[another_comment]

>> I think it might be time to define a new metric that more closely reflects the actual cost of living.

We had different inflation stats back in the day. The US government has been slowly changing them. This guy uses the old official methodology to calculate inflation today. He gets way higher inflation using the old methodology and plugging in todays numbers. https://en.wikipedia.org/wiki/Shadowstats.com

There is a lot of controversy here, but I think inflation is definitely higher than reported.

[perl4ever]

>I think inflation is definitely higher than reported.

I don't, and it's pretty straightforward why. You can argue about how high inflation was over the past year and about technical details. But if you think it's significantly and systematically off and has been for a long time, then it would compound and make a big difference.

Back when my sense of "normal" prices was established, a pound of spaghetti was a dollar. The official total amount of inflation in the last twenty years was a little over 50% or a little over 2% a year. I'm confident the official figures are in the ballpark, because if inflation had really been say 5%, then spaghetti would be twice the price it is.

Car prices are also a useful benchmark. A Honda Civic EX is a pretty standard unit of transportation. In 2001, it was $17350, and in 2021, $25495. That's an increase of about 47% vs. official CPI went up about 50%. And notice that we're not accounting for any sort of improvement in the product, such as all the safety equipment, more horsepower, etc. So you can't complain about "hedonic" adjustments.

A funny thing on the Wikipedia page on Shadowstats is:

"Shadowstats has not changed its $175 per year subscription fee since at least 2006"

If there's so much inflation, how come the price of reading about it hasn't gone up?

[jfengel]

Also durability. I just end-of-life'd my car at over 300,000 miles. It was a 2007 Honda Fit that cost around $15,000. My previous car was a 1992 Honda Civic that cost around $12,000 and got 250,000 miles. Before that was a 1984 Civic that got 150,000 miles. (No idea what the original price was; I got it used.)

I hope my new car will reach 500,000 miles. Maybe not. It's my first non-Honda, who appears to have passed its quality crown to Subuaru. Regardless, anecdotally, price-per-mile seems to be going down.

(I will note, though, that gas mileage doesn't appear to be improving. I should perhaps have gotten a hybrid, but I really want an electric.)

[ab_testing]

Interesting that you talk about sphagetti and a honda civic. Now do the same calculation for rent, mortgage or health insurance

[perl4ever]

The government already did, and they got the same answer. My point was that can be confirmed with something that is easy and simple to estimate.

I guess you're implying that these things could diverge and the inflation rate matching could just be a coincidence.

But I don't think it's possible, because they're all part of a feedback loop. The people who build cars have to pay for rent and health insurance. The people who manage apartments and work for insurance companies eat and drive cars.

Edit: The selling price of my home in 2019 was 53% more than the selling price in 1995. The CPI was up more, roughly 70%.

[tanseydavid]

>> "Shadowstats has not changed its $175 per year subscription fee since at least 2006"

>> If there's so much inflation, how come the price of reading about it hasn't gone up?

This is unserious analysis.

You were making a decent point up until you resorted to this type of rhetoric.

[perl4ever]

It's half-serious, I don't call it analysis; I think it's weak but genuine evidence. It was meant to be a Parthian shot; the rest of my comments don't hinge on it.

[joelwilliamson]

The Shadowstats methodology is to take BLS CPI and add 3%.

If your theory about what's wrong with CPI is "BLS is subtracting 3% from the true value" that is a useful fix. If you're concerned about the basket of goods they use, or deflators for quality improvements, or substitution effects, or any other technical concern it isn't clear that Shadowstats provides any value.