Transcript The OECD Human Capital Project: first findings

Human Capital and Sustainability:
preliminary findings from the OECD
Human Capital Project
Marco Mira d’Ercole
Household Statistics and Progress Measurement
OECD Statistics Directorate
1
Structure of presentation
1.
Background and motivation
2.
Genesis and features of OECD human capital project
3.
Preliminary results
4.
Planned developments and long-term challenges
2
1. Background and motivation
Human capital/education enter policy discussions under a variety of
headings, which shape the perspective taken on measurement:
 As determinants of economic growth (long OECD-tradition):
– Arnold, Bassanini, Scarpetta (2007): long-run effect on GDP of 1 additional year
of education about 6-9 points (within range of estimates of private returns to
schooling from micro-analyses)
– OECD (2010), High Costs of Low Educational Performance: an increase of one
standard deviation in PISA (maths) scores (100 points) would boost GDP growth
by 1.8 points (controlling for educational attainment), equivalent to increase of
OECD GDP by USD 115 trillion over lifetime of the generation born in 2010

As determinant of income and earnings inequality:
– OECD (2008), Growing Unequal? on trends and determinants of income
inequality (stressed role of market income inequality)
– Several Employment Outlook chapters on education and labour market
performance
– OECD (2007), No more Failures, proposed (PISA) measures on i) “minimum
standards of educational competences for students” (poverty) and ii) “personal
and family circumstances to achieving educational potential” (inequality)
3
1. Background and motivation (2)
Both of these perspectives largely rely on physical measures of human
capital . The perspective is different when assessing ‘sustainability’
(inter-temporal) of development path.
Some references:


2008, Joint UNECE/Eurostat/OECD WGSSD: necessary requirement for
sustainability is that the (total) capital stock (per capita) in each country is
not declining.
Similar perspective taken in 2009 by report of the SSF Commission , which
argued that we need separate indicators of “speed and gasoline left”.
Making this ‘capital approach’ operational requires:
– measures based on a common (monetary) metric for those types of capital that
can be substituted in production (man-made, financial, human capital)
– physical measures for those capital assets that are deemed to be ‘critical’ (i.e.
non substitutable) for development (e.g. natural capital)
4
2. Genesis and features of OECD Project


Joint UNECE/Eurostat/OECD WG SSD noted that “human capital values
are not directly observable.. but indirect methods exist for valuing them”,
suggesting the JF approach (life-time discounted income) as possible basis
To explore this option, OECD and Fondazione Agnelli convened joint
workshop in Turin in fall-2008. Main conclusions:
– Variants of the JF approach had already been implemented in several countries
(Norway, Australia, Canada, United States, others)
– A simplified , and easier to implement, variant of this approach could rely on the
grouped data (i.e. by age-bands) that are typically available at the OECD
– Scope for comparative exercise (based on common assumptions and databases)
led by the OECD and involving countries participating on a voluntary basis
5
2. Genesis and features of OECD Project (2)
OECD-project : launched in fall 2009
– Participating countries: 16 OECD (Australia, Canada, Denmark,
France, Italy, Israel, Japan, Korea, Mexico, Netherlands, New Zealand,
Norway, Poland, Spain, United Kingdom, United States), 2 non-members
(Russia and Romania), plus ILO and Eurostat
– Focus: i) formal education-system; ii) market-work (excl. leisure and
non-market work); iii) ‘realised’ human capital (employment
probabilities); and iv) specific type of return from education (i.e. higher
earnings accruing to the individual investing in formal education)
– Basic data sources: grouped OECD data on number of people aged
15-64 (by gender and 5-years age groups), cross classified by (ISCED
97) attainment-level. Supplemented by
i.
survey data on (gross) earnings (by gender, age-group, educ. attainment),
benchmarked on SNA values of “wages and salaries per employee”;
ii. enrolment rates (by gender and age group)
iii. survival rates (from one year to the next) for people of a given age (no
6
differences of survival by educational attainment)
2. Genesis and features of OECD Project (3)
Basic methodological assumptions:

An individual of a given age (s), gender and educational attainment will have
in year t+1 the same earnings and employment probabilities of a person
who, in year t, is one year older (s+1) but has otherwise the same
characteristics (e.g. gender and educational attainment)

Empirical implementation based on backward recursion:
 life-time income of a person aged 64 (1 year before retirement) equals current
earnings, i.e. life-time income at 65 is assumed to be 0 by definition;
 life-time income of a person aged 63 equals current earnings plus life-time
income of a person aged 64;
 etc..
7
2. Genesis and features of OECD Project (4)
Basic methodological assumptions:
 Three stages in the life cycle of each person aged between 15 to 64:
(3) retirement: for person aged 65 ad over, life-time income = zero (by
assumption, these persons have withdrawn from the labour market and do
not receive earnings)
(2) work only: for persons aged 41 to 64, life-time income is sum of:
(a) current earnings, adjusted by probability of being employed;
(b) life-time income in the next year, adjusted by corresponding survival
rate, real earnings growth and discount rate
(1) study and work: for persons aged 15 to 40, life-time income is sum of:
(a) current earnings, adjusted by probability of being employed;
(b) life-time earnings under two assumptions:
 they enter the labour market with their current attainment level; or
 they remain in school, reach higher attainment level, and then enter the
labour market with higher earnings and employment probabilities.
8
3. Preliminary results: levels
Benchmark estimates of human capital per capita in 2006:
common assumptions on real earnings growth (1.32%) and
discount rates (4.58%)
Human capital per capita (US$ in thousands)
700
GDP per capita (US$ in hundreds)
700
600
600
500
500
400
400
300
300
200
200
100
100
0
0
9
3. Preliminary results: levels (2)
Human capital (forward-looking) and physical capital
(backward-looking, PIM) as a share of GDP, 2006
Human Capital/GDP
Physical Capital /GDP
14
12
10
8
6
4
2
0
10
3. Preliminary results: levels (3)
Sensitivity analysis
800
800
700
700
600
600
500
500
400
400
300
300
200
200
100
100
0
0
HC per captial (r =1.32%, g = 4.58%)
HC per capital (r = country average 1997-2017, g = 4.58%)
11
3. Preliminary results: distribution
Measure: share of human capital of each group divided by share in
population (values greater than 1 imply “richer”)
 By gender
Distribution of human capital by gender
1.6
1.4
1.2
Female
Male
1.0
0.8
0.6
0.4
0.2
0.0
12
3. Preliminary results: distribution (2)
 By educational attainment
Distribution of human capital by educational attainment
2.5
2.0
Lower secondary or less
Above lower secondary and below tertiary education
Tertiary
1.5
1.0
0.5
0.0
13
3. Preliminary results: distribution (3)
 By age of individuals
Distribution of human capital by age of individuals
1.8
1.6
55 to 64
35 to 54
15 to 34
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
14
3. Preliminary results: volume changes
Divisia Index, weighted average of growth rates of each demographic group
based on three characteristics (age, gender, education levels): allows
measuring both total volume growth and contribution of various factors
Total volume growth: Growth rates of human capital, population and human capital
per capita, United States
115
110
105
100
VOL
95
POP
HCPERCAPITA
90
15
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
3. Preliminary results: volume changes (2)
Sensitivity analysis (discount rate set at 4.58%)
Baseline: annual real income growth rate = 1.3%
Scenario 2: annual income growth , around 1% (average over 1997-2017, OECD/MTB)
Scenario 3: annual income growth rate, around 0% (low-end over 1997-2017, OECD/MTB)
Scenario 4: annual income growth rate, around 4% (high-end 1997-2017, OECD/MTB)
115
110
105
100
95
90
VOL_1
VOL_2
VOL_3
VOL_4
HCPERCAPITA_1
HCPERCAPITA_2
HCPERCAPITA_3
HCPERCAPITA_4
Population
85
16
3. Preliminary results: volume changes (3)
Decomposition of total HC growth. The first order partial index by gender
captures changes in population structure between men and women (i.e.
doesn’t reflect population shifts among age groups or educational categories
within each gender)
Gender
Age
Educational attainment
1.4
1.4
1.4
1.2
1.2
1.2
1.0
1.0
1.0
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
1997/2002 2002/2007 1997/2007
Men
Women
0.8
0.6
0.4
0.2
0.0
1997/2002 2002/2007 1997/2007
15-34
35-54
55-64
-0.2
1997/2002 2002/2007 1997/2007
0/1+2
3
5+6
17
3. Preliminary results: volume changes (4)
Decomposition of HC per capita growth. The contribution to growth of HC per capita
by gender is the difference between the first order index by gender and the growth rate
of population; the sum of the contributions of all partial indices (by gender, age, and
educational) is a (first order) approximation to the growth rate of HC per capita
0.50
0.50
0.40
0.40
0.30
0.30
0.20
0.20
0.10
0.10
0.00
0.00
-0.10
-0.10
-0.20
-0.20
-0.30
-0.30
-0.40
-0.40
-0.50
-0.50
2002/2007
1997/2002
Gender
Age
Education
1997/2007
HC per capita
18
4. Developments and long-term challenges

Where we stand? WP to be released by year-end, containing
some of the results shown above (and more)

Second phase of the project in 2011-12 (as part of OECD
follow-up to the recommendations of SSF Commission).
Foreseen developments:
– Improving estimates and extending country-coverage
– Constructing accumulation accounts that will explain changes in human capital in
terms of investment, depreciation and revaluations
– Using HC estimates to construct an ‘educational account’ integrating data on the
various inputs entering its production as well as outputs produced
– Analysing how HC measures might be used to improve analysis of different
economic aggregates and accounting approaches: i) role of HC in measuring
MFP-growth; ii) measurement of the output of the educational sector; iii)
construction of extended household production accounts
19
4. Developments and long-term challenges (2)
Longer-term challenges:

How to incorporate quality? Qualitative measures of the cognitive skills of
the adult population will become more prominent in the future (PIAAC results
available in 2012). They provide direct measure of an important set of skills,
informing about both ‘average’ performance and inequality, and allowing to
assess how competencies change for a given attainment level. Integrating
these test scores into monetary measures of HC will be a challenge.

Can we extend monetary measures of HC to non-economic returns?
Monetary measures of human capital are based on view that main benefit
from investing in education is in the form of higher earnings. This also
applies to measures of human capital extended to non-market time (valued
at opportunity costs. What these measures exclude are the non-economic
benefits from education accruing to the individual and to society at large.
20
4. Developments and long-term challenges (3)
Two examples of importance of non-economic returns of HC: OECD
(2010), Improving Health and Social Cohesion through Education

Marginal effects of education on self-reported health
21
4. Developments and long-term challenges (4)

Marginal effects of education on political interest
22
Conclusion

Deriving monetary estimates of human capital (education)
based on a consistent methodology and assumptions is
feasible: they highlight the quantitative importance of human
capital, some of the factors that contribute to its accumulation,
the importance of considering expenditures on them as
investment rather than consumption

Monetary estimates are unlikely to fully displace physical ones;
some of the critical functions assured by education (e.g. better
parents, better citizens, greater tolerance to diversity) are not
captured by earnings-premia > physical measures of education
(quantity and quality) will remain important in the future
23