David R. Henderson  

Ron Paul's Misunderstanding of the CPI

PRINT
Macro Blog Bleg: What Should I... Hobbesian Misanthropy in Th...
One of the least discussed, but potentially most significant, provisions in President Obama's budget is the use of the "chained consumer price index" (chained CPI), to measure the effect of inflation on people's standard of living. Chained CPI is an effort to alter the perceived impact of inflation via the gimmick of "full substitution." This is the assumption that when the price of one consumer product increases, consumers will simply substitute a similar, lower-cost product with no adverse effect. Thus, the government decides your standard of living is not affected if you can no longer afford to eat steak, as long as you can afford to eat hamburger.

The problem with "full substitution" should be obvious to anyone not on the government payroll. Since consumers did not choose to buy lower-priced beef before inflation raised the price of steak, they obviously preferred steak. So if the Federal Reserve's policies create inflation that forces you to purchase hamburger instead of steak, your standard of living is lowered. CPI already uses this sort of substitution to mask the costs of inflation, but chained CPI uses those substitutions more frequently, thereby lowering the reported rate of inflation.


This is from Ron Paul, "Chained CPI Chains Taxpayers," November 11. It shows a fundamental misunderstanding of the CPI and the idea of substitution.

Start with the second sentence: "Chained CPI is an effort to alter the perceived impact of inflation via the gimmick of 'full substitution.'"

By using the words "gimmick" and "perceived," Paul makes it sound as if the purpose of the chained CPI is to trick people into thinking that their cost of living has not gone up as much as the old CPI said. That's not true. The purpose of the chained CPI is to get a more accurate read on how much the cost of living has gone up.

The idea of substitution is basic economics. When the price of one good rises a lot and the price of another good rises a little or not at all, and those goods are, in many people's minds, substitutes, people will tend to substitute out of the good whose price has risen more and into the good whose price has risen less. So the purpose of the chained CPI is to take account of that in order to estimate the cost of achieving a given level of utility.

I'll illustrate with a simple example. Imagine someone buys only two goods: chicken and steak. Each month he buys 10 pounds of chicken and 10 pounds of steak. The price of steak is $7.00 a pound and the price of chicken is $3.00 a pound. So he spends $100 a month.

Now the price of chicken rises to $4.00 a pound and the price of steak stays at $7.00 a pound. If the Bureau of Labor Statistics were to price the old bundle (10 pounds of each), it would say that the cost of that bundle is now $110. So the apparent increase in the cost of living is 10%. But some people would substitute a little out of chicken and a little more into steak. The fact that they would do that says that they don't need that $110 to get the same level of utility. They might need only $108 or $109. So the method of pricing a fixed bundle overstates the increase in the cost of living.

Why do I have the lower-priced item rise in price and the higher-price item not rise in price? To get at the fact of relative price changes and to offset a trick--a "gimmick?"--that Ron Paul pulls. The naive reader reading his quote above will probably think, "Those idiots (or worse) at the Bureau of Labor Statistics. Don't they know that hamburger, for the vast majority of people, is inferior to steak?" Of course, it's inferior for most people. And the BLS has never tried to kid us that it isn't. Substitution is about changing the proportions in the bundle you buy in response to relative price changes.

I've written about this before, when Brett Arends showed the same misunderstanding in a Wall Street Journal op/ed.

The BLS lays it out here.


Comments and Sharing





COMMENTS (43 to date)
Thomas DeMeo writes:

The problem with your analysis is that the CPI isn't a grocery store shopping list. Its an index of prices.

The underlying premise that this system provides an accurate measure of inflation is shaky enough to begin with. The economy is a massive exercise in substitutions, and its a huge conceit to try to adjust for it. What we can do is create a list of reproducible purchases, analyze the data and hope it tells us something useful.

There is some value in measuring exactly the same bundle in exactly the same way over time. There is no way that the bundle actually represents an accurate snapshot of the needs of the typical citizen. The more "adjustments" that are made, the less useful this all becomes.

ed writes:

[Comment removed for supplying false email address. Email the webmaster@econlib.org to request restoring your comment. We'd be happy to publish your comment. A valid email address is nevertheless required to post comments on EconLog and EconTalk.--Econlib Ed.]

Emily writes:

As BLS notes in the "common misperceptions" article, steak and hamburger aren't actually treated as substitutes at all: "Substitution is only assumed to occur within basic CPI index categories, such as among types of ground beef in Chicago. Hamburger and steak are in different CPI item categories, so no substitution between them is built into the CPI-U or CPI-W."

Jake writes:

Good post, Professor Henderson.

As much as I agree with Dr. Paul on issues like foreign policy, fiscal policy, and civil liberties, I find his dogmatic advocacy of hard money and a return to the gold standard to be increasingly troubling.

I don't think Paul's stance on monetary policy is due to dishonesty, and I definitely believe the US money and banking system could use a great deal of reform. But if Milton Friedman and Scott Sumner are correct about the negative effects of tight money -- and I think they probably are -- then Paul's monetary theory could be devastating for the world economy.

kebko writes:

I agree with Thomas. Anybody who says inflation has been low hasn't tried to shop for a decent buggy whip lately. (Sorry Thomas) ;-)

David R. Henderson writes:

@Jake,
Good post, Professor Henderson.
Thanks, Jake.
As much as I agree with Dr. Paul on issues like foreign policy, fiscal policy, and civil liberties, I find his dogmatic advocacy of hard money and a return to the gold standard to be increasingly troubling.
I agree with him on these things too. But actually, Ron Paul has been somewhat flexible in the last year or two on monetary policy. I don’t have the link handy but I saw him on TV last year making the case for allowing competing moneys, which is a reasonable case to make. Just as you make good distinctions between his foreign and fiscal policy views and his views on money, it’s important, within the category of money, to distinguish between hard money/gold and competing moneys.

Quinn writes:

What if they can't spend the extra money to reach the same utility? Or is that question simply pointless due to the nature of the CPI index?

Thomas DeMeo writes:

kebko - Think of this like the SP500. Once in a while over the years, the items in the index should change, and that makes sense. But you wouldn't want the mix adjusted every day, just because the stock price of IBM today is a better deal than Hewlett Packard. It wouldn't mean anything any more.

Hazel Meade writes:

I am in favor of using the chained CPI, BECAUSE I believe it actually does underestimate inflation in the way Paul suggests. I don't think these is any other way out of our entitlement mess except to cut benefits, and tying them to the chained CPI is an effective way to cut benefits while pretending you aren't.

The alternative is a far worse fiscal crisis and a mass currency devaluation, which I don't think anyone wants to see.

So please, even if you think the chained CPI is a big fat lie (and it is), just go along with it, because it's for the greater good.

Dave writes:

The problem with the example you used is that the consumer decisions actually cost more.

In your example, the chicken went up to $4, and the consumer perceived that at that price, they may as well buy a bit of steak.

So they buy 9 lbs of chicken at $4, and 11 lbs of steak at $7. The total in your example is now $113, not $110.

I won't argue that "Dr." Paul has it correct, since he obviously does not. But the example that you gave tends to support the argument that the chained CPI is not a useful measure of how consumers behave.

Most consumers live on incomes that are fixed from month to month. If the price of chicken goes up, consumers will be forced to buy more of it, not less. They only have $100 to spend, and they need 20 lbs of food, so the answer will eventually be 13 pounds of chicken and 7 pounds of steak, for a total of $101.

According to your simplified version of chained CPI, the individual only experienced 1% inflation. In actual practice, they experienced more than 14% inflation: They are now spending $101 for food that they could have bought previously for only $88.

By substituting the lower priced good, they experienced more inflation than they would have otherwise.

Tom E. Snyder writes:

Hazel, "telling a lie for the common good" is an even bigger lie.

William B. writes:

So... if people have to substitute 85% lean ground beef for 90% lean ground beef, we won't see that in the chained CPI. However, it really is a reduction in the quality of ground beef those people are eating.

Your argument isn't realistic. While it is possible that people will rate the utility of steak at $7.00/lb more than chicken at $4.00/lb, far more people are simply trying to minimize their grocery costs.

The chained CPI is a soviet-style tactic for lying to the consumer. We really are living in Orwellian times, where, like in 1984, the chocolate ration gets decreased and people like Mr. Henderson come out and try to convince us that it actually INCREASED.

Rusty writes:

Whether this makes sense depends on when the substitution is made. If made in the current period, I don't think it works. To follow your example with a few tweaks, let's assume that the price of steak increases to $1,000 per pound, resulting in almost total substitution. Consumers are now buying 20 pounds of chicken (for $60) and no steak. A CPI based on this substitution would show a 40% decrease in prices. Certainly not an accurate representation of inflation.

Michael writes:

William B wrote:

"So... if people have to substitute 85% lean ground beef for 90% lean ground beef, we won't see that in the chained CPI. However, it really is a reduction in the quality of ground beef those people are eating."

There are many who would argue that substituing from 90% lean to 85% lean ground beef would actually be an improvement in the quality of the ground beef. :)


Rick Hull writes:

It looks to me like Rusty has it right, at least in terms of disrupting Henderson's example. I'm curious to hear the resolution.

David R. Henderson writes:

@Rusty,
To follow your example with a few tweaks, let's assume that the price of steak increases to $1,000 per pound, resulting in almost total substitution. Consumers are now buying 20 pounds of chicken (for $60) and no steak. A CPI based on this substitution would show a 40% decrease in prices. Certainly not an accurate representation of inflation.
False. 20 pounds of chicken would give substantially less utility than before. How do we know? Because with the original $100, the consumer had the ability to buy 33 pounds of chicken, and chose not to. 20 is less than 33.

Russ writes:

33 pounds of chicken? I think you have missed the point. The consumer wanted steak or at least some steak and 33 pounds of chicken will never sate anyones desire to eat steak. Unfortunately the authoritarian in you never sees the decrease in utility of lost choices.

kharris writes:

[Comment removed for rudeness.--Econlib Ed.]

dpadd writes:

I think professor Henderson has done what many people do and taken Ron Paul's words to literally. I'm just speculating but I believe Paul's point is that the government uses chained CPI to mask the artifice increase in the price of steak because inflation. I don't think Paul is presenting his argument based in a perfectly competitive market. I guess a good test would be if the price of steak increases and people substituted for beef than at some point the price of steak should come down.

Justin writes:

I've heard a lot of libertarians make this argument. Curiously, they only apply the argument going forward in time, never backward.

For example, today we can buy a very powerful computer for $500. What would it take to buy that powerful of a computer in say, 1980? Let's generously assume $10,000 .

I don't think most people who apply Ron Paul's argument would ever say we had 95% deflation during this period because the substitution argument is perfectly clear in reverse. It doesn't matter that the computer cost $10,000 in 1980 because we could spend our money better on other things.

It seems that if you take the substitution argument seriously, it implies an infinite amount of deflation in the long run because nearly everything we consume today would have been exorbitantly expensive if you go back in time far enough.

Hazel Meade writes:

Justin, I don't think that's accurate.
We're talking about cost-of-living components, which include food and housing. Not electronic gadgets.

I don't know the statistics, but I doubt that the cost of housing has gotten dramatically cheaper. Particularly not the cost per square foot. Food prices have declined since the start of the century, but that mostly applies to cereals, and seems to have leveled off. I'm not sure you could claim that a pound of meat has gotten cheaper adjusted for inflation in the last 20 years (I'd like to see some data on this).

Alex Bollinger writes:

I've known a few of these "Inflation is actually really, really high but the BLS is lying!" cranks. Why don't they take out a loan and they buy a lifetime's supply of all the durable goods they think they'll need? Then they could make minimum payments and wait for the loan to be worthless. Considering some of their rhetoric, that'll only take a year or two.

If these people really think they've figured out the secret to inflation that investors, bankers, and bureaucrats don't know, then they could be making bank. They should put their money with their mouth is. I don't see Ron Paul doing that, though. It's almost like he thinks his followers are idiots that are easily played.

David R. Henderson writes:

@Alex Bollinger,
If these people really think they've figured out the secret to inflation that investors, bankers, and bureaucrats don't know, then they could be making bank. They should put their money with their mouth is. I don't see Ron Paul doing that, though.
Then you haven’t seen Ron Paul’s financial statement. His required disclosure when he was a Congressman showed that he really was putting his money [or his gold] where his mouth was. The guy walks the talk.

Chmee writes:

@David, you state to Rusty "20 pounds of chicken would give substantially less utility than before. How do we know? Because with the original $100, the consumer had the ability to buy 33 pounds of chicken, and chose not to. 20 is less than 33."

How do you measure 'utility'? Maybe in the world of economics it means one thing, but here in the real world it means another. Utility means I have a $100 food budget. Maybe I will need to buy more chicken than beef since beef steak has 16% more protein, so I would need to buy 23 lbs to compensate for the difference in the "balanced" mixture of 50/50.

So let's look at it from our real world perspective. If prices go up for one item that is of better quality, real or perceived, then a person has two choices to make, reduce his consumption and/or buy lower priced items (and sometimes even inferior products). Maybe one could buy less beef and more chicken to offset the cost as you illustrate. Or more likely, as Michael points out, buy lower quality ground beef instead, because reality is that the average consumer has already been pushed out of that steak market anyway. (Just for fun, ground beef has 9% more protein than chicken, assuming the 90% lean. 85% ground beef has 20% less protein than chicken, so a serious net loss here.)

So if I want to be smarter about spending my $100/month, in order to maintain the same protein intake I might scrap the beef altogether and buy just the chicken. Then I could buy 20 lbs of chicken at your now assumed price of $4/lb for $80 and have $20 left over to buy some other protein source commodity. Maybe use the savings to buy some veggies and fruits instead, which actually would be a good thing anyway since you can obtain the same protein levels per pound from most of them and have a healthier and more varied diet.

So here is my take away from the utility of the real world that Dr Paul alludes to: One can rely on a CPI to measure prices and economic health all one wants to, the bottom line is that my food bill goes up each year higher than that CPI, chained or not. People on fixed incomes at retirement, and in fact the average person in this country who is still working, are seeing their real income after inflation not simply remaining stagnant but actually going down.

Even if the average person could fully comprehend the math behind it, CPI is a meaningless number to them in terms of their real purchasing power for the commodities of basic survival. My cheap Chinese plastic toys cost me less than a year ago, but I don't need them. I need to feed, clothe and house my family, and make sure I can pay my utility bills to cook, light and heat my home, pay my real estate taxes to keep the sheriff from knocking at the door to take it all away from me. And the price for all of those (with the possible exception of clothing) keeps rising.

So maybe I'll just take the $20, buy a case of beer and drink to forget...

Casual Observer writes:

[Comment removed pending confirmation of email address. Email the webmaster@econlib.org to request restoring this comment. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

Enial Cattesi writes:

@Alex Bollinger:

If these people really think they've figured out the secret to inflation that investors, bankers, and bureaucrats don't know, then they could be making bank. They should put their money with their mouth is. I don't see Ron Paul doing that, though. It's almost like he thinks his followers are idiots that are easily played.

In other days this kind of comment would have been labeled ad hominem around this parts. And yes, these people had figured out the secrete to inflation: you print money, you get inflation. It is very simple.
@Justin:
It seems that if you take the substitution argument seriously, it implies an infinite amount of deflation in the long run because nearly everything we consume today would have been exorbitantly expensive if you go back in time far enough.

Why take only the PC as an example? How about software licenses, milk, butter, bread, pork, chicken breast, a college degree, cars, gas, housing, television sets, house ornaments, toilet paper, books, chewing gum, coffee etc.?

tom writes:

As others have noted your example doesn't make any sense. The rise in chicken prices might make steak more favorable relative to chicken but it also makes lower cost items than chicken more favorable. Given normal assumptions, such as limited budgets, the default position would be to expect substitution of lower priced goods than higher priced.

Substitution is about changing the proportions in the bundle you buy in response to relative price changes.

How does it account for the effects of the change in demand on prices? If chicken goes up to $4 a lb and 25% of chicken purchases are substituted because of that what would the expected price of chicken have been without the switches? $5?

CPI understates inflation already because rising prices lead people to more frequently consider less expensive options. Price increases are muted by consumer reaction, which, as far as I can tell is used as justification to lower the effect of inflation again.

Rick Hull writes:

@DRH

False. 20 pounds of chicken would give substantially less utility than before. How do we know? Because with the original $100, the consumer had the ability to buy 33 pounds of chicken, and chose not to. 20 is less than 33.

This is deeply unsatisfying, at least in terms of understanding the mechanics of chained CPI.

Can you walk through Rusty's example and do the math for us? Given just the info we have in your example and his adjustment, how do we spit out a useful number?

David R. Henderson writes:

@Rick Hull,
Can you walk through Rusty's example and do the math for us?
I did, but you said it was “deeply unsatisfying.” So I’m not sure what you want. Here’s what I would suggest as a starting point: read about the Laspeyres price index here. Once you’ve done so, tell me.

David R. Henderson writes:

@Rick Hull,
Can you walk through Rusty's example and do the math for us?
I did, but you said it was “deeply unsatisfying.” So I’m not sure what you want. Here’s what I would suggest as a starting point: read about the Laspeyres price index here. Once you’ve done so, tell me.

William B. writes:

If you preferred $3.00 chicken to $7.00 steak, and you switched to $7.00 steak because you preferred it to $4.00 chicken, there is an IMMEASURABLE loss taking place, because you still ended up not getting what you preferred out of the three choices. Hence, you not only suffer the financial loss, but your happiness also declines.

Justin writes:

@ Enial Cattesi

Why take only the PC as an example? How about software licenses, milk, butter, bread, pork, chicken breast, a college degree, cars, gas, housing, television sets, house ornaments, toilet paper, books, chewing gum, coffee etc.?

Fine, include all of those in a basket with the PC. The price of a modern PC's components will tend towards infinity if you date it back over the years. So unless you allow substitution to affect your inflation calculations, any calculation based on that basket will show massive deflation. It doesn't even matter if everything else is 10 times more expensive over the time period.

Rick Hull writes:
I did, but you said it was “deeply unsatisfying.” So I’m not sure what you want. Here’s what I would suggest as a starting point: read about the Laspeyres price index here. Once you’ve done so, tell me.

Fair enough. I haven't finished reading it. But I am not sympathetic to the notion that price indices are valid for how we would like to use them.

Each month he buys 10 pounds of chicken and 10 pounds of steak. The price of steak is $7.00 a pound and the price of chicken is $3.00 a pound. So he spends $100 a month.

I'll further assume that he is on a fixed food budget. We're only talking about price inflation. $100 can only go so far.

Now, we get some price inflation. Steak has shot up to $100 a pound. He can no longer afford 10 pounds of steak. We don't know exactly how his buying habits will change, but let's say he buys 20 pounds of chicken to compensate for the now-unaffordable steak. With the $40 left over, he can buy 0.4 pounds of steak for a special treat.

If the Bureau of Labor Statistics were to price the old bundle (10 pounds of each), it would say that the cost of that bundle is now $110. So the apparent increase in the cost of living is 10%. But some people would substitute a little out of chicken and a little more into steak. The fact that they would do that says that they don't need that $110 to get the same level of utility. They might need only $108 or $109. So the method of pricing a fixed bundle overstates the increase in the cost of living.

Under my revised example, the BLS prices the old bundle at $1030. So the apparent increase in the cost of living is over 10x. But our guy has substituted a lot out of steak and a lot into chicken. The fact that he would do that (forced, by the inflationists) says that he doesn't need that $1030 to get the same level of utility.

Sounds like hokum to me. Nonetheless, I will read and digest the Price index link and return. Thanks!

Mark V Anderson writes:

I agree 100% with Prof Henderson on this one.

Maybe a different example will explain substitution better:

Say every week I bought 2 sticks of yellow licorice and two sticks of blue licorice. I am completely indifferent to which color I buy, and they are the same price, so I arbitrarily buy half in each color. But then there is a supply shock to yellow dye, and that licorice goes up 30% in price. If we just took the continuing basket approach, then my licorice inflation went up 15%. In reality, I will substitute all my yellow licorice purchases for blue, and my true inflation equals 0%.

Obviously, we rarely have complete indifference to substitutions. But it is also true that when one product goes up a lot, we will mitigate some of this increase by substituting our consumption towards a product with less inflation. Sure the price increase hurts us, but not as much as our original consumption of that product implies.

I don't know how the BLS adjusts for inflation to account for substitution. But it seems to me they should do this by changing the basket of products purchased more rapidly. In contrast to Thomas DeMeo, I think the best inflation indicator would be to use some kind of instantaneous basket that changes constantly. Thus when inflation results in consumers changing their basket of consumption,which is almost always towards a lower inflation basket, the inflation would be measured by averaging the inflation of the old and new baskets. I have no idea how this could be done practically, but I can imagine it mathematically.

Micah writes:

Rusty's example explains it quite clearly (inadvertently I think): "let's assume that the price of steak increases to $1,000 per pound, resulting in almost total substitution. Consumers are now buying 20 pounds of chicken (for $60) and no steak. A CPI based on this substitution would show a 40% decrease in prices. Certainly not an accurate representation of inflation."

40% deflation is not the correct measure; there has obviously been some loss of utility. It seems equally obvious that 9930% inflation is not a useful measure. Therefore, substitution must be considered.

The extent to which it gets considered and the math that is used to incorporate it into the index are probably better left to professional economists (or at least to someone other than a guy on his couch eating Swedish Fish, whose utility of deeper understanding is suffering a serious diminishing returns problem), but it does need to be considered.

Micah writes:

Also, to Thomas DeMeo's point: it may be true that it is more fundamental to "create a list of reproducible purchases, analyze the data and hope it tells us something useful," but you're addressing the wrong question.

The appropriate question is this: if we presuppose that the government should give people checks each month and that those checks should be adjusted year-on-year to keep their standard of living constant, how big should the adjustment be? Clearly, in the example above, the adjustment should not be 9930%.

Adjusted for inflation, the shopper will receive $100*(1+X). X is presumably too small for him to afford steak at the new price, so he now buys 20 lbs of chicken. However, X is large enough that he can upgrade other purchases to keep his total utility constant.

Determining X is a difficult problem. The answer we reach will undoubtedly be wrong. The best answer may be different for different people. Those are not good reasons to avoid answering the question. Cost of living adjustments need to include substitution effects, even if the rate used is ultimately flawed and less fundamental than an unadjusted rate.

MikeDC writes:

Really the problem here is that price inflation and cost of living indices are often conflated. Including in the CPI. They're not the same thing. Strictly speaking, I agree that chain weighting is stupid for the purpose of measuring price inflation. If our benchmark was 10lbs of steak and the price of steak increased, then to accurately measure the inflation level you need to continue to measure 10lbs of steak no matter the cost.

Price inflation occurs across an entire economy and thus needs to contain all priced goods and services. Employing substitution doesn't make sense in that context.

More simply.... Substitution is a reaction to a price change, and thus should not be included in its measurement.

The CPI is right though, to include some level of chain weighting insofar as it measures the cost of living. While it's true there's a utility hit to be had from changing my market basket in response to, say, a rise in the price of steak, it's also implicit that the hit is less than I would suffer by not substituting (is we assume utility maximizing consumers).That is, the chain weighted CPI does a better job measuring cost of living than price inflation by itself would.

The problem is that the CPI is presented as an inflation index, not a cost of living index.

Mark V Anderson writes:

MikeDC --

I disagree. It is true that the price increase in the steak should be included in the CPI, but it is the weighting that should change because the price increase will result in less consumption of the item which increased in price. Thus the basket of items purchased should include a lower percentage of steak when it increases in price, assuming consumers do decrease their consumption in the face of a price increase.

Jacob A. Geller writes:

People are mixing up monetary and fiscal policy.

Chained CPI isn't really about getting an accurate picture of changes in the price level (which would be useful for conducting monetary policy), it's about getting an accurate picture of changes in the cost of living, a totally separate concept, which is useful for the conduct of fiscal policy, in particular Social Security Cost of Living Adjustments (COLA).

Social Security payments are periodically adjusted for inflation, as estimated by the (non-chained) CPI-W, so that seniors can afford roughly the same standard of living as before. But some observers (i.e. Republicans) have rightly observed that rising inflation (CPI-W) is not necessarily the same things as changes in cost of living. Right now some of the fastest growing countries in the world also have some of the highest rates of inflation -- they are experiencing decreasing costs of living alongside a rising price level. Inflation *can* translate unambiguously to a higher cost of living, but it doesn't necessarily, and it's really the cost of living that matters for fiscal policymakers in Congress (as opposed to monetary policymakers, who are concerned with inflation). If the price of steak skyrockets to $1 trillion per steak, for example, then Congressional Republicans wouldn't think it's a very good idea to increase Social Security payments by $1 trillion per steak consumed by SS recipients, because their *cost of living* won't actually increase by $1 trillion per steak consumed, obviously, because they'll substitute steak for chicken. Now, their cost of living *will* increase, because now they will have to substitute away from steak and into chicken, which means lower utility for them (because of revealed preference, as David explained), requiring a larger Social Security check to compensate for that huge price increase. But that Social Security check will simply NOT have to change by the same amount as "inflation," i.e. $1 trillion per steak consumed. That's just silly.

Rick Hull writes:

@DRH

I sort of understand the difference between Paasche and Laspeyres, and I've got a better sense of the nuance of your denial of Ron Paul's position. These are just two different ways of looking at giant sets of numbers. To the extent we are mindlessly applying transformations to big sets of numbers (facts) in order to get a "macro" sense of the data, both are "valid".

The problem Ron Paul has, I think, is that we are not mindlessly applying these transformations and using them in a scientific, politically neutral way. There are similar critiques of the usage of GDP, which I'm sure you are familiar with -- once we measure it, we start to optimize for it -- and we are vulnerable to the charge of the drunk looking for his keys under the lamppost.

I'd like to try to run through my revised example using Paasche versus Laspeyres:

  • P0(chicken): 3
  • P0(steak): 7
  • P1(chicken): 3
  • P1(steak): 100
  • Q0(chicken): 10
  • Q0(steak): 10
  • Q1(chicken): 20
  • Q1(steak): 0.4
  • Paasche(x): sum(P1(x) * Q1(x)) / sum(P0(x) * Q1(x))
  • Paasche(chicken): 3 * 20 / 3 * 20 = 1
  • Paasche(steak): 100 * 0.4 / 7 * 0.4 = ~14.3
  • Paasche(chx+stk): 60 + 40 / 60 + 2.8 = ~1.6
  • Laspeyres(x): sum(P1(x) * Q0(x)) / sum(P0(x) * Q0(x))
  • Laspeyres(chicken): 3 * 10 / 3 * 10 = 1
  • Laspeyres(steak): 100 * 10 / 7 * 10 = ~14.3
  • Laspeyres(chx+stk): 30 + 100 / 30 + 70 = 1.3

Look about right?

To the extent our intrepid eater has been ~forced~ away from buying steak due to inflation, I am not keen on on saying that he is in some sense indifferent to the steak versus chicken choice. It still feels like a way for the inflationists to convince us that we're just as happy (or nearly so!) as before the price inflation.

Rick Hull writes:

Whoops, big error. 100 * 10 = 1000, not 100. The following is incorrect:

Laspeyres(chx+stk): 30 + 100 / 30 + 70 = 1.3

Revision:

  • Laspeyres(chx+stk): 30 + 1000 / 30 + 70 = 10.3

This number seems much more relevant to human welfare, per my earlier comments, than the Paasche number.

Rick Hull writes:

@DRH

Essentially, I'm having trouble seeing the flaw in the following argument:

By reasoning that chained CPI measures the cost of living, we could inflate prices (theoretically holding wages constant) until everyone is forced to eat dog food and tree bark at a subsistence level, and still claim that the cost of living has not risen significantly.
Rick Hull writes:

In other words: I remain deeply unsatisfied with explanation X until explanation X can refute the argument above, which we should all agree must be fatally flawed in some way.

Comments for this entry have been closed
Return to top