course handouts spring 2010 - web hostinghschuetz/econ371/handouts371_10.pdfl’s indifference curve...

Economics 371 Economics of Work and Pay

Course Handouts Spring 2010

Professor H.J. Schuetze

1

Topic 4.3 - Compensating Differentials

Professor H.J. SchuetzeEconomics 370

Compensating Wage Differentials

Refers to the wage differentials that exist in equilibrium to compensate workers for undesirable job characteristicsExamples:

Unpleasant/unsafe working conditionsDesirable attributes like non-wage benefitsLong commute timeThus, wages can also reflect compensation to

workers for undesirable job characteristics not just

Professor Schuetze - Econ 370 2

j jproductivity

Let’s work through the theory thinking about compensating wages for the risk of injury that can result from an unsafe or unhealthy work environment

2

The Firm’s Iso-Profit Curve

The various combinations of wages and safety the firm can provide while maintaining the same level of profit

Downward Sloping:• Both wages and safety are

costly for the firm• Firm can maintain profit and

increase safety only if it can pay a lower wage

Concave:

Wage A

I0

I1


Concave:• There is a diminishing marginal

rate of transformation between wages and safetySafety

B

Concavity of Iso-Profit Curve

At A: little safety and high wagesWith little safety the firm can increase safety

inexpensivelyp ybetter lightingsignsguard railsThe firm is at a stage of increasing returns with

respect to the provision of safetyThe iso-profit is relatively flat


The iso profit is relatively flat

3

Concavity of Iso-Profit CurveAt B: lots of safety, low wage

The firm can only increase safety with more sophisticated and costly safety proceduressophisticated and costly safety procedures

Stage of diminishing returns in the provision of safetyThe iso-profit is relatively steep

Left to right: The firm starts with the cheapest forms of safety provision and moves to more expensive formsI0 to I1: Implies a higher level of profits.


Iso-profit curves with both lower wages and safety are higher because both are costly

Different Firms/Safety Technology

Different firms can have different abilities to provide safety at a given costThus different firms can have differently shaped isoThus, different firms can have differently shaped iso profit curves for the same level of profits

Wage Firm 1 (High Safety Costs)


Firm 2 (Low Safety Costs)

Safety

4


Firm 1: Safety is costlyCan provide additional safety only if wages drop rapidly to compensate for safety costs and maintainrapidly to compensate for safety costs and maintain profits

“rapidly diminishing returns to providing safety”e.g. inherently dangerous sectors

mininglogging



Firm 2: Low safety costsCan provide additional safety at a relatively cheaper

pricepricee.g. inherently safe industry

office managerprofessor

Competition implies that in equilibrium both firms will earn zero economic profits so that


I1 = I2 = 0

5

Employer’s Offer Curve

The maximum compensating wage that will be offered in the market for the different levels of safetyTherefore it is the outer limits of the two iso profitTherefore, it is the outer limits of the two iso profit curves

Firm 1

Firm 2

Wage

BW*

AW1

Points “inside” the offer curve will be dominated by points “on” the offer curve

Example:Worker can receive S* safety and a wage of W1 from Firm 1


Safety

AW1

S*

g 1(Point A)

Firm 2 can provide safety relatively cheaper and therefore can offer W* > W1 wage along with the same amount of safety

Individual’s Preferences

Workers have preferences for wages and safetyCan illustrate using indifference curves

• The level of utility is the same along the indifference curve.

• Utility increases up and to the right

• Convexity illustrates a diminishing marginal rate of substitution between wages and

f t

Wage

A

B

U1


safety

Safety

B U0

A: willing to give up a lot of wages to get small increase in safety (curve is steep)

B: not willing to give up wages for safety (curve is flat)

6

Different Individuals (Risk Preferences)

Different workers may be more or less willing to give up safety in return for higher wages (a risk premium)

A

UM

UL

Wage The individual with preferences ULis less risk averse while UM is more risk averse


Safety

Person L requires less of a compensating wage increase to give up safety at point A than person ML’s indifference curve is flatter than M’s

Equilibria

1. Single Firm, Single Individual:

WageWage

Safety

E

U0

S*

W*


SafetyS

For a single type of firm and single type of worker the equilibrium occurs at a point of tangency like EThis gives maximum utility given the firms ability to trade safety for wages at perfectly competitive profit = 0

7

Single Firm, Single Individual

To the left of E:

• Additional wage required by the worker to accept more risk (slope of IC) is greater than the firm is able to offer (slope iso‐profit)

To the right:

• The worker is not willing to give up enough risk

• If the firm tries to reach a lower iso profit (higher profits) workers will go to another firm


Equilibria2. Single Worker, 2 Types of Firms:


Wage

Safety

A

SA

WA

B

UA

WB

SB

Firm 1 (Costly)

Firm 2 (Cheap)

With preferences as drawn, maximum utility with Firm 1 is UA

8

Adjustment to Equilibrium• Suppose that Firm 1 is able to raise the price of its product

when the good becomes in short supply


Wage

Safety

A

B

Firm 1

Firm 2

• Now in equilibrium• The homogeneous workers would

be indifferent between working at either Firm 1 or Firm 2

Equilibrium

• We could measure how much workers value safety if we could measure safety

• In a cross‐section of workers the various combinations of wages and safety traces out an indifference curve


9

Equilibria3. 1 Type of Firm, 2 Types of Workers:


Wage

Safety

A

UMSA

WA

B

UL

WB

SB

Less Risk Averse

More Risk Averse

Here, a cross-section on wages and safety traces out the iso-profit curvetells us how firms are able to trade wages for safety

Different Individuals (Risk Preferences)4. Many Firms and Individuals:


Wage

Safety

UL Less Risk Averse

UM More Risk Averse

F1

F2

F3UA Average Risk Averse

10

Wage‐Safety Locus

•

• The set of tangencies between iso‐profits and indifference curves (combination of wage and safety that will prevail in the market)

• The wage‐safety locus can change depending upon:

(i)

(ii)

•


Wage‐Safety Locus

• The wage premium paid for risk is often called a “shadow” price

• This is because the price is embedded in the market wage rather than being attached explicitly to a job characteristic


11

Effects of Safety Regulation• Might expect the effects of safety regulation to be

“negative”

• Workers and firms efficiently allocate themselves in terms of worker’s risk aversion and employers adopting cost‐effective safety standards


• Suppose the competitiveequilibrium is at E0Wage

SafetyS0

W0U0

E0

Single Firm, Single Worker

Effects of Safety Regulation


Wage

SafetyS0

W0U0

E0

Suppose regulatory board requires the level of safety to increase to sr

12

Effects of Safety Regulation• Suppose instead the firm is able to absorb the cost

‐

‐


Wage

SafetyS0

W0

U0

Sr

E0

In this case the firm is worse off

Not All Firms Will Be Affected The Same


Wage

Safety

UL Less Risk Averse

UM More Risk Averse

F1

F2

F3UA Average Risk Averse

sa sr

For Firm 3 the safety regulation is redundantFirm 2 could meet the standard only if workers are willing to give up wages for safety or if the firm can absorb some costs

13

Imperfect Information

• The safety standard “weeds out” firms whose safety technology cannot meet the standard

• At a cost to workers and firms

Imperfect Information:

• Could be that firms have more information about risks than workers

e.g.

•


Imperfect Information


Up (perceived)

Wage

Safety

Wa

E0

SpSa

Ua (actual)

In fact, she is only getting Sa safety with the wage Wa (Ua)

The worker perceives that she is getting Sp safety and knows she gets the actual wage (wa) (on Up)

14

Why Safety Regulation?Workers and firms are worse off so why?

1.

2.

3.

4.

5.


Empirical Evidence• Difficult to estimate compensating differentials(i) (ii)

• Can be explicit:

• Can be implicit:

Studies have looked at premiums for:• risk of unemployment• mandatory overtime• commuting time• risk of injury and death (most)


15

Problems

1. Increase wage due to increase risk may work the other way••

•

2. “errors in variables”•3. Sample selection bias•


Results

1.

• risk of death pays more than risk of injury

2.


1

Topic 2 ‐ Human Capital


Economics 371

Human Capital

•Under labor supply we emphasized the quantity of labor supplied but there is also a quality dimension•In the section on compensating wages we talked about how negative job attributes like risk involved compensating wages to intice workers to those jobs•This theory can also be applied to jobs that require workers to go through the costly process of acquiring human capital•Thus, human capital can influence the quality of labor supplied to the economy as well as the wage that workers are paid•We will look at two types of human capital(1) Formal education(2) Training


2

Human Capital TheoryInvestments are made in human resources to improve their productivity and their earnings

Why “investment”?

•Costs are incurred in the expectation of future benefits•Like all investments we need to ask if it is economically worthwhile (i.e. benefits>costs)

Costs (2 components):

(i)

(ii)


Human Capital Theory

• We must also distinguish between the consumption and investment components of human capital

Consumption:

Investment:

• Distinguish between private and social costs and benefitsPrivate:

Social:


3

Human Capital Theory

•Distinguish between real and pecuniary costs and benefitsPecuniary:

e.g. savings in UI that result from a retraining program‐

‐Real:‐ should be included in social costs/benefits parties making the investment


Formal EducationWe can represent the education decision as follows:


A: did not complete high school (10 yrs ed. At age 16)B: Completed high school (start at age 18)C: University/College degree (start at age 22)

Present Value of Earnings

AgeT16 18 22

4

Notes(1) Earnings increase with age but at a decreasing rate•

• Adding more to productivity and earnings early in their careers and then diminishing returns set in late in their career

(2) Earnings of those with more years of education generally lie above those with fewer years of education


Optimal Lifetime Income Stream?

