3 ms·

> 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

> 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.

CrazyStat 5y ago
> 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.
- lend000 5y ago
  > 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. Further recommended reading: https://www.lynalden.com/inflation/ https://www.lynalden.com/inflation/
  - CrazyStat 5y ago
    > 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. > Further recommended reading: https://www.lynalden.com/inflation/ 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 https://fred.stlouisfed.org/series/M2V
    - 5y ago

CrazyStat 5y ago

> 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.

lend000 5y ago
> 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. Further recommended reading: https://www.lynalden.com/inflation/ https://www.lynalden.com/inflation/
- CrazyStat 5y ago
  > 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. > Further recommended reading: https://www.lynalden.com/inflation/ 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 https://fred.stlouisfed.org/series/M2V
  - 5y ago