Content uploaded by Stefan C. Wolter
Author content
All content in this area was uploaded by Stefan C. Wolter
Content may be subject to copyright.
IZA DP No. 733
Wage Elasticity of the Teacher
Supply in Switzerland
Stefan C. Wolter
Stefan Denzler
March 2003
DISCUSSION PAPER SERIES
Forschungsinstitut
zur Zukunft der Arbeit
Institute for the Study
of Labor
Wage Elasticity of the Teacher Supply
in Switzerland
Stefan C. Wolter
Swiss Coordination Centre for Research in Education,
University of Berne and IZA Bonn
Stefan Denzler
Swiss Coordination Centre for Research in Education
Discussion Paper No. 733
March 2003
IZA
P.O. Box 7240
D-53072 Bonn
Germany
Tel.: +49-228-3894-0
Fax: +49-228-3894-210
Email: iza@iza.org
This Discussion Paper is issued within the framework of IZA’s research area Internationalization of
Labor Markets. Any opinions expressed here are those of the author(s) and not those of the institute.
Research disseminated by IZA may include views on policy, but the institute itself takes no institutional
policy positions.
The Institute for the Study of Labor (IZA) in Bonn is a local and virtual international research center
and a place of communication between science, politics and business. IZA is an independent,
nonprofit limited liability company (Gesellschaft mit beschränkter Haftung) supported by the Deutsche
Post AG. The center is associated with the University of Bonn and offers a stimulating research
environment through its research networks, research support, and visitors and doctoral programs. IZA
engages in (i) original and internationally competitive research in all fields of labor economics, (ii)
development of policy concepts, and (iii) dissemination of research results and concepts to the
interested public. The current research program deals with (1) mobility and flexibility of labor, (2)
internationalization of labor markets, (3) welfare state and labor market, (4) labor markets in transition
countries, (5) the future of labor, (6) evaluation of labor market policies and projects and (7) general
labor economics.
IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion.
Citation of such a paper should account for its provisional character. A revised version may be
available on the IZA website (
www.iza.org
) or directly from the author.
IZA Discussion Paper No. 733
March 2003
ABSTRACT
Wage Elasticity of the Teacher Supply in Switzerland
In order to learn more about the wage elasticity of the teacher supply in Switzerland, this
paper estimates wages for teachers and non-teachers. The data used are ten surveys of
graduates of all Swiss universities for the period of 1981-1999. The data allows us to
estimate the wage elasticity for entry wages of upper secondary school teachers. In the
cases examined, the wage differential is highly significant and shows a wage elasticity for the
teacher supply of 0.12% to 0.18%. This elasticity is substantially lower than what has been
found in the literature for other countries. It is probably due to the fact that teachers’ salaries
are generally high in Switzerland - teachers here enjoy a positive wage differential over those
in other occupations, the opposite of what is normally the situation elsewhere.
JEL Classification: I2, J24, J45
Keywords: wage elasticity, labour supply, teachers, university graduates, Switzerland
Corresponding author:
Stefan C. Wolter
Swiss Co-ordination Centre for Research in Education (SKBF)
Entfelderstrasse 61
CH-5000 Aarau
Switzerland
Tel.: + 41 62 835 23 90
Fax: + 41 62 835 23 99
Email: stefanwolter@yahoo.de
2
Introduction
1
Reports of teacher shortages in most of the OECD countries have accumulated in recent years
(see OECD, 2002 or Santiago 2002). This problem raises questions that we all too often are
unable to answer without greater knowledge of the labour market for teachers, knowledge that
is generally not at hand outside the English-speaking countries.
One easily available avenue for eliminating the shortage of teachers, or for better balancing
supply and demand, is via wage formation. Although a number of studies have been
conducted on wages and teacher quality, and teachers’ unions never tire of emphasizing that
better salaries would result in better teachers, less has been said about whether higher wages
would also motivate more people to choose teaching as a profession. The relationship
between teachers’ salaries and their quantitative supply has never been researched in most
countries, primarily because the statistics needed to do so are often lacking.
The labour market for teachers in those countries for which we do have such studies
(primarily the United States and Great Britain), is not especially comparable with the market
in Switzerland, and it is thus questionable how far the results of these studies hold for other
countries. The main difference is that in English-speaking countries, teachers normally tend to
earn less than people who have a similar education but are employed in other occupations.
Since this is not the case in Switzerland, it is interesting to examine whether the purely
quantitative high wage elasticity of the teacher supply found by these studies also applies
here.
1
The authors would like to thank the Federal Statistical Office of Switzerland for the provision of the data and
the support that came with it.
3
The goal of this study is to empirically test the wage elasticity of the supply of upper
secondary school teachers in Switzerland. To do this, we use data collected between 1981 and
1999 on graduates of all of Switzerland’s traditional academic university institutions. The
paper is divided into six sections. We begin with a theoretical analysis of supply and demand
in the labour market for teachers, and derive from this the importance of the wage elasticity of
the teacher supply. We then briefly describe the state of research as set forth in the
international empirical literature. This is followed by a description of the data, presentation of
the model employed, and a discussion of our findings. The last section summarizes the results.
The teacher supply and wages
An educational system’s demand for teachers (D) is determined by two factors primarily (see
Figure 1). The first is the assumption that the number of teachers sought will depend in the
short term on the number of school classes, so that the number of pupils at a given moment,
combined with the standard class size set by the educational authorities, gives us a more or
less predetermined level of demand for teachers (Q
g
). The second factor is the amount of
money (E) that the political process provides for the overall educational system in the budget.
This budget allocation enables varying numbers of teachers to be hired if their salaries are not
fixed in advance. However, the educational system’s quantitative demand for teachers (D), in
combination with its budget allocation (E), implies a wage (W
S
).
A normal teacher supply curve (S) does not, of course, automatically mean that the supply and
demand curves will intersect at equilibrium wage (W
e
) that would result solely from the two
curves because the government’s budget restrictions would not permit this. In the case shown
in Figure 1, budget restrictions would produce a wage of W
*
, which would lie below the
4
equilibrium wage W
e
. Under the given circumstances, the equilibrium wage would be located
at the intersection of S and E, i.e. at W
*
. However, substantially fewer teachers (Q*) would be
employed at this salary than originally planned (Q
g
). The difference would be a shortage of
teachers, TS
g
. Such a shortage could, of course, be eliminated by raising the wage, but this is
only possible if the educational system’s budget is increased from E to E’ (see also Figure 2;
which would yield equilibrium EQ
1
).
Figure 1: Teacher supply and demand (1)
Since this is probably a rare occurrence even in times of teacher shortages, educational
policymakers have to resort to other means to alleviate the problem. Another possibility is to
look for more teachers while keeping the wage at W
e3
. However, this will only succeed, and
then only partially, if there is a supply (S’) of potential teachers who are prepared to work for
a lower wage (equilibrium EQ
2
). These people cannot be the same as those described by the
D
S
E
Q
*
Q
g
Q
W
W
e
W
*
TS
g
W
s
5
usual supply curve (S). It also means that the new teachers normally will not be able to
provide the same quality of instruction as the original group of teachers.
Finally, educational policymakers can, of course, use still other means to lower the demand
for teachers (new D’), by increasing class size, for example. The smaller demand that results
from this organizational measure enables equilibrium EQ
3
to be achieved with the same
budget (E).
Figure 2: Teacher supply and demand (2)
Thus policymakers have various means at their disposal to eliminate an initial disequilibrium
(TS
g
in Figure 1), but they will only be able to do so without damaging the quality of the
educational system if they increase the budget and thus teachers’ wages. The two other
solutions described above will cause the quality of instruction to deteriorate.
If one decides to raise wages in order to eliminate a shortage of teachers, the next question
that immediately arises is how elastically will the supply of teachers respond to changes in the
D
S
E
Q
*
Q
g
Q
W
W
e1
W
e2
E'
D'
S'
EQ
1
EQ
2
EQ
3
3
2
1
W
e3
6
wage? Without this knowledge, policymakers are unable to judge how much the budget has to
be increased in order to eliminate the shortage.
A brief review of the literature
A number of authors, primarily in English-speaking countries, have demonstrated the
existence of wage effects on what people decide to study and their choice of occupation. Thus
the relative incomes of teachers strongly influence whether a student decides to go into
teaching. Manski (1987) has calculated a wage elasticity for the supply of teachers of between
2 and 3. Dolton (1990) has shown in various analyses using graduate data from England that
the relative income of teachers and relative wage growth are both determining factors when
choosing a profession. The higher that teachers’ wages are relative to other occupations, the
more likely it is that graduates will go into teaching (Dolton 1990; Dolton and Makespeace
1993; Chevalier et al. 2002). Hanushek and Pace (1995) find that in addition to the length of
the educational training programme, and the certification examination, relative wages
2
also
influence the supply of teachers, but they dispute whether this effect has a strong influence on
the choice of studies (which is in this paper equated with occupational choice).
Most authors agree that income influences how long someone will remain a teacher. Almost
all studies show that increasing teachers’ wages would result in a significantly lower turnover
and/or exit rate
3
(Chevalier et al. 2002; Hanushek et al. 2001, 1999; Dolton and Van der
2
Using probit estimates the authors find that relative income has a weak but still significant effect: a wage
increase of 10% would result in a rise of 0.7 percentage points among graduates of universities (Hanushek and
Pace, 1995, p. 112ff.).
3
Chevalier et al. (2002), for example, have calculated that a 10% rise in the relative teacher’s wage would
improve the probability of the average UK university graduate still teaching after six years in the profession by
10 percentage points. This means that a good 24% (or approximately 11,000 more) of the 1990 graduate cohort
would still have been teaching in 1996 instead of the 14% that in fact remained in the profession. A 10% relative
wage increase would also reduce the likelihood of quitting by 9% (Dolton and Van der Klaauw 1995, 442).
7
Klaauw 1995, 1996, 1999; Gritz and Theobald 1996; Mont and Rees 1996, Murnane 1991;
Murnane and Olsen 1990).
However, Hanushek et al. (1999, 2001) argue that local school factors and pupils’
characteristics (origin, social class, performance, etc.) have a stronger influence on the
mobility of teachers than the wage or wage differential.
4
The literature cited shows that, in general, teachers and students seem to react to income
differences as well as to non-monetary incentives in the same manner as people in other
occupations, and that they also take financial factors into consideration when deciding
whether to become or remain a teacher.
However, this does not mean that higher wages automatically also attract better teachers. The
effect of income on teaching quality is a controversial issue in the literature and is difficult to
test methodologically. Little empirical evidence has been offered one way or the other up to
now; most studies have not been able to demonstrate an effect (Hanushek 1997 and 1999;
Manski 1987). In theory, general wage increases expand the supply of teachers in the labour
market, but the quality of instruction will only improve if the employment process is
selective.
5
Loeb and Page (2000) argue that because different school districts show different
teacher supply curves, the lack of empirical evidence is due to the methods employed being
inadequate. They go on to say that we thus have to control for alternative occupational
opportunities and for non-pecuniary school-related parameters (profiles of the pupils, their
parents, the school, community, etc.) if we want to be able to estimate the relevant wage
elasticities. Using an appropriate model (a wage-quality diagram), the authors find that
4
Hanushek et al. 2001, 12: “Teachers systematically favor higher achieving, non-minority, non-low income
students.”
8
teachers’ incomes have a significant effect on their pupils’ school performance (as measured
by the high school dropout rate and the proportion who go on to college).
6
Because teachers’
qualifications correlate strongly with school characteristics and the demographic makeup of a
student body,
7
Loeb proposes specific incentives (such as targeted wage increases and/or
improving the facilities of less attractive schools) to offset such discrepancies (Loeb 2000,
15).
Temin (2002) relates the sharp drop in teachers’ relative wages that took place in the United
States in the 1970s and 1990s to both changes in the labour market for women and cutbacks in
the public funds available to local communities. Women’s wages have gone up as their
employment opportunities have broadened, so that non-teaching occupations are now much
more attractive to them.
8
This means, however, that relative teacher wages have declined
especially for female teachers, a factor that partially explains the shrinkage in the supply of
women in teaching. Temin defines the demand for teachers as a function of both price (wage)
and quality,
9
arguing that a market in which quality depends on wages (among other factors)
will not exhibit a single equilibrium but rather multiple equilibria of differing degrees of
stability (see also Figure 2 of this paper).
10
The problem is how to move from a stable
low-wage, low-quality equilibrium to one of high wages and high quality.
5
Regarding the problem of teacher selection, see primarily Ballou and Podgursky 1995, 1998; and Miles and
Darling-Hammond 1997.
6
A 10% wage increase (which would offset the relative decline in wages that occurred in the 1980s) would
reduce the high school dropout rate by more than 3%, and lift the number of those enrolling in college by 8%
(Loeb and Page 2000, 406).
7
Teachers in urban schools are substantially less qualified than those in suburban and rural areas; the same is
also true for schools with a high proportion of poor, black, or Hispanic pupils (Loeb 2000, 12ff.).
8
The wage premium of women with some graduate education compared to female teachers was less than 20% at
the end of the 1970s, versus more than 40% today. Women with only a college education but no graduate
education (i.e. comparable to primary school teachers) earn about 10% more (Temin 2002, 13ff.).
9
Analogously to Akerlof’s “lemon” model.
10
Direct wage effect: declining demand. Indirect effect: rising demand. If the indirect effect is stronger than the
direct, we have a rising demand curve with multiple equilibria (Temin 2002, 5).
9
The data
Since there is little statistical material available on teachers and their wages in Switzerland,
and what exists is often inadequate, it was not that easy to find a data set suitable for
answering our questions. The data material employed here is thus limited in various respects.
Because, however, it is longitudinal in nature, it enables us to separate some of the economic
influences from those that are structural in nature.
The first limitation concerns the type of teacher for whom data is available. Since educational
policies in Switzerland are the responsibility of the cantons, the training that teachers receive
is not uniform, a fact that tends to negatively influence the statistics that are collected. The
situation is a bit better for upper secondary teachers, not least because teachers at this level are
normally required to have a traditional university education.
Switzerland has conducted a full census of graduates of traditional universities every two
years for more than 20 years now, and of graduates of the universities of applied science, for
about 10 years. Our study focuses on the data collected for the former group, whom we will
refer to as “university graduates”, and for whom we use the complete series of surveys
conducted from 1981 to 1999 (to our knowledge, this is the first time this has been done). The
2001 survey had been completed by the time of our study, but the data not yet released. Since
the data set available to us contains ten waves of cross-sectional data for university graduates,
we are thus able to combine the advantages of cross-sectional and longitudinal observation by
means of a pooled cross-section analysis. In particular, the ability to observe the same
question over a period of 20 years allows us to supplement the data set with other longitudinal
data like economic indicators. We can also test the influence of systemic changes such as the
impact of educational reforms and similar measures.
10
The data describing systemic changes were gathered by surveying all of the traditional
universities. They were first sent a written questionnaire; we then followed this up in many
instances with oral interviews. The questions asked concerned any reforms of their study
programmes, changes in admission requirements, or other similarly far-reaching
modifications of their institution’s organization that had taken place during the previous
20 years. For changes which we thought might have an effect on the question of interest to
us - the occupations chosen by their graduates - we constructed dummy variables for the
institution starting from when the change was expected to have influenced the graduates.
One disadvantage associated with this data set was that we have had to limit our analysis to
the first job held after graduation, i.e. we are only able to determine whether wage elasticity is
significant for the graduate initial decision concerning whether to go into teaching. We are
unable to examine whether wage elasticity plays a role in deciding at a later date whether to
remain a teacher or change one’s occupation. For this we would need regular follow-up
surveys. Such surveys were not carried out for the period covered here, but they are planned
for the future.
Empirical model
The empirical procedure employed is a relatively simple and classical one, well adapted to the
data. In the end, the proposition of interest to us and which we want to test is whether the
wage differential between teaching and other occupations influences the likelihood of
becoming a teacher. To do this, we estimate a probit model that includes, in addition to
personality characteristics (vector X), an individually estimated wage differential for each
11
person observed (lnW
T
– lnW
O
) as an independent (i.e. explanatory) variable for the
probability of the person choosing teaching as a profession after graduation (T).
1210
)ln(ln uXWWT
t
o
t
T
tt
++−+=
(1)
The individual wage differential can be estimated in various ways. We chose the following
method, which is in line with those employed by Dolton (1990) and by Chevalier et al.
(2002).
11
It can be broken down into three stages.
We first estimate a structural equation which we then use to calculate the selection correction
in the subsequent wage equations. The purpose of the equation is to capture the composition
of the teaching population and of non-teaching university graduates so that we can account for
wage effects related to differences in the makeup of the two groups. Since one does not
become a teacher by sheer accident, we cannot assume that the wages of non-teachers are a
good predictor for teachers’ wages if the latter do not choose to go into teaching, and vice
versa.
We estimate a reduced form probit model and calculate from this (equation 2) the inverse
Mills ratio ( ), which we insert into the wage estimates (equations 3 &4) to correct for
selection bias.
110
uXT
tt
++=
(2)
22
'
10
ln uXW
o
t
ooT
t
+++=
(3)
11
Chevalier et al. (2002) also employ matching, which we do not use.
12
32
'
10
ln uXW
o
t
ooo
t
+++=
(4)
One of the most important points concerning this selection correction is that the vector (X) in
equation 2 must not be identical with the vector (X’) in equations 3 and 4 (identification
problem). In other words, we need to have factors that only influence the likelihood of
becoming a teacher and which we can assume do not directly affect teachers’ wages. Whether
this identification process succeeds is entirely dependent on how well we choose which
personality characteristics and other variables are to be included or excluded.
In our empirical analysis we now take account of two variables in vector X’ of equation 2 that
do not form part of vector X’ in the subsequent equations: the university at which the diploma
was earned, and the type of certificate required for admission to the university. This may be
surprising at first, but all of the Swiss universities proved to be so similar in quality in the past
that we can assume that while the nature of the diploma earned was relevant for wages, the
university attended was not. On the other side we can assume that due to specialisation of
some of the universities, the choice of the university has an impact on the likelihood to
become a teacher.
Most other studies have included parents’ education as a variable influencing a graduate’s
choice of occupation, but not directly influencing wage. Since our data set did not contain any
information on parents’ education, we replaced this with the certificate required for
admission. In the past Switzerland has exhibited a high degree of social reproduction,
especially with regard to access to universities. We can thus assume that those who use a
gymnasium diploma to enter university - our reference group - normally have parents with
university level degrees, and that the minority of students who enter university with some
other certificate most probably have parents who do not have university level degrees. We can
13
also assume that the certificate providing admission to a university has some influence on
which occupation is chosen after graduation, but no direct influence on wage.
We also have variables which we expect to directly affect the wage but which should not have
any direct influence on occupational choice. These are the two economic variables GDP and
unemployment, and the workplace’s location (because there is a certain gap between the
wages paid in large cities and in other areas). We also take into account whether the person
holds a full-time or part-time job. It is interesting to note that university graduates normally
receive a better entry wage for a part-time position than for one that is full-time, a situation
exactly the opposite of what applies for the average of all wages. Although certain
occupations unquestionably have more part-time positions than others, we do not assume that
this directly influences the choice of occupation.
Thus we can use equations 3 and 4 to calculate the entry wages for teachers (lnW
T
) and
non-teachers (lnW
O
). For each graduate included in our data set we estimate a teaching and a
non-teaching wage as well as the difference between the two, which we then insert into
equation 1. The influence of this variable describes the wage elasticity of the teacher supply,
at least as regards the first job after graduation.
The results
The estimates for equations 2 to 4 were performed for the group of teachers as a whole and
broken down into gymnasium and non-gymnasium teachers (see Tables 1 and 2). While they
show that the qualitative statements are correct for the aggregate group and for the two
subgroups, the quantitative results vary. This is to be expected: although the graduates have a
14
wide variety of teaching occupations at different school levels to choose from (generally
primary through upper secondary), gymnasium teachers on average have the best wage
prospects.
The labour markets for the two categories of teachers have also moved in different directions
(see Figures 1 and 2 in the Appendix). In the case of gymnasium teachers, the proportion of
graduates who chose this for their first job dropped sharply in the mid-1980s and then held
steady at about 6% thereafter; but while their wage differential over non-teachers also
declined at first, it then rose sharply again, especially in the first half of the 1990s. The
proportion of graduates who chose to go into other areas of teaching did not go up or down
dramatically but fluctuated continuously around approximately 5%. Their wage differential,
however, which is lower on average than for gymnasium teachers, plunged in the 1980s and
then recovered somewhat in the following decade. If we were to base our interpretation on
Figures 1 and 2 solely, we would not necessarily conclude that the proportion of graduates
who choose to become teachers would react sensitively to the anticipated wage differential.
We can interpret the regressions in Tables 2 and 3 as showing that the profile of graduates
who tend to become teachers differs from that of other graduates on the basis of age and field
of study, and that men predominate significantly among gymnasium teachers. This contrasts
sharply with the predominance of women at the primary and lower secondary levels. It is also
interesting that the certificate providing access to the university sometimes plays a significant
role, i.e. students who come to the university already equipped with a primary or secondary
school teaching certificate are more likely to go back to teaching after graduation. The easing
of university admission requirements that took place in the early 1980s (more students with a
teacher diploma before entering the university) stimulated a stronger inflow of people that
15
were more likely to take up teaching after university. But this inflow (see also Figure 4)
lessened again in the following decade.
Most of the system variables we constructed are not significant, and for this reason, the results
are not shown separately, with one exception: When the University of Berne changed its
admission requirements to allow primary school teachers to study in any field they wanted to,
the number of graduates who later decided to become gymnasium teachers went up
significantly (but not, however, the proportion who chose other areas of teaching). This
created somewhat of a dilemma for the Canton of Berne. The good thing was that primary
school teachers admitted to the university were more likely to return to teaching after
graduation, but the bad thing was that they were also more motivated to leave their profession
as primary teachers to attend university. Although this filled in holes at the upper secondary
level and even prevented their arising, it created some problems for the canton’s primary
schools.
Lambdas were constructed from the initial reduced probit estimate and then employed as
variables in the relevant wage regressions to correct for selection bias. However, the lambda
is significant in only one of the four wage regressions. The wage regressions do not present
any major surprises when compared to the wage estimates for Switzerland contained in the
literature even though we estimate entry wages solely. One element of interest, and which
agrees with earlier results, is that men earn more than women even when we control for all
observable differences (however, we did not test this effect for possible time dependence).
Comparing the wages of teachers and non-teachers, it also striking that nationality seems to
play a significant role only for the teachers, and that their wages do not exhibit any
dependence on economic conditions. Unemployment has a significant negative influence on
non-teachers’ wages, but not on those of teachers. It is precisely this difference in the
16
dependence of wage formation on economic conditions that explains the sharp increase in the
wage differential between gymnasium teachers and non-teachers that took place in the first
half of the 1990s (Figure 1 in the Appendix). The differential is explained by the stagnation in
teachers’ wages, and a sharp drop in entry wages for non-teachers.
If we now use the two wage regressions to calculate for each graduate a separate wage for
teaching and non-teaching, we have the mean results shown in Table 1.
