WORKING PAPER
DEPARTMENT
OF ECONOMICS
The Palestinian Labor IVIarltet Between
Gulf
ttie
War and Autonomy
Joshua Angrist
No.
98-05
May
1998
MASSACHUSEHS
INSTITUTE OF
TECHNOLOGY
50 MEMORIAL DRIVE
CAMBRIDGE, MASS. 02142
Digitized by the Internet Archive
in
2011 with funding from
IVIIT
Libraries
http://www.archive.org/details/palestinianlabor9805angr
c:\projects\kschool\kdraft3.wpdMay 18. 1998
THE PALESTINIAN LABOR MARKET BETWEEN THE GULF WAR AND AUTONOMY'
by.
Joshua D. Angrist
MIT
"This study
is
part of the Project
and
on Palestinian Refugees
NBER
at
Harvard University.
I
tiiank the Project for financial
support and the project organizers, George Borjas and Dani Rodrik, for their comments. Thanks also go to Jon Guryan and
Samer Haj Yehia
Archive.
for research assistance.
The data used here
are available
from the Hebrew University Social Science Data
Abstract
were an eventful time in the Middle East. Key events include the Gulf War, the Madrid
peace conference, and the introduction of Palestinian autonomy in all of the Gaza Strip and in most West
Bank population centers. This period also provides the last chance to look at economic conditions in the
territories through the lens of the long-running Israeli survey program, which ended with the transfer of
The
early 1990s
civil authority to the Palestinians.
territories
I
use the Israeli survey data to describe labor market conditions in the
on the eve of autonomy. Special attention
is
given to the impact of changing employment
opportunities for Palestinians in Israel and to the situation of Palestinian refugees. Policy implications of
the findings are discussed.
I.
Introduction
The 1991 Gulf War marked a low
territories
point in Palestinian political and economic fortunes both in the
and abroad. By 1994, however, the Gaza Strip and the West Bank town of Jericho had been
turned over to the newly created Palestinian Authority (PA). In 1995, Palestinian cities and towns in the
West Bank were
also gradually brought under
PA is
relationship between Israel and the
issues
is
how
still
PA control,
although the final economic and political
being negotiated.
One
.
of the most important unresolved
the flow of Palestinian labor to Israel will be regulated or even if it will continue at
The
all.
years 1991-95 were a period of repeated sharp restrictions on Palestinian access to the Israeli labor market,
and, in
These
some
cases,
restrictions
on the flow of goods
were
in
to
and from the
territories (Israel
Employment
Service, 1996).
response to events such as the Hebron massacre and terrorist attacks in
Israel,
sources of variation in economic contacts that need not characterize the Israeli-Palestinian relationship
indefinitely. In the
mean
time,
it is
possible to use the experience of these years to better understand
how
the Palestinian and Israeli labor markets interact.
In three earlier papers (Angrist 1992, 1995, 1996),
Palestinian
for
wage and employment
working
the Israel
in Israel fell
wage premium
working
from roughly 18 percent
rose steeply.
By
in
in Israel increased similarly.
Israeli elasticity
demand
of
model
shows
in
wage premium
that the daily
1981 to zero in 1984. Beginning in 1986, however,
These changes
wage
differential.
fairly inelastic Israeli
for Palestinian labor
The monthly wage premium
parallel the pattern of Palestinian absences
demand curve
specification, but the best-fitting
demand
from
for Palestinian labor.
was estimated by using supply shocks
with security curfews as instrumental variables. FV estimates of
to the details of
number of important changes
1989, daily wages paid to Palestinians working in Israel were 37
work, and are consistent with movements along a
The
described a
patterns since 1981. Angrist (1996)
percent higher than local wages, nearly doubling the 1987
for
I
that originate
equations are somewhat sensitive
models consistently suggest
that the
observed
covariance between Palestinian wages and days worked can be rationalized by a short-run Israeli elasticity
1
of
demand
for Palestinian labor ranging
These
results suggest that labor
possibility that Israeli
demand
between -
1
and -2.
market policies
for Palestinian labor
in the territories
may be
inelastic
should be analyzed
enough
to considerably
earnings loss from reduced access to the Israeli labor market, at least for a while.
On
in light
of the
dampen
the
the other hand,
higher Palestinian wages and an uncertain supply create a strong incentive for Israeli employers to develop
alternative labor resources. Israel's increasing use of non-Palestinian guest-workers to
by Palestinians
now
is
includes as
note that the
fill
an attempt to develop such resources. Recent reports suggest that the
many
as 250,000 guest workers
number of Palestinians working
from Europe and Asia.^ To put
in Israel
positions vacated
Israeli labor force
this figure in perspective,
was probably never more than 120,000
at
any one
time; official statistics indicate between lOO-1 10 thousand workers (Angrist, 1996). Another relevant
piece of background information
is
that in 1995, the entire
only about 1.2 million (Israel Central Bureau of
In light of these recent changes,
it
male civilian labor force
now
seems natural to ask whether
proportional to changes in days worked?
workers on Palestinian wages? This
Another important change
is
numbered
Statistics, 1996).
consequences of negative demand shocks are moderated by
earnings
in Israel
it is still
inelastic Israeli
More
generally,
true that the earnings
demand. Or are changes
what has been the
in
effect of guest
the first set of issues discussed here.
in the Palestinian labor
market that
I
noted in earlier work was a
dramatic reduction in the wage premia paid to more educated workers between 1981 and 1987. Angrist
(1995) provides evidence which suggest that large
shifts in the
supply of educated workers combined with
restricted capital investment explain this decline in returns to schooling. Angrist
different types of workers
schooling coefficients
^This figure
is
were employed
may
reflect
from a recent
in Israel at different times,
(1992) also shows that
which suggests
that these
worker heterogeneity. But the panel data estimates reported
article in the
Hebrew
daily Maariv (1996).
A possible alternative rough
changes in
in
Angrist
indicator of the
flow of guest workers might be had from the number of "temporary residents" (a visa classification) entering Israel. This figure
reached 85,000 in 1996, up from 44,000 in 1995, and 18,000 in 1993 (Israel Central Bureau of Statistics, 1996; Table 4.1).
(1995) show that sharp declines in the returns to schooling are apparent even after allowing for individual
The second
heterogeneity.
more recent
human
issue
look at here
I
is
whether the returns to schooling have recovered
period, perhaps in response to the expectation of increased opportunities for the utilization of
capital
under the autonomy regime.
on the returns
Finally, in addition to updating earlier results
shocks on wages, the third issue discussed in
who Hve
in the
in the
West Bank and Gaza
Strip.
whose
resided in Israel in 1948 and people
this
paper
is
the
economic condition of Palestinian refugees
The TLFS micro data
sets for
1992-95 identify people who
father resided in Israel in 1948.
the children of refugees. Information on refugee status
is
refugee/non-refugee wage and employment gap since
992.
1
and the impact of supply
to schooling
These people are refugees and
used to estimate and characterize the
n. Data and descriptive statistics
The TLFS
In 1967, the Israel Central
Bank and Gaza
territories.
Strip
Bureau of
which was used
The sampling frame was
Because of the
TLFS
rotation
as a
Statistics
(CBS) conducted
a population
Census
in the
West
sampling frame for quarterly labor force surveys in the
periodically updated to reflect migration and
new
construction.
group design, each surveyed household was interviewed four times.
TLFS
interviews were conducted by local Palestinian enumerators employed by the Israeli Civil Administration
in the territories.
The
last
the last interviews in the
survey interviews in the Gaza Strip were conducted in the
West Bank were conducted
in the third quarter
sizes are reported in
Table
1
for selected years through
descriptive statistics and
1989 and every year from 1991-1995.
1994 when the Gaza survey ceased.
Roughly 30-39 thousand interviews were conducted each year
until
The
from a high of 27 percent
fraction of the
sample residing
in the
Gaza
strip fluctuates
quarter of 1994;
of 1995.
The sample used here contains observations on men aged 1 8-64. Basic
sample
first
in
1981-84
to a
low of 22 percent
in
1988.
The mean age of men
Approximately half of the men
in the
in the
sample
is
stable at about 33.
sample are wage laborers, the remainder being mostly
employed (including work on a family farm or business), unemployed, or
constitute 2/3 of the labor force, with participation rates ranging
men
in the labor force
who were employed
This ranges from 37-39 percent until 1993,
shows
that
skilled
and unskilled jobs, and
workers from the
territories
in Israel or
when
who
are
it
Jerusalem
sample rose from 7.7 years
in
shown
is
in Israel
and agriculture
Kleiman (1992) and Semyonov and Lewin-Epstein (1987). Table
level in the
to
Wage
laborers
82 percent. The fraction of
in the last
column of Table
1.
drops sharply, to 32 percent. The lower half of Table
employed
in the construction
from 72
in school.
self-
1981 to 8.7 years
and Jerusalem are concentrated
industries.
1
in 1991,
also
For more on
shows
1
in semi-
this point, see
that the average schooling
and then continued
to rise to
over 9
years in 1993.
For the purposes of this paper,
in the State
of
Israel.
I
define a first-generation refugee to be
someone who
1
indicate that in
Bank and Gaza
Strip,
1992 and 1993, the
when complete
last years
about 5 percent of the
men
in the extract
percent were second-generation refugees. Weighting this by
proportion of second generation refugees to 30.5 percent.
Gaza
Strip than in the
1
948
A second-generation refugee is someone whose father lived in the State of Israel in
1948. These definitions correspond to the information collected in the TLFS. Estimates not
Table
lived in
first
generation refugees.
in
surveys were done in both the West
were first-generation refugees and about 23
CBS
sampling weights increases the
The proportion of refugees
West Bank. About 54 percent of Gazans
refugees and 8 percent are
shown
is
much
higher in the
are estimated to be second generation
The corresponding
figures for the
West bank
are 16
and 3 percent, respectively.
The usual
TLFS.
types of sampling and non-sampling errors in labor force surveys are also present in the
In addition, data collection after
territories that
1988 became more
were part of the Palestinian uprising or
difficult
Intifada.
because of civil disturbances
Because
TLFS
in the
data on the territories were
collected
by employees of the
Israeli civil administration, these
may be
data
less accurate
than data from
comparable labor force surveys carried out elsewhere. Moreover, although there are many sensible
patterns in the
TLFS
data on labor force status
(e.g., seasonality;
wages are
see Angrist, 1996), data on
often poor even in the best circumstances.
As
a check on the
TLFS
data.
Table 2 reports estimates of average wages from the
first
two waves
of the Palestinian Labor Force Survey (PLFS), which was conducted in the third quarter of 1995 and the
PLFS wages
second quarter of 1996.' Average
employed
49 shekels
locally (including Jerusalem),
for Palestinians
employed
in
1995 were 35 shekels for residents of the West Bank
27 shekels for residents of the Gaza Strip employed
in Israel (excluding Jerusalem). Estimates
second quarter of 1995 are about 31 for Palestinians employed
The most
shekels for Palestinians employed in Israel.
quarter of 1994;
when
the average
wage
is
in the
recent Israeli
from the
and
locally,
TLFS
in the
West Bank or Jerusalem and 51
wage data
estimated to have been 26 shekels.
for
Gaza
So
the
is
from the
PLFS and
first
the
TLFS
data seem to generate estimates that are in the same ball park.
