14
THE ECONOMIC RETURN ON INVESTMENTS IN HIGHER EDUCATION: Understanding The Internal Rate of Return Presentation OISE, HEQCO, MTCU Research Symposium Defining and Measuring Student Success Toronto November 22, 2013 Torben Drewes Department of Economics Trent University

THE ECONOMIC RETURN ON INVESTMENTS IN HIGHER EDUCATION: Understanding The Internal Rate of Return Presentation OISE, HEQCO, MTCU Research Symposium Defining

Embed Size (px)

Citation preview

THE ECONOMIC RETURN ON INVESTMENTS IN HIGHER

EDUCATION:Understanding The Internal Rate of Return

Presentation

OISE, HEQCO, MTCU Research Symposium

Defining and Measuring Student Success

Toronto

November 22, 2013

Torben Drewes

Department of Economics

Trent University

Background

• Much public discussion about the financial viability of investments made by young Canadians in higher education (esp. university)

• The term “returns to higher education” occurs often in that discussion– What exactly does that term mean?– How are such returns estimated?– How useful are such estimates?

Definitions• “Return to education” often used loosely.• 2011 NHS – average annual wage and salary income for:

– high school graduates: $31,400– college graduates: $50,646– university graduates: $62,340

• Average college incomes are 61% higher than high school incomes. The “return” to a college diploma is a 61% gain in income.

• Should be termed the “earnings premium”, not the return. – After all, income comparisons do not consider the cost of the investment.

• The internal rate of return on an investment or project is the "annualized effective compounded return rate" or discount rate that makes the net present value of all cash flows (both positive and negative) from a particular investment equal to zero.– I.e. the interest rate at which the net present value of costs (negative cash

flows) of the investment equals the net present value of the benefits (positive cash flows) of the investment.

Definitions• Why the confusion between earnings premium and irr?• Regressing the natural logarithm of earnings against measures of

schooling attainment produces the proportional increase in earnings for each unit change in education: the earnings premium.

• Jacob Mincer (HCEF): under certain assumptions (eg., no tuition, schooling is measured in years, etc.) the coefficient on the measure of education is the IRR.

• Those assumptions do not typically hold.

– earnings premium from a regression model are not equal to IRR.

The IRR Calculation• Consider an annual series of cash flows, Ct, that can be negative or

positive.• The IRR solves for the value of ρ so that • Your brother-in-law wants to borrow $100. Promises to repay you $110

in one year. What is the IRR to you on this investment?• Solve So ρ = 0.10, or 10%• Why is the IRR useful? The brother-in-law proposes a different

repayment schedule: $56 in one year and $58 in two years. How to compare?

• Solve. ρ = 0.09, or 0%• Let’s try something a little more complicated ….

Internal Rate of Return Calculations

Assume Tuition = $6,000 University IRR to University 16%$3,000 College College

Assume Schooling Takes 4 years University3 years College

Decision year: 18 years old

Annual Earnings Net Gains

AgeHigh SchoolGraduate

CollegeGraduate

Bachelor'sGraduate College Bachelor's

18 31400 0 0 -$34,400 -$37,40019 31400 0 0 -$34,400 -$37,40020 31400 0 0 -$34,400 -$37,40021 31400 50646 0 $19,246 -$37,40022 31400 50646 62340 $19,246 $30,94023 31400 50646 62340 $19,246 $30,94024 31400 50646 62340 $19,246 $30,94025 31400 50646 62340 $19,246 $30,94026 31400 50646 62340 $19,246 $30,94027 31400 50646 62340 $19,246 $30,94028 31400 50646 62340 $19,246 $30,94029 31400 50646 62340 $19,246 $30,94030 31400 50646 62340 $19,246 $30,94031 31400 50646 62340 $19,246 $30,94032 31400 50646 62340 $19,246 $30,94033 31400 50646 62340 $19,246 $30,94034 31400 50646 62340 $19,246 $30,94035 31400 50646 62340 $19,246 $30,940

Let’s Increase University Tuition

Internal Rate of Return Calculations

Assume Tuition = $10,000 University IRR to University 15%$3,000 College College 16%

