« Negatively-valued (red) money in an OLG model | Main | Ontario Finances Not Quite On Target This Year »


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

I guess we actually still think of inflation like pollution, but instead of Pigou taxes or tradeable quotas, we pay subcontractors to efficiently pick up the litter.

As another a nostalgic old guy remembering that 70's show, I sympathise. I remember some things differently. But then your experience in UK and Canadian academia would have been different from my experience trying to protect an Irish exporter from getting screwed in the increasingly chaotic FX market. (One of the horrors which nobody under 40 has experienced is the telephone system of that era; the Reuters Monitor was a major breakthrough.) I became interested in economics because it was apparent to me that journalists like Samuel Brittan were clued into that market, whereas economists were (mostly) totally clueless. It was fascinating to figure out how that situation had arisen.

Were price controls such a hot topic? In Ireland there was much talk of wage controls. Wouldn't that have made more sense? Even today, Olivier Blanchard (another veteran of the 70's) is very supportive of policies which encourage wage moderation as a means of combining high employment with low inflation. Milton Friedman may turn in his grave, but the IMF's Chief Economist is apparently quite happy to let him rotate.

david: right now, keynesians both new and old are more worried about deflation, though just like in the olden days, they are doubtful that monetary policy alone can solve the problem. Some things change, and some things stay the same.

Kevin: In the UK in the 1970's, it was taken for granted that students didn't have phones. If you wanted to see someone, you just dropped by. I was stunned to arrive in the US as an exchange student, and find I was expected and able to get a phone connected the very next day!

It was both price and/or wage controls. But I ducked wage controls in this post, because wages play a subsidiary role in standard NK models, and I wanted to keep it simple. But if we assumed monopoly power in the labour market (unions), as well as in the output market, the model would say we would want both. I didn't know about Blanchard, but it makes perfect sense. (The 1987 paper I hint at above was Blanchard and Kiyotaki, which came out just a few months after mine, with a much better model, so my poor little paper got sadly, but justifiably, ignored.)

"do you agree that your model shows that price controls are a good thing; or if not, why not?"

No, I don't agree. The reason is that with wage and price controls while output would be permanently higher, in the presence of any sort of productivity shocks or growth welfare would be permanently lower.

As an aside, you do need wage and price controls. Price controls alone don't get what you want done.

> Now ban the Calvo fairy, by imposing price controls on all goods, so firms are not allowed to increase prices. Then have the central bank increase aggregate demand. Provided the central bank does not increase aggregate demand too much, firms will increase output and employment and will not ration customers and create shortages of goods. The result is a permanent increase in output and employment, which the model says is a Good Thing.

Wouldn't this happen in any model with price stickiness? Here, the price controls just cause "infinite" stickiness of prices.

However, I think that such a maneuver would fail in the real world because it would cause the money market to not clear. By imposing price controls, the government would fix the exchange between goods (in aggregate) and money. That makes "goods" the alpha and money the beta, and maintaining that exchange rate is not compatible with *also* stimulating AD via monetary policy.

Adam: market power creates a wedge between the Marginal Product of Labour and the Marginal Rate of Substitution between Consumption and leisure. MPL > MRS. (Or MRT > MRS, if you like.) Which means that output and employment will be lower than in the planner's problem (where the planner maximises utility of the representative agent). A small expansion of output and employment, along the existing production function, increases the utility of the representative agent. The first-best solution matches the competitive equilibrium (what the model would predict if all firms took prices as given). I'm not saying anything new there about NK models. That's why booms are good times in NK models, because the natural rate of output and employment is too low.

With a standard labour supply curve, flexible wages, and competitive labour market, price controls would be sufficient. W/P would rise and Y and L increase, with P fixed by law.

Majro: in a model with competitive markets, if you start in equilibrium, then hold prices fixed and increase AD, the only way you get Y to increase is if you assume that firms increase output more than the profit-maximising amount rather than ration sales. Because by assumption in competitive equilibrium every firm is already selling as much as it wants to sell at the given price. So yes, your next paragraph is correct. Firms would ration buyers of goods.

I'd add - many input costs can't be controlled domestically.
How do you tell Starbucks that they can't raise prices when there is a drought in Brazil and bean prices just tripled. There would be no coffee, and without coffee, GDP would be zero, as no one would make it to work!

derivs: yep. Sort of like my real shock, but not exactly the same. It was hard enough with fixed exchange rates, and would be harder still with flexible.

I think the apparent desireability of price controls has a lot to do with assumptions made for convenience. I tend to be most convinced by your objection 4) there are real shocks so relative prices should change.

I think the efficiency costs of price controls are greater than the efficiency costs of moderate inflation.

There is another issue -- imperfect competition means inefficiently low output. Price controls seem to be an odd way to deal with this issue -- they are regulating changes of prices when we are interested in levels compared to marginal cost. I think the obvious approach is to subsidize inputs and tax value added. An increase in markups would imply higher tax liabilities. This would be a tax based price policy.

I think William Vickrey was quite enthusiastic about this, but Google isn't being very helpful sending me here

Coincidentally, Bill Mitchell also revisits the 1970s / 1980s price control era in his blog today:



W/P won't go up (without a positive productivity shock). The firms in an NK model choose their prices as monopolists do, they maximize their profits at a fixed mark-up over marginal costs, which is just wages in the basic model (no capital). To pay a higher wage without ever being able to increase price would permanently reduce their profits, they wouldn't do it.

I know that p > mc makes it look like firms are happy to increase output but not if they have to pay a higher wage. In the standard model (without the price controls) the role of p > mc is to prevent exit. Since firms make some monopoly profits they are willing to keep operating at a point that doesn't actually maximize their profits, since profits can still be positive even while less than the max possible. Under perfect competition profits are zero so operating at less than max profits means operating at a loss, in which case firms would exit.

So, in the standard model (no price controls) an increase in demand induces those firms that can increase prices to hire more labour, paying a higher wage. Since everyone pays the same wage, the firms that can't raise prices have to pay the higher wage and operate at a lower profit. Wages transmit the shock among firms. With price controls nobody can increase price, so nobody pays a higher wage and if that means they can't increase output so be it, increasing output by paying a higher wage reduces their profits.

There are two possibilities in your example, either firms can hire more labour at the same wage, which is where I got the idea you need wage controls, or output won't rise.

If they hire more labour at the same wage then workers are off their labour supply curves, that is welfare reducing. Otherwise output doesn't increase.

Actually, I think you're correct on the welfare point. Booms are good times in the model. If you W increases then workers are better off.

However, as I said above, I don't think W will go up so I think the answer is that output will not increase.

Output is lower than the planner would set because of the monopolistic behaviour of firms, increase demand with price controls won't make them stop behaving as monopolists, nor would it make stop maximizing their profits.

To put this another way, imagine that we did the same exercise but imposed wage controls but not price controls.

When demand increases each firm's demand curve shifts out, the optimal price is determined by the curvature of the demand curve which hasn't changed. So, firms happily increase employment at the fixed wage and none of them increase their price.

This just emphasizes the point that in the standard model it's wages that drive the inflation from an increase in demand. Firms don't react to higher AD by increasing their desired mark-up over marginal cost.

"Provided the central bank does not increase aggregate demand too much,..."


"So, here's a question for all New Keynesian macroeconomists, just for fun: do you agree that your model shows that price controls are a good thing; or if not, why not?"

With price controls it is difficult to see if agregate demanding is increasing to much (after all, then you don't have an inflation rate to study and conclude "inflation is raising/lowering, then we are below/above the natural rate of unemployment, and we should decrease/increase demand") - replacing inflation by other variables like, for example, the size of the queues for bread is problematic because these alternative variables are not easy to measure in objective values.

Robert: "I think the efficiency costs of price controls are greater than the efficiency costs of moderate inflation."

I think that's the wrong way of saying it. When we think of the costs of inflation, we tend to think of shoe leather, menu costs, relative price distortions, etc.

For simplicity, imagine a central bank in an NK model initially targeting 0% inflation. Then impose price controls, and tell the central bank to increase AD. Inflation stays at 0% (by assumption), and we get (micro) efficiency losses due to relative prices being unable to change when relative/real shocks, but (macro) gains from aggregate output and employment being higher.

Back in the 1970's, before we had NK models with monopolistic comp, we just didn't understand this very well, and discussions got terribly confused (especially around that dreadful cost-push vs demand-pull inflation distinction, because where did monopoly power fit in?).

Yep, relative price shocks (4) were probably the biggy, but I also remember 1975/6 in Canada, where the posties were allowed to break the "6 and 5" (??), but the teachers weren't, so the teachers were really mad, and all kept striking or threatening to strike to try to force the government to give them an exemption. The whole thing was just a mess in practice, IIRC, due to reason 5.

Adam: if you have a competitive labour market, then you need an increase in W/P to get an increase in L and Y in a boom, because workers are on their labour supply curve. The way booms work in NK models is by squeezing the markup of P over MC (or P over W/MPL), so W/P increases, so Ls increases. In a boom, the representative firm wants to increase P given W, but the Calvo fairy won't let 90% of them do this, so Y, L, W/P, and welfare all go up (and, depending on the exact production function, the representative firm's profits may or may not go up too, because the AC curve has shifted up, but it is moving down along that AC curve too).

Price controls plus AD increase are exactly the same as a boom, except the Calvo fairy is banned from flying around, so it's 100% of the firms that have their markups squeezed. (Just take the limit of the standard model, as the Calvo fairy visits fewer and fewer firms per period.)

Miguel: in principle, if you saw firms start to ration customers, so Y stops rising and there is excess demand for goods, then you would know the central bank had increased AD too much. But in practice it's harder, because all firms are different, so some will ration customers while others will not.

" if you have a competitive labour market, then you need an increase in W/P to get an increase in L and Y in a boom" - yes, this is what I'm saying.

"The way booms work in NK models is by squeezing the markup of P over MC (or P over W/MPL), so W/P increases, so Ls increases. In a boom, the representative firm wants to increase P given W, but the Calvo fairy won't let 90% of them do this" - yes, this is what I'm saying.

And a direct implication is that under price controls there will be no boom no matter what the central bank does.

When AD increases nothing forces the firms to increase L and Y. Those 90% of firms that couldn't change price are not happy to pay a higher wage and operate at a higher output, they would like for W to stay the same and to simply refuse to meet the extra demand. Why do they pay a higher wage? Because the 10% of firms that can increase price are happy to expand L and this drives up the wage for everyone. The unhappy 90% can pay the higher wage or employ nobody, paying the higher wage is better because they still make a positive (but not maximized) profit. (Under perfect competition they'd exit and employ nobody).

If no firms are happy to increase L then it won't increase, so W won't increase and so there is no boom.

I doubt anybody's model addresses this, but I'll bet you lunch that if limits are announced that no firm may increase wages or prices more than 2% in each calendar year, every CEO in the land will spend time thinking about whether he/she should raise one or both now in order not to get caught in a situation where the firm needs to increase prices/wages but is forbidden to do so. The legally-mandated upper time and percentage limits can serve as a benchmark that will induce market players to modify their behavior.

Adam: take a monop comp firm in profit-maximising equilibrium. The MR curve is below the demand curve. MR < P. Now ban it from increasing P, and shift the demand curve right. The new MR curve now becomes horizontal, at the existing P, until it hits the demand curve. Then it drops vertically, down to the regular MR curve.

Providing P > MC, a firm increases profit by meeting the demand at that P, if it is not allowed to raise P. Subject to the constraint that it cannot change P, MR and P are the same thing.

Roger: Yep. It only works if you surprise firms. If they know or even suspect that the government will do this, and can set new prices in anticipation, it won't work at all. Good point.

It seems like the an (the most) important consideration wasn't mentioned at all. New tech/industries don't start out anywhere near monopolistic competition so no matter what you do with price controls you are going to ruin those industires.

Nick, yeah you're right. For some reason I can't possibly fathom now I was implicitly giving the firms monopsony power in the labour market (so they'd account for their own effect on wages when making a hiring decision). That's of course all wrong, the model has a standard competitive labour market.

So I think you're right that you'd get a boom and it'd be good for welfare, in the model that is.

FWIW, my understanding of why this usually turns out to be a bad idea in real life is that firms start diverting resources to circumventing the controls, forcing the government to divert resources towards enforcement, forcing firms to work even harder on evasion and so on. This is a dead weight loss and quickly exceeds the size of the output gain.

Adam: thanks. Yep, if you assumed monopsony power in the labour market, things would be very different (like in my old post on that subject). The whole thing is clearer of you actually assume monopoly power in the labour market, and sticky wages, then have both price and wage controls.

Yep, we get an arm's race on circumventing and enforcing controls, and it gets worse over time.

Hi Nick-

Not on topic- but I think there's something interesting about the fact that the major chartered banks haven't decreased their Prime rates in response to the Bank of Canada dropping the target for the overnight rate.

The press on the whole has been acting as apologists for the Banks, pointing out that as you approach the zero lower bound, the only way they can sustain margins is by not passing through the rate cut. Other rationales have included the current credit environment (using rate to address credit issues resulting from high levels of indebtedness and the level of housing prices) and the current regulatory environment (higher levels of required capital resulting from Basel III and higher costs of securitization through CMHC). These all seem inadequate to me. Putting my cards on the table, to me it feels like there's a problem with market structure, and this is another example of the 1% finding a way to extract rents from the rest of us.

I think it would be informative for an economist with a focus on monetary policy to enter the fray.

The naive non-economists's view is

Matt: those explanations (and we need to distinguish between explanations and justifications) don't seem adequate to me either. But "market structure/rent extraction" doesn't work either. A profit-maximising monopolist will set price above marginal cost, but will still cut price when marginal cost falls.

I don't have any obviously better explanations though.

Nick: oh no! not the return of the kinky demand curve?! Vade retro satanas said the humble I/O guy...

Jacques Rene: that kinky oligopoly demand curve was running through my mind too!

Hey, I once did a whole keynesian macro model with kinky demand curves!

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad