– the economy was made more competitive and was opened up to
international trade in a sustainable way,
– a superannuation savings and pension scheme was created, and
– a free health care systems was put into place.
• Looking beyond the immediate interests of current union
membership, goals such as employment creation, social security
gains and international competitiveness were seen as important
objectives by the ALP and ACTU.
• “Simple wage gains can be quickly eroded by inflation, but
institutional changes have the potential to benefit workers forever”
– claimed the ACTU.
Analysis of why wage accord is difficult for unions
• If there are many unions competing in the economy then
there is an incentive for unions to defect from the wage
accord and try and secure higher wages (even if this is not
optimal from the point of view of employment levels in the
economy)
• Game theory’s Prisoner Dilemma framework is useful to
understand this incentive to defect / confess:
– If one confesses and the other does not he gets 0 years and the
other gets 20 years
– If both confess they both get 10 years
– If neither confess they both get 2 years
– The dominant strategy is to confess (get 0 years for one and 20 for
the other or 10 years each) even though combined utility is
maximised if neither confess (2 x 2yrs in jail)
– 0 + 20 or 10 + 10 > 2 + 2 i.e. loss is minimised at 2 + 2 but prisoner’s
cannot co-ordinate their decision for the optimal outcome
Unions incentive to defect as prisoners dilemma
• Game theory’s Prisoner Dilemma framework is useful to
understand this incentive to defect from the wage accord:
– If one union defects from wage accord and the other does not
the defecting union gets higher wages (and reduced
employment)
– If both unions defect they both get higher wages (but reduced
employment)
– If neither unions defects they both get lower wages but
increased employment (which is the best social outcome)
– The dominant strategy is to defect with the result of higher
wages and lower employment even though the best
employment outcome is for no defection
• For this reason the wage accord is more possible/more
likely in economies where there are fewer competing
unions and where there is a recognised dominant union
leading highly centralised bargaining (the co-ordination
problem is easier to resolve optimally)
3.4 Unions, wage-setting and the ERU
• Calmfors and Driffill (1988) posited a humped-shaped (or
convex) relationship between the degree of centralisation
of wage setting and the ERU (see Fig 4.7)
• A low ERU is consistent with highly decentralised or highly
centralised wage bargaining. In between is the
intermediate degree of centralisation (where there may be
strong competition between unions undermining wage
accords) where the ERU is higher. (Post 94 SA is probably
best characterised as being in this zone of intermediate
centralisation – not only due to union competition but also
policy and political contestation)
• An implication of Calmfors and Driffill’s thesis is that union
strength is not necessarily associated with an upward push
in the WS curve
Calmfors and Driffill (1988)
• Union utility increases with employment and real wage, maximise E x W/P
• There are three context for wages setting:
– Firm level (decentralised) – if W/P rises, E falls (due to high substitutability
between firms)
– Industry level (intermediate) – if W/P rises, E not vulnerable (due to low
substitutability between industries)
– Economy wide (centralised) – if W/P rises, E falls (due to rising inflation economywide)
• Two forces for wage moderation:
– The way employment will respond (negatively) to wage increases (this is more of
an issue for decentralised firm-level negotiation than for industry-wide
negotiations as the degree of substitutability between the product of different
industries is less than between firms)
– The effect of wage increases on economy-wide prices (this is not relevant for firmlevel wage setting were p is assumed as given for setting W/P but for centralised
economy-wide negotiations wage setting impacts on the economy-wide price
level) as a result a centralised union does not maximise its utility by increasing
wages (as for W/P changes in W lead to changes in P) but by maximising
employment. (Referring to the Prisoners dilemma – the centralised union
overcomes the co-ordination problem by acting as a single decision maker
maximising the utility of its members)
Figs 4.5 and 4.6 represent wage setting under
different institutional arrangements
• Union / worker indifference curves slope downwards as maximum wage
bill is sought (W/P x E) and then upwards as eventually there is a
disutility of work
• For industry level bargaining the industry level labour demand curves
are rather steep (as demand is not highly responsive to wage changes as
there is a low degree of substitutability between industries)
• The 3 labour demand curves are shown for the different levels of AD
• By joining the points of tangency (between labour demand curves and
union indifference curves) the WSIND curve for industry-level bargaining
is derived (fig 4.5)
• At firm level there is more concern about the negative impact of wages
on employment, so firm-level labour demand curves are flatter
(reflecting a higher elasticity of demand) implying a WS curve WSFIRM
which is below the WSIND curve
• At a more centralised level, the central union overcomes the
coordination problem and adopts policies aimed at maximum
employment at a satisfactory real wage (there is a higher responsiveness
to the wage employment trade-off)
Implication of different institutional arrangements
• In Fig 4.6 the implication of institutional
arrangements is shown:
• Employment for WSIND is below employment for
WSFIRM
• Where there is a centralised union – it must take
into account the effect on prices of setting wages,
so the unions uses the labour supply curve Es in
setting wages and employment (leading to an
employment outcome Ee(C), where:
Ee(Centralised) >Ee(Decentralised)>Ee(Industry-wide)
3.5 Hysteresis – how actual U affects ERU
• Hypotheses already established:
– ERU is shifted by supply-side factors
– AD shocks can have s-r effects on actual U but no
effect on the ERU
– AD policy can stabilise the economy around ERU but
cannot effect the level of ERU
• New hypothesis (U effects ERU):
– If actual U is high for a long period this may result in
higher ERU (due to damaging effects on supply-side of
the economy) (referred to as path-dependence where
equilibrium of a system depends on history of the
system)
– Hysteresis can be explained by insider-outsider effects
(influencing WS curve), by long-term unemployment
effects (influencing WS curve) (See Colin Bundy on
history of SA)
Hysteresis – insider-outside model
• In Fig 4.8 when AD falls employment falls from Ee to E1
• Usually CB will cut interest rates (as inflation falls) and employment will
return to Ee, but if CB does not cut rates, or if the impact of falling
inflation on AD is weak then the economy may stay at E1 for some time
• => two groups of workers emerge – unemployed ‘outsiders’ and those
‘insiders’ employed at E1
• Insiders are in a strong bargaining position due to their skills and they are
assumed to want to push up real wages with no concern for increasing
employment levels => WS curve becomes vertical at E1
• Any increase in AD will be reflected in a rising real wage, equilibrium
employment has fallen to E1=E’e
– i.e. normally increased AD will lead to higher employment and higher wages,
but with hysteresis wages rise but employment does not rise (as outsiders are
not, or cannot be, brought back into employment)
• Policy implication: even though the rise in U is due to falling AD, only a
supply-side change altering wage setting arrangements can reduce
equilibrium U (raise E back to Ee)
Hysteresis: Long-term unemployment
• Long-term unemployed effectively withdraw from the
labour market due to progressive loss of skills and
erosion of psychological attachment to working life
(highly pertinent to South Africa’s “structural
unemployment”)
• The higher is the ratio of long-term unemployed to the
overall pool of unemployed the less impact has the
level of U on wages
• => WS curve shifts upwards (for given E wages higher
as long-term unemployed do not put downward
pressure on wages)
• Hysteresis effect- emergence of long-term U leads to
prolonged period of high U
Too few people work in South Africa
The employment to population ratio in South Africa and other emerging market economies
Aggregate employment ratio
Youth employment ratio
Adult employment ratio
80
Employment ratio (%)
70
60
50
40
30
20
10
0
Source: Statistics SA, ILO
36
Fig 4.9
• At point A (with equil. employment Ee) AD falls (e.g. in SA Case
under apartheid is low / text = disinflation policy) and employment
falls to E1 (after some time the share of long-term unemployed
rises, therefore WS rises to WS(LTUh) move from B to C
• When govt wishes to raise AD again the new equilibirium level of
employment is at E2 (point D)
• In the long-run, the economy will return to A again as at point D the
share of long-term unemployed will begin to decline (given the
increase in employment) and gradually the WS(LTUh) curve will
begin to shift down to WS(LTUm) and then to WS(LTUi)
• If the scarring has been very serious (over decades) then active
labour market measure may be required to reintegrate the longterm unemployed back into the market (eg re-training programmes,
grants fro travel to job interviews, etc)
• The flatter (dashed) WS curve shows the WS cure where the share
of long-term unemployed is constant – this WS curve intersects
with the PS curve at the long-run ERU at A (but due to the fact the
the share of long-term unemployed is not constant and WS shifts
up the return to A takes longer i.e. is a more gradual process)
3.6 Conclusions from chapter
•
•
If WS is pushed up or PS is pushed down then E falls and ERU rises
WS is pushed up (and equilibrium output and employment fall):
– When unemployment benefits become more generous
– Unions have more power – legal rights of industrial action (or union density or collective
bargaining coverage)
– Unions exercise less wage restraint (collapse of wage accord)
– There is more intermediate level wage setting (as opposed to centralised or decentralised)
•
PS is pushed down and equilibrium output and employment fall), when:
–
–
–
–
–
