Statliga ingrepp i lönebildningen och löneåterhållsamhet

Den senaste studien av efterkrigstidens påstådda löneåterhållsamhet -- kanske mest inflytelserikt teoretiserad av Charles Maier (1981) och Barry Eichengreen (1999a, 1999b) -- är av Chris Minns och Marian Rizov och lite förvånande publicerad i Business History. Deras teoretiska referenser är väldigt standard: Eichengreen. De säger att vi inte har så mycket empiriska belägg för ifall tripartism och andra former av statlig intervention ledde till löneåterhållsamhet. Wallerstein visade att centraliserad lönebildning korrelerade med lönespridning, men med föga utforskande av instituioenrna. Traxler och Brandl har kollat på en senare period. Brandl har i forskning centrerad på institutioner kollat på effekter på lönenivåer och arbetsmkostnader "rather than the key dimensions of wage moderation and wage dispersion". Vidare så är en metodologisk fråga hur man ska kunna separera ut effekter av centralisering -- som rör sig långsamt och därför uppvisar föga variation inom länder, jfr detta inlägg -- jämfört med country fixed effects. De kollar på perioden efter 1970 och framar uppsatsen så här:

"We focus on the effects of bargaining institutions on wage moderation and wage dispersion, which allows our results to speak to the debate between Eichengreen and co-authors who argued that new bargaining methods led to wage moderation, versus scholars of industrial relations who claimed that centralised unions were able to secure larger wage increases for their members." (s 2)

Om varför staten är viktig refererar de Anke Hassel och Jelle Visser.

Centraliserad lönebildning kan ha två olika effekter på löneutfallet, säger de. Å ena sidan kan den öka internaliseringen vilket minskar lönetrycket. Å andra sidan kan den öka fackets möjlighet att plocka ut löneökningar. Så här förklarar Minns och Razov det:

"High level (often national) unions and employer organisations may internalise inter-firm or inter-industry tradeoffs that are external to negotiating parties under fragmented bargaining. Once wages are agreed, the centralised nature of labour market institutions is key to achieving wage moderation, by preventing shirking on agreements by either workers or firms. On the other hand, decentralised bargaining may bring about greater wage discipline in unions for whom higher wages will reduce demand for labour in the firms in which they are active. A further counterpoint is that centralised unions may be able to secure higher wage settlements if the effect of collective arrangements is mainly to increase the bargaining power of organised labour"

Calmfors och Driffills U-kurva är logisk utifrån detta, säger de. (s 3)

Flanagan har rest frågan ifall löneförhandlingar påverkas av kultur och sociala normer som har djupare historiska rötter. Här knyter de an till Maiers (1984, i Goldthorpe red Order and Conflict in Contemporary Capitalism) forskning om 1920-talet och Panitch (1977, "The development of corporatism in liberal democracies") som hävdar att efterkrigstidens korporatistiska uppgörelser grundades i 1930-talets diskussioner om massarbetslöshet.

"Early examples of successful social compacts (including widespread wage controls) are seen in smaller nation-states (particularly in Scandinavia), for which export production was key to economic performance, and where social democratic parties generally had a presence prior to 1945. In West Germany, organised labour accepted wage restraint due to the primacy of reconstruction and postwar recovery that was in the interest of their members in the long-term. In short, the above explanations suggest that there may be a link between historical events and current labour market outcomes." (s 4)

De gör två hypoteser. 1, statlig inblandning ökar löneåterhållsamheten. 2, vad som hände i den tidiga efterkrigstiden påverkar senare löneåterhållsamhet.

De använder data från 1969 till 2008. Deras källa för de viktiga institutionella variablerna -- centralisering av lönebildningen, statlig inblandning i densamma -- är Visser, som vanligt. Deras lönedata kommer från ILO och deras produktitivtetsdata från Groningen:

"Comparable international wage data is drawn from the databases of the International Labour Organisation (ILO). The ILO database provides wages for up to 10 sectors for most of the countries also covered by the labour market institutions database. Finally, to allow for an examination of trends in wage growth relative to productivity (a key point in any arguments regarding moderation, see Baccaro and Simoni), we use statistical evidence on aggregate labour productivity from the Groningen Growth and Development Centre (GGDC) 10-sector database."* (s 5)

Så här förklarar de sina löne-mått/beroende variabler:

