You can follow this conversation by subscribing to the comment feed for this post.

Nick,

I don't think you've accomplished what you wanted to here. The relevant comparison is not between an efficiency wage world and one without efficiency wages (which is what you appear to be doing).

The relevant comparison is an efficiency wage world without a deficiency of AD and one with low AD. That's what Krugman's talking about.

So, when you say this "The intuition is that efficiency wage models say that firms need either high wages or high unemployment or a combination of the two" you're actually making Krugman's point, not your own.

Krugman's whole point is that each form takes unemployment as given to them, but once they have high unemployment they have less need for high wages. The coordination mechanism is that high unemployment (due to low AD) allows them to "coordinate" on paying low wages. In the absence of low AD firms can't coordinate on having high unemployment with low wages because if they get that then they all have a private incentive to deviate, but if the high unemployment falls into their laps from lack of demand then there's no incentive for individual firms to deviate and hire more so they can coordinate on the low wage equilibrium.

I don't think you've got a viable counter argument here.

Adam: But efficiency wage models explain why the natural rate of unemployment is positive. It's not a "recession" if we are at {W^,N^}. If Paul is simply arguing that efficiency wage models explain why the natural rate of unemployment > 0, he doesn't need to work on a formal model to show that. It's in all the textbooks.

Put it this way: in an efficiency wage model, there exists some U^, such that if U < U^, individual firms will try to raise relative wages, so we get ever-accelerating inflation. Just like in a model that doesn't have efficiency wages. Nothing new there. The only thing that's new is that U^ > 0, and U^ depends on different stuff.

OK Nick, I’ll bite.

1. This is irrelevant to your conclusion, but I have problems with your explanation of employer-employee bargaining at Nr. Wages are prices. Capital rental rates are also prices. Expenditures are incomes. It makes no sense to me to say that “prices” are flexible but “wages” are not, or vice versa. If there is a distributional shift it is due to *relative* price changes which should be modeled in those terms, for instance changes in bargaining power that give rise to agents’ returns. (I am also uncomfortable with the use of aggregate supply and demand models for labor or other factors, but I’ll put that aside for now.)

2. The main issue here is with the way you’ve drawn your efficiency wage locus, such that it is always more elastic than the labor supply curve. As you correctly say, “The intuition is that efficiency wage models say that firms need either high wages or high unemployment or a combination of the two.” As you move to the left in your diagram the unemployment rate is getting higher. Nevertheless, as you’ve drawn your curves, the efficiency wage premium, the gap between the efficiency wage itself and the reservation wage denoted by the supply curve, is also getting larger. This violates the intuition you’ve appealed to and that Paul Krugman’s conjecture is also based on. So, yes, the locus should be more inelastic: as you get closer to full employment the efficiency wage locus should be further above the labor supply curve. (Note this has nothing to do with endogenous unemployment; Nr is not an equilibrium outcome

As an amateur I don't know about the modeling or variables/equations/mechanics. As far as the concepts go, Krugman sets the issue out as "... one way to think about it is that it keeps firms from competing too hard for workers, enabling them to exert more monopsony power. This effect would have to be weighed against the direct adverse effect of slack demand on profitability...."

Increased bargaining power versus an adverse effect of slack demand on profitability.

In the U.S., I do wonder how much businesses lobby politicians about macroeconomic conditions. No doubt more specific, localized issues like taxes and regulations take precedence. Krugman contrasts the existence of "Fix the Debt" with the absence of a lobbying group to "fix the economy" or close the output gap. A good way to "fix the debt" would be to "fix the economy."

Regarding monetary policy, Fed Bank Presidents are chosen by boards composed of business leaders. Perhaps the Fed is doing what it can given politics, the fiscal headwinds and ZLB.

