The 1990 methodology looks like "CPI-U + 4%" and the 1980 methodology looks like "CPI-U + .25%*(year-1984)." If we assume for a second that there is a methodology beyond hand-waving those offsets into graphs in order to obtain money from people like my parents, we would probably be looking for basket items with constant, high price increases over the last 40 years and (log-)linear escalating price increases over the last 40 years. Something to do with health care and education respectively, maybe?

That's almost certainly giving Shadow Stats more benefit of the doubt than they deserve, but after the manufacturing figure debacle (see below) I'm open to the idea of spelunking into these calculations and debating the merits of various choices they make.

https://qz.com/1269172/the-epic-mistake-about-manufacturing-...

(Summary: "American Manufacturing is doing better than ever" figures were obtained by including Intel in the basket and multiplying up their revenue using Moore's law as a hedonic adjustment.)

It's the latter, of course. Because if these numbers were actually correct, the market would have figured that out by now, and it would be reflected in the level of interest rates. Yet, it is not. So either people are losing money by lending at rates that are well below the level of inflation, and everyone is too stupid to have noticed that over the last 30 years or this article is nonsense.

Two economists are walking down the street. One sees a $100 bill on the ground and says, "Hey, isn't that a $100 bill?" The other laughs-- "Of course not. If it was, someone else would have taken it by now."

While I agree with you in this specific case, this joke always pops into my head whenever "efficient markets" are brought up.

The author did link to a much more detailed explanation: http://www.shadowstats.com/article/no-438-public-comment-on-...

I skimmed it, it’s a lot to take in. In general it seems like the author doesn’t explore deeply why the academics wanted to change the inflation measure and takes for granted that it was simply ivory tower twiddling combined with political machinations. I feel like I am not getting a full picture. But I haven’t understood the author’s thesis in enough detail to comment rigorously.

Subjectively my personal experience over the last 20 years seems more in line with government stats than the author’s shadow stats: generally low inflation with a lot of volatility in certain commodities, and upward pressure in housing, combined with stagnant wages. Inflation at 10% over the last 20 years doesn’t really pass the sniff test considering my memory of prices in 2000.

The author contends that the old fixed-basket-of-goods-with-fixed-weighting that doesn’t account for substitution is more appropriate that modern methods which account for substitution.

So if you started tracking inflation before automobiles were available, then your 2020 stats shouldn’t be based on costs of automobiles at all, but should reflect horses and carriages with weighting reflecting their use in your baseline year.

That this approach is both seriously flawed and increasingly so the more rapidly available products and consumer behavior evolves should be obvious.

The old method is simpler, has less room for difficult assessments like substitution and hedonic adjustments, but is fundamentally broken in dealing with the modern world.

Using an outdated fixed basket and weightings means you are ignoring what are currently essential goods but giving heavy weight to things that used to be essential goods but have become niche vanity consumables or luxury collectibles.

Does the new methodology include a negative hedonic adjustment for substitution of inferior goods? Anecdotally that sort of thing seems to have been on the rise for a decade and by my reckoning this scenario should definitely count as inflation:

    Wage stays constant
    Price of health care goes up
    Make do with worse health insurance
    Price of housing goes up
    Move in with parents / more roommates

Not quite. Remember that if you held the true value of a currency flat with increasing automation and technological efficiency you should get more for the same amount of money. Prices staying flat over time in the face of efficiency gains is its own sort of inflation.

I do agree that they should source things better.

> Remember that if you held the true value of a currency flat with increasing automation and technological efficiency you should get more for the same amount of money.

That...requires a very special definition of “true value”. If the “true value” is measured by consumer goods and services you can exchange it for, this is trivially false. If it is measured by labor hours of production output you can exchange it for, it is equally trivially true. Neither is uncontroversially the One True Meaning of “true value”.

> would make everything 6.73X more expensive

Not everything. The weighted average.