To be considered a millionare like when the term was coined (roughly 100 years ago) you would need to have about 26M in wealth today (inflation adjusted).
I don't understand what this is supposed to be showing. They seem to be only adjusting by a gas-specific CPI component: "1978 Gas Price x (2020 CPI for Gas / 1978 CPI for Gas) = Adjusted Gas Price in 2020 Dollars". Which seems like a tautology... The "X Year Gas Price" and "X Year CPI for Gas" seem so linked - if you adjust this number by a number that's supposed to reflect it's rate of change, you get a flat line, but... of course?
For instance, the spike between 2003 and 2006 certainly didn't feel like a flat line compared to the cost of other things, so wouldn't this graph be much more interesting if it was based on "CPI for everything BUT gas"?
If the cost of gas after inflation is as stable as that graph suggests, then most of the large increase in gas prices in the past year -- from ~$2 to ~$3 -- is the result of inflation.
They're adjusting gasoline prices by a gasoline-specific index as their "inflation estimate". Basically dividing one average gasoline price by a slightly different average gasoline price.
> If the cost of gas after inflation is as stable as that graph suggests
They are actually adjusting retail price of gas for the CPI component for gasoline. Which is a dumb thing to do: by definition that should be completely invariant, but it doesn't tell you anything except how well aligned the gas price average they are using is with thr one the BLS uses to calculate the CPI component for gasoline.
If the goal of excluding gas from the inflation analysis is to eliminate its volatile pricing, understanding its relationship to inflation should use the same exclusion to calculate inflation, and a moving average could smooth the volatility for a clearer trend line. The methodology used here essentially provides no meaningful information as the chart clearly shows. The only way that visualization is accurate is if the huge spike in average gas prices coincided with a huge spike in general inflation measures (it didn’t) or a huge reduction in gas expenses as a proportion of spending (which didn’t occur).
The calculation for CPI is not publicly available, but it has been believed for some time that gasoline is a huge part of the index. The interesting part is the relative flatlining of gasoline in unadjusted dollars since 2004 (for many logical reasons), and how the BLS could use that to their advantage to publish a lower CPI.
> The calculation for CPI is not publicly available, but it has been believed for some time that gasoline is a huge part of the index.
The BLS publishes the weights for the CPI calculations every year (see e.g. [1]). There's no need to speculate about it. Gasoline is about 3% of the index.
Fascinating! I didn't know these weights were published.
If you look at the data from 2005 to 2019, you can see substantial, seemingly arbitrary weight changes in gasoline (and other items) to come up with a stable CPI number. For example, 2.9% in 2003 to 5.2% in 2007 and back down to 3% with plenty of bumps. Legitimately curious: can you think of a reasonable justification for 60% weight fluctuations of a major element of the index?
Recreation is another weight on the list that seems to fluctuate back and forth without any noticeable trend except to smooth out the numbers.
And, from what I've read, the actual data used to multiply against these weights is not public [0]. For average gasoline prices, there isn't much manipulation they can do short of blatant fraud. But for something like recreation, photography, etc.... Well.
> Legitimately curious: can you think of a reasonable justification for 60% weight fluctuations of a major element of the index?
Yes. The weights are determined from survey data (Consumer Expenditure Survey, conducted by the census bureau). They aren't just made up.
As for why gasoline jumped from a weight of 2.9 in 2003 to 5.2 in 2007, probably because the price of gas doubled over that period and so it was a larger proportion of survey respondents' budgets, which results in a higher CPI weight.
> Recreation is another weight on the list that seems to fluctuate back and forth without any noticeable trend except to smooth out the numbers.
Again imputing motive without evidence.
Disposable income expenditures like recreation and photography naturally are highly responsive to economic conditions. When the economy is good people spend more on recreation. When the economy is bad they spend less on recreation. As the percentage of survey respondents' budgets that gets spent on recreation changes, so does the CPI weight. Nothing strange about that.
CPI naturally will be smoother than tracking individual items because of (1) random variations averaging out in the large basket of goods and (2) consumers adjusting purchasing behaviors in reaction to prices. There's no need to appeal to conspiracy theories about the BLS intentionally smoothing it by manipulating data.
> Yes. The weights are determined from survey data (Consumer Expenditure Survey, conducted by the census bureau). They aren't just made up.
That's sort of the point of the previously linked article. Rather than measure price changes of a standardized basket of goods and services, the CPI adopts an inflation metric that is reactive to prices (median Americans will tend to buy cheaper items, regardless of the change in prices of standard materials in the economy). So really it's more of a measure of the median American's budget, not inflation. The two are related of course, but inflation can look quite low while wealth inequality is rising, and vice versa.