Peter Dorman:

1. OK. True enough. I'm using "price" in the vernacular, to mean the price of output, as opposed to the price of labour.

2. Take the simple example of a reverse-L-shaped labour supply curve. The supply of labour is N*, provided W > V, where V is welfare, or UI, or something. That is the usual way labour supply is modelled in efficiency wage models. If you are fired, you suffer a loss of W-V per period, and the number of periods unemployed in proportional to the unemployment rate U, which = (N*-N)/N*. So the expected costs of being fired = (W-V).(N*-N)/N*. and this must equal the expected benefit of shirking. So we get an efficiency wage locus that is a rectangular hyperbola asymptoting towards the labour supply curve. Provided W* > V, we will always get my result.

Peter K: "As far as the concepts go, Krugman sets the issue out as "... one way to think about it is that it [the recession] keeps firms from competing too hard for workers, enabling them to exert more monopsony power. This effect would have to be weighed against the direct adverse effect of slack demand on profitability....""

But that would work just as well, and better, in a model without efficiency wages.

And you could make exactly the same argument for workers. A recession helps them bargain down prices of consumer goods.

No problem, Nick, with your example of a fixed labor supply at W>V. But what does this prove? Indeed, even in your special case the effwage premium rises without limit as you go from low N to N → N*. Aren’t we back in Krugmanland?

I think it's pretty tough to argue for monopsony power for businesses in hiring labour across the entire economy. Also - is this suggestion (businesses make more in recessions due to power over wages) compatible with large fiscal multipliers? I think not.

Nick,

I think you've missed my point, and Krugman's. Let's try again.

Imagine, as you did, that an efficiency wage world is one in which in order to maximize their profits firms see a trade off between the real wage they pay and the unemployment rate.

Consider the set of pairs (U,W) that are consistent with profit maximization. For any U there is a unique number W(U) that maximizes firm profits. Let the profit that results be P so P in principle is a function of both W and U, P(U,W).

Krugman's point is that P is NOT a constant function of (U,W). Some (U,W) pairs yield higher values of P. Krugman further says it's perfectly possible that pairs with high U and low W may yield higher P than pairs with low U and higher W. Suppose that's true:

Then the question is, what determines U?

If firms could collude they'd choose the (U,W) pair that maximizes P and so, by our supposition above they'd choose high U and low W. That's the coordination problem they face.

Now, in the real world firms can't collude effectively, so what determines U in real life? Krugman's gonna say aggregate demand.

Thus, AD determines U and then firms choose W(U) to maximize P.

Krugman is saying that if AD is high, so U is low, then firms get lower profits than they would if AD were low and so U was high. In both cases the firms are maximizing profits given the U they face, it's just that higher U may mean a higher maximum is available.

Of course this is predicated on AD being above some minimal level but Krugman is saying there is some range for AD that is high enough, but lower than what get's you full employment (U at it's NAIRU), where firm profits end up being decreasing in AD (equivalently profits are increasing in U for some range of U values).

I could have missed it, but I don't see you addressing this at all in the post or comment response.

What I mean is, you appear to get your argument mostly from this statement here: "Less obviously, in an efficiency wage model, real wages cannot fall by as much in a recession as they would in the same model without efficiency wages. The gap between W^ and Wmin1 is smaller than the gap between W* and Wmin2."

This compares the efficiency wage world to one without efficiency wages which is completely irrelevant. Further, firm profit is determined by the complete interaction between volume sold and margins.

Essentially Krugman is saying that (U,W) pairs where U is high and W low may give the same or higher productivity to the firm but at a lower wage. So total profit can be higher, even if volumes sold are lower, because margins are higher.

In fact Nick, we KNOW that in an efficiency wage world firm profits go up as we move from N* to N^. This is true of every efficiency wage model since it's the firms that impose the efficiency wages. The reason is that moving from N* to N^, with it's increase in the real wage, increases labour productivity by more than the rise in wages. That is the essence of the "efficiency" in efficiency wages.

All Krugman is saying is that perhaps firm profits would continue to increase for at least some values of N

Remember, the diagram you drew is for a given level of aggregate demand (say determined by MV=PY). Shift AD and you shift all your curves.

Correction: All Krugman is saying is that perhaps firm profits would continue to increase for at least some values of N less than N^.

That's what that was supposed to say.

Another Correction:

All Krugman is saying is that perhaps firm profits would continue to increase for at least some values of N less than N^ and perhaps a small shortfall in AD would get you there.