First we need some simplifying assumptions

1. Individual receives no direct utility or disutility from education

2. Hours of work are fixed (includes hours in acquiring education)

3. Income streams are known with certainty

4. Can borrow/lend at real interest rate (r)


5

Optimal Lifetime Income Stream?Given these, choose the quantity of education to maximize the net present value of lifetime earnings

Costs and benefits:

(i) Attaining high‐school education


CPresent Value of Earnings

Age

B

A

T16 18 22

Opportunity Costs:

Benefits:

Costs and Benefits(ii) High School grad contemplating university


CPresent Value of Earnings

Age

B

A

T16

18 22

a

b

c

d

e

f

Costs:

Benefits:

6

Example, Present Value Calculation

18 year old high school graduate considering university:

• Work until age T (retirement)

• Income at each age if she chooses high school is YH(age)

• Assuming she works now at age 18


If she goes to university income at each age is YU(age)>YH(age)However, she will earn nothing for four years and will incur direct costs of $D per year

We could compare PV(H) to PV(U) to determine whether or not it is rational to obtain a university degree

Example, Present Value Calculation

• These calculations can be tedious

• Instead, we might simply compare the PV of benefits to the PV of costs


This is precisely the same as comparing PV(H) to PV(U)

7

Alternative Portrayal

• The individual should increase years of education until the present value of benefits (year of ed.) equals the present value of the additional cost


MB declines with years of education because:

1.2.

MC rises with years of education because

P.V. of MC/MB

Years of Ed.

Problems

1. May not have this kind of detailed information

Models still might predict well

• Most people take into account at least some of the costs and benefits

2. Simplifying assumptions may not be realistic

(i)

(ii)

• We could relax these assumptionsProfessor Schuetze - Econ 371 14

8

Education and Market Equilibrium

• The wages and education levels we observe are a result of the interaction between individuals and employers

• To understand this relationship we must look at the interaction between individuals and firms

•

•


Workers’ Preferences for Education• Suppose there are 2 types of workers

Type A: Strong preferences for education (could be because of ability or utility derived)

Type B: Dislikes education


Wage

Years of Ed.

Requires less of a wage increase to get A to increase education than B while holding utility fixed

9

Firms’ Iso‐Profit Curves• Suppose there are 2 types of firms

Type 1: Values skilled workers highly

Type 2: Values skilled workers less


Wage

Years of Ed.

Positively sloped Iso-ProfitsHigher wage can be paid to higher educated workers (greater productivity

Firms’ Iso‐Profit CurvesFirm 1: Iso‐profit is steeper

•

Firm 2: Iso‐profit is flatter

•


Employers Offer Curve:Maximum wage for each level of education

I1

I2

Wage

Years of Ed.

10

Equilibrium


The worker with more education receives a higher wage (compensating differential)

I1

I2

Wage

Years of Ed.

UA

UB

EB

EA

Empirical Evidence• Most include men only because of intermittent labor

force participation of women• Basic idea is to run regressions with lots of things that

should affect wages (occupation, industry, experience etc.) and see what effect education has

Results:1. • Income streams of the educated are above the less

educated2.

3.


11

ProblemsThere are some problems with estimating the returns to

education1. Ability Bias• Difficult to control for ability• More able are inherently more productive and may also

get more education


Problems2. Selectivity Bias• You may get the right education to prepare you for a job

you simply have an aptitude fore.g. 2 people and 2 occupationPerson A: mechanically inclined Person B: bean counterCan choose to be either a mechanic or an accountant• Need college degree to be accountant• Both A and B could become an accountant•


12

Signalling/Screening Hypothesis• Higher education acts as a filter (screening the more

able) rather than enhancing productivity• Workers signal unobserved ability and firms use

education to screen workers• Bachelor’s diploma represents a “sheepskin”Model Assumptions:• Asymmetric information

–

–

• Employers do observe some characteristics of workers


Signalling/Screening HypothesisIndices:

––

Signals:

–• Employers may form “beliefs” about the relationship

between education and productivity• Perhaps based on past experienceMarket Equilibrium:•• Employers will offer higher wages to more educated

workers if they believe there is a positive relationship


13

Signalling ‐Model• Suppose there are two types of workers

– Low ability (L)– High ability (H)

••Education (the signal) is acquired at a cost• Education is measured in years (y)• The costs are both financial and psychic

––

• Could be that it takes able workers less time or that they simply dislike school less


Signalling ‐ Beliefs• Suppose that the employer’s beliefs are as follows:

• There is some critical value of education above which individuals are believed to be type H below which they are believed to be type L

Thus, the employer’s offered wage curve w(y) will be:


W, MPE

Education1 2

1

2

14

Signalling ‐ Equilibrium• Persons who choose y<y* will set y=0•• Persons who choose y≥y* will set y=y*•

• We can add in the cost schedules to see if this is true (i.e. if this is an equilibrium)


W, MPE

Education

w(y)

1 y* 2

1

2

Signalling ‐ EquilibriumType L will choose y=0•••Type H will choose y=y*••Can show that the equilibrium requires y* to be between 1

and 2 years of education•••


15

Notes on Signalling1. There are an infinite number of equilibrium values for y*••

•

2. The education level (y*) acts as an entrance requirement for the high‐salary job

•

•


Notes on Signalling3. The private and social rates of return will differPrivate Returns –Social Returns –•

• Ignores the fact that signalling serves the useful role of sorting workers into the right jobs–

• Within the confines of the model there are more or less efficient ways of getting sorting–


16

Empirical Evidence• The few studies that test between human capital and

signalling are based on the following notion•

•

Example, Kang & Bishop (86) look at high‐school grads:• Diplomas are generally homogeneous to employers, but

can be obtained taking easy or hard courses•

Overall, these tests have not been conclusive•


Training• Form of human capitalBecker ‐ 2 types1. General Training:

2. Specific Training:


17

General Training

• Firms can bid for general training by offering a wage that is higher by an amount equal to the value of the training

• Therefore, the trainee would be willing to pay for the training as long as the benefits (higher earnings) exceed the costs

• If the company paid for training they would still have to bid against other companies for the services of the trainee


General Training

No training:

Training:


W, MRPE

Time

W*=MRPE*

WA=MRPEA

MRPET

18

Who Pays for General Training?

Employee:• Could finance training if C<B and earn w* after trainingFirm:• Could pay for training and reap the benefits• i.e. pay the worker wA before and after trainingProblem?


Specific Training

• Other firms have no incentive to pay higher wages•

•


W, MRPE

Time

W*=MRPE*

WA=MRPEA

MRPET

19

Shared Investment?• The firm faces the risk that the worker will quit after the

training (can get wA elsewhere)• Thus, the anticipated return (B) is eliminated•


W, MRPE

Time

MRPE*

WA=MRPEA

MRPET

Thus, both the employer and employee incur costs and reap benefits of training

Shared Investment?• This also minimizes the risk that either party will

terminate the employment relationshipBoth earn rents after training:Employer:

Employee:

Note:• May not be easy to distinguish between general and

specific training


20

Role of Government• Are there situations when the private market does not

provide a socially optimal amount of training?1.

2.

3.

4.


1

Topic 3 – Wage Structures Across Markets


Economics 371

Wage Structures Across Markets

• Wages vary across a number of different markets (dimensions)

• We’ve looked at how wages vary across skills (education) but there are a number of other interesting cases

–––––––

• Here, and in later topics we will examine and attempt to explain why wages vary across markets

– Do such earnings differentials indicate that the labour market is not fully integrated?


2

Earnings Function • As we have shown, a simple way to summarize and

isolate wage differences is through regression analysis

e.g. Earnings Function(1) ln(w)=a + b*School + c*Exp + d*Exp2 + є•

–• Estimation of this equation by OLS gives estimates

of the returns to schooling and experience• Within this context we can allow wages to depend

on other characteristics as well


Earnings Function

(2) ln(w)=a + b*School + c*Exp + d*Exp2 + e*Occ + f*Ind + g*Prov + є

•

• Equation 2 is estimated using the 1996 census to summarize differences in wages across these dimensions

• Let’s start with the results looking across provinces

• The table gives the percentage difference in wages across provinces “holding constant” education, experience, industry and occupation

–


3

Results


Results

• The wage premia in this type of a regression are always relative to some base category

• In this case, the base category is Ontario (largest group)

•

• As we can see there is a great deal of variation in wages across provinces

• Wages are significantly lower in the Atlantic Provinces


4

Results• The base occupation is Trades• The highest paid occupation group is

Managerial/Business Professional•

• Finally, the base industry is manufacturing• There is a large premium associated with “other

primary” : 26.2 % higher than manufacturing controlling for region etc.

