Inflation past 4 percent

Note: Please continue to e-mail subscribe to Salil's Statistical Ideas.  While Google terminates some of these site functions in 2021, we will resume activities in some other format.  Also if you like, please purchase or review my books: Statistics Topics, or Colors and Numbers.

Modeling inflation utilizing modern econometric techniques is not without challenges.  As we have discussed at length, it is driven off of high volatility in a de-trended series.  For example, in the last century, headline inflation (on a 12-month basis) gyrated by +/-3% monthly!  Core inflation: by about +/-2½% monthly.  

Over the past couple decades, things have simmered down, with both headline and core indices gyrating only +/-0.1% monthly.  That was until right before the pandemic. 

But this is where things get interesting, as inflation collapsed for a brief period in 2020 Q2.  Prices fell as COVID-19 lockdowns slowed some sectors of business.  In this article we discuss to what degree prices have, or are recovering, and why we are forecasting a sharply elevated 3½%-4% inflation going through 2022.  This is also nearly 2σ (standard errors) higher than the 2.1% inflation we had in 2019.  

Inflation that is between transitory, and cyclical.  And the level of inflation that rallies the public into action.  

Background
In passage of the economic stimuli (similar to the $1.9 trillion American Recovery Act) there was concern it would lead to straight overheating of the economy.  Resulting is immediately escalating inflation. Such inflation, well over 2%, would cause the the Federal Reserve to panic and chaotically attempt to bring prices under control.  

Arguments by Chair Powell and Secretary Yellen meandered from any such inflation wouldn’t happen  Or if it happened it would be transitory (i.e., not last more than a some months).  They would later clarify again that if inflation occurred, they had tools to manage it.  More recently, they have conceded there might be inflationary times ahead but that it is still a plus for society.

We have polls suggesting for example that 4% over 9 months would be considered a bad inflation outcome.  An outcome beyond what people are comfortable with.  And not simply 4% for a few quarters, but a contour of other dour combinations of severity and duration.

Recent data

Recent pricing data has been high, without doubt.  On a headline basis we have had over 0.7% monthly inflation.  For three straight months!  Levels not seen since 2013, and even then it was just for a month.  Before that it was in 2012, when I was worried about inflation while working at the U.S. Treasury.  Then too, just for a month.

Core inflation, a tighter measure of prices that strips away noisier components of food and energy.  Core inflation is surprisingly more volatile on a multi-year view.  Core inflation steadily runs higher than headline inflation on a multi-year basis, when headline inflation runs high.  Conversely runs lowers, when headline inflation runs low.

Core inflation was been been coming in at a similar 0.7% reading for the past two months.  A first, since 1992 when we saw a lengthier streak of 0.7%-0.8% monthly inflation. 

 

So the recent data could look concerning, but only if you continue to extrapolate it well into the future.

And how rare are the combined core and headline readings?

There are multiple considerations about the severity of the last month’s inflation data.  For example, there have been 767 months since 1958.

In 48 of those months we had higher (or equal) joint headline and core 1-month % change:

And in 148 of those months had higher (or equal) joint headline and core 12-month change:

We can also think about the correlation of these tail risks, though roughly 48 of 767 is only the extreme 6% or so.  Not the traditional extreme 1% which we think of when we think of "extreme tails".

We would need just a 1.0% 1-month inflation, in headline or core, to be in the extreme 46 months since 1958.  Or need 1.0% 1-month inflation, in headline and core, to be in the extreme 14 months since 1958.  Recall the May inflation report we had was not far from these levels on both headline and core.  And we still have a long way to go this year.

But bear in mind that most times we see this 1.0% 1-month inflation, it's in the context of 9% 12-month inflation.  We simply don't see these levels often.  Usually one of the readings (i.e., headline, or core) comes a little but at the expense of the other.  And 9% inflation readings are never part of transitory events.

We have 1.0% 1-mo inflation, in Headline OR Core, in 46 months [since 1958].

We have 1.0% 1-mo inflation, in Headline AND Core, in 14 months [since 1958].

Most times we see 1.0% 1-mo inflation, it's in the context of 9% 12-mo inflation; & not transitory. Are we aiming for 1.0%?

— Salil Mehta (@salilstatistics) June 10, 2021

We are operating in increasingly rare space.  When will the current inflation spike end, and how?  

Our forecast
We don’t see either 9% inflation.  Nor do we see 2½% inflation.  Both take large assumptions that seem contrary to the underlying forces and statistical aberrations we operate in.  I instead forecast troubling price growth in the 3½%-4% range.

We start by looking more into the noisiest statistical part of this recent inflation data.  And we go back to where we said this inflation data fell off the tracks to begin with.  2020 Q2; and the suppressed business activity then.  The pricing data across different measures shows the most impacted months center around April and May.  A little of March, and a little of June...  three months in total.  

And we have already shown a recovery not only of prices from the pre-COVID-19 period, but the rate of inflation is simply higher now versus it was in this year “prior to coronavirus” (i.e., COVID-1 on the chart).

Prices are indeed growing fast right now, and on trend to finish the year well above 3½%, as this transitory window moves further into the background, but we still have the shadows of non-transitory elements lingering on for months to ahead.

The long view

We also anticipate a long-view of what is the driving catalyst for the current inflation.  Business owners stating customers having excess cash at this time, with an interest to spend.  This carries us, non-trivially, into 2022.  Doesn't the Federal Reserve and the Biden Administration have this at the heart of their design?  Lingering inflation that appears inevitable for a short while, beyond the dwindling effects from the transitory months we are now phasing out of.  

There is no reason to assume extreme prices beyond 2022.  It's not in the data right now; nor do we know how the economy will shape up in 2023 and beyond.  Even as 2020 was starting, most were not predicting a recession.  Same in 2008.  Same in 2000.  Hopefully the point's been made.  

  

Now 3½%-4% inflation for 2022 is already an upside stretch as noted at the start of this article.  It's lower than some double-digit inflation numbers some prominent academics have forecasted.  But it's inflation that would anger the public (as well as the financial markets) as it did during similar times in the past.  

It would also catch policy makers inexplicably off-guard (political economists in Washington showing inflation expected to be just a tad over 2% for years to come).  Inflation that would have to lead to an implicit concession of responsibility.  

What do you think the 10-year Treasury rates would look like in this scenario?  More like 2% where we are likely going to be this year, or perhaps double that?