I. Introduction
Teacher shortages have been a long ongoing concern on
discussion, with the root cause such as difficulties in recruitment,
availability of fully certified teachers, etc. varying over time and across
logistics. In trying to resolve these, school officials opt to giving out
compensations like higher loans, condoning student loans in exchange for teaching
commitment, housing reserved for teachers, etc. – all of which aim to encourage
existing teachers to stay, and appeal to new teacher applicants. However the
success of such policies is contingent on the receptiveness of the supply (the
teacher pool).
The inability to understand the factors of teacher labor
supply is found to be one of the hurdles in developing efficient teacher labor
market policies. The teacher labor supply comprises of various decisions based
on a range of information and influences in different times. The paper focuses
on those who are already teaching and consider their decisions.[1]
According to some studies[2]
higher teacher compensation decreases the possibility that they leave the
profession, specifically once the differences in alternative earnings opportunities
are taken into account. However one setback for these studies is the limited
availability of information on working conditions that may be linked with
salary – that may actually lead to a biased result and a further decline in
understanding the actual cases.
To gain a better understanding of the impact of salary and
other school determinants on teacher movements a matched student/teacher panel
data on Texas public elementary schools is used. Provided with the large number
of teachers and teacher shifts in the data, teachers can be partitioned on the
basis of experience, type of school community, ethnicity and other determinants
and study differences in the sensitivity to salary and student attributes on
the basis of teacher experience, race and ethnicity. It shows that teacher
movements are much more robustly related to certain student features than to
salary differentials. Schools attending to large numbers of academically
disadvantaged, black or Hispanic pupils have a tendency to lose an ample portion
of teachers annually both to other districts and out of the Texas public
schools completely. A suggestion is that the supply curve seen by these
districts noticeably varies from that experienced by middle and upper middle
class communities wherein a far lower fraction of teachers aim to better their
employment arrangement by transferring to another public school.
II. Determinants of
Supply of Teachers
The labor supply of teachers for a particular district in a
certain area is set to be a function of the wage and working condition in that
district and the amenities and employment opportunities in the area.
A. Salaries
Teachers’ salaries are expected to differ within a district
and such differences reflect the various components of teacher wage contracts
involving experience, education, and other factors.
B. Alternative
Earnings Opportunities and Amenities
One must consider the variations in the alternative
opportunities for teachers, particularly competitions for specific teachers
such as math and science teachers as opposed to those in other areas of
expertise. There are empirical evidences that affirm the thinking that
alternative earning opportunities affect teacher labor supply. Dolton and van
der Klaauw (1995, 1999) and van der Klaaw (1997) examined the impact of such on
teacher transitions and found evidence that opportunity wages have an impact on
the likelihood of both entry in and exit from the profession. Their results
were consistent to the work of Murnane and Olsen (1989, 1990).
C. Working
Conditions
There have been discussions that there is more to teaching
job than the general salary or incentives levels. According to Greenberg and
McCall (1974) and Murnane (1981), teacher preferences somewhat influence the
selection of schools. From Antos and Rosen (1975) study, it was observed that
teacher may be willing to have lower wage in exchange for better conditions in
their schools. Failure to consider the disparities in working and labor market
conditions may serve as an explanation to the failure of other studies in
identifying the core relationship between achievement and salaries (Loeb and
Page, 1998).
The main component of the empirical analysis in the paper is
the description of the movements of teachers across various types of schools
and student populations in order to examine the inclinations of teachers and
the how the compensations and earnings degree of differences is executed.
Included are the four measures of student attributes that are possibly
connected to teacher labor supply which are percent low income, percent black,
percent Hispanic, and average student achievement grade. However, it is yet
indefinite if these characteristics directly affect or serve as proxies for
other factors in teacher decisions. Regardless, the estimates will recognize
the schools experiencing the most difficulties in teacher labor market.
