Printer Friendly

Gaps in monitoring wages and industrial relations.

Gaps in monitoring wages and industrial relations

In recent years, academic researchers and other users of Bureau of Labor Statistics' data on wages and industrial relations have become increasingly concerned about the future availability of such information. The Industrial Relations Center Directors--an informal group of more than 60 university research programs--protested impending budget cutbacks at BLS in 1982.(1) Although the worst of the budget problems that befell BLS are past, issues about priorities still remain.

Influence of macro-economics

Following World War II, macro-economic policy came into ascendency. Policy makers needed aggregate indicators of unemployment, productivity, labor costs, and inflation. BLS was able to accommodate these needs, while also expanding its offerings of traditional wage and industrial relations data. In retrospect, the late 1970's were a golden age in which the two needs--macro and micro--both received adequate funding. But when the budget pressures of the early 1980's developed, a "revealed preference' for the macro side became apparent. The traditional price series were protected, a program of import and export price indexes was expanded, and productivity measures were refined. Those wage and industrial relations data which were macro-oriented were preserved and expanded, but micro-level indicators were cut back or eliminated.

The macro-policy influence is clearly illustrated by the development of the Employment Cost Index (ECI) in the mid-1970's. Through the 1960's, hourly and weekly earnings data from the establishment survey were the prime measure of wage costs available from BLS. These data covered only production and nonsupervisory workers and omitted fringe benefits. They were affected by shifts of employment between industries and occupations and by changes in the mix of overtime and regular hours. For econometricians interested in aggregate wage-change equations, these deficiencies were unfortunate.

One solution was to use the more comprehensive measure of hourly compensation which included all occupations and fringe benefits. But this index, too, suffered from employment shift and overtime effects. Initially, the BLS offered its hourly earnings index (HEI) as a partial solution. The HEI controlled for interindustry shift and overtime effects in manufacturing. But the more-refined Employment Cost Index (ECI) paints a different picture of wage trends than any of its predecessors.

The total compensation ECI shows a lower peak wage inflation rate in 1980 than the more volatile compensation per hour index and a higher peak than the indexes which omit fringes. It also shows a higher rate of wage inflation by 1983 (after the economic slump had taken its toll) than the alternative indexes. With the addition of public-sector data in 1982, the ECI is the best macro indicator of wage change available.

Series abandoned

Prior to the ECI, the only time series available with a union/nonunion cut was a series on wage developments in manufacturing (WDM). But this series was seriously flawed. In the nonunion sector, the omission of "merit' pay adjustments was known to bias its estimate of wage inflation downward. But it also apparently underestimated union wage increases. Because the series covered adjustments in small union units as well as the "major' (1,000 workers or more) agreements, it created the impression that "minor' union agreements were not keeping up with their major counterparts. After the ECI became available, this impression was contradicted.

Given its inaccuracies, it is not surprising that BLS abandoned the wage development series after 1978. But the series did offer information on the dispersion of wage decisions at the micro level not available from the ECI. For the nonunion sector and smaller union bargaining units, lack of dispersion information is an important gap in monitoring wage developments.

The abandonment of the wage development series was based on its deficiencies rather than on budgetary considerations. But other wage series, particularly in the union sector, were dropped because of the budget crunch. And it could not be said for these that superior alternatives had become available. For example, the now-abandoned series on union wage-rate changes in construction can be compared with the still-available series on effective wage adjustments in "major' construction union agreements. During the latter half of the 1970's, construction wage settlements went through a period of comparative moderation after two earlier wage explosions. A comparison of the two series indicates that the wage moderation was more dramatic in the agreements covering relatively small numbers of workers. Construction has been a center of concession bargaining in the 1980's, but now it is impossible to make such comparisons with BLS data.2

Also lost during the crunch was the wage chronology series. It provided useful information on wages and other conditions in selected union-employer settlements. As econometricians became more interested in the micro side of wage decisions, the chronologies were used to provide insights not available from aggregate Phillips curves. Without the chronologies, researchers must use the original contracts (not always easy to obtain retroactively) or other less-detailed sources such as Current Wage Developments. Research efforts--in short--have been and will be impeded.

As of 1980, almost 8 of 10 private-sector wage earners were not in unions. Thus, if any criticism could be leveled at the wage chronology series, it would be for the neglect of nonunion companies. Research interest in the personnel practices of large, nonunion firms grew in the 1970's. Thus, a widening of the chronologies to include such employers-- rather than their abandonment--was indicated.

Further curtailments

Collection of data on strikes dates back to the late 19th century. Regular (annual) surveys of such information began in the World War I period. The data gathered were not limited to aggregate tabulations. Detailed tables were available by industry, issue of the dispute, means of settlement, and so on. In 1982, however, reporting was cut back to disputes involving 1,000 workers or more and detailed analyses were eliminated. Limiting coverage to disputes involving 1,000 workers or more is in keeping with the macro emphasis.

Abandonment of comprehensive strike surveys has caused a loss of information which--unlike the wage chronologies --cannot be retrieved retroactively. Using the Current Population Survey (CPS) as a substitute source is not satisfactory. CPS estimates of individuals not at work, or forced to work part time, due to an industrial dispute, fell well below the prior work stoppage survey's estimates. Moreover, the CPS sample is too thin to provide industrial detail and contains no information on the issue of the dispute or the other information categories previously collected.

Also with the budget crunch, BLS dropped its union membership survey. The Bureau first published union directories in the 1920's, and during the post-World War II period, substantial statistical detail on union membership was added. Because the data were based on claimed membership, their accuracy was questioned. In 1980, for example, the CPS estimate of labor organization membership was 20.1 million compared with a claimed membership of 23.9 million. However, the claimed membership data provided the only tabulation of membership by organization.

In addition, no CPS data on union membership have been published since the May 1980 survey. Fortunately, the Bureau of National Affairs has maintained part of the directory, but statistical detail has been lost.3 Ironically, this loss of information came at a time when union membership fell dramatically. BLS' own estimates of the number of workers under major private union agreements fell from 9.3 million in 1979 to 7.9 million in 1983. Thus, at a critical period for the collective bargaining sector, an important data source was dropped.


1 Discussions of the IRRA Executive Board are reported in the Proceedings of December 1981 and 1982. The Board considered a resolution urging continued statistical service in industrial relations at BLS and other agencies. Although the Board voted to approve the resolution by 11 to 3, no official action was taken due to opposition by management members. The Industrial Relations Center Directors' letter appears in the May 1982 IRRA Newsletter.

2 Related specialized wage series in other industries were also eliminated.

3 Courtney D. Gifford, ed., Directory of U.S. Labor Organizations, 1984-85 edition (Washington, The Bureau of National Affairs, 1984). Another directory has been advertised by Industrial Relations Data and Information Services but was not available at the time this paper was prepared. (After this paper was given, BLS released CPS-based estimates of union membership for 1983-84.)
COPYRIGHT 1985 U.S. Bureau of Labor Statistics
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1985 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Thirty-Seventh Annual Meeting of the Industrial Relations Research Association, Dallas, December 1984
Author:Mitchell, Daniel J.B.
Publication:Monthly Labor Review
Article Type:transcript
Date:Jun 1, 1985
Previous Article:Productivity trends in the machine tool accessories industry.
Next Article:Innovative approach to plant closings: the UAW-Ford experience at San Jose.

Related Articles
Airline union concessions in the wake of deregulation.
Labor-market data: supplementary sources.
BLS compensation programs: what will users need?
Jobs guaranteed at Cummins.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters