What is it with Microeconomists?

They are certainly not stupid. And they are certainly not ignorant either. I know that the ones I'm complaining about are smarter than me, and more knowledgeable than me. And that includes economics smarts and knowledge. Some of them make me feel totally inadequate on a daily basis (I read their blogs daily). Some of my best friends are microeconomists. But they just don't get macro!

I'm talking about money wages and employment. I can't be bothered to link to the posts I'm complaining about. And I can't be bothered to go through those posts and explain why their reasoning is wrong. Others have done this, and have failed. Or at least, have failed to make any impression on the 'microeconomic miscreants'. They seem to be preaching to the choir; and the choir is composed of macroeconomists.

I want to try a different tack. I'm not going to try to show that they are wrong. I want to try to understand why they keep going wrong.

But I'm not altogether sure why they keep going wrong. I have three theories, and am going to run through each in turn. Actually, I think we probably need all three theories to explain why microeconomists are just so confused about macro.

Just to set the microeconomic scene: draw demand and supply curves for labour. Suppose wages are above the equilibrium, so there's an excess supply of labour. Cut wages, and you move along the demand curve for labour, and employment expands. Or does it?

1. They don't understand that macroeconomic demand (and supply) curves are fundamentally different from microeconomic demand (and supply) curves.

Why do demand curves slope down? The microeconomist answers: income and substitution effects. Let's take them in turn.

If I cut the price of apples, holding money incomes constant, holding all other prices constant, then consumers are better off in real terms. They can afford to buy more goods, and if apples are a normal good, they will spend some of their increased real income on buying more apples.

This "income effect" is total rubbish in macroeconomics. If there are 100 apples sold, and I cut the price of apples by $1, then consumers of apples are $100 richer. But producers of apples are $100 poorer. No effect on aggregate real income. Sure, there might be distribution effects, but distribution effects are the last refuge of a .... person who can't think up a better argument.

And, if apples were the only good produced in the economy (as they would be if this were macroeconomic analysis), then it's even worse rubbish. What's on the horizontal axis of the supply and demand curve graph? Real income is the quantity of apples. It's GDP! And money income is the price of apples times the quantity of apples. It's what's on the vertical axis times what's on the horizontal axis. It's logical nonsense to draw a demand curve for apples holding money income constant, unless you draw a rectangular hyperbola demand curve for apples.

OK. What about the substitution effect? A fall in the price of apples, relative to the price of other goods, means consumers will substitute away from other goods and into consuming more apples. That makes sense in micro, where we draw the demand curve of apples holding the prices of other goods constant. But in macro, where the price of all goods is on the vertical axis, it doesn't make any sense at all. If the prices of all goods fall, what are they substituting away from? And why?

2. They don't understand the difference between a notional and constrained (effective) demand for labour.

Here's the simplest way to derive a labour demand curve for macro. Write down a short run production function Y=F(L) showing output as a function of employment, holding land, capital, and technology constant. Assume positive but diminishing marginal product of labour: F'(L)>0, F"(L)<0. Assume competitive firms maximise profits taking money wages W and prices P as given. The first order condition is that the marginal product of labour demanded equals the real wage: F'(Ld)=W/P.

That looks like a nice downward-sloping labour demand curve with the real wage on the vertical axis. But it isn't. It's only a "notional" labour demand curve. It only tells us how much labour a firm will want to hire if the firm is able to sell as much output as it wants. But if there's excess supply in the output market, firms won't be able to sell as much output as they want. They can only sell the quantity demanded, Yd. That's an additional constraint on the firm's maximisation problem. If that constraint is binding, the solution to the firm's maximisation problem is Yd=F(Ld): hire as much labour as is needed to produce the output demanded. W/P does not appear in that "constrained" (effective) labour demand function. For a given Yd, the labour demand curve is vertical.

Let's simplify the whole thing. Forget about diminishing marginal product. Make the production function Y=L. One worker produces one apple. Output and employment are the same, and are the same as real income. The notional labour demand curve is a horizontal line, because the marginal product of labour is one, at all levels of employment. Forget about firms. All workers are self-employed apple-growers. Their money wage and their price of apples are the same thing. The real wage is one.

How can there possibly be an excess supply of labour in such a world? If one worker tried to sell his apples for a higher price than the other workers, he might be unemployed. Nobody would want his apples. But then there would be an excess demand for other workers' apples, and an excess demand for their labour. There is no way the labour market/apple market as a whole could be in excess supply, because on average, the price of apples must equal the average price of apples, duh!

Which brings me to:

3. They don't understand monetary exchange.

This is the really important thing they don't understand.

Walrasian general equilibrium theory ignores monetary exchange. If there are n goods (n varieties of apple), then each person submits a list of demands or supplies for the n goods to the Walrasian auctioneer. That list has to conform to the person's budget constraint, which means that the values of goods demanded has to equal the values of goods supplied (Walras' Law). The income from the goods they want to sell must be sufficient to buy the goods they want to sell. All goods are bought and sold in one big market.

But macroeconomists assume monetary exchange instead (though some of them may not realise it, I admit).

If all apples were identical, there would be no need for trade. Each worker would eat his own apples. No worker could be unemployed; he could just grow as many apples as he wanted to eat, and eat his own apples.

Even if apples came in different varieties, or there were a tabu against workers eating their own apples, if barter were easy there could never be unemployment. The unemployed workers could all just get together and swap all the apples they wanted to produce and sell. In barter, a supply of apples is a demand for apples.

It is monetary exchange (or rather, the high transactions costs of barter that make monetary exchange essential) that is the root of all deficiencies in aggregate demand. Each worker's apples are sold in his own private market. And they are sold for money, the medium of exchange. To demand apples is to supply money in exchange. And if people want to hang onto their money, rather than buy apples with it, the demand for apples, and the demand for labour, will be deficient.

A deficiency of aggregate demand has got nothing whatsoever to do with a deficiency of income. Income is always sufficient. It's always the same as goods sold. A deficiency of aggregate demand is a deficiency of peoples' willingness to get rid of money. The "Paradox of Thrift", and the "Paradox of Toil", are merely corrupt versions of, or way-stations to, the Paradox of Money. Each individual can increase his stock of money by buying less; but in aggregate they fail, but cause unemployment as a side-effect. 

The Aggregate Demand curve is a locus of points at which people in aggregate are just willing to hold the stock of money they do hold. If microeconomists don't think about monetary exchange they can't think about aggregate demand. Cuts in wages will improve or worsen unemployment due to deficient aggregate demand if and only if they increase or reduce people's willingness to get rid of money.

 

Comments

[Too Much Fed]

Nick said: "And, if apples were the only good produced in the economy (as they would be if this were macroeconomic analysis), then it's even worse rubbish. What's on the horizontal axis of the supply and demand curve graph? Real income is the quantity of apples. It's GDP! And money income is the price of apples times the quantity of apples. It's what's on the vertical axis times what's on the horizontal axis."

What about a z axis for time?

[Too Much Fed]

I think you need to figure in productivity growth, a fungible money supply (currency and currency denominated debt), and durable goods into your model along with corporation(s) and workers for the corporation(s).

[Too Much Fed]

Do you have any links to graphs about macroeconomic demand curves? Thanks!

[Adam P]

Nick, this is a nice post.

I'd add

4) they don't appreciate the consequences of nominal stickiness.
5) they don't appreciate the interactions of 1, 2, 3 and 4.

However, speaking of the interactions of 1-4, your statement here:

" But in macro, where the price of all goods is on the vertical axis, it doesn't make any sense at all. If the prices of all goods fall, what are they substituting away from? And why?"

is quite wrong. Aggregate demand curves can be understood by substitution effects, it works through the real interest rate (which requires the interaction of 3 and 4) but the the thing they are substitutiing away from or to is future consumption/investment, that is the macro level substitute for current consumption/investment.

[Mike Moffatt]

Agreed with this:

"Cuts in wages will improve or worsen unemployment due to deficient aggregate demand if and only if they increase or reduce people's willingness to get rid of money."

But how many macroeconomists understand this point? I don't think it's just us micro guys who don't get macro!

[MattM]

Too Much Fed, he can't simply include one axis for time; time is a cube! Don't you know anything?

[Bill Woolsey]

Nick:

I don't deny that aggregate income and output are the same.

However, I still think that a lower minimum wage raises the employment of unskilled workers. If demand in that sector is unit elastic, then the flow of expenditure for the products of unskilled labor is the same. The flow of expenditure in the rest of the economy is the same. More employment and more output in one sector. No change in the rest of the economy. More employment and more output together.

If demand is inelastic in the minimum wage sector, then the flow of expenditure towards unskilled labor is less, and it is greater to the rest of the economy. And that appears to be the situation, but it is a small effect.

Does this contradict the logic of monetary disequilibrium? No, because the prices of the products of unskilled labor are lower.

I think the error among the macro apologists for the minimum wage is to correctly understand that the lower minimum wage is only going to expand aggregate demand if it lowers prices. And the lower prices only work through the real balance effect. Base money is small, so the real balance effect is small. The small impact of the the minimum wages on prices can only have a tiny effect on aggregate demand.

Well, the reality is that minimum wage workers are such a small part of the economy that it takes only a tiny increase in aggregate demand to greatly expand their employment. (They are also making a stock-flow error in seeing the increase in real balances as a one shot increase in the flow of expenditures.)

The lower minimum wage slightly lowers the price level and slightly raises real balances and slightly raises raises real expenditure and slightly raises real output and slightly raises real income. But because of the substitution effect--because the relative prices changed, the increase in demand and real output is concentrated on the products of unskilled labor. The increase in employment is concentrated on unskilled labor.

There is an income effect. Real income rises because real output rises. And it is small.

The microeconomists go wrong when they go from this partial analysis to thinking about cutting all wages. Are they really thinking that there is an aggregate shift to more labor intensive technologies? Of course not. I know Caplan has in mind a lower price level and a real balance effect. I think he is just assuming that it works perfectly, more or less, and it is just wages that are failing to make the adjustment.

The thought experiment (which I got from Patinkin) where we suppose nominal wages fall and the price level stays the same, and so employment doesn't rise is just not something most people worry about. Prices fall too.

Further, I don't think most of these folks really favor the lower price/wage solution to the problem. It is just that they don't see price and wage floors to be helpful. And it bugs them when apologists for wage floors make bad arguments about these things.

I know it bugs me.

In the U.S. the unemployment rate among black males between 16 and 19 if 57%. Raising nominal expenditure would do wonders, of course. But the 50% increase in the minimum wage probably didn't help much.

[Frances Woolley]

Nick, doesn't this assume that microeconomists have some sort of vague interest in questions such as unemployment, GDP growth, etc? Sure, some labour economists do (and the sins of labour economists are a topic for another post), but your average microeconomist would probably rather spend a Christmas party hiding in the bathroom than give some sort of intelligent opinion on what's going to happen in the macro economy.

So most microeconomists are not guilty of these sins simply by virtue of being totally uninterested in the macro economy. But there must be some microeconomists out there blogging and motivating this post. So why are there microeconomists, or perhaps more accurately, people with micro-based conceptions of supply and demand, out there talking about these ideas?

Perhaps micro-style supply and demand is part of it? Even though over 1/3 of my department is nominally macro, several people work in areas like finance or trade with essentially micro paradigms/methods. Perhaps micro people talk about these issues because there's not enough macro economists supplying good analysis? (see previous discussions on economics in the media).

And what's the appeal of micro methods? 1 to 3 make perfect sense when you explain them, Nick, but how easy is it to write down a model? And would that model have micro-foundations? And if you can't write down a model, can you get it published? And if you can't get it published, what's the point? So do people really get much exposure to this type of macro-thinking in school?

And if you're reasoning things out on your own, what's easier than to think "o.k., what makes sense to me?" and then generalize to the entire economy? Going from individual to social level analysis - isn't there some impossibility theorem about that?

[Nick Rowe]

Adam P: Thanks!

I agree with your 4. Though I would say that this is *because* they don't understand your point 5.

I also agree with your point 5. I was sort of saying that at the beginning, when I said "Actually, I think we probably need all three theories to explain why microeconomists are just so confused about macro." But I should have been more explicit.

I also basically agree with you on substitution effects. There are substitution effects driving the macro AD curve; it's just that they are very different than the substitution effects driving micro demand curves. Substitution between present and future goods (Euler-IS); and substitution between present goods and money (AD). Also, these macro curves aren't demand curves in the same sense as micro demand curves. I prefer to think of them as equilibrium conditions. A true demand curve tells you what quantity demanded would be even when the economy is *off* the curve; an equilibrium condition only tells you what quantity demanded is when the economy is *on* the curve. (Imagine that output was supply constrained for example, so we are off the IS curve; the IS curve does not necessarily tells us how much output is demanded; it can only tell us what output would have to be so that output=output demanded.)

Mike: unfortunately, you may be right.

MattM: You lost me there!

Too much Fed: Sure, you can add time as the third dimension. That's just saying how the curve shifts over time. And you can add lots of other things that affect demand as the 4th, 5th, etc. dimensions. Partly we think of 2 dimensions because damned universities only supply us with 2-dimensional chalkboards and paper. Partly it's because we want to focus on the variable that bears the main burden of adjusting to ensure equilibrium.

[OrganicGeorge]

Nick you are attempting to have a logic discussion about hubris.

What you really want is for the macro guys to admit that their life's work is crap. With all their accolades and education they are wrong, they cannot admit it and so they continue to repeat what they know.

This is an interesting point in time, we have found ourselves at an financial, social, and governmental dead end. Looking back we see the wreckage of the policies past 20 years that must be cleared away, like debris from a storm. There is no easy way to clean up the mess, which includes all the wrong ideas and ideals.

Nobody really knows what will replace the old ideas, but clinging to them for the sake of pride is a sure sign of failure.

[Adam P]

Bill, when you say this:

"However, I still think that a lower minimum wage raises the employment of unskilled workers. If demand in that sector is unit elastic, then the flow of expenditure for the products of unskilled labor is the same. The flow of expenditure in the rest of the economy is the same. More employment and more output in one sector. No change in the rest of the economy. More employment and more output together."

are you sure about the no other changes part? It seems to me that you're assuming that legacy minimum wage workers keep their original (higher) wage. However, I'd imagine that this wouldn't happen, that all minimum wage workers would get the new, lower minimum wage rate and thus a class of workers would suffer a fall in income. This fall in income is partially offset by the increase in employment but it seems likely that total consumption expenditure could still fall and this would drive the fall in prices that drives the real wage back up.

[Nick Rowe]

Bill:
"However, I still think that a lower minimum wage raises the employment of unskilled workers. If demand in that sector is unit elastic, then the flow of expenditure for the products of unskilled labor is the same. The flow of expenditure in the rest of the economy is the same. More employment and more output in one sector. No change in the rest of the economy. More employment and more output together."

First, the assumption of unit-elastic demand for unskilled workers only matters because it means we can simplify by ignoring income distribution effects. (You probably agree, I'm just making sure.)

Second, why does the demand curve for unskilled workers slope down? Is it because firms are substituting unskilled for skilled workers? Or is it because consumers are substituting the products of unskilled workers for the products of skilled workers? In either case, there's a substitution *away* from skilled workers.

Third, suppose, just suppose, there were no real balance effect, and that aggregate real output demanded stayed the same? How can there possibly be more output in the unskilled sector, and the same output in the skilled sector, and so more output overall? All there could possibly be would be a change in the composition of output. (Employment would rise, but only because unskilled are less productive).

In other words, your analysis only really starts when you talk about the real balance effect. Everything before that adds nothing to the story (except the composition of demand, output, and employment).

I agree with what you say thereafter.

[Andy Harless]

Perhaps the explanation (in some cases, anyway) is that microeconomists are not talking about a blanket wage cut but about a relative wage cut that would affect the distribution of labor demand in such a way as to increase total employment. See Robert Waldmann on this point. Or, to put it a bit differently than Robert does, perhaps they are talking about a change in the relative price of labor vs. human capital, which would increase the demand for labor and reduce the demand for human capital, holding aggregate demand constant. Since labor is more severely underutilized than human capital (i.e., the unemployment rate is much lower for educated workers), a reduction in the price of labor relative to the price of human capital could bring us closer to the optimum. It would also presumably (if you choose parameters that certain microeconomists think plausible) raise measured employment.

[Nick Rowe]

Hi Frances! I was hoping you would join in. (As I said, "Some of my best friends are microeconomists"!)

I am *really* pleased to see your: "1 to 3 make perfect sense when you explain them, Nick,..." That was what I was really hoping to achieve in this post.

And you can model those things, with microfoundations, though it's not always easy. Macroeconomists do model them. 3 is probably the hardest thing to model, and most macroeconomists usually build 3 into their models implicitly, without even realising they are doing it. So no wonder it confuses microeconomists.

In normal times, microeconomists can talk sensibly about most macro topics even while not understanding 1,2,3. That's because the Bank of Canada does its best to make sure that 1, 2 and 3 are irrelevant for practical purposes. It tries to keep the economy in monetary equilibrium, so there is always just enough aggregate demand, that we can ignore aggregate demand. And it tries to keep inflation steady at 2%, so we can ignore what's happening to the overall level of prices. The job of the Bank of Canada is to make macroeconomics irrelevant for everyone except macroeconomists.

But now we are in a "Special Period in Time of Peace", as Castro would say.

But it's not all about these abnormal times. As I was writing this post, I remembered a conversation we once had about productivity and living standards. You were working through a micro analysis of the effects of productivity on wages, employment, prices. And I was thinking "output=income, output/person=income/person".

"Going from individual to social level analysis - isn't there some impossibility theorem about that?"

There's the Fallacy of Composition", that what is true for each of the parts isn't necessarily true for the whole. That's a large part of my point 1.

[Nick Rowe]

Organic George: "What you really want is for the macro guys to admit that their life's work is crap."

Sometimes I think that too. But the one thing guaranteed to cheer me up is to read non-macroeconomists talk about macro. Because they usually make the same mistakes we figured out years ago. Which is NOT to say we can't sometimes learn from them.

[Nick Rowe]

Andy: I skimmed Robert Waldmann's post yesterday. On a brief reading, what he says is right. And what you say sounds right too. But I don't think that's what the microeconomists are talking about.

[Rogue]

When you have an open economy in a globalized environment, the micro assumptions can hold, and macro assumptions of a closed economy break down.
1. If you make apples $100 cheaper, domestic producers can potentially sell more apples, regardless of the level of demand domestically. It becomes a substitute to apples (or oranges) everywhere else in the world.
2. Hence, even if there is an excess supply of apples in the domestic market, it can still (potentially) sell as much output as it wants.
3. And even if the “paradox of thrift” holds in the domestic economy, the excess supply of apples could clear elsewhere in the global economy. (It also helps if your government controls the value of your currency, which in actuality is an integral part of what you trade in a globalized economy).
Macroeconomists will need to go beyond closed economy assumptions if they want to be able to address the confusion of microeconomists.

[Nick Rowe]

Rogue: If you are talking about a small open economy, (under fixed exchange rates, with a common currency) you are absolutely right. In fact, I would say that SOE macro isn't really macro at all. It's micro. The only true macro is closed economy, and therefore global macro.

But if we have national currencies, then I think there are some aspects that can only be handled from a macro perspective.

[Frances Woolley]

other examples of sloppy micro to macro generalizations:
- run government like a business (or a household) (that's from another of your best friends, Nick)
- by saving one person can ensure they will be secure in retirement therefore by everyone saving everyone can be secure in retirement (demographics affect wages and other stuff)
- "productivity" (people talk about productivity as an explanation of wages, and it is true in some sense that productivity=output/person, but this ignores the fact that labour's share of output (or the typical worker's share of output - I don't mean CEOs salaries) is changing. Even if - and this is a big if - markets are competitive so wages=marginal product of labour, who gets the surpluses? And to the extent that productivity is just output/person, why think that you're doing something really special and unique talking about productivity when all you're talking about is per person GDP like everyone else?)

[Bubble Meter Blog]

Nick Rowe said:
"If there are 100 apples sold, and I cut the price of apples by $1, then consumers of apples are $100 richer. But producers of apples are $100 poorer. No effect on aggregate real income."

This seems wrong from the outset. The consumer doesn't just have $100 more, he has $100 more to deploy elsewhere, such as buying oranges. So, the apple growers are $100 poorer, the consumer is $100 richer (measured in $100 worth of oranges), and the orange growers are $100 richer. You are being led astray by looking at dollars rather than looking at how much stuff those dollars can buy. The loss to the apple growers is compensated by an equivalent gain to the orange growers. However, the consumer has more stuff than he otherwise would have (an extra $100 of oranges), thus the overall gain in wealth in the society is $100 of oranges.

[Nick Rowe]

I think the savings and retirement is the best one of those. Each individual can save enough to pay young people look after him in retirement, but that doesn't mean that all people can do so.

And the strange thing is, my guess is that macroeconomists would be just as likely (more likely?) to miss that point as microeconomists.

Another example: that book on marriage, where you had to explain to me why the arithmetic didn't work in aggregate.

You know, I don't think it's the Fallacy of Composition thing that differentiates micro from macro. Microeconomists can do, and do do General Equilibrium, after all. I think it's monetary exchange.

[Nick Rowe]

Bubble Meter: OK, suppose people spend that extra $100 freed up on oranges. So consumers of apples of apples demand $100 worth more oranges. But producers of apples demand $100 worth fewer oranges. It's a wash. The consumers of apples get more oranges, and the producers of apples get fewer oranges.

[Bill Woolsey]

Nick:

Quit starting with the scenario of monetary disequilibrium and some kind of notion that a lower minimum wage has something to do with ending the recession. The way to do microeconomics is to assume monetary equilibrium. Either the price level adjusts to keep the real supply of money equal to the demand, or the nominal quantity of money adjusts to meet the demand. Or, really, just abstract away from the problem.

http://monetaryfreedom-billwoolsey.blogspot.com/2009/11/more-about-productivity-norm.html

Suppose there is monetary equilibrium and the minimum wage is cut?

The products of unskilled labor have negative slopes because of both substitution and income effects. If the demand for the products of unskilled labor is unit elastic, then the combination of both those effects leaves total expenditure on the products of unskilled labor unchanged. Total expenditure on everything else is unchanged as well.

There is an increase in real output, the products of the unskilled workers. More are hired. The produce stuff. More output. And so more real income.

The increase in real income raises the demand for all normal goods. The income effect on the rest of the economy raises the demand for those products enough to offset the substitution effect of the increased relative prices of their products. The net effect is zero.

If the demand for the products of unskilled labor were elastic, then expenditure rises for the products of the unskilled labor and falls in the rest of the economy. That means that income effect was smaller than the substitution effect in the rest of the economy. The substitution effect of their higher relative prices means less is sold, and the income effect of the higher real income only partly offsets that.

How does monetary equilibrium fit in? Well, it only fits in to the degree that the demand for money depends on real income. The story above involves an increase in real income and that raises the demand for money. If the quantity of money is assumed given, the price level must fall. Thankfully, the price of the products of unskilled labors has fallen. But there is no guarantee that this is exactly right. Of course, by making the claim that the demand for unskilled labor (and the demand for the products of unskilled labor) is unit elastic, I have already embedded that assumption.

Since the demand for unskilled labor appears to be inelastic, and if that is because the demand for their products are inelastic, then lowering the minimum wage results in lower income for unskilled workers and less spent on their products. That increases demand in the rest of the economy. But this is a small effect on the rest of the economy. The big effect is more unskilled workers working and producing more of their products.

AdamP:

If the demand for unskilled labor is unit elastic, then the total income of all the unskilled workers is unchanged. The existing workers make less, but the new ones who are hired make more. Total income for them is the same.

I think that it is inelastic, and that the total income of the minimum wage workers falls. The existing workers make less, the new workers make more, but the decrease is greater than the increase. You know, marginal, average, that stuff.

But everyone else's real income rises. Their nominal income stay pretty much the same really, and the prices of the products of unskilled labor are a bit less, and so the real incomes of everyone else are a bit higher. They spend a bit less on the products of unskilled labor and a bit more on other things. That raises the demand for everything else. But, none of these effects are much worth worrying about. The big effect is that more product of unskilled workers is produced and more of them are employed doing it.

Now, if the minimum wage were high enough so that it applied to a large proportion of the work force, then these small effects become big. As Nick explained, the substitution effects disappear.

[Lord]

I wonder how much is not understood and how much is not believed. It is even possible for a non Keynesian to believe in macro? When someone argues that macro has to be the sum of all micro decisions and no more than that, they don't really understand it. Others can formulate the arguments but can't really think about it without reverting to micro as if these refute them.

[Nick Rowe]

Bill: OK. I think I follow you now. Start out by assuming AD adjusts to whatever is needed, work out the micro, then go back and check what's happening to AD and whether it adjusts as needed.

My tactics are the opposite. Start out by assuming no change in AD, work out the micro, then see what happens to AD.

We should end up in the same place.

Lord: I think it's quite possible for a non-Keynesian to believe in macro. It's just that it's hard to disentangle the two, since our macro is so wrapped up in Keynesian modes of thinking.

[Jim Sentance]

I wasn't going to dignify this discussion with a post - imagine, given recent history, asserting that microeconomists not getting macro implies a deficiency with micro thinking - but then you wouldn't be aware that I was doing that. So...

[Mattyoung]

I don't understand why Price in in the vertical axis of Supply and Demand. The economist uses price comparison as the normative behavior, then claims that price does not always work. Figure out what the consumer Norm was in Truck and Barter days, then use that Normative function in the vertical axis. Treat money separately as a highly utilitarian technology.

[Geoffrey]

Very nice post and very helpful on the intuition! Any links or suggestions for more on the monetary/macro connection? (i.e. other resources that do such a good job on the intuition?)

[scott sumner]

Nick, Good post, I've noticed the same problem. I think we confuse micro people by calling AD "aggregate demand." Let's draw it as a rectangular hyperbola and call it "NGDP." Then they wouldn't be so confused.

How would your discussion of the labor market change if you believed (as I do) that wages are much stickier than prices? I.e. the typical oil company can sell all the oil it wants at the world price, but the wages of oil field workers is sticky.

[Nick Rowe]

Scott: Thanks!

"How would your discussion of the labor market change if you believed (as I do) that wages are much stickier than prices?"

Good question. One that ought to be answerable, but to which I don't have an immediate answer.

Here is my initial thought: suppose we had an ISLM theory of Aggregate Demand, for example. Y is on the horizontal axis, and the LM curve assumes M/P is fixed. Use W/P=F'(Ld) and Y=F(L) to re-write the LM curve in {r,L} space, with M/W fixed, and re-write the IS curve in {r,L} space too. Then the AD curve would be in {W,L} space.

I must play around, and see what it looks like. Actually, in the simple case where the production function has constant returns to labour input, so Y=L, the ISLM diagram looks exactly the same, only with L on the horizontal axis. And the AD curve looks exactly the same, only with L on the horizontal and W on the vertical axis. So the labour demand curve becomes the standard AD curve.

That's a neat way of thinking about it.

Geoffrey: Thanks! The trouble is, I have pulled together a number of old ideas which aren't discussed much nowadays. I would start out by reading an old essay by Robert Clower: "A reconsideration of the Microfoundations of Monetary Theory" 1967. Maybe also "The Keynesian Counter=revolution: a Theoretical Appraisal".1963/65. Both in Money and Markets, ed. Donald A Walker. And these are not easy reading, because Clower was stumbling towards understanding this stuff.

Somebody else must have a better answer to Geoffrey's question than I do.

Matt: I'm not following. Let me guess. In a barter economy, with n goods, there are n(n-1)/2 markets. In each of those markets one pair of goods is exchanged. Write the relative price of those two goods on the vertical axis, and the quantity of one of them on the horizontal axis.

You know, the weird thing is: I have never thought about how one would model a true barter economy with more than 2 goods. All our theories are about Walrasian economies, with one big market for all n goods. Or for monetary exchange economies with n-1 markets.

[Nick Rowe]

Arrr, Jim! Sure, but if you want to make us macro guys look good, just let the micro guys do macro!

[Jim Sentance]

Thats the whole point Nick, why would we want to do something as messed up as that?

[pireader]

Professor Rowe --

You've offered an intellectual explanation for the micro guys' difficulties with macro--What they don't understand. I'd favor a psychological explanation--Why they don't understand.

In my experience, people get deeply into micro-economics because they find the rational-actor view of behavior intuitive and satisfying. It fits their sense of 'how the world should be'. And it works well for a variety of actual behaviors, validating their aesthetic pre-disposition empirically.

Now comes the problem. Since the 1930s (at least), it's been quite clear that the macro-economy has emergent properties of great practical importance which rational-actor models just don't describe very well. But that reality provokes deep revulsion in these people, contradicting not only their view of how the world is, but how it should be.

So for the past 30 years they've struggled to come up with a rational-actor account of the macro-economy's emergent properties ("micro-based macro"). Not because that's the most-promising scientific approach but rather because that's the aesthetically-required approach. With the "Great Moderation", it looked like maybe they were getting somewhere. But now we've all been forcibly reminded that there's a lot more to the macro-economy than the micro-based models can account for.

Must be disheartening ... no wonder such smart guys have repeatedly said such dumb things over the past year.

[TGGP]

Pete Boettke says macroeconomics is bad for your brain, the problems are all micro:
http://austrianeconomists.typepad.com/weblog/2009/12/jobless-recovery-or-why-macroeconomics-is-bad-for-the-brain.html

[reason]

It did occur to me that Rogue was right, but the problem is (as I have been saying all along) that the international financial system is broken and relative international prices are unreasonably sticky.

[reason]

Or we wouldn't have got in this mess in the first place!

[reason]

I could even put it another way, to make the point clearer - the problem is that micro-economists think prices are always right (because each market moves towards an individual clearing price in isolation). But let the adjustment towards clearing prices occur at different rates in different markets and short term movements can be AWAY from global market clearing prices. Prices were not only wrong, they were persistantly wrong. Hence large imbalances and sudden swings in the propensity to save.

[Frances Woolley]

I agree with Pireader on the aesthetic appeal of micro-based approaches. Glenn Ellison’s 2002 Journal of Political Economy article on “Evolving Standards for Academic Publishing: A q-r Theory” argues that papers are now judged less by the quality of their main ideas and more by other dimensions of quality - i.e. aesthetics.

Like I said earlier, this debate isn't about microeconomists, but about people with micro-based conceptions of (aggregate) supply and demand. A better title for the post would be 'Why people shouldn't use micro-methods to explain macro concepts' - but that would make it obvious that, as someone else pointed out earlier, that what you're really doing is arguing for throwing out much of modern macro.

Your average microeconomist just spends their days happily thinking about whatever little rational choice model is playing in their mind right now and totally ignores all of the things being discussed here. Or at least I do - and here I'm being a good microperson and unhesitatingly generalizing from a sample of one to the entire economy.

[Mike Moffatt]

For what it's worth, my International Trade course contains a pretty heavy dose of macro, and I spend part of the first lecture describing how the AD curve differs from a microeconomic demand curve. But perhaps I'm the exception, rather than the rule.

Nothing I've read so far has changed my opinion that *nobody* understands macro, so I'm still puzzled why us micro folks are being singled out.

[Ralph Musgrave]

Bill Woolsey (Dec 19, 8:01am and 2.43pm) is concerned about AUTOMATIC effects of minimum wage cuts on AD (e.g. via real balance effects (Pigou effect)). He admits these effects are small.

Adam P (8.35) doubts these automatic effects.

Given the wild gyrations in AD (we’re in a recession, aren’t we?) aren’t these arguments near irrelevant? Seems to me the important and long term point (i.e. not of any great relevance to the recession) is that minimum wages (and union wage rates come to that) form a barrier to employment expansion.

I.e. given falling unemployment, the marginal product of labour falls because employers find it increasingly difficult to find the types of labour they want. When marginal product of labour equals the minimum wage (or union wage) further expansions in AD will simply be inflationary. Ergo, cut the minimum wage (and/or abolish unions!) and that will facilitate a rise in AD and aggregate employment (with less inflation than would otherwise have been the case).

[Bubble Meter Blog]

Nick Rowe said...
"Bubble Meter: OK, suppose people spend that extra $100 freed up on oranges. So consumers of apples demand $100 worth more oranges. But producers of apples demand $100 worth fewer oranges. It's a wash. The consumers of apples get more oranges, and the producers of apples get fewer oranges.

It's not a wash. Keep in mind that producers and consumers are often the same people playing different roles in the economy. On the consumption side, assuming that everyone eats a basket of apples and oranges, everyone gains from lower apple prices. Their savings from lower prices are redeployed to purchase more apples and oranges. On the production side, the loss to apple growers is completely offset by the gain to orange growers and to unemployed workers who become either apple growers or orange growers. Only the production side is a wash.

Let's try saying this using textbook economics. Only two things can cause the price of apples to fall, increasing supply or decreasing demand. Since the change in apple prices is being caused by lower marginal costs (specifically lower wage costs), it should be obvious that the number of apples being produced will increase. That is, the supply curve for apples has shifted to the right. All other supply curves have remained fixed. Since the supply curve for apples has shifted to the right while all other supply curves have remained fixed, it is obvious that the AS curve has shifted to the right. At this point the economy is more productive than it was before. Also at this point, the rightward shift of the AS curve has caused the price level to decline. Since the nominal money supply has remained fixed, the real money supply has increased.

[Bubble Meter Blog]

I should also add that with high unemployment and much unused productive capacity, there's no way we're currently sitting on the long-run AS curve. The long-run AS curve is to the right of our current position. The question is how quickly can we move the short-run AS curve or the AD curve to meet it.

[Bubble Meter Blog]

Nick,

Looking back at your original scenario, it seems to rest on several dubious assumptions.

First, you assume that if the price of apples declines due to lower marginal costs, the quantity sold remains unchanged. Keeping the quantity unchanged is the only way you can get the gain to consumers to equal the loss to producers. That assumption violates the basic laws of economics. Since your price change is occurring completely due to lower marginal costs, the quantity sold will increase. What producers lose from lower prices, they will partly make up through higher quantity. You can't say that cutting the price by $1 will cause a loss of $100 if 100 apples are sold, because the number of apples sold will change.

Second, you assume that a lower minimum wage will cause currently employed minimum wage workers to have their salaries cut. In practice, sticky wages are the rule.

Third, your assumption that quantity remains fixed leads you to assume that unemployed workers are not put to work. In practice, they would be hired at a rate below the minimum wage, but that is still preferable to being unemployed. (If it weren't preferable, they wouldn't take the jobs.)

[Nick Rowe]

Pireader:
That's an interesting perspective on micro-based macro, but I'm not sure it's right. One of the forces driving micro-based macro was the disconcerting discovery, in the 1970's, that empirical relationships we thought were structural (like the traditional Phillips Curve) suddenly shifted on us, which we interpreted as meaning they were not structural relationships after all. So we sought the theory underlying those empirical relationships, to try to figure out when they would shift and why. The only deep theory we knew was micro; so we ran with it.

But, to agree with you, that doesn't mean that macro is just "micro writ large", or that there aren't emergent properties. But our job as macroeconomists is to find and explain those emergent properties. And what I was trying to do in this post was to do just that, sort of. To expalin to microeconomists why macro had to be different.

TGGP: Dunno. It always seemed to me that the best Austrians (e.e. von Mises) were always doing macro all along, even if they didn't call it that. Ditto Hayek. Of my 3 points above, the best Austrians always seemed to understand 1 and 3, at least. Maybe not 2. Just my impression, of course. I can't give you chapter and verse.

reason: but suppose the US had been a closed economy. Would it have gotten into the same financial crisis? Maybe without the flow of savings from the rest of the world. But then maybe the countries with the big savings would have gotten into the same mess if closed. Like Japan, for example.

"I could even put it another way, to make the point clearer - the problem is that micro-economists think prices are always right (because each market moves towards an individual clearing price in isolation). But let the adjustment towards clearing prices occur at different rates in different markets and short term movements can be AWAY from global market clearing prices."

IIRC, that argument was floating around in the late 1960's and early 1970's (Leijonhufvud?). But I don't know if anyone ever fleshed it out and built a model showing it worked. Then that whole approach got passed over.

Frances: "A better title for the post would be 'Why people shouldn't use micro-methods to explain macro concepts'..." Maybe. But lacks the eye-appeal, no? Could you see The Sun running with a headline like that?

"...- but that would make it obvious that, as someone else pointed out earlier, that what you're really doing is arguing for throwing out much of modern macro."

I'm really not sure about that. There is some truth to it. It really bothers me, for example, that I can't give a good reply to Geoffrey's question above (On where can he read more about this stuff?). I can't decide whether what I wrote about here has been lost to modern macro, or whether it is just implicit in modern macro. I might do a post on this, if I can get my head clearer. On the other hand, I'm writing too many damned theory posts, and should probably write something sensible and useful. If I could.

Mike: "...and I spend part of the first lecture describing how the AD curve differs from a microeconomic demand curve. But perhaps I'm the exception, rather than the rule."

That's exactly what I do too, whenever I do my first macro lecture! (And repeat for the AS curve.) And it's great fun doing this, from first year to graduate level. But we may be alone.

"Nothing I've read so far has changed my opinion that *nobody* understands macro, so I'm still puzzled why us micro folks are being singled out." Fair enough. As Frances says, it's more that I'm criticising people who use micro arguments to answer macro questions, without understanding the difference. The Devil made me pick on microeconomists!

Ralph: If we weren't in a recession, so that AD were no problem, then I agree that the analysis of the effects of minimum wages on unemployment would be very different. I did keep repeating "unemployment due to deficient AD".

Bubble. Sorry. I disagree. You are begging the question that a fall in the price of apples causes an increase in quantity demanded. Assuming the conclusion.

[Bubble Meter Blog]

Nick,

Your response is unacceptable. I'm not begging the question. The scenario you presented was a drop in prices caused by a drop in wage rates. A drop in wage rates causes a drop in marginal costs. Crack open any intermediate or advanced macroeconomics textbook and tell me what is says happens to aggregate supply and demand curves in that situation. You're not suggesting the AS curve is horizontal or downward sloping, are you? Or that the AD curve is vertical or upward sloping? If that's what you're suggesting, then you should be upfront about it. I challenge you to back up your claims with aggregate supply and demand curves.

[Too Much Fed]

MattM, let me enlighten you about what I was saying. I was referring to a time line of spending.

[Too Much Fed]

Nick's post said: "Too much Fed: Sure, you can add time as the third dimension. That's just saying how the curve shifts over time. And you can add lots of other things that affect demand as the 4th, 5th, etc. dimensions. Partly we think of 2 dimensions because damned universities only supply us with 2-dimensional chalkboards and paper. Partly it's because we want to focus on the variable that bears the main burden of adjusting to ensure equilibrium."

Actually, I wanted to look at it from the fungible money supply perspective. That would be:

1) savings spent is past demand brought to the present

2) currency spent is current demand in the present

3) currency denominated debt spent is future demand brought to the present

I'm going to try an example although I may not get it completely correct the first try.

There is a one-person owner and a cut in the minimum wage. The owner is able to cut wages and decides to save the difference to retire earlier. He doesn't need to hire any more workers. He decides his best return on his savings is to invest it as bank capital. The bank then decides to make loans to the employees whose wage(s) were cut to maintain demand.

With that scenario, can you see a likely (but not only) outcome of the owner being able to retire earlier and the workers retiring later if they take the loans to maintain their standard of living in the present?

[Too Much Fed]

I'm not going to challenge Nick, but I would like to see the curves and the assumptions.

[Too Much Fed]

Frances said: ""productivity" (people talk about productivity as an explanation of wages, and it is true in some sense that productivity=output/person, but this ignores the fact that labour's share of output (or the typical worker's share of output - I don't mean CEOs salaries) is changing."

Good comment. Wages are set by supply and demand. If the labor market is oversupplied, do productivity gains go into corporate profits?

[Nick Rowe]

Bubble: read that macro text on *why* the AD curve might have a downward slope. It has nothing whatsoever to do with the income effect in a micro text.

In a micro text, the income effect holds money income constant, and holds prices of other goods constant.

In a macro text (using ISLM for example), a fall in *all* prices, holding the nominal money supply constant, causes an increase in the real money supply, shifting the LM to the right, lowering interest rates and increasing C and I as it causes a movement along the IS curve.

If the AD curve does slope down, it does so for very different reasons from the reasons for the downward slope of the micro demand curve. It's all to do with the demand and supply of money.

The slope of the AD curve is precisely what is at issue in the current debate over whether wage cuts would increase employment in current circumstances. Those who say it would have no effect argue that the AD curve is indeed vertical in current circumstances. They argue that interest rates are already as low as they can go, the LM curve is horizontal, so shifting it to the right will have no effect on AD.

Whether they are right or not is debatable. But arguing from micro-theory on the downward slope of micro demand curves is to miss the point.

[Bubble Meter Blog]

Nick, thank you for the clarification. Perhaps we should follow the scientific method and cut the minimum wage to test the theories. (Half-joking. I'm not actually advocating cutting the minimum wage, but the more empirical knowledge, the better.)

[reason]

Bill Wooley,
it occurs to me, that you might be right - we should cut the minimum wage at the top of the trade cycle. Does that mean we should raise it at the bottom of the cycle?

(P.S. Personally I think a citizens basic income is a better solution than a minimum wage to solve the power disadvantage of low paid workers in the labour market.)

[bill woolsey]

Nick, if you assume a 10% wage cut, and then say that the prices of final goods drop 10%, then you get a 10% drop in nominal income. That isn't because of the 10% drop in wages. It is rather from the simplifying assumption of a 10% drop in prices (and given output, I guess.)

It seems to me that the micro approach shows an increase in real income. I think nominal income is unchanged. What happens to the real demand as real income rises?

The liquidity trap approach is assuming an incease in the real quantity of money. Will the increase in the real quantity of money lead to more spending? No, because it will all be hoarded.

Stock and flow issues?

[Nick Rowe]

Bill:

My preferred approach is to assume initially no change in real output/income, then see if I can derive a contradiction.

In normal times, with an inflation-targeting central bank, it is easy to derive a contradiction. A cut in money wages, means firms want to hire more workers and sell more output, which creates an excess supply of output and puts downward pressure on prices. The central bank responds by doing whatever is necessary to prevent that happening, which means loosening monetary policy if needed to increase AD by the same amount that AS expands. Therefore I have derived a contradiction to my initial assumption that real output stays the same. If both AS and AD increase, real output will increase.

If someone who believed in the liquidity trap tried o run the same argument, they would argue that they cannot derive a contradiction. Since the CB can't expand AD.

Now, like you, I don't believe that a liquidity trap necessarily means "game over". There *is* a real balance effect, that real balance effect *isn't* necessarily just a wealth effect, and anyone who says it must be tiny is in any case possibly making a stock-flow mistake, as you say. In principle, a $1 excess supply of money could create $ trillions of excess demand for goods.

But I ducked the question of whether or not the AD curve does or does not slope down in a liquidity trap. I'm going to keep on ducking it, till I get my head clearer.

[Too Much Fed]

bill woolsey said: "The liquidity trap approach is assuming an incease in the real quantity of money. Will the increase in the real quantity of money lead to more spending? No, because it will all be hoarded.

Stock and flow issues?"

Could it be wealth/income inequality issues?

[Too Much Fed]

Nick said: "A cut in money wages, means firms want to hire more workers and sell more output, which creates an excess supply of output and puts downward pressure on prices. The central bank responds by doing whatever is necessary to prevent that happening, which means loosening monetary policy if needed to increase AD by the same amount that AS expands."

Depending on how money wages is defined, that looks to me like negative real earnings growth on the workers.

[Too Much Fed]

Nick said: "Therefore I have derived a contradiction to my initial assumption that real output stays the same. If both AS and AD increase, real output will increase.

If someone who believed in the liquidity trap tried o run the same argument, they would argue that they cannot derive a contradiction. Since the CB can't expand AD."

It seems to me it depends on whether the group experiencing negative real earnings growth is willing to go further into currency denominated debt and if that group can make the interest payments.

[Doc Merlin]

@reason

How do you know if you are at the top or if you are at the bottom of the cycle. Generally these things are only observable in retrospect. (This is another problem I have with attempts at stimulus).