D. Personnel
Policies
Although the link between wages and employment can be
detected, it is not possible to theoretically conclude that the relationship is
a “supply function” for teachers. Instead the district hiring and retention
practices are significant components in the labor market for teachers. Ballou
and Podgursky (1995, 1997) and Ballou (1996) used this in the interpretation of
movements seen in the market, for instance, the observed negative relation between
retention and salaries is consistent with the positively sloped supply curve
and a positive relation between wages and likelihood that districts will not
re-employ teachers.
The paper illustrates the changes in the wage from shifting
districts in addition to the study of the likelihood that teachers leave
schools. Furthermore, it is uncertain that the increase in pay or improvements
in non-monetary compensations would be as great for involuntary as for
voluntary job transitions, hence the observed changes in pay and other
characteristics for school transfers must understate the gains of those who opt
to shift schools. Likewise the relation between the possibility of resigning
and salaries should be negative than that between the possibility of being laid
off and salaries. Also the estimated relationship of quitting to salaries must underestimate
the supply relationship.
III. The Texas
Database
The ability to comprehend the nature and results of teacher
labor market activities originates from the unique database developed under the
UTD Texas Schools Project. In coordination with the Texas Education Agency
(TEA) the project has combined various data sources to gather matched panel
datasets on students and teachers. The samples comprise the Texas public school
teachers in each year along with entire groups of Texas students which allow
precise descriptions of the schools of each teacher’s employment.
The Public Education Information Management System, the
TEA’s state-wide educational data base covers key demographic data (race,
ethnicity, gender) for students and teachers, and also student’s eligibility
for subsidized lunch. In addition it reports annual information on teacher
experience, wage, education, class size, grade, population served and subject.
Significantly, this database can be combined with information on both student
and teacher accomplishment.
In 1993, the Texas Assessment of Academic Skills (TASS) was
administered each spring to qualified students enrolled in grade three to
eight. Several special education and limited English proficient pupils are
excused from the assessment, as are other students for whom the test would not
be educationally appropriate. In each year about 15 percent of students do not
take the tests, which may be due to an exception or as a result of repeated
absences on testing days. Empirical salary schedules are developed for each
school district by means of the teacher micro data for 1993 to 1996. The salary
information does not include the various special compensations for individual
teachers.
IV. Teacher
Mobility, Salaries and Student Demographics
There is a significant ambiguity on the weight that a school
district has to enhance the supply and performance of the teachers through
compensation, class size and other non-monetary job attributes.
A. Transitions
Between and Within Districts (1993-1996)
One of the main goals is to distinguish the varying weights
of salary and other factors of job attractiveness. It was observed that
teachers were transferring.
Teachers are observed moving within districts, between
districts, and out of Texas public schools entirely annually between 1993 and
1996. Importantly, we have information about salaries and student
characteristics for both the sending and receiving schools for each transition.
On a general yearly basis of 79 percent of teachers stay in
the same school, 14 percent leave Texas public schools, 4 percent transfer
schools within districts and 3 percent change district. The estimates for
leaving Texas public schools involve people exiting teaching in addition to
people teaching either in private schools or outside of the state. Furthermore,
the rapid expansion of the Texas economy and student population during the
period certainly had influenced the teacher mobility.
Just like the general job turnover pattern found in the
labor market, movements vary sharply according to teacher experience.
Probationary teachers having zero to two years of experience have a higher
mobility within and across districts, and almost twice more likely to exit
Texas public schools than prime age teachers with 11 – 30 years experience and
were found to have a higher percentage of staying. However, mobility
accelerates as teachers approach retirement age as almost one-fourth of teachers
with over thirty years of experience leave the Texas public schools each year.
The national patterns of transitions across experience categories are somewhat
similar to that of Texas.
On the other hand the patterns of mobility have a tendency
to oppose the conventional knowledge that large urban districts are the proving
ground for teachers, who transfer to suburban jobs when feasible. According to
the data (limited to teachers transferring districts), among the teachers in
large urban districts, most of those switching districts relocate to suburban
schools, but as a whole, less than two percent of teachers in large urban
school change districts each year. The total number of those entering is just
faintly fewer than the number of those exiting urban districts. In this period
the share of Texas teachers in urban districts increased, which means that the
small net outflow of teachers from urban district is not just a reflection of
the variation in the distribution of teaching positions across types of community.
