American Economic Association
Unemployment and Labor Market Rigidities: Europe versus North America
Author(s): Stephen Nickell
Source: The Journal of Economic Perspectives, Vol. 11, No. 3 (Summer, 1997), pp. 55-74
Published by: American Economic Association
Stable URL: http://www.jstor.org/stable/2138184
Accessed: 13-04-2018 11:08 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
American Economic Association is collaborating with JSTOR to digitize, preserve and extend
access to The Journal of Economic Perspectives
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
Journal of Economic Perspectives-Volume 11, Number 3-Summer 1997-Pages 55-74
Unemployment and Labor Market
Rigidities: Europe versus North America
Stephen Nickell
H T ere is the received wisdom. The European job market is rigid and inflexible.
Result: high unemployment. The North American job market is
dynamic and flexible. Result: low unemployment. So Europeans had better
do something about their labor markets unless they want permanent double
digit unemployment.
In fact, this is not totally wrong. There are features of the labor markets in some
European countries that help sustain high levels of unemployment. Some of these
features can be thought of as rigidities. However, there are many other so-called
rigidities that do not cause high unemployment and, indeed, may serve a useful
purpose. So it is important to know which features of the labor market cause high
unemployment and which do not. This is the subject of what follows.
Labor Market Outcomes in Europe and North America
While it is sometimes convenient to lump all the countries of western Europe
together in order to provide a suitable contrast to North America, most of the time
it is a rather silly thing to do. Different European countries are effectively different
labor markets with the intercountry movement of labor being very small, mainly
because of language and cultural barriers. Partly as a consequence of these differneces,
labor markets in Europe exhibit enormous diversity; in fact, differences
within Europe are much greater than are the difference between the European
average and North America. This section looks at some of these differences, first
* Stephen Nickell is Professor of Economics, Institute of Economics and Statistics, University
of Oxford, Oxford, United Kingdom.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
56 Journal of Economic Perspectives
Table 1
Unemployment Rates in the OECD
1983-96 1983-88 1989-94
Total Total Short-term Long-term Total Short-term Long-term
Austria 3.8 3.6 na na 3.7 na na
Belgium 9.7 11.3 3.3 8.0 8.1 2.9 5.1
Denmark 9.9 9.0 6.0 3.0 10.8 7.9 3.0
Finland 9.1 5.1 4.0 1.0 10.5 8.9 1.7
France 10.4 9.8 5.4 4.4 10.4 6.5 3.9
Germany (W) 6.2 6.8 3.7 3.1 5.4 3.2 2.2
Ireland 15.1 16.1 6.9 9.2 14.8 5.4 9.4
Italy 7.6 6.9 3.1 3.8 8.2 2.9 5.3
Netherlands 8.4 10.5 5.0 5.5 7.0 3.5 3.5
Norway 4.2 2.7 2.5 0.2 5.5 4.3 1.2
Portugal 6.4 7.6 3.5 4.2 5.0 3.0 2.0
Spain 19.7 19.6 8.3 11.3 18.9 9.1 9.7
Sweden 4.3 2.6 2.3 0.3 4.4 4.0 0.4
Switzerland 1.8 0.8 0.7 0.1 2.3 1.8 0.5
U.K. 9.7 10.9 5.8 5.1 8.9 5.5 3.4
Canada 9.8 9.9 9.0 0.9 9.8 8.9 0.9
U.S. 6.5 7.1 6.4 0.7 6.2 5.6 0.6
Japan 2.6 2.7 2.2 0.5 2.3 1.9 0.4
Australia 8.7 8.4 5.9 2.4 9.0 6.2 2.7
New Zealand 6.8 4.9 4.3 0.6 8.9 6.6 2.3
Source: OECD Employment Outlook, UK Employment Trends, U.S. Bureau of Labor Statistics.
with regard to unemployment and then with regard to other labor market outcomes,
notably job and worker mobility, and wage flexibility.
Table 1 sets out some information on unemployment' where we focus on the
recent past, namely the period following the major recession of the early 1980s.
The first column provides an up-to-date summary picture; the other columns present
averages over two subperiods, which will be used for more detailed analysis.
The immediate point that stands out is the enormous variation in European rates.
Taking the period 1983-1996, these stretch from 1.8 percent in Switzerland to
19.7 percent in Spain. This variation means that around 30 percent of the popu'Table
1 uses OECD standardized rates, with the exception of Austria, Denmark and Italy. For Austria
and Denmark, the table presents national registered rates. For Italy, the table presents the unemployment
rate as calculated by the U.S. Bureau of Labor Statistics "on U.S. concepts." Aside from Italy, the OECD
rates and BLS rates are very similar. For Italy, the OECD rates appear to include the large number of
Italians who are registered as unemployed but have performed no active job search in the previous four
weeks. Finally, the unemployment rate here is for West Germany, both to maintain comparability across
time and because including a "transition economy" in the data would weaken comparability across
countries.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
Stephen Nickell 57
lation of OECD Europe lives in countries and operates in labor markets with average
unemployment rates lower than that of the United States.
A closer look at Table 1 raises two additional points. First, the European countries
with the lowest unemployment rates (Austria, West Germany, Norway, Portugal,
Sweden and Switzerland) are not noted for the flexibility of their labor markets.
Britain, on the other hand, has always had the most flexible labor market in Europe
on standard measures and yet has an average unemployment rate higher than half
of its European neighbors.
Second, it is worth remarking on the fact that the variation in short-term unemployment
is substantially smaller than that in long-term unemployment, where
long-term is defined as a duration of more than a year. Thus, while countries require
some short-term unemployment, long-term unemployment appears to be an optional
extra. The reason is that long-term unemployment, in contrast to the shortterm
variety, contributes very little to holding down wage pressure and hence inflation
(OECD, 1993, p. 94). The long-term unemployed are far enough away from
the active labor market that their presence has little influence on wages. So if some
suitable microeconomic policy can eliminate long-term unemployment, this will
have few adverse macroeconomic implications. That is, it will not require much of
a rise in short-term unemployment to maintain stable inflation.
Instead of concentrating on unemployment rates, some commentators prefer
to focus on total employment, noting, for example, that North American employment
has risen much faster in recent years than has European employment. Such
a contrast is not helpful, however, because there is no control for different rates of
growth in the population of working age. Controlling for this by normalizing on
the size of the labor force takes one back to unemployment. A more reasonable
alternative is to focus on employment/population ratios, although these tend to be
strongly influenced by all the social and cultural factors that affect the labor market
participation of married women. Table 2 presents evidence on alternative labor
supply measures, like the employment/population ratio.
The first two columns of Table 2 show the ratio of employed persons to the
total working-age population and the ratio for males ages 25-54. The cross-country
variation in overall employment/population ratios is due to a variety of factors.
Particularly important are variations in the participation rates of married women
(which are very low in southern Europe), variations in the retirement rates of those
over the age of 55 (OECD, 1996, p. 188) and variations in the employment rates
of prime-age men, shown in the second column of the table. The third column of
the table shows annual hours worked by the average worker in these different economies.
Differences in this column are dominated by the extent of part-time working
and by variations in weekly hours and annual holiday entitlements. Many countries
in continental Europe have low annual hours actually worked even excluding parttime
workers, because of their low weekly hours and long annual holidays compared
to those of the United States andJapan. This does not imply that European workers
would like to work more paid hours per year. Indeed, across the EC, more people
would like to work fewer paid hours than would like to work more paid hours at
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
58 Journal of Economic Perspectives
Table 2
Alternative Labor Supply Measures
Employment/Population
Ratio (%) Employment/Population Annual Hours Overall Labor
(whole working age Ratio (%) Worked per Supply
population) (males age 25-54) Worker (%)
Austria 67.3 86.6 1600 51.6
Belgium 56.1 87.4 1580 42.6
Denmark 75.0 86.6 1510 54.5
Finland 67.1 82.4 1770 57.1
France 59.8 87.9 1650 47.4
Germany (W) 65.2 87.0 1600 50.0
Ireland 53.2 80.3 1750 44.8
Italy 54.0 84.3 1730 44.9
Netherlands 62.2 86.5 1510 45.2
Norway 73.3 87.4 1430 50.4
Portugal 69.3 90.6 2000 66.6
Spain 47.5 81.5 1820 41.6
Sweden 75.6 88.2 1510 52.0
Switzerland 78.6 94.7 1640 62.0
UK 69.6 86.7 1750 58.6
Canada 70.6 84.7 1740 59.0
U.S. 73.1 88.2 1940 68.2
Japan 73.4 95.9 1960 69.2
Australia 68.2 86.5 1870 61.3
New Zealand 68.0 86.6 1830 59.8
Source: OECD Employment Outlook (1996), Tables A, B and C.
given hourly rates (European Economy, 1995, Table 25a). The final column, "overall
labor supply," combines the annual hours worked and employment/population
ratios. Take the annual hours worked as a percentage of 2080 hours, which represents
a full-time year of working 40 hours a week for 52 weeks. Multiply this by the
employment/population ratio. The result can be thought of as the proportion of
total "potential" hours worked in the economy. Total labor supply varies enormously
across countries, with Japan, Portugal and the United States all supplying
about two-thirds of potential hours, while Spain and Belgium supply barely 40 percent
of potential.2
Another way of putting the unemployment/rigidity story into a broader perspective
is to look at job and worker mobility. Job turnover is defined as the sum
of the gross job creation and job destruction rates across companies; that is, the
total of all new jobs generated plus all old jobs destroyed. Worker mobility includes
2 Of course, these numbers exclude unmeasured labor input into, for example, the "black economy."
However, these total labor supply numbers are worth bearing in mind when comparing GDP per capita
across countries.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
Unemployment and Labor Market Rigidities: Europe versus North America 59
all job turnover, but also includes the numerous occasions where workers enter or
leave a job in a company when the overall number of jobs remains fixed, because
of quits, retirements and so on. Of course, there are problems of comparability with
cross-national data such as these (Contini et al., 1995), but there is no evidence
that jobs are created and destroyed at a more rapid rate in North America than
they are in Europe. However, workers do appear to circulate faster through the
existingjobs in North America (OECD, 1996, Tables 5.1, 5.2). This is also consistent
with the finding that the United States has relatively high levels of regional mobility:
about 3 percent of U.S. households change their region of residence in a year,
compared to closer to 1 percent in the United Kingdom, Germany and France, and
even lower in Italy and Spain. However, regional mobility rates in Norway and
Sweden are similar to those in the United States (OECD, 1990, Table 3.3). Since
the encouragement of regional mobility has always been a feature of Norwegian
and Swedish labor market policy, this outcome is no surprise.3
A final perspective on the aggregate labor market is to look at the evidence on
wage flexibility. Table 3 presents some measures of the responsiveness of overall
wages to unemployment, derived from both aggregate time series and individual
survey data. Of course, this is only one feature of wage flexibility; for example, it is
not informative about the flexibility of relative wages across different groups. However,
for this particular aspect of wage flexibility, there is no dramatic contrast
between Europe and North America. If anything, Canada and the United States
veer toward the inflexible end of the spectrum.
To summarize, the contrast between Europe and North America is more complex
than is commonly realized. Unemployment is higher in the majority of European
countries than in the United States, but there is considerable variation across
Europe. Rates ofjob turnover are no higher in North America than in Europe and
neither are overall wages any more flexible, but it does seem that U.S. workers are
more mobile than are many Europeans both geographically and between jobs. The
next step is to focus on a large number of separate features of the labor market
and to try to isolate those that have some responsibility for the high levels of unemployment
in many European countries.
What Features of the Labor Market Generate High Unemployment?
Our aim in this section is to pinpoint precisely which features of the labor
market generate unemployment and which do not. Then we can discuss how these
facts relate to the view that high unemployment in Europe is due to rigid and
inflexible labor markets.
The first step is to look at labor market characteristics in different countries.
Table 4 presents direct measures of labor market rigidities and summary statistics
'The "regions" in all these countries are comparable in size, so these comparisons have some meaning.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
60 Journal of Economic Perspectives
Table 3
Wage Flexibility: The Percentage Increase in Wages in Response to a One
Percentage Point Fall in the Unemployment Rate
Aggregate Time Series Measure
Microeconometric
Short-run Long-run Measure
Austria 1.43 3.11 2.43
Belgium 0.65 4.06
Denmark 0.66 1.74
Finland 0.48 1.55
France 2.22 4.35
Germany (W) 0.55 1.01 2.06
Ireland 0.80 1.82 2.35
Italy 2.07 12.94 1.32
Netherlands 0.66 2.28 1.98
Norway 1.96 10.59 1.95
Spain 0.17 1.21
Sweden 2.31 12.16
Switzerland 1.32 7.33 7.06
U.K. 0.98 0.98 0.82
Canada 0.50 2.38 0.92
U.S. 0.32 0.94 1.52
Source: Aggregate time series measures; Layard et al. (1991), chapter 9, Table 2. Microeconomic measures;
Blanchflower and Oswald (1994), Table 9.1. These later numbers are derived by dividing the Blanchflower/Oswald
numbers by the average unemployment rate, because they only report the elasticity of
wages with respect to unemployment.
on the treatment of the unemployed. The employment protection index in the first
column was drawn up by the OECD and is based on the strength of the legal
framework governing hiring and firing. The countries are ranked from 1-20, with
20 being the most strictly regulated. The countries of southern Europe have the
toughest regulations and, roughly speaking, these regulations get weaker as one
moves further north. Switzerland, Denmark and the United Kingdom have the
weakest laws in Europe, and these laws are comparable to those in place outside
Europe.
The labor standards index in the second column was also drawn up by the
OECD and refers to the strength of the legislation governing a number of aspects
of the labor market. The index ranges from 0 to 10, with each country being scored
from 0 (lax or no legislation) to 2 (strict legislation) on each of the five dimensions:
working time, fixed-term contracts, employment protection, minimum wages and
employees' representation rights (on works councils, company boards and the like).
The scores are then added up. The picture is similar to the employment protection
column. The United Kingdom and the United States have very weak legislation in
this area, whereas Spain and Italy have many strict rules and regulations. So it is
undoubtedly true that if we are to think of inflexibility as referring to legal restricThis
content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
Stephen Nickell 61
Table 4
Features of OECD Labor Markets I, 1989-1994
Direct Rigidities Treatment of the Unemployed
5
1 2 3 4 Active Labor
Employment Labor Benefit Replacement Benefit Duration Market
Protection Standards Rate (%) (years) Policies
Austria 16 5 50 2 8.3
Belgium 17 4 60 4 14.6
Denmark 5 2 90 2.5 10.3
Finland 10 5 63 2 16.4
France 14 6 57 3 8.8
Germany (W) 15 6 63 4 25.7
Ireland 12 4 37 4 9.1
Italy 20 7 20 0.5 10.3
Netherlands 9 5 70 2 6.9
Norway 11 5 65 1.5 14.7
Portugal 18 4 65 0.8 18.8
Spain 19 7 70 3.5 4.7
Sweden 13 7 80 1.2 59.3
Switzerland 6 3 70 1 8.2
U.K 7 0 38 4 6.4
Canada 3 2 59 1 5.9
U.S. 1 0 50 0.5 3.0
Japan 8 1 60 0.5 4.3
Australia 4 3 36 4 3.2
New Zealand 2 3 30 4 6.8
Source: OECD Jobs Study (1994), Part II, Table 6.7, column 5. OECD Employment Outlook (1994), Table
4.8, column 6 (extended by author). U.S. Department of Health and Social Services, Social Security Programmes
Throughout the World (1993). OECD Employment Outlook (1995), Table T.
tions on the operation of the labor market, southern and continental Europe are
the most inflexible. As an offset to this, however, it is worth remarking that southern
Europe also has the highest rate of self-employment in the OECD (OECD, 1994,
Table 6.8). The self-employed are, presumably, among the most flexible of all
workers.
Benefit systems vary quite dramatically. The "replacement rate," which shows
what share of income is replaced by unemployment benefits, and the duration of
these benefits (four years means indefinite duration) are typically fairly generous
by U.S. standards (50 percent replacement rate for six months). Italy, however,
barely had an unemployment benefit system at all for most of the postwar period.4
'Until recently, the unemployed in Italy were entitled to 800 lira per day (around 50 cents). A small
proportion of the "unemployed" would be covered by the CIG scheme for industrial workers who are
in danger of being laid off. These typically do not amount to more than 1 percentage point of
unemployment.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
62 Journal of Economic Perspectives
Some of the countries with the most generous benefit levels have strictly timelimited
systems, notably in Scandinavia, like Sweden's 80 percent replacement rate,
which is limited to 1.2 years. The next column, "active labor market policies," refers
to expenditures on activities for the unemployed that are geared to help them back
into work and are popular in many, although not all, European countries. These
include labor market training, assistance with job search, subsidized employment
and special measures for the disabled. The numbers in this column are derived by
taking active labor market spending per unemployed person as a percentage of
GDP per member of the labor force. Thus, Sweden's figure of close to 60 shows
that expenditure on active policies per unemployed person is nearly 60 percent of
national output per potential worker, which is extraordinarily high. Spain, on the
other hand, is notable for its combination of a generous benefit system and a low
level of expenditure on active labor market policies.
The first few columns of Table 5 present variables that summarize the structure
of wage determination systems. In most European countries, with the exception of
the United Kingdom and Switzerland, trade unions play a very significant role in
wage determination. The union density column shows the proportion of trade union
members as a percentage of all wage and salary earners. However, this does not
tell the whole story. In many nations, union wage negotiations determine the wages
of workers who are not explicitly part of the union. In Spain and France, for example,
only about 10 percent of workers are union members, but the wages of over
70 percent of all workers are covered by union bargaining. Thus, the "union coverage
index" presents a summary of the share of workers actually covered by union
bargaining, where 3 means over 70 percent covered, 2 means from 25-70 percent,
and 1 is under 25 percent.
The next column of the table shows the extent of coordination in wage bargaining,
on the part of both unions and employers. In each country, the degree of
union and then employer coordination is ranked from a low of 1 to a high of 3. In
some of these countries, both unions and, more significantly, employers coordinate
their wage bargaining activities, particularly in central Europe and Scandinavia. In
those countries where unions play a lesser role, although still an important onelike
the United Kingdom, Switzerland, and all non-European countries except the
United States-there is very little coordination over wage bargaining, with the notable
exceptions of Switzerland and Japan, where employer coordination is very
important.
