Labour Supply - Text: Benjamin, Gunderson, Lemieux and Riddell, Chapters 2-4 Labour supply decisions: (1) Work or not: labour force participation decision (Ch. 2,4) (2) How much to work: number of hours to work (Ch. 2) (3) What kind of occupation to work at? What kind of skills to acquire? (this is considered later in the course) Relevance? - Explanation of trends in labour force participation and hours of work over time. - Explanation of differences in labour force participation, hours of work between groups. - Evaluation of the labour supply effects of government policies (taxation, income support Ch.3) 1
72
Embed
Applications of Labour Demand - Lakehead Universityflash.lakeheadu.ca/~mshannon/lab19fall2.docx · Web view- Attainable Y, le combinations: anything on or below the budget line.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Labour Supply
- Text: Benjamin, Gunderson, Lemieux and Riddell, Chapters 2-4
Labour supply decisions:
(1) Work or not: labour force participation decision (Ch. 2,4)
(2) How much to work: number of hours to work (Ch. 2)
(3) What kind of occupation to work at? What kind of skills to acquire? (this is considered later in the course)
Relevance?
- Explanation of trends in labour force participation and hours of work over time.
- Explanation of differences in labour force participation, hours of work between groups.
- Evaluation of the labour supply effects of government policies (taxation, income support Ch.3)
1
Theory of Labour Supply
- Assume people act rationally to maximize their own well-being (utility).
- Rationally? actions are consistent with the objective and with each other.
- Labour supply decision: decision to allocate time between different uses.
- Uses for time? Two in the simplest model
(1) Work for pay
(2) Leisure (any use of time other than paid work)- house work- raising children- school- entertainment
- The decision to work involves a cost-benefit comparison:
- Benefit: value of income earned by working
- Cost: value of the foregone non-work activities
disutility of work.
- How to model this logic?
- Introduce concept of “utility” (shorthand for idea that people can compare and rank possible choices)
- Introduce the “Budget line” to show what choices are possible.
2
- Measure these values in terms of "utility" or well-being
- Value of choices depends upon the person's preferences.
- Captured in the form of the person's “utility function”.
- Utility is a tool for thinking about choices: not measurable.
3
Preferences: value and ranking of different choices
- Utility function:U = U( Y , le )
Y - income le - leisure (hours) U - utility
- more “Y” or more “le” raises utility (both are “goods”).
- indifference curve: shows combinations of Y and le that give exactly the same level of U.
- Graph utility function in Y, le space (Y on vertical axis, le on horizontal).
- An indifference curve for each value of ‘U’ (curves for U0, U1, U2, U3 in diagram)
- Higher indifference curve means higher utility (U3>U2>U1>U0).
4
- Shape of indifference curves:
- Negatively sloped: if Y is reduced the person must be given more le if U is to remain the same.
- Interpreting the slope: value of extra leisure in terms of income. (marginal rate of substitution: “MRS”)
- Convex shape:
- curve becomes flatter as you move down it.
- this assumes: - Value of something is high when scarce and low when abundant.
- Indifference curves cannot cross if the person is rational.
5
- Example preferences:
- Workaholic: flatter indifference curves (low value of leisure, high value of income) .
- Couch potato: steep indifference curves
- People differ in their preferences.
- Such differences may help explain differences in labour market outcomes.
6
Budget Line: What Choices are Possible?
- Budget line shows what combinations of Y and le are attainable.
- Shows constraints the person faces.
- Income : Y = wH + I
w = wage rate (per unit of time worked)
H = time worked
I = non-labour income (interest, rental income, government transfers, maybe income of other family members)
- Time Constraint: T = H + le
T = total number of hours available.
- Combining the income and time constraints gives:
(Note: Income is on the vertical axis not the wage rate !!)
- Hours of leisure: measured left-to-right.
- Hours of work: measured right-to-left.
- Attainable Y, le combinations: anything on or below the budget line.
- People differ in the wage rates (w) and in their levels of non-labour income (I). These differences may help explain differences in labour market outcomes.
Best Choice?
8
- The attainable Y,le combination that puts the person on the highest indifference curve, i.e. gives the highest utility.
- Two types of outcome:
(1) Labour force participant: best choice has hours of work >0.
- note: budget line and indifference curve are tangent.
- implies marginal value of leisure (MRS) = wage at best choice.
i.e. a the margin time has the same value in both uses.
9
(2) Non-participant: best choice H=0, le=T.
- Choice as a benefit-cost comparison? start on budget line at le=T
- benefit of working an hour: W
- costs of working an hours: value of foregone leisure (MRS).
- work if benefit>cost: if budget line steeper than indifference curve.
What determines the best choice?
10
(1) Preferences: indifference curves become steeper (blue ones replace red ones) – result? work less (more leisure).
(2) Non- labour income:- rise in I: - work less if leisure is a normal good.
- work more if leisure is an inferior good.
- empirical studies: leisure almost certainly normal.(3) Wage rate:
11
- a rise in W has two effects:
- income effect: richer so raise le is leisure normal, reduce le if leisure is inferior.
- substitution effect: work time is more valuable so substitute toward work.
Total effect = Income effect + Substitution effect (text Figure 2.9)
- Wage rise raises H if leisure is inferior
- Wage rise could raise or lower H if leisure is normal
i.e., is income or substitution effect stronger?- Individual person’s labour supply curve:
- plots the relationship between wage and H chosen.
12
- See Fig. 2.10 of text for example derivation.
- Empirical studies of effects of wages on hours worked?
- See Benjamin, Gunderson, Lemieux and Riddell Table 2.3
13
- Studies for Canada: wage effect on hours of work is small
- Close to vertical labour supply (men especially)
- Maybe slightly backward bending.
14
- Wage effects: some complications
- After-tax wages:- relevant wage for supply decisions: after-tax wage i.e. w(1-t)
t = tax rate (proportion of wage taxed away)
- effect of a tax change: like a wage change.
- progressive tax system: “t” is higher at higher levels of income
- can create a disincentive to work longer hours.
( diagram: tax rates t1<t2<t3)
- if the tax also taxes non-labour income: tax has an additional income effect.
- Overtime premium (p):- wage per hour rises at higher work hours (over Hot).
- steeper budget line for hours of work over Hot : say w(1+p)
(most provinces mandate “time and a half” for hours over some limit, i.e. p=.5)
- creates an incentive to work longer hours. (see text Figure 2.13)
15
Labour Force Participation Decision
- Participant (in labour force):
best choice: hours worked>0
- Non-participant (out-of-labour force):
best choice, hours worked =0 (le=T)
What Affects Participation? Theory
Wages:
- A rise in the wage pivots the budget line upward around the point where le=T.
- Consider a non-participant.
- continually increase the wage.
- eventually the non-participant becomes a participant.
- Higher wages mean that labour force participation is more likely.
- Reservation wage (wR): wage rate at which the person is indifferent between working and not working.
- slope of indifference curve at H=0 (le=T) with Y=I.
- the person works at any wage higher than this.
- determinants of wR: preferences and non-labour income,
Non-labour Income:
16
- A rise in non-labour income shifts the budget line upwards by the amount of the increase in I.
- The effect on participation of a rise in I is unclear.
- can draw cases where:
it induces non-participant to participate (leisure must be an inferior good)
it induces a participant to not participate (leisure must be a normal good).
Preferences (Indifference curves):
- Flatter indifference curves are more likely to imply participation,
i.e., implies a high value on extra income relative to the value of extra leisure.
- What factors lie behind preferences?
(a) Person's attitudes- How hard, tedious, etc. does the person find work?- How strong is the person's work ethic?- How materialistic is the person?- How much pleasure from recreational time?
- Some of these attitudes may be related to observable characteristics like age, family status or sex.
(b) Factors affecting the value of non-work uses of time.e.g., Presence of young children: may raise the value of
time at home.- Students: studying as a valuable alternative use
of time, raises the value of time in non-work.
17
An Extension: Labour Supply with Home Production
- Break leisure into: “pure leisure”= le, “home production”=HP.