A similar pattern of mobility is observed for the subsample
of probationary teachers, wherein the net outflow from urban districts is
minute.
In general comparison to probationary suburban teachers,
probationary urban teachers are less probable to stay in the same schools by
one percentage point and this is a gap identical for teachers of all levels of
experience.
Two crucial distinctions between new teachers starting in
urban and starting in suburban:
1. Probationary
urban teachers are more likely to leave teaching than those in suburban
districts by 3.5 percentage points
2. Probationary
suburban teachers are rather more likely to just transfer schools within
districts
The switch from rural districts shows a very clear pattern,
as the bulk of movers transfer to a different rural district. Considerably
smaller number of rural teachers goes to urban districts than the case wherein
teachers are initially in urban or suburban districts.
These rates of movement amongst community types present no
information on the definite changes in salary and student composition.
Additional observations deal with relationship between
pre-transfer and post-transfer wage and student attributes for teachers who
change school districts.
The changes in salary are calculated by single years of
experience, for instance the change in wage for transferring teacher with four
years of experience is equal to the difference of the district mean salary of
fifth year teachers in the new district and district mean salary of fifth year teachers
in the old district, both computed in the year of the change. Note that the
consistent salary schedule data for teachers with ten or fewer years of
experience (which is approximately three-fourths of teachers transferring
districts) is the only available data thus teachers with more experience are
excluded from some of the discussion.
It was observed that generally probationary teachers who
transfer increase their salaries relative to what they would get in their
initial district. Male participants gain by about 1.4 percent in wage with a
move, while women gain half that amount. The mean salary gain decreases with
experience and the disparity is statistically insignificant for more
experienced teachers. Amongst transferees with less than ten years of
experience, the mean gain is 0.4 percent of the yearly salary at the time of
the move.
Another finding is that there’s a strong evidence that
teacher steadily prefer higher achieving, non-minority, non-low income students.
The clearest and most steady finding is that of achievement across gender and
experience groupings which illustrate that district mean achievement increases
by about .08 standard deviations or three percentile points on the state
distribution for the mean transferee.
Even though there is disparity across experience categories,
black and Hispanic compositions of district decreased by 2.5 and 5 percent,
respectively and the portion eligible for free or reduced lunch dropped by 6.6
percent. However, the mean changes cover significant heterogeneity, some may
appear to be systematically linked to origin and destination community types,
such as the strongest support for existence of compensating differentials comes
from teachers who move between urban and suburban districts.
The result shows that teachers who shift from large urban to
suburban schools experience average nominal salary losses of 0.65 percent,
however the mean adjusted salary rises by 1.4 percent. Likewise, the adjusted
salary[3]
increment is three times as large as the raw salary increase for teachers who
move across suburban districts.
Just like that of salary patterns there are striking changes
in the district mean student attributes for teachers who transfer from urban to
suburban districts, such as a 0.35 standard deviation or 14 percentile increase
in average achievement and 15 to 20 percent decline in racial and ethnic
concentrations. Also, among suburban districts the average achievement
increases by more than one tenth of a standard deviation and there is a decline
in the proportion of black, Hispanic and eligible for a subsidized lunch.
Another observation
shows little evidence those teachers who transfer from urban to suburban
districts experience changes that go beyond the differential between district
means thus implying that urban to suburban transferees seem to maintain the
same relative positions in the two districts. However those who move within
urban districts demonstrates a significant rise in mean achievement and a fall
in percent minority and percent eligible for subsidized lunch. Those who change
schools within urban districts seem to search for schools with fewer
academically and economically disadvantaged students. These patterns appear to
coincide with the commonly hypothesized placement of new teachers in the most
difficult teaching conditions within urban districts along with an ability to
switch locations as their experience progress (cf. Raymond, Fletcher, and Luque,
2001).
Note that while the typical movement increases average
student achievement and lessens economically disadvantaged and nonwhite by a
certain percent, there is a considerable differences among black, Hispanic and
white teachers. Black teachers tend to
transfer school with higher black enrollment shares than their previous school
not considering whether or not they switch districts. Alternatively, the mean
percentage change for Hispanic teachers is quite similar in direction and scale
to the changes experienced by teachers as a whole. Looking at the yearly
changes, Kain and Singleton (1996) demonstrate that the moving patterns accrue
and interact with new hiring to produce significant disparities in teacher
characteristics for black and white students, even across campuses within
individual districts. Also, the variation in the average test scores is much
smaller for black and Hispanic teachers.
Since pattern of teachers’ mobility may reflect various
teacher preferences, or may be derived from very diverse preferences for
factors related to race or ethnicity, it is difficult to unravel the probable
underlying mechanism for the outcomes of race/ethnic pattern in transferee. For
instance if there is widespread residential segregation and teachers favor to
work closer to their homes, blacks may rank predominantly black schools much
more highly than Hispanic and white colleagues, other things equal. However,
the teacher preference may not entirely drive the disparity by teacher
ethnicity.
Another apparent pattern shows that teachers belonging in
the top quartile of adjusted salaries are 3 percentages less likely to leave
the public schools and roughly 1 percentage point less likely to change
districts than teachers found in the bottom quartile schools.
The most striking disparities in school rates of movement
are linked to student achievement. Schools in the bottom achievement quartile
show a much higher teacher transition rates (with over 25 percent of teacher
leave each year) than those in the top quartile (with only less than 20 percent
exit).
The most prominent difference is in the likelihood of
entirely leaving public schools. These differences show that the lowest
achieving pupils are more likely to have teachers new to the school and to the
profession.
B. Transition
Regressions
Some of the observations also take into account the joint
impact of the different influences, such as setting the likelihood of leaving a
district as a function of the both teacher and district characteristics.
Specifically, those with higher experience levels have generally opted to stay
in their present district for a number of years regardless of district
qualities, which would tend to decrease the connection between change
probabilities and the involved district characteristics. Moreover, the
calculated link between transitions and percentages of black and Hispanic are
permitted to differ by teacher race and ethnicity.
Additional finding illustrates that higher salaries decrease
the probability of leaving a district and the impact tends to reduce with
experience.
The movement rate is also substantially linked to a number
of student attributes involving average achievement, percent of black and
Hispanic. Having higher mean student achievement considerably declines the
probability of moving or leaving Texas public schools at experience levels.
Non-black or Hispanic teachers are more prone to move given higher black and
Hispanic enrollees, even though the only major effects are connected to percent
black students for younger teachers. Conversely black and Hispanic teachers are
less likely to move given higher the minority enrollment shares.
Finally there is little or no evidence that the likelihood
of moving or leaving is systematically related to mean class size in any
conditions thus causing doubts about the commonly hypothesized influence of
smaller classes on teacher decision.
In terms of the salary variable, the estimated influence for
probationary teachers declines by approximately two thirds in size following
the inclusion of the fixed effects, while the estimates for teachers with six
to ten years of experience increases by more than a factor of three.
Again with the inclusion of fixed effects, the estimated
impacts of student mean test score on the probability of exiting a district
incline to fall a little by 15 to 20 percent. It seems that some other
characteristics of district are correlated with student mean achievement and
teacher movements.
An additional result indicates that the salary relative to
those staying is much more strongly linked to the probability of changing
districts than to the probability of leaving the Texas public schools – such
wage sensitivity for district transitions holds across all experience
categories. On the contrary, student
achievement seems to be a much more crucial determinant of the likelihood of
leaving schools. Whether teachers consider it harder or less gratifying,
teaching lower achieving pupils is a strong factor in them deciding to exit
Texas public schools, and the size of the influence holds across the full range
of teacher experience.
As illustrated earlier, teaching in racially concentrated
districts has a substantial impact on both the likelihood of exiting public
schools completely and on the likelihood of changing districts. For white
teachers, the influence on changing districts continues across the experience
distribution, while the influence on leaving public schools is observed more in
the earlier years. For black teachers, the feedbacks to varying concentrations
of black pupils are almost precisely the opposite that for whites in both sign
and size.
V. Conclusions
The outcomes substantiate the complexity that schools
serving academically disadvantaged students have in maintaining teachers,
mainly those early in their careers. There is also significant evidence that
non-black/non-Hispanic teachers systematically favor non-black/non-Hispanic
pupils, while the converse seems to be the case for black and Hispanic
teachers. These observations conform to the widely publicized teacher
scarcities affecting many of the nation’s inner city schools. The main concern
is the size of the added return required to compensate for the disadvantages
some schools must overcome so to compete for teachers. There was also little or
no evidence that the impact of salaries diverges with student attributes and
the likelihood that the effects of student attributes were non-linear, hence,
the salary coefficients calculated provide the best estimates of the
compensating differentials required to counterbalance the labor market
shortcomings of certain schools. The estimates imply that higher salaries
lessen the likelihood that teachers exit a district, and the scale of the
impact has a tendency to be rather analogous across the experience levels with
an interval from -0.1 to -0.3, indicating that, say, a ten percent salary
increase (about two to three thousand dollars per year depending upon the
district) would reduce moving out by at most three percent. Equivalently, a one
standard deviation reduction in school mean achievement raises the probability
of moving out by about one to two percent, and a ten percent rise in the
proportion of black or Hispanic pupils increases the likelihood that non-black,
non-Hispanic teachers leave further by one to two percent. Accordingly, schools
serving a high percentage of pupils that are either black or Hispanic who are
academically lacking might have to pay an extra 20, 30 or even 50 percent more
in wage than those schools serving a primarily white or Asian, and academically
advantaged students. However, evidently the availability of black or Hispanic
teachers may considerably decrease the expenses of hiring for such schools.
Significantly the pattern of multinomial logit estimates
implies that across the board wage hikes are unlikely to offset for the labor
market shortcomings. It seems as if salaries relative to other districts rather
than the absolute level of teacher salaries is the crucial factor, as wages
appear to have a bigger impact on the likelihood of changing districts rather
than leaving teaching in general.
Moreover,
enhancements in academic grounding in the means of better preschools or child
care services may possibly have the indirect benefit of making schools more
appealing to potential teachers. Learning more about the specific source of the
link between teacher labor supply and the precise student features would
present relevantly vital, policy information.
Lastly, this paper centers on the quantity of teacher
movements with little or no consideration to quality. The actual cost of
enhancing the quality of teaching, critically counts on the aspects of district
hiring, retention and other personnel policies. Ballou (1997) questions that
district systematically employ the best applicants available, indicating that
instructional quality could be mended at little or no cost in terms of higher
wage. The extent to which improved personnel policies would simply transfer
teachers amongst schools rather than increase the general mean level of
teaching is an enormously crucial issue that values additional research.
Source:
Eric A. Hanushek,
John F. Kain, and Steven G. Rivkin, “Why Public School Lose Teachers”, NBER Working Paper 8599, November 2001.
[1] Variants in decision begins with
pre-teaching phase (wherein it includes decision to undergo training for
teaching), then proceeds to the application and job matching procedure, if and
when hired at a certain school the career path is ascertained by the
continuation and retention decisions of both teachers and schools. (Hanushek,
Kain, and Rivkin (2001)).
[2] Such as Murnane and Olsen (1989,
1990) and Dolton and van der Klaauw (1995, 1999).
[3] Residuals from the regressions
of the data provide salary measures adjusted for differences in working
conditions,
amenities and local labor markets.
0 comments:
Post a Comment