Interest rates and Aggregate Demand

What happened in 2008? Why didn't the cut in interest rates prevent Aggregate Demand from falling? Was it just that the cut in interest rates wasn't big enough? Or is the rate of interest the wrong thing to look at? Because it's only a relative price, and relative prices only matter for relative demand?

Suppose we had a monetary economy that produced apples and bananas. Let (Pa/Pb) be the relative price of apples in terms of bananas. Would we draw an AD curve with (Pa/Pb) on the vertical axis? Would we say that a cut in (Pa/Pb) would cure a deficiency in AD? Of course we wouldn't say that.

Suppose we had a monetary economy that produced goods this year and goods next year. Let (1+r) be the relative price of goods this year in terms of goods next year. Would we draw an AD curve with (1+r) on the vertical axis? Would we say that a cut in (1+r) would cure a deficiency in AD? Of course we wouldn't say that. Hmmm. That's exactly what many of us do say.

What's the difference?

Say's Law is false in a monetary exchange economy. In a monetary economy that produces only apples and bananas, it is perfectly possible to have an excess supply of both apples and bananas. In a monetary economy that produces only goods this year and goods next year, it is perfectly possible to have an excess supply of both goods this year and goods next year. Because there's an excess demand for money, the medium of exchange. And if that's the problem, a cut in a relative price like (Pa/Pb), or a cut in a relative price like (1+r), is not the cure.

There is no difference.

Suppose we had a model that just assumed, without explaining why, that the market for bananas always cleared, so there was never any excess supply or demand for bananas. In that case we might put (Pa/Pb) on the vertical axis of the AD curve. But that would be a very strange macro model of a monetary economy. It's a micro, partial equilibrium model, of the market for apples.

Suppose we had a model that just assumed, without explaining why, that the market for goods next year always cleared, so there was never any excess supply or demand for goods next year. In that case we might put (1+r) on the vertical axis of the AD curve. But that would be a very strange macro model of a monetary economy. It's a micro, partial equilibrium model, of the market for goods this year.

Hmmm. That's the standard New Keynesian model. It just assumes, without explaining why, that the future output gap approaches zero. The standard New Keynesian model is a very strange macro model of a monetary economy. It's a micro, partial equilibrium model.

Be that as it may: if the market for bananas did in fact always clear, a micro partial equilibrium model which just assumed it always cleared, without explaining why, might work quite well in practice. And if the market for goods next year did in fact always clear, a micro partial equilibrium model which just assumed it always cleared, without explaining why, might work quite well in practice.

Micro, partial equilibrium models of the market for apples, or the market for goods this year, usually work quite well in practice.

Until one year, when the market for bananas does not in fact clear, and the model fails. Then we need a truly macro model of a monetary economy. Because a cut in (Pa/Pb) might simply reduce the demand for bananas, rather than increasing the demand for apples? We need a truly macro model to answer that question.

Until one year, when the market for goods next year does not in fact clear, and the model fails. Then we need a truly macro model of a monetary economy. Because a cut in (1+r) might simply reduce the demand for goods next year, rather than increasing the demand for goods this year? We need a truly macro model to answer that question.

I've talked about only two goods, apples and bananas, but there might be many goods.

I've talked about only two years, this year and next year, but there might be many years.

If people always do in fact expect "full employment" (zero output gap) in the near future, the New Keynesian model might work quite well in practice, even though it is a partial equilibrium model and not a truly macro model of a monetary economy. A cut in real interest rates will cure an excess supply of goods this year.

But if people do not expect a return to full employment in the near future, and fear a continuing recession, the New Keynesian model might fail. A cut in real interest rates might not cure an excess supply of goods. Interest rates are the wrong thing to look at. Interest rates are a relative price that only affect relative demand, not Aggregate Demand for goods this year and goods next year.

Comments

[Ralph Musgrave]

To put Nick's final paragraph another way, an interest rate cut pulls demand forward in time. As Mervyn King pointed out, that effect works only for a limited time: at rough guess a couple of years. And that makes our continued low interest rates a nonsense.

Worse still, I’m baffled as to where the logic is in channelling stimulus into an economy just via extra investment. Why not do it via massage parlors, car production, and/or subsidised skiing holidays?

[Nick Rowe]

Ralph: without a truly macro model of a monetary economy, all we know is that a cut in real interest rates *may* only pull demand forward in time. We know a cut in real interest rates will increase the ratio (A/B). But there are three ways to increase a ratio of (A/B):1. increase A leaving B constant;2. reduce B leaving A constant;3. increase A and reduce B leaving A+B constant. (Or loads of other ways too, of course).

The NK model assumes 1. You are assuming 3. It might be 2. It could be anything. We need a truly macro model to know which it will be. It depends on what the central bank is holding constant.

Your second paragraph is wrong, but it's a red herring here.

[Frank Restly]

Nick,

"Interest rates are a relative price that only affect relative demand, not Aggregate Demand for goods this year and goods next year."

That would depend on your credit model for interest rates. If I borrow from you at 10% per year and you borrow from me at 10% per year then our aggregate income rises by 10% per year. The 10% interest payments we make to each other do not cancel each other out in aggregation, they sum.

If instead, I borrow from the central bank at 10%, then the money that I borrow and spend adds to aggregate demand in the period it is borrowed but the interest and principle repayment subtracts from aggregate demand in future periods.

[Nick Rowe]

Frank: no.

[Tom Brown]

Nice... you just had to write half and article, cut and paste, and then change a few words.

[Nick Rowe]

Tom: yep, that's pretty well exactly what I did, except I did it paragraph by paragraph. I learned that from Dierdre McClosky: express parallel ideas in a parallel way (something like that), so the reader can see the parallel. And my whole argument in this post (my rhetoric, or way of persuading the reader) was based on that parallel.

[Mike]

"Or is the rate of interest the wrong thing to look at?"

The rate of interest on reserves is the wrong thing to look at. The rate of interest on broader money is more relevant to the broader economy. The rate of interest on broader money is nowhere near zero.

[Nick Rowe]

Mike: sure, but this post is not about interest rate differentials!

Noses down on the scent trail you beagles, to hunt the hare, and not sniffing around to find what else smells interesting!

[Tom Brown]

It works well! If I knew Woodford's email (like JP Koning does apparently... have you seen his latest?) I'd alert him. Since I don't I alerted commentator "DOB" who seems to have an NK flavor of model going on (from what I can tell). I'd like to see his response to this:

http://catalystofgrowth.com/

Plus I responded to a question about your comment at pragcap. You may want to correct my libelous statements.

[Kevin Donoghue]

"But if people do not expect a return to full employment in the near future, and fear a continuing recession, the New Keynesian model might fail."

I'm not sure you need the "might" in there. If the NK model somehow predicts correctly when a key assumption is false that's a fluke, not a success. I think you ask too much of NK. It's a model of small variations around a steady state. (Didn't Leijonhufvud have some notion of a corridor?) If you want to think about what happens when things really go to hell, Keynes is your only man.

[Paul Friesen]

Nick:

"Suppose we had a monetary economy that produced apples and bananas. Let (Pa/Pb) be the relative price of apples in terms of bananas. Would we draw an AD curve with (Pa/Pb) on the vertical axis? Would we say that a cut in (Pa/Pb) would cure a deficiency in AD? Of course we wouldn't say that."

Actually, I would say that, under certain circumstances. If bananas were the commodity that people used as a store of value, such a relative price cut could cure a demand deficiency. If bananas don't spoil and people keep more of them around than they want to eat just to be able to buy apples with them whenever they want apples, a cut in the price of apples will make those stores of bananas worth more. If we assume that people have a certain level of those stocks of bananas that they want to keep around, and that this amount goes down when some overall price index falls, the stocks of bananas will be too large after the apple price cut, so people will spend more bananas than they get to draw down the stocks.

If people don't eat bananas at all, and use them only as a medium of exchange and a store of value, they become like money. I think the store of value function of money is more important than you seem to be saying in some previous posts.

[Nick Rowe]

Kevin: "[The NK model] It's a model of small variations around a steady state."

I would put it slightly differently (though not inconsistent with what you wrote): It's a model of *short-lived temporary* variations around a steady state.

Am I asking too much of the NK model? maybe. But in the olden days we used to talk about M/P needing to adjust to get the economy to full employment. And this mechanism would apply just as much if we are talking about permanent deviations as it would to temporary deviations. Actually, if P is slow to adjust, it should be even more applicable to permanent deviations.

Interpretation of Leijonhufvud, to make it fit with this?: if there's a small deviation, people don't change their expectations of the future. If there's a big deviation, they all think they might be eating squirrels in future. ?

Paul: "If bananas were the commodity that people used as a store of value, such a relative price cut could cure a demand deficiency."

Change "store of value" in that sentence to "medium of exchange" and you've got it. It doesn't matter if bananas are a store of value, if people don't buy apples with bananas. But if people buy apples with bananas, it doesn't matter if they are a store of value.

[Mike]

"In a monetary economy that produces only goods this year and goods next year, it is perfectly possible to have an excess supply of both goods this year and goods next year. Because there's an excess demand for money, the medium of exchange"

Why is there an excess demand for money? How does an excess supply of goods result in excess demand for money?

[Nick Rowe]

Mike: I find it easier to think of it the other way around: start in equilibrium, then burn half the money in everyone's pockets (or double demand for money instead, if you like). Hold prices sticky. There is now an excess demand for money. There are two ways to get more money in your pocket: sell more goods; or buy fewer goods. The apple producers supply (try to sell, but they will fail) more apples, and demand fewer bananas. The banana producers supply (try to sell, but they will fail) more bananas, and demand fewer apples. There is now an excess supply of both goods. Since demand for both apples and bananas drops, actual sales of apples and bananas falls, even though supply increases. Since trying to sell more apples and bananas failed, people will give up on that strategy to get more money, and will demand even fewer apples and bananas.

[Paul Friesen]

Nick:

Don't get it.

Surely the store of value function is critical. It's the fact that the buying power of all those stores of value goes up that gives the stimulus. The only reason that the medium of exchange function is important here is that people tend to use the medium of exchange as a store of value.

[Nick Rowe]

Paul: we are talking about money, not wealth. Money is a very small part of our wealth. But it's a very special and important part. Because: all prices are measured in money; we buy and sell everything else for money.

[Frank Restly]

Paul,

The value of any medium of exchange is a function of many things:

1. Is it easily portable? - Boulders versus Sea Shells
2. Is it easily divisible into identifiable smaller units? - Half a rock is still a rock, half a cow is a bloody mess
3. Is it durable - corrosion and weather resistant?
4. Does it retain purchasing power relative to other goods?
5. Is it liquid - commonly used as a medium of exchange for other goods?

Purchasing power and liquidity can move in tandem or they can move in opposite directions.

[Tom Brown]

Frank, don't mean to butt in here, but you're item 1. made me think of the Yapese money stones:

http://upload.wikimedia.org/wikipedia/commons/0/0f/Presentation_of_Yapese_stone_money_for_FSM_inauguration.jpg

I saw some when in Palau (the Palau people didn't use them, but the Yapese would canoe over to use Palau's rocks to carve their "money" with... they left the broken ones behind). Here's another good pic:

http://media.npr.org/assets/img/2010/12/09/yap-3de5582da4f8b2e30812beb5aee6e672d3edde00-s6-c30.jpg

[Tom Brown]

"2. Is it easily divisible into identifiable smaller units? - Half a rock is still a rock, half a cow is a bloody mess"

Yes, a delicious bloody mess. :D

[Nick Rowe]

Frank: stop.

[Determinant]

Right on, Nick. Now if we can only get past this "Central Bank as God" thing, and recognize that in a poker game, er, expectations management, people can and do call the central bank's bluff, then the central bank is not omnipotent. Very strong perhaps, but not infallible. So if we need demand management, and the central bank has a busted flush, we need to roll out the Fiscal Cannon and fire off some good old deficit stimulus.

[Nick Rowe]

Determinant: nope. You have no reason for believing fiscal will work either. Because you don't have a theory of what determines AD.

Plus, monetary policy is not interest rate policy.

[Ritwik]

"Interest rates are a relative price that only affect relative demand, not Aggregate Demand for goods this year and goods next year"

Umm, no. A decision to consume less today is not fully expressed as a decision to consume more tomorrow.

Why? Incomplete asset markets. Lack of self-coordinated optimal equilibria in the numeraire. Rate of adjustment of output and employment may be faster than rate of adjustment of interest rates, so that at the lower level/path, there is no further tendency for interest rates to adjust (which is turn, ma be driven by the interest-elasticity of numeraire demand). Etc...

You may be right that a standard neo-Walrasian New Keynesian model finds it hard to explain, within its own circumscription, why an "excess demand for bonds" does not auto-correct. But Woodford's model provides the first clue. And then if you tack the simple idea that the strategy space of financing is different from the strategy space of saving, leading to multiple equilibria in risk premia, some sub-optimal, you're there.

Keynesian theory, old or new, is about why the economy refuses to move to the new IS curve without shouting and screaming. Granted, NK theory doesn't do a exceedingly good job of explaining exactly why. Neither did the NK synthesis of Solow-Samuelson-Modigliani.

But if you have a theory where the economy is always on the IS curve because the IS curve is only about relative prices, and today vs tomorrow is analogous to barter between apples vs bananas, you're not quite engaging with the core of Keynesian theory, old or new.

[Nick Rowe]

Ritwik: "Umm, no. A decision to consume less today is not fully expressed as a decision to consume more tomorrow."

That is sorta like the point I was trying to make.

"Keynesian theory, old or new, is about why the economy refuses to move to the new IS curve without shouting and screaming."

I don't read it that way. Most Keynesian models, Old or New, assume the economy is always on the IS curve. Old Keynesians had a story about how the economy gets onto the IS curve (the multiplier process, where a fall in desired expenditure causes a fall in income which causes a further fall in desired expenditure etc, that eventually converges on the IS curve). New Keynesians just assume the economy is always on the IS curve, but that Euler IS curve is very different from the IS curve of ISLM. With the IS curve of ISLM, if you knew the path of interest rates, and you knew the path of shocks to the IS curve, you knew the path of Y(t). With the Euler IS you don't, unless you just assume without explanation that the economy always approaches full employment. Because that Euler IS only determines the *ratio* of this year's to next year's demand, given the rate of interest.

[Nick Rowe]

Frank: since you won't take a polite hint: too many of your comments are very low quality comments. Like that last one. You are wasting space and reducing the quality of the conversation here, because people have to sift out your comments to get to the better quality ones. Plus, you are wasting my time.

[Nick Rowe]

Ritwik: let's put it this way:

In a barter economy with only 2 goods, a supply of apples *is* a demand for bananas, and a supply of bananas *is* a demand for apples. If you have an excess supply of apples you must have an excess demand for bananas, (they are the same thing), and a cut in (Pa/Pb) is the cure.

Now change "apples" to "goods this year", and "bananas" to "goods next year", and "(Pa/Pb)" to "(1+r)", and it's also true.

In a barter economy, a decision to consume (or invest) less than you produce this year *is* a decision to consume more than you produce next year. So if there's an excess supply of goods this year, the cure is a cut in the real rate of interest.

But in a monetary exchange economy, that's all wrong.

So, somehow I think we are agreeing, but talking past each other. Dunno.

[Nick Rowe]

Ignore this comment if you already understand what real interest rates are.

Let r be the real interest rate, i the nominal, p the inflation rate, P0 the current year's price level, and P1 next year's price level.

r is defined implicitly by: (1+i) = (1+r).(P1/P0) = (1+r).(1+p) = 1+r+p+r.p which equals approximately 1+r+p when r and p are both small.

So r=i-p is an approximation (except in continuous time).

If you own 1 bottle of beer today, you can sell it for P0 dollars today, lend those dollars and get P0.(1+i) dollars next year, and use them to buy P0.(1+i)/P1 beers next year. So the price of one beer today in terms of beers next year is (1+r)=(1+i)/(P1/P0).

[Nick Edmonds]

Imagine an economy with a large number of overlapping generations. Each generation works and earns for a period of time, then lives off savings till they die. Assuming various things (life expectancy, productivity, etc.) stay the same, there is a steady state where aggregate income equals aggregate consumption and the balance of aggregate savings is constant.

Say savings take the form of IOU claims on a neighbouring economy. The balance of savings varies if there is a trade imbalance. If consumption in our economy drops, there will be an accumulation of IOUs. In steady state, current flows between the economies must be in balance.

Now, imagine a change in the interest rate paid on IOUs causes the agents of our economy to prefer to consume less in their early years and more in later life (and I'm not saying anything here about whether I think that is a realistic assumption). The new steady state still involves the same level of aggregate income and aggregate consumption (since these are dictated by the level that achieves balanced trade), but a higher level of aggregate savings is needed. There therefore needs to be a finite period of accumulation of additional savings (or an increase in real savings caused by price changes). The attempt to accumulate additional savings may cause a recession.

I know you like to have money in the story, so let's say the IOUs function as the MOE.

This tends to be how I look at things. In this framework, I can see that how interest rate changes might affect aggregate demand. Exactly how depends on what you belief about agents' preference functions. I've been trying to square this with your analysis, but I haven't really been able to.

[Nick Rowe]

Bill Woolsey has a very good clear post on this connecting all the dots.

[Nick Rowe]

Nick Edmonds: that is an analysis of a small open economy that is hit by an exogenous shock of an increase in world real interest rates (I think). That's fine, but a small open economy is partial equilibrium analysis. Analytically, it's the same as an individual experiment, where an individual responds to higher real interest rates by consuming less today and more tomorrow, assuming it has no effect on his time-path of income from production. We can't use that same analysis for a general equilibrium experiment, like the world economy.

[Nick Edmonds]

OK. Then replace the neighbouring economy with a government. Fiscal policy involves a fixed amount of spending and a fixed rate of tax, with any deficit funded by issuing IOUs. (Actually, this is the way I usually tell the story, for example in the second scenario here http://monetaryreflections.blogspot.co.uk/2013/08/financial-assets-and-national-wealth.html which was prompted by another post of yours). So we have a closed economy, but otherwise everything is the same - steady state income and consumption dictated by fiscal stance and steady state savings dictated by the life cycle and preferences of private agents.

[Nick Rowe]

Nick Edmonds: I had a quick read of your post. I read your model as a version of Samuelson 1958.

Here's my old post on that model

In Samuelson 1958, there is a big Pigou real wealth effect from "money". The bigger is M/P, the higher is the equilibrium real interest rate. It's a bit hard to translate that OLG model to compare it to a NK model, but one day I should perhaps do that.

[Nick Edmonds]

Thanks for the reference. I'll look at Samuelson and your post.

I just read your Bill Woolsey link as well and I found it quite useful for seeing how my approach fits with what you are saying. I tend to look at financial assets more generally where you look at money, but otherwise I can see parallels.

Just for the record, I am most definitely not suggesting here that I think that interest rates are an effective way of stimulating demand.

[Major_Freedom]

Say's Law is true in any market. There is no such thing as an "excess" demand for money and "insufficient" demand for goods.

There is no such thing as a general overproduction. There is only ever a partial relative overproduction, and a corresponding AND UNSEEN partial relative underproduction.

When you OBSERVE the economy suddenly developing a "glut" of inventory sitting on shelves that you LOOK, as well as an OBSERVED reduction in spending, such that "velocity" slows down over time, you are not OBSERVING a "general" glut.

What you are OBSERVING is everything that was partially relatively overproduced.

The partial relative underproduction of goods is what you CAN'T observe. These are the potential goods that could have been produced AND SOLD, but weren't produced and sold, because scarce resources and labor went into producing the goods you end up SEEING.

The reason why we OBSERVE reduced spending and shelves stocked with goods that aren't being sold as quickly as expected by the sellers every now and then, is not because "too many goods were produced and not enough money was produced." It is because too many goods were produced of certain types, available at certain time horizons, at certain places, and at certain relative prices.

If the world were different, and investors and producers made other choices, and brought to market goods of other types, at other time horizons, at other places, at other relative prices, then we would not have OBSERVED falling spending and rising inventory waiting times in the aggregate.

I know this because the desire for goods on the part of mankind is practically infinite. No matter how wealthy we get, there is always a desire for better goods, bigger goods, more goods, etc. We're more than ten times wealthier than people 300 years ago, and yet we do not feel like we're satisfied.

And why is that? Because we're ACTORS. Action is the purposeful attainment of ends that are superior than ends that would be achieved with no action. We act because we're not satisfied with the world as it is. That is where wealth generation comes from.

[jt]

Nick, given your recent posts on bank MOE etc. Do you think a monetary exchange economy for the comparison is the "right" model? I guess this is asking the same old question, viz a viz monetary exchange vs. credit economy with endogenous (inside) money. Wasn't sure whether your bank MOE posts were signalling a slight shift in thinking. Cheers.

[Mikael]

Nick, good post, but I must disagree. You say "A cut in real interest rates might not cure an excess supply of goods...Interest rates are a relative price that only affect relative demand, not Aggregate Demand for goods this year and goods next year.", which I would agree with if you were referring solely to nominal interest rates, not real interest rates. It seems to me that positive inflation expectations infer the expectation of future "full employment" (zero output gap), and a consequent bidding up of prices. Expecting positive inflation and expecting goods and services to remain unsold are opposing and mutually exclusive expectations.

[apt]

A few ideas/questions:

- the first one is when you say that "here's an excess demand for money, the medium of exchange. And if that's the problem, a cut in a relative price like (Pa/Pb), or a cut in a relative price like (1+r), is not the cure." then you are actually saying that back then more money should have been printed? MM-like analysis?

- the second concerns your analysis of the AD curve and changes in realtive prices. In standard, basic microeconomics, which is what you claim the model tends to behave like, a change in relative prices causes both an income effect and a substitution effect. You cannot despise the income effect, and that is why I do not think your analysis is precise.

- regarding the second point, I could only think of something that might just back up your idea partially. It is quite austrian I guess, but if we are changing relative prices artificially, we would just be causing a misallocation of spending, namely through investments. And that could be the cause why it is unnefective: maybe it is just increasing aggregate demand where that demand is not needed. And that could lead to, at least temporary or cyclical supply shortages in certain markets, and big gaps in the others, which are certainly the majority.

[Nick Rowe]

MF: "The partial relative underproduction of goods is what you CAN'T observe."

Yes we can. If I see a queue of people lined up trying to buy meat, and they run out, I am seeing an excess demand for meat. I saw this sort of thing a lot in Cuba. I rarely see it in Canada.

Start in equilibrium, where supply = demand in all markets. Now suppose that half the stock of money in people's pockets magically disappears, and that all prices are sticky. I say people want to sell more goods and/or buy fewer goods, as each individual tries to increase his stock of money again. So there's an excess supply of all goods. You are saying that's logically impossible?

Of course human wants may be insatiable, but so what. I would like a Lotus. If the price of a Lotus were lower, or if I were richer, I would go to the dealer and demand one. But they aren't, and I'm not, so I don't.

jt: you really lost me there.

1. My recent bank/MOE posts were signalling no shift in my thinking.
2. Of course I think it is right to model the economy as a monetary exchange economy, as opposed to a barter economy, or an economy with a single centralised Walrasian market. Because that's what we have.
3. Monetary exchange economy vs credit economy is a false dichotomy. You can have a monetary economy with or without credit, and a credit economy with or without money. And if it is a monetary economy, the good people use as money may or may not be an IOU.
4. When people talk about "endogenous" money, I don't think they know what that word means. All "endogenous" means is "depends on stuff in the model".

Mikael: I was talking about real interest rates. Keynesians, especially New Keynesians, say AD depends on real interest rates.

apt: (numbering your paragraphs):

1. Yes.

2. standard basic micro is wrong about income effects. If 100 apples are being bought and sold, and the price of apples rises $1, the buyers of apples are $100 poorer and the sellers of apples are $100 richer. There is a change in the *distribution* of income, but no change in income.

3. the producer of any good has to choose either price or quantity (or something) of the good he produces. There's nothing "artificial" about it. That's true whether or not the good is money, or the producer is the government.

[sustain_ability]

It's the vision thing; the economic development model post WWII (to which we all refer to, consciously or not) has faltered, and is threatening the entire planet's wellbeing.

Globalization without diversification* sees an increasing number of people chasing an increasingly smaller resource-based wealth dream.

As people lose their material dreams, few are able of quantifying their lives with a non-material lifestyle. Statistics are rarely correct; it's the immediate state or condition of the vision thing that matters, plus your ability to launch an initiative.

*Diversification (e.g the "energy" sector) implies complexity, yes, but it makes for a multiple moving target, i.e. the cost of disabling it becomes prohibitive.

We implicitly rejected complexity (e.g. externalized costs) when we decided to store nuclear wastes indefinitely. Had we diversified from the beginning, the amount of wastes would have been much smaller, and we likely would have found a cheap(er) way of getting rid of them, e.g. launching them into outer space.

[Major_Freedom]

Nick:

"MF: "The partial relative underproduction of goods is what you CAN'T observe."

"Yes we can. If I see a queue of people lined up trying to buy meat, and they run out, I am seeing an excess demand for meat. I saw this sort of thing a lot in Cuba. I rarely see it in Canada."

What I was talking about are all the goods that didn't get produced because the goods you see, like the meat in your example, were instead produced using the same scarce resources and labor. These are the goods you can't see. You can't see them because they don't exist. But the non-existence of what could have been produced is crucial to understanding Say's Law.

"Start in equilibrium, where supply = demand in all markets. Now suppose that half the stock of money in people's pockets magically disappears, and that all prices are sticky. I say people want to sell more goods and/or buy fewer goods, as each individual tries to increase his stock of money again. So there's an excess supply of all goods. You are saying that's logically impossible?"

Yes, because logically speaking, if you propose the existence of money, then you can't at the same time propose the non-existence of money. Does money supply M exist or does some other money supply exist? Prices are a function of money supply that exists. What you are doing in this example is proposing a matrix of prices that are a function of a money supply M that you also deny exists in the same example, because you're including money supply M/2. This is not logical.

Don't you see the significance of how you have to propose MAGIC in order to create this conception of a general overproduction? That should tell you that your critique of it is weak.

"Of course human wants may be insatiable, but so what."

So what? That is EXACTLY the foundation for why Say's Law is true. All Say's Law says is that a GENERAL overproduction of goods is impossible, precisely because human wants are insatiable!

"I would like a Lotus. If the price of a Lotus were lower, or if I were richer, I would go to the dealer and demand one. But they aren't, and I'm not, so I don't."

You can't afford a Lotus because there is a lack of capital. There is a lack of capital because human wants are greater than what any existing supply of capital is capable of helping laborers and capitalists produce.

Not sure how this is relevant to your point.

[Major_Freedom]

Nick:

A Say's Law type analysis of your relative underproduction of meat example would be that yes, partial relative underproduction of meat is possible, with a corresponding partial relative overproduction of other goods.

But if you consider all individuals, then there is not only a want for more meat, but there is also a want of other goods that could not have been produced by virtue of the fact that everything that was produced, including those goods that were partially relatively overproduced by virtue of the meat being relatively underproduced, were indeed produced.

[sustain_ability]

Major_Freedom:

You're in deep water if you apply your prescient analysis to "intellectual" assets!

[Patrick]

Can't resist a short OT.

Nick: "I would like a Lotus ..."

Me too! I drove past a shiny new red Evora yesterday. Such a pretty car. A little pricey new, but second hand and a couple of years old they're not so bad. And maybe one day we'll be able to import them from the US and save $25K.

[Nick Rowe]

MF: if all prices were perfectly flexible, we would never observe any market in excess supply or excess demand. But they aren't, so we do.

It is very easy to get all goods (except money) in excess demand. Cuba shows us how. Hold all prices fixed, then get more money into people's hands (one way or another) than they want to hold. Excess demand for all goods, and excess supply of money. Just do the opposite of what Cuba did, and do it quickly so prices don't have time to adjust, and you get the opposite results: excess supply of all goods, except money.

Sure, more production is great, as long as human wants are unsatiated. But if money supply is insufficient to satisfy money demand, those extra goods will rot unsold. Overproduction isn't the problem; it's underproduction of money.

[Mikael]

Nick, I know you are talking about real interest rates, which is why I was desputing your claim about real interest rates only being a relative price. The expectation of inflation which lowers the real interest rate below the nominal interest rate, infers the expectation of a future zero output gap. Consequently, the real interest rate is not just a relative price, but also holds information on expectations of absolute spending. Cheers!

[Nick Rowe]

Mikael: OK, I sorta see where you are going. But somehow I thinks that's adding a cause (r) and a consequence (inflation).

[Mikael]

Nominal r is just as much a consequence as inflation

[Major_Freedom]

Nick:

"MF: if all prices were perfectly flexible, we would never observe any market in excess supply or excess demand. But they aren't, so we do."

All you're saying with that comment is that in a world of perfect equilibrium, there is perfect equilibrium, but because we don't live in a world of perfect equilibrium, we don't live in a world of perfect equilibrium.

A world of infinitely flexible prices is a world where relative prices are ALWAYS reflective of general productive equilibrium.

At any rate, we would still live in a world where general overproduction does not occur. Imperfectly flexible prices are not something that brings about general overproduction. Imperfectly flexible prices is a reflection of the fact that humans act through time, rather than all at once.

"It is very easy to get all goods (except money) in excess demand. Cuba shows us how. Hold all prices fixed, then get more money into people's hands (one way or another) than they want to hold. Excess demand for all goods, and excess supply of money. Just do the opposite of what Cuba did, and do it quickly so prices don't have time to adjust, and you get the opposite results: excess supply of all goods, except money."

This is not an example of general overproduction of goods either. The fact remains that the desire for goods is practically infinite. Your example is just another example, like the first, that there is a partial relative overproduction of goods - that you can see, which there is an insufficient demand - and a corresponding partial relative underproduction of goods - that you can't see, which there is an excess demand. This insufficient demand for goods that you can see, and the corresponding excess demand for goods that you cannot see, is why you see dollar spending slow down (what you call an excess demand for money), and store shelves remain full.

What is happening here is that people are holding onto money for longer because they want more goods that don't exist, as the scarce resources and labor went into producing the excess goods that do exist. They are not holding onto money for longer because their desire for goods as such is satisfied. The wrong goods were produced. Too many goods X, not enough goods Y (that you cannot see).

"Sure, more production is great, as long as human wants are unsatiated. But if money supply is insufficient to satisfy money demand, those extra goods will rot unsold. Overproduction isn't the problem; it's underproduction of money."

No, the problem there is partil relative underproduction of goods. People do not want money because they want to eat money. People want money because they want to buy goods. People stockpile money when the world has too many goods they don't want, and not enough goods they do want.

If I lived on an island, and another inhabitant produced coconuts that he thought I wanted, but I actually do not want, then the fact that coconuts are the only goods that were produced that day does not at all provide you with an example of.a general overproduction of goods. For you would be making the same error as above, namely, you would be ignoring the goods that you can't see, namely, the goods that I want but don't exist. I don't want more coconuts. I want more meat. If instead the other inhabitant produced animal meat for me instead, then I would have accepted it, and then we could say that "the goods produced in the aggregate were demanded in the aggregate."

Just because I don't want the coconuts, it doesn't mean there is not enough medium of exchange! I am holding onto my money on that island not because there were too many goods produced in general. It is because too many coconuts were produced and not enough meat was produced.

Just because I say no to an unsold supply of goods it doesn't mean there are too many goods AS SUCH.

[Brian Romanchuk]

Hello Nick,

I have a fairly basic question, which is slightly related to your post topic. By way of introduction, I am applied mathematician who is trying to come to grips with the DSGE literature. (Actually, I have a lot more questions, but I will start with the easiest.)

In the models that I have seen (both RBC and New Keynesian), (e.g., the text by Gali), there is a market-clearing assumption Y(t) = C(t). This implies that there is no physical inventory of goods (or at least, it is fixed). Has this not just assumed away the existence of an inventory cycle?

This also makes it hard for me to understand how production is determined - since the markets clear by assumption, only one entity can really have a choice in the level of production. As far as I can tell, the entity that chooses is the household sector, and the firm has no choice but to go along with the household sector decision.

[Gene Callahan]

"Yes, because logically speaking, if you propose the existence of money, then you can't at the same time propose the non-existence of money. Does money supply M exist or does some other money supply exist? Prices are a function of money supply that exists. What you are doing in this example is proposing a matrix of prices that are a function of a money supply M that you also deny exists in the same example, because you're including money supply M/2."

Nick, the Major also thinks it is "illogical" to give an example in which yesterday it was raining a lot, but today it has stopped, because you would be proposing a supply of rain that both exists and doesn't exist.

[Nick Rowe]

Brian: yes, this does assume away the inventory cycle, for simplicity. You can think of all goods as being services like haircuts, or all goods being produced to order.

But this is *not*, in a NK model, a market-clearing assumption. (It is a market clearing assumption in RBC models). In a NK model, what C=Y means is that haircuts produced/bought/sold equals haircuts demanded. It says nothing about the *supply* of haircuts ("supply" means the quantity firms *want* to sell).

In the NK model, firms are monopolistically competitive, and set a price to maximise profits, and produce and sell as much as buyers want to buy at that price. But they cannot always adjust price in response to changes in demand (by assumption).

this old post, and this old post may help you make sense of this.

Gene: I confess I find it hard to follow MF's argument at times, and have given up. You may be right. I thought he was saying that prices are perfectly flexible, so that if M changed P would have to change immediately too. Which is at least logical, but then whether *all* markets can be in excess supply at the same time becomes rather a redundant question, because all markets would always clear.

[Kevin Donoghue]

"I confess I find it hard to follow MF's argument at times, and have given up."

Same with me and Heidegger.

[Brian Romanchuk]

In response to Nick Rowe,

Thanks for your response. I will try to digest the articles and compare to the basic models in the text by Gali. As an outsider to the literature, it does not appear elegant that the firms use marginal logic to detrmine their options, while the representative household looks at the global solution to determine the optimising solution.

[Nick Rowe]

Brian: "As an outsider to the literature, it does not appear elegant that the firms use marginal logic to detrmine their options, while the representative household looks at the global solution to determine the optimising solution."

? Individual firms set prices to maximise present value of profits profits taking all things into account; individual households choose consumption to maximise present value of utility taking all things into account.Both firms and households use marginal logic to solve their problems.

[Brian Romanchuk]

I will have to write out in full what is confusing me (with actual equations and stuff like that). Since it probably would take quite a bit of space to explain my thinking properly, I will probably stick the full version of my question on my blog (I need content...).

But to summarise what is puzzling me - the "no-ponzi" condition (as well as other constraints?) puts a global constraint on the solution, and so the maximisation needs to take into account those constraints, for both the firm and household. It may be that the constraints are embedded in the functions you are taking derivatives of, but if that is true, I missed how that was done.

The constraints should not matter for a linearised system, but you need to calculate the solution before you linearise. My copy of Gali is somewhere in a box (renovating my basement), so I will not be able to look at this for at least a few days.

It may be that your posts will have clarified my thinking; I will let you know after I look at the system equations again. Once again, thanks for the response.

[Nick Rowe]

Brian: "But to summarise what is puzzling me - the "no-ponzi" condition (as well as other constraints?) puts a global constraint on the solution, and so the maximisation needs to take into account those constraints, for both the firm and household."

BINGO! (I think)

That (in slightly different words) is what I have been complaining about, when I complain that NK models "just assume" the people in the model expect the economy approaches zero output gap/full employment in the limit as time goes to infinity. If every other individual has this expectation, then my satisfying my "no ponzi" condition means that my planned consumption path satisfies this condition too. But if we all violate this condition together, then no individual violates his own no ponzi condition. It's like a collectivist version of the no-ponzi condition: "let's all agree to choose a consumption path that doesn't cause the economy to explode or implode, OK?".

I don't think this has anything to do with linearity.

[Jacques René Giguère]

Nick: "? Individual firms set prices to maximise present value of profits profits taking all things into account; individual households choose consumption to maximise present value of utility taking all things into account.Both firms and households use marginal logic to solve their problems."
My parents were in business. 20 years later I was. We had no idea of our marginal costs, my parents didn't even knew the concept. We guessed about the future and groped the present. The lucky ones in the industrial park survived,helped by somewhat better management. (That company was sold and no longer exist...)
As said in yesterday's NYT "...statistician George E. P. Box: “All models are wrong, but some are useful.”
We get to the result of marginal thinking because it is efficient and only the efficient survives. But we don't think like that. That's why we have to teach it. We don't teach people to breathe.

[Jon]

Hey nick,

Can we get some further discussion on http://johnhcochrane.blogspot.com/2013/09/the-new-keynesian-liquidity-trap.html?m=1

?

[Nick Rowe]

Jon: done!

[Tom Whelan]

You say Say's Law is false in a monetary economy. Why? I'm not buying your argument today, and it's not because I want to hold onto the money.

[Kevin Donoghue]

Tom, your refusal to buy Nick's argument in no way implies a willingness to buy another, newly-produced.

[Nick Rowe]

Kevin: yep, because he might wish to hold more money instead.

Take a simple example. Two goods, apples and bananas.

In a barter economy, where people swap apples for bananas, if I am offering apples for sale I am by definition trying to buy bananas in exchange. So if there are people who are trying to sell apples but can't find buyers for their apples those same people must be trying to to buy bananas but can't find sellers. An excess supply of apples *means* an excess demand for bananas.

Now add a third good, money, and assume monetary exchange, so in the apple market people buy and sell apples for money, and in the banana market people buy and sell bananas for money. It is perfectly possible for people to be trying to sell apples and unable to find buyers and also trying to sell bananas and unable to find buyers. Because everyone is trying to "buy" more money.

Much more involved explanation here.

[Curmudgeon_Killjoy]

Mr. Rowe:

Say’s Law, properly understood, says production creates the producer’s capacity to demand (purchasing power) if, and to the extent, revenue (in goods or money) from production exceeds the costs of production. The efficient use of resources (input costs < output revenue) increases purchasing power and makes possible economic growth. The inefficient use of resources (input costs exceed revenues) reduces purchasing power and causes economic contraction. To the extent money makes the exchange or use of resources more efficient, money contributes to growth.

But money (as used in your example) does not subtract from aggregate demand for apples or bananas as “the structure of an economic model that . . . include[s] the interrelated balance sheets and income statements of the units of the economy” would show. The quote is from Minsky. The idea traces back at least to Fisher. And my little observation is that the inefficient use of use of resources, not the existence money, extinguishes demand. Money, in your example, just saves demand for a later day.

As a thought experiment, put labor costs in your apple-banana-money economy, give everyone a balance sheet, make the correct quadruple entries (double entry for buyer and seller or borrower or lender), and then see what happens to net worth on the actor's balance sheets from one cycle to the next when spending = income (revenue) but costs exceed revenue. If you want to introduce money in the model, include lenders, have the lenders make loans to produce or and consumers of goods, schedule debt payments to make sure you follow effect on purchasing power in subsequent accounting periods, postulate defaults along the way, and remember capital is never income. Before you add the complication that low interest rates (relative to what is a really good question) raise the price of stocks (capital assets) but do not increase the income flows from those stocks, the outlines of the financial crisis become vaguely comprehensible and the paradox of thrift reveals itself as the fallacy it is.

Spending may equal income, but a given volume of spending or revenue is not necessarily equal in balance sheet or economic effects. PK says he can save the economy by spending today to raise aggregate demand today. At the start of the day, PK has a net worth equal to $10 and 2 shovels. He pays one man $5 to dig a hole in the morning and another $5 to fill it up in the afternoon. At the end of the day, PK has no money to use to demand anything else, he has nothing to sell, he has what's left of 2 worn out shovels. But on the positive side, he does have a GDP report that says, since spending = income, measured GDP went up $10. Thus, to raise today’s measured GDP by $10, PK wasted the labor of 2 men for the day, lowered his own net worth by $10 and 2 shovels in exchange for nothing he could sell or use, transferred his $10 to the ditch diggers (the gift theory of growth), and lowered society’s aggregate net worth by 2 shovels. Potential purchasing power actually went down. PK transferred his $10 to the ditch diggers. The ditch diggers have $10 to spend instead of PK. But purchasing power and output potential both fell because PK didn't earn enough from the ditch digging exercise to employ the ditch diggers tomorrow for something productive or to replace the worn out shovels. Because the shovels got worn out, PK can't use them or sell them. And because PK lost all his money digging and filling a worthless hole he has no shovels to use or sell or money to spend, save, or invest in something else today, tomorrow, or the next day.

[Nick Rowe]

Curmudgeon: I would put it this way: in a barter economy, or if there were no excess demand for money, Say's Law would work just fine. If we produce $1 trillion worth of goods a year, we can afford to buy that same $1 trillion worth of goods a year. But if the quantity of money is too low, relative to the velocity of circulation, and relative to the money prices of goods, those goods won't get sold.

[Kevin Donoghue]

Don't ditch diggers spend at all? A man's got to eat.

[Nick Rowe]

Kevin: it makes more sense to just give the ditch diggers the money, and tell them to forget about the ditches. Worst case scenario: they dig their gardens instead, or lift weights, whatever, unless they really like digging ditches.

[Kevin Donoghue]

Nick, I agree. Atrios has been pushing that line for some time, and Benassy models fiscal policy as simple cash transfers. I don't know why that isn't the standard approach. It bypasses all the usual guff about broken-window fallacies and suchlike.

[Odie]

Kevin,
I think Curmudgeon raises an important point in his last paragraph although I would not put it in those dire terms. You just need to realize that all monetary assets equal all monetary liabilities to understand there cannot be any net monetary savings in a society. All monetary surpluses/savings equal all monetary deficits/debts. All balance sheets must add up to zero. Hence, what is then the net worth of our society? It is the non-monetary assets we have acquired over time using the money we created. Those assets can be material (e. g. houses, cars, tools) or immaterial (e. g. health, education) but they constitute the real wealth of our society.

Thus, when we talk about purchasing decisions we should not discuss how much debt that requires or think about how much money someone can save at the end (saving meaning having surplus later) but what is the non-monetary asset that we will acquire using that money? Will it really add to the net worth of our society? Now you can easily see why digging holes and filling them again is a pointless endeavor but digging a hole and planting a tree will actually increase our net worth. Money then becomes a tool for proper resource allocation. Given a fixed amount of money the less you can spend to create a new asset the more assets you can create overall. The less non-renewable resources you use per item the more assets you can create now and in the future. The more durable the asset is you create the more you increase the net worth over time.

This thinking seems to be completely absent in our policy leaders as you can see in the current discussions about the debt ceiling although it is an accounting triviality.

[Curmudgeon_Killjoy]

Nick, I respectfully disagree. The essential point is that if A produces $1 trillion in goods this cycle and B buys them for $1 trillion, but it costs A $1.1 trillion to make the goods, A's net worth at the end of the first cycle declines, and A's purchasing power, including A’s capacity to buy inputs for the next production cycle, declines. Demand destruction occurs on the producer side even if B still has $1 trillion to spend. Under this condition, spending = income at $1 trillion in the first cycle portends a lower equilibrium for spending = income (revenue) in the next cycle, even if B’s income and purchasing power remain the same. Raising B’s spending may raise A’s gross income in the current cycle, but if B’s spending "buys" A's revenue below the cost of production, aggregate demand for the next cycle declines to the extent of the difference.

Now, back to money. We agree, I think, that if there is an insufficient quantity of money to facilitate desired exchanges, the efficient use of resources suffers. But I think this is a different problem than the one you identify as the basis for the invalidation of Say’s Law in a money economy. The decision to hold money, rather than spend it, is a decision to spend now or to spend later, not a decision never to spend. A man with $2 dollars in his pocket can buy as many apples or bananas today or tomorrow as the apple or banana seller will sell for $2, whether the man holding the money spends $1 today and $1 tomorrow, $2 today, or $2 tomorrow. The decision to hold today and spend tomorrow divides the same potential contribution to aggregate demand over 2 days instead of one, but it does not reduce the holder’s potential contribution to aggregate demand or actual contribution over the relevant period. The decision to hold money today is not a decision to hold it forever. The money held remains on the holder’s balance sheet, and the holder can spend it as soon as she decides the utility of purchasing something today exceeds the value of having the $1 available to purchase anything today or tomorrow. Aggregate demand remains the same over the 2 day period. $2 + 0 = $2; $1 +$1 = $2; and $0 +2 = $2. I think the real difference between us is that my apple-banana economy has a calendar and, at least implicitly, yours does not, and I assume money held is eventually spent, and you do not.