Table 1: Predicted log gross annual earnings for teachers and non-teachers in their first
job following graduation (CHF in 1999 values in brackets)
Teachers Non-teachers Difference
Teaching job 3.877
(CHF 77’246)
3.769
(CHF 69’337)
0.108
Non-teaching job 3.755
(CHF 68’374)
3.700
(CHF 64’715)
0.055
Difference 0.122
(CHF 8’872)
0.069
(CHF 4’622)
We see that teachers in general have a higher entry wage than non-teachers. It is also the case
that non-teachers can expect to receive a higher wage if they choose teaching as their initial
occupation, but the anticipated difference in wages between the two options is substantially
higher (0.122 versus 0.069) for graduates who do in fact become teachers. This indicates that
the initial employment choice is a rational decision. This result most certainly is not to be
interpreted as meaning that all graduates should have become teachers because of the
generally higher entry wage; after all, we know only this initial wage and can only deduce
how their income will perform thereafter, something that can vary greatly between teachers
and non-teachers.
12
12
Although we have no information in this data set on wages throughout the life cycle of a person, we know
from other sources that the entry wages are representative insofar that the discounted life income stream that
17
Our regression results are shown in Tables 4 and 5 (in the Appendix). The expected individual
wage differential has a significant positive influence on the likelihood of choosing teaching as
one’s initial occupation for both subgroups, with the marginal effect larger for
non-gymnasium teachers than for gymnasium teachers. The latter already have a higher wage
differential. If the wage differential in favour of teachers is increased by 10%, the proportion
of a graduate cohort that chooses to go into teaching is 1.2 percentage points higher for
gymnasium teachers,
13
and 1.8 percent higher for other teachers. Although the field of study
remains one of the main factors determining whether teaching will be chosen as an occupation
even after the wage differential is accounted for, the teacher supply in the past has reacted
very sensitively to the relative wage at the time of entry into a profession. Comparison of the
two categories of teachers, and comparison with the wage elasticities found in other countries,
leads us to think that elasticity is high primarily where the differential is either very small or
even in favour of non-teachers.
After determining that a significant wage elasticity of supply exists for both gynmasium and
other teachers, the question arises - especially as regards gymnasium teachers (see also
Figure 1) - why the proportion of graduates did not increase strongly in the 1990s when the
wage differential in favour of teachers rose so sharply? In accordance with the results shown
in Tables 2 and 4, this is primarily because men have accounted for a steadily and notably
smaller proportion of graduates over the last 20 years (see Figure 4), and men become
gymnasium teachers significantly more often than women. Two other factors are that the
proportion of people who study literature, mathematics and the sciences, fields that supply
particularly large numbers of teachers, has been consistently small; and the proportion of
teachers can expect is higher than the life income of non-teachers (on average). While there is only a slight
difference for men, the difference is substantial for women.
18
graduates who already have a teaching certificate before entering the university has been
decreasing since the mid-1980s. Interpreted this way, we can assume that the proportion of
graduates choosing to become gymnasium teachers in the 1990s would have declined even
further if the positive wage differential had not offset this trend. This naturally raises the
question of how the labour market for upper secondary school teachers will develop should
the overall labour market enjoy another prolonged recovery and thus reduce the wage
differential between teachers and non-teachers.
Conclusions
The labour market for teachers is special in at least two respects. Firstly, in most of the OECD
countries the dominance of public schools means that the state is virtually the sole employer
and thus a monopsonist. Secondly, the market is characterized by frequent disequilibria that
take the form of a shortage or excess of teachers. One possible reason for this might be that
the state in its role of employer provokes these disequilibria through the rigidities of its
wage-setting process and education budgets. However, we also know that governmental
actors lack the knowledge needed for active wage formation. The wage elasticity of the
teacher supply is a critical point in this respect because it must be ascertained before we can
estimate the effects that changes in the employment rate for teachers will have on the
education budget.
In order to learn more about the wage elasticity of the teacher supply in Switzerland, this
paper has estimated wages for teachers and non-teachers, with these estimates including a
selectivity correction for wage distortions that might possibly be caused by the occupational
13
For 1999, this would mean some 100 more graduates would choose to become gymnasium teachers if their
19
selection process. However, the data allows us to estimate wage elasticity only for entry
wages and for upper secondary school teachers. These estimates provide us with two expected
entry wages for each university graduate, one for the case that the graduate decides to become
a teacher, and one for the case that she/he does not. These individual wage estimates enable us
to calculate individual wage differentials which we then use as the explanatory variable for
estimating the likelihood of a graduate going into teaching. In the cases examined, the wage
differential is significant and shows a wage elasticity for the teacher supply of 0.12% to
0.18%. This elasticity is substantially lower than what we have found in the literature for
other countries. It is probably due to the fact that teachers’ salaries are generally high in
Switzerland - teachers here enjoy a positive wage differential over those in other occupations,
the opposite of what is normally the situation elsewhere.
Our results also show that the expected wage differential between teaching and other
occupations is larger for teachers than for non-teachers. This result leads us to suppose that
university graduates choose their initial occupation on a rational basis, and that they take
monetary factors into consideration when making this decision.
Finally, our analyses show that the number of graduates who choose to become teachers (and
especially gymnasium teachers) would of necessity have gone down in the 1990s in light of
the changes in the demographic composition of the student population that have taken place
over the last 20 years. That this did not occur was a consequence of the recession resulting in
a significant increase in the wage differential in favour of teachers.
entry wage were raised by CHF 1’300 (860 Euro) and wages for other occupations remain unchanged.
20
Literature
Ballou, D., Podgursky, M. (1995). Recruiting smarter teachers, Journal of Human Resources, 30 (2), 326-338.
Ballou, D., Podgursky, M. (1998). Teacher Recruitment and Retention in Public Schools, Journal of Policy
Analysis and Management, 17 (3), 393-417.
Chevalier, A., Dolton, P., McIntosh, St. (2002). Recruiting and Retaining Teachers in the UK: An Analysis of
Graduate Occupation Choice from the 1960s to the 1990s, London: Centre for Economics of Education.
Dolton, P. (1990). The economics of UK teacher supply: the graduate’s decision, Economic Journal, 100,
Conference papers, 91-104.
