[throw0101a]

> If milk is $4/gallon today and still $4/gallon 12 mos. from now, that's 0% YoY inflation.

Conversely, if there was a one-time jump in a particular item, it will take a year before it gets 'removed' from the inflation numbers.

Extremely contrived example: if gas/petrol was $1/L in December 2021 (and generally in all of 2021), but $1.20/L in January 2022, then there will be a 20% YoY jump in inflation for the January number comparing Jan 2021 to Jan 2022.

Now if gas stays at $1.20/L in February 2022, it will still register as 20% YoY even though the price has not changed month-to-month. That 20% (YoY) is "stuck" in the system until January 2023 when we're comparing $1.20/L to $1.20/L.

A one-time jump can 'skew' the numbers if all you look at is YoY metrics.

[JumpCrisscross]

Which is partly what we’re seeing. CPI rose 0.4% MoM in October, or 4.8% seasonally adjusted. That’s above target, but better than YoY.

[fizx]

Really wish we reported the 4.8% number. Seems more reasonable for a monthly report to focus on the changes of the month.

[Bootvis]

I agree it's confusing. I think they do it to remove seasonality and the experts on the topic probably don't understand anymore why the rest of us think their way of reporting is confusing to lay people.

[treis]

Seasonally adjusted != multiply by 12. Prices naturally fluctuate throughout the year due to weather, consumer patterns, and business cycles. Seasonal adjustment is an attempt to account for those changes.

[JumpCrisscross]

> Seasonally adjusted != multiply by 12

My brain swapped annualised with seasonally adjusted. Sorry.

[reedjosh]

Beyond seasonal price fluctuations, wouldn't it be 1.04^12 = ~6% annually?

[Izkata]

0.4% is 0.004, so 1.004^12 = 1.049... or 4.9%

[reedjosh]

Derp, yeah. Thanks.

[orra]

I agree the annualizing the monthly figure would be handy, as an additional figure. Looking at annualized monthly figures, it was also apparent (say) six months ago, that the monthly annual figures would almost certainly continue to get worse.

[thunky]

Also note the four most recent (seasonally adjusted) monthly CPI numbers:

  Jul 2022: 0.0%
  Aug 2022: 0.1%
  Sep 2022: 0.4%
  Oct 2022: 0.4%

[soared]

It’s not really skewing anything though, since the price is 20% yoy?

[danielmarkbruce]

It is skewing if you are looking for an up to date change. It's like, is it better to get two points on a curved line and draw a straight line through them, or is it better to calculate the derivative and the the slope at a specific point?

[joshuahedlund]

That depends on how noisy or smooth the changes are. That's why it's valuable to look at both.

[danielmarkbruce]

It may be - but to the above poster's point - if gas went up 20% in Jan of this year and then stays flat... is it helpful to think of prices as "going up" ? It seems misleading.

[jerf]

"is it helpful to think of prices as "going up" ? It seems misleading."

Here you are encountering a common human cognitive failing, which is the belief that there is some sort of objective answer to the question "are prices going up?" that we should all be able to totally agree on, somehow floating in Platonic space without reference to any particular measure of "prices increasing".

The problem is that if you drill down to the question of "what does it mean for prices to be 'going up'?", you must admit to the fact that there are multiple valid definitions of that. It just isn't possible or plausible to create one true definition.

In the presence of that fact, it becomes inevitable that there will be senses in which the price is going up, and senses in which they are not, and senses in which prices are going up more than other senses. That is the reality, which is complicated.

(One propaganda technique is to take one of these numbers, which really exists and is perfectly defensible on its own terms, and then use it in a context in which you know people are generally going to interpret it as one of the other senses of the term. Excitingly, by controlling which "sense" you anchor your listeners to, both "sides" of a debate can push the numbers in whatever direction favors them at the same time.)

According to our nationally-used, generally-accepted metric, if gas is the same price today as it was 365 days ago, inflation is zero. But does that mean the metaphorical person on the street is "lying" if they say prices are generally going up because gas is 20% more expensive than it was three hundred and sixty six days ago? There is a fundamental arbitrariness both to our metrics, and how we all feel about things. I've seen plenty of "How can annual inflation be %8 if my eggs are 2.5x more expensive than this time last year?" posts around lately. The literal answer to that question is obvious, but if someone's expenses involve more eggs than mine, either because their food is a bigger percentage of their home budget or they are a business for whom eggs is a major input cost, they may feel a higher level of inflation than I do, and they're not wrong. They've just got their own inflation metric that disagrees with the national one, but their own metric may well be more relevant to their life than the national one is.

[danielmarkbruce]

Most people in this thread are using prices and CPI interchangeably, for better or worse. My read of the thread is that most people in the thread know they aren't actually the same thing.

[dcow]

If the last time you looked was last year, then yeah. They went up this year.

What’s misleading about reporting yearly percentage increases in prices? I’m baffled.

[MobiusHorizons]

The misleading bit would be reporting yoy price increases every month. Still not necessarily misleading to someone trained in reading this data, but the average lay person would probably interpret yoy statistics reported monthly incorrectly when comparing adjacent months

[unethical_ban]

Sigh.

The YoY is useful for businessmen and economists.

People aren't checking their statements for milk prices from Nov 2021 right now. They just know "man it jumped up".

The point is that the everyday person cares about jumps and trends, and hearing "20% YoY increase" for 12 months is misleading to someone not thinking about it in economists' terms.

The fact a bunch of people here are debating what it means and how to interpret it proves my point.

[UncleMeat]

Imagine a more extreme case where we reported inflation over the past 50 years each month. Every month we compute the price of the basket 50 years ago and the price of the basket today and compare. This number would go up when the most recent month had more price increases than the month 601 months ago. It would go down when the situation was reversed. It would be a number, but it would tell you very little useful information.

It is also the case that going extreme in the other direction is ridiculous. Imagine a daily inflation measure. Also nearly useless. "Oh, prices were flat today so everything is fine."

Published inflation numbers in media are usually used to make either the claim "everything is fucked, you should be mad about public policy" or "everything is fine, you should be happy about public policy." To me, this means that the reported should ideally be tied to some cadence that matches public policy. I'm not sure what that cadence is.

[danielmarkbruce]

You appear to be saying: factually reported data cannot be misleading.

This is wrong, but so obviously so that you must be making another point?

[ncallaway]

I don’t think anyone was saying it was misleading, rather they were pointing at a nuance of the data.

[cwilby]

That literally leads to the conclusion that prices rose, in your example.

[rflrob]

The conclusion that “prices rose” is the correct one. What you should not conclude, even though the inflation has been at 20% all year is that “prices are (still) rising”.

[dcow]

But they have risen since last year. YoY they are still rising.

[zhte415]

It isn't, yet it is.

The price is still 20% you, and still 0% mom. If reported as inflation still at 20%, does that affect perceptions of inflation, and if so, expectations of inflation? Because expectations of inflation often turn out to be self fulfilling drivers of inflation.

Reported as inflation stable even if it is 20%? Better? Not that 0% is a great target, but that's a separate issue.