By Atul Bhatia, CFA
“This time it’s different” has been an expensive phrase for investors, whether it’s switching to a price-to-click valuation model in the early days of the internet or ignoring Asian currency imbalances in the late 1990s.
At the same time, investors and policymakers need to be aware when underlying assumptions are no longer valid; one of the leading contributors to the global financial crisis was a failure to recognize that changing mortgage lending standards had altered the nature of default risk and potential losses in the housing sector.
While the stakes today do not begin to approach those of the late 2000s, we believe there is a similar error in how policymakers are talking about inflation, labour market tightness, and the appropriate policy response.
Take this week’s release of the January Consumer Price Index (CPI) inflation data. Most of the reported price increase came from shelter costs, a segment that inherently lags economic reality since it only recognizes rental inflation when a new lease is signed. Federal Reserve Chair Jerome Powell has previously acknowledged the deficiency and emphasized the importance of CPI core services less housing, a measure that rose a relatively sluggish 0.27 percent in January. Even with the artificially elevated shelter costs, January marked the seventh consecutive month of slowing annual inflation.
Fed officials, however, are choosing to ignore these positive inflation indicators – and the multiple signs of a broadly slowing economy – and are instead emphasizing what they see as signs of an overheated labour market to justify more restrictive policy. We believe that, properly understood, these signals are less inflationary than policymakers appear to think, even though institutional biases leave the central bank ill-suited to recognize the opportunity.
Key inflation component and wage growth slow
Federal Reserve effectively silent on its own preferred metric
The line chart shows the change in the CPI core services less housing, the Conference Board’s U.S. Leading Economic Index, and the Employment Cost Index from 2017 through Dec. 31, 2022. All three measures have declined from 2022 peaks and the first two measures are now at or below pre-pandemic readings. The Employment Cost Index is above pre-pandemic levels, but is at 1% at the end of 2022 vs 0.7% at the end of 2019.
- Conference Board U.S. Leading Index (LHS)
- Employment Cost Index (RHS)
- CPI core services less housing (RHS)
Source - RBC Wealth Management, Bloomberg; quarterly data through December 2022, plus January 2023 CPI less housing
The most recent nonfarm payroll report showed over 500,000 net jobs created in January 2023, above even the highest projection in Bloomberg’s pre-release survey. Part of the reported growth is likely a statistical aberration related to seasonal adjustments, and we would expect that to reverse in upcoming months. But the payroll count – which is derived from employer data – was reinforced by the household survey, which showed a growing labour force having no difficulty finding jobs. We feel comfortable saying that the number of employees grew meaningfully in January, even if the degree of growth was likely overstated.
What is open to question is what is driving this employment: is it an inflationary pull by employers paying up for additional workers, or is it a less inflationary supply push as sidelined workers – stressed by rising consumer prices – return to the labour force? We believe both factors are at play, but, on balance, the less inflationary supply push is what likely drove much of the January hiring. This view is supported by increasing labour participation rates and rapidly rising household credit card debt.
It is also difficult to square highly inflationary hiring with other corporate behaviour. Bank senior loan officers are reporting falling commercial loan demand, and survey data shows declining confidence across business owners and senior managers; sentiment and behaviour that does not easily track with paying higher wages. Nor is the trend highly inflationary. The Bureau of Labor Statistics’ quarterly Employment Cost Index – which we believe is the best data the U.S. provides on employee compensation – showed quarterly wage gains slowing to one percent in Q4 2022. It is possible that employers switched to more aggressive pay packages in January, but it would be inconsistent with their stated views on growth and their borrowing behaviour.
Worrying more, borrowing less, and laying off
Credible backdrop for inflationary growth?
The line chart compares the Federal Reserve’s Senior Loan Officer Survey demand for commercial and industrial loans at medium and large firms with the job cuts announced in the Challenger report and the NFIB Small Business Optimism Index. It shows loan demand dropping to the lowest level since the start of the pandemic as announced job cuts have risen from 21,000 in March 2022 to over 100,000 in January 2023. The Small Business Optimism Index is at the lowest level since June 2022 (90.3 today vs. 89.5 in June).
- Growth in loan demand, large & med. firms (%, LHS)
- Announced job cuts (thousands, LHS)
- NFIB Small Business Optimism Index (RHS)
Source - RBC Wealth Management, Bloomberg; commercial and industrial loan demand based on Federal Reserve’s Senior Loan Officer Survey; quarterly data through December 2022 plus January 2023
JOLTing the data
While the recent nonfarm payrolls data painted an ambiguous picture, we believe the Job Openings and Labor Turnover Survey (JOLTS) presents a misleading one. Powell has called out the openings per available worker as a measure policymakers monitor closely. The problem, in our view, is that the data is almost certainly wrong.
The issue stems from the survey’s construction. Designed in the 1990s, the Bureau of Labor Statistics applies a three-part definition of a “job opening,” of which the main constraint is that an employer must be taking active steps to hire. In the last century, this typically meant spending money on newspaper ads. Today, a basic internet job listing is free. Unsurprisingly, this makes it much easier for a business to report an opening; while hard numbers are scarce, one study found a 40 percent drop in job postings after a website introduced a listing fee. A 40 percent reduction in openings would bring the Fed’s measure down from a historically high 1.9 openings/worker to a more reasonable 1.15.
Adding to our concern is the interplay between job openings and wages. For at least two decades leading into the pandemic, year-over-year wage growth was relatively subdued; this gave employers a reasonable idea of what compensation would allow them to hire. In the pandemic era, that visibility decreased as prevailing wages rose quickly. As a result, at least some of the current job openings are likely real, but are only available at lower, less inflationary wages.
Bias to believe
What makes these potential errors particularly pernicious, in our view, is the institutional and personal incentives at the Fed. The bottom line is that central bankers can cause an unnecessary recession through overly tight monetary policy and emerge with their reputations – and their institution’s credibility – intact and even enhanced. The converse is not true. Loosening policy early and letting inflation reignite risks permanent reputational harm to the central bank and its leadership. In those circumstances, it’s unlikely that Powell and company will take a truly risk-neutral view of the incoming data and more likely that they will latch on to the most pessimistic interpretation to justify restrictive policy.
As painful as it may be to recognize, we think there is something different happening this time, and investors are forced to choose which path they believe: either U.S. hiring managers are aggressively adding workers despite slowing growth and a pessimistic outlook, or the reported levels of job creation are deceptively strong and less inflationary than prior history would lead us to believe.
We believe institutional blinders have closed the Fed to the latter possibility, raising the risks of overly tight policy.