Report finds that wage rigidity remains high, even during high unemployment

There is "no evidence that the increase in unemployment has softened the mechanisms generating wage rigidity" in Sweden - in other words, pay has not fallen despite an unemployment rate of some 10%. This is among the main findings of research published in June 1999, based on 1991 and 1998 surveys of manufacturing companies conducted by two Swedish economists, Jonas Agell and Peter Lundborg.

A report entitled Survey evidence on wage rigidity and unemployment: Sweden in the 1990s was presented on 29 June 1999. The study is based on two surveys, one conducted in 1991, the other in 1998, aiming to explore among managers from 157 companies in the Swedish manufacturing industry how a severe macroeconomic shock affects wage rigidity and unemployment. The research was carried out by two economists, Jonas Agell and Peter Lundborg, and funded by the Swedish government's Office of Labour Market Policy Evaluation (Institutet för arbetsmarknadspolitisk utvärdering, IFAU). In 1998, when the second survey was conducted, the unemployment rate was much higher and the inflation rate much lower than when the first survey was carried out in 1991.

Comparing the situations in 1991 and 1998, the authors conclude, in short, were that there is no evidence that the increase in unemployment "softened" the mechanisms generating wage rigidity. On the contrary, because of severe downward nominal wage rigidity - ie wages in cash terms never tend to fall - real wages have become more rigid during Sweden's move to a low-inflation environment. The authors have unearthed new evidence on issues such as "underbidding" - ie workers outside a company seeking employment at lower wages than those paid to the company's employees - employment security legislation, and the extent to which long-term unemployed people may be subject to discrimination in recruitment.


From 1945 to 1990, the unemployment rate never exceeded 4% in Sweden. However, between 1991 and 1993, GDP fell by more than 5% and total unemployment, including those enrolled in labour market training and employment programmes, increased from almost 4% to more than 12% of the labour force. The rate of job destruction was particularly high in the manufacturing sector, where between 1991 and 1993 the number of employees decreased by 22%. Inflation fell from 10% in 1991 to almost 4% in 1993, reaching zero in 1996. By the time of the second survey interviews with managing directors and staff managers in late 1998, inflation was still very close to zero and unemployment remained high, at more than 10%.

Wage rigidity has been a recurrent theme among economists during the 20th century. Nearly 70 years after the famous economist John Maynard Keynes took the position that wage rigidity was a social fact of life, economists still argue about the subject. There is still little consensus over the substantive mechanisms, the report's authors observe.

Does unemployment lead to wage cuts?

Has high employment made employees and local unions more willing to relax their wage claims? Does the labour market function less well when inflation is very low, since "downward money wage rigidity" - in other words, the tendency for wages in cash terms not to fall - prevents real wages from falling in times of slack capacity? These were two of the questions that the authors of the IFAU study brought up for discussion with the management of the 157 manufacturing companies surveyed. Specifically, they asked the firms whether at any time during the 1990s, covering a period of six to seven years of very low inflation and high unemployment, they had reduced nominal wages.

Out of 153 responding firms, only two had experienced cuts in wages. In one firm with several hundred employees, two office clerks were given reduced pay by being reassigned to less qualified duties. The other firm renegotiated and lowered the piece rates for some of the employees, as a part of a general revision of its pay system.

"This deeply rooted nominal wage rigidity", as described in the report, arises for several reasons, according to the two economists. One is an absence of "underbidding". In the survey, firms were asked if they presently had external job applicants who offered to work for less than the going wage for employees with the same qualifications and experience in their company. In 1991, such underbidding was found to be not at all uncommon, with 10% of respondents replying that they presently encountered underbidding blue-collar workers, and 13.6% reporting underbidding white-collar workers. In 1998, the researchers expected to find that underbidding had become more common. However, they found quite the opposite. In spite of much higher unemployment, wage competition had become much less common. Only 3% of the firms reported that they presently encountered underbidding blue-collar workers, and 5.2% reported underbidding white-collar workers.

"It is easy to think of reasons why wage competition ought to be uncommon in a welfare state", the two economists argue. Due to unemployment benefits and income-dependent taxes and transfers, many unemployed Swedes gain relatively little from acquiring a job. However, this does not explain why underbidding became less intense between 1991 and 1998. It is suggested that the answer must be that the circumstances of a severe macroeconomic shock discourage active job-searching. During and shortly after a period when companies reduce their workforces, applying for a new job might not appear to be a worthwhile activity.

The survey also contained questions about the recruitment policies of the companies. In both 1991 and 1998, around 20% of the respondents were of the opinion that job applicants who had undergone long-term unemployment were likely to be less productive than others. Unemployment's "stigmatising" effect on people was found to be a little weaker among those that had taken part in a labour market programme.

Employment security legislation

What influence has the Employment Protection Act (lagen om anställningsskydd, LAS), which seeks to give employees greater job security, had on companies' decisions on new recruitment and on redundancies? As many as 36% of the managers interviewed in 1998 answered that the risk of losing competent, newly engaged personnel because of the "last-in, first-out" seniority rules in the LAS made them reluctant to issue notices of redundancies during a downward economic trend. A majority of managers - 57% in 1991 and 55% in 1998 - stated that they preferred not to employ new people during an upward economic trend, because of the LAS, and would rather let their existing workers work overtime.

It thus seems that Sweden's employment security legislation tends to level out the variations in employment levels during upward and downward economic trends. In 1998, 58% of the managers indicated strong support for the proposition that employment security legislation made their firm more prone to offer flexible short-term employment contracts. In 1991, this figure was 34%.


"We find no evidence that high and persistent unemployment has softened the mechanisms generating wage rigidity," the two authors conclude, adding that "in 1998 employees seem as determined to resist wage cuts as in 1991, even at the cost of additional - and substantial - job losses." There was a near absence of nominal wage cuts during a prolonged period characterised by very high unemployment and low inflation. The authors trace "this pervasive nominal wage rigidity to the compound influences from various legal provisions of the wage contract, and from employees? concerns about relative income and fairness". Certain aspects of the "safety net" provided by the Swedish welfare state are also claimed to aggravate wage rigidity. The solid "wage floor" - ie the fact that nominal wages will not fall - seems to suggest that Sweden's move to a low-inflation environment have been accompanied by increased rigidity of real wages, the authors conclude.

The report also conclude that many managers view people with an unemployment record as being less productive. Finally, Sweden's stringent employment security legislation is claimed to be an important reason for the weak recovery of demand for labour after the initial macroeconomic "bust" of the early 1990s. (The issues of "efficiency wage mechanisms" and workers' wage norms are also discussed in the report, but this debate is not covered in this article.)


On the issue of employment security, the survey indicates that the Employment Protection Act functions in the way that was intended. In a period of swiftly changing economic trends, the Act has an important stabilising effect on the rate of unemployment. Companies hesitate to issue notices of redundancies to some of their staff, being frightened of losing the most recently recruited employees and workers with a high level of skills. Seen from the point of view of some employers, however, their claims for more flexibility have been partially satisfied, with new forms of "weaker" recruitment now allowed, no longer involving the full LAS commitments. (Annika Berg, Arbetslivsinstitutet).

Useful? Interesting? Tell us what you think. Hide comments

Eurofound welcomes feedback and updates on this regulation

Add new comment