Forthcoming Review of Economics and Statistics, 1999.The Effect of Collective Bargaining Legislation on Strikes and WagesPeter Cramton, Morley Gunderson, and Joseph Tracy*University of Maryland, University of Toronto, and Federal Reserve Bank of NY25 May 1999AbstractUsing Canadian data on large, private-sector contract negotiations from January 1967 to March1993, we find that strikes and wages are substantially influenced by labor policy. The data indicate thatconciliation policies have largely been ineffective in reducing strike costs. In contrast, general contractreopener provisions appear to make both unions and employers better off by reducing negotiation costswithout systematically affecting wage settlements. Legislation banning the use of replacement workersappears to lead to significantly higher negotiation costs and redistribution of quasi-rents from employers tounions.*We have benefitted from comments by two referees and seminar participants at Johns Hopkins University,Stanford University, the University of Maryland, the University of Montreal, and the University ofWisconsin. We are particularly grateful to John Budd of the University of Minnesota for helpfulsuggestions and for providing us his coding of policy variables based on an examination of the originalstatutes. Financial support is gratefully acknowledged from the National Science Foundation for Cramtonand Tracy and a Social Sciences and Humanities Grant for Gunderson. Most of this work was completedwhile Cramton was a National Fellow at the Hoover Institution, Stanford University, and Gunderson wasvisiting the National Bureau of Economic Research, Stanford, California. The views expressed in thispaper are those of the authors and are not necessarily reflective of views at the Federal Reserve Bank ofNew York or the Federal Reserve System.JEL Classifications: D82, J52, J58IntroductionIn this paper, we investigate the effect that collective bargaining legislation has on strike incidence,strike duration, and wage outcomes. From a policy perspective, it is important to assess whether laborlegislation has its intended effects on bargaining outcomes. Often, collective bargaining legislation isenacted to improve the efficiency of negotiations, for example by reducing the incidence and duration ofstrikes. We evaluate the extent to which these policies are effective at reducing strike activity. Variations inthe bargaining rules as determined by labor legislation are also likely to influence the balance of powerbetween the union and the firm, and hence affect wage settlements. Indeed, the primary motivation ofadvocates of a particular legislative initiative may be distributional, even when their stated intentions are toimprove efficiency. For this reason, we analyze the wage effects and potential welfare consequences of thevarious labor policies.We utilize a data set that includes all private-sector contract negotiations in Canada involving 500or more workers from January 1967 to March 1993. Canada is a good setting to carry out this type ofstudy. First, its labor law is determined mainly at the provincial level, rather than the national level. Hence,there is variation in labor policy both among the provinces at a point in time and within provinces overtime. Second, Canadian data are systematically collected through mandatory reporting requirements whichshould increase data reliability.We find that conciliation policies have been largely ineffective in reducing strike costs. In contrast,general contract reopener provisions appear to make both unions and employers better off by reducingnegotiation costs without systematically affecting wage settlements. Legislation banning the use ofreplacement workers leads to higher negotiation costs both by increasing the frequency and duration ofstrikes. In addition, replacement bans also result in significantly higher real wage settlements. Theestimated average cost to employers from a replacement ban is around $4.75 million (1993 Canadian)dollars per contract.We present our theoretical perspective and a description of the key Canadian collective bargaininglegislation in the next section. In section 3 we describe our data more fully. Section 4 discusses the effectsof labor policy on strike incidence, strike duration, and wages. Section 5 compares our results to the earlierliterature. In section 6, we calculate the welfare implications of bargaining legislation for unions andemployers.2 Canadian Labor Legislation – A Theoretical PerspectiveSee Kennan and Wilson (1989, 1993) for surveys of bargaining theory and its relation to strike1data.2In this section, we briefly outline the structure of a bargaining model of labor negotiations. We usethis model to help frame predictions for various Canadian collective bargaining legislation. For each type oflegislation, we describe the nature, intent, and likely effect on strike activity and wage settlements.2.1 A Precis on Strategic Bargaining ModelsA detailed exposition of our theoretical wage bargaining model appears in Cramton and Tracy(1992, 1994a). The union and firm are negotiating the wage to be paid over the life of the contract. The1firm has private information about its willingness to pay. Offers are exchanged until a settlement isreached. Prior to settlement, the parties receive their threat payoffs. In the simplest model, the union decidesbetween two possible threats — the union can strike or it can apply pressure on the firm while continuingto work under the expired contract, which we call holdout. The threat payoff to the union under a holdout isthe wage from the expired contract.The settlement wage consists of the union’sthreat payoff plus a share of the inefficiencies inducedby the threat (the “pie”).The union strikes if the higher bargaining costs associated with a strike are morethan made up for by a higher expected settlement wage. If the wage under the expired contract issufficiently high, this is not the case and the union prefers the holdout threat. Policies that reduce the size ofthe pie should reduce settlement wages, while policies that increase the attractiveness of the strike threat tothe union should increase settlement wages.Dispute activity (strikes and holdouts) occurs because of the union's uncertainty about the firm’swillingness to pay. The firm has an incentive to falsely claim that times are tough and therefore a lowsettlement wage is appropriate. The firm’swillingness to endure a costly strike or