• Accomodation/food/beverage industry groups have a large penalty


Reasons for these Differences

Why do wages vary across these markets?Clearly differences in educational requirements for the jobs matterBut differentials still exist even after controlling for schoolingThere are a number of reasons we might consider that apply broadly to most cases

1.

Wages will be higher in unpleasant occupations


5

Reasons for these Differences

2. Immobility Across Sectors:

e.g. ‐

‐

– Problematic because this suggests that the labour market is not integrated

3. Short‐run Disequilibrium:

4. Unobserved Heterogeneity:

With these in mind let’s examine each differential more closely


Occupational Wage Structures


• The term occupation is used to refer to a number of jobs that have “the same basic work content”

• These jobs may be found in different establishments and industries

• Occupations can be broadly defined (e.g. clerical) or very narrowly defined (e.g. typist)

•

• Otherwise workers at the margin will move from jobs of low net advantage to ones of high net advantage

6

Occupational Wage Structures


•

• The net advantage includes non-pecuniary aspects of the job, giving us our first reason

I. Occupational wage differentials may exist to compensate workers for pleasantness, safety, fringe benefits and stability of the job

••

A Simple Example

• Upward sloping supply –

• May require costly skills, preferences for certain occupations

• Supply ranks workers according to reservation wages for that occupation

•


Wage

Employment (skilled)

Ss

Ds

ws

Skilled WeldersWage

Employment (unskilled)

Su

Du

wu

Unskilled Labourers

7

Short‐run Adjustments

• Suppose technological change induces an increase in demand for skilled labour


Wage

skilled

Ss

Ds

ws

Skilled WeldersWage

unskilled

Su

Du

wu

Unskilled Labourers

II. This example can illustrate how short-run adjustments may also affect occupational wage structures

Supply schedule for skilled workers may be fairly inelastic –Wages will rise in the skilled sector in the short-runIn the long-run supply will increase in the skilled sector (reducing the wage premium)

Imperfect Competition


III. Finally, imperfect competition, which hinders adjustment to the long-run equilibrium, will also affect the occupational wage structure

Examples,1. Occupational Licensing:

2. Unions:

3. Minimum wages and pay equity work directly on wages and affect the wage structures of low-wage and female dominated occupations

8

Regional Wage Structures

• Once again, theory predicts equality of net advantage at the margin for identical jobs across regions

• Wage and non‐pecuniary aspects of the job must be the same otherwise workers would move to the region with higher “pay”

• Wages need not be the same, however, because of compensating differences, short‐run adjustments and non‐competitive factors

Compensating differences include:–––––


Regional Wage Structures

• As before, short‐run wage differentials may exist as the labour market adjusts to long‐run equilibrium

• Non‐competitive factors that hinder labour movement across regions will also impact the wage structure

Some examples are:

1. Occupational licensing – some provinces may not recognize training completed in other provinces or may have residency requirements

2. Unions –

3. Social transfer programs –

e.g. EI benefits are an increasing function of the unemployment rate in a region


9

The Migration Decision

• While we will discuss the migration decision in greater detail under immigration a simple treatment here will highlight some aspects of internal migration

• One can think of the migration decision in a similar manner as the human capital decision

•

Benefits?••Costs?••



• This framework has a number of implications which have been tested empirically – it implies:

i. Younger workers should be more mobile than older workers

•••••ii. Mobility should be from areas of high unemployment to

areas of low unemploymentiii. Mobility should increase at the “peak” of business

cycles and decline in recessions–


10


iv. Mobility to and from Quebec is likely to be lower–

v. Distance should matter–

vi. Mobility to provinces with more generous social programs, lower taxes and higher expenditure on health and education should be higher

• The empirical evidence to date appears to support these propositions

• There is also evidence that regional wage disparities are eroding over time (as theory predicts)––


Evidence

U.S. – wage differential between the North and the South (controlling for the cost of living, human capital etc.) disappeared by the 1970’s

Canada – Interprovincial differences in per capita income levels are falling

• There is some debate about whether or not regional differences still exist but there has been some movement towards equalization


11

Interindustry Wage Differentials

• The term “industry” refers to the principal kind or branch of economic activity of the establishment in which the individual works

• Industries can be broadly defined (e.g. non‐durable manufacturing) or narrowly defined (e.g. textiles)

• Because industry classification is based on the establishment, interindustry wage differentials reflect a variety of factors

–––

• To get “pure” industry differentials we need to net these out



• These difference reflect, once again:i. Compensating differentials: • Some industries are safer than others• Some industries are characterized by seasonal

employment (e.g. construction)ii. Short‐run adjustments to changes in demandiii. Non‐competitive factors••••


12


• Interindustry wage differentials likely also reflect a fourth factor; “Efficiency Wages”

•

• They may do so in order to reduce shirking, improve morale, reduce turnover and absenteeism and to elicit effort from employees

• The desirability of paying efficiency wages depends on the cost of shirking, turnover and absenteeism to the firm

•

• Econometric results suggest that efficiency wages exist – pure interindustry wage differentials exist


Public‐Private Wage Differentials

• Because the public sector is simply one industry, we might think of this differential as an interindustry wage differential

• However, the public sector has a number of peculiarities that merit special attention– It is a very large sector– Wage determination is different– Outcomes in this sector impact all sectors

• The public sector includes all sectors that are funded or owned by the government, including– Education – Health– Government enterprises– Government employment (federal, provincial, municipal)


13


• Many of the same reasons for wage differentials discussed earlier apply here

Compensating differentials:•

••

• These suggest that wages should be lower in the public sector, all else equal

Short‐run adjustments:• Recent contraction of the public sector may have lead

to downward pressure on wages making it difficult for the government to recruit and retain employees



Non‐competitive factors:• It is perhaps here that the public sector deviates most

from the private sector• We will discuss four peculiarities of the public sector

under this headingi. Political rather than profit constraintii. Monopsony power of governmentiii. An inelastic demand for labouriv. A high degree of unionizationPolitical rather than profit constraint:•

• Political constraint is generally viewed as less binding


14


•

• Future taxpayers have little or no representation in today’s political process

• That said, working against public sector workers is the fact that even during a strike revenues come in

• This isn’t true in the private sectorMonopsony:•

• There is some evidence with regard to teachers and nurses which suggests governments do utilize this power



Inelastic demand for labour:• If demand is inelastic, unions can bargain for higher

wages without having to worry about the effects on employment

•

•

• Empirical evidence suggests that the demand for labour in the public sector is highly inelastic

Unionization:•• Such non‐competition restricts the ability of wages to

adjust


15


• Most of the factors examined suggest that wages will be higher in the public sector than in the private sector

• There is another reason to believe that wages will be higher in the public sector

•

• The government would not be able to lower wages below those in other industries–

• Competitive forces do not preclude the government from paying wage that are higher, however––



• The empirical evidence suggest that public sector worker in the 90’s earned a 9% premium over private sector workers

•••


1

Topic 4 ‐ Immigration


Economics 371

Immigration

•Immigration policy has been a hotly debated topic in Canada for some timeSome of the issues discussed are non‐economic:•

–Many of the issues are economic:•

– Some concern that immigration may be leading to higher unemployment and growing wage polarization

•

•– Wages tend to be lower than native born upon arrival


2

Some Facts About Immigration•The importance of immigration in policy debates is not surprising given the flow of immigration to Canada and the increasing diversity

•Since the late 1980’s Canada has admitted over 200,000 immigrants per year

‐

‐


Some Facts About Immigration

1960’s:

2004:


3

Some Facts About Immigration

•By 2001 18.3 percent of all Canadian residents were immigrants (born outside of Canada)

•Immigrant concentrations are highest in Toronto and Vancouver

•Most new immigrants also reside in these cities

•

•


Canadian Immigration Policy

• In order to have effective immigration policy, policy makers must answer two questions

• First, what are the impacts of immigration• Second, what “tools” are available to set appropriate policy

in response to the first question• Let’s start by answering the second question and we will

return the first later• Policy makers really have two mechanisms for controlling

immigration1)2)


4

Canadian Immigration Policy

• Prior to 1967 in Canada the focus was on the first of the two mechanisms (the number)

••

• Prospective immigrants from other countries met a stiffer series of requirements for admission (Immigrants from China were not permitted)

• Only the very sick, destitute and criminals were disallowed


Canadian Immigration Policy• By the early 1960’s the influx of unskilled Europeans

(particularly from the south) led to a shift in policyThe Point System:• In 1967 the point system was introduced• Potential immigrants to Canada are categorized based

on motivation for immigrationNon‐assessed:

‐May also be for humanitarian reasons (escape political persecution or poverty)

•


5

Canadian Immigration PolicyAssessed:

• Such individuals are evaluated based on their likely success in the labour market

•

• Relatively few immigrants are assessed


The Immigration Decision • How well immigrants perform in Canada depends, in

part, on whether it is the most skilled or least skilled workers from other countries who migrate

• To look at how potential immigrants make such a decision let’s work through the “Roy Model”

Roy Model:• For simplicity consider two countries

––

• Assume that earnings in both countries depend on a single factor

•


6

The Roy Model • Let s denote the number of efficiency units (skills)

embodied in a worker• The frequency distribution of skills may look much like

the standard normal distribution• Assume, for now, that there are no costs to migration•

• There are a number of possible outcomes to this model – two of which we will illustrate


Case 1: Positive Selection


The wage-skill lines give the payoff to skills in Canada and the source country

With no costs to migration a worker will migrate to Canada whenever Canadian earnings exceed the source country

$

S (Skills)

7

Case 2: Negative Selection


Here the payoff to skills in the source country exceeds the payoff in Canada

$

S (Skills)

Implications of the Roy Model• The key implication of the Roy Model is that:

• Workers flow to markets where they can get the highest price

•

• Changes in the base income levels do, however, affect the size of the flow

• To see this consider the following example


8

Example: Change Income with Positive Selection


Suppose there is a decrease in the base income in Canada relative to the source country

Source Country

$

S (Skills)SP

Canada

Notice that we still get positive selection

However, the decrease in income reduces the size of the flow

Adding in Migration Costs• The previous result allows us to relax the assumption

that there are no costs to migration• Suppose it costs $20,000 to migrate to Canada from a

given source country• And that these costs are the same regardless of the

worker’s skills• These costs reduce the net income the worker can

expect to receive in Canada–

• This is equivalent to a reduction in income in Canada•


9

Evidence• The available data seems to suggest that income is more

unequally distributed in Third‐World Countries (e.g. India, Mexico, etc.)

••

The “Brain Drain”• Canada has a relatively egalitarian income distribution

compared to the US– Partly as a result of the tax and social system

• Relative to the US we tax able workers and insure the unskilled against poor labour market outcomes

•


The Brain Drain• Not surprisingly, this topic has received a great deal of

media attention• However, data limitations have meant that very little

research has been done on this topic• The work that has been done examines:i)ii)Size of the brain drain:• It is difficult to estimate because the Canadian

government does not collect data on those leaving Canada

• There are, however, a number of indirect sources of data


10

Indirect Sources of Data“Reverse Record Check”: Statistics Canada double checks the

coverage of the Census• Establishes whether or not people have moved to the

US• Gives no further information on immigrantsTax Records: Can identify if tax‐filers switch from Canadian to

US addresses• Again little demographic informationThe Decennial US Census: Collects information on country of

birth• This is only collected every 10 yearsUS Current Population Survey: contains lots of demographic

information• The sample sizes tend to be smallUS INS Administrative Data: Visa data


The Brain Drain•

• Most alarming is that these migrants are disproportionately highly educated

Impact of the Brain Drain:• It is unlikely that an annual outflow of 25,000 workers

will have much of an impact on the overall Canadian labour market

• However, in more narrowly defined markets emigration may matter–


11

Impact of The Brain Drain• Some have measured the impact by calculating the loss

of goods and services produced by Canadians in the US• e.g. there are approximately 60,000 highly skilled

Canadians living in the US (NAFTA Visas)•

•• Of course, with this type of calculation it is easy to “cook

the books”Final Analysis:•

• We don’t seem to lose any sleep over this!


The Impact of Immigrants• As we have already noted, effective policy requires

information about the impact of immigrants• Of course, we will focus on the labour market impacts

although there are certainly others• Let’s start by looking at what simple theory suggests and

then move on to the evidence• Theory• The “simple” model assumes that immigrant and native

workers are perfect substitutes in production•


12

Impact When Workers are SubstitutesThis case can be illustrated as follows:


As immigrants enter the labour market the labour supply curve shifts out

Wages fall from w0 to w1

Total employment increases from N0 to E1

w

NN0

w0

S

D

Impact When Workers are Compliments• However, if immigrants and natives are complements, they do not compete in the same labour market

• For example, immigrants may be more adept at certain tasks than natives

• Immigration frees up the native workforce to perform tasks that they are better suited for

•


13

Impact When Workers are ComplimentsThis case can be illustrated by looking at the market for native workers:


An increase in immigration raises the marginal product of natives

Native workers who previously did not find it profitable to work now see higher wages and enter the labour market

w

NN0

w0

S

D

Other Considerations1) Immigrants are often selected based on skills that are in

short supply•

2) Immigrants also consume goods and services•

3) Importing labour may be a substitute for importing the goods produced by that labour in the country of origin

4) Immigrants may disproportionately invest in Canadian firms, creating employment

• Thus, the answer to the question “What is the impact of immigration?” is ambiguous and left to empirical studies


14

Empirical Studies on Impacts

Estimating the impact of immigration turns out to be quite difficult

Problems:1. What is the relevant labour market?•• In any case, there are likely linkages2. Result depends on whether immigrants are compliments or

substitutes etc.3. Immigrants do not arrive randomly to a market••Nevertheless there are a large number of studies that attempt to

estimate the impacts


Empirical Studies on ImpactsMost studies find that immigration has little impact on the

labour market outcomes of nativesApproaches:1. Exploit cross‐city variation in immigrationExample: compare Toronto to Montreal•

• This has been done using US data and the results suggest that there is little impact

Clearly there are some problems with this approach•

–• There is still the endogeneity problem –


15

Empirical Studies on Impacts2. “Natural experiments” to avoid endogeneity• To get an exogenous change in immigration David Card

looks at the impacts of the Mariel Boatlift on the Miami labour market

• Between May and September of 1980, 125 thousand Cubans were permitted to leave Mariel Cuba for the US.

• Many settled in Miami• Again, however, there was little evidence of an adverse

effect

• Another “branch” of immigration research examines the labour market performance of immigrants


Immigrant Assimilation• There is a general concern that immigrants do not

succeed in the labour market and become a burden• This research examines whether or not these concerns

are founded• We might expect the earnings profile of an immigrant

and native worker to look as follows:


Controlling for hours workedThe workers are similar in

every aspect except originAssume the immigrant arrives

at age 20 and retires at age 65YSM = years since migration

Native Born

earnings

Age20(YSM=0)

65(YSM=45)

16

Immigrant Assimilation

• With age (experience) both workers’ earnings will rise•

• In this case wages catch up at age T but this age could exceed 65• If there is “positive selection” the earnings of immigrants may

overtake those of comparable native born workers•


Initially the immigrants earnings may be below a similar native-born worker. Why?

Native Born

earnings

AgeT

{

Immigrant

entryeffect

20(YSM=0)

65(YSM=45)

Empirical Findings

• Need panel data to sort between cohort effects and assimilation

Men:

•–

• The assimilation rate is very small; Bloom et al estimate it to be 0.25% per year

–Women:

•


1

Topic 5 ‐ Discrimination


Economics 371

Discrimination

What is labour market discrimination?•Occurs when participants in the marketplace take into account such factors as race and sex when making economic exchanges•Even after accounting for the skills of workers and job characteristics race or sex influences decisions•When thinking about wages could be discrimination in setting wages or in providing access to human capital or labour markets

What leads to discrimination in the labour market?1.•


2

Discrimination2.

• Could come from employee who underestimates his/her labour market capacity

• Could come from employer who consistently underestimates value of group of workers

3.

•• More likely to occur when information on individual

workers is costly– e.g. employers judge pay decisions of a woman based on the

turnover rate of all women

– Difficult to signal your true “type”


Discrimination4.

One group of individuals may try to protect their high‐wage jobs from low‐wage competition–

Who discriminates?

• Through wage, hiring and promotion policies

• Through unions, apprenticeship systems etc.

• Simply do not purchase goods from establishments that employ certain groups


3

Labour Market Discrimination Theories

•We can classify theories of discrimination based upon whether the focus is on demand, supply or noncompetitiveaspects of the labour market

Demand Theories:

•The demand for labour of a group being discriminated against is reduced relative to other workers

Outcome Without Discrimination:

•Workers with different racial background or gender and similar skills are perfect substitutes

•q=f(NA, NB)

•


Demand Theories of Discrimination

• Thus, the marginal product of workers is the same as well• The firm’s hiring decision is simple:• Hire workers up to the point where MRPN equals the wage• If the wages of group A are less than group B hire only

worker from group A With Discrimination (against group A):• The employer acts as if group A’s wage is not WA but

WA(1+dA); dA is the “discrimination coefficient”Employee discrimination: Consumer discrimination:


4

Demand Theories of Discrimination

• The employer’s hiring decision is not based on a comparison of wages WA to WB

• Instead, the employer compares the utility adjusted price

• The decision rule is:

• Hire only group A workers if

• Hire only group B workers if

• Clearly, this will lower demand for those workers who are discriminated against and lead to segregation in the labour market


Demand Discrimination Illustrated


N*A

wA

w

NNDA

w(1+dA)

MRPN

Group A

N*B

w

N

wB

MRPN

Group B

Clearly, for both types of firms discrimination will not be profitable

5

Supply Theories of Discrimination1. Crowding Hypothesis•

– e.g. women into “pink‐collar” jobs• The relatively high supply in these sectors leads to low

wages even though workers are paid their marginal product

2. Dual Labour Market• Suggests there are two distinct labour marketsCore labour market: Secondary market: • Discrimination forces workers into the secondary market

through the use of unions, barriers to education and training etc.


Noncompetitive Theories of Discrimination• The predictions of the perfectly competitive model we

reviewed lead to segregation not wage differentials• Competitive forces should lead to an equalization of

wages yet we don’t observe this for some groups1. Arrow (1973) suggests equalization might occur in the

long‐run but the long‐run is a very long time•

2. Queuing of workers, such as that caused by efficiency wages may allow firms to ration jobs on the basis of discrimination


6

Noncompetitive Theories of Discrimination3. Discrimination may occur in noncompetitive sectors

4. Historically determined practices•


Measuring Discrimination• In measuring discrimination you want to record

evidence of unequal pay or promotion that is net of productivity differences

• However, productivity is difficult to measure• It is easy to identify discrimination in cases where two

individuals are doing the same job but are paid differently

• There are likely relatively few cases like this and we rarely get this kind of data

• Thus, most studies look at differences among groups and try to control for productivity

• Any difference that can’t be accounted for by productivity is attributed to discrimination


7

Oaxaca Decomposition• This procedure decomposes the overall difference in earnings between two groups into two parts1. 2. • For example, suppose we wish to explain the earnings gap between men (M) and women (F)Earnings gap = YM/YF or YF/YM

• Using human capital models of earnings:

XF andXM are productivity characteristicsβF and βM are the rates of return to these characteristics


Oaxaca Decomposition• Taking logs we can rewrite the male‐female wage ratio as:ln(YM/YF) = lnYM – lnYF =βM•XM ‐βF•XFAdding and subtracting βM•XF we get:lnYM – lnYF =βM•XM ‐βM•XF ‐βF•XF +βM•XF

•

(Both men and women are paid at the men’s rate βM)•

• This is usually labelled discrimination


8

Graphical Illustration ‐ Oaxaca


Equations give the relationship between income and characteristics for men and women, separately

Steeper male line suggests that men receive a higher return to productive skills

Intercepts suggest men also receive a higher base pay

lnYm

XXf

lnYm=ßmXm

lnYf=ßfXf

Xm

lnY

lnYfA

BC

D}

}

Problems?

1) What are the appropriate X variables?i) Education and experience?•

ii) Industry and occupation?•

iii) Demographic variables like marital status and number children?

•2) What variables are missing?•


9

Empirical Results • Despite some of its problems the Oaxaca decomposition

has been used to examine a number of earnings differentials

Male/Female Differentials:• Lots of variation in the results; depending on approach

and x‐variables includedGenerally find:1. Pure wage gap remains even after controlling for a number

of wage‐determining characteristicsBaker et al. ––Kidd and Shannon––


Empirical Results 2. Potential discrimination outside the labour market is

importante.g. unequal division of labour with respect to childraising; affects

hours worked, absenteeism, mobility, training, career interruption

Kidd and Shannon (1997)––Hersch and Stratton (1997)–

3. The occupational segregation of women into lower paying jobs does not explain much of the earnings differential

• Perhaps too coarse to measure segregation into low‐wage firms and industries


10

Empirical Results 4. The unexplained portion tends to be smaller in the union

sector and in the public sectorEthnic/White Differentials:• Confounded by the fact that many nonwhites in Canada

are also immigrants–

• Some compare earnings of whites in Canada to various ethnic groups born in Canada

• The unexplained gap tends to vary by ethnic group–

• Unexplained gap is smaller in the US• Rankings are also different


Policies to Combat Discrimination

1. Conventional Equal Pay Legislation

• Deals with wage discrimination within the same job in the same establishment

••• Only minor differences are allowed especially when offset by

other tasks

Example: equal work even if males must do heavy lifting especially when women do other tasks

• All Provinces and Territories in Canada are covered


11


2. Pay Equity (Comparable Worth) Legislation:• Recognizes that Conventional Equal Pay Legislation is limited

in scope••

Example: Male‐Female Procedure• Compare jobs that are predominantly male to those that are

predominantly female (say 70% or more)• Points are allocated to job attributes such as skill required,

effort, responsibility and work conditions


Pay Equity• The points are summed for each job and “female” jobs

are compared to “male” jobs• Wages are then adjusted to reflect the “value” of the job

done•

• It is felt that market forces reflect discrimination• The equal value concept is determined by the notion of

value‐in‐use•• Rationalized on grounds of being able to deal with both

wage discrimination and Occ. segregation


12

Pay Equity• Who is covered by Pay Equity?

• In Canada, all jurisdictions except Alberta and Saskatchewan

• In most jurisdictions coverage is restricted to the public sector

• Quebec and Ontario are the only two that require it to be applied to the private sector as well



3. Equal Opportunity Legislation:

• Designed to prevent discrimination in recruiting, hiring, promotion and dismissal

••

• Hoped that this will increase demand for female employees and, in turn, increase wages

• Likely to benefit new recruits and those changing jobs more than incumbent females


13


4. Employment Equity (Affirmative Action)

• Focuses not on providing equal opportunities for work but on results or quotas

•

• Employment equity applies to four designated groups:

• Essentially, goals are set so that the internal representation of workers match their representation in the relevant external market

• Employment equity covers only federal workers in Canada


Estimated Impacts of Policies

Conventional Equal Pay• Most studies find equal pay for equal work policies have very

little effect on wages–

Pay Equity (Comparable Worth)• A number of studies have examined the impact of

comparable worth• Typically examine its application in the public sector in

Canada and the USe.g. Public Service Alliance of Canada PSAC alleged that female

clerical workers were paid lower wages than the male dominated Program Admin. Group

• Ended in a 3.6 billion dollar settlement


14

Estimated Impacts of Pay Equity

• The female/male earnings ratio in the overall public sector before comparable worth adjustments was around 0.78

•

• Note that these studies don’t likely give a good indication of their likely impacts on the broader economy

– Gap in public sector is just over half of that in the economy as a whole

• Impact is likely to be smaller because the gap reflects the segregation of women into low‐wage firms and industries

– Comparable worth doesn’t compare across firms and industries


Estimated Impacts of Policies

• Evidence of the impact of comparable worth in the wider labour market confirms this

• e.g. Baker and Fortin (2000) examine the impact of Pay Equity in Ontario

••Affirmative Action:• Studied in the US by examining the impact of the US federal

contract compliance of Executive Order 11246• Minority groups targeted did benefit to a certain extent• The effects were stronger during economic expansion than

contraction in the overall economy


1

Topic 6.1 ‐ Union Growth and Incidence


Economics 371

Unions:


“Collective organizations whose primary objective is to improve the well-being of their members”Achieve this through “collective bargaining”

In “North America” bargaining is done primarily at the firm levelUnions are also often involved in government lobbying In European countries the ties between unions and political parties are particularly close

e.g. The Labour Party in the UK

2

Unions in Canada


In Canada Unions have relatively close ties with the New Democratic PartyThere are two types of Unions in Canada

Craft Unions:

Industrial Unions:

Union Membership in Canada


3





Unlike in the US, the percentage of workers covered by a “collective agreement” in Canada has increased (or stable) in recent yearsTable 14.1 presents a number of different measures of “union density”

Non-agricultural paid workers exclude a number of individuals who are not eligible for unionization

e.g. self-employed and agricultural workers

In fact, these measures likely understate the fraction of workers influenced by collective agreements

4



First, there is a distinction between union “membership” and union “coverage”

Once certified the union is the exclusive bargaining representative for a bargaining unitThus, union coverage is likely even higherSecond, a unions influence may reach far beyond those covered by a collective agreement

e.g. Steel industry in Hamilton – Stelco (unionized) while Dofasco (not unionized)

Legal Framework


The law with respect to collective bargaining in Canada has passed through three main phases

Phase 1 (prior to confederation):

Phase 2 (Trade Unions Act of 1872):Saw the removal of many of the restrictions on union formation and collective action

5

Legal Framework


Phase 3 (1944 National War Labour Order):Modeled on the Wagner Act of 1935 in the USLabour law facilitated the formation of unionsEstablished employees’ rights to union representation and collective bargainingAlso established the certification process and limited the interference of employers with union establishment

Certification process:The process by which unions come to represent workers is somewhat complex and varies by province (power is divided between federal and provincial governments)

Certification Process


Basically, an application is made by the union to the province’s Labour Relations BoardCertification depends on the percentage of workers who indicate their desire for organization by signing union membership cards

In some provinces, even if this threshold is not met an election may be held to determine if representation is given to the union

6

Determinants of Growth and Incidence


What explains the large variation in union density?We have seen that union density in Canada varies through timeThe extent of unionization in Canada also varies by industry, occupation, region and other characteristics

Union density and growth also varies by country Most notable are the divergent trends in unionization between Canada and the US. (Figure 14.1)

Demand and Supply Framework


We can analyze the reasons for variation in union density using a demand and supply framework

Demand for Union Representation:Lies with the employeesDepends on the benefits and costs associated with union representation

Benefits: Costs:

7

Demand and Supply Framework


Supply of Union Representation:Lies with union leaders and their staffUnions allocate their scarce resources to achieve the unions objectiveAdministering contracts and organizing new workplaces are obviously costly activities

Determinants


Thus, the extent of unionization depends on:1. Demand –

Demand will be high in organizations where many employees perceive the benefits of unionization to be substantialDepends on the actions of the employer and the political, social and economic environment at any point in time

2. Supply –Supply will be high in organizations in which per worker costs of organization/representation is lowDepends on the legal and political environment also

8

Empirical Evidence


The demand and supply framework has been used as the basis for a number of empirical studiesSeveral factors suggested by the demand and supply framework have been investigated using cross-sectional data

These include:i) Social Attitudes Towards Unions

The idea here is that the resistance to unionization and the receptiveness of employees to union representation are affectedThe difficulty is measuring social attitudes

Empirical Evidence


Example: Ashenfelter and Pencavel (1969)

Find that the percent of Democrats is positively correlated with union growthThis can’t help to explain Canada/US differences in unionization as attitudes towards unions has become increasingly unfavourable in both countries

9

Empirical Evidence


ii) Legislation Towards UnionizationThis is somewhat related to the previous factor as legislation is often influenced by social attitudes

Changes in the legislative framework in Canada have been used to explain the growth in unionizationThese changes led to decreases in the costs of unionization for employeesDifferences in legislation have also been used to explain the divergent trends in unionization between Canada and the US

Empirical Evidence


iii) Other Economic and Social LegislationLegislation that raises employment standards can narrow the gap between union and non-union workers

iv) Aggregate Economic ConditionsThe Business cycle may also influence the rate of union growth

10

Empirical Evidence


Economic conditions work through:The number and growth rate of employees who are eligible for unionizationIts influence on the resistance of employers to union formation

The unions ability to get wage increasesThese suggest that union growth is procyclical

Empirical Evidence


v) Firm and Industry CharacteristicsStudies find that union density is higher in industries where

a) It is more difficult for a single person to be heard in a large firmLikely that the need for work rules, grievance procedures etc. is greaterPer worker cost of union organizing is lower

b) Implies labour costs are a small fraction of total costs

11

Empirical Evidence


c) Barriers to entry may limit the threat of union competition in the form of new entrantsPotentially greater benefits due to barriers to entryOrganizing costs may be lower

vi) Personal CharacteristicsThe characteristics of the individual affect the demand for union representation

e.g. part-time workers are less likely to be unionizedThe benefits are likely lowerAge, gender and the individuals earnings level have also been found to influence the demand

1

Topic 6.2 ‐ Unions:Wage and Employment Determination


Economics 371

Wage and Employment Determination

How do firms and unions interact in the setting of employment and wages?•The view taken in the economic literature is that unions maximize an objective (utility) function subject to firm behaviourUnion Objectives:What do unions want?•“more” – Samuel Gompers, founder of the American Federation of Labor•


2

Union Objectives

•Note that the union’s objectives are not necessarily the same as those of the union members•

•

•Nonetheless it is useful to think of the union acting as a single decision making unit in search of higher wages and employment


The Union’s Indifference CurvesThus, the union’s indifference curves have the usual shape


Downward Sloping:• High wage is necessary to

compensate for low employment

Convex:• With a high wage and low

employment willing to give up a lot of wages to increase E

•

Real

Employment (E)

Wa/PU0

U1

U2

Wage(W/P)

The worker’s alternative wage also matters

3

Alternative Objective Functions• Other union objective functions have also been suggested1. Maximize the wage rate• Place all the weight on wages and none on employment


Real

Employment (E)

Wa/P

Wage(W/P)

• Utility increases “vertically” only

Alternative Objective Functions2. Maximize employment


• Utility increases “horizontally”

Real

Employment (E)

Wa/P

Wage(W/P)

Recall that union membership matters to the union because it increases union dues

4

Alternative Objective Functions3. Maximize the real wage bill (W/P•E)• The real wage bill is just total labour income• Here the union places weight on both wages and employment


• W/P•E =constant•

•

Only makes sense if the income is shared between employed and unemployed members

Real

Employment (E)

Wa/P

Wage(W/P)

Alternative Objective Functions4. Maximize real economic rent (W/P‐Wa/P )•E• Similar to profit maximization for a monopolist

–• The alternative wage represents the opportunity cost to each member


W/P•E = total revenueWa/P•E = total cost

• Indifference curves are rectangular hyperbolas

•

Real

Employment (E)

Wa/PU0

U1

U2

Wage(W/P)

5

Union Constraints• The “choice” of wages and employment by the union is constrained by the firm’s behaviour

• Assume that the firm is dealing with a profit‐maximizing competitive firm

• Also assume (initially) that the determination of wages and employment is carried out in two stages

•

•

• The firm need only to look to its labour demand curve which specifies the profit maximizing employment level at each wage


Union ConstraintsIn a sense, the firm’s labour demand curve can be viewed as a constraint on union behaviour


• If the firm can not be induced off of the demand curve

• Union utility maximization occurs at a point of tangency (a0)

a0 represents one of the possible outcomes from collective bargaining in this model

Real

Employment (E)

Wa/P

Wage(W/P)

DL

6

The Firm’s Preferred Outcome• The firm prefers outcomes that yield higher profits

• Profits vary along the firm’s labour demand curve


• To see this consider the firm’s isoprofit curves

Isoprofits lower on the demand curve are associated with higher profits (preferred by the firm)

a

∏0

W* ••

•

• b

c

d

∏1

∏2

E*

•

•

E

Wa

Wage

DL

Bargaining Range• The range of wages along the demand curve over which bargaining can occur is constrained


As drawn the zero profit constraint is not bindingThe union prefers a wage less than W0 (IU) because of the negative employment effects of a higher wage

1. The lowest possible wage the firm can negotiate is Wa

-

2. The zero-profit isoprofit limits the wage the union can ask for

-

∏0=0WU •

•

∏*E

Wf =Wa

Wage

DL

U*

W0

7

Some Implications I. Elasticity of Demand:• In the absence of the union the competitive wage is given by Wa• Thus, if the union has any bargaining power wages will be higher and employment lower with the union• Because of this the likelihood of a successful union drive and union utility increase when the labour demand curve is inelastic

–

–


Some Implications • Not surprisingly, a number of union practices are aimed at making labour demand more inelastic

These include:

Reducing the number of substitutes for union labour

••

Reducing substitutes for union‐made products

•••


8

Some Implications II Labour Market Efficiency:

• The outcome implied by the model is inefficient because unions reduce the total value of labour’s contribution to national income

• If unions are able to raise wages (reduce employment) in union sectors employment increases in nonunion firms (if available etc.)

• The last worker hired by a nonunion firm would have greater productivity if she moved to the union sector

•


Efficient Contracts• The fact that the previous equilibrium is inefficient suggests there might be a “better” contract off of the demand curve


When will all the gains from trade be exhausted?No change can make one party better off without

making the other party worse off

Point A: represents the “ideal” union outcome in previous model

∏A

WU •

•

∏B

E

Wa

Wage

DL

A

UAB

W0

UA’’

Point B: Again there is a shaded region in which both parties benefit

9

Efficient Contract (A’’)

• A’’ is a “Pareto‐efficient” contract•

•


•A’’

∏A

WU •

•

∏B

E

Wa

Wage

DL

A

UAB

W0

UA’’

At A’’ all of the gains from trade are exhausted

This occurs at a point of tangency between the union’s indifference curve and the firm’s isoprofit curve

Contract Curve

• Points up on C‐C’ are preferred by the union and those lower on C‐C’ are preferred by the firm

•Bargaining range is determined by zero economic profit and the workers’ alternative wage


Contract Curve = C - C’

The locus of all of the relevant Pareto-efficient wage-employment contracts

•

••

E

Wa

Wage

DL

W0

The contract curve lies to the right of DL

10

Implications 1. Featherbedding• Outcomes on the contract curve contain higher levels of employment than the firm would choose on its own• The firm is overstaffed

• The firm and the union will be forced to negotiate “make‐work” or featherbedding practices to share tasks•


Implications 2. Relationship between wages and employment

• The negative “ceteris paribus” relationship between wages and employment need not hold

•

3. “Efficiency”

• Although the term “efficient contract” is used for all contracts on the contract curve they may not be allocatively efficient

••


11

Obstacles to Reaching Efficient Contracts• There are obvious incentives for firms and unions to reach an agreement on the contract curve• These agreements may, however, be difficult to reachWhy?1. Imperfect Information• May not realize that there are gains to be made if there is not full information about willingness/ability to trade•


Obstacles to Reaching Efficient Contracts2. Difficult to enforce employment contracts

• The firm has an incentive to reduce employment at the negotiated wage (reach the demand curve)

•

•


12

Bargaining Theory


The union models we have examined so far suggest that there is a range of possible outcomesSome of these outcomes are more preferred by the firm and some are better for the union

Bargaining Theory is used to:1.2.

The basic idea is that the union and firm will engage in strategic behaviour (like a card game or chess)Both parties conjecture about the potential actions of their collective bargaining partner

The Basic Bargaining Problem


There is a set of characteristics that is common to all bargaining situations

For the bargaining problem of a collective agreement between a firm and a union the problem might be illustrated as follows:

13

Firm/Union Bargaining Problem


The area bounded by the curve represents the set of feasible outcomes

Bounded by the firm’s profitsPoints on the boundary are Pareto EfficientTo see this consider point C (inside the boundary)

Union

Firm Utility

A

B

C

Utility

d

d = disagreement or “threat point”

Firm/Union Bargaining Problem


Conflict arises because the set of Pareto Efficient outcomes yield higher utility for one party at the cost of the otherPoint B is also Pareto Efficient but yields higher utility to the firm than A and lower utility to the union

Union

Firm Utility

A

B

C

Utility

d

14

Bargaining Problem Solutions


There are two classes of solutions:1.

