Minimum wages and predistribution: Extract from Economics in Two Lessons

by John Quiggin on May 15, 2016

A bit out of order, this is another draft extract from my book-in-progress, Economics in Two Lessons. It’s part of the chapter on income distribution, meant to follow the section on unions, and precede the Australia-US data point and the discussion of corporate profits. After this, I plan to conclude the “predistribution” part of the chapter with a discussion of intellectual “property”, then move on to “redistribution” through taxation and public expenditure.

As always, encouragement is welcome, constructive criticism even more so.

Minimum wages and predistribution

The most direct way for government to influence the distribution of market income is to set minimum wages.

The effects of minimum wages on the distribution of income has been the subject of a vast economic literature. Much of this literature starts from a simple (or simplistic) version of Lesson 1. The starting point is the assumption that the price of labor (that is the wage) is the product of a competitive market of the kind we discussed in Chapter 2.

If this is correct, then a minimum wage involves setting a price above the opportunity cost of labor. This means that some workers who would be willing to work at a wage below the minimum will remain unemployed, while potential jobs which yield less production than is needed to cover the cost of a minimum wage worker will remain unfilled or will not be created at all.

Even within this framework, workers may benefit from an increased minimum wage. Suppose for example that the minimum wage is increased by 10 per cent, and that employers respond by reducing the hours of work, for all minimum wage workers, by 5 per cent. In this case, workers would get 5 per cent more total pay, and work 5 per cent fewer hours, gaining both more income and more leisure.

Economists working in this framework point to a number of reasons to doubt this favorable projection. First, the gain to the workers here is associated with a larger increase in cost to the employer. Not only does the employer pay more and get less, but as shown in Lesson 1

Second, typical estimates of the change in hours of work associated with a given change in wages (referred to as the elasticity of demand for labour) are derived for small changes in the wage. Larger proportional effects might arise with a large and rapid increase in the wage.

Third, the idea of a uniform reduction in hours of work for all minimum wage employees is clearly unrealistic. More likely, many workers will experience no change in their hours (getting the full benefit of the increase) while others will lose their jobs, or fail to find jobs when they enter the market.

The third of these points is the most important. However, far from strengthening the case for an analysis based on Lesson 1, it undermines it. Hours of work are not a commodity that can be supplied and demanded so as to match prices and opportunity costs. Rather, each worker is typically matched with one job which largely determines their living standards. With the allocation of property rights to employers that normally prevails in the US, referred to as ‘employment at will’, the job is the property of the employer who can withdraw it at any time, for any reason, or none. Donald Trump’s catchphrase, ‘You’re Fired’ is the simple and brutal expression of this reality.

Because of this imbalance of power, Lesson 2 is just as relevant to the determination of wages as Lesson 1. In the absence of offsetting institutions like unions and minimum wages the imbalance of bargaining power will ensure that most of the benefits of the bargain go to the employer.

(except where they have to patch two or three jobs together, almost invariably ending up with worse terms)

Approaches based solely on Lesson 1 dominated the economics literature until the early 1990s. The central concern of this literature was to estimate the elasticity of demand for minimum wage workers. The elasticity of demand is the ratio of the percentage change in hours worked a given percentage change in the minimum wage. In the example above, where minimum wage is increased by 10 per cent, and that employers respond by reducing the hours of work, for all minimum wage workers, by 5 per cent, the elasticity would be 0.5 (that is, 5/10).

Economists working in this approach expected to find a moderately elastic demand for labor, and they did so. Econometric analysis undertaken in the 1970s and 1980s typically yielded estimated elasticities above 0 (no response) but below 0.5. However, over the course of the 1980s, the estimates tended to decline. Moreover, with the re-emergence of chronic high unemployment after the economic crises of the 1970s, the idea that wages could be regarded as prices emerging from a competitive equilibrium (for which full employment is a pre-requisite) became less and less plausible.

The debate changed radically in the 1990s. The biggest single event was the publication of research by two leading young economists, David Card and Alan Krueger. Card and Krueger examined differential changes in minimum wages in neighbouring states and found that they had no discernible effect on employment in the fast food industry. These estimates were subject to lots of reanalysis, the majority of which tended to confirm the original Card and Kreuger analysis.

More importantly perhaps, Card and Krueger shifted the terms of the debate to include the key point of Lesson 2, that market prices do not always reflect social opportunity costs. In particular they stressed the imbalance of bargaining power between employers and potential workers. This is reflected in what is called, in the jargon of economics, ‘monopsony power’. Monopsony is the other side of monopoly: literally interpreted it means that there is only a single buyer for the good or service in question, in this case labor hours. But more generally, monopoly and monopsony are relevant whenever one of the parties to a transaction has sufficient bargaining power to influence the price (in this case, the wage)

The central implication of the Card-Kreuger is that the primary effect of higher minimum wages will be to redistribute the benefits of the wage bargain from employers to workers, rather than to raise the opportunity cost of hiring to a level exceeding the private and social benefit.

Minimum wages are not a panacea. There must exist some level of minimum wages at which the wage is greater than both the opportunity cost of working and the social value of the output produced. At this point Lesson 1 would be more relevant than Lesson 2.

There is, however, no reason to believe that the current US national minimum wage of $7.25 an hour (far lower in real terms than the level prevailing 50 years ago) is high enough to produce such effects. A comparison with Australia, a country very similar in many respects to the US, suggests that an adult minimum wage of $15/hour could be achieved over time with few, if any, adverse effects on employment.

{ 42 comments }

1

Robert 05.15.16 at 2:34 am

I have all sorts of problems with this. But I’ll put most of them aside for this comment.

As I understand it, A. Dube (often with coauthors) has validated and extended Card and Krueger’s analysis of natural experiments. Can you mention this work?