The final two columns of the table give information on the tax burden on
labor. First we have the payroll tax rate, defined as the ratio of labor costs to wages
(less unity) and then we show the total tax rate, which is the sum of the average
payroll, income and consumption tax rates. The latter are based on aggregate tax
and income data. The payroll tax rate varies dramatically across countries, with
Denmark levying no payroll taxes and France and Italy with a rate close to
40 percent. The total tax rate is less variable and represents a crude measure of the
tax wedge between real labor costs and real take-home pay. This is arguably the
correct measure of the tax burden on labor.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
Unemployment and Labor Market Rigidities: Europe versus North America 63
Table 5
Features of OECD Labor Markets II, 1989-1994
3
1 2 Co-ordination 4 5
Union Density Union Coverage Payroll Tax Total Tax
(%) Index Union Employer Rate (%) Rate (%)
Austria 46.2 3 3 3 22.6 53.7
Belgium 51.2 3 2 2 21.5 49.8
Denmark 71.4 3 3 3 0.6 46.3
Finland 72.0 3 2 3 25.5 65.9
France 9.8 3 2 2 38.8 63.8
Germany (W) 32.9 3 2 3 23.0 53.0
Ireland 49.7 3 1 1 7.1 34.3
Italy 38.8 3 2 2 40.2 62.9
Netherlands 25.5 3 2 2 27.5 56.5
Norway 56.0 3 3 3 17.5 48.6
Portugal 31.8 3 2 2 14.5 37.6
Spain 11.0 3 2 1 33.2 54.2
Sweden 82.5 3 3 3 37.8 70.7
Switzerland 26.6 2 1 3 14.5 38.6
U.K. 39.1 2 1 1 13.8 40.8
Canada 35.8 2 1 1 13.0 42.7
U.S. 15.6 1 1 1 20.9 43.8
Japan 25.4 2 2 2 16.5 36.3
Australia 40.4 3 2 1 2.5 28.7
New Zealand 44.8 2 1 1 34.8
Source: Layard et al. (1991), Annex 1.4, and OECD Employment Outlook (1994), p. 175-85. Centre for
Economic Performance (LSE), OECD data set.
Overall, therefore, there are quite substantial differences between European
and North American labor markets as well as important differences within Europe.
The consequences of these differences for unemployment and labor supply form
our next topic.
The Labor Market and Unemployment
Our purpose in what follows is to investigate the relations between unemployment
and other measures of labor supply, and labor market institutions. Table 6
presents three regressions relating to unemployment. Each regression is based on
two cross-sections dated 1983-88 and 1989-1994. The dependent variables are the
unemployment rates reported in Table 1, and the values of the independent variables
for the time period 1989-1994 are from Tables 4 and 5. The corresponding
values of the independent variables for 1983-88 are not presented here but are
available from the author. Some variables take the same values for both periods,
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
64 Journal of Economic Perspectives
Table 6
Regressions to Explain Log Unemployment Rate Percentage
(20 OECD countries, 1983-88 and 1989-1994)
1 2 3
Total Long-term Short-term
Unemployment Unemployment Unemployment
Employment Protection (1-20) -0.0032 (0.03) 0.051 (0.034) -0.046 (0.024)
Replacement Rate (%) 0.011 (0.0050) 0.011 (0.0080) 0.011 (0.0060)
Benefit Duration (years) 0.088 (0.055) 0.25 (0.089) 0.043 (0.062)
Active Labor Market Policies' -0.024 (0.0087) -0.039 (0.013) -0.012 (0.0098)
Union Density (%) 0.012 (0.0063) 0.010 (0.0096) 0.0082 (0.0071)
Union Coverage Index (1-3) 0.45 (0.22) 0.83 (0.35) 0.39 (0.24)
Co-ordination (Union + Employer) (2-6) -0.46 (0.087) -0.54 (0.15) -0.37 (0.11)
Total Tax Rate (%) 0.026 (0.0087) 0.023 (0.013) 0.025 (0.010)
Change in Inflation (% pts. p.a.) -0.17 (0.11) -0.30 (0.17) -0.18 (0.10)
Dummy for 1989-94 0.20 (0.095) 0.30 (0.16) 0.17 (0.089)
R2 0.76 0.84 0.60
N (countries, time) 40 (20, 2) 38 (19, 2) 38 (19, 2)
Notes: Estimation is by GLS random effects using two time periods (1983-88 and 1989-1994). Standard
errors are in parentheses.
'The variable is instrumented. Because the active labor market policies variable refers to percentage of
GDP normalized on current unemployment, this variable is highly endogenous. So we renormalized the
current percentage of GDP spent on active labor market measures on the average unemployment rate
in 1977-79 to create the instrument. Insofar as measurement errors in unemployment are serially uncorrelated,
this will help with the endogeneity problem.
but many are different. We chose to use six-year averages in order to smooth out
both the cycle and year-on-year noise. On the other hand, we felt there was enough
useful information here to warrant the use of two cross-sections rather than one
12-year average. The regression coefficients are estimated using the standard random
effects generalized least squares procedure, which is essentially ordinary least
squares corrected for the fact that the two successive observations for each country
cannot be treated as independent random draws. Finally, note that the dependent
variables are the logs5 of the unemployment rate (column 1), the long-term rate
(column 2) and the short-term rate (column 3). Thus, if the right-hand side of the
equation increases by 0.1, log unemployment goes up by 0.1, so unemployment
rises by just over 10 percent. From a baseline unemployment rate of 5 percent, this
would represent an increase of half a percentage point to 5.5 percent.
In Table 7 we report similar regressions explaining other aspects of labor supply,
notably the employment/population ratios and overall labor supply reported
'The use of the log of the unemployment rate follows from the fact that many investigations of wage
determination find that the use of log u in a wage equation is preferable to the use of u. See Blanchflower
and Oswald (1994), for example.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
Stephen Nickell 65
Table 7
Regressions to Explain Labor Supply Measures
(20 OECD countries, 1983 and 1989-1994)
Employment/Population Ratio (%)
1 2 3
Whole Working Males Aged Overall Labor
Age Population 25-54 Supply
Employment Protection (1-20) -0.94 (0.30) 0.040 (0.18) -0.70 (0.39)
Replacement Rate (%) -0.026 (0.072) -0.052 (0.043) -0.037 (0.091)
Benefit Duration (years) -1.26 (0.63) -0.61 (0.43) -0.32 (0.73)
'Active Labour Market Policies 0.16 (0.11) 0.081 (0.073) -0.028 (0.14)
Union Density (%) -0.082 (0.086) -0.11 (0.053) -0.18 (0.11)
Union Coverage Index (1-3) -0.96 (2.54) -1.36 (1.74) -2.24 (2.84)
Coordination (Union + Employer) (2-6) 5.03 (1.23) 2.71 (0.74) 4.20 (1.58)
Total Tax Rate (%) -0.24 (0.12) -0.16 (0.075) -0.26 (0.16)
Change in Inflation (% pts. p.a.) -2.12 (0.93) -0.97 (0.72) -2.02 (0.97)
Dummy for 1989-94 1.87 (0.79) -2.09 (0.63) 0.041 (0.83)
RS 20.81 0.63 0.51
N (countries, time) (20, 2) (20, 2) (20, 2)
Notes: Estimation is by GLS random effects using two time periods (1983, 1989-1994). Standard errors
are in parentheses.
a Active labor market prices are instrumented as in Table 6.
in Table 2. Again we use two cross-sections with the same independent variables.
In this case, the dependent variables are not in logs.
Before we go on to discuss particular rigidities, it is worth commenting briefly
on the status of these results. First, we see them as a helpful overview of the correlations
in the data and nothing more. Like all simple cross-section correlations,
care must be taken with their interpretation because of issues of reverse causality
and the like. Second, despite the use of six-year averages, there may still be significant
long-term variations across countries in the stance of macroeconomic policy.
We control for the average change in inflation as one attempt to deal with this
problem. Third, there may be factors that explain cross-country differences in unemployment
that are not associated with the labor market. For example, it can be
argued that higher levels of product market competition tend to reduce unemployment
(Layard, Nickell and Jackman, 1991, chapters 7 and 9, for example).
Unfortunately, it has not proved possible to obtain measures of product market
competition that are consistent across enough countries to include in the
regressions.
Finally, and most importantly, why focus only on the 1980s and 1990s? Underlying
this question is the reasonable argument that in the 1960s, the unemployment
rankings across countries were completely different but, roughly speaking, the labor
market institutions were the same. So how can the labor market institutions have
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
66 Journal of Economic Perspectives
anything to do with unemployment? Part of the answer has to be that the institutions
had a big impact on the way in which each of the economies of the different
countries responded to the major adverse shocks of the 1970s and the way in which
some of these responses, notably unemployment, persisted through the 1980s and
1990s. In part, this effect is what our regressions are picking up. There remain a
number of unanswered questions concerning the evolution of labor markets since
the 1960s. Here, our main concern is much more limited, namely the question of
which institutions, for whatever reason, appear to be important in understanding
recent unemployment levels across the OECD. So let us consider various institutions
in turn.
Direct Rigidities
Labor market legislation is typically put in place to protect employees from
arbitrary, unfair or discriminatory actions on the part of employers. In so doing, it
may raise the effective cost to firms of employing workers and/or raise the effective
cost of adjusting levels of employment. The impact of the former on unemployment
depends crucially on the extent to which the extra costs are shifted onto employees
by a suitable adjustment of the wage. The general evidence on payroll taxes (as we
shall see) is that the major part of the burden of such costs is typically shifted onto
workers in the long run, thereby nullifying their impact on unemployment. While
this obviously cannot be the case for minimum wages, there is no evidence in our
data that high labor standards overall have any impact on unemployment whatever.
For example, if we add our labor standards variable (Table 4, column 1) to our
unemployment regression (Table 6, column 1), it has a negligible and completely
insignificant coefficient.6
Laws that raise the cost of employment adjustment, notably those relating to
employment protection, will tend to reduce the inflow into unemployment and,
because they make firms more cautious about hiring, will also reduce the flow out
of unemployment into work. This will almost certainly reduce short-term unemployment
(via the reduced inflow) and raise long-term unemployment (via the
reduced outflow). The overall impact on unemployment is likely to be rather small,
as these effects would tend to cancel out. The results in the first row of Table 6 are
entirely consistent with this discussion and confirm the analysis of Bentolila and
Bertola (1990).
However, as the coefficients in the first row of Table 7, columns 1 and 3, indicate,
there is some evidence of a negative correlation between employment protection
and measures of labor supply that go beyond unemployment (see also Lazear,
1990). Much of this correlation arises, in fact, because participation rates
among married women in southern Europe are very low and employment protection
laws in these countries are very tough (OECD, 1994, Table 6.9). Thus, as the
first row of Table 7 also indicates, if we focus on prime-age men (column 2), there
'The coefficient is 0.019 with a standard error of 0.063.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
Unemployment and Labor Market Rigidities: Europe versus North America 67
is no effect. A speculative hypothesis might be that low female participation and
tough employment protection laws in southern Europe are both consequences of
a culture that places a great deal of weight on the position of the (male) head of
household, which is not to be undermined either by the presence of a high-earning
wife or by the loss of a job.
The Treatment of the Unemployed
There are two aspects of the treatment of unemployed individuals, which might
be termed passive and active. The passive is exemplified by the payment, as of right,
of unemployment benefit for a given period. Active policies, on the other hand,
consist of measures that attempt to ensure that the unemployed individual is able
and willing to take up work.
On the passive side, generous benefit systems influence unemployment via two
mechanisms. First, they reduce the fear of unemployment and hence directly increase
upward pressure on wages from employees (via unions, for example). Second,
they reduce the "effectiveness" of unemployed individuals as potential fillers
of vacancies, by allowing them to be more choosy. The impact of a high benefit
replacement ratio on unemployment is well documented (Layard, Nickell andJackman,
1991; OECD, 1994, chapter 8) and is confirmed by the significant coefficient
on the replacement rate in Table 6. The other important feature of the benefit
system is the duration of entitlement. Long-term benefits generate long-term unemployment
(Table 6, row 3; OECD, 1991, Chart 7.1B). Of course, it can be argued
that countries might introduce more generous benefit systems when unemployment
is a serious problem, so that in cross-country correlations, the causality runs
from unemployment to benefits rather than the other way round. However, the
microeconometric evidence on the positive impact of benefit levels and entitlement
durations on the duration of individual unemployment spells (Narendranathan,
Nickell and Stern, 1985; Meyer, 1990) suggests that at least part of the observed
cross-country correlation can be taken at face value.
The impact of a relatively generous benefit system might be offset by suitable
active measures to push the unemployed back to work. Such policies seem to work
particularly well when allied to a relatively short duration of benefit entitlement,
reducing long-term unemployment while alleviating the social distress that might
be caused by simply discontinuing benefits without offering active assistance toward
ajob. Their effects are well summarized in OECD (1993, ch. 2), and their significant
impact in reducing long-term unemployment is illustrated in the fourth row of
Table 6.
