Hey there folks! Today, I’ll be examining some economic theory from this paper by Jonas Agell and Kjell Erik Lommerud from 1993. It models solidaristic wage bargaining (SWB): the practice by sectoral unions of bargaining pay equality not just within or across firms, but across industries and sectors as well. This practice was especially strong in the height of social democratic power in the Nordics in the 50’s to 80’s, and brought tremendous gains. My own notes will be in brackets [like this]. Let’s get started!
Egalitarianism and Growth
Post-World War 2, Scandinavia had tremendous growth coupled with very low unemployment. Strong, centralized unions [union membership in Sweden in Denmark was around 95%] had tremendous wage-setting power. Yet traditional economic theory suggests this equalized wage-setting should lower efficiency, productivity, and employment. Much of this theory neglects the difference between firm-specific and sectoral unions: the latter, by encompassing many firms and even multiple industries, can factor in external effects of wages into the wage-setting process. We’ll examine the effect of wages on growth here.
To quote trade union economist Gösta Rehn, labor is not “like mercury, requiring only small level differences between two areas in order to float quickly, and in large quantities, from one of them to the other”. What that means is labor will not easily flow from old, less productive industries to new, more productive ones in the free market. As such, new industries will need to pay higher wages to attract workers. This leaves old industries at an advantage, because they don’t have to pay these higher wages. We’ll call the difference between these wages the wage premium.
II. The Basic Model
Our model in this paper has 2 sectors that produce goods, a traditional and a modern sector, both with perfect competition. The model assumes that workers always prefer to work in the traditional sector, all else (especially wages) being equal. This is of course a strong assumption, that should if anything work against SWB. But it can be partially justified by the additional training costs that could be needed to enter the modern sector, or the pain of having to switch jobs, including moving. This preference is the origin of the wage premium. While economics often focuses on capital that is slow to move to new industries, this model interestingly centers on labor that is slow to move to new industries! We assume the existence of “learning-by-doing”: as more work is done in the modern sector, workers become more efficient, simply through practice and minor innovations. This learning speeds up if more people are employed in the modern sector.
III. Competitive Wage Premia, Growth, and Distribution
Labor in the modern sector must be paid a wage premium, but this is not true for capital. Since the cost of labor is higher in the modern sector, less of it will be used, and more capital will be used. Here we see a core reason why the wage premium is bad: it encourages higher capital usage and lower labor usage in the modern sector, yet more labor usage would increase productivity and growth. The conclusion from this model is that the free-market equilibrium, which leaves a wage premium, is not efficient. One more interesting note: a higher capital intensity in the modern sector drives inequality compared to if the wage premium was zero, because (under capitalism at least) capital ownership is very unevenly distributed. Even if we look to the far future in this model, “though the growth rate of wages eventually catches up to that of capital, the inequality gap created during the transition will not be closed”.
IV. Solidaristic Wage Policies
The argument for SWB flows naturally from this result. If the wage premium harms productivity and growth, it must be ended. That is, wages in the traditional sector must be forced to be about equal to those in the modern sector. Of course, many in the traditional sector likely can’t afford this level of pay over time, compared to the increasingly productive modern sector. This means low-productivity firms will be forced out of business, pushing workers towards the modern sector. Workers should be, and were in the height of social democratic power, assisted and compensated in this transition through active labor market policies (ALMP) to train folks and generous unemployment insurance are needed. Crucially, if ALMP works as intended, every worker who loses employment in the traditional sector will gain a job in the modern one, meaning that SWB is perfectly compatible with full employment. And the data confirms this: in 50’s and 60’s Sweden, for example, unemployment averaged 1.5%.
V. Omissions and Extensions
This simple model does not take international trade into account. If you do, the case for SWB grows stronger. In an open economy, there’s more of a risk of an economy with free-market wage setting getting “trapped” in traditional employment by specializing in it. Meanwhile, when taking mobility of labor and capital into account, the case for SWB may be weaker. The authors suggest inflows of immigrants triggered by the promises of higher wages could lower wages [though more modern economic research suggests otherwise]. By pushing out low-productivity firms and preventing high capital intensity in the modern sector, SWB could also cause the exit of capital. This may explain why Scandinavian countries at the time often had capital controls to restrict movement of capital.
VI. The Demise of the Scandinavian Model?
“In Sweden, the highly centralized wage bargaining system dissolved in 1983”, though it maintains somewhat centralized bargaining to this day. In Norway in the 1980’s, the main union confederation was challenged by two competitors, undermining the ability to centrally negotiate wages. An important lesson is that politics matters here. SWB being good economic policy does not guarantee its existence: you must have high union membership, cohesion among the union movement, and tight ties between unions and government to maintain proper ALMP and wage-setting. How to get there is a crucial question, but beyond the scope of this paper.