"We focus on explaining two outcome measures that are in keeping with the existing (theoretical) literature on bargaining institutions. The first is the (real) wage–productivity gap, which we take as the ratio between wages and output per worker. This measure has obvious links to ideas regarding economic efficiency, as well as to the classic debate on wage moderation – if workers are accepting smaller wage increases so that firms can invest excess profits, this ratio should be smaller. The second outcome we examine is wage dispersion. Greater equity in wage outcomes is often an explicit aim of the actors engaged in wage negotiations, and ideally our results would speak to whether or not particular bargaining arrangements tend to achieve this aim. To compute the wage–productivity gap, we use sector-level wages for all countries and all years for which we have labour market data and evidence on bargaining institutions (as outlined below), and divide these by national output per unit of labour.30 We use the standard deviation of the sectoral wage observations in the country in each year as our wage dispersion measure.31 The wage data used refer to ILO series that combine male and female workers. We have much wider country coverage in observations that combine both genders than for only male or only female workers, but the data suggest similar patterns (and in particular similar regression results) if we extend coverage to all available data." (s 5)

Så här sammanfattar de data i en tabell:

För perioden mellan 1950 och 1960 använder de data på löneförhandlingar från Allard och Lindert (2006); man undrar ju varför de inte använder Golden Lange Wallersteins välkända dataset?

De kör några regressioner.

Deras slutsatser är:

"We find some support of our first hypothesis (H1) that tripartism and other forms of government intervention in the bargaining process have consistent, but modest effects on labour market outcomes. Wage moderation is greater in the presence of tripartism, with workers accepting wage increases that are lower relative to productivity growth, as compared to states without government involvement in bargaining. Wage dispersion is much less affected by tripartite bargaining, which suggests that the influence of government may have been primarily on overall wage levels, rather than targeting the pay of particular sectors.

Historical bargaining arrangements, measured here as the degree of corporatism in national labour markets in the immediate post-war settlement between 1950 and 1960, appear to have larger effects than current institutions – a finding strongly in support of our second hypothesis (H2). One explanation for this finding is that post-Second World War arrangements reflect labour market ‘culture’ in the early twentieth century, something that is observed to some extent in terms of international differences in work time and work sharing in the long-run. Alternatively, one could argue that the Second World War served as a form of structural break in the national labour markets under consideration. Subsequent changes in bargaining institutions matter on the margin, as our results indeed show, but expectations surrounding acceptable negotiated pay outcomes are strongly related to the institutional framework that operated when formal industrial relations first emerged as an important feature of many economies in the early 1950s." (s 12)

Fotnot
*I ett kort data-appendix i slutet förklarar de lönedata så här:

"Information on wages is drawn from the ILO’s LABORSTA database. We began by extracting wage information from 1969 to 2008 (when available for those years, as discussed below) for the following countries: Australia, Austria, Belgium, Canada, Denmark, Spain, Finland, France, Great Britain, Germany (West then unified), Greece, Ireland, Israel, Italy, Japan, Netherlands, Norway, New Zealand, Portugal, Singapore, Sweden, Switzerland, and the US. Of this initial selection, we retain all earnings series that consist of actual earnings or wages paid – we leave aside pay measures derived from the scales of collective agreements, minimum wages or salaries, and the like. In the next step, we convert all pay series into US dollar equivalents, using exchange rate data to convert from local currencies. The wage series are then converted to 2000 dollars using similar information about historical US consumer price indices. A final conversion relates to units of pay, which were recorded as earnings or wages per hour, day, week, or month. All series were converted to hourly in the work that follows. Visual inspection of the surviving series suggested that wages for some series were implausibly high or low, regardless of adjustments made for currency or time worked. Several series for Austria, Belgium, Canada, Germany, Finland, and the United States were discarded. All data for Italy was omitted due to the implausible wage numbers. The final task was to gather the wage series by sector. The indicators used by the ILO differ somewhat by survey methodology, but data was coded into the following groupings to the best of our ability: manufacturing, construction, trade, transportation, finance, community, primary, mining, energy, and ‘other'." (_s 17)

Och produktivitetsdata så här: "Labour productivity is from the GCDC. It consists of GDP per hour, expressed in 1990 Geary- Khamis US dollars."

Referens

Chris Minns och Marian Rizov, "Institutions, history and wage bargaining outcomes: international evidence from the post-World War Two era", Business History