Tuesday, November 22, 2011

Forms of employment, unions and wages

Loath as I am to disagree with an economic historian as eminent of Peter Temin, his paper The Great Recession in Historial Context (pdf) has a claim about wage stickiness and its source I disagree with.

In explaining the development of stickiness of wages and the rise of unions. Temin writes:
As the size of production units, whether mines or factories, became larger, the ability of labor markets to be optimally competitive also diminished. Large employers yielded little bargaining power to workers to negotiate wages and working conditions. If a factory, for example, was the only large employer in town, the options for workers were even more limited and the market power of the employer more obvious. Workers formed unions to countervail the market power of employers, and wage bargaining and strikes supplanted the individual wage negotiations implicit in Hume’s and Smith’s analyses.
This is a wonderful (indeed popular) “just so” story. The trouble is, it is clearly wrong. Large employers tend to pay more than small employers, just as large supermarkets tend to be cheaper than corner stores. Size does not equal market power and does not determine comparative wages or prices.

There is a much simpler reason why unions arose in response to large employers. The workforces of large employers are easier to organize. The rate of unionization increases with the size of the employer (hence the public sector is far more unionised than the private sector) because the bigger the employer, the easier its to organize the employees—they are easier to identify, collectively talk to and have more commonality of interests. Moreover, the power of unions comes from their ability to exclude competing workers. The true enemy of a union is not the company, it is the “scab”, the non-unionised competing worker. The more centralised the workplace, the easier to exclude.

Similarly, wages are “sticky” outside unionised workplaces and across employers regardless of size. The “stickiness” comes from the labour relations being across time periods and asymmetric information. Reliable workers who understand how the firm operates are worth keeping, are valuable. Massively undermining their status as bargaining agents—and your own reliability as a bargaining agent—by unilaterally cutting wages is not the way to have a good relationship with your employees. Particularly given they have obligations set in money terms, so cutting their money income makes their situation worse regardless of what money prices are doing generally. Nevertheless, that money is how contracts “keep score”—so go directly to employee status as bargaining agent and employer reliability as bargaining agent—is even more important.

Moreover, with the development of extensive regional, national and global markets, and increased complexity of products, it becomes harder to workers to judge employer claims. A medieval peasant could see how good the harvest was, and could observe grain prices, so variability in income was much more manageable because far less trust was involved. A modern employee has far less information to directly observe about inputs and outputs in what they produce. That leads to more emphasis on what workers can judge as “proxies” for what they cannot. The reliability of employer behaviour, the respect (or lack) of employee status as bargaining agents has to be increasingly relied upon in an ongoing interaction (a repeated game, if you like).

It is not the size of the company that determines “stickiness”, but the form of the employment contract. “Spot” markets in labour allow much more flexibility in wages since there is no ongoing relationship. It also provides an example where unionisation provided large gains to (some) workers—the unionisation of the waterfront. But that is a case where unionisation changed the form of the labour contract. It is also a case where exclusion of competing labour is particularly intense—employment on the Australian waterfront, for example, has practically become hereditary.

It was not unionisation, but changes in the forms of labour contracts, in the structure of labour relations so that labour became much more an across-time interaction with increased information asymmetries, which increased the “stickiness” of labour. Unions are a symptom of that change far more than they are a source of wage “stickiness”.

Contemporary unions confront a range of problems. First, with the growth of two-income households, variation in income became rather less of an issue for many households, so workers become more willing to shift to forms of labour provision not susceptible to unionisation. Second, the increase in incomes and growth of regulation and other government interventions has meant that legal mechanisms (lawyers) and political ones (politicians and media) became increasingly competitive to unions as bargaining mechanisms. Third, the interests of unions is to make employment remuneration as complex as possible—both because that increases the need for a bargaining agent and because that provides various “victories” for unions to trumpet. The problem is that such a strategy increasingly generates wasted resources that can be harvested by moving to different forms of labour provision or contractual arrangements. Just as it was not employer market power which drove the rise of unions, nor is it driving their decline.

As to why large corporations tend to pay more than small employers, consider why large supermarkets have lower prices than the corner shop. The corner shop is, indeed, the corner, that is local, shop. A large supermarket has to make it worth your while to go that extra distance. Once you have decided to travel further (typically drive) to shop, then it is competing with all the providers in reasonable driving distance. It offers range and low prices to compensate for more travel time and less personalised service.

A large corporation finds it easier to spread/manage risk than a small employer but harder to tie worker effort to productive outcome. So, it pays a “hostage premium”—more than the employee can get elsewhere so that they police themselves more, as they have more to lose. This “hostage premium” is not merely basic salary; it includes a range of benefits and common activities to try and encourage self-policing. So, even in the absence of a unionised labour-exclusion premium, wherever the corporation finds it hard to tie employee effort to productive outcome, we can expect a “hostage premium” to encourage self-policing.

This does not explain what we observe of CEO pay, however. Tying the pay of CEOs to performance should be a lot clearer than executives further down the corporate hierarchy. Yet, what we observe is pay rates that seem unconnected to performance. A (very high) premium that is apparently often not hostage to productive behaviour.

So, consider the mechanism that selects pays for CEOs. In political science terms it is like rigged election autocracies. What we get is an "insider's game": insiders agree that you should be rewarded for being an insider, a game they all hope to benefit from so seek to maximise the return for being an insider. Benchmarking just increases the “gaming” (pdf), since it just increases information to insiders without increasing effective accountability since it does not breach the insider dominance of such decisions: the problem is not information asymmetries, it is insider privilege. The real puzzle is not why CEOs are paid so much, it is why their compensation can be notoriously unconnected to actual performance.

Forms of employment (including expected length of interactions), degree of centralisation of workplaces, information assymetries, insider privilege/outsider exclusion: they explain a lot more of how labour markets work than alleged employer market power. Including wage stickiness and the rise and decline of unions.


  1. Typo in 3rd paragraph … "Large employees tend to pay more than small employers, ". I think you meant "Large employers", unless you are suggesting that physical girth of the individual is a factor here.