I. Introduction
Two empirical puzzles motivated this study:
1. Controlling
for student’s background, student- and school-level data frequently show little
evidence of a relationship between pupil performance and school inputs.[1]
2. Metropolitan
districts with little opportunities for competition among public schools are
likely to have greater school inputs but may also yield worse student
performance.[2]
These results imply the presence of some school attributes,
such as teachers’ unions, that potentially raise inputs while at the same time
tends to reduce efficiency of each input.
Theory proposes two reasons as to why teachers demand a
union.
1. Teachers
are assumed to maximize similar objective function as that of parents, however
informational and market imperfection stirs teachers to desire different school
input levels.
2. (On
the other hand) teachers may desire a different objective function than that of
parents or administrators, most likely school policies which directly affect
them, like teacher wages, are assigned
more weights than polices that indirectly affect them.
Teachers’ unions may impact the educational production
function through at least three channels:
a. Unions
are expected to change (specifically, increase) the total budget that finances
school inputs
b. Unions
are expected to transfer any available budget among alternative inputs. Such
redistribution is efficiency enhancing if the union’s different objectives show
superior information, but efficiency reducing if the union is rent seeking.
c. Unions
may have an effect on each input’s productivity.
All of the possible unions’ impacts on schools are projected
to be magnified when the market for schooling is improperly competitive, since
monopoly rents will be accessible for rent-seeking unions and in the absence of
active choice among schools by parents and teachers, the market will convey
less information.
The empirical study on the impact of teachers’ union on
education production function experiences four primary obstacles:
1. Getting
data on unionization, student achievement, and demographics for a big
representative sample of schools across the time span of the era of
unionization.
2. Teachers’
unionization is difficult to determine and measure since teacher organizations,
for instance, National Education Association (NEA), carry out union function
like collective bargaining in some schools while maintaining a purely
profession associations in other schools.
3. Associating
the impacts of teachers’ unions to measure of competition among schools.
4. The
identification problem due to the difficulty of distinguishing between the
union’s impact on a school and the school attributes that likely lead to the
existence of a union. The unobservable attributes that induce unionization may
also affect the education production function.
After dealing with the obstacles, this study aims to build
consensus on teachers’ unions by broadening the best aspects of previous
literatures, such as largely extending the coverage of individual school
districts and utilizing deeper longitudinal methods. Another goal is to
determine whether or not (and why) school inputs matter. Lastly, it is to clarify
the mechanism that connects competition among schools to better pupil
achievement and school productivity.
II. Teachers’ Unions and the Education Production Function
Model is rent-seeking in which teachers’ unions favor
different inputs than parents do since the unions’ goal is not purely student
achievement maximization. Efficiency-enhancing is when unions favor different
inputs since they have the same essential goals as parents but have greater
information on student and input efficiency or internalize externalities in
education production that parents disregard.
Both models have varying implications for education production function.
Considering first the rent-seeking unions, by monopolizing
the services of current teachers and creating a political alliance in local
election, they may be able to raise the budget to a level unlike before. The
first likely impact of teachers’ (rent-seeking) unions is an increase in budget
that may be social welfare enhancing. Such increase would occur if budgets are
otherwise too low due to not having internalized positive externalities
connected with schools or imperfect capital markets for human capital
investments. The second potential impact of teachers’ unions is through
distribution of the budget among inputs. Third is through the productivity of
measured school inputs (which do not always include a level of teacher quality,
unlike actual school inputs). If unions lower teacher effort for any
combinations of measured school inputs, then the budget efficiently contracts
since the price of an actual unit of school input increases with unionization.
Taking to account efficiency-enhancing teachers’ unions that
maximizes student achievement and reflects superior information; it uses
monopoly or political strength (or both) in local elections to raise the school
budget to the point that is most favorable when positive externalities of
education are considered. Teachers may have more precise information on the
school inputs productivity or on the interactions of school inputs, thereby
seeing a different student achievement than what parents observe. Lastly, if a
union motivates teachers to act more professionally it may improve the effort
they share to any given set of measured school inputs. Enhanced teacher quality
raise actual school inputs and efficiently expand the budget constraint.