Remember, the diagram you drew is for a given level of aggregate demand (say determined by MV=PY). Shift AD and you shift all your curves.

Adam P: We must be talking past each other.

"This compares the efficiency wage world to one without efficiency wages which is completely irrelevant."

I think it is relevant, because I read Paul Krugman as saying his argument depends on an efficiency wage model.

My efficiency wage locus shows pairs (U,W) that are consistent with individual firms maximising profits. But, as you say, it is perfectly possible that a point of my efficiency wage locus with lower W and lower N (higher U) gives all firms higher profits than than at the point {W^,N^}. But they would have to act collectively to coordinate on that point. So I am agreeing with you and Paul Krugman on that.

But it is even more likely to be true that profits would increase in a recession if we ignore the efficiency wage model and replace the efficiency wage locus with the labour supply curve. If the labour supply curve is inelastic (it almost certainly is) if all firms could act collectively to cut wages and employment then profits would increase.

Peter D: " Indeed, even in your special case the effwage premium rises without limit as you go from low N to N → N*. Aren’t we back in Krugmanland?"

Take the simple reverse-L-shaped labour supply curve: We get a larger cut in W for the same sized recession in a standard model than we would in an efficiency wage model. Firms have a smaller incentive to conspire to create a recession in an efficiency wage model than in a standard model.

Adam P: "All Krugman is saying is that perhaps firm profits would continue to increase for at least some values of N less than N^ and perhaps a small shortfall in AD would get you there."

Agreed. It might. But it would be even more likely in a non-efficiency wage model.

"Remember, the diagram you drew is for a given level of aggregate demand (say determined by MV=PY). Shift AD and you shift all your curves."

I disagree. Here's the model: MV=PY (or whatever); Production function Y=F(N) ; labour demand curve F'(N); labour supply curve W/P = MRS(N,Y); Shapiro-Stiglitz curve (damn, maths is too hard). The labour demand curves and labour supply curve depend only on technology and preferences. shifts in the AD curve, given sticky P and/or W move you along the curves (or off the curves, if both are sticky.

Adam P: "In fact Nick, we KNOW that in an efficiency wage world firm profits go up as we move from N* to N^. This is true of every efficiency wage model since it's the firms that impose the efficiency wages."

It may be true, but we don't know that. It will depend on the model. With a very inelastic labour demand curve it will not be true. Starting at W*, it is profitable for the individual firm to raise its wages. But when other firms raise their wages, it might help or harm our firm. Collectively, they might lose more from higher wages than they gain from productivity.

Actually, in the reverse-L model workers don’t take any pay cut at all as unemployment increases if employers can reduce wages to their reservation level. (I use the definition of the reservation wage as the lowest wage that will induce a given number of workers to accept employment.) In the efficiency wage model there is a pay cut.

Peter: That is only true if W=V when there is no recession (if the labour demand curve cuts the horizontal bit of the labour supply curve).

Huh? It is true at every level of employment less than or equal to N*. Remember the definition of reservation wage. Even if you want to argue for some reason that employers will pay above the reservation wage at N* there is no wage cut beginning at N = N* minus epsilon.

Nick,

I won't claim to have figured this out, but my suspicion is that Krugman is right. The thing is, I don't see how you can make an argument by comparing the size of "consumer surplus" when the whole concept ignores the fact that wages (and unemployment) are in the production function itself. In the context of an efficiency wage model (at least some of them), if you raise wages, or increase unemployment, you actually get more output from a given amount of labor. So your X-axis isn't measuring something proportional to output, where you can just multiply a horizontal distance by a vertical distance (or integrate a curve across a horizontal distance) and get a meaningful surplus value. We're in some kind of warped space, where both the horizontal and vertical positions affect the meaning of the vertical distance. If a recession means that you produce more from a given amount of labor (as it would in an efficiency wage model), then the smaller vertical distance might actually reflect a greater amount of profit.

Of course micro economists are going to say that the efficiency wage is more elastic - since that would rationalize higher unemployment at lower levels of demand - and since micro economics is all about creating circular justifications for their more basic crackpot theories.

Also, does not aggregate supply of labor decrease with higher income. That is at least what it looks like to me, whether you look at one country over time or between different countries at the same time (our ancestors did, after all, put in a lot of backbreaking work in order to simply receive a chance at survival). The reason a single firm or industry is facing an upward sloping supply is that people might look for work in other firms/industries.

And you could make exactly the same argument for workers. A recession helps them bargain down prices of consumer goods.

But the effect of bargaining power is not equal for firms and workers. Workers have to have money in order to spend, and generally get it from one source, a paycheque from their employer. The effect of a single bargaining session is much more drastic on labour income then on expenditure, where workers have to nickle-and-dime merchants down. It is not a contest of equals and it isn't a very linear exchange, either individually or in aggregate.

PS: I think that the statement that labor supply decrease with income also would be supported by looking at different groups of people in the same country, at the same time, when one group only has access to relatively low paying jobs due to e.g. legal status.

Sorry for spamming, but the claim that the efficiency wage curve is more elastic than the labor supply curve would have to imply that unemployment would be falling over time as a country becomes richer (which, to some extent, generally, seem to be the case - but surely is due to better institutions rather than a closing between the efficiency wage and labor supply curve).

Andy Harless,
"I won't claim to have figured this out, but my suspicion is that Krugman is right."

My suspicion is always that Krugman is right unless it's obvious to me that he's not (in such cases I'm almost always among the first to comment). In this particular case I was wondering why Krugman felt that it was desirable to bring up efficiency wages, and now I'm wondering if Andy has figured this out.

Andy: OK: take a simple version, where the worker's choice is binary: work or shirk. As long as we are at a point on or above the Efficiency wage locus, the production function and labour demand curves won't shift. Can PK (or anyone) rig up an efficiency wage model where a recession increases profits? Sure he can. But you don't need efficiency wages to do that. A simple model with flexible wages and a relatively inelastic labour supply curve can do exactly the same, or give an even bigger increase in profits. With a perfectly inelastic labour supply curve, for example, an infinitely small recession would cause a large drop in wages and a large increase in profits.

Aha! Unless PK is (implicitly) assuming that both nominal P and nominal W are fixed. If real wages are fixed, and there's a recession, then profits fall in the standard model. They would fall less in the efficiency wage model, for a given drop in employment, if labour efficiency increased. Could profits conceivably rise? You could probably rig a model to make it work.

Maybe that's what he has in mind?

Or maybe he just hasn't got the point about firms being able to coordinate on the efficiency wage equilibrium without a recession?

I've (once again) re-read what he wrote, but it's just not clear enough to figure out what he's saying.

Determinant: suppose a competing supplier of apples offered you better apples at the same price, or equally good apples at a lower price, than your existing supplier of apples. Would you feel free to stop buying apples from your existing supplier and switch to the competitor? Now replay that argument with "labour" instead of "apples".

Suppose a seller of labour got a better price from an competing buyer of labour, compared to his existing buyer. Would he feel free to stop selling to his existing buyer and switch to that competing buyer? Now replay that argument with "sell" and "buy" switched.

Suppose husbands could easily divorce wives, but wives could not easily divorce husbands. Or vice versa. Who would have the bargaining power?

Cut through the framing and see the reality. Join is in the joys of the autism spectrum!

Nick, I have no idea what you are trying to say. Let's not be ungracious to those "out there" on the autism spectrum, but one of the characteristics of the autism disorders is that a person can't handle nuance very well.

If you don't do nuance, then you'll utterly despise non-linear results (by which I mean that you can't reverse an operation and get your original result back). Andy Harless' observation neatly describes a non-linear function.

I'm with Paul Krugman and Andy P on this one. Additionally, firms do "conspire"; it's called "Industry-average compensation". At least for non-executives, most firms I have seen have tried to stay near or hopefully below the industry average in terms of pay, to preserve their profit margin. When everybody tries to be "average", the result looks a lot like a conspiracy, though in this case an innocent one.

Unless PK is (implicitly) assuming that both nominal P and nominal W are fixed.

