There is a general understanding of what happened last week. The 2.9% rise in average hourly earnings in the US reported, the fastest since 2009 spurred fears of rising inflation. The jump in US interest rates triggered equity sales and a spike in volatility, which in turn spurred the unwinding of low vol bets that had been paying off handsomely.
While this consensus narrative has much to recommend itself, there is a major discrepancy. The rise in US interest rates since both the end last year and since the January employment data appear to reflect mostly an increase in real rates and not the inflation premium.
This preliminary assessment is borne out by reviewing the change in nominal yields and the change in market-based measures of inflation expectations. The US nominal 10-year yield finished at 2.79% the day before the latest employment data were released. The yield finished last year a little above 2.40%. The high yield print has been almost 2.90%–a 50 bp increase from the end of last year and a 11 bp since before the employment data.
There are two readily available market-based measures of inflation expectations. There is the 10-year breakeven, which is the difference between the 10-year conventional yield and the 10-year inflation protected security (TIPS). The 10-year breakeven was at 1.98% at the end of last year and 2.12% before jobs report. It reached 2.14% last week, and is now near 2.08%.
Another market-based measure of inflation expectations is the five-year five-year forward breakeven, which is similar to the 10-year breakeven, but uses a five-year forward rate of the conventional and inflation-linked securities. A similar picture arises. The five-year five-year forward was at 1.99% at the end of 2017. It was at 2.22% before the jobs data and now is near 2.21%.
Most of the rise in nominal yields took place before the average hourly earnings surged. However, the increase in these two market-based measures of inflation expectations do not explain the bulk of the increase. Both measures of inflation expectations are now actually unchanged or lower than prevailed the day before the data were reported while nominal yields are higher.
There are other reasons to doubt that inflation expectations became unanchored by the average hourly earnings data. First, nearly as soon as the data came out and economists had time to look at the details, it appeared to be exaggerated. The increase may reflect the increase pay for supervisory workers, for example. Second, the increase was in line with the Q3 average (0.3% a month) after a soft (0.2%) average in Q4. Third, other measures of labor costs, like the employment cost index, did not bear out the acceleration. Fourth, and importantly, the relationship between average hourly earnings and the core PCE, which the Fed targets, is weak at best, which, in part, speaks to the frustration policymakers are experiencing over the Philips Curve that links unemployment and inflation.
The US reports January CPI on Wednesday. The consensus narrative says that the CPI is the key report this week, but if our analysis is right and it has been an increase in real rates rather than the inflation premium, but 10year yield may not ease as much as one might expect if the headline and core rates slip economists expect. The Bloomberg survey found a median forecast for the headline pace to ease to 1.9% from 2.1% and the core rate to slip to 1.7% from 1.8%.
Another way to gauge inflation expectations is simply to ask people. The market-based measures may be compromised by liquidity premium, for example. Surveys have their own methodological challenges to be sure, like the channel (internet, landline, mobile) and gender bias (women reportedly typically see more inflation than men when controlled for income and education). The University of Michigan reports its preliminary February figures at the end of the week. In January 5-10-year inflation expectation measure ticked up to 2.5% matching the highest level since August 2017. It is also the average for 2017 and 2016.
The Federal Reserve conducts a quarterly survey of professional forecasters. The results of the Q1 survey were reported last week. Essentially, the professionals revised up Q1 18 forecasts but left the long-term forecasts largely unchanged. Specifically, the Q1 headline CPI a revised to 2.7% from 2.1%, but this largely reflects the past increase in oil prices. The expectation for the core rate was tweaked to 2.2% from 2.0%, and the core PCE deflator expectation was increased to 1.8% from 1.7%. At the end of the year, headline CPI is expected to be a 2.2% instead of 2.1%, while the core PCE deflator expectation was not changed from 1.9%.
Rather than inflation expectations being the critical driver lifting nominal rates, we suggest an alternative hypothesis, which we think matches the facts better. The increase in nominal rates reflects a small increase of inflation expectations but a more significant rise in real rates. Real rates are rising in the face of changing supply and demand considerations. Specifically, one of the largest buyers of US Treasuries, the Federal Reserve has pre-announced it will buy $420 bln less than in 2017. At the same time, the combination of the tax cuts (loss of revenue) and spending increases ($300 bln over the next two years) means that issuance is going to rise sharply.
Even for those who advocate fiscal support during economic downturns, like in 2008-2009, and who rarely see deficit reduction as a key policy goal, there is a strong distaste for purposely running 4%-5% deficit when the economy is growing above trend. The proverbial chickens come to roost, in such a scenario, with the end of the business cycle. The deficit will swell more during slower economic growth and the base is already high. Higher interest rates may be necessary to attract global savings at a moment in time where the economy may not justify higher rates.
In some corners, there are already concerns about the re-emergence of America’s twin deficits. The fiscal deficit gooses the economy, which has a high propensity to consume imports. The non-energy trade deficit is deteriorating, and the current account deficit is set to widen. We note, however, the corporate repatriation of earnings kept offshore, induced by the tax changes, may show up in the investment income line item of the current account. This would give the appearance of reducing the current account deficit–for a short period of time.
We suspect that with the ECB and BOJ still buying all the new net issuance by their respective governments, and the US the only net provided of net new core paper, the demand for US Treasuries may hold up well this year. That is to say that if 50 bp increase in the 10-year yield in the first six weeks of 2018 means the bulk of the near-term move is likely already in place. Price pressures may become more evident in Q2, when last year’s anomalies drop out of the year-over-year comparison.