Assume Schooling Takes 4 years University3 years College

Decision year: 18 years old

Annual Earnings Net Gains

AgeHigh SchoolGraduate

CollegeGraduate

Bachelor'sGraduate College Bachelor's

18 31400 0 0 -$34,400 -$41,40019 31400 0 0 -$34,400 -$41,40020 31400 0 0 -$34,400 -$41,40021 31400 50646 0 $19,246 -$41,40022 31400 50646 62340 $19,246 $30,94023 31400 50646 62340 $19,246 $30,94024 31400 50646 62340 $19,246 $30,94025 31400 50646 62340 $19,246 $30,94026 31400 50646 62340 $19,246 $30,94027 31400 50646 62340 $19,246 $30,94028 31400 50646 62340 $19,246 $30,94029 31400 50646 62340 $19,246 $30,94030 31400 50646 62340 $19,246 $30,94031 31400 50646 62340 $19,246 $30,94032 31400 50646 62340 $19,246 $30,94033 31400 50646 62340 $19,246 $30,94034 31400 50646 62340 $19,246 $30,94035 31400 50646 62340 $19,246 $30,940

Estimation• Use regression or other smoothing techniques applied to Census data

to produce age-earnings profiles.• Out-of-pocket (C1) and opportunity (C2) costs are compared to

earnings gains (B) to compute IRR.

18 22 26 30 34 38 42 46 50 54 58 62 T

High School

University

Ear

nin

gs

AgeC1

B

C2

The Caveats1. Parameters Matter.

– Estimates use “prototypical” decision-maker – age 18.

– Eg.: returns to adult learning will be lower

2. Only earnings/tuition data are used.

– Non-monetary costs and benefits are important, but hard to measure

– Implies that IRR calculations are only part of the story.

3. Only private returns

– Correct if we are trying to model individual decision or inform young decision-makers.

– Incorrect if we are trying to evaluate public policy to promote higher education as

social investment.

• Need to compare social costs (different than tuition) with social benefits (gross income,

reduced health expenditures, implications for productivity of co-workers, etc.)

The Caveats

4. Future stream of earnings unknown.

– Proxied by current earnings of older workers.

5. The Missing Counterfactual

– The high school profile is what the university graduate would have earned had she

entered the labour market after high school.

– University (and college) graduates may differ from high school completers in other

ways that are related to earnings. The ability bias issue.

6. Average vs Marginal Returns

– Estimated returns are averages across all individuals participating in higher

education.

– Policy question may need to estimate returns at the margin. Eg. should we expand

access to higher education?

The Caveats7. Estimates are average rates of return. There will be large differences

across and within fields of study

Distribution of real earnings (20th, 50th and 80th percentiles) by field of study for Canadian-born male bachelor’s graduates in 2005 dollars, population aged 25-64 in the labour force: Census 1985 to 2005Bell, King, Afshar, (2011), POSTSECONDARY EDUCATION AND THE LABOUR MARKET: THE EVOLUTION OF SUPPLY AND DEMAND, 1985 TO 2005, HRSDC

$0

$20,000

$40,000

$60,000

$80,000

$100,000

$120,000

1985

1990

1995

2000

2005

1985

1990

1995

2000

2005

1985

1990

1995

2000

2005

1985

1990

1995

2000

2005

1985

1990

1995

2000

2005

1985

1990

1995

2000

2005

1985

1990

1995

2000

2005

1985

1990

1995

2000

2005

1985

1990

1995

2000

2005

Education Fine Arts Humanities SocialSciences

Commerce Life Sciences Engineering Health Math,Computer,&physicalSciences

Conclusions

• Given these difficulties, should we dump the IRR?• No!

• Conceptually, IRR to higher education is a valuable theoretical construct in understanding demand for PSE

• For good policy design, statistical evidence trumps anecdotal evidence.– More work needed on empirical implementation, but we need the estimates.

• What does that statistical evidence look like?

OECD Estimates of IRR: Canada

1996 2000 2004 2009

Women 7.3% 9.5% 9.1% 12.2%

Men 7.3% 8.7% 9.4% 12.3%Sources: Boarini & Strauss (2010), OECD JournalOECD, Education at a Glance, various years