Posts Tagged The Labour Market
Chapter 5 of Steve Keen’s Debunking Economics takes a look at labour market economics, an area where I feel there are almost too many different criticisms to know where to start. Keen spends a relatively short amount of time on this chapter, and though his expositions are sufficient to question many of the standard stories economists tell us about labour, I feel he misses some low hanging fruit.
Keen primarily builds on his earlier approach to demand-supply analysis. He first notes that the relatively uncontroversial backward bending labour supply curve, when aggregated, can mean a labour supply curve can have any shape at all – this aggregation issue is not fully explored on economics courses. He then applies his earlier analysis of demand curves, which implies the same for labour demand. The definitive PHD textbook , Mas-Collell, has also noted the latter, but assumes it away by supposing that:
…there is a benevolent central authority, perhaps, that redistributes wealth in order to maximise social welfare…
Naturally, this assumption isn’t true in reality, and since it’s a domain assumption – for which conclusions only follow as long as it applies (more on this in a later post) – we can safely ignore it. The result is that there may be any number of equilibrium points in a labour market.
On top of this, there are many more points that make the clear cut story of a single equilibrium highly questionable:
- The intertwined nature of work and leisure – the latter relying on the former if, as Keen puts it, you want to do anything more than sleep.
- The existence of nominal debt contracts.
- Keynes’ argument at the beginning of TGT, which is that workers cannot control their real wages, as it is dependent on price, which are controlled by their employers. Keen weaves around this argument at the beginning of the chapter but never states it explicitly.
- The fact that wages are an essential component of aggregate demand, and, similarly, that the labour market is so broadly defined that treating demand and supply as independent is not possible.
- The fact that, since markets do not approximate perfect competition, even standard neoclassical analysis teaches that minimum wages and unionisation can be beneficial to combat market power.
Clearly, there are a multitude of reasons that ‘interfering’ with wages through demand management, legislation and unionisation is not obviously a bad thing.
This is actually fairly uncontroversial stuff, and most neoclassical economists would endorse some of it. Having said that, many still think real wages should fall, so clearly haven’t fully thought the implications through.
But maybe Keen should have been more controversial. For me his criticisms don’t go deep enough – he repeatedly refers to reasons that “workers will not be paid their marginal productivity”, without questioning the concept itself.
As I have noted before, labour only has productivity when combined with capital. Shops assistants need a shop, tills, hangers and bags; builders need tools, machinery and materials. Furthermore, some labourers only have productivity when combined with other labourers, too. Two people carrying a heavy box cannot be said to have a discernible individual productivity. The productivity of a McDonald’s relies on the cooks, the till operators and the supervisors combined – take away a cook and what’s the point in having an extra till operator when the cooks can’t keep up with the orders?
The result is that productivity can only be said to be a result of combined, rather than individual, factors. The relative shares are then determined by bargaining power.
I consider this argument a strong one, and it also fits quite nicely with many of the Sraffian arguments throughout the book: that firms do not hold capital fixed and employ more labour – they need to employ both simultaneously; that much of neoclassical analysis cannot take place independent of income distribution, which is dependent on political power (more on this in a later chapter). For these reasons I’m guessing Keen is unaware of this particular criticism, rather than rejects it – it would significantly strengthen his critique, which, though sufficient to complicate the neoclassical story, does not completely ‘debunk’ it.
Simon Wren-Lewis discusses the large gap between mainstream and heterodox economics, and asks why the heterodox economists are so willing to throw out almost every aspect of neoclassical theory. Allow me to offer an explanation.
The reason heterodox economists remain dissatisfied with mainstream economics, no matter how many modifications the latter adds to its core framework, is that there is always an implication that, in the absence of various real world ‘frictions’, the economy would function like a smoothly oiled machine. That is: assuming perfect information, mobility, ‘small’ firms, no unions, flexible prices/wages and so forth, the economy would achieve full employment, with near perfect utilisation of resources, and stay there, perhaps buffeted by mild external shocks.
New Keynesians and New Classicals sometimes act like bitter rivals, but mainly they only differ on which ‘frictions’ should be present or not (this is an oversimplification of the disagreement, of course). The original New Classical models started with economies that are always in equilibrium, preferences are constant, and competition is perfect. New Keynesian models add imperfect competition, sticky prices, transaction costs and so forth. The newest papers go further and add heterogeneous agents (which generally means two), changing preferences, and other ‘frictions.’ However, it is assumed that if the economy were rid some specific features/characteristics, it would function similarly to one of the core Walrasian or Arrow-Debreu style formulations.
So is it not true that real world mechanics prevent things from going as smoothly as they might do in absence of those mechanics? Well, partially. But according to heterodox economists, capitalism has inherent tendencies to crisis, unemployment and misallocation anyway.
A key example of where this is evident is finance. Generally the mainstream analyses of why finance is unstable focus on irrationality, imperfect information, externalities and other such modifications. If only everyone had access to information, if transactions were cost less, and if people were rational self maximisers, then finance would be stable.
Minskyites, on the other hand, argue that this isn’t the real problem. Even if the economy starts stable, the resultant strong returns on investments will cause capitalists/investors to take more risk. This process will continue and the economy will endogenously destabilise itself as higher returns are sought and more risk is taken on, until eventually the capacity to make a return on these risks is outrun and we face a collapse. There is no need to invoke a specific ‘friction’ for this process to occur.*
Another prominent example is the labour market. Generally, economists presume that without ‘search costs’, oversized firms/unions and sticky wages, the economy would achieve full employment. But heterodox economists disagree on a number of counts: the Marginal Value Product Theory is faulty, so higher wages will not necessarily cause unemployment to rise; wages are also an essential component of aggregate demand, so reducing them may well be counterproductive. In fact, Keynes argued that sticky wages were far from a barrier to full employment; they actually stabilised aggregate demand. Steve Keen’s model also produces less severe business cycles when sticky wages and prices are added.
So the reason heterodox economists want to throw the proverbial baby out with the bath water (and also redecorate the bathroom and possibly even move house, or something), is that they think the core of mainstream economics has dug itself too deep into a ditch. The inevitable ad hoc modifications of ‘perfect’ models sometimes have so many ‘frictions’ introduced that the supposed ‘deep’ mechanics that underlie them become questionable. But they are still never abandoned. Heterodox economics is not just about adding a few real world mechanics here and there; it’s about throwing out the entire core and starting over.
*It could be said that this might not occur if Knightian uncertainty were not a factor in the real world, but I think calling this a ‘friction’ jumps the gap between friction and fundamental reality.
The debate over libertarianism and the workplace (if you can call it a ‘debate’, when libertarians make responses like this, here is a summary of what Cowen and Tabarrok are saying) seems like as good a time as any to post on the bargaining power relationship between labour and capital.
I have posted before about how the idea that wages are determined by productivity is indefensible; capital and labour only have productivity when combined, so it is impossible to separate their relative contributions, which are instead determined by bargaining power. As Daniel Kuehn also notes, a ‘job’ is generally what is bargained over, rather than specific aspects. So it would not be unreasonable to say that working conditions, hours and pay are generally all determined by bargaining power, though not separately. It is also not unreasonable to say that employers generally have the edge in this. But why?
The first reason, noted by Paul Rosenberg, is that labour requires wages to subsist every day, whereas those sitting on capital can produce for themselves. This means that labour’s situation is generally more urgent than capital’s. Now, libertarians might respond that people can save money, inherit money, and so forth. But this begs a lot of questions: what if you are born poor? Where do you get your savings from initially, if not wages?
Libertarians also might respond, as the BHL libertarians have, by advocating a universal income (something that strikes me as trying to make the world behave like an economics textbook, where workers can smoothly trade off leisure for work, from 0 hours to 24). This would indeed improve labour’s bargaining power. However, it is also the case that, even under this system, many workers would incur obligations such as debts, families, and of course social obligations, that require money. Whether these people ‘choose’ to do this is irrelevant: what we are asking is if, at the moment somebody tries to get a job, they have more bargaining power than their employer.
The second reason is that employers are fewer than employees, making the latter more readily substitutable, particularly in low skilled jobs. This starts from the obvious observation that not everyone can be a capitalist. Since wages tend to be consumed, but profits don’t, it is fair to say that an increase in the amount of capitalists over workers will reduce consumption and therefore available profits. This will result in capitalists going bankrupt. Obviously, if there are too few capitalists then opportunities will also open up, and we will go in the other direction.