> Rather than measure price changes of a standardized basket of goods and services, the CPI adopts an inflation metric that is reactive to prices (median Americans will tend to buy cheaper items, regardless of the change in prices of standard materials in the economy). So really it's more of a measure of the median American's budget, not inflation.
An unchanging standardized basket of goods would be a remarkably stupid way to try to measure inflation over a long time period, given the introduction of new goods, retiring of old goods, and changing preferences of consumers. If I want to measure inflation from 1970 to 2021 I don't want to do that by measuring the average prices of '69 vettes, slide rules, and Beatles LPs.
Over a short period of time you can make the assumption that most individual products are relatively stable, and this is in fact what the BLS does: weights are recalculated every two years and the products in the price survey are kept constant to the extent possible.
If you have a better approach to measuring long-term inflation than the current approach of periodically adjusting your basket of goods to make sure it still reflects the expenditure patterns of consumers, I'm sure your local univerity economics department would love to hear it.
> weights are recalculated every two years and the products in the price survey are kept constant to the extent possible
Perhaps you mean that the weights follow a 2 year moving average? They clearly change significantly each year, if you look at the weights for consecutive years at your original reference.
Your overall negative tone suggests you are emotionally invested in the CPI being correct, or perhaps just seeming smart. I'm more interested in the truth. While I agree that using consumer spending surveys is superior to having bureaucrats manually pick out weights, I think the budget data should constitute less of the current (100%) weight in the index, the weights should follow a much smoother 10 year SMA, and an index of raw commodities and business expenses should account for the rest of the index. As you say, this would better reflect expected short term price stability and register deviations from the norm.
> If you have a better approach to measuring long-term inflation than the current approach of periodically adjusting your basket of goods to make sure it still reflects the expenditure patterns of consumers
Collecting the data to calculate CPI the way I suggested is a difficult solo exercise. But if you are actually interested in an alternative estimation, start by looking at the growth of SMA(3) of GDP ^ (2/5) * M2 ^ (2/3). This is the baseline growth heuristic I have been using to consider one to be keeping up with inflation.
> Perhaps you mean that the weights follow a 2 year moving average?
No, I mean what I said. The weights used for CPI calculations are updated every two years. In between they are constant.
The BLS also updates weights more often based on estimates of consumer behavior (rather than surveys), which are used for C-CPI calculations. The published annual weights include these updated (estimated) weights for the inbetween years.
> Your overall negative tone suggests you are emotionally invested in the CPI being correct, or perhaps just seeming smart. I'm more interested in the truth.
I have fairly low tolerance for people who go around posting things like
(1) Average price of gasoline adjusted for average price of gasoline
(2) CPI weights are secret but it's widely believed (by who?) that gasoline is a huge factor, which the government uses to manipulate inflation numbers
(3) The government is manipulating weights to smooth out the CPI
It's fairly clear that you're well out of your depth on this material.
> an index of raw commodities and business expenses should account for the rest of the index.
This would no longer be a Consumer Price Index. There are other types of indices, such as the Producer Price Index (also produced by the BLS).
> Collecting the data to calculate CPI the way I suggested is a difficult solo exercise. But if you are actually interested in an alternative estimation, start by looking at the growth of SMA(3) of GDP ^ (2/5) * M2 ^ (2/3). This is the baseline growth heuristic I have been using to consider one to be keeping up with inflation.
Looking at money supply is pretty meaningless without also looking at velocity. M2 has been going up but the velocity has been going down [1].
Besides that, the exponents add to more than one, so even ignoring the extreme dubiousness of attributing GDP increases to inflation this is going to outpace (by an exponential factor) actual inflation under normal economic regimes.
Back to this
> Your overall negative tone suggests you are emotionally invested in the CPI being correct
I can say almost the same from my point of view: you seem to be emotionally invested in CPI being flawed and/or manipulated in such a way as to intentionally underestimate true inflation. You've made a number of false or at best dubious claims in support of that position. I've tried to answer them even-handedly.
> Looking at money supply is pretty meaningless without also looking at velocity. M2 has been going up but the velocity has been going down [1].
Are you not aware of the GDP component of velocity?... The purpose being to incorporate economic activity/effects of growing population into M2. It seems like you're just googling as you go (mostly effectively).
You're clearly an intelligent person, and I learned some things about the BLS's process that I haven't found in previous googles. Hopefully you learned something in this thread, too. However, the slew of logical fallacies (character assassination, twice, and appeal to authority) and overall negativity are serious hindrances to intellectual discussion. Positivity and skeptical open-mindedness will go a long way if you ever want to make progress on a controversial topic.
The graph says it all. There haven’t been any technological changes in oil refining to make gas change in value. That it costs more per gallon in raw dollar terms means that the value of the dollar relative to the value of a gallon of gas has dropped massively.
Even more if you figure that today’s cars are pretty damned efficient and use way less gas than a car from the 70s, so we extract more value from a gallon of gas in terms of work performed.
No wonder Americans drive old gas-guzzler planet destroying trucks with those prices.
A US gallon of fuel costs US$7 in Denmark. Which is A Good Thing for the environment. Only very environmental conscious people would buy a non gas-guzzler with those prices (which is also exactly what the numbers show on sold cars).
32 comments
[ 4.8 ms ] story [ 69.4 ms ] threadTo be considered a millionare like when the term was coined (roughly 100 years ago) you would need to have about 26M in wealth today (inflation adjusted).
https://www.usinflationcalculator.com
This is an entirely pointless exercise.
For instance, the spike between 2003 and 2006 certainly didn't feel like a flat line compared to the cost of other things, so wouldn't this graph be much more interesting if it was based on "CPI for everything BUT gas"?
https://www.gasbuddy.com/charts
A near 50% annual inflation rate? I sure hope not. Someone please tell me what's wrong with this graph or my interpretation.
Unsurprisingly this results in a flat line.
They are actually adjusting retail price of gas for the CPI component for gasoline. Which is a dumb thing to do: by definition that should be completely invariant, but it doesn't tell you anything except how well aligned the gas price average they are using is with thr one the BLS uses to calculate the CPI component for gasoline.
For which the answer is “close, but not perfect”.
The BLS publishes the weights for the CPI calculations every year (see e.g. [1]). There's no need to speculate about it. Gasoline is about 3% of the index.
[1] https://www.bls.gov/cpi/tables/relative-importance/2019.pdf
If you look at the data from 2005 to 2019, you can see substantial, seemingly arbitrary weight changes in gasoline (and other items) to come up with a stable CPI number. For example, 2.9% in 2003 to 5.2% in 2007 and back down to 3% with plenty of bumps. Legitimately curious: can you think of a reasonable justification for 60% weight fluctuations of a major element of the index?
Recreation is another weight on the list that seems to fluctuate back and forth without any noticeable trend except to smooth out the numbers.
And, from what I've read, the actual data used to multiply against these weights is not public [0]. For average gasoline prices, there isn't much manipulation they can do short of blatant fraud. But for something like recreation, photography, etc.... Well.
[0] https://www.forbes.com/sites/perianneboring/2014/02/03/if-yo...
You're imputing motive here without any evidence.
> Legitimately curious: can you think of a reasonable justification for 60% weight fluctuations of a major element of the index?
Yes. The weights are determined from survey data (Consumer Expenditure Survey, conducted by the census bureau). They aren't just made up.
As for why gasoline jumped from a weight of 2.9 in 2003 to 5.2 in 2007, probably because the price of gas doubled over that period and so it was a larger proportion of survey respondents' budgets, which results in a higher CPI weight.
> Recreation is another weight on the list that seems to fluctuate back and forth without any noticeable trend except to smooth out the numbers.
Again imputing motive without evidence.
Disposable income expenditures like recreation and photography naturally are highly responsive to economic conditions. When the economy is good people spend more on recreation. When the economy is bad they spend less on recreation. As the percentage of survey respondents' budgets that gets spent on recreation changes, so does the CPI weight. Nothing strange about that.
CPI naturally will be smoother than tracking individual items because of (1) random variations averaging out in the large basket of goods and (2) consumers adjusting purchasing behaviors in reaction to prices. There's no need to appeal to conspiracy theories about the BLS intentionally smoothing it by manipulating data.
That's sort of the point of the previously linked article. Rather than measure price changes of a standardized basket of goods and services, the CPI adopts an inflation metric that is reactive to prices (median Americans will tend to buy cheaper items, regardless of the change in prices of standard materials in the economy). So really it's more of a measure of the median American's budget, not inflation. The two are related of course, but inflation can look quite low while wealth inequality is rising, and vice versa.
An unchanging standardized basket of goods would be a remarkably stupid way to try to measure inflation over a long time period, given the introduction of new goods, retiring of old goods, and changing preferences of consumers. If I want to measure inflation from 1970 to 2021 I don't want to do that by measuring the average prices of '69 vettes, slide rules, and Beatles LPs.
Over a short period of time you can make the assumption that most individual products are relatively stable, and this is in fact what the BLS does: weights are recalculated every two years and the products in the price survey are kept constant to the extent possible.
If you have a better approach to measuring long-term inflation than the current approach of periodically adjusting your basket of goods to make sure it still reflects the expenditure patterns of consumers, I'm sure your local univerity economics department would love to hear it.
Perhaps you mean that the weights follow a 2 year moving average? They clearly change significantly each year, if you look at the weights for consecutive years at your original reference.
Your overall negative tone suggests you are emotionally invested in the CPI being correct, or perhaps just seeming smart. I'm more interested in the truth. While I agree that using consumer spending surveys is superior to having bureaucrats manually pick out weights, I think the budget data should constitute less of the current (100%) weight in the index, the weights should follow a much smoother 10 year SMA, and an index of raw commodities and business expenses should account for the rest of the index. As you say, this would better reflect expected short term price stability and register deviations from the norm.
> If you have a better approach to measuring long-term inflation than the current approach of periodically adjusting your basket of goods to make sure it still reflects the expenditure patterns of consumers
Collecting the data to calculate CPI the way I suggested is a difficult solo exercise. But if you are actually interested in an alternative estimation, start by looking at the growth of SMA(3) of GDP ^ (2/5) * M2 ^ (2/3). This is the baseline growth heuristic I have been using to consider one to be keeping up with inflation.
Further recommended reading: https://www.lynalden.com/inflation/
No, I mean what I said. The weights used for CPI calculations are updated every two years. In between they are constant.
The BLS also updates weights more often based on estimates of consumer behavior (rather than surveys), which are used for C-CPI calculations. The published annual weights include these updated (estimated) weights for the inbetween years.
> Your overall negative tone suggests you are emotionally invested in the CPI being correct, or perhaps just seeming smart. I'm more interested in the truth.
I have fairly low tolerance for people who go around posting things like
(1) Average price of gasoline adjusted for average price of gasoline
(2) CPI weights are secret but it's widely believed (by who?) that gasoline is a huge factor, which the government uses to manipulate inflation numbers
(3) The government is manipulating weights to smooth out the CPI
It's fairly clear that you're well out of your depth on this material.
> an index of raw commodities and business expenses should account for the rest of the index.
This would no longer be a Consumer Price Index. There are other types of indices, such as the Producer Price Index (also produced by the BLS).
> Collecting the data to calculate CPI the way I suggested is a difficult solo exercise. But if you are actually interested in an alternative estimation, start by looking at the growth of SMA(3) of GDP ^ (2/5) * M2 ^ (2/3). This is the baseline growth heuristic I have been using to consider one to be keeping up with inflation.
Looking at money supply is pretty meaningless without also looking at velocity. M2 has been going up but the velocity has been going down [1].
Besides that, the exponents add to more than one, so even ignoring the extreme dubiousness of attributing GDP increases to inflation this is going to outpace (by an exponential factor) actual inflation under normal economic regimes.
Back to this
> Your overall negative tone suggests you are emotionally invested in the CPI being correct
I can say almost the same from my point of view: you seem to be emotionally invested in CPI being flawed and/or manipulated in such a way as to intentionally underestimate true inflation. You've made a number of false or at best dubious claims in support of that position. I've tried to answer them even-handedly.
> Further recommended reading: https://www.lynalden.com/inflation/
I think I'll stick with economists whose job is to understand inflation rather than an engineer dabbling in investment advising.
[1] https://fred.stlouisfed.org/series/M2V
Are you not aware of the GDP component of velocity?... The purpose being to incorporate economic activity/effects of growing population into M2. It seems like you're just googling as you go (mostly effectively).
You're clearly an intelligent person, and I learned some things about the BLS's process that I haven't found in previous googles. Hopefully you learned something in this thread, too. However, the slew of logical fallacies (character assassination, twice, and appeal to authority) and overall negativity are serious hindrances to intellectual discussion. Positivity and skeptical open-mindedness will go a long way if you ever want to make progress on a controversial topic.
I can get 1.4gallons of milk and 5.5 stamp for the price of one gallon of gas (3.04).
In summary, gasoline is more expensive now then it was in 1978 and does not match the graph.
Even more if you figure that today’s cars are pretty damned efficient and use way less gas than a car from the 70s, so we extract more value from a gallon of gas in terms of work performed.
A US gallon of fuel costs US$7 in Denmark. Which is A Good Thing for the environment. Only very environmental conscious people would buy a non gas-guzzler with those prices (which is also exactly what the numbers show on sold cars).