3 ms·

This strikes me as lazy data science and full of overreaching conclusions, because nowhere is uncertainty/spread/variance discussed. What are the errors/varianc

This strikes me as lazy data science and full of overreaching conclusions, because nowhere is uncertainty/spread/variance discussed. What are the errors/variance on these estimations for each year and each spending category? Are these inflation differences statistically significantly different? Comparing 7% vs 8% is already cutting it close without even knowing this crucial information. The headline vastly oversteps the analysis itself, there is simply not enough information presented to make these conclusions. This is one reason why peer-review is essential. This author can throw out a claim like millenials get inflation 14% worse than everyone else, failing to mention it's a difference of 1% over 7-8%, and moreover he did not even look at variance or statistical significance. This would not pass muster in a statistics 101 course, never mind a reputable journal.

betahiker 8y ago
> nowhere is uncertainty/spread/variance discussed This is the standard for reporting on inflation data. Do you know of a reporting on inflation where they mention an uncertainty measure? For example when reporting on the difference between this year and last year's inflation number? > failing to mention it's a difference of 1% over 7-8% Quoting from the article: > Over the past five years, millennials have seen their goods become 8.7% more expensive, Generation X 8%, and baby boomers 7.6%.
repiret 8y ago
In general I agree that more reporting should include measures of uncertainty. But I also think there are cases when talking about inflation where uncertainty measures are inappropriate. You can have uncertainty in the average price of avocados, or in the number of avocados that millennials consume on average. But those aren't very holistic views of inflation. If I define my holistic measure of inflation to be the CPI, then there's no margin of error. Its just a number that the Bureau of Labor Statistics publishes every month. I can say, for example, that the CPI for June 2009 was exactly 215.693. Or that the CPI increased by 1.47%, exact to the nearest 0.01%, between March 2012 and 2013. On the other hand, if you want to ask whether the CPI is any good as a measure of inflation, then in general inflation isn't precisely enough defined to talk about that with a simple margin of error. There is no exact definition of inflation that we're trying to approximate, so we can't say things like "we're 95% confident that the actual inflation value was between 1.4% and 1.5%".
- skybrian 8y ago
  It looks like there's a principled way to compute it: "Variance is a measure of the uncertainty caused by the use of a sample of retail prices, instead of the complete universe of retail prices. Each month the U.S. Bureau of Labor Statistics collects prices from a sample of approximately 77,400 commodities and services (C&S) quotes in approximately 21,500 outlets around the United States for the Consumer Price Index (CPI). In addition, we collect approximately 5,100 housing quotes each month." [...] "Margins of error are usually expressed as a statistic’s point estimate plus or minus two standard errors, so the margin of error on this CPI’s 1-month change is approximately 0.15 percent plus or minus 0.06 percent." https://www.bls.gov/cpi/tables/variance-estimates/home.htm https://www.bls.gov/cpi/tables/variance-estimates/home.htm
AnthonyMouse 8y ago
> This strikes me as lazy data science and full of overreaching conclusions, because nowhere is uncertainty/spread/variance discussed. It's not just that, the premise itself is nonsense. Inflation measures the value of money. We measure inflation using the price of stuff because measuring it in nominal dollars is obviously self-canceling. The relative difference in the price of stuff has nothing to do with inflation. Housing costs are high because of supply not increasing to match demand. If restricted housing supply increases prices by 5%, while supply and demand for clothing is unchanged but there is 5% inflation, then housing prices increase by ~10% while clothing prices increase by 5%. If there was 0% inflation then housing prices increase by 5% and clothing prices stay the same. The fact that housing prices increase by 5% relative to clothing prices has nothing to do with inflation at all. What would help millennials isn't less inflation, it's more housing. If anything inflation helps them, because they hold significantly more debt (in the form of student loans and mortgages) than older generations, and inflation allows the debt to be paid with cheaper money. It also increases the nominal but not real price of housing, which allows implementation of policies that cause real housing prices to decrease without putting existing homeowners underwater.

betahiker 8y ago

> nowhere is uncertainty/spread/variance discussed This is the standard for reporting on inflation data. Do you know of a reporting on inflation where they mention an uncertainty measure? For example when reporting on the difference between this year and last year's inflation number? > failing to mention it's a difference of 1% over 7-8% Quoting from the article: > Over the past five years, millennials have seen their goods become 8.7% more expensive, Generation X 8%, and baby boomers 7.6%.

repiret 8y ago

In general I agree that more reporting should include measures of uncertainty. But I also think there are cases when talking about inflation where uncertainty measures are inappropriate. You can have uncertainty in the average price of avocados, or in the number of avocados that millennials consume on average. But those aren't very holistic views of inflation. If I define my holistic measure of inflation to be the CPI, then there's no margin of error. Its just a number that the Bureau of Labor Statistics publishes every month. I can say, for example, that the CPI for June 2009 was exactly 215.693. Or that the CPI increased by 1.47%, exact to the nearest 0.01%, between March 2012 and 2013. On the other hand, if you want to ask whether the CPI is any good as a measure of inflation, then in general inflation isn't precisely enough defined to talk about that with a simple margin of error. There is no exact definition of inflation that we're trying to approximate, so we can't say things like "we're 95% confident that the actual inflation value was between 1.4% and 1.5%".

skybrian 8y ago
It looks like there's a principled way to compute it: "Variance is a measure of the uncertainty caused by the use of a sample of retail prices, instead of the complete universe of retail prices. Each month the U.S. Bureau of Labor Statistics collects prices from a sample of approximately 77,400 commodities and services (C&S) quotes in approximately 21,500 outlets around the United States for the Consumer Price Index (CPI). In addition, we collect approximately 5,100 housing quotes each month." [...] "Margins of error are usually expressed as a statistic’s point estimate plus or minus two standard errors, so the margin of error on this CPI’s 1-month change is approximately 0.15 percent plus or minus 0.06 percent." https://www.bls.gov/cpi/tables/variance-estimates/home.htm https://www.bls.gov/cpi/tables/variance-estimates/home.htm

AnthonyMouse 8y ago

> This strikes me as lazy data science and full of overreaching conclusions, because nowhere is uncertainty/spread/variance discussed. It's not just that, the premise itself is nonsense. Inflation measures the value of money. We measure inflation using the price of stuff because measuring it in nominal dollars is obviously self-canceling. The relative difference in the price of stuff has nothing to do with inflation. Housing costs are high because of supply not increasing to match demand. If restricted housing supply increases prices by 5%, while supply and demand for clothing is unchanged but there is 5% inflation, then housing prices increase by ~10% while clothing prices increase by 5%. If there was 0% inflation then housing prices increase by 5% and clothing prices stay the same. The fact that housing prices increase by 5% relative to clothing prices has nothing to do with inflation at all. What would help millennials isn't less inflation, it's more housing. If anything inflation helps them, because they hold significantly more debt (in the form of student loans and mortgages) than older generations, and inflation allows the debt to be paid with cheaper money. It also increases the nominal but not real price of housing, which allows implementation of policies that cause real housing prices to decrease without putting existing homeowners underwater.