I like the whole idea of meta-analysis. Card and Krueger had a paper on such. More recently, H. Doucouliagos and T. D. Stanley had another analysis. I appreciate describing the whole concept in a work intended for a popular audience is beyond my capabilities.

John Schmitt, of the CEPR – a think thank, put out a summary not too long ago.

2

John Quiggin 05.15.16 at 2:47 am

Thanks. I hadn’t seen Dube’s work, but I will cite it (my plan, as with Zombie Economics is to avoid endnotes and to include a “sources and further reading” section at the end of each chapter.. I’ll also look up Schmitt. Doucouliagos is a fellow-Aussie and I know his work well.

3

sanbikinoraion 05.15.16 at 12:20 pm

5th para under your first heading ends rather abruptly.

4

J-D 05.15.16 at 12:47 pm

The third and (apparently) final sentence of the paragraph beginning ‘Economists working in this framework …’ has lost at least one concluding clause (‘as shown in Lesson 1 [what?]’).

The parenthesis ‘(except where they have to patch two or three jobs together, almost invariably ending up with worse terms)’ has been cut adrift from whatever sentence it belongs to and pasted in as a paragraph on its own.

The word ‘for’ is needed after the word ‘worked’ in the sentence ‘The elasticity of demand is the ratio of the percentage change in hours worked a given percentage change in the minimum wage.’ In the following sentence, beginning ‘In the example above’, the first instance of the word ‘that’ should instead be a second instance of the word ‘where’.

You have ‘Krueger’ three times but ‘Kreuger’ twice. In the sentence beginning ‘The central implication of the Card-Kreuger is that …’ a noun is needed after ‘Card-Kreuger’ (or ‘Card-Krueger’) — ‘findings’? ‘analysis’? ‘work’? ‘research’? ‘results’?

5

paul wolfson 05.15.16 at 6:27 pm

Economists working in this approach expected to find a moderately elastic demand for labor, and they did so. Econometric analysis undertaken in the 1970s and 1980s typically yielded estimated elasticities above 0 (no response) but below 0.5. However, over the course of the 1980s, the estimates tended to decline.

My recollection of Lesson 1, or rather Economics 101, is that if an elasticity exceeds 1 in absolute magnitude, we term the response “elastic”. If it does not, we term the response “inelastic”. In what world (other than the MW – okay, you got me!) is an elasticity less than 0.5 in absolute magnitude “moderately elastic”?

Also, if we are going to plug work on the minimum wage, I will take this opportunity to plug my own survey (see listing for 2014 near the bottom). You can find the conclusion here.

6

John Quiggin 05.15.16 at 9:38 pm

Thanks for all these comments and links.

7

david 05.16.16 at 2:52 am

C-K succeeded because it hit at a time when New Keynesians were fiercely tussling with the RBC folks. C-K built their model in terms of search frictions and matching problems, and invoked very envelope-theoremy reasoning in their argument (e.g., arguing that the stock market response to minimum wage legislation is ‘flat’, hence…). It is Burdett/Mortensen/Pissarides/etc. inventively applied. This is why the reactions were sharpest from macroeconomists rather than labour economists.

With that in mind, the C-K model is not a static monopsony. Most obviously, the welfare of the C-K firm increases under a binding minimum wage, as turnover is reduced and less is lost to the solving matching problem. The search externality is internalized between employers and between jobseekers. Any further costs are simply passed on customers via prices. You have to be careful about how you invoke its reasoning.

8

david 05.16.16 at 2:59 am

e.g., “central implication of the Card-Kreuger is that the primary effect of higher minimum wages will be to redistribute the benefits of the wage bargain from employers to workers” is not straightforwardly true in a dynamic-monopsony, monopolistically-competitive matching model. the minimum wage functions in the model by improving allocative efficiency rather than imposing predistributive transfers

9

RNB 05.16.16 at 3:49 am

Another way to see this. A minimum wage aids workers in appropriating a larger share of the value added (v+s in Marx’s terms) and that higher minimum wages are likely to cut into that part of the value-added that capitalists had wanted to reserve for their own luxury consumption (that is, surplus value is divided up into additional constant capital expenditures, additional variable capital expenditures and capitalist consumption).

If markets are competitive, then capitalists will have to compensate for higher wages through reducing their own consumption funds (that is, higher wages will result in smaller yachts and shorter luxury vacations and fewer mistresses) since they are compelled to accumulate, i.e. spend on additional constant and variable capital ( this compulsion to accumulate Marx famously called the Moses and Prophets).

Since workers have a higher marginal propensity to consume than capitalists, it may well be that effective demand and therewith employment levels will rise with an increase in the minimum wage.

This is not to say that there is no point at which higher wages would not reduce expenditures on additional constant and variable capital and thereby reduce employment levels or at least the rate at which the absolute level of employment increases.

10

Alex K. 05.16.16 at 5:15 pm

Any comments on David Neumark’s WSJ article (and published research) that claims the opposite, that in fact minimum wage increases do increase unemployment?

http://on.wsj.com/1O7D0uT
(you can probably pass the paywall if you google the article title:
“The Evidence Is Piling Up That Higher Minimum Wages Kill Jobs”)

What exactly is the science behind clutching the Card and Krueger study and ignoring the opposite evidence?

11

paul wolfson 05.16.16 at 7:09 pm

@10: Any comments on David Neumark’s WSJ article …?

Yes.

12

Trader Joe 05.16.16 at 9:11 pm

The problem I always have with these economic studies is that they think like utility optimizing machines and not like an actual business owner. If I’m a business owner and I’m paying the minimum and intend to keep doing it I don’t really have any choice in the matter. I have burgers that need flipped and trash cans to empty and now I have to pay $8 or $10 or $12 or whatever instead of $7.65.