It is reasonable to conclude that there is a rough ratio of capital to labour around which the economy oscillates, something similar to what Phillips was actually saying with his ‘curve.’ Capitalism generally finds it hard to deal with true full employment, as it diminishes the capital available for investment. This results in lay offs, and diminishing bargaining power for labour. Historically, capitalism appears to spend a lot more time in period of unemployment than periods of full employment.
There is the final point that under modern capitalism, labour is free to organise and create collective bargaining power. However, in the absence of legislation to assist this, unionisation falls into all the familiar problems with collective action, problems that capital doesn’t have: coordination, aligning different interests, the incentive for individual members to cheat. This is reflected by the fact that countries with strong unions generally have legislative support of those unions, too.
Obviously I’ve been assuming that neither capital nor labour ‘hijack’ the state to further their own interests (questions over whether capitalism is a system characterised by capital’s hijacking of the state aside), but I don’t think it’s necessary to invoke these to understand why labour often seems to be on the losing side of the bargain, particularly for low skilled workers.
Bringing it back to the debate over libertarianism and the workplace, it’s worth noting that ‘voluntary’ versus ‘coerced’ is not a binary distinction but a spectrum, with one end representing virtually no costs for choosing something different, whilst the other represents death/torture. In between you can have anything from walking down the road to another shop to social pressure to moving country – all are costs of not taking a particular choice, and hence reduce the ‘voluntariness’ of the decision itself. If employers generally have more bargaining power, this is a reflection that the costs of them choosing another employee (or no employee at all) are lower than the costs of the employee taking another job (or no job at all). This means the spectrum is tilted further away from ‘voluntary’ for the labourers, and the mere axiom that they have agreed to it so it’s OK will not suffice.
The labour market, along with finance/banking and development, is one of the areas where neoclassical theory has proven to be most off the mark. Although more advanced models of the labour market try to address some of the flaws in MVP theory, they generally do this by adding ‘frictions’ such as heterogeneity and job search problems. The general premise – as with much of neoclassical economics- is that without these ‘frictions’, the labour market would operate as described in the textbooks. But this is not true.
Production, by definition, requires that all the factors of production be brought together. As a result, labour is employed at the same time as capital (and land, but we can lump the two together for the purposes of this post) in every circumstance. A taxi driver without his taxi is worthless, and vice-versa. Adding an extra worker to an office necessitates an extra computer, desk and stationary. Adding another builder to a site necessitates tools. Hence any ‘marginal productivity’ can only be applied to the labour and capital as a whole (what Ricardo called a ‘bushel’).
Not only this, but the Division of Labour (DoL) means that it is often impossible to separate the produce of one worker from that of his colleagues. A construction site requires carpenters, plumbers, bricklayers, supervisors, semi-skilled labourers, labourers and many more. But what if you remove only the carpenter? You wouldn’t be able to construct the house. But it would be incoherent to claim his MVP were an entire house – you can only evaluate the product of the entire team together, including their capital. So the marginalist approach makes absolutely no sense, particularly in a society where the DoL spans global borders.
This is also a prime example of an area where historical context in teaching would be appropriate, for the fact is that John Bates Clark created his ‘Marginal Value Product’ theory – the reasoning for which is entirely circular – at the turn of the century, a time of unrest among the working classes, in order to try and settle them. Basically, it was a justification for the status quo: you are paid what you are worth, sorry that isn’t as much as you thought. This context is strangely absent in economics classes, though surely its presence would lead students to ask some questions.
Given that the only coherent way to think of produce is as a result of all of the factors of production combined, what determines each factor of production’s share of the produce? Each wants as much as possible, but each requires the others in order to gain any produce at all. So the share for one factor of production is determined by its relative ability to replace the other factors of production. Or, to put it another way, the produce is distributed by bargaining power. In absence of government, capital, being scarcer, generally holds this, and hence receives higher returns. History also suggests that when capital finds itself in a position where this isn’t true, it will use government apparatus to correct this injustice (increasing interest rates, busting unions).
The empirical failure of the standard theory with respect to the minimum wage is well documented*. Furthermore, employment was generally higher during the post-WW2 age of rising real wages and strong unions, and has been lower post 1980, the age of stagnant real wages and weak unions. This doesn’t suggest a causal link, but it does show that high employment is not incompatible with higher wages, and, along with Card-Krueger and the outright logical inconsistency of the MVP theory, makes me inclined to align with the bargaining power story.
*I’m aware that Card and Krueger do not adopt a classical bargaining power perspective, but their evidence still corroborates with it.