Since increased budgets are expected in either type of union
and there is difficulty in determining between inputs that merely benefit
teachers and inputs that parents underrate, the school inputs productivity is
the distinguishing factor that separates the behaviors mentioned above.
III. Teachers’ Unions in the United States
Teachers’ unions are by large a post 1960 occurrence in the
United States. Although not considered as “unions”, the teachers’ unionization
movement started when the teachers’ professional organizations of few central
city districts began to use union strategy, such as strikes.
Presently, schools are unionized on a district-by-district
basis and most teachers’ unions are associated with the American Federation of
Teachers (AFT) or National Education Association (NEA). Almost all public
school districts have teachers’ organizations, however many schools with
organizations still remain nonunionized.
A. Measuring Unionization
This study will focus on the strict definition of
unionization which comprises of collective bargaining, contractual agreement,
and 50 percent union membership.
Generally AFT is perceived to act more aggressively than the
NEA. However many of the disparities between the two diminish using a refined
definition of unionization. This is due to the fact that AFT associates have a
greater tendency to function as unions while a considerable portion of NEA
associates fundamentally remain professional affiliations. Furthermore, strict
definition provides greater consistency of union behavior across union
affiliation.
B. Laws Facilitating Teachers’ Unionization
One of the most significant aspects of the history of
teachers’ unionization is the shift in the legal environment for public sector
unions following 1960. Before, collective bargaining by teacher was explicitly
illegal in various states and possibly implicitly forbidden in many
others. However, between 1960 and 1990,
states extended progressively ample collective bargaining rights to teachers’
unions.
The study uses passages of three types of laws as
instruments for unionizations, which are explicitly extending the right to
participate in collective bargaining, and allowing teachers’ unions to have
agency shops (this exists if there is collection of dues from all teachers in
the bargaining unit, whether or not they are union members) and union shops
(this exists when the school district cannot hire teachers who are not union
members). Laws allowing agency and union shops assist in assertive collective
bargaining, since they significantly weaken the position of teachers in a
district who are against the union.
Only the passages of laws that allow union activities are
used as instruments, instead of passage of laws that prohibit such activities,
such as ban or strikes. This is because “permitting” laws seem to lead, rather
than delay, bursts of union activity.
IV. Data
The 1972, 1982, and 1992 Censuses of Governments are matched
to generate a panel on expenditures, teacher employment, teacher wage, and
student enrollment for each public school district in the United States over
the three decades.
To expand the panel data on unionization to the 1960s, the
1966 data on individual school districts’ negotiation agreements from
Negotiation Agreement Provisions (NEA 1967) and 1963 statistics on unionization
from Perry and Wildman (1966) were used.
For the demographic data and a measure of student
performance, the school district tabulations of the 1970, 1980 and 1990 Censuses
of Population and Housing were matched to one another and then to the Census of
Governments. The demographics data include populations of black, Hispanic,
urban, adults having at least 12 years of schooling, adults having at least 16
years of schooling, and population in poverty, it also includes total K-12
enrollment, private K-12 enrollment, black K-12 enrollment, median household
earnings, median gross monthly rent, the unemployment rate, and the percentages
of population aged 16 to 19 and aged 18 to 19.
The 585 districts that could not be successfully matched
were omitted from the analysis and the study was left with 10,509 school
districts, which is roughly 95 percent of the total in the United States in the
1990.
The final primary source of data is the NBER Public Sector
Collective Bargaining Law Data Set (Valletta and Freeman, 1988), which sums up
state laws governing teachers’ unionization for each year after 1954.
V. Empirical Strategy
Various district attributes having an impact on both unionization
and school trend somewhat steadily, for instance, housing stock in a district
may slowly decay or job opportunities may eventually become more oriented
toward professionals and managers. Acceleration or deceleration in the time
trends of school inputs or pupil performance associated with the discrete event
of unionization is what motivates the differences-in-differences results. Differences-in-differences results are suitable
to the analysis, since unionization occurs through a teacher vote, the unionization
event is discrete even if its determinants trend continuously.
Meanwhile the possible solution for the potential remaining
identification problem due to the likely correlation between shocks to
unobserved district-time-specific attributes and the district change in
unionization is the use of instrumental variables that are uncorrelated with
shocks in school districts but cause unionization to change discretely when the
variables discretely alter.