I’d like to try to raise my prices, but most businesses don’t have complete control over that. I might like to get by with fewer workers, but that pre-supposes that I wasn’t already running my business as lean as I could from a labor standpoint. If I could have cut workers I’d have done so well ahead of the implementation of the higher minimum wage – more likely than not I can’t in the short-run. In the long-run maybe I can labor substitute, but in the short run, particularly in service businesses, its all but impossible.

To cope with the problem there’s going to be a little wage theft, some maintenance corner cutting, and input/output quality reduction and none of these seems to get accounted for in the studies either – although any person who’s worked minimum wage knows with 100% certainty that all of these will happen in response to an increase with any of them individually having a greater or lesser transparency to the customer, suppliers and staff.

I guess what I’m proposing is, if the question is ‘What’s the impact of a higher minimum wage’ – I’m not sure the answer involves whether there is a net plus or net minus in employment – I think in the aggregate there’s not likely to be a lot of shift either way. But its also clear that corporate earnings don’t go down dollar for dollar to absorb the added cost, so the question is what constituencies absorb the net difference – suppliers, customers, employer and employee all seem to absorb some small share.

13

RNB 05.16.16 at 9:19 pm

I am not sure the economists care, but a lot of people rightly or wrongly don’t think higher minimum wages will come at the expense of job growth. As I noted above, they think higher wages will mean that CEO’s, upper management and big shareholders will have to do with a bit smaller houses or even one fewer house, one less luxury car, a shorter luxury vacation, fewer Michelin dinners, maybe short-lived affairs rather than concubines; and they may now have as much excess to gamble on derivatives or buy up real estate and use zoning codes to keep rents and their ROI on real estate high.
A lot of the people who want higher wages also have a strong sense that workers deserve them due to the nature of the work they do, which really isn’t much described in an economics textbook as opposed to sociology books.
It does not seem that economists want to address popular understandings of the issue due to the technical apparatus they have, which I grant looks scientific due to the equations and jargon.

14

RNB 05.16.16 at 9:20 pm

typo: they may noT have as much excess to gamble on derivatives or buy up real estate and use zoning codes to keep rents and their ROI on real estate high.

15

Alex K. 05.16.16 at 11:03 pm

Bellow is one of the papers that David Neumark references, in case people want to dig deeper:

http://www.nber.org/papers/w18681.pdf

16

eric titus 05.16.16 at 11:18 pm

Some responses come to mind:
First, it’s not entirely accurate to say that Card and Krueger settle on Monopsony as an explanation. The present monopsony AND search cost models as possible explanation, but their findings don’t really match up with either. There has been strong evidence that classical models don’t explain the outcomes of minimum wage increases. But there is considerable debate about whether Monopsony models do any better. Adding to the confusion, certain forms of monopsony models (notably Manning 2005) use search costs as an explanation for monopsonist behavior. Basically, a company may be able to find 2 good workers if it offers a wage of $7.25 but 5 good ones at a wage of $9. The result is a firm supply curve that looks like a monopsony model without necessarily the same power dynamics.

I’d second the Dube et al paper as a more recent study. There does continue to be some debate about youth employment, low-skill employment, and potential lagged effects. But I would agree that the current consensus is that effects are practically insignificant.

For academics interested in this topic, I’d recommend checking out some of the work looking at China, Indonesia, and other countries with a lot of regional variation in minimum wages. Non-western countries often get overlooked when people talk about minimum wage increases but it’s where a lot of the interesting action has been happening.

17

Kurt Schuler 05.17.16 at 12:52 am

As you may know, the city of Seattle has raised its minimum wage substantially and is scheduled to do raise it further in the coming years. Here is a site following the data, which are skimpy since the law was fairly recent:
http://evans.uw.edu/centers-and-projects/minimum-wage-study

No matter what any particular study may show, though, demand curves do slope downward, even for labor. Readers who are employed and who don’t believe me can try this little experiment: refuse to work at less than double your current salary, creating a self-imposed minimum wage. Do it for a year and report the results here.

Finally, it is not accurate to say that “The most direct way for government to influence the distribution of market income is to set minimum wages.” Minimum wages only affect workers with low wages. There is a way to affect higher-wage workers: comprehensive wage controls, such as the United States had in World War II and under Nixon. In the absence of a global war, one should expect the same results that Nixon got.

18

paul wolfson 05.17.16 at 12:57 am

@16:

For academics interested in this topic, I’d recommend checking out some of the work looking at China, Indonesia, and other countries with a lot of regional variation in minimum wages. Non-western countries often get overlooked when people talk about minimum wage increases but it’s where a lot of the interesting action has been happening.

Another selfy (self-plugging, that is).

19

John Quiggin 05.17.16 at 2:13 am

I’m aware of Neumark’s work and plan to cite it in the “sources and further reading” section. Still, I stand by what’s in the OP

These estimates were subject to lots of reanalysis, the majority of which tended to confirm the original Card and Kreuger analysis.

If there’s a credible survey reaching the opposite conclusion (as opposed to individual studies running against the majority view) please point it out.

20

Peter T 05.17.16 at 2:53 am

Following up on trader joe’s point, there’s an implicit assumption that employees can be added or subtracted one by one. But most (not all) production is done by teams, where the additional unit is the size of the required team – and no work at all takes place below the minimum number. It takes 5 employees to staff a single-person slot over 24 hours, it takes some minimum number greater than one to run even a small fast-food joint, it takes hundreds to run a factory or steel mill. The unit of employment is variously 5, 2-3 or 300+. Minimum wages can be expected to affect sectors differently according to the minimum size of the production unit. Larger production units will absorb the cost and seek ways to raise productivity.

I’m surprised no one has mentioned Singapore’s policy of varying the minimum wage to force changes in the sectoral composition of the workforce (Singapore wanted to move to higher value-added; one way to force capital to move from low-wage sectors was to raise the wage).

