Subjective Expectations - PowerPoint Presentation

Slide1

Subjective Expectations, Labor Market Choices and MigrationEvidence from University Students in Maputo

Mariapia Mendola

(

University

of Milano Bicocca, IZA and

LdA

)

Luigi Minale

(

Universidad

Carlos III de Madrid and

LdA

)

Ines

Raimundo

(

Universidade

Eduardo

Mondlane

)

IGC, Maputo

March, 24°, 2016

ObjectiveThis project aims at studying the role of subjective expectations about employment and earnings

in 

influencing

occupational and migration choices

of university students in Mozambique

.

By

collecting detailed survey data from students at UEM we aim at shedding

light

on the underlying causes of important socio-economic issues such as

high university drop out, youth unemployment and skilled

migration

(brain drain)

in Mozambique

.

Background -1People take forward-looking decisions

depending on

expectations on their own returns (

Kimbal

, 1990)

Subjective expectations

are heterogeneous (we are agnostic on how they

are formed)

and people react to them making up their own mind over intertemporal choices

In

contexts

of

incomplete markets and high

risk, uncertainty

on

ex-ante

economic returns (as opposed to

ex-post

ones) may play a significant role in shifting

key lifetime decisions - such as education, occupation, migration

(

e.g.

Manski

, 2004; Jensen 2010;

Attanasio

2009;

Delavande

et al. 2011).

Human capital investment decisions have been shown to have a long-lasting impact on

wage inequality, economic growth, and employment

, leading to interest in the subject from policy makers as well as researchers

Background -2Not only expectations are important for people’s sorting in specific economic outcomes (e.g. occupation), but

perceived

expected returns can easily become

self-fulfilling

(Reuben,

Wiswall

and Zafar, 2015)- For

example

Youngsters

with low-expectations will have a smaller incentive to perform well academically, or subsequently they will be willing to accept a low-paying job offer and less likely to negotiate for higher salary because it is in line with their

beliefs

Students with

high expected

probability of working in the agricultural sector may believe such occupation is low-skilled and low-paid, therefore they will drop out from school and end up having a low wage.

Policy relevance..In developing contexts, getting hard evidence on individual expected returns on key human capital investments may have important implications in terms of both education and

labour

market

policy.

Using choice data, while assuming

homogenous

earnings

’

expectations,

is problematic since observed choices might be consistent with several combinations of expectations and preferences- especially in contests of high heterogeneity and

uncertainty.

Getting

to know people’s expectations and how they act upon them may provide,

ceteris paribus,

direct evidence on the relevance and severity of imperfections (e.g. in the information or insurance market) in determining actual investment behavior.

..and questions:Are Mozambican youngsters informed about the actual labor market opportunities?Do young people have correct expectations (information) about returns to schooling (and dropping-out)?To what extent expected income uncertainty affect youngsters’ education and occupational choices?

To what extent over- or under-estimated expectations play a role in the skill-mismatch in the labor market?

And on migration choices?

MozambiqueMozambique has been growing steadily over the last decade (7%), mainly due to an extractive resource boom.

Yet, job creation has been slow and unemployment rates hardly declined over the same period, especially among youth.

On the supply side, more than 300,000 people enter the labor mkt every year and this is poised to increase at about 500,000 by 2025 (WB 2015)

Rates of return to higher education has been increasing from 6% in 2002 to 18% in 2008.

A higher education degree increases wages by 148% compared to those with secondary

education

(WB 2015)

Employment rates by gender and educational attainment (INCAF 2012)

Employment rates by occupation / income source (INCAF 2012)

MozambiqueExpanding the share of skilled workforce is a crucial

condition for

long term and sustainable

growth.

The country has already

prioritized the development of higher education with a focus on science and

technology (PARP, CPS-WB 2014, WB-HEST, Let’s Work etc.).

The

number of higher education institutions have increased from 3 to 48

btw 1992 and 2014, also thanks to strong links with private sector.

This

is intended to address a problem of skills mismatch, to respond to a fast demand of highly specialized skills and to improve productivity

both in the formal and

informal

sector

(WB 2014

).

MozambiqueYet,

so far the

demand for tertiary educated workers has

been still

growing more rapidly than supply in Mozambique.

Completion

rates at higher education institutions are still low

:

only about one out of three of those who enroll into university do actually

graduate (UEM 2012).

Moreover,

among those who graduate many decide to emigrate

abroad (45%).

This translate in companies

often

importing

skilled

labour

from abroad (e.g. from

neighboring

countries

and even overseas).

Data collectionWe plan to collected information on a random sample of (about 800) students from different years and majors (Arts&Social

Sciences, Science, Engineering, Medicine, Architecture, Law, Economics, Mathematics, Informatics)

in UEM through a unique survey

We collect information on students’ current or intended major (if junior)

For each major, we

elicit from the students their

expectations about their likelihood of choosing future careers (including migration) and how much they

expect

to earn in them

, where the period of reference may be

10

years after they graduate (to recover the present value of lifetime earnings

).

More specifically, we ask students what are their expectations with respect to possible future outcomes

in each occupation under

different scenarios (i.e. university degree, drop-out)

the focus will be

both on

expected

average

income and

on income

variance.

To measure the variance of income, the survey will elicit

the

probability density function of

students’

expected earnings. It does this using five “bins” of different income levels and asks the students to allocate 100 “points” across the bins to reflect what they believe is the probability they would receive that level of income, conditional on successfully pursuing that occupation.

We also ask students what is the probability they will end up in certain

occupations

(and in the migration status).

We shall use broad

groups

to characterize possible careers: e.g. Science/Technology, Health, Business, Government/Non-Profit, Education and Law (Arcidiacono et al.

2014)

Alternatively, we may characterize occupations as wage work,

entrepreneurship and

inactivity (Osman 2014).

Importantly, for all students in the sample, those probabilities and expected earnings are elicited for all possible occupation-major combinations, i.e. both for the chosen (or intended) majors and the counterfactual majors

.

This will allow us to examine the possible complementarities btw majors and expectations, both in terms of earnings and preferences (i.e. how attractive working in a particular occupation is with different majors). For example, expected returns for business carriers are highest for economics major, which in turn leads people to report higher

prob

of pursuing a business occupation in the (sometimes)

hypothetical

case that they were an economics major

We further collect

a

wide range of background characteristics (including financial aid/constraints) of students as well as measures of risk aversion.

Analysis -1By using subjective expectations data on occupations for all counterfactual majors, we are able to

isolate

the separate influence of monetary returns from non-pecuniary factors

quantify

the importance of sorting across occupation on ex-ante

monetary returns,

keeping all the rest constant.

Identification: using

stated choice is advantageous because we rely on expectations elicited at the time of the choice in a hypothetical scenario that we seek to explain; this is equivalent to a situation in which

respondents’

preferences are elicited

before

they make their educational or career decision. S

uch ‘timing’

mitigates the

endogeneity issue.

Analysis -2Hence, this kind of data allows us to do

3 things:

identify

the

ex-ante

impact (treatment effect) of particular occupations (relative to a reference

one)

on earnings, for any given college

major (and for both graduating and dropping-out)

identify

the

ex-ante

impact of migration on earnings, for any given college

major

identify

the

ex-ante

impact of particular majors on the probability of working in any given occupation and on migration

the

differential

effect of a treatment on a 'treatment group' versus a 'control group' in a natural experimen

Intended Q&A- 1Do youngsters have well-informed or ‘correct

’ expectations about future earnings?

We shall compare

elicited expectations (in particular

wrt

earnings) with actual data from IOF

2014-15

(most recent data containing some labor market information)

From this set of information we see where UEM students believe they rank relative the population (if

possible, by

comparing earnings by major-occupation combinations or at least by educational attainment)

Intended Q&A- 2To what extent expected economic returns (pecuniary and non-pecuniary) affect people behavior with respect to human capital investment decisions?

We do so by looking at the difference between the

ex-ante

treatment effect of occupation

k

on the treated (i.e. by weighting the difference in the reported earnings by the probability

p

i

the individual reports she will work in occupation

k

) and the

ex-ante

treatment effect on the untreated (i.e. by using the weighting

(

1

-

p

i

))

The latter

gives a measure of the importance of

selection

on the expected returns to each occupation (a positive difference is consistent with positive

sorting

on expected earnings in different occupations- the magnitude of the difference is also important though)

Intended Q&A- 3What is the role played by expected earnings in the choice of students’ occupation?What is the role played by expected earnings in the choice of students’ migration?

We can look at this by estimating the probabilities of choosing particular occupations as a function of the (log) expected earnings controlling for individual-occupation preferences (dummies) and major-occupation fixed effects.

This may be derived by as simple model of occupation choice, which provides a link between subjective expectations and preferences, making it possible to tell apart the role of expected earnings and non-pecuniary factors in this context.

Alternative follow-up interventionsStudents are re-interviewed after they graduate, allowing for the measurement of how their actions in the

labor market line up with their

reports in the first round of the survey

Students are followed over time in order to collect their actual occupational choices and realizations

(Part of the) students are provided with information about the distribution of incomes in the different occupations (either domestically or abroad). Students are asked again

to provide their expectations of outcomes in the different occupations and

the corresponding

updates

to their

occupational/migration

intentions. This allows for the measurement of how the

expectations and

intentions

are

impacted

by an information

intervention.

THANKS!