Give a set of properties that describe the outcome2.

Model the process of bargaining along with giving predictions about the outcomeWhat follows are examples of each of these

The Nash Bargaining Solution:Follows from the work of John NashAssumes perfect information about the possible payoffs and preferences

Not about what the other party will do

The Nash Bargaining Solution


Outlines four axioms that a solution to the bargaining problem must obey

1. The outcome must be Pareto Efficient

2. If the bargaining set is “symmetric” the solution must give equal utility increments to each party

Bargaining power depends on possible outcomesWith symmetry both parties have the same amount of bargaining power

15



3. The solution is not altered by a linear transformation of either party’s utility functionThe units that utility is measured in should not matter

4. Independence of Irrelevant Alternatives

The basic idea is as follows:Suppose you “play the game” with all possible outcomes and come to a solution



If we remove some of the possible outcomes (other than the solution) we should get the exact same outcome

Example: Deciding on how to get to schoolChoose between: bus, car and bikeSuppose you choose to ride your bikeYou find out that, in fact, the buses are not running

It is a little more complex in a two person situation and it is an axiom that is often violated in experiments

16

Nash Equilibrium


It turns out that these four axioms imply a unique solution to the bargaining problemThe “Nash Equilibrium” is such that the product of the two parties’ utility increment is maximized

Union

Firm Utility

NUN

Utility

d

FN

UdFd

In terms of the graph

Rubinstein’s Bargaining Theory


Rubinstein models the bargaining processClearly there are a number of different ways in which firm’s and unions will interactThe interaction will depend, in part, on the “rules” of the game

The rules in Rubinstein’s game are as followsThe bargainers take turns making offersThe offer is either accepted or rejected

17

Assumptions


Perfect informationEach party cares only about the utility derived at the end of the processEach round in which an agreement is not reached is costly to both parties

Potential profits from bargaining (boundary) decrease

First period bargaining boundary = U1 – F1

Second period falls to U2 – F2

U

F

U1

d

F1

U4

F4

U2

U3

F2F3

Assumptions


Each party acts rationally and can expect the other party to do the same

Example: The union starts the bargaining by offering A and only A in the following graph

U

F

U1

d

F1

U4

F4

U2

U3

F2F3

A

B

i.e. otherwise end up at dThis is not a credible threat

18

Equilibrium – Union Makes First Offer


Period 2: the firm moves, period 3: the union moves, period 4: no decision to make The best the firm can do in period 4 is F4

Period 3:

U

F

U1

d

F1

U4

F4

U2

U3

F2F3

R

FR

UR

The equilibrium will occur at R if the union moves firsti.e. the union will offer and the firm will accept RTo see this we must use backward induction

Equilibrium – Union Makes First Offer


Period 2:

The firm will offer (U3,FR)Period 1: The same rational leads to the union offering

and getting (UR,FR)

U

F

U1

d

F1

U4

F4

U2

U3

F2F3

R

FR

UR

19

Delay Costs


Notice that delay costs are important in that the threat of these costs give the negotiators powerIn fact, what really gives a party power in negotiations is the relative delay costs

To see this consider the following diagram:

U

F

U1

d

F1

U4

F4

U2U3

F2F3

The feasible set in period 1 is the same as beforeDelay costs to the firm are also the same

Delay Costs


U

F

U1

d

F1

U4

F4

U2U3

F2F3

R The relative delay costs for the firm have increasedThis decreases the bargaining power of the firm and increases that of the union

20

Some Final Words


Clearly, Rubinstein’s model is overly simplistic

Captures the importance of relative delay costs in determining the equilibrium outcomeMight help to explain why strikes or lockouts are important in bargaining

Doesn’t help to explain why strikes actually occur

Some Final Words


The strike is irrational

The irrationality of strikes is known as the “Hicks Paradox”Most models that attempt to explain strikes assume that there is asymmetric information

In such circumstances strikes may make sense

21

More on Bargaining Power


So far, we have thought of bargaining power as the ability of one party to raise (lower) wages

However, union bargaining power is also related to the elasticity of labour demandHere the notion of bargaining power is associated with the union’s willingness to raise wages

It is possible for a union to be powerful in one respect but not in the other

More on Bargaining Power


Of course, unions can be powerful or weak according to both meanings as well

w

E

A

wa DL

wU

wC

w

E

B

wa DL

wU

wU=wC

1

Topic 6.3 ‐ Union Impacts


Economics 371

Union Impact on Wages and Benefits


A great deal of research has focused on estimating the impact of Unions on wagesA simple approach to doing this is to compare the average wages in the two sectorsUnion –nonunion wage differential:

There are a number of problems with using this simple method

2

Problems/Difficulties


1. The nonunion wage that prevails in the presence of unions may differ from that which prevails in the absence of unions (endogenous)

As we noted previously, the presence of unions may also influence wages in the nonunion sector

a.

To illustrate, consider the following two-sector general equilibrium model

Two‐Sector General Equilibrium Model


Both sectors employ the same type of labourIn the absence of unions the labour market is competitive with

equilibrium wage w0

NA0

w0

w

N

DA

A UnionSA

NB0

w0

w

N

B Non-UnionSB

DB

Suppose unions enter and raise the wage in sector A to wU

Employment decreases to NA1

3



Notice that the estimated union-nonunion wage differential is bigger than the true union impact

Other possibilitiesb.

We need to distinguish between those that are willing to pay higher wages and those that are notSome may face very little threat of unionization.

General Equilibrium With Threat Effects


Sector B raises its wage to compete (wBN)

Employment declines in sectors A and B

The average non-union wage can be higher or lower than the previous case

However, the union-nonunion wage differential still doesn’t measure the union impact on wages

NA1

wU

NA0

w0

w

N

DA

A UnionSA

NB0

w0

w

N

B ThreatenedSB

DB

NB1

wBN

S’B

}a

}a

}b

NC0

w0

w

N

C UnthreatendSC

DB

S’C

}a

4

Vertical Contract Curve


c. It is possible that the firm and union will negotiate off of the demand curveAn interesting case occurs when the contract curve is vertical

wU

NA0

w0

w

N

DA

A UnionSA

NB0

w0

w

N

B Non-UnionSB

DB

C

Union Sector Queuing


d.

Wages are higher in the union sector, thus, it may be rational to wait

e.g. --Here the increased supply of workers to the non-union sector will be less than the reduction in employment in the union sectorWages in the nonunion sector won’t be depressed quite as far

5



2. Selection BiasWhat we want to do is compare differences in wages controlling for other wage determining differences

Examples:Unionized sector may utilize skilled workers more than the nonunion sector

Unionized firms may be larger (have more capital) than non-unionized firmsSome of these differences can be controlled for in a regression setting because they are observable



The problem arises when the unobserved factors that influence wages also influence selection into the union sectorThe estimated effect will be biased

Example:

This is likely to be the case if unions are able to create queues of workers wanting higher wagesThe solution is to use panel data

6

Empirical Results


Estimates of the union-nonunion differential vary depending on statistical method and data usedEarly studies, primarily using US aggregate data estimate the union-nonunion differential to be around 10-15 percent (Lewis 1963)

Because these studies are based on aggregate data they are not very reliableStudies based on individuals or individual firms estimate the impact to be greater than 10-15 percent

Empirical Results


However, it has been argued that such studies may overstate the true union wage impact

Recent studies attempt to deal with these issuesAccount for the joint determination of union status and the wage impactThese studies find the impacts to be smallerLongitudinal studies have been used to control for person-specific characteristics (not observed)

7

Empirical Results


US longitudinal studies find the union-nonunion wage differential to be around 10 percent

Others have suggested that the lower estimates may result from mismeasurement of union status in longitudinal data

Variation in Impacts:1. The differential tends to be higher the higher the

proportion of the industry/occupation organized

Empirical Results


2. The impact is larger the larger the firmThis is even more true in the non-union sectorTherefore, the wage differential declines with firm size

3. The union impact is higher for blue collar than white collar workersIn general, the impact is larger at low skill levels and smaller at high skill levels

8

Empirical Results


4. In Canada, it appears that women benefit more, in terms of wages, than men from unionization

5. The union impact tends to be smaller in the public sector than in the private sector

6. The union wage impact varies countercyclicallyWidening in recessions and narrowing in booms

Non‐Wage Impacts


i) Unions tend to have a larger impact on non-wage benefits than they do on wages

ii) Unions also tend to reduce worker “turnover”Quit rates are lower in union firms even after controlling for wage differentialsMay reflect the fact that workers are reluctant to leave these high wage jobs

9

Non‐Wage Impacts


iii) Empirical studies find that unions may either increase or decrease worker productivityThe overall impact depends on the magnitude of various offsetting factorsIt is important to distinguish between

a.

b.Union firms will respond to an increase in the wage by:1.2.3.

