160 E. Cohn, Y.C. Ng Economics of Education Review 19 2000 159–168
Kong. Data obtained from the 1986 Hong Kong By-cen- sus and the 1991 Hong Kong Census allow us to study
the labor market behavior of overschooled and under- schooled workers in Hong Kong.
There are two main purposes for the present study. First, the incidence of overschooling and underschooling
in the Hong Kong market for 1986 and 1991 will be examined. Second, the study will address the effects of
overschooling and underschooling on earnings and how these effects change over time. Finally, we will also
examine the effect that labor-market experience might have on the economic returns to overschooling and
underschooling.
2. Methodology
Two principal methods have been employed in the literature for classifying individuals as overschooled or
underschooled: the “subjective” and the “objective” methods. In the subjective method, a worker’s self-evalu-
ation of the required schooling or skill for adequate job performance is observed. An individual’s actual school-
ing level is compared with the self-evaluated schooling requirement to determine whether she is overschooled
or underschooled.
The objective method is employed when data on required schooling for a job are unavailable. Such a
methodology is found in Verdugo and Verdugo 1989, and several other studies. The criterion for the classi-
fication of overschooling and underschooling is based on the years of schooling completed SCHOOL by individ-
uals within occupations disaggregated at 2-digit or 3- digit levels. In the Verdugo and Verdugo scheme, when
an individual’s SCHOOL falls within plus or minus one standard deviation of the mean value for the occupation,
she is considered to be adequately schooled. Those whose SCHOOL is higher than one standard deviation
above the mean for the specific occupation are said to be overschooled. Conversely, underschooled individuals
are those whose SCHOOL is less than one standard devi- ation below the mean for the specific occupation.
An alternative method to measure overschooling and underschooling has been proposed by Kiker et al. 1997.
Instead of defining adequate schooling ADSCH by plus or minus one standard deviation around the mean of
SCHOOL for a given occupation, Kiker et al. suggest that ADSCH is equal to the modal value of SCHOOL
for each occupation. That is, an individual is said to be adequately schooled if hisher SCHOOL equals the
modal value for one’s occupation. Those with SCHOOL greater than the modal level of schooling for their spe-
cific occupation are classified as overschooled. Similarly, individuals whose SCHOOL is lower than the modal
value of the SCHOOL for their specific occupation are said to be underschooled. Since the Hong Kong data only
permit employment of the objective method, we utilize the objective method modal value for all of the calcu-
lations shown in this study.
1
To investigate the effects of overschooling and under- schooling on earnings, a standard Mincer-type wage equ-
ation is used. Following Sicherman 1991, two basic specifications are used in the present study.
lnINC
it
5 X
it
b 1 a
1
ADSCH
it
1 a
2
OVERSCH
it
1 1
a
3
UNDERSCH
it
1 m
it
and lnINC
it
5 X
it
d 1 g
1
SCHOOL
it
1 g
2
OSCH
it
2 1
g
3
USCH
it
1 e
it
where lnINC
it
is the natural logarithm of labor income of individual i at time t and X is a row vector of inde-
pendent variables including experience, experience squared, marital status, and industry dummy variables.
2
ADSCH is the number of years of adequate schooling for a given job defined according to the objective method
discussed above. SCHOOL is the actual number of years of schooling of the individual. OVERSCH is the number
of years of overschooling, that is, OVERSCH 5 SCHOOL 2 ADSCH, for SCHOOL ADSCH. Simi-
larly, UNDERSCH 5 ADSCH 2 SCHOOL, for SCHOOL , ADSCH, indicating the number of years
of underschooling.
In the second specification, dummy variables are used for overschooling and underschooling, and ADSCH is
replaced by SCHOOL. OSCH indicates that an individ- ual is overschooled, and has a value of 1 if OVERSCH
0, and
otherwise. Likewise,
USCH being
underschooled has a value of 1 if UNDERSCH 0, and 0 otherwise.
We also wish to examine whether the wage effects of overschooling and underschooling interact with experi-
ence. We therefore add interaction terms for experience EXP and the schooling variables. Thus we have,
lnINC
it
5 X
it
b 1 a
1
ADSCH
it
1 a
2
OVERSCH
it
1 a
3
UNDERSCH
it
1 a
4
ADSCH· EXP
it
3 1
a
5
OVERSCH· EXP
it
1 a
6
UNDERSCH· EXP
it
1 m
it
. and
lnINC
it
5 X
it
d 1 g
1
SCHOOL
it
1 g
2
OSCH
it
1 g
3
USCH
it
1 g
4
SCHOOL· EXP
it
4 1
g
5
OSCH· EXP
it
1 g
6
USCH· EXP
it
1 e
it
.
1
The definitions of adequate schooling, overeschooling and underschooling are based on 2-digit level occupational codes.
2
For brevity, we omit the “it” subscripts in the discussion.
161 E. Cohn, Y.C. Ng Economics of Education Review 19 2000 159–168
To test whether there are significant differences between the effects of overschooling and underschooling
on earnings between 1986 and 1991, the two samples of observations are pooled. All independent variables are
interacted with the year dummy variable, T91 which has a value of 1 for observations of the 1991 sample, and 0
otherwise, and both the original variables and the inter- action
terms are
included in
the new
earnings regressions. The dependent variable lnINC for both
years is expressed in constant 1990 Hong Kong dollars. Also, a Chow test is performed to test whether the struc-
tural equations for earnings differ between the two years.
3. Data