While benefits affect unemployment, our evidence suggests that the benefit
system seems to have little impact on overall labor supply as shown in Table 7.
There is a suggestion here that while high benefits lead to high unemployment,
they also lead to high participation because they make participation in the labor
market more attractive, because participation is necessary to be eligible for the high
benefits. This is consistent with a weak impact of benefits on employment/popuThis
content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
68 Journal of Economic Perspectives
lation ratios, because the higher unemployment effect and the higher labor market
participation effect tend to cancel out.
Wage Determination and Unions
The key features of wage determination systems are the extent to which wages
are determined collectively, via union bargaining (union coverage), and the degree
to which employers and unions coordinate their wage bargaining activities given
that wages are determined collectively. Of course, if wages are not generally determined
collectively, as in the United States, the extent of coordination simply does
not apply.
Unions tend to raise pay, and thus one would expect the extent of union activity
in an economy to influence unemployment. This is confirmed by the results in rows
5 and 6 of Table 6, where greater union density and especially union coverage tend
to raise unemployment. However, Table 6, row 7, also shows that this is offset if
unions and employers can coordinate their bargaining activities. For example, leapfrogging
is a common feature of decentralized, uncoordinated, union-dominated
systems; that is, each union tends to take an earlier pay settlement in a related
sector as a baseline to be exceeded in its own negotiations. This generates an additional
source of inflationary pressure that requires more unemployment to quash
it. If unions and employers can coordinate their wage bargaining activities, such
leapfrogging may be eliminated.
It is important to note that coordination does not mean centralization, which
typically implies government involvement in wage bargaining. Both Japan and Germany
have a high degree of coordination in wage bargaining, particularly across
employers, but neither system is centralized. And as OECD (1994, Table 5.16)
makes clear, coordination appears to have a significant negative impact on wages,
whereas the centralization of wage bargaining does not. To summarize, therefore,
unions are bad for jobs, but these bad effects can be nullified if both the unions
and the employers can coordinate their wage bargaining activities.
Labor Taxes
Lowering payroll taxes is a very popular recommendation by those concerned
with reducing unemployment (OECD, 1994; Phelps, 1994). It is easy to understand
this advice if a payroll tax is viewed as a tax on jobs. Things are not, however, quite
as they seem. The first point to recognize is that, broadly speaking, the key tax rate
for the labor market is the sum of the payroll tax rate, the personal income tax rate
and the consumption tax rate. Switching between these taxes will not have an important
impact, so payroll taxes, per se, are of little consequence. This result has
nothing to do with the incidence of these taxes, which we shall address later. It
derives from the logic of supply and demand.
Consider a simple example. Suppose we have a labor market where total labor
costs per employee are $100, payroll taxes paid by the employer are $10 (so pretax
wages are $90), income taxes paid by the employee are $10, and post-tax wages are
$80. Suppose this labor market is in equilibrium. Thus, firms are just willing to
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
Stephen Nickell 69
employ at $100 all the workers who are willing to work at $80. Now suppose that
income taxes are reduced to $5 and payroll taxes are raised to $15 to maintain
revenue. Further, suppose that as a result of this change, firms pay pretax wages of
$85. Then labor costs per employee are $100; post-tax wages are $80. This remains
an equilibrium because firms are still willing to employ at $100 all the workers who
want to work at $80. Nothing substantive has changed except that pretax wages
have fallen from $90 to $85. But this is irrelevant; the only prices that interest the
agents in this economy are labor costs per employee and post-tax wages.
But what about consumption taxes? Employees are interested in what their
wages can buy. So if their income taxes are cut by 10 percent and the cost of
consumption is raised by 10 percent, post-tax real wages are unchanged and so is
labor market behavior. So, broadly speaking, what really counts is the sum of payroll
taxes, income taxes and consumption taxes; the total tax burden on labor. Of
course, this is not exactly correct for a variety of reasons. For example, income tax
is charged on nonlabor income whereas payroll tax is not, so that a cut in payroll
tax and a rise in income tax will reduce nonlabor income, raise labor supply and
reduce unemployment. But, in practice, this is not important because individuals
who are likely to become unemployed have little or no nonlabor income.7 Our
conclusion is that payroll taxes, per se, can be expected to have little impact on
unemployment but the total tax burden might.
The fundamental question, therefore, is whether or not this total tax burden
is entirely shifted onto labor. That is, does real labor cost per worker remain unaffected
by variations in the total tax burden, at least in the long run?
If capital is internationally mobile and labor is not, then we should expect to
see labor bearing all of the tax burden. In this case, employment and unemployment
will, in the long run, remain unaffected by changes in the overall tax rate on
labor. There is, however, one situation where it is impossible to shift payroll taxes
onto workers. That is where there is a rise in the payroll tax and an employee is
already receiving the minimum wage. The burden of the extra tax must then fall
on the employer because the wage cannot adjust.
What happens in practice? The balance of the evidence suggests that lowering
payroll taxes and raising consumption taxes will have no long-run impact on unemployment
(OECD, 1990, Annex 6A; OECD, 1994, Table 9.5).8 This result is confirmed
by the fact that if we include the payroll tax rate in any of the regressions
in Table 6 or 7, its coefficient is always negligible.9 It also helps to explain why
Denmark, which uniquely has no payroll taxes, has unemployment on a par with
7For example, in Britain in 1987-88, only 7 percent of the unemployed had savings in excess of 3,000
pounds, enough to produce an annual interest income of around 10 percent of unemployment benefit
(Layard, Nickell and Jackman, 1991, Table A6).
8 There are some individual country time series results that appear to give a role to payroll taxes in
individual countries-see OECD (1994, Chapter 9, p. 247) for a summary. However, in relatively short
time series, it is often very difficult to distinguish between long-lasting short-run effects and long-run
effects.
9 For example, in columns of Table 6 its coefficient is -0.014, with a standard error of 0.06.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
70 Journal of Economic Perspectives
the European average and appears to derive no special employment benefit from
its lack of these taxes.
The evidence on the total tax burden is less clear. One careful cross-country
study has ruled out any long-run impact of the total tax burden on employment
(OECD, 1990, Annex 6A). However, the results in Tables 6 and 7, row 8, which are
in agreement with the findings of Bean, Layard and Nickell (1986), suggest that
the overall tax burden may raise unemployment and reduce labor supply. A
10 percentage point fall in the total tax burden reduces unemployment by around
25 percent and raises labor supply by around 2 percentage points on every measure.
