Hoping For Boring Inflation

I have been racking up a lot of consulting hours recently, so my writing has been on a back burner. I have also been looking at my two manuscripts (the inflation book, and the banking primer), without having new content to publish here.

I have decided that I am not going to add new content to my banking primer. I need to consolidate all my “banking theory” pieces into one or two chapters, and that would make the book at my target length (novella, not novel length). Rather than tackle that now, I have turned back to my inflation manuscript. The advantage of a being a long time separated from the text is that I can read it as a “new” text. This makes it much easier to edit — I have to re-read what I wrote, and mangled text stands out much better in text that I have been recently rewriting. It also made me happier with the manuscript when compared to when I was last heavily revising it.

The chapter that probably still needs the most attention is the last one — the discussion of the post-pandemic inflation kerfuffle. Until that chapter, the text reflects my attitude towards inflation that I had when I was working in finance: developed country inflation post-1990 is generally boring. Although I am not attempting to do a forecasting exercise in a book, it could easily be somewhat embarrassing to publish a book whose theme is “inflation is boring, actually” when the economy is in the process of entering some secular inflationary maelstrom.

def sooper_sekrit_inflation_forecasting_model(**kwargs):
    """
    Proprietary inflation forecasting model.

    So proprietary that the function arguments are themselves proprietary.
    :param kwargs:
    :return:
    """
    return 2.0

The above code snippet is a variant of one of my standing jokes: my proprietary one year ahead inflation forecasting model that just return 2%. The ugly reality is that this model outperformed most of the inflation forecasting models I had the misfortune of having to look at (so long as the look-ahead period is 12 months or more, so that a model cannot take advantage of known information when computing the 12-month rate-of-change).

The pandemic blew up that forecasting model, albeit there are not a great number of independent 12-month intervals after 2020. If we look back at the chart at the top of this article, headline inflation is back at fairly “normal” levels for the post-1990 period. Median inflation remains elevated (where it was in 1990, which was before inflation stabilised near the 2-2.5% level (depending on which inflation measure you use, CPI tended to be higher than PCE).

I have not yet started digging through charts of updated data, but my educated guess is that the housing component of CPI “explains” the stickiness of median CPI. That is, the inflation story is now largely one about housing (which is directly measured via rents, but house prices presumably influence rents).

The difficulty with dealing with housing inflation is that the internet is filled with people who have very strong opinions about the subject (to the point of it defining their online personality). I have reached the stage in life where I am not going to pursue the details of what Californians are doing to themselves in housing regulation. All I can offer is the insight is that housing inflation is high, and it may or may not come down over the coming decade.

But if we make the possibly incorrect assumption that rents cannot outstrip wages for a decade, inflation movements returns to being a business cycle story. The non-housing part of the economy seems to be closer to the 1990-2020 norm, although the spectre of a structurally tighter labour market (courtesy of demographic trends) remains to keep the inflation worriers scared this Halloween season. Although I was historically an inflation dove, I did see the story about a ageing population raising the dependency ratio being a plausible driver for changing the long-term inflation trend. (Admittedly, one can go look at all the failed “Japan is heading towards hyperinflation!” stories to see that demographics is not necessarily destiny with respect to inflation forecasting.)

The problem with the “structurally tight labour markets will lead to higher inflation” story is that business cycles move faster than structural trends. We got a big inflation pop in 2020 that can be tied to tighter labour markets — but that pop required a lot of short-term shifts all pointing in the same (higher inflation) direction. Outside the panicked pandemic reactions, governments and firms have time to react to changing labour market conditions. Modern management has gotten extremely aggressive in wasting workers’ time in practices like gig work, and so there is a lot of malinvestments to be culled in response to a relative price shift in workers’ favour. And policymakers will react to inflation target misses (with the primacy of monetary or fiscal policy left to the readers’ preferences), so policy should be expected to lean against inflation overshoots. (Inflation was left below target in the 2010s, but nobody really cared, and the attempts to use monetary policy to nudge inflation upwards was just a classic string pushing exercise.)

Admittedly, my benign inflationary outlook does face an immediate risk — in the United States at least. Although I have done my best to avoid discussing the upcoming American Presidential election (out of deference to my cardiovascular system), it is clear that some of the err, wackier, economic proposals that have been floated by one major candidate might disrupt the economy. Obviously, I will have time to react to this before the manuscript is finally published.

The other angle to the text I need to revisit is the “transitory debate.” Although I might just grab some articles that fit my priors, I think the debate can summarised as possible.

• If we define “transitory” as inflation immediately returning to “normal” levels, it was not transitory. However, this is certainly not how I interpreted “transitory.”

• The definition I leaned towards was that inflation would return to “normal” levels without requiring a recession or much higher inflation. Although this definition seems weak, it is in opposition to the possibility of entering a phase of structurally rising inflation (like the 1970s).

• One camp of people argue that “inflation was not transitory” because central banks raised interest rates to reduce it. However, this requires the belief that interest rates control inflation, which is not easily falsified. This episode can be explained by central bankers following their historical practice of setting the policy rate at a lag to inflation trends. Such behaviour will always results in the policy rate peaking just before transitory inflation spikes reversing. Since I am not an interest rate believer, this is not a definition I would use.

I should have more time to devote to writing. Although I should be focussing on getting my manuscripts out the door, I might be putting up some updated inflation charts as part of a post-mortem on the pandemic inflation spike.

Email subscription: Go to https://bondeconomics.substack.com/ 

(c) Brian Romanchuk 2024