Impacts on Productivity


Thus, we might expect the union wage impact to raise productivityTo get the underlying impact of unions, therefore, we must control for capital, worker characteristics and job characteristicsThrough work stoppage and featherbedding we might expect unions to have a negative impact on productivityOn the other hand, reduced turnover and improved worker morale may increase productivity among union workersEstimates of the impacts of unions have been both positive and negative

10

Impacts on Productivity


The estimates depend on:

1

Topic 7.1 Unemployment Measurement


Economics 371

Unemployment

• In general, those not currently employed but who indicate that they want to work at the prevailing wages and working conditions.

• The most common measure is obtained from the monthly Labour Force Survey (LFS)

• In this survey people are asked about their labour market activities in the week prior to the survey

•

• The unemployed also includes those who were on temporary layoff or had a job starting within 4 weeks

•


2

Unemployment

Why is there unemployment?• The Neoclassical model suggests that prices should adjust to

reach an equilibrium in which all those wishing to work find employment.

• To see why some unemployment might be expected even with our simple supply and demand model let’s look at the different types of unemployment.

• However, we will need more than this to explain the existence and persistence of unemployment.

• We will examine some of the explanations proposed to explain persistent unemployment as well.


Types of Unemployment:

(i) Frictional Unemployment:

• Unemployment associated with normal turnover in the labour force.

• Thus, unemployment would prevail even in a well functioning labour market.

• New jobs open up, others disappear, workers return to school.


3

Frictional Unemployment

• Often associated with job search which takes time.How much time? → “Optimal Search”• Job vacancies and job seekers exist because of imperfect

information.Costs:•

•

Benefits:••


Optimal Search• Assume that individuals choose search activities to

maximize expected utility–


Initially MC of search is low

More costly methods will be necessary as search continues

MB likely declines with search time

Equilibrium search time = Se

“expected” amount of search

B

CP.V.

totalC/B

P.V.

MCMB

Search Time

Search Time Se

4

Frictional Unemployment

It may not be beneficial to reduce frictional unemployment.•••

Temporary layoffs due to• changes in product demand• weather• work stoppagesare part of frictional unemployment



(ii) Structural Unemployment:

• Results when the skills or location of the unemployed are not matched with the characteristics of job vacancies.


Like with frictional unemployment, unemployed workers and job vacancies co-exist.

5

Structural Unemployment

Here, successful matching requires more than the acquisition of information.

••

•Solutions:

•••


Structural Unemployment

Whether frictional or structural depends on whether changes in demand are permanent or temporary.

Permanent →Temporary →• Not always clear which is the case.


6


(iii) Demand Deficient Unemployment:• When there is insufficient aggregate demand in the

economy to generate sufficient job vacancies.• Doesn’t arise because of a lack of the correct skills or

by being in the wrong labor market or because of search.

Supply =Demand =


Demand Deficient Unemployment

• Usually associated with the business cycle “cyclical unemployment”.

• Could be chronic as opposed to short‐term.


7


(iv) Seasonal Unemployment:• Associated with insufficient demand in a particular

season.• Different from demand deficient unemployment in

that it is not a shortage of aggregate demand but rather a shortage in a particular season.

• Usually predictable over the year and specific to particular industries.

e.g. Unemployment increases in the winter in:•••


Seasonal Unemployment

Labour supply may also increase

• Seasonal unemployment is analogous to temporary layoffs in several ways – often included in frictional unemployment.

Seasonally adjusted unemployment rate:•

•


8


(v) Involuntary Unemployment:

• It is important to distinguish between voluntary and involuntary unemployment.

Involuntary unemployed:

Voluntary unemployed:


Unemployment in Canada


9



Canada’s unemployment rate fluctuated a great deal during this periodTypically because of cyclical fluctuationsGreat depression began in 1929 and the unemployment rate soared up to almost 20%



From the 1960’s on unemployment rates have trended upward

This trend broke somewhat in recent years as unemployment rates fell to levels not seen in almost 30 years

10

Hidden Unemployment (marginal labor force attachment)


Refers to situations in which individuals may be without work yet desire to work but are not classified as unemployed

Examples:1. “discouraged worker” phenomenon:

Individual would work at the going wage but does not seek work because he/she believes there is no work available

Examples of Hidden Unemployment


2. Those waiting for recall after 6 months of layoff3. The underemployed

Those working fewer hours than they desire

There is evidence of significant hidden unemploymentSome labour force surveys identify those who want to work, are available but not seeking

11

Should they be “unemployed”?


Yes:

No:

Note: the number of discouraged workers increases during recessions

Labour Force Dynamics


The measures we have discussed so far provide a snapshot at a point in time

i.e. an estimate of the stock of persons in each labour force stateHowever, even if the stocks remain constant it could be a mistake to say that the labor force hasn’t changedFigure 17.2 gives the monthly flows between three labor force states: employed, unemployed and the non market sector between 1976 and 1991

12





Even though the number of employed changes only marginally from month to month, many of the unemployed in one month are no longer unemployed the next month

13

Incidence and Duration


It is important to understand how stocks and flows influence unemployment

Incidence of unemployment:

Duration of unemployment:

i.e. Incidence measures the probability of becoming unemployed and duration the length of time the individual can be expected to remain unemployed

Incidence and Duration


The unemployment rate is affected by both incidence and duration

In general, we find that those with the highest incidence and low duration of unemployment have the highest unemployment rates

e.g. high youth unemployment

low unemployment among older workers

1

Topic 7.2 Unemployment Causes


Economics 371

Why is there longer term unemployment?


We have shown that a certain amount of frictional unemployment will existWe have not offered any explanations for longer-term unemploymentLet’s work through some explanations that have been proposedIt’s important to note that no one explanation is widely accepted

2

1. Intertemporal Substitution Hypothesis


Suggests that some portion of cyclical unemployment may be voluntaryBasic idea is that real wages rise during economic expansion and fall when the economy contracts (procyclical)

Problems


3

2. Sectoral Shifts Hypothesis


Helps to explain structural unemploymentShifts in demand do not affect all sectors of the economy equallyAt any point in time some sectors of the economy are growing while others are in decline

e.g. e-business and service sectors growing while manufacturing is declining

This is likely the case

Sectoral Shifts Hypothesis Continued


Can test this empirically by looking at the relationship between aggregate unemployment and the dispersion of growth rates across industries

4

3. Efficiency Wages


With shirking (high costs of monitoring) the firm might be willing to offer higher wages to encourage workers not to shirkThe wage/employment combination the firm is willing to offer depends on the unemployment rate (shirking cost)

High Unemployment

Low Unemployment

Equilibrium


The equilibrium is given by the intersection of demand and the no-shirking supply curve

Wage

Employment

W*

E

S

D

NS

WNS

ENS

5

4. Implicit Contracts


Labour contracts often specify wages and the number of hours of work given aggregate economic conditions

Could be:“fixed employment contract” -

“fixed wage contract” -

Implicit Contracts


Helps to explain rigid wages and why firms tend to adjust using layoffs instead of wagesThe typical implicit contract is a fixed wage contract (incomes are relatively stable)Reflects risk-sharing between the employee and the employer

Thus, workers prefer arrangements where they receive a lower average wage with more certainty The firm likes such contracts because average costs are lower

6

Implicit Contracts


Thus, the firm provides “insurance”Why is insurance not provided privately?1. Moral Hazard -- Insurance may affect the workers’

behaviour such that the likelihood of unemployment increases

2. Adverse Selection -- Insurer doesn’t observe the likelihood of unemployment for a particular worker

Example


Suppose workers’ utility is a function of the wage only u(w)

There are N0 workers who sign a contract in period 1 There is uncertainty about the product market such that there are two states, each occurring with equal probability

Wage

Employment (N)N0

S

Db=MPN X Pb

Wb= kDa=MPN X Pa

An unemployed worker receives utility u(k)k=value of leisure time

7

Example


We know the outcome is “efficient”However, if the firm is risk neutral and workers are risk averse there are potentially Pareto improving contracts that involve “risk sharing” Pareto optimal contracts involve

1.2.

Example


Workers are better off

The firm may be better off

= expected wage with market clearing= ½ wa + ½ kW* is the contract wageNb* is the number of workers hired in the bad state under the contract

Wage

Employment (N)N0

S

Db=MPN X Pb

Wa

Nb

k

Da=MPN X Pa

w

8

Implications


As long as the savings are high enough in the good state the firm is better offNotice that the workers income is not fully stabilized

The optimal contract represents a tradeoff between risk-sharing and production efficiency

Implications


Even though they would prefer to be working at the going wage once laid offSuggests that employment will fall during recessions and rise during expansionsWeakness is that the model suggests that unemployment with such wage contracts will be less than if wages were allowed to adjustHowever, models that allow for asymmetric information (which is likely) give unemployment that is greater than under market clearing

9

Insider‐Outsider Model


Wage setting is determined through bargaining with “insiders” – existing work forceThe “outsiders” – unemployed – have little influence on the outcomeRequires costs to the firm in order to replace existing workers

These costs give the insiders bargaining powerThus, the insiders are able to set wages above the market clearing level

course handouts spring 2010 - web hostinghschuetz/econ371/handouts371_10.pdfl’s indifference curve...

Documents