There were no
at all locations
following the
Israeli
surveys in the territories in 1996.
between 1995 and 1996. This drop
November 1995 Rabin
is
in
wages
probably due to the disruption of economic activity
assassination, the early
subsequent closure and quarantine of the
The PLFS data show a big drop
1996 bus bombings
territories until the elections in
in Israel,
and the
May.
Graphical analysis
After an
initial
decline during the Palestinian uprising, the proportion of Palestinians working in
Israel fell sharply in 1993, largely as a
Israel.
The change
in probability of
consequence of an increasingly
working
in Israel can
be seen
'These estimates were derived from Palestinian Central Bureau of
reports suggests an effort has been
sometimes reported on a different
made
basis.
to maintain
in
Figure
Statistics
some comparability with
restrictive access policy
1
(1996a,
,
enforced by
separately for residents of
b).
The
introduction to these
the Israeli surveys, although the
PLFS
results are
the
West Bank and Gaza
employment/population
Gaza
Recent changes
Strip.
ratios.
These also
fell
Strip to levels not substantially different
Bank employment
unemployment
rates also
rates,
seem
to
employment
in
in the
West
On
have recovered by the end of 1995.
data
show
in
Figure
from those prevailing during the preceding
which are affected by movements
TLFS unemployment
can be seen
2,
which
plots
sharply in early 1993, but recovered very quickly in the
in
and out of the labor
early 1993 than they had been previously, especially in the
most recent
rates
rates
the other hand,
force, stayed higher after
West Bank. This can be seen
around 4 percent
in the
West
years.
Gaza
Strip
in Figure 3.
The
and around 7 percent
Bank."*
Figure 4 plots average days worked by region of residence.' Like the employment data, this figure
also
shows a sharp drop
in
1993 that was almost as large as the Gulf War drop. Subsequently, average
days worked seem to have returned to the 18-19 day level prevalent during the Intifada period. Figure
which
plots average days
worked by work
shows
location,
that reductions in average
5,
days worked
continued to be most severe for those employed in Israel or Jerusalem. In contrast, average days worked
appear to have been on a gradual upward trend for those employed
In recent years, the pattern of fluctuations in
days worked, although
it is
somewhat
quarter changes in average days
I
have also confirmed
this
6,
which
worked occurred
at
in
plots
by comparing quarter-to-
worked and average monthly earnings. This comparison shows
recent years at least, changes in days
that
5.
monthly wages largely mirrors the fluctuations
This can be seen by comparing Figure
less volatile.
monthly wages by work location, with Figure
locally.
roughly constant real (daily) wage
changes in monthly wages are roughly parallel to changes in days worked
is
that in
rates.
The
fact
important because this
implies that fluctuations in days worked are not associated with changes in wages.
"•The
TLFS
labor force status recode from which the
similar to those used in the
'Figures that plot information by region of residence
work
location.
unemployment
rate is constructed uses "actual status" definitions
CPS.
Figures that plot information by work location
regardless of region of residence.
show averages for all workers who live in the region, regardless of
show averages for all workers employed at that location,
The
fell
wages of Palestinians increased steeply from 1985
to 1989, fell
from 1989
to 1991,
and
again in 1993. This can be seen in Figure 7, which plots average daily wages by work location. This
figure also
the
real
first
shows
that the
wages paid
employed
to Palestinians
half of the 1990s and declining in Gaza.
in the
But wages paid
West Bank were more or
to Palestinians
working
less flat in
in Israel
or
Jerusalem exhibit an increasing trend from 1991-95. Only in 1995 do wages paid to Palestinians working
in Israel turn
down
the increase in
again, for the first time since 1991 (except for first quarter 1993). Note, however, that
wages paid
to Palestinians
working
in
Jerusalem
is
somewhat more gradual
after 1992.
Interpretation
The
years 1992-95 were
migrant labor to
under curfew, as
on the
Israel.
marked by extended and repeated
In 1992, 1993,
in the Intifada years.
territories instead
interruptions in the flow of Palestinian
and 1994, there were periods when large numbers of people were
It is
also important to note that Israel increasingly
of curfews, so that curfews were not the only or even the most important source of
supply shocks. There were
total closures
of the territories in April 1993, February 1994, April 1994,
October 1994, January 1995, July 1995, August 1995, and September 1995
1996).
The
imposed closures
closures lasted from one
Palestinians were allowed to
work
week to up
to
(Israel
Employment
Service,
two months, during which time few and sometimes no
in Israel or Jerusalem.
During the Intifada years (roughly 1988-91), exogenous supply shocks were associated with an
increase in daily wages. This
wage response
partly offset the decline in earnings
so that earnings were less variable than wages.
wages
in the
more recent period suggests
workers as they used to be.
It
The
that Israeli
fact that earnings appears to
from fewer days worked,
be about as
volatile as
employers are no longer as dependent on Palestinian
should be noted, however, that the period discussed here largely predates the
massive inflow of guest workers to
Israel,
which probably did not begin
apparent insensitivity of wages to fluctuations in days worked
in earnest until 1995.
may therefore
The
reflect a structural
and
perhaps long-lasting change in the Palestinian-Israeli employment relationship.
Real wages earned by Palestinians working in
95 period, only beginning
to decline in 1995.
probably reflects post Guif
wages
in
The growth
War economic growth
1995 may indeed be due
Israel increased substantially
wages paid
in
in Israel.
