human capital - rongsheng tang 唐荣胜rongshengtang.weebly.com/.../ch6_human_capital.pdf ·...

Human Capital

Rongsheng Tang

Washington U. in St. Louis

July, 2016

Rongsheng Tang (Washington U. in St. Louis) Human Capital July, 2016 1 / 57

Overview

Stylized facts of educationThe schooling modelSchooling as a signalPostschool human capital investmentsOn-the-job training


Education in the Labor Market: Some StylizedFactsThere are only slight differences in educational attainment between menand women, but that there are substantial differences among racial andethnic groups.

Figure:Rongsheng Tang (Washington U. in St. Louis) Human Capital July, 2016 3 / 57

Education in the Labor Market: Some StylizedFacts

Figure:


Education in the Labor Market: Some StylizedFacts

Education is strongly correlated with labor force participation rates,unemployment rates, and earnings.Education has substantial beneficial impact on the labor marketexperiences of women and minorities.


The Schooling Model

Education choice

Why don’t all workers get doctorates or professional degrees?In other words, what factors motivate some workers to getprofessional degrees while other workers drop out before they finishhigh school?


The Schooling Model

Assume that workers acquire the education level that maximizes thepresent value of lifetime earnings.

In general, the present value of a payment of y dollars next is given byPV = y

1+r

the present value of y dollars received t years from now equalsPV = y

(1+r)t


The Schooling ModelConsider the situation faced by an 18-year-old man who has just receivedhis high school diploma and who is contemplating whether to enter thelabor market or attend college and delay labor market entry by anadditional four years.

Figure:


The Schooling Model

Present value of age-earnings profiles

PVHS =t=46∑t=0

wHS

(1 + r)t

PVCOL = −t=3∑t=0

H(1 + r)t +

t=46∑t=4

wCOL

(1 + r)t

worker attends college if the present value of lifetime earnings when hegets a college education exceeds the present value of lifetime earningswhen he gets only a high school diploma, or

PVCOL > PVHS


The Schooling Model

e.g. there is a two period model,

PV0 = 20, 000 +20, 0001 + r

PV1 = −5, 000 +47, 5001 + r

r = 15%

then PV0 = 37, 391, PV1 = 36, 304 and the worker would not go to school.

How about r = 1% ?


The Schooling Model

When deciding to go to college, there are two cost associated withinvestment and waiting,

r is interest rate and also assume to be person’s rate of discount.Larger r means investment opportunity cost is higher, and also peopleare less patient.


The Schooling Model


The Schooling Model

The wage-schooling locus has three important properties:

The wage-schooling locus is upward slopingThe slope of the wage-schooling locus tells us by how much a worker’searnings would increase if he were to obtain one more year ofschoolingThe wage-schooling locus is concave


The Schooling Model

The Marginal rate of return to schooling:

Percentage change in earnings resulting from one more year of schoolThe marginal rate of return to schooling gives the percentage increasein earnings per dollar spent in educational investmentsThe marginal rate of return to schooling must decline as a person getsmore schooling.


The Schooling ModelThe stopping rule, or when should I quit school?


Education and Earnings

The previous model isolates two key factors that lead different workers toobtain different levels of schooling and, hence, to have different earnings:

workers have different rates of discountworkers face different marginal rate of return schedules.

I Differences in ability



Differences in ability may also affect schooling and earning

all workers have the same rate of discount, but each worker faces adifferent wage-schooling locus.It is often assumed that higher ability levels shift the marginal rate ofreturn schedule to the right, so that the earnings gain resulting froman additional year of schooling outweighs the increase in forgoneearnings.In other words, more able persons get relatively more from an extrayear of schooling.



In the figure, wage gap between Ace and Bob arises for two reasons.

Bob has more schooling than Ace and is getting the returns toadditional schooling.Bob also earns more than Ace because Bob is more able.


Policy Application: School Construction inIndonesia

The wage gap between high-educated and low-educated workers indeveloping countries is even higher than the gap in industrializedeconomies.It is tempting to infer from these finding that developing labormarkets offer a high rate of return to schooling, and that these highrates of return justify sizable investments in the educationinfrastructure



In Indonesia, children typically go to school between the ages of 7 and12.In 1973, the Indonesian government launched a major schoolconstruction program(INPRES) designed to increase the enrollment ofchildren in disadvantaged areas.By 1978-79, more than 61,000 new primary school had been built,approximately two schools per 1,000 children.This construction program cost almost $700 million(2002 US dollars),representing 1.5 percent of the Indonesian GDP as of 1973.As a way of grasping the scale of the construction, a similarcommitment by the US would require an expenditure of around $150billion.



The result were immediate: enrollment rates among children aged 7to 12 rose from 69 percent in 1973 to 83 percent by 1978.A recent study uses data drawn from the Indonesian labor market in1995 (two decades after the school construction) to determine if thehuge investment increased the educational attainment , earnings ofthe targeted Indonesians, and to calculate the rate of return toschooling to Indonesia.


Policy Application: School Quality and Earnings

Conventional wisdom has it that today’s high school graduates are notas good as yesterday’s graduates.The media often report that a large fraction of high school graduatesare “functionally” illiterate despite the fact that expenditures onprimary and secondary education rose dramatically in the past twodecades

I per-student real expenditures in public schools increased from $4,600 in1980 to $9,500 in 2008.

Does “throwing money” at the public school system raise the rate ofreturn to schooling?Put differently, does school quality, as measured by teacher salaries orpupil/teacher ratios, matter?


Policy Application: School Quality and Earnings

Some answers

Prior to 1992, the consensus was that high levels of schoolexpenditures had little impact on educational or labor marketoutcomes.The 1992 publication of an influential study by David Card and AlanKrueger, showing that school quality is indeed positively correlatedwith the rate of return to schooling, sparked a heated debate over theeconomic importance of school quality.


Self-selection

We want to determine if the schooling model provides a useful “story”of how students actually go about the business of deciding whether tostay in school.Ideally we could observe the age-earning profile of a particular workerboth if he were to go to college and if he were to stop after high school.

I It is impossible!

Can we just observe the earning of a group of college graduates andthat of a group of high school graduate?


Self-selection

A two-period model,Blue-Collar job requires no schooling,White-Collar job requires one period schooling.Suppose that both Willie and Wendy have a discount rate of 10percent.

Figure:


Self-selection

the present values of Willie are 38,182 and 36,364 if he does not goand go to school respectively,Willie will decide that he should not go to school and will be ablue-collar worker.the values for Wendy are 28,636 and 37,273 respectively.Wendy will go to school in the first period and works in a white-collarjob in the second.


Self-selection

The data we observe in the labor market is that earnings of personswho do not go to school and work in blue-collar jobs (like Willie) is38,182,while the earning of persons who do go to school and work inwhite-collar jobs(like Wendy) is 37,273.A comparison of the two numbers that can be observed would suggestthat Wendy made a terrible mistake.Selection bias: workers self-select themselves into jobs for which theyare best suited.


Schooling as a Signal

The schooling model is based on the idea that education increases aworker’s productivity and that this increase in productivity raiseswages.An alternative argument is that education need not increase theworker’s productivity at all, but that “sheepskin” level of educationattainment signal a worker’s qualifications to potential employers.



Suppose there are two types of workers in the labor market, lowproductivity workers and high productivity workers,

Figure:



Symmetric information: employer can determine easily distinguishworkers

I high-productivity worker would pay the worker 300,000 over the lifecycle,

I for low-productivity ones, he would pay the worker only 200,000.

Asymmetric information: one of the parties in the transaction knowsmore about the terms of the contract.

I employers can not observe employee’s quality, but just know theproportion

I Employers simply pools all job applicants and treats them identically.I The average productivity and salary of the workers hired by the firm is

average salary = (200, 000×q)+[300, 000×(1−q)] = 300, 000−100, 000q



low productivity workers prefer a pooled equilibrium because they arebeing pooled with more productive workers, who push up their salary.Neither employer nor high productivity workers like the pooledequilibrium.High-productivity workers have an incentive to provide and firmshave an incentive to take into account credible information that canbe used allocate the worker into either productivity group.This type of information is called a signal.



Suppose a firm chooses the following rule of thumb for allocating workersto the two types of jobs.

If a worker has at least y years of college, the firm assumes that theworkers is a high productivity worker, allocate him to a job thatrequires a high level of skills, and pays him a salary of 300,00.If a worker has fewer than y years of college, the firm assumes thatthe workers is a low productivity worker, allocate him to an unskilledjob, and pays him a salary of 200,00.

Obtaining the college credits is more expensive for less-able workers, inparticular,

a year’s worth of college credit costs 20,000 for a high-productivityworker,but 25,001 for a low productivity worker.



Figure:



A worker will decide either not to go to college at all or to go to collegefor y yeas.The low productivity worker will not attend college if200, 000 > 300, 000 − (25, 001 × y), that is y > 3.999The high productivity workers get y years of college whenever200, 000 < 300, 000 − (20, 000 × y), that is y < 5.



The human capital model suggests that human capital investmentssuch as education provide a way out of low incomes and poverty.The signaling model says that education does not really increase aworker’s innate productivity.

I from a worker’s point of view, education still has a positive private rate ofreturn, since the high productive worker gains from signaling that he ishighly productive.

I From a social point of view, it looks like education expenditures arewasteful, national income is not increased because of the worker’sproductivity is the same both before and after the investment ineducation.

I In fact, however, education serves the very useful role of sorting workersinto the right jobs.


On-the-Job Training

Most workers augment their human capital after completing theireducation, particularly through on the job training(OJT) programs, e.g:

secretaries learn word processing skill,lawyers get courtroom experience,investment bankers concoct new financial instruments


On-the-Job Training

Two training: general vs specific

General training is the type of training that, once acquired, enhancesproductivity equally in all firms.

I e.g. how to drive, typing, etc.

Specific training is the type of training that enhances productivityonly in the firm where it is acquired and the productivity gains areslot once the worker leaves the firm.

I e.g. how to drive a tank in the army


On-the-Job TrainingConsider a simple framework where the employment relationshipbetween a competitive firm and the worker lasts two periods.Suppose that in the first period, the total labor costs equal TC1and inthe second period, the costs equal TC2.Similarly, the value of marginal product in each of the two periods areVMP1and VMP2 respectively.Let r be the rate of discount.The profit-maximizing condition giving the optimal level ofemployment for the firm over the two periods is

TC1 +TC2

1 + r= VMP1 +

VMP2

1 + r

Suppose OJT takes place only in the first period, it costs firm Hdollars to put worker through the training, then

w1 + H +w2

1 + r= VMP1 +

VMP2

1 + r


On-the-Job Training

Who pays for general training?

In the post-training period, the worker’s value of marginal productincrease to VMP2in all firms,many firms are willing to pay the worker a wage equal to VMP2,the firm that provided the training must either increase the wage toVMP2or lose the worker,therefore w2 = VMP2, hence w1 = VMP1 − H,


On-the-Job Training

therefore, workers pay for general training by accepting a lower “traineewage” during the training period.

There are many examples of workers paying for general trainingthrough lower wages.It is common for trainees in formal apprenticeship programs toreceive low wages during the training period and to receive higherwages after the training is completed.

Competitive firms provide general training only if they do not pay any ofthe costs


On-the-Job Training

Who pays for specific training?

The productivity gains resulting from specific training vanish once theworker leaves the firm.As a result, the worker’s alternative wage is independent of thetraining and equals his pretraining productivity.


On-the-Job Training

Suppose firm pay for specific training:

setting w2 = w1 = VMP1

firm’s cost= Hsuppose the worker stay in the second periodfirm benefit= VMP2 − w2=VMP2 − VMP1

risk: if the worker were to quit in the second period, the firm would suffera capital loss.


On-the-Job Training

Suppose worker pays for specific training:

setting w1 = VMP1, w2 = VMP2

worker’s cost = Hsuppose firm continue to hire workerbenefit= w2 − VMP1= VMP2 − VMP1

risk: if the worker were to get laid off, he would lose his investmentbecause specific training is not portable


On-the-Job Training

Both the firm and the worker are reluctant to invest in specifictraining because there does no exist a legally binding contract thatties together workers firms “until death do them part”.Need to find some way to share the benefit as well as the cost.


Age-Earnings Profile



Three important properties of age-earnings profiles:

1 Highly educated workers earn more than less-educated workers2 Earnings rise over time, but at a decreasing rate3 The age-earnings profiles of different education groups diverge over

time



At every age, we will want to invest in human capital up to the pointwhere the marginal revenue of the investment equals the marginalcost of the investment.To “measure” the human capital, we introduce efficiency unitsefficiency units: standardized units of human capital, the total humancapital stock of the worker equals the total number of efficiency unitsembodied in him or her.



assumptions:

an efficiency unit of human capital can be rented out in the labormarket, and the rental rate per efficiency unit is R dollars.the market for efficiency units is competitive, so the per-unit rentalprice is R dollars regardless of how many efficiency units a worker hasall training is generalthere is no depreciation of the human capital stock over time



suppose that the worker enter the labor market at age 20 and plans toretire at age 65, the marginal revenue of acquiring one efficiency unit ofhuman capital at age 20 is

MR20 =t=45∑t=0

R(1 + r)t

similarly

MR30 =t=35∑t=0

R(1 + r)t

The marginal cost curve has the usual shape: marginal cost rise as moreefficiency units are acquired



What is the age earning profile pattern theoretically?

worker’s age earnings profile is upward slopingI Since workers pay for on-the-job training through reduced wages,I older worker earn more than younger workers because older workers

acquire fewer efficiency units of human capital and have lower forgoneearnings.

I The old worker still collect “rent” from earlier investment.

worker’s age earnings profile is concaveI year to year wage growth depends partly on how many additional

efficiency units the worker acquires,I because fewer units are acquired as the worker ages, the rate of wage

growth declines over time.


Policy Application: Evaluating GovernmentTraining Programs

Since the declaration of the War on Poverty in the mid-1960s, a largenumber of government programs have indeed attempted to providetraining to disadvantaged workers.

I the Manpower Development and Training Act of 1962 (MDTA),I the Comprehensive Employment and Training Act of 1962 (MDTA),I the Comprehensive Employment and Training Act of 1973 (CETA),I and the Job Training Partnership Act of 1982 (JTPA).

Each of these programs spent a lot of money trying to “expose”minority and other low-income groups to formal training programs.Federal expenditures on job training programs now exceed over $4billion per year.



Do these programs increase the human capital and earnings of thetrainees ?

A number of studies have made kind of before-and-after comparisonand have found that there are some earnings gains associated withthe training programs,typically, trainees earn about $300 to $1,500 more per year after theprogram than before the programThis calculation may not be very meaningful because of self-selectionproblem.



The National Supported Work Demonstration(NSW) estimate the impactof the training program is about 1,400 versus the program cost 12,500.

Figure:


human capital - rongsheng tang 唐荣胜rongshengtang.weebly.com/.../ch6_human_capital.pdf ·...

Documents