•
Rise in the tax wedge
Rise in the price mark-up (when monopoly power rises)
Fall in productivity
A rise in real interest rates
More employment regulation (net of any productivity gains)
How does the CB respond to such shocks (or supply side changes) in the IS-PC-MR
model:
– As ye changes the position of the MR line changes (as it intersects with (y’e; πT)
– The interest rate will overshoot if the CB is sluggish in its response to the supply side shock
– Inflation will always remain a πT if the CB fully anticipates the supply side shock
3.7 The Nairu or natural rate of
employment
• Ball and Mankiw propose a method for measuring how
the Nairu shifts over time
– Their conclusion for the US economy is that the Nairu was low in
the 60’s, then rose and peaked in the late 1970’s and then
declined into the 1990’s (see Figure 1)
– In IS-PC-MR model:
– 60’s ye high, Nairu low
– 70’s ye low, Nairu high
– 90’s ye high, Nairu low
• The authors then discuss what the causes of the shifting
Nairu could be
– They conclude that demographic issues and government policies
play some role, as well as an important role for changes in
productivity growth
Measuring shifts in the NAIRU
Inflation – Unemployment I-U trade off:
π = k – aU (Philips rels – see quote from Hume)
Expectations augmented I-U trade off:
π = πe – a(U – U*)
(i.e. if π > πe, therefore U < U* i.e. high inflation is associated
with low U)
With supply shocks:
π = πe – a(U – U*) + v
(i.e. π > πe could be associated with oil price shock or
exchange rate depreciation = v)
(Also longer term shifts in U* means that stable inflation can
be associated with lower levels of U)
Measuring NAIRU
Authors make use of adaptive expectations:
π = πt-1 – a(U – U*) + v
(Justified on basis that inflation has followed a random walk in
recent decades i.e. πt = πt-1+ ut (ut = random error term),
therefore, system approximates optimal rational behaviour)
Re-write as ∆π = aU* - aU + v
Rearrange and divide by a: U* + v/a = U + ∆π/a
Therefore, shift in U* + v/a can be computed on basis of
observables U + ∆π/a (with U* representing long-run trends
and v/a representing short-run shocks)
Using a technique called Hodrick-Prescott Filter to estimate the
Long-run trend in the series (Fig 1)
Causes of Shift in NAIRU
• US 1960 to 2000 NAIRU is found to be time varying (5.4% in
the 1960, peaking at 6,8% in 1979 and declining to 4,9% in
2000). Why?
• Demographic explanation:
– Changing composition of the labour force i.e. changes in sizes of
groups with relatively high or low rates of unemployment can change
the aggregate unemployment rate
– e.g. increase of numbers of (relatively high unemployment) young
workers before 1980 (baby boom generation) led to increase in NAIRU
and decrease in numbers of young workers led to decrease in NAIRU
after 1980
– But, if fixed weight is given to different demographic groups (Perryweighted index - not adjusted according to labour force shares) i.e.
what would have happened to the NAIRU if there had been no
demographic shift, it is found that the impact of demographics shifts
has been modest (Figure 2)
Impact of Government Policies
• Government disability and incarceration policies can also
impact on the NAIRU e.g. if government policies cause people
to leave the work force due to paying better disability benefits
or due to increased incarceration then the NAIRU will tend to
decline (as these people are removed from groups likely to
experience high unemployment)
• Changing policies on disability and incarceration in the 80’s
and 90’s could explain a reduction of the NAIRU by only
about 0,8% (with disability policy playing a more important
role with numbers rising from 3,1% of workforce in 1984 to
5,3% in 2000
Productivity Acceleration
• An approach is that rising productivity (e.g. growth of
output per hour of work) is associated with a decline
in the NAIRU (in the 1990’s) and falling productivity
is associated with an increase in the NAIRU
(“productivity slowdown” of the 1970’s)
• Explained by the fact that “wage aspirations” adjust
slowly to shifts in productivity growth
– i.e. when productivity slumps real wages must fall, but this
is resisted, real wages are relatively high and the NAIRU
increases (in 1970’s λ falls => PS down while WS downward
adjustment is sluggish so employment falls/NAIRU rises)
– Or, when productivity accelerates, workers accustomed to
slow wage growth, mean that real wages do not increase
accordingly and employment levels rise, NAIRU falls (in
1990’s λ rises => PS rises while WS upward adjustment is
sluggish so employment rises/NAIRU falls)
Explaining the 1990’s
• What is important in explaining the fall of the NAIRU in the
1990’s is not the rate of productivity growth (which was
actually slightly higher in the 1960’s than 1990’s), but the
increase in productivity relative to the recent past (i.e. the
slowdown of the 1970’s)
• Past experience meant that wage aspirations were not high
when productivity increased in the 1990’s, so adjustment was
slow resulting effectively in a situation of a mismatch between
productivity levels and real wages
• The 1990’s experienced a period of falling unemployment
and low inflation (reflected in reduction in the NAIRU)