- Let: f(HP)=amount of home production (measured in $ equivalents: so treat it like extra income).
- assume slope of f(HP) becomes flatter higher is HP (call it f’)
- Time constraint is now: T=le+H+HP H=paid work time.
- Budget line (if worth doing both home production and paid work):
Y = W (T-le-HP) + f(HP) + I
- Measure HP right to left (just like H earlier).
- Budget line has two segments:
- hours where HP has higher return than work (f’>W)
- hours where work has high return than HP (W>f’).
- Best choice as before: highest indifference curve consistent with budget line.
- Shifts in f(HP) captures changes in value of HP (a new determinant of outcomes)
18
- Logic is the same as the simple model: at the margin allocate time to its highest value use.
- Marginal value of time in pure leisure (MRS): slope of indifference curve.
- Marginal value of time in paid work: wage (w)
- Marginal value of time in home production is slope of f(HP), i.e. f’
- Outcomes are determined by preferences, wages, non-labour income and productivity of home time (f).
- Further extensions? Many types of ‘le’ (each affecting U differently), many types of ‘HP’ (each with their own home production
function).
19
Labour Force Participation: Empirical Measures
Statistics Canada: some basic labour market measures (not seasonally adjusted)
Labour Force Source Population:- non-institutional population, age 15 and over.
30,803,800 in Canada in August 2019 (+452,000 2018-19)
Labour Force: employed plus unemployed
- unemployed:- those who do not hold a job but are looking for work.
- For those age 15 and over in August 2019: labour force = 20,623,000 (+446,000 2018-19)
- See Chapter 3: Benjamin, Gunderson, Lemieux and Riddell.
- Objectives:
- Anti-poverty programs: provide some minimum level of income.
e.g. social assistance (welfare)
- Insurance programs: aimed at replacing lost income.
e.g. Workers' compensation, unemployment insurance.
- Both types of programs raise similar incentive issues.
Social Assistance (SA) or Welfare
- Tweddle and Aldridge Welfare in Canada, 2017 (Maytree website) for a description of Canadian SA programs and benefits.
https://maytree.com/welfare-in-canada/canada/
- Provinces are primarily responsible.
- Municipalities share responsibility in some provinces.
- Federal government social transfers to the provinces are partly to fund SA.
31
- Elements of a basic social assistance program:
Income guarantee:
- geared to personal circumstances (e.g., number and age of kids, disability)
- cost of living: urban vs. rural; shelter allowance.
- Non-pecuniary benefits: drugs, dental care, etc.
- May be linked to "employability".
Means test: - Value of liquid assets must be below a certain level.
Tax-back rate.
- Social assistance benefits are reduced when income from other sources is earned.
- there is often an exemption for the first few dollars of income.
- The proportion by which benefits are reduced per dollar earned is the tax back rate.
32
The Effects of Social Assistance on Labour Supply: Theory
- S.A. alters the form of the budget constraint.
- If the person doesn't work and has I=0 receives: G (guarantee)
- For each dollar of other income S.A. benefits are reduced by "t" (tax-back rate).
- “t” can be quite high: even 1.
( see graphs for Quebec from: Duclos, Fortin and Fournier (2009))
- So a person will receive S in social assistance:
S = G - t (wH + I) (if >0, otherwise S=0)
H = hours worked, I = non-labour income other than S.A., w= wage rate.
- Adding this to the budget constraint gives:
Y = wH + I + S
Y = w(1-t)H + I(1-t) + G = w(1-t) (T-le) + I(1-t) + G if social assistance payment is
positive.
- Breakeven point: income (hours of work) at which S becomes 0.
wH + I = G/t at incomes above this S=0.
(draw: budget with and without the S.A. program)
33
Effects: some illustrative cases
Case 1: Person with a low wage, earns more with S.A. than if they work.
- Illustrate: - this person does not work if S.A. is available.
- Implications for the shape of the labour supply curve? - with SA: 0 labour supply at low wage rates.
34
Case 2: Person with low wage, earns more if they work but are better off (higher utility) with S.A.
-S.A. induces this person to not work.
- The value of the extra leisure is greater than the value of the lost income.
- Implications for labour supply curve? leftward shift at low wages.(for some people)
35
Case 3: Person who does not work without S.A., still does not work with S.A.
- Illustrate. Better off, but labour supply unaffected.
Case 4: Most people’s choices are unaffected by the program: high wage and non-labour income make S.A. irrelevant.
36
- Programs are successful in the sense that they raise “utility” of the poor.
- Labour market policy concern is typically with cases 1 and 2.
- Why? - efficiency loss from reduction in work:
- surplus from this work no longer realized (see discussion of Supply-Demand model)
- loss: value of lost output (wage) minus value of leisure.
- possible long-run effects on work habits and attitudes.
- Do preferences change? Does dependency develop?
- equity / fairness issues: someone else must pay
- program cost: value of alternative uses of funds?
37
Welfare reform: Designing programs that minimize efficiency and equity losses
- See text discussion and Figure 3.3 for illustrations of a number of cases.
- Tradeoff: efficiency and program cost vs. the anti-poverty goals of the program.
- Some Possible Design Changes:
- Lower the guarantee:
- make welfare less attractive.
- may induce some to work (illustrate).
- Problem? Makes the group the program aims to aid worse off!
- Lower the tax-back rate:
- Idea: reduce the penalty to working for those on SA.
- Negative income tax (NIT) proposals: - set tax-back rates in line with the income tax system tax
rates.
- Effects in theory?
- Lowering the tax-back rate is like raising the wage.
- Positive effect on participation.
- Ambiguous effect on hours worked (for those already working).
- BUT: Low tax-back means a high breakeven point (point at which S=0).
- Large number of the working poor may collect SA.
- Will this program be too costly?
38
- Workfare (SA with a Work Requirement):
- Impose a minimum work requirement (Hm) on SA recipients.
- work? in practice may be training/education.
- Usual arguments made for workfare?
- Fairness / equity: do something for the money
- Value of output can offset benefit costs.
- may develop skills, develop workplace connections.(could this raise w in the long run?)
- Makes welfare less attractive too: some disutility fromworking.
- Budget Constraint with workfare:
If H>= Hm : Y = (wH + I)x(1-t) + G
if H< Hm : Y = wH + I
39
Possible Problems?
- Unintended effects: perverse case, leaves SA and does not work.
- Why was the person not working?
- incentives of the welfare system?
- Very low-wage, high cost of time? e.g., lone mothers
- Administrative cost- Successful workfare programs are expensive.
- counseling, childcare, training costs.
- Subsidies for the working poor:
- Make working a more attractive option vs. welfare.
- U.S.: Earned Income Tax Credit (EITC)
- very low income (under $7000) effectively raises wage:
w(1+c)
(c = rate of subsidy about 0.34 )
- maximum payment received at income of $12,500: - wage again “w”
- credit of $2500 paid: like non-labour income in this range.
- tax-back applies: income $12,500 to about $25,000- wage: w(1-t)
- tax back rate of about 0.16
40
- Effects on labor supply?
- could induce some welfare recipients to work.
- working poor?
- some ambiguity: - raises effective wage for some lowers it for others.
- non-labour income: could encourage to work fewer hours.
- Similar policies in Canada?
- National Child Benefit (1998):- intended to raise payments to working poor families
but not to families on social assistance.
(Provinces encouraged to reduce SA payments by the increase in benefits)
- “lowering the welfare wall”
- Provincial supplements to the working poor (Quebec, Alberta, BC, Saskatchewan and NB): similar.
41
Do People Respond to the LS Incentives of Income Support Programs?
Welfare Program Changes:
- Significant changes in Canada and US in the mid-1990s.
- Ontario: - benefit cuts and eligibility restrictions after 1995, workfare after 1998 (see text Fig 3.10; also diagrams from Richards; data from Finnie and Irvine)
- Alberta: - benefits cuts and program changes starting 1993.
- BC: - big changes as well.
- Other provinces: - benefit cuts in real terms in all provinces.
- US: benefit cuts, eligibility restrictions (lifetime limits), Earned Income Tax Credit (EITC) (text pp. 97-98)
- Caseloads have dropped dramatically in US, Ontario, BC and Alberta.