Dolton, P., Klaauw, W. van der (1995). Leaving teaching in the UK: a duration analysis, Economic Journal, 105,
431-444.
Dolton, P., Klaauw, W. van der (1996). Teaching Salaries and Teacher Retention, in Becker, W. E., Baumol, W.
J., Assessing Educational Practices: the Contribution of Economics, Cambridge MA: MIT Press, 239-272.
Dolton, P., Makespeace, G. H. (1993). Female labour force participation and the choice of occupation. The
supply of teachers, European Economic Review, 37, 1393-1411, North Holland.
Dolton, P., Klaauw, W. van der (1999). The turnover of teachers: a competing risks explanation, Review of
Economics and Statistics, 81, 543-552.
Gritz, R.M., Theobald, N.D. (1996). The Effects of School District Spending Priorities on Length of Stay in
Teaching, Journal of Human Resources, 31 (3), 477-512.
Hanushek, E.A. (1997). Assessing the effects of school resources on student performance: An update,
Educational Evaluation and Policy Analysis 19, 2, 141-164.
Hanushek, E.A., Pace, R.R. (1995). Who Chooses To Teach (and Why)?, Economics of Education Review, 14
(2), 101-117.
Hanushek, Kain, J., Rivkin, E.G. (1999). Do Higher Salaries Buy Better Teachers? NBER Working Paper Series,
No. 7082, National Bureau of Economic Research, Cambridge MA.
Hanushek, Kain, J., Rivkin, E.G. (2001). Why Public Schools Loose Teachers? NBER Working Paper Series,
No. 8599, National Bureau of Economic Research, Cambridge MA.
Loeb, S. (2000). How Teachers’ Choices affect what a Dollar can buy: Wages and Quality in K-12 Schooling.
Paper prepared for the Education Finance Research Consortium.
Loeb, S., Page, M.E. (2000). Examining the link between teacher wages and student outcomes: The importance
of alternative Labor Market Opportunities and non-pecuniary variation, The Review of Economic and Statistics,
82, 3, 393-408.
Manski, C.F. (1987). Academic ability, earnings, and the decision to become a teacher: evidence from the
National Longitudinal Study of the High School Class of 1972, in Wise, D. (ed) Public Sector Payrolls, Chicago
IL: University of Chicago Press.
Miles, K.H., Darling-Hammond, L. (1997). Rethinking the allocation of teaching resources: some lessons from
high performing schools, Paper prepared for the University of Wisconsin-Madison, Center for Education
Research, Consortium for Policy Research in Education.
Mont, D., Rees, D.I. (1996). The Influence of Classroom Characteristics on High School Teacher Turnover,
Economic Inquiry, 34 (Jan), 152-167.
Murnane, R.M. (1991). Who will teach? Policies that matter, Cambridge, Mass.: Harvard University Press.
Murnane, R.M., Olsen, R.J. (1990). The effects of salaries and opportunity costs on length of stay in teaching:
evidence from North Carolina, Journal of Human Resources, 25, 106-124.
OECD (2002): The Teaching Workforce: Concerns and Policy Challenges, Education Policy Analysis 2002,
Paris: OECD, 65-87.
Santiago, P. (2002): Teacher Demand and Supply: Improving Teaching Quality and addressing Teacher
Shortages - A Literature Review and a Conceptual Framework for Future Work, Education Working Papers No.
1, Paris: OECD.
Temin, P. (2002). Teacher Quality and the Future of America. NBER-Working Paper No. 8898, Cambridge MA.
21
Appendices
Figure 1: Percentage of graduates who become gymnasium teachers, and the wage
differential between teachers and non-teachers
Figure 2: Percentage of graduates who become non-gymnasium teachers,
and the wage differential between teachers and non-teachers
0
2
4
6
8
10
12
14
16
18
81 83 85 87 89 91 93 95 97 99
Year of Graduation
Percent
Graduates
Wage Difference
-4
-2
0
2
4
6
8
10
12
14
81 83 85 87 89 91 93 95 97 99
Year of Graduation
Percent
Graduates
Wage Difference
22
Figure 3: Percentage of graduates in fields that do and do not supply large
numbers of teachers
Figure 4: Percentage of male graduates, and graduates who enter university with a
primary or lower secondary school teaching certificate
0
5
10
15
20
25
30
35
40
81 83 85 87 89 91 93 95 97 99
Years
% of graduates
Exact Sc. & Literature
Law, Soc. Sc. & Medicine
0
1
2
3
4
5
6
81 83 85 87 89 91 93 95 97 99
Years
% of graduates with primary of secondary teacher diploma
40
45
50
55
60
65
70
75
80
% of male graduates
Primary and Seconary I teacher diploma
Male
23
Table 2: Determinants of entry wages for gymnasium teachers
Variables Selection Teachers Non-teachers
Male 0.122* 0.049* 0.060*
Age 0.017* 0.020* 0.025*
University of Berne -0.679*
University of Basel -0.171*
University of Lucerne -0.004
University of Fribourg -0.190*
University of Geneva -0.230*
University of Lausanne 0.151**