On
to competition with the guest
over most of the 1992-
to Palestinians
the other hand, the
working
dechne
in Israel
in Palestinian
workers since the IsraeU economy
continued to grow in that year. In contrast with the decline in Palestinian wages between 1994 and 1995,
enjoyed real wage growth of about 2 percent over
Israeli citizens
Statistics,
1996; Table 12.24).
It is
still
period (Israel Central Bureau of
and 1993, when the wages paid
also worth noting that in 1992
Palestinians working in Israel and Jerusalem were
essentially
this
wages of Israeli
increasing, the
citizens
to
were
flat.
in. Regression estimates
and changes
in schooling coefficients
This section uses regression models to show that the pattern in relative wages seen in the figures
is
apparent after controlling for basic demographic characteristics, and to look at recent changes in the returns
to schooling.
Table 3 reports the results of estimating the following equation (computed by weighted
squares using
CBS
sampling weights):
log(wJ =
where
is
6^, is
in-Israel effect in year
term
Xi(aot
-i-
+
Xi,ao,
is
a
Xi,^(x,t is
more
t, aj^
is
a
+
^
aiePc
dummy
+ Zg
+
bigYg.
f.Qt
variable indicating if
i
t, fj
is
a
dummy
variable indicating if
i
works
Cj
is
a
dummy
variable indicating if
i
lives in
variable indicating
work
in
+ C-X +
variable that indicates if
dummy
is
i
in schooling
Jerusalem and
i);
],\\i+ u^,
is in
are:
(1)
age group c and P„
group g and
in Israel in year
t
and
Ygt is
6t is
an urban area and X^
is
is,
8
a work-
a city effect
The
Xit=ageit-years of
25-34, 35^44, 45-54, 55-64. The schooling groups indexed by g
years.
is
a
a work-in-Jerusalem effect.^
a time-varying quadratic function of potential experience (that
'The age groups indexed by c
years, 16 or
Xi,'ai,
a
t,
dummy
+
bjg is
t,
schooling effect in year
t, Jj
Sq,
a quarter effect in year
an age effect in year
in year
least
are:
13-15
schooling|,-6).
Column
Table 3 reports estimates of the daily wage premium for working
(1) of
of residence. The premium
fall
sharply from about 20 percent in 1981 to -.03 for residents of
Bank
1984, and to .056 for residents of the West
.25 for
1994
Gaza
at
residents in
1987 and
both work locations. In
in Israel
to .20 for
fact,
in 1984.
The
West Bank. The
the 1992-1995
Israel
premium
by region
Gaza
in
rose sharply after that, to
increase continued every year from 1991 to
wage premium
for
work
in Israel
was
larger than
it
has ever been. The 1995 premium was about 50-60 percent for Palestinians from both locations.
The schooling
locations,
coefficients
and continued
shown
to decline in the
in
Table 3 decline monotonically
West Bank
thereafter.
But beginning
schooling began to increase substantially for residents of the Gaza Strip.
also increased in 1995 for residents of the
investment and the increase in
until
The
West Bank. This turnaround may
1991 for residents
in 1992, the returns to
returns to a college degree
reflect
1
993
in
an inflow of
and administrative jobs generated by the autonomy process.
civil service
PA administrative activities were concentrated in the Gaza Strip during this period,
in the returns to schooling in
both
at
Gaza pre-dates
the official beginning of
although the increase
autonomy
there.'
Finally, note that in contrast with the specification reported here in Table 3, potential experience
was not included
in the
models reported by Angrist (1995) because exact years of schooling are poorly
measured. But the specification used here
otherwise identical to the ones
is
I
used
potential experience necessarily raises the estimated returns to schooling but has
period comparisons. This can be seen in the second panel
equation (1)
when
the potential experience quadratic
rV. Changes in Refugee/non-Refugee
wage
'The Declaration of Principles was signed
The preparatory
transfer of
power
of Table
3,
little
effect
Including
on cross-
which reports estimates of
dropped (and age dummies are retained).
differentials
Table 4 reports estimates of differences
1994.
is
B
earlier.
in
in other areas
in daily
wages between refugees (men who residing
September 1993 and the Gaza/Jericho withdrawal was completed
was completed in December 1994.
in the
in
May
State of Israel in 1948), the children of refugees
and other Palestinians. Refugee
effects)
fathers resided in the State of Israel in 1948),
models without controlling for covariates (other than period
were computed by estimating:
log(W;,)
where
effects in
(men whose
R,i, is
a
dummy
=
6^,
+
+
R,i,(t)„
for refugees
and
R2i,(|)2,
Rj,, is
a
+
(2)
Ui„
dummy
corresponding refugee effects. Estimates with covariates were produced by adding
equation
1992 and 1993, refugees earned
The 1992 refugee wage gap was about
the
wage gap was about
Strip,
less than
8.5 percent in
4.5 percent in
model (including controls for work
Gaza
Gaza and
Unlike
in the
West Bank. The
first
covariates.
Gaza and
15.7 percent in the
15.3 percent in the
+
(j)2,
are the
R2i,4'2i to
gap
in
Bank and Gaza
Strip.
West Bank. The 1993
West Bank. Adding
covariates to
both years to less than 2 percent in the
Gaza
Strip in
In the
1
covariates-adjusted gap between refugee and non-
percent.
Gaza
generation refugees residing in the
covariates are not taken into account,
in the
non-refugees in both the West
location) reduces the
refugee wages in 1992 and 1993 was around 8-1
men
RntCJ),,
and
an amount less than the standard error of the estimates. Including covariates also reduces, but
does not eliminate, the gap
when
(J),,
(1).
In
refugee
and
for the children of refugees,
1992 and 1993. This appears
West Bank, however,
same when estimated with or without
other Palestinians in the
men whose
West Bank
From 1992-1995,
the
is
the
wage gap
covariates.
Strip,
fathers
to
who
earned less than other Palestinians
were refugees earned no
less than other
be the case with and without controlling for
for first
and second generation refugees
is
about the
The wage gap between second generation refugees and
also largely unaffected
raw refugee wage gap
in the
by control for covariates.
West Bank
fell
from -.16
to -.05. This
is
apparently due in part to changes in personal characteristics. Controlling for covariates, the gap shrinks
much
less.
years.
So
The 1995 gap
for second generation refugees in the
the overall pattern suggest that refugee
West Bank
wage gaps have been
10
is
also smaller than in earlier
closing.
The data
for residents of
the
Gaza
Strip actually
show an
than other Palestinians in that year. But this
1994 survey
in
more
interesting turnaround in 1994, with refugees earning substantially
Gaza was conducted
may be an anomalous
result
which
reflects the fact that the
Also, the 1994 interviews in
in the first quarter only.
Gaza were
the
before the Israeli survey program ended there, and this quarter marks the transition from Israeli to
last
Palestinian authority.
Table 5 reports estimates for employment similar to those reported in Table 4 for wages. The
employment models with covariates do not control
story here
similar in
is
many
refugees residing in both the
Palestinians.
much
of the
for
work
respects to that for wages.
West Bank and Gaza
location since
When
Strip are
some men do not work. The
comparisons are not adjusted for covariates,
much
less likely to
be working than other
But adjusting for education, age, urban residence, and potential experience accounts
first
generation refugee employment gap, especially in the
covariates also reduces the refugee
employment gap
for residents of the
Second-generation refugees in both the West
gap remains
substantial.
employment
rates than other Palestinians, with
difference between the results for wages and
Gaza
Strip.
Controlling for
West Bank, although
Bank and Gaza
and without adjusting for demographic
employment
have been shrinking, the employment gap does not appear
is that,
to
unlike the
wage
for
Strip
the adjusted
have lower
characteristics.
One
gap, which appears to
have changed systematically
in the
1992-95
period.
V.
Summary and
policy discussion
Summary
Three important characteristics of the Palestinian labor market are the wage premium for working
in Israel, the
economic returns
to schooling,
and the refugee wage and employment gaps. For most of the
period since 1991, the wages of Palestinians employed in Israel were rising, while local wages were
The
(log)
premium
for
working
in Israel in
1992-95 reached 50 or 60 percent as the
11
Israeli
flat.
economic
improved while
situation
roughly proportional to changes
change
begun
wage
in
to
fall.
served to
in
Moreover,
rates.
Recent changes
that in the territories did not.
monthly earnings have been
days worked, suggesting that labor supply shocks did not induce a
wages of Palestinians employed
in 1995, the
in Israel
appear to have
were associated with increasing wages, which
In the past, negative labor supply shocks
dampen
in
the consequences of these shocks for earnings and offered evidence of Palestinian
market power in the market for construction and agricultural workers. Overall, the new findings suggest
that the
dependence of
Israeli
now been
employers on Palestinian labor has
reduced, perhaps because of
The question of whether
guest workers, sectoral shifts, or substitution in favor of capital or local labor.
new
situation
is
permanent or transitory remains open.
Other findings suggest that the returns to schooling
sustained decline, although there
pre-Intifada levels.
On
is still
way to go
a long
Palestinians have been concentrated in the
West Bank, once
The
(1996),
Strip.
I
last set
Gaza
of relationships looked
the other hand, there are
at
Gaza
Strip
West Bank
new employment
Perhaps a change
Strip.
began
before the returns on
the Palestinian political infrastructure there
do not find dramatic differences
On
in the
the other hand, educated workers in the
for their schooling. Together, these findings suggest that
the
is
more
to recover after a
human
still
capital reach their
receive almost no
in opportunities will also
percent.
to
established.
market conditions by refugee
some important
however, the adjusted refugee wage gap
come
here are refugee/non-refugee contrasts. Like Pedersen
in labor
than other Palestinians, although the refugee
premium
opportunities for skilled
status, at least in the
differences, especially for residents of the
wage gap appears
in the
to
have been declining. As
West Bank was about 7
The adjusted employment gap
in the
Gaza
12
Strip in
percent, while the
Gaza
West Bank.
Refugees and the children of refugees in the West Bank have lower wages and lower employment
was 4-6
the
late as
rates
1995,
employment gap
1994 was about 7 percent.
Implications for access policy
Structural
Israeli labor
economic changes are almost by definition
long-lasting.
If a structural
market has occurred, then the Palestinians should probably not count on returning
arrangement that allows large numbers of migrant workers to commute to work each day
other hand,
change
if
the
new
relationship
between Palestinians and
by guest workers, prospects for a
Palestinians have been replaced
makers and many commentators
(see, e.g,
Maariv,
1
homes
in the
an
in Israel.
On
the
employers has developed because the
reversal
seem
better.
Israeli policy
996) appear to be uncomfortable with the notion of a
permanent guest worker population; Palestinian workers,
the most part return to their
Israeli
to
in the
at least,
West Bank and Gaza
do not aspire
to Israeli citizenship
Strip every evening.
This
is
and for
a remarkable
and perhaps even unique feature of Palestinian labor migration.
The question remains whether either
number of Palestinians working
Palestinians
employed
in Israel
in Israel as a desirable
must surely be
employers benefit from low cost employees
Moreover,
little
Israelis or Palestinians
outcome. The large wage premium
attractive to workers.
who
are, in
premium apparently need be paid
for
should look to an increase in the
many cases,
more educated
And even
at these
now enjoyed by
wages
Israeli
highly skilled and productive.
Palestinian workers.
On
the other
hand, employment of workers from the territories involves substantial transaction costs on both sides. For
the workers,
employment
in Israel involves long
commutes even though
distances are short (see, e.g.,
Angrist, 1996), as well as arbitrary search and seizure, risk of personal attack, and lockouts in response to
attacks or as a
consequence of local quarantines and curfews. Employers pay a price
in terms of
uncertainty and employee absences, sabotage, and the risk of personal attack.
In addition to these considerations, data
from other countries show
economic integration has increased, migrants from developing countries
analogy appropriate in the
Israeli case) constitute
In a recent study of labor in the world
that while international
to industrialized countries (an
only a modest proportion of the world stock of migrants.
economy. Bloom and Brender (1993)
13
note:
"World economic integration has proceeded even though international labor mobility has
not increased in recent decades
.
.
.
Developed country governments generally do not
encourage international migration and often inhibit
legal or financial barriers
.
.
.
limitations
through the imposition of various
it
on the physical mobility of labor, however, have
not prevented the emergence of a global market for labor services as services can be
exchanged
in a variety of
ways
that
do not require the physical proximity of buyers and
sellers."
This suggests that there are a variety of economic and social factors that limit the benefits from migrant
labor.
In fact, the experience of developed countries around the world
at
who have employed
one time or another has been mixed. Recent examples of non-citizen labor
include
US employment of Mexicans, German employment of Turks,
citizens of
in
migrant labor
developed countries
Western European employment of
Southern and Eastern European nations, and Malaysians employed
in Singapore.
In each of
these cases, pressure has been generated for change in one or both of the following directions: (1) Granting
the full range of civil rights and social benefits to migrant workers, with the possibility of eventual
citizenship.
A leading example is the situation of long-time resident Turkish guest workers in Germany;
(2) Substitution of capital mobility for labor mobility.
manufacturing from
Hong Kong
one hand, under the
1
here
is
the recent flow of capital and
to Southeastern China.
Recent American experience
On
An example
is
especially interesting because
990 Immigration Reform and Control Act,
it
reflects
moves
illegal aliens
in both directions.
who have been
country from a long time can in some cases apply for amnesty and permanent residence.
hand, efforts have been
the
made
Rio Grande. In addition
program
is
the Offshore
to
to
encourage American firms to
low labor costs and proximity
Assembly Provision of the
US
14
tariff
set
to
On
in the
the other
up establishments on the Mexican
side of
American markets, a key component of this
schedule which stipulates that products
assembled or processed abroad with
The Mexican government
value added.
when
the
US
US
closed
down
parts are levied an import duty (to the
originally
promoted
US
US) only on
the foreign
offshore assembly to offset job losses
a program that permitted seasonal entry of Mexican farm workers (Mendez,
Murray, and Rousslang, 1991).
An
analogous "offshore assembly" effort for Israel would be to promote Palestinian industry and
agriculture in 4he territories, integrated with Israeli distribution and marketing, as an alternative to labor
migration. This process will probably develop faster
employing Palestinians or foreign guest workers
employers do not have the option of
if Israeli
in Israel.
In the past, Israeli policy
makers seem
to
have
found local Palestinian economic development a threatening prospect but lessons from other countries
suggest that this need not be the case.
Some
of the features that make Gazan labor attractive to
employers apply equally well to the employment of Gazans
Gaza using
in
Israeli
Israeli
and other foreign
capital.
Recent research on the determinants of foreign direct investment also suggests that for developing
countries, infrastructure, stable international relations, rapid industrial development,
domestic market are the factors most important for attractive capital from abroad
than tax concessions and low labor costs (Wheeler and
Two
Mody,
and an expanding
- more
important even
1992).
other policy issues
I
conclude with a brief discussion of the policy implications of the two others
reported here.
My study of the returns to schooling (Angrist,
political reasons for the decline in
capital
1995) in the
economic returns on human
have begun to recover for workers
who
live in the
Gaza
capital.
territories
returns to
human
capital, either
may want
discussed possible
In fact, the returns on
Strip, possibly
human
because of autonomy.
the other hand, the returns to schooling remain had not yet turned around in the
the near term, the Palestinian leaders
sets of findings
West Bank
to consider efforts directed at reducing the
On
as of 1995. In
imbalance in
by changing the geographic distribution of public sector employment or by
15
making investment
in the
West Bank more
attractive.
Of course,
the success of such efforts depends on a
range of external factors that are impossible to control.
Finally,_the results for refugees suggest that refugees
only moderately worse off than other Palestinians.
(1996).
While the
results reported here
rates than non-refugees, the
wage gap
is
show
who
reside in the
West Bank and Gaza
A similar conclusion was reached recently by Pedersen
that refugees
closing and the
have somewhat lower wages and employment
employment differences
are not dramatic. This
suggests that resources devoted to further integrating refugees into the local Palestinian
most effective
overall
if
are
economy way be
they are focused on improving schools and physical infrastructure and on boosting the
macro economy,
as
opposed to programs of targeted subsidies and
16
transfers.
References
"Wages and Employment in the West Bank and Gaza Strip: 1981-90," Falk Institute for
Economic Research in Israel, Discussion Paper No. 92.02, Hebrew University, Jerusalem, July
Angrist, J.D.,
1992.
"The Economic Returns to Schooling in the West Bank and Gaza Strip," American
Economic Review 85 (December 1995), 1065-1087.