Of course, a 10 percentage point fall in the total tax burden is enormous. Most
countries find permanently reducing expenditure by 1 percent of GDP an extremely
difficult task. To generate a 10 percentage point shift would mean, for
example, transferring the whole of the UK health service to the private sector.'0
Minimum Wages
While it is impossible to produce a single cross-country variable that captures
the impact of minimum wage laws or related legislation (like extending union pay
bargains to the nonunion sector), it is still worth discussing the potential impact of
minimum wages on unemployment. A reading of Card and Krueger (1995) and its
various reviews in the July 1995 issue of the Industrial and Labor Relations Review
reveals that there is no consensus on the impact of minimum wages on unemployment.
However, the following conclusions do seem to be consistent with the evidence.
First, where the minimum wage applies, it is low enough not to have an
important effect on the unemployment rates of adult men. Second, minimum wages
do have a significant though small adverse impact on youth unemployment rates,
particularly in countries like France and Spain where payroll taxes are high and
there is little in the way of an age adjustment to the minimum wage (Dolado et al.,
1996; Abowd et al., 1996).
Labor Supply Measures
Two much-canvassed solutions to unemployment are reduced hours of work
and early retirement. Advocates of these measures often seem to imagine that there
is some exogenously given level of work to be done. In fact, all historical evidence
shows that, for a given institutional structure, the amount of work to be done tends
to adjust in line with the available supply of labor, leaving equilibrium unemployment
unaffected. So we can expect that an imposed cut in hours or reduction in the
labor force will raise wage pressure in a way that can only be offset by an equivalent
cut in jobs. Indeed if, in a standard wage equation, we allow wages to depend
"' Even if there were some macroeconomic benefits to this, there could easily be substantial costs; for
example, total health expenditure in the United Kingdom is 4-5 percentage points of GDP less than
health expenditure in the United States, without there being notable differences in the overall health
of the two populations. Moving health care to the private sector might impair efforts to hold down costs,
or result in greater inefficiency.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
Unemployment and Labor Market Rigidities: Europe versus North America 71
separately on (the logs of) labor force and employment instead of on unemployment,
we typically obtain equal and opposite coefficients. This indicates that a fall
in the labor force relative to employment raises wage pressure just as much as a rise
in employment relative to the labor force (Layard, Nickell and Jackman, 1991,
p. 504; Jackman, Layard and Nickell, 1996, p. 28). Similarly, if one adds measures
of labor supply like hours worked per worker to the unemployment regressions in
Table 6, no significant effect is found.
Shifts in the Demand for Skills and Unemployment
It has become commonplace to argue: "The rise in joblessness in Europe is
thus the flip side of the rise in earnings inequality in the U.S." (Freeman, 1995,
p. 19). This view is based on the notion that first, in all countries, there has been
an increase in the relative demand for skilled workers (as against unskilled workers)
that has been greater than the increase in their relative supply. Then the argument
goes that in Europe, the inflexibility of the labor market has turned this shift into
higher unemployment whereas in the United States, labor market flexibility has
translated this shift into increased inequality. Finally, this shift explains the majority
of the rise in European unemployment relative to that in the United States.
Despite this being a commonly held view, a variety of facts cast doubt on it
(Card, Kramarz and Lemieux, 1995; Nickell and Bell, 1995, 1996; Nickell, 1996;
Jackman et al., 1996). First, it appears to be the case that in Britain and the United
States the demand for skill outran the supply by more than in the rest of Europe.
Second, for a variety of European countries including Britain, the evidence suggests
that skill shifts account for between 0 and 20 percent of the rise in unemployment
from the 1970s. There is no evidence that this number is lower in "flexible" Britain
than it is anywhere else in Europe. In any event, the vast majority of the rise in
European unemployment is due to other factors. Third, there has been a substantial
rise in unskilled unemployment in the United States since the early 1970s (over
100 percent) despite (because of?) the fall in unskilled real wages. Fourth, the
adverse impact of the fall in the relative demand for unskilled workers on the wages
and unemployment of this group is strongly attenuated in those countries whose
education and training systems are particularly effective at raising the human capital
of those at the lower end of the ability range (notably middle Europe" and
Scandinavia).
Overall, therefore, there is no evidence that these skill shifts have made a
substantial contribution to the rise in European unemployment nor that labor market
inflexibility per se is associated in any simple way with such effects as have been
observed.
Special Cases and the Demand Side
Our aim has been to understand what generates high average levels of unemployment
over long periods. Business cycle effects and autonomous demand shocks of
" That is, Switzerland, Austria, Germany and Holland.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
72 Journal of Economic Perspectives
various kinds should wash out if we take a long enough period-and our focus has
been on 1983-1996, a 14-year stretch. Despite the length of this period, it is possible
to argue that because of exceptional problems, policy mismanagement, very high levels
of hysteresis and the like, the average unemployment figures give a distorted picture
of the underlying equilibrium rate. If we were just considering the 1990s, this argument
might be applied to a number of countries, such as Sweden. But over the longer period,
there is only one country where truly exceptional problems have distorted the long
period average dramatically, namely Finland. In the three years from 1990 to 1993,
Finnish unemployment rose from 3.4 to 17.7 percent. This increase was generated first
by the collapse of an enormous domestic credit boom, which was, in its turn, brought
about by a mismanaged deregulation of the financial sector. Real house prices fell by
over 50 percent between 1990 and 1993. This disaster was reinforced by the more or
less complete elimination of Soviet trade over the same period, which had previously
been responsible for about one-third of Finnish exports. Without these exceptional
events, there is no question that average unemployment would have been substantially
lower over the relevant period and this lower number would more accurately reflect
the equilibrium rate in Finland.
Conclusions
High unemployment is associated with the following labor market features:
1) generous unemployment benefits that are allowed to run on indefinitely, combined
with little or no pressure on the unemployed to obtain work and low levels of active
intervention to increase the ability and willingness of the unemployed to work; 2) high
unionization with wages bargained collectively and no coordination between either
unions or employers in wage bargaining; 3) high overall taxes impinging on labor or
a combination of high minimum wages for young people associated with high payroll
taxes; and 4) poor educational standards at the bottom end of the labor market.
