Thread by @Mssr_le_Baron

To reiterate, the Feds have a number of tricks to hide the football in inflation, and probably have even more that I miss because they're sneaky about it.

Let's go over some. t.co/Nfb4MqP5zk

1. Hedonic Adjustment: This is when the price changes are adjusted to reflect real (or, importantly, fictive) increases in quality or product enjoyability. It's a technique invented and most suited for electronics, where there was and is massive improvement each year.

2. Government-set Prices: The components of the CPI include markets, like healthcare, where the government has a large say in the prevailing prices. By refusing to change Medicare and Medicaid rates, the government can hold elements constant or increase them less.

3. Consumer Basket Substitutions: When prices for some goods increase, the basket they use to measure CPI is reweighted to substitute cheaper goods for more expensive. This is invalid because goods are rarely true substitutes, but exist as preference substitutes.

This is a little harder to explain, but if you like pork and only switch to chicken when pork is too expensive, you have experienced a degradation in your subjective quality of life. The *bundle of goods* American consumers really consume is non-arbitrary.

True costless substitution is only valid in industrial inputs, where different oils are treated as perfectly interchangeable inputs, different carb sources are just building materials, etc. There can be real substitution only caring about cost. But consumer value is subjective.

4. Cost of Housing: Owner's Equivalent Rent, aka voodoo. This is a measure of the hypothetical rent the owner would have to pay if renting. Instead of using YoY home price increases and YoY same-unit rents, weighted by home ownership %, they use this hypothetical number.

Let me put it to you this way: How often are rental and sale Zestimates off? And that's with all the resources of Zillow behind it, and Zillow has all the revenue sharing of the whole MLS services funding that. Huge amounts of ML go into it.

It still can't account for factors.

OER fails because of the calculation problem: prices are a fast heuristic to measure true value because they have huge incentives to correct mispricings, so they have bounded wrongness. We are not good at calculating hypothetical situations and values.

This problem is much, much worse when you calculate OER by doing a survey of homeowners (which they do, they literally ask people how much they think rent in their area is). People *consistently* underprice rents for the same reasons Boomers think candy should be a quarter.

Our emotional memory of our home price is built when we actually cared about these things, which is when we bought them. The less it matters to us, the less the average person will bother keeping their idea of market rent or market price accurate (they won't, ask Boomers prices).

The CPI print for June says YoY shelter was 5.5%. In Manhattan, 23.1% own and the rest rent. If we take a 10% drop in YoY closed sale prices and a 36.9% increase in rents, we get a ownership weighted increase of 22.4% in cost of shelter.

If we could do that for every market and weight by population, we would have a true and traditional cost of shelter index. OER is a new method of calculation that fundamentally and systematically underestates housing cost inflation, which gets pegged to close to 0% by design.

5. Shrinkflation, Classic and Hidden: Classic shrinkflation is when product sizes go down or become less substantial. It's relatively easy to catch size decreases if they care to (they don't). Watering down is hardy, but you can measure by active ingredients or headliners.

What we are seeing, new to this economic depression, is Hidden Shrinkflation. This operates by quality fade. Every product has a QA process which determines what products are rejected or accepted. Products that receive less QA or pass lower bars are graded accordingly.

How does this operate in practice? For a branded product, the best grades will end up in their flagship brand. Slightly less high end, but still high end, is often resold to Costco under the Kirkland brand. The middling cuts go to a middle market brand. Below that, store brand.

Below that, you get factory rejects sold at their factory outlet as "funny vegetables" or other similar products. Or they get industrially reprocessed - dog food veggies, product bag textiles. The worst stuff is discarded. What happens with QA fade? Everything moves down a step.

What would be rejected is processed, what would be processed is sold as store brand and sometimes literally rots on the shelf (don't eat dog food, kids). Everything seems to go bad faster or break more often than it should. That's because QA costs money through reject rate.

6. Direct Intervention Analysis: The BLS has discretion to factor out "outlier events", discounting massive one time, temporary shifts in price.

Good thing we don't hear anything about that lately. Eheheheh.

It's hard to know how much they've adjusted with this technical method, because the details are proprietary. Assume, however, the worst, and that the magnitude is similar to other adjustments listed.

In conclusion, these statements do not represent the views of my employer, a bulge bracket bank, nor do they constitute investment advice. They are legally not statements of expertise, they are personal opinions.

I'm shitposting, lads.

@threadreaderapp unroll