21

Marc 05.17.16 at 1:12 pm

@17: High marginal tax rates for very high incomes discourage corporations from funneling very high salaries to their executives, as the latter see less utility from them. And, yes, once the salaries of executives hit the tens to hundreds of millions of dollars a year range, then they really do start to impact the salaries of ordinary employees. The government can set minimum standards for benefits.

The government has plenty of tools to nudge wages in the direction that it wants.

22

jake the antisoshul soshulist 05.17.16 at 1:14 pm

Kurt Schuler @17:
There may or may not be good research, but your claim that minimum wage only
effects low wages is over simplistic. If the employer now paying $7.65 has to increase that
to $10.00, then the employer now paying $10 will have to raise his compensation platform to something higher to maintain his advantage in attracting “better” workers.

23

eric titus 05.17.16 at 3:35 pm

John @19

Neumark/Wascher are the main people doing “surveys” that don’t support the minimum wage. However, I would say that there are some areas where the jury is still out, particularly long-term (ie “lagged”) effects and hiring of entry-level workers. In addition, there has been surprisingly little conclusive evidence showing positive effects from raising the minimum wage. And certainly little evidence that increases lead to higher productivity or spending. In general, US studies tend to run up against shortcomings in the data.

Speaking to Jake @22’s point, this point is called “spillovers” in the literature.

In my opinion, the evidence is pretty clear that small minimum wage increases don’t have a big negative effect more broadly. But, most research looks specifically at the restaurant industry because it is (presumably) the most heavily affected. At businesses with fewer minimum wage employees. However, I have seen (non-academic) indications that there may be effects in government/non-profit industries (healthcare, childcare, retirement, social work, etc) because prices cannot be adjusted upward.

24

Bernard Yomtov 05.17.16 at 4:42 pm

I wonder if part of the effect is reductions in overtime work.

If you pay $7.25/hr then an overtime hour costs $10.88. Raise the minimum to $10.00 and the overtime hour costs $15.00. The response might be either to reduce overtime by reducing opening hours or to reduce it by hiring extra workers at $10 to work those hours.

25

bruce wilder 05.17.16 at 4:45 pm

eric titus In addition, there has been surprisingly little conclusive evidence showing positive effects from raising the minimum wage.

I am having difficulty understanding what that sentence is supposed to mean.

The positive effect of raising the minimum wage is that it increases the incomes of the lowest-paid workers. Once you’ve established that the negative effects of reduced employment, etc are sufficiently modest, that positive effect follows as necessity from the arithmetic. Even in Neumark/Wascher, where employment is reduced by an increase in the wage, the numbers are such that the increase in income is substantial: the arithmetical implication is that a minimum-wage worker has a larger income and shorter workweek. Earning more and working less are both positive effects from the standpoint of a minimum-wage worker.

What’s going unargued here, but maybe should be confronted directly, is whether the perspective of the minimum-wage worker is sufficient proxy for the public interest.

26

eric titus 05.17.16 at 7:59 pm

Bruce @25
It is “common sense” that minimum wages should have positive effects in terms of higher incomes. Perhaps because of this, there is not much research actually showing positive effects. Researchers actually have not found a strong impact on poverty level (likely because of shortcomings in how poverty is defined).

While it is clear that there *should* be higher incomes, there are reasons to ask how large these benefits are. Employers may reduce hours or remove perks such as staff meals. There are also potentially concerns about exceeding thresholds for certain programs (one of the arguments for raising the minimum wage, after all, is to make these public costs private). So there is still research to be done about benefits.

27

bruce wilder 05.17.16 at 9:15 pm

Trader Joe @ 12: The problem I always have with these economic studies is that they think like utility optimizing machines and not like an actual business owner. . . . if the question is ‘What’s the impact of a higher minimum wage’ – I’m not sure the answer involves whether there is a net plus or net minus in employment – I think in the aggregate there’s not likely to be a lot of shift either way. But its also clear that corporate earnings don’t go down dollar for dollar to absorb the added cost, so the question is what constituencies absorb the net difference – suppliers, customers, employer and employee all seem to absorb some small share.

Peter T: . . . there’s an implicit assumption that employees can be added or subtracted one by one. But most (not all) production is done by teams, where the additional unit is the size of the required team – and no work at all takes place below the minimum number. It takes 5 employees to staff a single-person slot over 24 hours, it takes some minimum number greater than one to run even a small fast-food joint, it takes hundreds to run a factory or steel mill. The unit of employment is variously 5, 2-3 or 300+. Minimum wages can be expected to affect sectors differently according to the minimum size of the production unit. Larger production units will absorb the cost and seek ways to raise productivity.

The points that Trader Joe and Peter T raise go directly to some of the core issues in contrasting Lesson 1 with Lesson 2.

The economics of Lesson 1 — also the economics of conventional textbook economics, aka Econ 101 — are basically the economics of practical certainty and complete knowledge or perfect information. The economics of Lesson 2 are, or ought to be, the economics of uncertainty, incomplete and imperfect knowledge.

The imagined world of Lesson 1 is a world on a knife-edge — a metaphor I used in earlier comments on JQ’s drafts to describe Hazlitt’s village with a broken window. Belman and Wolfson used that or a similar metaphor in general reference to business operations in the conclusion Wolfson linked to, upthread .

In the world of practical certainty imagined for the purposes of Lesson 1 reasoning, labor (services) are a quantity metered out at the wage rate to be a quantitative input to a production process. You need more labor to produce more restaurant services, and both are added to the point of diminishing returns.

In a world of practical uncertainty, things are a bit different. Labor isn’t a quantity so much as it is a behavior. The quantity is still there — so many labor hours and muscular effort and so on, but the critical elements become behavior and control of the production process. The wage isn’t a simple meter placed on the quantity of labor used as an input; the wage is part of a contingent contract designed to elicit effort to follow rules.