Labor market rigidities that do not appear to have serious implications for
average levels of unemployment include the following: 1) strict employment protection
legislation and general legislation on labor market standards; 2) generous
levels of unemployment benefit, so long as these are accompanied by pressure on
the unemployed to take jobs by, for example, fixing the duration of benefit and
providing resources to raise the ability/willingness of the unemployed to take jobs;
and 3) high levels of unionization and union coverage, so long as they are offset
by high levels of coordination in wage bargaining, particularly among employers.
Suppose we define high unemployment as above 120 percent of the U.S. rate
over the 1983-1996 period (7.8 percent). Then, looking at Table 1, we see there
are eight European countries in this category out of 15, as well as Canada. These
eight include three major countries (France, Spain and United Kingdom) of which
the last has far and away the most flexible labor market in Europe, as normally
measured. The remaining countries with high unemployment are Belgium, Denmark,
Finland, Ireland and the Netherlands.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
Stephen Nickell 73
Unemployment is high in these countries (excluding Finland, for reasons already
explained) partly because, on average, they have reasonably generous benefits
with very long periods of entitlement and little in the way of active policies to push
the unemployed into work. Wages are typically bargained collectively, so unions
apply pressure on wages, but coordination is not high, particularly among employers.
Education levels at the lower end of the ability range are generally weak. Of
course, not all of these apply to every country, and the country to which they apply
least, the Netherlands, is now moving out of the high-unemployment group. Most
importantly for the topic of this paper, many features of the labor market that are
popularly viewed as serious rigidities apply no more to this high-unemployment
group than they do to the low-unemployment group. These include high payroll
taxes, high overall taxes, strict employment protection legislation, high labor market
standards (legally enforced), high unionization and high benefit replacement
rates.
It is clear that the broad-brush analysis that says that European unemployment
is high because European labor markets are "rigid" is too vague and probably
misleading. Many labor market institutions that conventionally come under the
heading of rigidities have no observable impact on unemployment.
* I am most grateful to Tracy Jones and the Leverhulme Trust (Programme on Unemployment
and Technical and Structural Change) for their help in the preparation of this paper. My
thanks are also due to Alan Krueger, Bradford De Long and Timothy Taylor for their very
useful comments on an earlier draft.
References
Abowd, J., F. Kramarz, T. Lemieux, and D.
Margolis, "Minimum Wages and Youth Employment
in France and the United States," mimeo,
Cornell University, 1996.
Bean, C., R. Layard, and S. J. Nickeli, "The
Rise in Unemployment: A Multi-Country Study,"
Economica, Supplement 1986, 53, S1-S22.
Bentolila, S., and G. Bertola, "Firing Costs
and Labour Demand: How Bad is Eurosclerosis,"
Review of Economic Studies, 1990, 57:3,
381-402.
Blanchflower, D., and A. Oswald, The Wage
Curve. Cambridge: Massachusetts Institute of
Technology Press, 1994.
Card, D., and A. Krueger, Myth and Measurement:
The New Economics of the Minimum Wage.
Princeton, N.J.: Princeton University Press, 1995.
Card, D., F. Kramarz, and T. Lemieux,
"Changes in the Relative Structure of Wages
and Employment: A Comparison of the United
States, Canada, and France." Industrial Relations
Section, Working Paper No. 355, Princeton University,
December 1995.
Contini, B., L. Pacelli, M. Filippi, G. Lioni, and
R. Revelli, "A Study ofJob Creation andJob Destruction
in Europe," study for the European
Commission, D.G.V., Turin, R. & P., 1995.
Dolado, J., F. Kramarz, S. Machin, A. Manning,
B. Margolis, and C. Teulings, "The Economic
Impact of Minimum Wages in Europe," Economic
Policy, October 1996, 23, 319-72.
European Economy, "Performance of the EU
Labour Market: Results of an ad hoc Labour Market
Survey." Reports and Studies, No. 3, European
Commission, Directorate-General for Economic
and Financial Affairs, B-1049, Brussels,
1995.
Freeman, R. B., "Are Your Wages Set in Beijing?,"
Jourmal of Economic Perspectives, Summer
1995, 9:3, 15-32.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms
74 Journal of Economic Perspectives
Jackman, R., R. Layard, and S. Nickeil, "Combatting
Unemployment: Is Flexibility Enough?"
Centre for Economic Performance, DP No. 293,
London School of Economics, 1996.
Jackman, R., R. Layard, M. Manacorda, and B.
Petrangolo, "Skills Mismatch: The Race Between
Demand and Supply," mimeo, Centre for Economic
Performance, London School of Economics,
1996.
Layard, R., S. Nickeli, and R. Jackman, Unemployment:
MacroeconomicPerformance and the Labour
Market. Oxford: Oxford University Press, 1991.
Lazear, E. P., "Job Security Provisions and Employment,"
Quarterly Journal of Economics, 1990,
105:3, 699-726.
Meyer, B. D., "Unemployment Insurance and
Unemployment Spells," Econometrica, 1990, 58:4,
757-82.
Narendranathan, W., S. Nickeli, and J. Stern,
"Unemployment Benefits Revisited," Economic
Journal, June 1985, 95, 307-29.
Nickeli, S. J., "Unemployment and Wages in
Europe and North America." Leverhulme Trust
Programme on the Labour Market Consequences
of Technical and Structural Change,
Discussion Paper No. 6, University of Oxford,
1996.
Nickell, S. J., and B. Bell, "The Collapse in
Demand for the Unskilled and Unemployment
across the OECD," Oxford Review ofEconomic Policy,
Spring 1995, 11:1, 40-62.
Nickell, S.J., and B. Bell, "Changes in the Distribution
of Wages and Unemployment in OECD
Countries," American Economic Review, May 1996,
86, Papers and Proceedings, 302-8.
OECD, Employment Outlook. Paris: OECD, 1990.
OECD, Employment Outlook. Paris: OECD, 1991.
OECD, Employment Outlook. Paris: OECD, 1993.
OECD, Jobs Study: Evidence and Explanations.
Paris: OECD, 1994.
OECD, Employment Outlook. Paris: OECD, 1996.
Phelps, E. S., "A Program of Low Wage Employment
Tax Credits." Russell Sage Foundation
Working Paper No. 55, 1994.
This content downloaded from 194.228.79.113 on Fri, 13 Apr 2018 11:08:32 UTC
All use subject to http://about.jstor.org/terms