Finally, news reports that month-over-month CPI has dropped

In a recent post, I complained that the media tends to report just the year-over-year inflation figures even though those numbers can remain high even when inflation has stopped or prices have even fallen. I said that the month-over-month figure is a better indicator of current inflation and should be reported as well.

Well today, they reported just that, confirming my suspicion that inflation likely peaked back in June and since then has ceased or even declined.

Prices dropped in the US in December for the first time since May 2020, in an encouraging sign that the inflation crisis may be easing.

According to the latest consumer price index (CPI) – which measures a broad range of goods and services – the cost of living dropped 0.1% in December compared with a rise of 0.1% in November. The annual rate of inflation fell to 6.5% from 7.1% in the previous month, the sixth straight month of yearly declines, according to the Bureau of Labor Statistics.

Falling gas prices were by far the largest contributor to the monthly decrease, falling 9.4% over the month, more than offsetting increases in shelter indexes, which rose 0.8% over the month and were 7.5% higher than a year ago.

US inflation peaked at 9.1% in June, its highest rate since 1982, as the war in Ukraine drove up energy costs and supply-chain issues in the wake of the coronavirus pandemic continued to push prices higher.

Despite the fall, the inflation rate remains more than three times as high as the Fed’s annual target rate of 2%, and is expected to remain elevated through 2023.

The professionals at the Federal Reserve know all this of course and likely take both CPI rates and other factors into account but the general public may not be as aware and giving only the year-over-year figure may make them fearful that inflation is still high when it might have stopped or even declined.


  1. xohjoh2n says

    Note that today was BLS CPI figures. The Fed pay more attention to PCE CPI figures which are out on the 27th, in time for their next meeting on the 1st Feb.

    (Also the figures came in at about the expected levels, but are still far above the target 2%. That they appear to be improving increases the odds that the Fed will keep to a 0.25 pt rate increase on the 1st, rather than the more aggressive 0.5 pt had the numbers been worse, but you’re still a ways away from no rate increases or even a rate decrease.)

  2. Bruce says

    Yes. Apparently, we are supposed to support the Fed as they put the economy into recession to fight the inflation that has already gone away. It’s called (mob) leadership. The nice villagers with the pitchforks and torches will see you now. I guess?

  3. consciousness razor says

    It’s actually very normal for it to curve back down somewhat around the end of every year (roughly Oct., Nov., Dec.), despite the overall trend being upward of course, since prices do generally increase over the long run. That shouldn’t be surprising at all, as many businesses do run big sales in order to encourage a reduction in inventory before the end of the year, mainly for tax purposes. So maybe don’t take this as some kind of important sign that things are better, because it’s not very meaningful. And obviously, the super high inflation we experienced for quite a long time is baked in, so people are still continuing to suffer from that, even if (for the time being) they’re not currently turning the screws on us even more.

    Of course, sometimes it curves down a bit more at the end of the year, sometimes less or not at all. If you look at the monthly CPI-U data, it’s not as if December 2022 stands out as a big or sudden “drop” or anything of the sort.

    It’s more or less just doing the regular December thing that the CPI-U usually does, which is at least a welcome change of pace compared to the last couple of years. It had been consistently higher than average increases from about June 2020 all the way until about June 2022, when things started to get back to normal (for certain values of “normal”).

    At this point, since this is partly the issue being raised, I’ll note that I’m talking about the “average” data since the year 2000, since I don’t think it’s especially helpful to look back even more into the 1990s or earlier…. It would just mean there’s more calculating to do, and it would be about stuff that has little bearing on our lives these days. The economy in general is of course very different from what it was like way back when. So, I don’t think this is a bad choice or one that is particularly misleading. But it is a choice, and now you know which one I made when I decided to paint the very simple picture in the previous paragraph.

    I said that the month-over-month figure is a better indicator of current inflation and should be reported as well.

    But ideally you’d work with all the data, not just one hand-picked set of numbers that you happen to fancy, so it’s a little weird to say you only want one. With just data about “current” situation, you’re not getting a different view of the bigger picture, which also matters to at least some people, just like the current situation does.

    I mean, it’s not clear that you’re really grasping that all of these numbers are (1) accurate and (2) consistent with one another. So, it’s not as if any of them are “better indicators” of the relevant phenomena. They are just different indicators and have various uses in different contexts. How do you plan to use them and for what purpose?

    Also, it’s worth pointing out that with any data at all (or none), reporters and talking heads and politicians will engage in propaganda no matter what, so it’s not like we’ll somehow benefit in that way with a different set of numbers. The idea that monthly CPI data is a way to address that sort of systematic problem is at best silly.

    Maybe think about this a little differently. Suppose you took the last 20 years (or whatever … that was my choice) and got the CPI-U, looking at how it changes each month. You could simply see what’s going on with it, if you put it on a chart, which is probably the least misleading thing to do for the purposes of conveying this kind of information to the general public, since they don’t want a giant table of numbers like in the link above and will not spend the time to honestly figure out what it may or may not tell us about anything.

    No matter how you decide to report on smaller and more digestible bits of the data (whatever you may care most about), you’re tossing out some information and being selective about what you show people. But, the thought is, who are you to do something like that? If you want to avoid introducing those kinds of biases, you can give them all of it, but just represent it in a form that people can easily use, like a nice simple graph. That’s my suggestion. Then we don’t have to be all that picky about which things we choose to tell people, how that’s going to be phrased, how we want them to interpret that, etc., because it’s all plainly there in front of them.

    A headline containing a single number (or words like “bigger” or “smaller”) is also relatively easy to digest, but it won’t really be offering nearly as much useful/relevant information. That’s the sort of choice you make because you have a particular tool (such as your ability to write words in headlines) and make do with that tool. But it’s totally possible to just use a different set of tools.

Leave a Reply

Your email address will not be published. Required fields are marked *