You don’t need rules in a world of practical certainty. Local knowledge is enough to make the right decision and the right decision is known to be optimal and all that. Hayek and Friedman and Hazlitt have exploited this aspect of a simplified economics of distributed decision-making in constructing their ideology.

In the real world, you need rules. You need rules and learning to achieve technical efficiency in the production process. And, the rules may be embodied as technology in very useful machines and systems of operating procedures. Peter T mentions that people are organized in teams. But, it is more than that. The organization of production, with its teams and machines and rules, results in a cost structure of a completely different character from the one envisioned in Lesson 1 economics, a cost structure where costs at the margin of production are low (relative to average unit cost) and often declining.

In short, firms in a Lesson 2 world are pre-occupied with the problems of technical efficiency and waste. They want to sell more stuff at the current price, because their costs for the last meal served are lower than the average. The restaurant wants the servers to upsell the customer on another drink, another dessert, but also to deliver the check and get the customer out the door in order to seat the next customer, when that’s possible. There’s slack and unused capacity; food left over at the end of the night.

In a world of certainty, the marginal product of labor determines the wage. But, a world of uncertainty, the wage determines the marginal product. Causality, if you like, reverses. Because in a world of uncertainty, someone is managing production and the production process. Management will take the wage as given and manage in such a way as to make marginal product equal the wage. If the wage is low enough, managers may well hire people to supply slack in the system, a use of labor with a particularly low marginal product, even if inframarginal uses of the same labor is quite productive.

If the wage is low enough, managers will hire people and waste their time. And, though economists are remarkably reluctant to do the obvious and model managers slacking, I will tell you that managers do slack. So, when wages rise to some modest degree, part of what disappears from the system is waste. Managers will tighten schedules, retain workers longer and train less often.

28

J-D 05.17.16 at 9:23 pm

bruce wilder (@5) observes that increases in the minimum wage provide benefits to minimum-wage workers in the form of earning more and/or working less. eric titus (@26) expresses doubt, noting that employers may respond to increases in the minimum wage by reducing hours. Yes, exactly! That’s not an objection, that’s part of the point! If I earn the same amount for working a shorter time, that’s a clear benefit to me.

29

Alex K. 05.18.16 at 4:26 pm

“The economics of Lesson 1 — also the economics of conventional textbook economics, aka Econ 101 — are basically the economics of practical certainty and complete knowledge or perfect information. The economics of Lesson 2 are, or ought to be, the economics of uncertainty, incomplete and imperfect knowledge.”

I would say that dividing the arguments for different policy options in this way is a straw man argument.

Although it is true that silly introductory models have a “knife edge” aspect to them, you’re not arguing against the best case for laisse-faire when you assume that the only defense of such policies are the silly little Econ 101 models, GE models and the like.

I would say that the opposite is the case: the decentralized nature of markets, the selection process of the markets, is what enables such a system to function in the sea of uncertainty and intractable economic problems, what gives the market system robustness to the general lack of knowledge present in most economic actors.

Letting a large number of actors come up with solutions to different economic problems and then having the market selection process triage those solutions is an insanely general problem solving mechanism. Such a mechanism is robust to the lack of knowledge even of the majority of the participants, and is robust to a large spectrum of intractable uncertainty — you don’t need assumptions about god-like knowledge and instant adaptation to advocate for such a problem-solving system.

By contrast, expert driven, top-down solutions are extremely fragile to the lack of knowledge of the supposed “experts” — and in the real world, we have mostly experts on economic literature, rather than experts on the actual, concrete, economy.

Of course, insanely general is not yet universal, but I will not elaborate, on Crooked Timber, on the possible systemic failures of market systems. But I will say that the remedies to such systemic failures would better be similarly robust to the lack of knowledge of the “experts” about the concrete economy, otherwise the solutions may be worse than the cure.

(This is just a general point, triggered by BW’s comments above, only vaguely related to the minimum wage issue)

30

Alex K. 05.18.16 at 4:46 pm

Doesn’t a lot of the minimum wage literature (on the side of increasing the min. wage) claim that the effect of modest minimum wage increases on unemployment is not statistically significant once you adjust for various factors?

If so, then the actual claim advanced is more along the lines of “In the highly complex economic world, from the limited data that we have, the effects of modest minimum wage increases on unemployment can not be disentangled from the myriad other possible effects”

This is a much weaker claim than the claim that modest minimum wage increases do not increase unemployment. In fact, the claim is more about the complexity of the system than anything else.

31

bruce wilder 05.18.16 at 7:17 pm

Alex K @ 29: I would say that dividing the arguments for different policy options in this way is a straw man argument.

It would be, I suppose, if I had had to stuff “Lesson 1” economics myself, but I didn’t. It is an already existing and extensive set of texts. You can find many examples, from Hazlitt to Hayek, Friedman to Mankiw. (I looked at Mankiw’s textbook discussion of minimum wage policies once, but I couldn’t recount it now; I recall it as mildly misleading and ideological.)

For whatever it is worth, I agree with you that it is possible to make sensible arguments for conservative and even propertarian laissez faire views, while adopting a realistic appreciation of the concrete economy. It is absolutely not my view that a full appreciation of uncertainty would lead to advocacy for fully centralized direction of the economy; quite the opposite. Top-down solutions, as you label them, designed by mandarins steeped in the fantasies of the economics literature, would be as much a nightmare scenario for me as it would be for you.

When you write, “the decentralized nature of markets”, I suspect you have not quite left Lesson 1 behind. We have an economy in which decision-making is extensively distributed and somewhat loosely coupled, coordinated in important respects (but by no means exclusively) by money and finance. Hierarchies, public and private, are extensive and salient. Actual markets are very few. None of that necessarily obviates what you wrote about the importance of a kind of trial-and-error and evolutionary selection in the way institutions develop and are modified in this system of distributed decision-making that we label the political economy.

It does suggest that when economists tie themselves into intellectual knots pattering on about “labor market monopsony” — just to take one example from the present topic of discussion — they are practicing a kind of doctrinal idiocy, substituting ritualized manipulation of the conventions of the economics literature for investigation of the concrete economy. I don’t think that informs the policy debate, even if I am sympathetic to the policy point of view.

32

bruce wilder 05.18.16 at 7:40 pm

For those playing at home without a program, I should probably note that “labor market monopsony” is a label usually applied to theoretical models that involve neither market nor monopsony. What people who say that typically have in mind is something like a take-or-leave it contract offer to work under schedule and supervision. Game theory is used to support the claim that take-it-or-leave-it as a first move is sufficiently powerful as to be akin (furious waving of hands to obscure the craziness) to monopsony market power allegedly understood from theoretical analysis. A take-it-or-leave-it offer is a powerful first move, but introducing the market metaphor where no market exists just obscures the economic reality of hierarchical organization of production processes. A lot of struggle to escape the quicksand of Lesson 1, even among professionals who should know better . . . and more about the actual economy.

33

Greg Mckenzie 05.18.16 at 11:48 pm

I would only add two points to this excellent chapter:
Firstly, Karl Marx would have been as critical of monopsony power as the laissez-faire capitalist writers were of monopoly power. Classical, or even modern day orthodox economists, implicitly assume that local domestic based employers are somehow ethical in labour negotiations but foreign multinationals are unethical in labour negotiations. The “take it or leave it era” (my way or the highway approach) under the old liberal government’s Workchoices legislation suggests otherwise. As for value added theories justifying higher minimum wages, Marxian
economists would suggest that marginal physical product should determine wage rates.
The idea that only productivity increases can justify any wage rise is so debased when you look at the cross elasticities of labour supply and demand theory. Income elasticity of supply considerations are relevant here. The efficiency first argument that states only technical optimum is acceptable for production will always predicate the lowest possible minimum wage.

Secondly, if the labour supply curve behaves like all other supply curves then it can be seen to be a backward bending supply curve. The current penalty rates debate in Australian politics seems to ignore this theoretical constraint. If a work unit has already worked a forty hour week (still the statutory maximum for normal rates) and is asked to do overtime on a Sunday, they should get penalty rates. There is now some valid arguments for overtime rates for Saturday work when it is less than eight hours a day. But after that penalty rates should also apply on Saturday. The backward bending supply curve can support this argument in theory as labour elasticity changes over the curve. As for the elasticity of 0.5, the whole idea is that it is a relative elasticity calculation. Simply put, perfect elasticity of supply almost never exists and this goes for perfect inelasticity of supply.

34

paul wolfson 05.19.16 at 1:16 am

33:

I don’t think it’s accidental that the most successful liberal-corporatist models (Sweden, Denmark, Switzerland) use a different, more sensible method: establishing and dynamically adjusting minimal wages though negotiations between unions and employer associations

1) I am not familiar with Switzerland, but the Nordic countries all have strong unions and a tradition of co-operation between employers and unions. That would not work well in the US, given the current state of labor relations.

2) It sounds good, but administering &/or enforcing these MWs can be a mess. The problem is that the MW differs along a several dimensions so both employers and employees can have difficulty figuring the applicable minimum. To give an idea, below I cut and paste from a very nicely done study of Sweden — Minimum wages and employment in Swedish hotels and restaurants by
Per Skedinger in Labour Economics, 13 (2006):

Like the contractual wage increases, minimum wages are not subject to local bargaining. During the period of study, 1979–99, some changes have occurred in the differentiation of the rates. In 1979, the minimum wages were defined for around 15 occupations and were also dependent on experience and region. In some occupations, minors (i.e., under the age of 18) were subject to lower minimum wage levels. The regional differentiation ceased in 1986, when the two former wage regions (Greater Stockholm and outside Greater Stockholm) were merged into one. In 1993, minimum wages were defined for only two occupational categories, skilled and unskilled workers, retaining the dependence on experience. At the same time, lower minimum wage levels were introduced for all unskilled workers below the age of 20. Minimum wages for 18- year-olds (19-year-olds) were defined to be 85 (95) percent of the minimum wage for unskilled workers at least 20 years of age.

In the collective wage agreements, jobs are defined as unskilled if previous training is not required and introductory training is provided at the workplace. Formal education, above primary school level (grundskola), is not required in order to perform the tasks. Hotel chambermaid, hotel assistant and restaurant assistant are examples of unskilled jobs. Skilled work is defined as jobs which require vocational training at the high-school level (gymnasium), equivalent formal education, or a long period of practical experience. Among the jobs defined as skilled are receptionist, night-porter, cook, cold-buffet manageress, chef and waiter/waitress.

Fig. 1 shows the development of real hourly minimum wage in 1995 SEK (deflated by the consumer price index for hotels and restaurants) for the four unskilled occupations in the hotels and catering industry which are included in the econometric analysis. In order to achieve comparability over time and across occupations, the minimum wages pertain to workers outside Greater Stockholm, at least 20 years old and with job tenure less than one year… Minimum wages were initially somewhat higher for restaurant assistants (level I) than for the other groups, but the differential has diminished over time and from 1993 all unskilled workers have been subject to the same minimum.

Fig. 2 presents the regional differentiation of minimum wages over time. For expositional convenience, the figure contains data for only one occupation, hotel chambermaids, but the same pattern applies to the other three occupations in the analysis. According to the figure, regional differentiation of minimum wages in Greater Stockholm vs. the rest of the country has been quite modest and has not been used at all after 1985. In Fig. 3, the differentiation of minimum wages by job tenure is displayed. Different minimum wages have applied to workers with job tenure less than one year, one year, two years, at least three years, and, from 1987, at least four years. While these differentials were small until the mid-1980’s, they have since widened. For example, in 1981 a hotel chambermaid with four years’ experience was subject to a 3.6 percent higher minimum wage than a chambermaid with no experience and in 1999 this difference had increased to 10.5 percent.

Fig. 4 shows how minimum wages have been differentiated by age. Minors, i.e., workers below the age of 18, are not considered, since minimum wages for these workers are not included in the econometric analysis. Until 1992, the same minimum wage applied to all hotel chambermaids above the age of 18, with job tenure less than one year. From 1993, different minimum wages have applied for 18-year-olds, 19-year-olds and workers at least 20 years old.

… The minimum wage applicable to the month of September was assigned to each worker and year, based on his or her occupation (1979–92) or occupational group (1993–99), job tenure, region (1979–85) and age (1993–99).

Employers can be well intentioned and still have difficulty keeping track of all the determinants of the pertinent minimum wage for a position since it depends on demographic details of the employee, including not only age but tenure. Typically, enforcement of the MW depends on the employee knowing what the MW is and reporting wage theft to the appropriate administrative agency. This becomes much more difficult when the minimum wage is exquisitely tailored. It used to be common in the US for the minimum wage to vary not only across states (as it currently does) but also to some extent by employee age and size of establishment. There is still a bit of this, but considerably less than in the past (though perhaps that has to do with the MW not having kept up with the price level over the last 35 years).

Germany had an arrangement similar to what you describe up until a couple of years ago, at which point it went from a negotiated to a legislated MW. From discussions I have had with a couple of German academics who study the issue, a major reason was the difficulty of enforcement.

(I hope my HTML is correct: apologies for errors).

35

Alex K. 05.19.16 at 1:48 am

“It would be [strawman], I suppose, if I had had to stuff “Lesson 1” economics myself, but I didn’t. It is an already existing and extensive set of texts. You can find many examples, from Hazlitt to Hayek, Friedman to Mankiw.”

There’s supposed to be an “as if” justification behind the unrealistic models, although the justification is rarely provided. Ideally models would have robustness analysis, where you would find that say, arbitrage models, while highly dependent on liquidity assumptions, are very robust to relaxing rationality assumptions: a handful of actors with limited and partial rationality are enough for obtaining roughly the model’s results. On the other end of the spectrum, some General Equilibrium models are robust to virtually nothing, and the slightest wind breeze would completely throw off their predictions.
But such robustness analysis of models does not seem to be very popular.

Hayek does have a lot of work which fits the ecological view of the economy, with trial and error and market selection processes playing important roles. He also had periods when he exposed a more simplistic view of equilibrium. He had a long and productive life and it is unrealistic to expect 100% coherence in his thought across decades.

I guess my objection is that focusing criticism on unrealistic and unjustifiable models is in fact breaking the norm of arguing against the best case of your ideological counter-party.

“When you write, “the decentralized nature of markets”, I suspect you have not quite left Lesson 1 behind. We have an economy in which decision-making is extensively distributed and somewhat loosely coupled, coordinated in important respects (but by no means exclusively) by money and finance. Hierarchies, public and private, are extensive and salient. Actual markets are very few. ”

I’m never sure what you mean by markets being few. The space for human action left over by the market discipline is not a measure zero set, on which economic actors have to do an elaborate tightrope walking exercise. There is plenty of space remaining for human agency — including space for organizing a firm’s behavior in different ways, hierarchically or not. But I don’t see the value in claiming that market discipline essentially does not exist.

To put it concretely, the behavior of a plumber in pricing, getting and executing his work is not well described by perfect markets. But the plumber’s behavior –and the deal you can make– will be quite different when you can call ten other plumbers (the market) as opposed to when he knows that he is the only one available.
A large organization will also operate quite differently in an environment without competitors as opposed to an environment with free entry and with free exit as a “reward” for bad decisions.

I’m sure all of the trivialities above are known to you, so what you mean by there being very few markets is still a mystery.

36

Nate Roberts 05.19.16 at 6:52 am

37

Nate Roberts 05.19.16 at 7:02 am

Looks like the comment I attempted to post before the preceding one got lost. It was about privatization, a bit off-topic, but something to John Quiggin has written about in the past and I shared because it illustrates one of his central arguments. The quote I meant to post from the previous link is:

“The Housing and Planning Bill, which returns to the Commons this week, will make it even more difficult for anyone either to get a council home or to keep it once they do. Councils will be forced to sell off higher-value properties that could have been let to people on their waiting lists. Most of the money will go to the Treasury. It’s nominally aimed at funding the new right to buy for housing association tenants, but it now looks likely that the ‘higher value’ threshold will be set so low that the sales will create enough money to fund other policies, too, such as the programme to build ‘starter homes’ for first-time buyers. According to Shelter, ‘the only group it appears to help on a significant scale will be those already earning high salaries who should be able to afford on the open market without government assistance.’

38

bruce wilder 05.21.16 at 12:24 am

Alex K: . . . what you mean by there being very few markets is still a mystery.

I suppose one could hear that assertion in any or all of several differing registers. In some range, I am protesting emphatically against the ideological mythology of “the market economy” that originated in the polemics of Hazlitt, Hayek, Friedman et alia, as propaganda opposed to the supposed central planning doctrines of socialism and social democracy. It is not a question of finding the strongest arguments as much as it is a matter of challenging the Biggest Lie, the Big Lie being the core of any exercise in propaganda.

In a completely different register, I am saying most prices are administered and are subordinated to administrative systems of control. “Market price”, if it means anything, ought to mean prices that are formed in market processes akin to something like an open cry auction or a brokered tatonnement. I think many economists believe that the market idea is robust enough that we are justified in the poesy of talking metaphorically about exchange taking place in administered systems (like a supermarket) “as if” they, too, are markets and market analysis is adequate to those cases. I think they are wrong to imagine that.

I am not denying that rivalry affords a kind of competition; I just don’t think that strategic competition is akin in any way to the assumed absence of strategic behavior that defines perfect competition in theory. Moreover, I think economists often mislead themselves by assuming a production function (a theoretical construct that hides problems of technical and managerial efficiency). Administered pricing is associated with increasing returns to scale and the necessity of price discrimination to recover costs and achieve returns on sunk-cost invested capital.

In short, if there were a market in most cases where administered pricing predominates, that market would not clear and could not find an equilibrium in price. This is the essential point in refuting the idea that exchange is taking place in circumstances sufficiently akin to markets that we are justified in thinking of exchange as taking place as if a market were operating, and therefore thinking in terms of metaphoric markets. It is precisely because markets are not compatible with what has to be done to achieve sufficient technical efficiency or the cost structures that result that markets are abandoned and corporate (and government) bureaucracies substituted.

If you need a concrete example, think of movie tickets. The theatre chains and movie distributors cum movie studios lock themselves into contractual agreements that constrain competition and a key reason is that while they want the rivalrous competition promoting this movie or that at locations near you, they are also aware that there will be many empty seats at nearly every showing — the ticket price is not a market-clearing price and cannot be. Administrative information processing constraints mean holding ticket price schedules mostly constant across movies and administratively allocating screens so that the popularity of a movie is transparently known to all parties.

The way economists talk about “market discipline” in the case of banking, where the term is applied frequently, well illustrates how misleading and foolish it is, to not recognize that prices are administered. These are administrative systems that issue mortgages or package mortgages into financial securities and it is the integrity, or lack thereof, of the administrative systems that drove everything in the recent financial crisis. Economists who focused exclusively on market analysis frequently could not find a two trillion dollar bubble and even afterward could not reason out how control fraud might have figured in driving events.

39

Greg McKenzie 05.21.16 at 5:44 am

The problem I have with the term market discipline is the implication that some self discipline is exerted due to ethical standards. So for minimum wage setting, the inference here is that the employer (who demands the labour hours worked and choose the working conditions they like) somehow limits their behavior to accommodate the needs of the workers (who supply the labour). In free enterprise market theory, the usual market model is the monopolistic competition model. This implies a degree of competition from rival bidders for labour units but suggests that each bidder regards their operation as uniquely desirable to those who “win” employment contracts. This sense of entitlement seems to somehow negate ethical considerations. The monopolistic competitor sees themselves as the exception to any ethical rule. In Australia we have the franchise industry that seems reluctant to negotiate minimum wages on ethical grounds. The Seven Eleven cases that have made news in Australia are just one case in point. There is also the last collective bargaining deal for the McDonald’s franchise workforce.
Also there are cases of below minimum wages paid to seasonal workers in fruit picking. These workers can be backpackers who have little knowledge of Australian labour laws and even less bargaining power.
The elasticity of demand for low paid workers can be illustrated by many diagrams but understood only through bitter experience. Try taking a small elastic band and see how much this can expanded, the physical limitations encountered can represent the futility encountered by workers on minimum wages who try to get a substantial wage rise. The term “working poor” was first used in the USA to describe the lowly paid workers who had little choice but to accept a wage rate that was so low that they had no ability to save any income.

40

bruce wilder 05.21.16 at 6:13 pm

(I ask for forgiveness in advance of my sin for my obsessiveness.)

To explain my assertion that markets are “few” in the organizing of the economy as a system, simply consider how many interstices there are, across which exchange occurs. How many can be said to be governed by markets and market-formed prices?

Even if a generous assessment of the robustness of the theory of market price and the applicability of “as if” operationalization criteria lead you to regard many circumstances of exchange as constituting at least metaphoric markets, if the concept of market is to mean anything, there must be some which do not qualify. What is the distinction?

The employment relation has to be among the most problematic. Once you are employed at a wage for work under schedule and supervision, further exchange under that regime surely does not qualify as participating in a market. So why do we talk about a “labor market” if ~90% of the labor force is not transacting in the labor market at any particular time? And, when 10% are, what is the market process? The currently popular analytic idea isn’t the quintessential market process, bidding; instead, it is search. Search is stretching the conceptual envelope a bit, but leave that aside. It is also actively avoiding, it seems to me, the extent to which the customs and conventions of wages and employment are shaped, not by any kind of “market”, but are shaped by the needs of hierarchical supervision and the economic functions hierarchical supervision serves (which I assert is primarily a matter of technical and managerial efficiency, not allocative efficiency; or stated in another way, cybernetic control of the production process).

The point is that the market metaphor gets in the way of focusing attention on what we are actually doing, what problems we are actually struggling with while addressing the economic problems of organizing. Pure market analysis says price is allocating resources; that’s simply wrong, as a matter of fact in an economy in which hierarchies are predominant — prices are held constant, as part of a system of control. That’s as true of a bank’s schedule of relative mortgage rates as it is of the price of a new iPhone or the starting wage at McDonald’s.

Too often we are talking nonsense because the customs and norms of economic dogma require us to talk nonsense.

41

Brett Dunbar 05.22.16 at 3:12 pm

Most of them. They tend not to be ideal markets but are close enough that modelling them as an ideal market s a non misleading simplifying assumption. The number of competing products is limited. However provided that collusion between producers is prevented you still get a fairly well functioning market even with a surprisingly small number of competitors. The actual markets that Smith was discussing were imperfect they still function as markets.

42

Brett Dunbar 05.22.16 at 6:50 pm

Greg McKenzie @41

I think the implication that there is some ethical element tom the term market discipline is entirely in your head. All it means is that the market punishes certain types of behaviour. For example setting a price substantially away from the marginal price, either you lose sales due to setting it too high and are undercut by your competitors or if you set it too low you fail to cover your costs.

Comments on this entry are closed.