Basing on Heckman and MaCurdy (1985) and Angrist (1991) the
study uses a linear probability model for the first stage of the instrumental
variables estimation, which then allows for the use of Lagrange Multiplier test
of identifying restrictions.
VI. Results
Unless stated otherwise, all estimates mentioned in the
study are asymptotically significantly different from zero at a 0.05 or lower
level.
A. Effects on per-Pupil Spending and School Inputs
Since the results are consistent with the expectations,
other per-student spending will be first discussed. Per-student spending is
considerably greater in school districts having bigger populations, higher
median household earnings, more educated populations, larger enrollment shares
in private schools, and smaller public school enrollments.
The results indicated that per-pupil spending increases by
some percent when teachers unionize. It grew rapidly in schools that unionized
in the period between 1972 and 1982 as wells as between 1982 and 1992.
Furthermore, the fastest growth in per-student spending is in states that are
presently passing laws assisting unionization.
Moreover, teachers’ unionization leads to roughly 5 percent
increase in teacher wages. Meanwhile, unionizing teachers decrease the
pupil-teacher ratio or increase teacher and classroom inputs per pupil; such
result is not only statistically important, but economically significant as
well.[3]
Overall, teachers’ unionization seems to raise school
budgets distributed mostly to teacher wages and the pupil-teacher ratio. The
increases are coherent with either an efficiency-enhancing or rent-seeking
union models.
B. Effects on the Dropout Rate
The observations show that in 1970 and 1980, higher teacher
wages are connected to lower dropout rates or improved pupil performance. For
the 1990 results, although per-student spending has statistically insignificant
impact on the dropout rate, the dropout rate is improved by a small amount.
Meanwhile, whether or not unionization is considered,
teacher wages seems to have no impact on pupil performance. However, what’s
more interesting is that the result suggesting that inputs’ effectiveness is
lower in unionized schools, although the differences are insignificantly
different from zero. Furthermore, unionization is predicted to have a direct
worsening impact on student achievement (higher dropout rate is expected).
On the other side, allowing unions to both directly impact
the dropout rate and alter the input productivity, it seems that inputs are
effective in nonunion schools while unionized schools reflects the traditional
pattern of inefficient inputs. In nonunion schools, a decrease in the
pupil-teacher ratio as well as an increase in teacher wages reduces the dropout
rate. Meanwhile, unionization main impact is to increase (worsen) dropout rate.
In summary, the findings show that teachers’ unions are
successful in increasing school budgets and school inputs, but have a generally
negative impact on pupil performance (rising dropout rate). According to the
results, much of the negative effect is brought about by the decreased school
inputs’ productivity.
It is remarkable that unionization is connected with both
increased school inputs and poor student performance. These results imply that
teachers’ unions, at the very least, serve partially a rent-seeking purpose.
VII. Teachers’ Unions and Schools’ Market Power
Both union models suggest that teachers’ unions will be able
to exercise greater influence when schools experience less competition – when
school districts residents are less likely to react to increased or reallocated
school spending, via a transfer to another district. If the movements across
school district are costless, then teachers‘ unions would face difficulty
either in extracting rent or increasing school spending to the socially optimal
degree if it were above the privately optimal level.
Considering the rent-seeking model, it seems that unions
have a stronger impact in areas having less competition among schools (more
concentrated enrollment). In addition parents that can easily move may be able
to constrain teachers’ unions to go for lower budget hikes, maintain greater
levels of effort, and add lesser administrative burdens.
VIII. Conclusions
The study finds teachers’ unions to be mainly rent-seeking,
increasing school funds and inputs but decreasing pupil performance by lowering
the inputs productivity. Teachers’
unions may be one of the reasons, educational production functions estimated
before 1960 showed the importance of school quality. Also, teacher unions’ may
be a primary means through which a lack of competition among public school
transforms into more ample school inputs and worse student achievement.
Source:
Caroline M. Hoxby, "How
Teachers’ Unions Affect Education Production",
Quarterly Journal of Economics, Vol.
111, No. 3, August 1996, pp. 671-718.
0 comments:
Post a Comment