Angrist, J.D., "Short-run Demand for Palestinian Labor," Journal of Labor Economics 14 (July 1996),
Angrist, J.D.,
425-453.
Bloom,
D.,
and A. Brender, "Labor
in the
World Economy,"
NBER Working Paper No. 4266, January
1993.
Israel Central
Bureau of
Statistics,
Labour Force Survey
in
Judea and Samaria: January-September 1995,
Jerusalem: September, 1996.
Israel Central
Israel
Bureau of
Employment
Abstract of Israel 1996, Jerusalem: 1996.
Payments Division, Dates of Closures 93-94-95, Memo dated January
Statistics, Statistical
Service,
15,
1996.
Lewin-Epstein,
Noah and Moshe Semyonov. The Arab Minority
Inequality.
in Israel's
Economy: Patterns of Ethnic
San Francisco: Westview Press, 1993.
Kleiman, Ephraim. "The Flow of Labour Services From the West Bank and Gaza to
Israel." Department of
Economics Working Paper No. 260, Hebrew University, Jerusalem, July 1992.
Maariv (Hebrew Daily), "Foreign Workers: Underground Israel '96", Saturday Supplement, pages 14-20,
October 25, 1996.
A., T. Murray, and D.J. Rousslang, "US-Mexico Employment Effects of Repealing the
Offshore Assembly Provision," Applied Economics 23 (March 1991), 553-566.
Mendez, Jose
Palestinian Central
Bureau of
Statistics,
US
Labour Force Survey: Main Findings (September-October 1995
Round), Labour Force Survey Report Series (No.
1),
Ramallah: April, 1996.
Bureau of Statistics, Labour Force Survey: Main Findings (April-May 1996 Round),
Labour Force Survey Report Series (No. 2), Ramallah: August, 1996.
Pedersen, Jon, "West Bank and Gaza Living Conditions: Are Refugees Different? (the answer is: not very
much), paper prepared for the Harvard Refugee Project, October 1996.
Wheeler, David, and Ashoka Mody, "International Investment Decisions: The Case of US Firms,"
International Economic Review 33 (August 1992), 57-76.
Palestinian Central
17
;
'
E
»rj
Tt
a
*
6
—
o
^
-*
—
^
O
o
rf"
fi
—
oo
•a
Q
5
°-
—
tN
—
—
5'E
2
t_
t?
13
c«
OS
oi
2
o
o
f^
-
c
u O
cr^
c
o
c
p
CT;
—
c*S
r^
vo
^
0>
00 •a
4>
rt
" o
o
a
'd
,:
>>
p
so
r*i
00
q;
•H
u « 5
S *
.^
C3
E'd
.,-
i5
a B o
o
to
8
u c
=*
lis
U~t
C
—
°
.°
S2
^
CO
set
o o
s E a
a*
—
OS
—
^
o
—
<
—
o
oo
—
S a o
lic
o
V]
tf
o
u « g
a.
?J
E a
i
_-
=03
CL
E
OS
in
—
'
—
—
^^
i
T3
O =
ca
c
03
U
—
S
E S «
en
n
as
o^
-*
—
*
03
d
-'
n
2
2
d
o^
1
o ?
c
^^
C
C 3
"E.
iSt
ia
<u
&
2
"5
^
in
a
u
a.
£
o
<
z
<
z
p
O
•a
c
o
o o
Z
.t
I
la
^^
m <3
—
o
U-l
"^
to
o
-*
o
^
— o
r^
O Tt
— o
— "^
oo
*
— o
CO
ifi-a
vn
M
;S
r^ oo
<N
o
(N »n
<N
O
^-1
o
O
u
oo oo
oo
*-
^ o
o
^-
D
u
O
<
vj->
^ & o ;_
uU O ;^
(N
Tt
„H
m
o
^H >i-l
\D (N
Tf
O
—
CS
csi
O
O O
— o
'
O ON
—o
O
T
S
o
mm
o
§
•^
oo
—o
«
mo
m
(N
oo (N
oo
rr\
— —
Tj-
I
rj oo
^
m
O o
—
o
(N —
•* O
^- oo
XT) ^~
OO
oo
o
—
O
—O
^"
D
— o
CM
»n
O
o
s:00 ,o
o
U
^§
k.
OJ
-r
s
rncNt
'OfNl
TtO
o
f*^
^-
(SO
r^—
f^O
On^O
oo—
— O
c^^/^
oo—
—
m
-o —
xn
—
^O
^ (N
o o
82
q
o
_
°o
P o
oooo
-^
OO
P
OO
—
oo
^
P O
oo
CN OO
»o
o
oO
5 ^
MD -J
P O
<2
O
<
z
o
—
mo
u
cs
oo (^
-r
oo
en
O
S •o:9'
li-)
p-i
o>n
r^(N
^o
-rt
oo rs
<M
O
r--—
•
ONCN
— o
oo
oo
—
CN(N
— o
^-.
— o
moo
On —
m
—o
CO
oo ^-
CO
2
^s
Po
a
a E
— o
,/
E
c
-ii
;::
mo
o
Tj-
o
m <N
o
-^
v£J
c^
Ov
»n
O
^
o
E u w
o
Z
(N
o
po
^
1^
"^
s
•a
>
o
CJ
Jr
£
u
«
-c
OOCT
r
:
^
°oS
Oo
t^f^
oR
"^O
CT^^/^
vo^
of-^
<^»n
prS
Po
'^O
frg
HO
O
— o
o
OJ
":
o
o
Oo
— o
—
SiT!
Po
Oo
O
ts
'^
o
o
— OS
2o
—
O
go—
o g^
P o
S2o
:
Sox
Sf^
Oo
U
n<^
^O
S<^
Oo
**
"".