- Empirical problem?
- Separating out effects of reforms from effects of strong labour markets.
- What will happen in a recession? (likely to find out soon)
- Finnie and Irvine (2008) “Welfare Enigma”:
- Canada 1993-2005.
- Big decline in caseloads: both SA and falling unemployment rate were important.
- US literature: similar findings.
42
Self Sufficiency Project
- Text discussion pp. 102-105, handouts from Card and Robins (1996) and Making Work Pay (2002).
- Experimental program in BC and New Brunswick (single mothers)
- large supplement payments to those on S.A. who managed to find full-time jobs.
e.g., working 30 hrs/week at $7/hour would make $10,500 per year and receive a $13,250 supplement.
- this would leave participants with much higher incomes if working than if they relied solely on S.A..
- payments were temporary: eligible to receive it for 3 years.
- Effects:- Found sizable effects on the probability of being employed in the
1-2 years after incentive provided.
(see Card and Robins diagrams)
- experimental effect is measured relative to a control group.
- Suggests that people on social assistance can and do respond to incentives
- it is not purely a case of them being constrained because jobs do not exist.
- Longer-term effects:
- employment effect diminishes over time (4 years after effect not so large)
43
- Why fall off?
- mainly because control group employment eventually rises too.
- main benefit of the subsidy: accelerated movement into the labour market.
44
Unemployment, Underemployment and Labour Supply
- The model of labour supply assumes that:
- people can work if they decide to participate (no unemployment)
- people can work the number of hours they desire (no under- or over- employment).
- Problem?
- There are a lot of people seeking work who (it appears) cannot find it.
- Many people seem to be constrained in the hours they work (see reasons for part-time work).
Part-time employment by reason, annual averages for 2018 (000s of people)
Both sexes Men WomenPart-time employment, all reasons 3,482.70 1,193.20 2,289.50Own illness 159.5 57.8 101.7Caring for children 316.7 20.9 295.7Other personal or family responsibilities 145.6 29.8 115.8Going to school 1,028.00 433.3 594.7Personal preference 992.8 338.9 653.9Other voluntary reasons 81 30 51Business conditions, did not look for full-time work in last month 345.3 133.9 211.4Could not find full-time work, did not look for full-time work in last month 118.9 35.8 83.1Business conditions, looked for full-time work in last month 161 64 97Could not find full-time work, looked for full-time work in last month 134 48.7 85.3
Source: Statistics Canada Table: 14-10-0029-01 (formerly CANSIM 282-0014), Labour Force Survey estimates.
- Behind the assumption that work hours are a choice variable lies the model of supply and demand.
- Equilibrium: total hours supplied equals total hours demanded by employers at the equilibrium wage.
- If hours supplied exceeds hours demanded, wages will fall and bring them into equality.
45
- Constraints seem to exist:
- Limited choice of hours of work in practice, even across employers.
e.g., full-time, part-time, non-participation, spike at 40 hours.
- moonlighting / multiple job holding can ease this somewhat.
-In practice, the budget line may consist of a couple of points not a line.
- This does not change the model that much.
- Hours of work and participation will be less responsive to wage and income changes
- Changes must be large enough to induce the person to jump between points.
- Unemployment:
- Potentially a severe problem for the model:
- implies no choice for the unemployed.
- Handling unemployment in the labour supply model:
- frictional unemployment: unemployment is temporary and reflects the time it takes to match workers and employers.
- measurement: could the measured unemployed really be non-participants?
46
Two measurement stories:
(1) mistaken wage expectations:- unemployed overestimate their actual wage
- They are participants based on the expected wage but would not be based upon the actual wage.
(2) unemployed as cases of income-support induced non-participation.
e.g. Casey Mulligan: current US unemployment and unemployment insurance.
- These arguments say unemployment does not reflect serious constraints on
choice.
- Alternative view: unemployment reflects a shortfall in labour demand (relative to supply)
- Maybe wages are rigid and "too high" or there is a job shortage.e.g. Keynesian macroeconomics.