The US recession that isn’t: the contrarian take on 2023
Plus, US January non-farm payrolls
Recession calls are stock and trade for macro punditry. But the traditional year-ahead outlooks that accompanied the turn of the year were remarkable this year for making a 2023 US recession close to a consensus view. That in turn underpins expectations for a renewed Fed easing cycle, as well as heavily conditioning the global outlook.
Such a consensus naturally invites contrarian scepticism. Recessions are for the most part not forecastable because if they were, economic actors, not least policymakers, would take action to avert or at least mitigate the downturn. Implicitly, a recession forecast is a forecast of a central bank policy error. Pandemic and other potential real shocks notwithstanding, recessions are always and everywhere caused by monetary policy. Even then, monetary policy can give a real shock a worse starting point, as was the case with Australia going into the pandemic.
Rather than a policy error, the consensus seems to view a recession as a necessary element of the Fed achieving its inflation goals. But as supply-siders have always argued, strong economic growth can be perfectly consistent with moderating inflation, so long as the supply-side can keep pace with demand.
The US is at least better placed than other economies, like the UK and NZ, where a downturn is the central bank’s official forecast. The BoE and RBNZ seem to have adopted the view that a recession is necessary to contain inflation, although the official forecasts are also partly a reflection of market pricing. As Ben Southwood argues in the case of the BoE, this is the monetary policy strategy equivalent of controlled flight into terrain. The Fed is still at least notionally committed to avoiding a recession.
The strongest evidence for a near-term recession in the US is the inverted term structure. As Menzie Chin notes, simple probit models based on the term structure imply a recession in January 2024 is anything from an even bet to a near certainty, which is uncomfortably high either way. Of course, these market implied expectations are not independent of the consensus view described above and so not necessarily confirmatory. But the predictive power of the yield curve is largely attributable to its ability to neatly summarise the effective (as opposed to notional) stance of monetary policy. If recessions are always a monetary phenomenon and the yield curve summarises the stance of monetary policy, then we should listen-up.
The current juncture is one in which positive macro data is perversely taken to confirm the necessity of a monetary policy-induced downturn, even as the data itself defies those expectations. The recession narrative can’t lose.
Last week, the advanced estimate for US Q4 GDP came in at an annual rate of 2.9%, after increasing 3.2% in the third quarter. Growth has moderated sharply from the post-pandemic rebound, but that’s all for the good. Nominal GDP growth moderated from 7.7% in Q3 to 6.5% in Q4. This still left the Mercatus NGDP gap at a record (for the period since 1997) 5.9% compared to 5.7% in Q3. On this measure, the US still has an excess demand problem, at least in expectation. The NGDP gap is not expected to close until Q2 2026,
The Berger, Morely and Wong real output gap came in at 4.6%, with a Q1 2023 nowcast of 3.5%, with the output gap not seen turning negative until Q2 2024 (based on their conditional forecast). Again, this is a larger deviation in output from potential than any seen since the late 1960s, assuming their modified Beveridge–Nelson decomposition based on a mixed-frequency Bayesian VAR is a good description of the data generating process for US potential output.
Yet the continued moderation in US inflation to target-consistent rates suggests the supply-side is stepping-up to meet moderating demand. There is nothing inevitable about a US recession and I would certainly not hang my asset allocation hat on what is now a very shop-worn view (the Commonwealth’s macro-driven prop trading operation, for example, is fully bought-in). The no-recession contrarian trade looks more attractive.
Central bank gold buying and US dollar dominance
The World Gold Council reported the strongest central bank gold buying since 1967, which prompted the usual suspects to argue this was somehow a knock to US dollar dominance. But to state the obvious, central banks were not buying other currencies, but what is effectively a US dollar substitute in terms of having a liquid global market, but without the US stranglehold over payments and settlement infrastructure. Attempts to circumvent US dollar dominance only underscore rather than undermine that dominance.
US January non-farm payrolls
Our payrolls forecast for December did remarkably well despite (or perhaps because!) we subbed in estimate values for some model inputs. Sometimes modelled data actually works. The Chicago Fed National Activity Index is back-up, so we are back to BAU on those inputs. The model inputs are up across the board for January, which has us once again looking for an above-market outcome of 235k for non-farm payrolls (consensus is 185k). This would be the first acceleration in payrolls growth since July last year. We expect the unemployment rate to be steady at 3.5% (market is expecting an uptick to 3.6%). Needless to say, this will be seen by markets as raising the bar on the Fed’s remaining tightening task, especially in the wake of Fed Chair Powell’s comments that the labour market is ‘out of balance.’
A Brookings write-up of New Zealand’s housing reforms. It is odd that New Zealand’s reforms are getting more attention in the US than in Australia.
Twitter is becoming a ghost town. My experience is similar.
Where do your fellow IE readers come from? 55 countries and 30 US states. Out of interest, if you are reading from a country or US state that is missing from these maps, drop me a message in the comments below. It would be an interesting test of the accuracy of Substack’s metrics.
Memes and themes: