Who forecast inflation correctly? Bringing money back into business cycle analysis

Plus, US January non-farm payrolls and Canada’s real return bonds

Feb 8, 2023

Very few central banks, professional forecasters or macro pundits picked the global acceleration of inflation in 2021 and its persistence into 2022. This has prompted a lot of hand-wringing and claims that we do not understand the inflation process.

Yet as Joshua Hendrickson shows in a recent paper for the Mercatus Center, a simple model of inflation in terms of Divisia M2 growth does a good job of forecasting the acceleration in the inflation rate in the United States, although underpredicts the actual rate. The point is not that this is the best model of inflation, but the claim that we do not understand the inflation process and that inflation is not forecastable doesn’t bear much scrutiny. We have previously pointed to the real-time inflation predictions of monetarists like Scott Grannis and Tim Congdon, who anticipated higher inflation rates largely on the back of observed growth in M2 and trends in money demand.

It could be that M2 is just a proxy for other variables. As we have noted here previously, monetary aggregates are both supply and demand-determined and so we should be wary of identifying money aggregates with the stance of monetary policy. But the point about forecastability still stands, even if you view M2 as endogenous and a proxy for other variables.

As Hendrickson notes, money has essentially been banished from the workhorse New Keynesian model that informs monetary policy, even though that model can be equivalently written down in quantity theoretic terms by writing the interest rates equation in terms of monetary aggregates and including de-trended velocity in the inflation equation alongside an output/unemployment gap. The difference is that whereas the standard New Keynesian model becomes unstable at the zero lower bound, the quantity theoretic re-specification is robust to that issue.

David Laidler, in his prescient 2004 critique of Michael Woodford’s neo-Wicksellian opus Interest and Prices: Foundations of a Theory of Monetary Policy, said that:

If they allow the elegance and rigour of his exposition to distract their attention from the narrow scope of his theory, and under its influence begin to lose sight of the neoclassical theory of money that their predecessors, including Wicksell, created, and of the broader set of issues with which monetary policy sometimes has to deal, Woodford’s work will be a source of trouble.

Scott Sumner subsequently showed exactly how much trouble the New Keynesian framework exemplified by Woodford would cause in the wake of the 2008 financial crisis.

Money is arguably making a comeback, however. Carl Walsh recently took up blogging (blogging is also making a comeback) and linked to a recent paper of his with Roberto Billi and Ulf Söderström on the role of money at the zero bound in a New Keynesian model. Carl Walsh was one of my suggestions for the RBA review panel. He participated in a panel on monetary-fiscal policy interactions with me, the ECB’s Jacopo Cimadomo, Jan Libich and former RBNZ Governor Don Brash on the sidelines of the 2010 Australian Conference of Economists:

L-R: Carl Walsh, David Gruen, Don Brash, Jan Libich, Stephen Kirchner, and Jacopo Cimadomo.

The immediate relevance of all this is that M2 in the US is showing some of the weakest growth since the Great Depression. The unofficial Divisia M2 measure is showing the weakest annual growth since the depths of the Volcker disinflation in 1981 (the Divisia aggregates go back to the late 1960s). This is consistent with the inflation process having peaked. While markets may doubt Powell’s resolve, the monetary aggregates at least are consistent with a Volcker-like determination. Also worth recalling that the Volcker disinflation laid the foundations for a bull market in equities starting in 1982. That is not to ignore the recession of 1981-82. But this time really is different, with long-term inflation expectations remaining well-anchored.

US January non-farm payrolls

We were very comfortable with our above market forecast for US January non-farm payrolls. In the event, we saw a 517k gain in employment relative to our expectation of 235k and market median of 185k. The unemployment rate fell to 3.4%, its lowest since May 1969, relative to our expectation of 3.5% and the market median’s 3.6%. For anyone under 54 years of age, the unemployment rate in the US has never been lower.

ʎllǝuuop ʇuǝɹq @donnelly_brent

Here's Joe Biden the last time the US Unemployment Rate was 3.4%

There is a lot going on methodologically with both the household and establishment surveys in the month of January, which strictly speaking makes the January 2023 and December 2022 unemployment rates non-comparable, at least for now. For a deeper dive into some of the methodological issues, see Joey Politano’s newsletter. But that does not change the fundamental picture of a highly resilient US labour market against the backdrop of moderating inflation.

One of my regular criticisms of pre-pandemic monetary policy was the extent to which the Fed and RBA would hang their policy hat on labour market indicators as predictors of inflation. Markets also bought into that view. We are still doing it. The Phillips Curve cannot fail, it can only be failed.

Canada’s real return bonds (RRBs)

Canada’s finance minister has announced that Canada will discontinue issuance of real return bonds (RRBs), ostensibly because of a lack of liquidity, which can also be attributed to a lack of demand. Of course, the conspiracy theorists will see this as a prelude to inflating away the debt.

This is an unfortunate development in that it deprives us of a market-based implied inflation forecast. That forecast is less reliable to the extent that the market is suffering from liquidity issues, but that is an argument for building liquidity, not scrapping the market.

It is fair to say that inflation-indexed bonds globally struggled in the low inflation environment pre-pandemic. In the recent high inflation environment, investors have still taken a hit in terms of higher real yields. This might explain the lack of demand and hence liquidity.

Of greater concern is that policymakers seemingly do not value the information about future inflation contained in inflation-linked securities. In the wake of the forecasting failures in relation to the global surge in inflation, you would think policymakers would be looking to expand and deepen market-based insights into future inflation, even with all the limitations associated with inflation-linked bonds.

ICYMI

The lights wink out in Asia.
Building More Housing Makes It Cheaper. Really.
Supply vs Quantity Supplied: Housing Edition.
BoJ Deputy Governor Masayoshi Amamiya has been approached to take over from Governor Kuroda, according to the Nikkei, which would provide an element of continuity in Japanese monetary policy given his key role in giving effect to Kuroda’s monetary strategy. The Cabinet Office denies the story.
Yes, Noah Smith still hates Australia: