Low vs falling prices, and relative vs absolute prices, Intro Econ again

The discussion of Intro Economics in my last post got a bit esoteric at times. I want to bring us all back down to earth.

This morning I got an email from a student who took my Intro Economics course last year. He works in retail, selling hi-tech goods. He asked me this question: hi-tech prices have been falling, and yet sales have been growing, so why does the Bank of Canada target 2% inflation? Why would deflation be such a bad thing?

He didn't ask that question because he thought it might be on the exam. He asked it because he wanted to know. He wanted to understand the world around him. And he wanted to know if the Bank of Canada's 2% inflation target made sense, given his own personal experience.

There's a demand out there for knowledge and understanding. Our job, as suppliers, is to try to meet that demand. And my ex-student's question is one we economists can answer. And our job, as teachers of Intro Economics, is to try to answer questions like that without setting up a two sector dynamic stochastic general equilibrium model of a monetary economy. Our job is to try to explain it simply, and clearly. And that is what makes it a very hard job to do well.

Forget all that hi-falutin ideological crap. We're no good at that stuff anyway. Or no better than anyone else in Poli Sci, Sociology, Philosophy, or wherever. This is what we are better than anyone else at doing. So we should be doing it. This is our comparative, and absolute, advantage. This is our core business. This is economics. This is what Intro Economics is really about.

But damn it's hard to do it well.

Here's my attempt to do my job and answer his question:

There are two distinctions you need to be clear on.

1. The distinction between low prices and falling prices. Between high prices and rising prices. Between levels, and rates of change. Fred is higher up the mountain than Mary, but Mary is climbing the mountain faster than Fred.

2. The distinction between relative prices and absolute prices. Between the price of cars relative to the price of houses, and the price of both cars and houses measured in money. Between what is happening to the price of one good relative to the average price of goods, and what is happening to the average price of goods. Both cars and houses have higher prices than 30 years ago, but car prices have increased less than average and house prices have increased more than average.

OK. Now let's start on the question.

Draw a regular supply and demand diagram for hi-tech goods. Show the equilibrium price and quantity.

1. Relative prices vs absolute prices.

Now suppose there's an improvement in technology, so the marginal costs of production fall, and so the supply curve shifts right. The result is a fall in the equilibrium price and a rise in the equilibrium quantity. And that, basically, is what has been happening to hi-tech goods like computers and cameras over the last several decades. Except we really have to measure quantity not just in the number of computers and cameras, but in the number of bits and pixels (or whatever) too. We are buying an increased quantity of computing and camera power. And it's the price of computing and camera power that has decreased.

Now, why does that demand curve slope down? Why does a lower price cause a higher quantity demanded? Because the price of hi-tech has fallen relative to the average price of goods, and fallen relative to our incomes. So we substitute away from other goods and buy more hi-tech goods. The demand curve slopes down if a lower relative price causes a higher quantity demanded. Which it almost always will.

But suppose the price of everything fell, and fell by the same percentage? There would be no reason to substitute away from some goods into other goods. And even if we did substitute away from some goods into other goods, this wouldn't increase the aggregate demand for goods. Plus, if the price of everything fell, our money incomes would fall too, because ultimately we get our incomes from producing and selling things. So if all prices fell by 10%, the immediate effect is that our money incomes would be 10% lower too, so we wouldn't be able to afford more stuff.

The effect of a lower relative price of hi-tech goods on the quantity demanded of hi-tech goods is very different from the effect of a lower average price of everything on the demand for everything.

2. Low prices vs falling prices.

Go back to your supply and demand diagram. Now suppose the supply curve stays fixed, but people suddenly hear news that prices will be 20% lower next year. (Before hearing the news, suppose people had thought that prices would be the same next year as this year.) What happens?

The news causes the demand curve to shift left. Some people decide to wait until next year to buy a computer or camera, because they will get more power for their buck. This year's equilibrium price and equilibrium quantity will be lower than they would have been if people had expected next year's price to be the same as this year's.

3. Putting 1 and 2 together.

This is what (I think) has been happening to the demand and supply of hi-tech goods over the last 30(?) years.

Every year technology improves, marginal costs fall, and the supply curve shifts right. So prices have been falling, and quantity demanded has been rising, year after year, as the equilibrium moves down along the same demand curve. And because prices have been falling at a roughly constant (say) 20% per year, people have always expected next year's prices to be 20% lower than this year.

Every year, for the last 30 years, the level of prices (relative to other goods) has been lower than the year before. And we have observed the effect of lower prices on increased quantity demanded. But, I would guess, the rate at which prices have been falling, and the rate at which people expect them to be falling, has stayed about the same over the last 30 years. We have seen the movement down along the demand curve, but we haven't seen a leftward shift in the demand curve, because there hasn't been one. If my guess is right, people have always expected falling prices, and haven't changed their beliefs about falling prices, so the demand curve hasn't shifted.

But suppose people suddenly changed their expectations about falling prices. Just imagine what would happen if people suddenly got the news that the engineers had simply run out of ideas, so that technology would stop improving, the supply curve would stop shifting right, and so next year's prices wouldn't be any lower than this year's. Some people, who had been postponing buying computers or cameras, would suddenly stop waiting for next year's better cheaper model, and want to buy this year's instead. The demand curve would shift right. The equilibrium quantity sold this year would increase even more than it normally would have, because the rightward shift in the demand curve adds to the rightward shift in the supply curve.

4. Back to macro.

Suppose the Bank of Canada suddenly announced that it would be targeting 0% inflation in future, instead of the 2% inflation it currently targets. And suppose people believe this, and expect 0% inflation from now on, instead of the 2% they had previously expected. What would happen to the demand for goods?

It would fall. With 2% expected inflation there used to be an incentive to buy goods now because their prices will be 2% higher on average next year and your savings will be worth 2% less. With 0% expected inflation that incentive disappears, and so the demand curves for many things will shift left as people save more and postpone spending. [This paragraph edited for clarity; thanks Gregor.]

But, if nominal interest rates fell by the same 2 percentage points, from 5% to 3% say, that drop in interest rates would exactly offset the drop in expected inflation, leave real interest rates the same, and so would mean exactly the same incentive to save or spend today. So, a lower rate of expected inflation would not cause demand to drop if nominal interest rates fell by the same amount.

But squared. Could nominal interest rates drop by the same 2 percentage points? Well some couldn't, because they are already so low that dropping by 2 percentage points would make them negative. And that can't happen, because people would just hold their cash under the mattress. Therefore, a reduction in the inflation target from 2% to 0% would cause a fall in demand.

End of answer.

That answer took me over an hour to think up and write down. I could have done it quicker and much clearer if I could talk instead of write, draw diagrams on the board, wave my arms around, and make full use of my voice to convey meaning.

But it was hard. It is nowhere near as clear and simple as I would like it to be. I could maybe have made it shorter, simpler, and clearer....by bullshitting. I could have said something that sounded good, but wasn't even approximately right.

Damn but teaching Intro Economics is hard. And learning Intro Economics is hard too. But this is what teaching and learning Intro Economics is all about, not just BSing some poncy stuff to keep the OWS gang happy. Sure, I could do that. So could anyone. Not everyone can do this.

Comments

[Min]

The average price of goods has a definitional problem. What weight do we give different goods? OC, there are things like milk, which has no natural unit. And even when there are units, how do we compare them? If there are two types of cars for sale, one costing $100,000 and one costing $10,000, is the average cost $55,000? Suppose that one of the $100,000 cars is sold, and nine of the $10,000 cars are sold. Is the average cost $19,000? That seems reasonable. But what if 500 chickens are also sold at $2 each? Is the average cost now $191,000/510? Adding cars and chickens is questionable.

[Nick Rowe]

Min: and that is another good question to which we economists have a (reasonably) good answer. We economists *can* (more or less) add apples and oranges (and we also know a lot about the limits to adding apples and oranges too). And we answer that question in Intro Economics too.

But I'm a bit too burned out to answer it now. Who wants to have a stab at it? Basics of CPI or GDP deflator?

*That's* what Intro Economics is about. We can answer that question. Them others can't.

[Lord]

I would emphasize why tech prices fall, progress. If other prices fell for the same reason it would be good, but we shouldn't confuse the result with the cause because falling prices of other goods would not be due to progress, we just don't make that kind of progress in other areas.

[Jeremy Fox]

Just wanted to say great post, Nick. Very much echoes my feelings about teaching introductory ideas in my own field (which you've now inspired me to post about at some point). I hope I'm as good at introducing students to population and community ecology as you are at introducing them to economics.

[W. Peden]

"Basics of CPI or GDP deflator?"

Surely it depends on why you're measuring prices. A macroeconomic strategy based on stabilising the economy through stabilising inflation should be interested in the GDP deflator, since it gives a wider perspective on price movements in the economy. Another example where the GDP deflator is good is for (shock!) deflating GDP in order to remove inflationary distortions such that we can derive real GDP stats from nominal GDP stats.

On the other hand, the CPI may have a place in assessing changes in the cost of living for benefits systems like unemployment benefit and state pensions, since the CPI can be used to measure changes in living costs for those who don't spend on housing and the like. In fact, in the case of pensioners, a very narrow index of prices might be useful e.g. one that heavily weights heating, clothing and food, and excludes electronics and the like which (poor) pensioners don't spend money on.

[W. Peden]

Lord,

I agree. I like to think of prices as like an appearance of water: it can be real water or a mirage. Falling prices caused by increased productivity are the appearance of something fantastic. Falling prices caused by falling real wages are the appearance of something awful.

[Gregor Bush]

Nick,
"But suppose the price of everything fell, and fell by the same percentage? There would be no reason to substitute away from some goods into other goods...So if all prices fell by 10%, the immediate effect is that our money incomes would be 10% lower too, so we wouldn't be able to afford more stuff.
The effect of a lower relative price of hi-tech goods on the quantity demanded of hi-tech goods is very different from the effect of a lower average price of everything on the demand for everything."

Wait, why did the price of everything fall? If it was because of a positive AS shock and the central bank was targeting NGDP, then our nominal income would stay constant but our real income would rise. We would buy more of everything at lower prices. The way you've described it, the decline in the price level is a result of a negative AD shock (because our nominal incomes are falling). But this need not be the case. As Sumner would say, never reason from a price (level) movement.

[K]

How about...

64 words:

Imagine that prices are declining so fast that NGDP also is declining. Then, all things being equal (labour/capital productivity, labour force participation, etc - most industry is *not* computers), wages will have to decline. But lets imagine that wages won't go down. Then the labour market won't clear. And then unemployment, animal spirits, uncontrollable debt-deflation (because of ZIRB), blah, blah, blah... Great Depression.

Or (49 words):

If inflation plus the natural rate drops below zero, then the nominal equilibrium rate would have to be negative to avoid a negative cumulative Wicksellian process. But it can't be negative. So inflation must be maintained a safe margin above the lowest reasonable outlook for minus the natural rate.

Not enough detail? Wrong?

[Nick Rowe]

Lord: good point. My answer missed that bit, and I should have included it. When the prices of high tech goods fall because of improved technology, our real incomes really do rise, but they rise as a result of that technological progress, not the fall in prices, which is just a symptom. We can produce more (better) goods with the same inputs as before. It's very different if all prices fall because of tighter monetary policy.

Jeremy: thanks! I'm not sure I'm that good. But I have had a helluva lot of practice, which helps.

W. Peden: Yep. It's always a good thing to ask what we want to use that index for, before we construct it. Yep, a lot of the talk about measurement bias in the CPI is very relevant if you are talking about adjusting payments for the cost of living, but is totally irrelevant if you are talking about the BoC targeting 2% inflation to avoid the zero lower bound.

[Nick Rowe]

Gregor: yep. I was talking about a permanent change in the growth rate of the AD curve, caused by a lower target inflation rate for the Bank of Canada. So all the AS stuff was being held constant. And I would treat this as a Long Run experiment, so we are on the LRAS curve in both cases, and so wouldn't use the words "AD *shock*" to describe it (unless I wanted to distinguish the short run from the long run effects of moving to a lower inflation target).

K: Not wrong, but your answer is all macro, whereas mine is mostly micro, about the hi-tech sector. And you've got the short run and long run effects all in there together. Plus, a first year student won't understand half of what you are saying.

[Ryan]

I have a couple of bones to pick...

First Bone:
This passage: "Just imagine what would happen if people suddenly got the news that the engineers had simply run out of ideas, so that technology would stop improving, the supply curve would stop shifting right, and so next year's prices wouldn't be any lower than this year's. Some people, who had been postponing buying computers or cameras, would suddenly stop waiting for next year's better cheaper model, and want to buy this year's instead. The demand curve would shift right. The equilibrium quantity sold this year would increase even more than it normally would have, because the rightward shift in the demand curve adds to the rightward shift in the supply curve."

There is a big difference between falling prices and improving quality. Just we expect quality to improve year over year doesn't mean we expect prices to increase. What you have expressed as one concept (higher expected prices) is actually two concepts (improving quality which only means rising prices if the company wants next year to be a carbon-copy of this year).

There are so many assumptions embedded in this scenario that it really begins to break down upon closer scrutiny. In reality, assuming they want to compete, technology companies can: (1) Change product quality, (2) change product prices, (3) change production costs, or (4) any blend of the previous 3.

The point is that (1) implies that prices could stay the same if companies choose to compete on quality. We have no reason to automatically assume a price increase. It all depends.

Second bone:
This statement: "Suppose the Bank of Canada suddenly announced that it would be targeting 0% inflation in future, instead of the 2% inflation it currently targets. And suppose people believe this, and expect 0% inflation from now on, instead of the 2% they had previously expected. What would happen to the demand for goods?

"It would fall. The demand curve for nearly everything would shift left, as people postponed buying goods. There's no longer the incentive to buy goods now because their prices will be 2% higher on average next year and your savings will be worth 2% less."

I assert the following instead: Demand for currency would decrease because 2% less currency would be required to purchase the same quantity of goods and services. This would probably result in lower prices and thus lower NGDP, but it will not necessarily result in a shift of the AD curve and therefore may not necessarily change RGDP.

My point is this: Isn't the real issue not the price change, but rather what's driving the change in price? If prices are constantly forced to adjust to the central bank's inflation targeting rule, those price increases must contain two components: Part of the price change is driven by the conditions of the market in question; the other part is driven entirely by the central bank's inflation targeting rule.

It is this second part of the price change that Austrian-leaning economists cite as a major market distortion enacted by central banking policy.

[Gregor Bush]

Thanks Nick. I'm confused about this too.

"There's no longer the incentive to buy goods now because their prices will be 2% higher on average next year and your savings will be worth 2% less."

Did you mean "prices will be 2% LOWER than previously expected and your savings will be worth 2% MORE than was previously expected?"

[Nick Rowe]

Gregor: damn! My writing wasn't clear. I'm going to edit the post. Just goes to prove my point!

[anon]

Since you've mentioned the distinction between relative and absolute prices, I would just state that the absolute price level measures the value of money in terms of goods, and that the BoC cares about keeping that value on a stable path--maybe I'd briefly mention price frictions. Then your supply-and-demand analysis follows from general principles, and can mostly be left as an exercise.

[Nick Rowe]

anon: "Then your supply-and-demand analysis follows from general principles, and can mostly be left as an exercise."

But elucidating and applying those general principles is precisely the bread and butter of Intro Economics! If you leave it as an "exercise for the reader", and the reader is an Intro Economics student, it won't be done. That's precisely what they need help with! I bet you are some very bright young economist, who has totally forgotten, or never learned, just how hard this is for most first year students. Young guys like you are totally vicious, without even realising it, if we let you loose on poor unsuspecting little students! That's why old guys like me exist.

[anon]

"But elucidating and applying those general principles is precisely the bread and butter of Intro Economics!"

OK, you're right. I was being somewhat facetious here--so we should be doing that execise after all. However, I did find your answer somewhat awkward, in that it seemed to deal with a special case without elucidating general rules: namely, that money can be treated as an economic good like any other - albeit one with a very special status in a real-world, monetary economy; and that price is best thought of as a relative concept, regardless of the unit of account. (Even distortions arising from nominal prices can be thought of as relating to a shift in the value of money.)

IMO, expressly stating and applying these general rules would have made your answer more straightforward and easier to memorize. Sometimes there are tradeoffs between generality/abstraction and simplicity or directness, which instructors should be mindful of. But I don't think this was the case here.

[K]

I'm with anon. The student asked why deflation is bad. Now it's fine to mention, as an aside, that massive productivity gains can cause falling prices for specific products. And that kind of deflation is harmless. But *general* inflation is a monetary phenomenon. It obviously depends on some kind of arbitrarily determined nominal anchor of the unit of account. Then you need to  provide *some kind of justification* of non-monetary super neutrality (if your story is about liquidity demand) or non-monetary neutrality (if your story is about price/wage stickiness). The micro stuff seems beside the point. Once you establish the basics of non-clearing markets at the ZLB, then you are pretty well done.

Best to tell a *really* short (even if slightly cryptic) story with 2 or 3 important concepts to focus on. Eg: "You must understand 1) inflation is a change in the nominal *monetary* anchor, *not* a specific product price change 2) nominal rigidities 3) the evil ZIRB." The shorter the initial explanation the more the student will be focused on those critical bits when you elaborate the details later.

[Determinant]

You could make a terms-of-trade difference story. Better technology means less input for a given quantity of output. That means prices per unit of output can fall, profit margins can be maintained, and the supply inputs and money that would have been spent on these inputs can be spent somewhere else.

Better technology improves the seller's terms of trade, letting her maintain margins while cutting prices and it vacates supply space to be used elsewhere. Good thing. The dog wags the tail.

If general prices fall while the requires supply inputs to produce a given level out computer output stays the same, the seller's terms of trade are deteriorating. Profit margins are squeezed and at best can stay steady. Any debt contract that has a fixed price and a future settlement date means one party runs the strong risk of not having the funds to pay their obligations because of deteriorating terms of trade.

Decreasing general prices don't lead to supply space being vacated for additional production. The result is stagnation. The link between supply improvement and money is broken. This is the tail wagging the dog and it doesn't work.

[JL]

Great post

"With 2% expected inflation there used to be an incentive to buy goods now because their prices will be 2% higher on average next year and your savings will be worth 2% less."

Empircal question: at what income level do people start taking this into consideration? Or is it more structural?

"Not just BSing some poncy stuff to keep the OWS gang happy"

I am as left as they come when it comes to political values but on a scientific level I have been disapointed by some professors who are just using OWS to wave around their usual mantras. Still its not always the case. I realy hope economics and the other social sciences set aside their differences and start working on confronting their ontologies in a sythetic spirit... one day maybe.

[Ian Lippert]

Nick I am just curious why people would reduce present consumption in the face of future prices being 2% lower? If we look at things like credit card usage we can see that people don't defer gratification often, even when faced with costs upwards of 20%. If 2% decrease in prices was due to an economy wide increase in aggregate productivity it seems like our gains in our real income would overcome the possibility that we will be
Missing out on lower prices in the future.

For example I bought a big flat screen tv last year that is now $2000 cheaper this year. If I could go back in time I would still make the purchase because that $2000 is not a significant portion of my income. I think that deflation caused by productivity increases can be more than compensated by the diminishing marginal value of our increasing real income.

[Nick Rowe]

Ryan: your comment got stuck in the spam filter, and I only just fished it out. Sorry.

"There is a big difference between falling prices and improving quality. Just we expect quality to improve year over year doesn't mean we expect prices to increase. What you have expressed as one concept (higher expected prices) is actually two concepts (improving quality which only means rising prices if the company wants next year to be a carbon-copy of this year)."

There are three dimensions: price, quantity, and quality. (Actually, there's a lot more than 3 dimensions, since quality can have more than one dimension). And I have forced this 3 dimensional world into a 2 dimensional supply and demand framework, by multiplying quantity of cameras by pixels per camera so we get total quantity of pixels on the horizontal axis, and price per pixel on the vertical axis.

Sure, what I did is a simplification, which leaves stuff out, and won't be exactly right under all circumstances. I would have to do some serious analysis to figure out exactly what those circumstances are. But economics is a collective enterprise. I'm very confident some microeconomist somewhere has done that analysis, and if the results were totally different, someone would probably have told me by yelling it from the rooftops that his brilliant new model showed that supply and demand was all wrong once you introduced varying quality. (Indeed, they did shout it from the rooftops in the "Market for Lemons" version, but I don't think that version applies to new hi-tech goods.)

That's what's neat about economics, as a discipline. We're all in this together, and we are all out there testing and checking different bits, and letting each other know when something's wrong. Any sentry can raise the hue and cry, like Akerlof did with Market for Lemons.

"I assert the following instead: Demand for currency would decrease because 2% less currency would be required to purchase the same quantity of goods and services. This would probably result in lower prices and thus lower NGDP, but it will not necessarily result in a shift of the AD curve and therefore may not necessarily change RGDP."

A 2% lower *price level* means a 2% lower demand for *nominal* currency. But a 2% lower *rate of inflation* (matched with 2% lower nominal interest rates) means an *increased* demand for *real* currency (M/P), because the opportunity cost of holding it is less.

Md=P.L(Y,i) where dL/dY positive and dL/di negative (where i is either inflation rate or nominal interest rate).

"It is this second part of the price change that Austrian-leaning economists cite as a major market distortion enacted by central banking policy."

And that is the bit of Austrian economics I have never understood.

[Nick Rowe]

Determinant: "You could make a terms-of-trade difference story. Better technology means less input for a given quantity of output. That means prices per unit of output can fall, profit margins can be maintained, and the supply inputs and money that would have been spent on these inputs can be spent somewhere else."

I would be tempted to show that with a PPF and Indifference curve analysis. Hi-tech goods on one axis, all other goods on the other axis, then swivel the PPF outwards, to a tangency with a higher indifference curve, and the slope at the tangency being the relative price ratio. That's a good way of showing it. It would work well for second year students, but not for first year, where the basic micro model is supply and demand. Not easy to put money into that model though.

JL: Thanks!
"Empircal question: at what income level do people start taking this into consideration? Or is it more structural?"

I'm not aware of any theoretical or empirical research that says the effect of real interest rates on demand only kicks in above a certain income level. Maybe it couldn't work for someone on or close to the margin of bare survival. Dunno.

Ian: "If 2% decrease in prices was due to an economy wide increase in aggregate productivity it seems like our gains in our real income would overcome the possibility that we will be Missing out on lower prices in the future."

Agreed. There are two effects here. Suppose you suddenly get news that there will be a big gain in productivity next year. People anticipate this increase in real income, and want to consume more, and save less, now. At the macroeconomic level (closed economy) the equilibrium real interest rate would have to rise to offset this reduction in desired saving. And there are two ways this increased real interest rate could come about: higher nominal rates; lower inflation. Which one of those two we get depends on what the central bank is targeting: inflation; or NGDP.

[JL]

Nick: "Maybe it couldn't work for someone on or close to the margin of bare survival"

Well I guess I should point out that I am a student, but I just don't know anyone who makes enough money to be like "I'd better spend this now before I lose it all to inflation" A salaried employee doesn't necessarily live pay check to pay check if he has a good job, but people buy things because they need them or want them or they invest their money. I just don’t know anyone who treats money like too many air miles about to expire.

I can see how it might apply if you are an investor, a medium or large business or if you have invested your money and someone else is "spending/investing" for you though, in that sense maybe it's structural or indirect? IE: More big players who take this into consideration invest; economy goes better, more demand?

So I guess what I meant to ask is:
Is the relation between inflation and spending/saving an assumption of optimum behaviour or is this empirically verified? If it is empirically verified is it a correlation, a direct causation or an indirect causation?

[Steve Roth]

From one in the center of your target audience: Bravo. It brings me close to the point that I feel that I could explain what you just explained (which I've understood generally, but not concisely or cogently).

But here's what I don't understand:

"if nominal interest rates fell by the same 2 percentage points, from 5% to 3% say, that drop in interest rates would exactly offset the drop in expected inflation, leave real interest rates the same, and so would mean exactly the same incentive to save or spend today."

There's an unstated assumption, implication, here that interest rates follow inflation in lock-step. It's not clear to me that they do, or even that they should theoretically. Ceteris paribus, I get that, but still. Might be that you just strayed outside your explanation with this, that it's not necessary.

But: Even if rates do follow inflation (theoretically or empirically) It's not at all intuitively obvious to me that or why the aggregate incentive to spend is unchanged. Because:

The incentive effects of A) price changes and B) interest-rate changes would not be uniform. Because some people/businesses have a lot of debt, some a lot of wealth (credit). (Ditto, or IOW, "savings" [as a stock].) And there are far more debtors than there are creditors. Can we just assume that disparate incentives net out/aggregate up to zero?

Price changes affect the distribution of wealth (where wealth equals the buying power to purchase real goods that can be consumed and valued by humans -- including productive assets that are also consumed, just more slowly). It's the overwhelming effect of price changes.

And people with different levels of "savings"/wealth have different tastes/preferences/indifference maps/demand curves/utility functions/marginal propensity to spend. *Because* they have different stocks of savings.

So price changes cause the *relative* demand for different products to change. So, ditto the relative prices of different products.

Price changes affect wealth distribution (quite profoundly). And wealth redistribution affects *relative* prices/demand curves.

So even accepting the price-change/interest-rate-change lock-step idea, it's not clear to me that price changes (and presumably accompanying interest-rate changes) result in "exactly the same incentive to save or spend today" -- that it aggregates up, homogenizes, in that way. It just seems darned unlikely. But I haven't done the math...

P.S. I don't seem to be able to post from Chrome/Mac. Error: "We're sorry, we cannot accept this data." Works from Firefox.

[Mandos]

But this is what teaching and learning Intro Economics is all about, not just BSing some poncy stuff to keep the OWS gang happy.

...

That's what's neat about economics, as a discipline. We're all in this together, and we are all out there testing and checking different bits, and letting each other know when something's wrong. Any sentry can raise the hue and cry, like Akerlof did with Market for Lemons.

*wipes bloody mouth after having bitten tongue too hard*

[Jon]

Hmm.

1) The essential lesson is that the price-level is a 'fiction'. That's the main lesson that undergrad econ students need to learn. You need to puncture money illusion.

2) You can make a good case for a stable inflation rate

3) Your story is weakest when defending the particular 2% target. That part of the story is weak but that isn't your fault.

As an aside, your story made me wonder about why were stuck with this nasty 'deflation' risk. When some people started proposing negative IOR, the response was a concern about currency hoarding. Then some people proposed a lottery where certain serial numbers would be invalidated. Rightly, I think people saw that sort of behavior as unethical as it imbues an element of risk to currency.

But we don't have money anymore. We have bank notes. It seems curious that these have been imbued with so many of the legal characteristics of money that people forgot they were something rather different: a note. This was easy to forget, after all they can only be redeemed for themselves (which you can literally do).

With that in mind, it makes you wonder why it would be so challenging to legally designate what now circulates as currency as an ETN with an indexing formula designated by the CB. Every note is marked with a year and is redeemable for further notes less the IOR on excess reserves (positive or negative as it may be).

Then this would be a quite powerful tool--it could be quite readily introduced during normal times where the nominal rate was high and thus currency would 'pay' an interest rate to control high-side inflation by driving cash demand up. Then when the next deflation comes, the index would go negative and the interest-rate on cash would go negative and thus cash-balance demand could be driven down.

I don't want to stress the index part too much--in the sense of being mechanical. It would well be that the CB publishes the next index number as its approach to implementing its inflation (or NGDP target).

I think it might be perceived that the big trouble with this scheme is that face-value of the note is not indicative of its exchange value, but that could readily be dealt with by using only bills issued in the current year for general exchange. Older bills would be accepted by the banks only. Since people regularly go to ATM machines, this would hardly be a problem.

[Nick Rowe]

JL: "Is the relation between inflation and spending/saving an assumption of optimum behaviour or is this empirically verified? If it is empirically verified is it a correlation, a direct causation or an indirect causation?"

It's both. Theoretically, it's no different from saying that demand curves slope down. If the price of apples rises relative to bananas, people demand fewer apples and more bananas. If the price of present apples rises relative to future apples (if the real interest rate rises), people demand fewer present apples and more future apples.

Empirically, the beginning of the 1982 recession was enough for most of us. Tighter monetary policy caused real interest rates to rise, and demand fell. And this is exactly the empirical relationship that the Bank of Canada has used to keep inflation on target. If you Google scholar something like "estimate 'interest elasticity' consumption (or investment)" you get a whole slew of articles by econometricians arguing about who has the best estimate of the magnitude of this effect, and whether it's bigger or smaller than everyone else thinks.

[Nick Rowe]

Steve Roth: (Sometimes it won't accept my comments either, and I have to copy and save my comment, log out, log in again, and then paste and post.)

1. Do nominal interest rates adjust one-for-one with expected inflation?
2. If they did, would saving and investment be unchanged?

These are really the same question, if we are talking about what happens in equilibrium.

The best answer to this question is: "Yes, but...".

We start out with a very simple model in which they must adjust one for one, because rational savers and investors care only about real, not nominal interest rates, so the equilibrium real interest rate, at which desired saving equals desired investment, is independent of the expected rate of inflation. Therefore "yes".

Then we say "but", and think of all the possible reasons why that simple model might not work exactly. The nominal interest rate on currency can't adjust. What does the government do with the extra revenue it gets from the inflation tax, does it spend it or save it? The tax system on interest income is not indexed for inflation. Some people might get confused by inflation (money illusion). An unexpected change in inflation will redistribute wealth between creditors and debtors who might have different marginal propensitities to save, so it might not cancel out in aggregate (as you said). Etc.

"Yes, but" (or "No, but") is often the best answer to a lot of questions in economics. Give the simple answer first, then consider cases where it might not work exactly.

[Nick Rowe]

Jon: "1) The essential lesson is that the price-level is a 'fiction'. That's the main lesson that undergrad econ students need to learn. You need to puncture money illusion."

I agree. First we have to explain the "yes" in the "yes, but" answer (or the "no", in the "no, but" answer). Only then can we go on and talk about the exceptions. We start out by saying money is neutral, and superneutral, then we talk about reasons why it might not be, under some circumastances. One of the hardest thing in talking to non economists is that they don't look at it this way, and you get into endless conceptual confusions, because they want to talk about everything at once, without even having a model.

Not sure I fully follow what you are saying about notes, but if notes dropped discretely in value at the end of the year (rather than smoothly), then that would seem to create problems. And the whole point of monetary exchange is the lower transactions costs caused by the fungibility of money, and the ease of determining the value of one note relative to another, so we don't worry about Gresham's Law (which is a Market for Lemons problem).

[Nick Rowe]

Nobody has tackled Min's question!

"The average price of goods has a definitional problem. What weight do we give different goods?"

There are two answers: Laspeyres' and Pasche (and I can never remember which is which, or how to spell them). The CPI uses one, and the GDP deflator uses the other.

CPI: Look at the basket of goods that people consumed last year. Ask how much that same basket of goods would cost this year. Calculate the percentage increase in price. That's the CPI inflation rate.

If Q0 and P0 are the vectors of quantities and prices for year 0, and Q1 and P1 are for year 1, then the CPI is defined as:

CPI = Q0.P1/Q0.P0 x 100%

The CPI uses last year's quantities consumed to get a weighted average of price increases. ("Last year" really means a couple of years back, depending on how frequently they update the basket.)

The GDP deflator uses this year's quantities as weights (and it's quantities produced, not consumed):

GDP Deflator = Q1.P1/Q1.P0 x 100%

Because people substitute between goods when relative prices change, using last year's quantities as weights creates an upward bias, and using this year's quantities as weights creates a downward bias.

A very small number of economists spend their lives working on this problem, and working out different weighting systems. The rest of us just remember what I have written above. IIRC (I probably don't) as you update the basket more and more frequently, then in the limit the two measures converge to eliminate bias (unless there are discrete price jumps).

The biggest practical and theoretical problems are how you handle new goods (and a change in quality of an existing good is like a new good). You can even argue that there has been no inflation over the last 100 years, because new goods have made people better off today than someone with the same money income 100 years ago. Internet access had an infinitely high price 100 years ago.

[JL]

Ok that clarifies things thanks!

[Jon]

Nick, yes discretely like a dividend is discrete or moving the gold peg one day is discrete, but known in the sense that the date of the change is known but the acculumating apr is varying with each CB meeting.

I don't see this being outside of the scope of human experience, for instance during the wildcatting era notes from many different banks circulated. Nor is it that different from now where the real exchange value fluctuates with the exchange rate and the inflation rate.

Th difference is that currently these variations are only indirectly controlled by the CB. I propose a mechanism where there would be direct control.

[JeffreyY]

I haven't read all of the comments, so sorry if you've already discussed this, but I agree with Jon's (3) that your answer is weak on the "why 2%" point, even while it answers the "why not change from 2% to 0%".

I know of two plausible answers for "why 2%":
1) Money illusion: Some workers necessarily need to take real wage cuts from time to time, but everyone resists nominal wage cuts, so the economy runs smoother if the real wage cuts can be hidden behind inflation.
2) Too many savers: Lots of people (baby boomers) currently want to save money for the future, and the market has established that there aren't enough productive uses for the savings to give them a positive real rate of return. With a 0% inflation goal today, they appear to be keeping full value by keeping their money in a mattress, but in the future if all of them try to spend their money at once, the economy will either need to deal with a burst of inflation, or the central bank will need to withdraw that extra currency then, leading to a transfer of wealth from future-people to today's-savers. Alternately, with a small-positive inflation rate (e.g. 2%), the savers obviously lose some real wealth continuously, but in the future they get to spend at par. (<--- this feels like a the core of a good argument, but is pretty shaky as it stands.)

[Too Much Fed]

"Now suppose there's an improvement in technology, so the marginal costs of production fall, and so the supply curve shifts right."

Let's call that productivity growth.

"Plus, if the price of everything fell, our money incomes would fall too, because ultimately we get our incomes from producing and selling things. So if all prices fell by 10%, the immediate effect is that our money incomes would be 10% lower too, so we wouldn't be able to afford more stuff."

So where did the productivity growth go?

"Now, why does that demand curve slope down? Why does a lower price cause a higher quantity demanded? Because the price of hi-tech has fallen relative to the average price of goods, and fallen relative to our incomes. So we substitute away from other goods and buy more hi-tech goods. The demand curve slopes down if a lower relative price causes a higher quantity demanded. Which it almost always will."

I can think of many goods that if they fell in price, I don't believe "we" would buy more of, meaning they were not supply constrained to begin with. I believe it is important to put more emphasis on wage incomes vs. prices.

What does the scenario look like if hi-tech goods prices fall 10%, wage growth in hi-tech industries falls 10%, other goods prices change by 0%, and wage growth in other industries changes by 0%?

"But suppose the price of everything fell, and fell by the same percentage? There would be no reason to substitute away from some goods into other goods. And even if we did substitute away from some goods into other goods, this wouldn't increase the aggregate demand for goods."

Why isn't wage growth of 0% and price deflation of –2% possible (not necessarily preferred)?

Most of this comes down to should productivity growth and other things be evenly distributed between the major economic entities and be evenly distributed in time. That brings up the medium of exchange question.

For more on those see:

When will the workers wake up?

http://bilbo.economicoutlook.net/blog/?p=13193

The origins of the economic crisis

http://bilbo.economicoutlook.net/blog/?p=277

mark thoma actually gets one right.

http://economistsview.typepad.com/economistsview/2011/12/structural-adjustment-for-the-middle-class.html

"I'd prefer for the economy to solve this problem on its own, but this may not be a problem that self-corrects."

Actually, that means a medium of exchange problem.

"In addition, I don't think that the idea that increasing income inequality in recent years is attributable to merit, i.e. people getting rewarded purely for their contributions, is defensible. Wages for the working class have not increased with their productivity, but that income went somewhere and it's not hard to figure out where. One argument is that it would be unfair to redistribute income. However, to the extent that redistribution simply claws back the income that should have gone to the working class, it is not unfair at all. In fact, it would be unfair not to do this. Another argument is this would harm economic growth. There's little evidence that redistributing income affects economic growth in any case, but if the existing rewards are mal-distributed, as I think they are, then the distribution is not growth maximizing to begin with. Redistributing rewards so that the distributuion of income better reflects the contribution to production should enhance growth, not harm it."