Paul Krugman is a huge believer in sticky wages and sticky prices. You may have got to the root of your problem.

I think it may be that Paul had the right intuition but went down a blind alley by bringing efficiency wages into it. The possibility he mentions may hold in a more standard Pissarides type search model, with wages set by bargaining rather than the "no-shirking condition". However, IIRC, in that model wages are strongly pro-cyclical, which is how you can get that profits increase - (real) wages end up falling by more (%) than the shock (%). But in the recession real wages have been pretty much constant.

Prices, nominal wages stagnant. Employment down. Demand down. Total output down. That doesn't leave much in the firm's profit function to play with. Productivity I guess, but that's already in output. Hmm, let's see. Profits (including capital income) = pQ-wL. p, w, fixed per data and usual sticky assumption. dQ<0 per recession, dL<0 per unemployment going up. Whether profits go up depends on dQ/dL vs. w/p, or whether workers were getting paid marginal product. Basically they would have had to have wages above MPL pre-recession (I think) for it to work. Then the recession somehow does the job of maximizing the firm's profits for them by bringing MPL more inline (by reducing labor input, diminishing returns etc) with existing (fixed) real wages. I dunno, I can believe wMPL is going to be hard to squeeze out of some model.

Mmmm, scratch that, the sign flips when you take dL to the other side.

OK, this is becoming clearer to me. There are price effects, and there are productivity effects. You choose a model in which both of these exist, or you can choose one in which one or the other does not exist. If you assume wages and prices are equally sticky, then you rule out price effects, and you can focus on productivity effects, which is a stylized description of what Krugman does. If you choose a class of efficiency wage models in which productivity is always maximized (as in a typical no-shirk-condition model), then you can rule out productivity effects and focus on price effects, which is what Nick does.

My conception of the real world is that Krugman's implicit assumption is more relevant. Although there is some evidence for procyclical real wages, I'm pretty sure the procyclical part comes from booms rather than recessions. A firm doesn't want to piss off either its workers or its customers, so (if it has market power, so that the zero profit condition doesn't apply) it resists raising prices, and it resists cutting wages. Obviously there is no harm in raising wages when the marginal cost curve tells you to do so (presumably during a boom), and there is no harm in cutting prices when the marginal revenue curve tells you to do so (presumably during a recession). Since we're talking about a recession, we're talking about a time when wages are going to be stickier than prices, so if the recession benefits the firm at all, it has to be through productivity effects, not price effects. It seems plausible to me that it might, but somebody's still going to have to show us the model and convince us that it works with reasonable parameters.

Nick Rowe: "Suppose husbands could easily divorce wives, but wives could not easily divorce husbands. Or vice versa. Who would have the bargaining power?"

The wives, always! ;)

Let me run the danger of making more sloppy math mistakes. If Output=capital income + labor income + profits, and prices and capital stock are constant, and we measure "firms's income" by capital income + profits then for firms's profits to increase we need g-s(new)*(1+g)+s(old)>0 where g is the % change in total output, s(old) is previous labor share and s(new) is new labor share. Or, rewriting, g>%ds*(s(new)/(1-s(new)), or the growth rate in total output needs to be greater than about 1.38 times the % change in labor's share.

During the actual recession, 2007-2009 (or so), approximately g=0. But, going by BLS data, %ds=0. In other words, the fall in labor's share didn't actually occur till *after* the recession, as profits rose, while labor and employment stayed constant. And Krugman's graph shows this if we look at 2007, and then 2009. Of course there's that big "dip" in profits in the middle. Post recession we have g>0 and %ds<0 which is also born out in Paul's graph.

Here's the thing. And it bothers me. It's the definition of a "recession". It's implicitly being used in two different ways. A NBER recession is one where output falls for more or two quarters. A "macroeconomic" recession is where output is below trend, potential, full employment, etc. even if it is growing. Properly speaking we're no longer in the first kind of recession but we are in the second kind. And Paul's post is titled "Why Corporations Might Not Mind Moderate ***Depression***" (that could be "recession").

So the question is, what kind of "recession" does our model model? If you're a macro guy using usual frameworks (old or new keynsian) then you're talking about the second kind and when you put y's, p's and w's in your graph you're implicitly assuming that they're all measured in "deviations from trend" and such. But you *can't* make that implicit assumption if you're doing micro models, like Shapiro-Stiglitz. There, output is output not "deviation of output from trend".

Since late 2009 output has been increasing. It's still below trend, but the right shock to analyze in the efficiency wage model (or whatever search micro model) is NOT a fall in demand (or productivity) but an increase in it. Sure, that increase is "not big enough" to get us back to where we were in 2007 - hence we're still in a "macro recession" - but it is still an increase in demand.

So what you're need to be looking at is not a change from N^ (or N*) and Nr, but a change from Nr *up* to some level closer to N^.

The "standard explanation" works quite well for the (NBER) recession. Profits actually do take a big dip, before they level back out. It's what's going on afterward that needs explaining. And there you get to have output going up, so it's not that crazy. In fact, to explain that portion you don't need a demand side story, a supply side story will do. You can just make capital and labor very complimentary (think close to Leontief PF) and have the increasing productivity improvement be labor augmenting, as is standard. You get an imperceptible (zero with Leontief) increase in wages, an imperceptible (zero with Leontief, on that kink) increase in employment and a big rise in profits. Done.

OK. If we assume:

Both nominal W and P stuck, or equally sticky, so that real wages W/P do not change,

The shirking function is continuous

The economy is otherwise perfectly competitive (no monopolistic competition in the output market)

The recession is "small"

Then it works.

Proof: start at W^ and N^ where N is individually profit-maximising. Hold W constant at W^, and decrease N by dN. The direct effect of firm i decreasing its own Ni is second order small, since it had already chosen Ni to maximise profits. But there is a first order positive externality from other firms cutting N, since this increases U and reduces shirking.

The proof doesn't work under monopolistic competition, because of the aggregate demand externality (the natural rate is suboptimal).

On the other hand, that proof also proves that small recessions are Pareto efficient! I'm not sure if Paul wants to go there!

Maybe, if you assume monopolistic competition, and rig the parameters just right, you could get a model where small recessions are Pareto inefficient yet increase profits. You probably could.

Nick: I think I see what you wanted to say. Recession + sticky prices/wages is very similar to what is going on with tariffs/taxes. There is some deadweight loss so the economy as a whole loses and then there is some consumer and producer surplus. The size of these respective surpluses depends on the shape of supply/demand curves. It can go either way and efficiency wage makes it so that it is more likely to go in favor of workers, or at least it makes recession less costly for them.

However despite efficiency wage there is still a possibility that it goes so that producers capture more of the surplus to compensate their share on DWL from recession. That they are better off. This is where Krugman's graph of profits/wages comes in which suggest that workers are losing on this recession more then "capitalists"

So the process is like this "Data show that firms are doing much better than workers in this recession. Why is it so? Let's take a look at this via labor market model to find out what story supports this data."

JV: Yep. But if I were building some sort of conspiracy theory of recessions, to ask who gains from the silver lining that every cloud has, I would look at real interest rates, not real wages. We know that econometricians have a hard time figuring out whether real wages are pro or countercyclical. The simplest stylised fact is that real wages don't do anything much in recessions. But real interest rates have done a lot in this recent recession. Firm's borrowing costs have shrunk. The relevant class analysis is between capitalists and rentiers. Owners of stocks vs owners of bonds.

Determinant: it looks like Paul Krugman reads your comments! He's wrong. It's buyers who gain, and sellers who lose, bargaining power.

notsneaky: "A "macroeconomic" recession is where output is below trend, potential, full employment, etc. even if it is growing."

Isn't the word for that "depression"?

Nick Rowe: "On the other hand, that proof also proves that small recessions are Pareto efficient! I'm not sure if Paul wants to go there!"

Well, why not? Doesn't it make sense to say that a small recession, or one with a quick recovery, is part of a healthy homeostatic process, while a deep recession, or one with a slow recovery, is dysfunctional?