.
p:
P o
OS
r^ Co"
~ o
f:^
"".
O
So
Po
-:
o
~o
qj
'S-
~:
o
" o
ii-S^
M
—o
oo
o
U
^o
Po
>o
o
(S o
—
po
CJ
—
^
Po
'
o
o cs
o
CsJ
CN
>
— o
P O
^
OS
O^
PO
>n Co
o
U
Po
^' oo
»*
J.
Po
OO ^OO so
—o
Po
^
o o
OS f^
V~i
P o
as
OS
^-'
S2
Po
O
CN
in t^
oo
CQ
"^
U
P o
Po
Po
p o
Po
~;
o
Po
P O
P o
°o
P o
M Co
^
po
o
5
P
So
po
S--3
u
>•
-2o
3
-:
S
S'
p o
—
—
:
So
Po
S'
o
o
Po
p)
en
eg-
P o
o
iji
p
P o
S-
£
CSS
-J
ws
D
Z
c=
-a -a
&
§ i
o
o
^
o
£
^
=
2o
Po
-c
"o
Po
—
P o
"^
?°
SI"
PS
"S S^
P8
P o
.g
g
H
to
"a
<U
5 " S
c
1
3 5
c5
IK
^
CN
o S
s
3 ^
o
PO
:
Po
t~ o
Po
P o
rn oo
P o
P o
U 3 ^
i3
u
S
~o
Po
-3 •=
Po
a
p E
Po
= s s
o [^
U a
o w v.
c 60 C/1
u.
^
OC
.o
-S s
^
^
^3
_
'r7
c^
rt
ca
^ u
3
"rt
C
c
1-^
£i
5
-a
C3
uT
^
g s
i
is
.2?
'5
.70
.
65
.
60
0.55-
0.50w
u
m
0.45
•-1
0.401
u
o
^ 0.35
o
Q
0.30-
§0.25
4-)
U
n3
fc
.20
.
.
15
10
0.05-
.
00
'
'
'
'
'
'
I
I
1
82
'
'
'
'
I
'
I
I
83
84
'
'
'
'
I
'
'
'
85
Region of Residence
Figure
1.
'
I
37
'
'
'
'
'
'
'
'
'
'
'
'
'
'
I
'
I
'
'
'
'
'
'
'
'
'
'
I
I
I
I
I
I
I
88
89
90
91
92
93
94
Year
Gaza
'
'
'
'
'
'
'
I
I
95
96
West Bank
Fraction of employed working in Israel by region
Ji^
0.85-
.
80
CO
Pi
c
o
a,
°
0.75
o
QJ
e
>,
o
j—i
a
e
w
.
.
70
65
'
'
I
1
82
'
I
'
i
I
83
84
85
,
I
I
1
I
I
I
86
87
Region of Residence
Figure
2.
Employment rate by region
'
'
88
89
Year
Gaza
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
I
I
I
I
I
I
90
91
92
93
94
95
'
'
I
96
West Bank
c^^S
{
0.10-
.
09
.
08
07
.
0)
>i
o
1—
ex
%
0.06
Q)
U
>-l
O
05
.
O
(0
04
.
o
u
03
.
.
02
0.01-
.
00
Region of Residence
Figure
3
.
Unemployment rate by region
Year
Gaza
West Bank
iL
24
22
20
T3
O
16
-
14
-
CO
Q
12
10
'
'
'
I
'
'
'
I
81
82
'
'
'
'
I
83
'
'
'
'
I
84
'
'
I
85
Region of Residence
Figure
4.
1
I
I
I
I
'
'
I
I
6
87
88
89
Year
Gaza
Average days worked by region
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
I
I
I
I
I
90
91
92
93
94
'
'
'
'
I
95
'
'
'
I
96
West Bank
^7
26
24
22
-
20
-
0)
u
o
16
Q
14
12
-
10
-
'
'
'
I
1
'
'
I
82
'
'
'
'
I
83
'
'
'
'
I
84
'
'
'
'
I
'
85
Work Location
Figure
5.
'
'
'
'
I
86
'
'
'
'
I
87
'
'
I
I
39
Year
Gaza
Jerusalem
Average days worked by work location
'
'
'
'
'
'
'
'
'
'
'
'
'
I
'
'
'
'
'
'
'
'
I
I
I
I
I
I
I
90
91
92
93
94
95
96
XXX
'
'
Israel
West Bank
3.?
1200
1150
/I
1
1100
1050
1000
950
^
900
(0
^
850
(0
QJ
PS
800
750
4-1
G
o
'^
tn
-
700
650
m
u
CD
>
600
<;
550
-
500
450-
400
-
350
-
300-1
81
-T—i—I—r—i—r-
82
83
84
85
16
87
88
89
91
90
92
93
94
95
96
Year
Work Location
Figure
6.
Gaza
Jerusalem
><
X
X
Israel
West Bank
Average monthly wages in 1991 NIS by work location
Male employees aged 18-64
2f
60
55
-
50
-
Q)
(0
S
45
PS
en
en
(0
Q 40
0)
Di
(0
U
QJ
>
35
30
25
I
82
'
'
'
'
I
83
'
'
!''''!'
'
84
85
Work Location
86
—
'
'
'
'
38
7.
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
I
I
'
'
'
'
'
I
I
I
I
I
I
I
I
89
90
91
92
93
94
95
96
Year
Gaza
Jerusalem
Figure
I
I
'
I
^—^—^ Israel
West Bank
Average daily wages in 1991 NIS by work location
Male employees aged 18-64
3^