>Funny, for me it seems terrible exactly for the opposite reason. These kinds of baskets are too heavily weighted by things that go relatively down in costs like technology and airplane tickets.

Have you looked at the actual weights?

https://www.bls.gov/news.release/archives/cpi_10142026.htm

Airline fares make up a whopping 1.04% of the CPI basket. "Information technology commodities" makes up a whopping 0.752%. Meanwhile the weights of essentials are: Food (13.507%), Energy (7.418%), Shelter (35.323%), Medical care services (6.834)

You might as well complain about that the basket contains tomatoes because you don't like tomatoes.

It's a statistical tool concerning populations, not a sampling of what hacker news user cicko in their wisdom thinks increases quality of life.

It is just an example, but you come to the base of my argument: CPI basket is an average - and you cannot use it to make statements about individuals. If Netflix is nothing to you, then look at houses: My granparents (born even after 1920s) had an outhouse. That means, every time you need to use the bathroom, you left your house through the garden - come wind and weather. That meant in the winter, you were walking through the snow and do your business in a small, smelly outhouse without no access to running water whatsover at sub-zero degrees. Yet there are people running around saying that our grandparents could afford houses, while we can't anymore.

The quality of the house your grandparents (for 1920 probably your grand-gran parents) is vastly different. Those houses cannot be compared at all, when I take my dump in the heated warmth in a nicely smelling place listening to some music over my airpods.

It is not necessarily politically influence. You simply cannot measure inflation properly. See my siblin comment about technology - it is just impossible to compare the life of a 1920s person to a 2020s person based on tomatos, butter and rent. Because the access to products that didn't exist at the given quality level in 2020 is impossible to price in.

You can think of it with a thought experiment: A smartphone (with hypothetical internet) in 1920 would be worth almost all the money in the world. If you simply had one smartphone in 1920, the richest billionares would offer most of their wealth to have one. Hence in 1920, a smartphone would be worth at least $10M in todays money. Yet you can buy one for $500 today - so you are twenty-thousand times richer than 1920. The same applies for flatscreen TVs, antibiotics (really valueable when you need it), polio vacination etc.

You can't measure this with CPI, yet people complain about the CPI is fake. It is just a statistical average, no more no less. It always was irrelevant. It is the arguments people do with that number, that are bullshit. Basic statistics teach you never use an average to come to conclusions about the individual.

I kid you not, ipads. That is what the rich out of touch fed officials tried to claim to a chorus of boos.

You really couldn't do that yourself? I linked the data tables with all the information you need. I get you might be lazy but that's hardly an excuse now that we have AI. Anyways I asked AI, here's what it calculated:

    Owners' equivalent rent of primary residence (sub6): score=-9.9748, rel_imp=24.937, yoy=3.3
    Health insurance (sub4): score=-10.3886, rel_imp=0.818, yoy=-9.0
    New vehicles (sub4): score=-11.6188, rel_imp=3.748, yoy=0.6
    Used cars and trucks (sub4): score=-13.9776, rel_imp=2.688, yoy=-1.5
    Motor vehicle insurance (sub4): score=-22.0197, rel_imp=2.531, yoy=-5.0

OER seems to be there not because it's dropping, but because its weight is huge, so if it grew even slightly slower than overall inflation it would be a significant drag. All the vehicle related stuff are probably related to the car market being crazy during the pandemic and then normalizing. No idea what's going on with health or even car insurance. Of course this is all 1 year's worth of data so oddities like this are expected You'd probably want to redo this with 5, 10, or 20 year data.