Federal Institute of Technology in Lausanne -0.894*
University of Neuchâtel -0.343*
University of St. Gallen -0.388*
Federal Institute of Technology in Zurich -0.309*
Graduate year 1981 0.359* -0.035 -0.086*
Graduate year 1983 0.353* -0.064** -0.041*
Graduate year 1985 0.145** -0.025 -0.045*
Graduate year 1987 -0.038 -0.084* -0.020**
Graduate year 1991 0.032 Dropped Dropped
Graduate year 1993 -0.039 0.033 0.027*
Graduate year 1995 0.014 -0.003 -0.029*
Graduate year 1997 -0.038 Dropped Dropped
Graduate year 1999 -0.059 -0.052 -0.023*
French-speaking part of Switzerland 0.132** -0.232* -0.115*
Italian-speaking part of Switzerland -0.030 -0.115** -0.173*
Nationality Swiss 0.238* 0.072** -0.000
Gross domestic product -0.006 -0.001
Unemployment rate -0.010 -0.021*
Part-time worker 0.124* 0.089*
Zurich -0.058* 0.068*
Basel -0.128* -0.061*
Berne -0.038 0.080*
Zug 0.112 0.071*
Geneva 0.112* -0.014
Natural sciences 0.844* -0.010 -0.041*
Natural sciences + TD
14
2.500* -0.007 -0.099
Theology 1.564* -0.403* -0.134*
Theology + TD 2.494* -0.220 0.016
Literature and languages 1.564* -0.062 -0.100*
Literature and languages + TD 2.494* -0.030 -0.003
History 0.934* -0.076 -0.136*
History + TD 2.404* -0.041 -0.023
Social sciences -0.092 -0.020 -0.132*
Social sciences + TD 1.376* 0.209 -0.027
Economics -0.065 0.050 0.123*
Economics + TD 1.604* 0.079 0.162*
Law -0.929* 0.005 -0.367*
Exact sciences 1.167* -0.023 0.124*
Exact sciences + TD 2.871* 0.021 0.028
Engineering -0.558* -0.136 0.082*
Vocational certificate -0.160*
Certificate as primary teacher 0.109
Certificate as lower secondary teacher 0.414*
University of applied sciences diploma -0.242
Admission criteria change, University of Berne 0.346**
Lambda 0.652 -0.014
Constant -2.884* 3.283 3.088
F test 14.79 250.97
Log-likelihood -4874.8
Pseudo (or adjusted) R
2
0.3734 0.250 0.252
Observations 32931 1678 27634
*,** indicate significance at the 1% and 5% levels respectively. The reference graduate is female, had an A-level
university entrance certificate, graduated in 1989 in medicine from the University of Zurich, and works full-time in the
German-speaking part of Switzerland.
14
TD = Teacher diploma
24
Table 3: Determinants of entry wages for non-gymnasium teachers
Variables Selection Teachers Non-teachers
Male -0.042 0.081* 0.061*
Age 0.013* 0.017* 0.025*
University of Berne -0.127
University of Basel -0.039*
University of Lucerne 0.807*
University of Fribourg 0.151**
University of Geneva 0.214*
University of Lausanne 0.018
Federal Institute of Technology in Lausanne -0.762*
University of Neuchâtel 0.250**
University of St. Gallen -0.052
Federal Institute of Technology in Zurich 0.377*
Graduate year 1981 0.449* 0.044 -0.088*
Graduate year 1983 0.232* -0.042 -0.041*
Graduate year 1985 0.133** 0.000 -0.045*
Graduate year 1987 0.266* 0.016 -0.026**
Graduate year 1991 0.058 Dropped Dropped
Graduate year 1993 0.250* 0.034 0.027**
Graduate year 1995 0.112 -0.010 -0.027**
Graduate year 1997 0.173** Dropped Dropped
Graduate year 1999 0.123** -0.010 -0.025**
French-speaking part of Switzerland 0.325* -0.139* -0.121*
Italian-speaking part of Switzerland 0.351* -0.152** -0.178*
Nationality Swiss 0.173** 0.140* -0.005
Gross domestic product 0.011 -0.002
Unemployment rate -0.008 -0.021*
Part-time worker 0.114* 0.095*
Zurich -0.027 0.063*
Basel -0.096** -0.064*
Berne -0.027 0.073*
Zug 0.103 0.070*
Geneva 0.086** -0.023**
Natural sciences 0.313* -0.103 -0.030**
Natural sciences + TD 1.054* -0.107 0.033
Theology -0.242 -0.083* -0.141*
Literature and languages 1.069* -0.126* -0.068*
Literature and languages + TD 1.290* -0.054 -0.044
History 0.767* -0.147* -0.125*
History + TD 1.152* -0.014 -0.009
Social sciences 0.557 -0.083** -0.142*
Social sciences + TD 1.274* -0.007 -0.019
Economics -0.069 0.017 0.124*
Economics + TD 1.890* 0.080 0.132**
Law -0.275* -0.180** -0.367*
Exact sciences 0.265* -0.050 0.121*
Exact sciences + TD 0.461** -0.073 0.102**
Engineering -1.222* 0.181 0.082*
Vocational certificate -0.004
Certificate as primary teacher 0.750*
Certificate as lower secondary teacher 0.700*
University of applied sciences diploma 0.137
Admission criteria change, University of Berne 0.160
Lambda 0.304** -0.020
Constant -2.951* 3.154* 3.091*
F test 8.44 272.86
Log-likelihood -5296.68
Pseudo (or adjusted) R
2
0.205 0.166 0.259
Observations 32929 1346 27964
*,** indicate significance at the 1% and 5% levels respectively. The reference graduate is female, had an A-level
university entrance certificate, graduated in 1989 in medicine from the University of Zurich, and works full-time in the
German-speaking part of Switzerland.
25
Table 4: Determinants of becoming a gymnasium teacher
Variables Probit Marginal effect
Male 0.158* 0.005
Age 0.036* 0.001
University of Berne -0.650* -0.014
University of Basel -0.220* -0.006
University of Lucerne -0.105 -0.003
University of Fribourg -0.200** -0.006
University of Geneva -0.553* -0.013
University of Lausanne 0.170** 0.007
Federal Institute of Technology in Lausanne -0.959* -0.014
University of Neuchâtel -0.306** -0.008
University of St. Gallen -0.445* -0.010
Federal Institute of Technology in Zurich -0.329* -0.009
Graduate year 1981 -0.030 -0.001
Graduate year 1983 0.303* 0.014
Graduate year 1985 0.008 0.000
Graduate year 1987 0.083 0.003
Graduate year 1991 -0.097 -0.003
Graduate year 1993 -0.340* -0.009
Graduate year 1995 -0.321* -0.009
Graduate year 1997 -0.319* -0.009
Graduate year 1999 -0.130** -0.004
French-speaking part of Switzerland 0.215* 0.008
Italian-speaking part of Switzerland -0.443* -0.010
Nationality Swiss 0.011 0.000
Natural sciences 0.726* 0.053
Natural sciences + TD 1.930* 0.376
Theology 0.719* 0.053
Theology + TD 3.371* 0.864
Literature and languages 1.376* 0.169
Literature and languages + TD 2.438* 0.564
History 0.702* 0.050
History + TD 2.230* 0.521
Social sciences -0.417 -0.010
Social sciences + TD 0.429* 0.024
Economics 0.204** 0.008
Economics + TD 1.800* 0.326
Law -2.289* -0.034
Exact sciences 1.622* 0.246
Exact sciences + TD 2.696* 0.672
Engineering 0.103 0.004
Vocational certificate -0.139** -0.004
Certificate as primary teacher 0.057 0.002
Certificate as lower secondary teacher 0.344* 0.018
University of applied sciences diploma -0.202 -0.006
Admission criteria change, University of Berne 0.382** 0.019
Expected wage differential (lnW
T
-lnW
O
) 3.345* 0.117
Constant -3.369*
Observed probability 0.059
Log-likelihood -4294.10
Pseudo (or adjusted) R
2
0.389
Observations 31144
*,** indicate significance at the 1% and 5% levels respectively. The reference graduate is
female, had an A-level university entrance certificate, and graduated in 1989 in medicine from
the University of Zurich..
26
Table 5: Determinants of becoming a non-gymnasium teacher
Variables Probit Marginal effect
Male -0.098** -0.005
Age 0.043* 0.002
University of Berne -0.166 -0.007
University of Basel -0.072 -0.003
University of Lucerne 0.649** 0.060
University of Fribourg 0.088 0.005
University of Geneva -0.117 -0.005
University of Lausanne 0.021 0.001
Federal Institute of Technology in Lausanne -0.879* -0.020
University of Neuchâtel 0.197** 0.012
University of St. Gallen 0.314* -0.007
Federal Institute of Technology in Zurich -0.191** 0.019
Graduate year 1981 -0.191** -0.008
Graduate year 1983 0.277* 0.017
Graduate year 1985 -0.117 -0.005
Graduate year 1987 0.032 0.002
Graduate year 1991 0.172** 0.010
Graduate year 1993 0.095 0.005
Graduate year 1995 -0.022 -0.002
Graduate year 1997 0.058 0.010
Graduate year 1999 -0.014 -0.001
French-speaking part of Switzerland 0.122** 0.007
Italian-speaking part of Switzerland 0.065 0.003
Nationality Swiss -0.366* -0.025
Natural sciences 0.475* 0.036
Natural sciences + TD 1.400* 0.237
Theology -0.450* -0.014
Literature and languages 1.087* 0.135
Literature and languages + TD 1.432* 0.252
History 0.716* 0.067
History + TD 0.982* 0.122
Social sciences 0.283* 0.018
Social sciences + TD 0.959* 0.118
Economics 0.318* 0.020
Economics + TD 1.665* 0.328
Law -1.046* -0.029
Exact sciences 0.827* 0.086
Exact sciences + TD 0.981* 0.122
Engineering -1.511* -0.026
Vocational certificate 0.001 0.000
Certificate as primary teacher 0.559* 0.047
Certificate as lower teacher 0.470* 0.037
University of applied sciences diploma 0.102 0.037
Admission criteria change, University of Berne 0.131 0.007
Expected wage differential (lnW
T
-lnW
O
) 3.656* 0.179
Constant -3.129*
Observed probability 0.047
Log-likelihood -4574.90
Pseudo (or adjusted) R
2
0.226
Observations 31142
*,** indicate significance at the 1% and 5% levels respectively. The reference graduate is female,
had an A-level university entrance certificate, and graduated in 1989 in medicine from the
University of Zurich..
IZA Discussion Papers
No.
Author(s) Title
Area Date
717 A. Hijzen
H. Görg
R. C. Hine
International Fragmentation and Relative Wages
in the UK
2 02/03
718 E. Schlicht
Consistency in Organization 1 02/03
719 J. Albrecht
P. Gautier
S. Vroman
Equilibrium Directed Search with Multiple
Applications
3 02/03
720 T. Palokangas Labour Market Regulation, Productivity-
Improving R&D and Endogenous Growth
3 02/03
721 H. Battu
M. Mwale
Y. Zenou
Do Oppositional Identities Reduce Employment
for Ethnic Minorities?
1 02/03
722 C. K. Spiess
F. Büchel
G. G. Wagner
Children's School Placement in Germany: Does
Kindergarten Attendance Matter?
6 02/03
723 M. Coles
B. Petrongolo
A Test between Unemployment Theories Using
Matching Data
3 02/03
724 J. T. Addison
R. Bailey
W. S. Siebert
The Impact of Deunionisation on Earnings
Dispersion Revisited
2 02/03
725 S. Habermalz An Examination of Sheepskin Effects Over Time 1 02/03
726 S. Habermalz Job Matching and the Returns to Educational
Signals
1 02/03
727 M. Raiser
M. Schaffer
J. Schuchardt
Benchmarking Structural Change in Transition 4 02/03
728 M. Lechner
J. A. Smith
What is the Value Added by Caseworkers? 6 02/03
729 A. Voicu
H. Buddelmeyer
Children and Women's Participation Dynamics:
Transitory and Long-Term Effects
3 02/03
730 M. Piva
M. Vivarelli
Innovation and Employment: Evidence from
Italian Microdata
2 02/03
731 B. R. Chiswick
N. DebBurman
Educational Attainment: Analysis by Immigrant
Generation
1 02/03
732 A. Falk
A. Ichino
Clean Evidence on Peer Pressure 5 03/03
733 S. Wolter
S. Denzler
Wage Elasticity of the Teacher Supply in
Switzerland
2 03/03
An updated list of IZA Discussion Papers is available on the center‘s homepage www.iza.org.