"On the other hand, that proof also proves that small recessions are Pareto efficient!"

Does it? I don't think you've modeled the disutility workers get from dialing down their shirk function. That would be kind of the point: in a weak economy, you produce a lot, but work is hell. Good for firms; good for consumers; but bad for workers. Since most households are both workers and consumers, whether it's net bad for them is, I think, ambiguous; at least it's not obvious.

Andy: the value of shirking to the worker must be less than its cost to the employer, otherwise they wouldn't ban it, and would cut wages and allow shirking for a Pareto-improving deal.

I didn't say that going into a recession would be Pareto improving. I meant that recessions would be Pareto Efficient. Gains to the gainers bigger than losses to the losers. The constrained second-best socially optimal level of employment would be lower than N^, given Paul Krugman's argument (and efficiency wages being the only departure from the first theorem of Welfare economics).

Min: "Well, why not? Doesn't it make sense to say that a small recession, or one with a quick recovery, is part of a healthy homeostatic process, while a deep recession, or one with a slow recovery, is dysfunctional?"

You Schmumpeterian you! Don't you know that recessions are bad? Aren't you a true Keynesian or Monetarist?

Min, no. I'm pretty sure I'm right here though. Paul and Nick are trying to explain a phenomenon which didn't occur. It's not "profits went up in a recession/depression" but rather "profits went up in a jobless recovery", which is more vanilla.

Surely, all this theorising fails in the face of NEW businesses entering the market? There is no incentive for them to collude with existing businesses. Isn't this just a new facet of the PSST thing?

I can see a reason why the "high wage" end of the spectrum would be self-extinguishing in a tight labour market.

Businesses don't compete within an industry, but against all other entities which extract capital from ordinary people. Supposing one modern day Henry Ford begins paying his workers more, "so they can buy his cars." Other entities are likely to restructure so as to capture that increased revenue before it ever gets back to the Ford showroom. Excess cash paid to Ford's workers can actually work against him, by strengthening the sales of his competitors.

Or other, unrelated sectors can capture that capital by dint of positioning themselves higher on the necessity rankings -- for instance, internet service can edge out cable TV and telephone both, as it can replace both to some extent, and can do things they cannot do.

Does anyone study this endless jostling to get upstream of non-sector competitors? Or is it consciously pursued as a business tactic, rather than just happening?

Noni

Determinant: it looks like Paul Krugman reads your comments! He's wrong. It's buyers who gain, and sellers who lose, bargaining power.

I'm flattered. ;)

Um, Nick, that "sellers of labour", aka workers, lose and "buyers of labour", aka employers, gain is exactly Krugman's point. I don't think you were trying to say this, but you did. Anyway Krugman's still right.

His characterization of the employment relationship is also exactly spot-on. It is the difference between a contract "for services" vs. a contract "of service". An example of the former is a plumber who has control over his tools, his work methods, can send his apprentice if he wants instead of him, and who has many customers. A contract of service are "contract" jobs, i.e. all those sessional lecturers.

One of the greater frauds of our time is "contract workers" who are in pith and substance employees being told to incorporate themselves and bill as a business, for which they get purported tax deductions. I have been in this situation myself. It cheats the government out of income tax and cheats the employee out of benefits like EI they get from an employment relationship. The fraud is passing off a contract of service as a contract for services.

Noni: "Supposing one modern day Henry Ford begins paying his workers more, "so they can buy his cars.""

That always puzzled me too. Why would they spend it on Henry Ford's cars? Why not just pay them in cars?

Then I read a couple of years back it's an urban myth. That's not why he paid high wages.

Determinant: this argument has moved on to the new post!

"the value of shirking to the worker must be less than its cost to the employer, otherwise they wouldn't ban it, and would cut wages and allow shirking for a Pareto-improving deal."

Is it true? The usual assumption is that shirking can only be monitored imperfectly, with low probability. If this is the case, an employer trying to determine whether her worker is shirking "too much" faces a hard noise-vs.-signal problem.  It may be that a no-shirk solution is only efficient in a quite constrained sense.

The comments to this entry are closed.

• WWW