Declines in energy prices pushed headline inflation sharply lower during the fourth quarter. But core inflation has also been uninspiring, with 12-month core CPI growing at only 2.2% in December – higher than in 2017, but roughly the same as in 2016. The puzzle is that a tightening economy and labor market have failed to meaningfully lift underlying inflation, compared to prior economic expansions.
Exhibit 1: Changes in US CPI inflation and US core CPI
As of 25/01/2019
Headline inflation, of course, is impacted by a number of forces – some cyclical and some unrelated to the economic cycle. These forces include:
- Fluctuations in wholesale energy prices;
- Output gap consideration that can impact domestically generated services inflation;
- Trends in rents and owners’ equivalent rent (OER);
- Currency strength, taxes and tariffs that impact prices of tradeable goods;
- Regulated or administered prices; and
- Methodological changes to the measurement of inflation.
We discuss these below.
The reader will be aware, headline US CPI reached 2.7% year-over-year in July, and has since been on the decline because of base effects, as well as the recent fall in energy prices. As of the December print (released in January), year-over-year CPI stood at 1.9%. The decline in energy prices is set to reduce annual headline inflation to perhaps barely 1.0% by summer 2019.
WTI crude oil futures prices began to slide from around USD 75 per barrel in October to around USD 50 in January as
- the Trump administration handed out more waivers to buyers of Iranian oil than had been anticipated,
- speculation mounted that the Saudi / Russia supply agreement might unravel, and
- the IEA lowered global demand forecasts.
Cyclical / output gap factors
With the economy having expanded at a rate faster than potential in 2018, the labor market has continued to generate payroll gains at a pace sufficient to reduce excess slack in the labor force and drive up wages. This reduction in labor market slack is driving up labor compensation rates. As the December labor market report showed, average hourly earnings rose by 3.2% year-over-year, matching the pace of growth in the employment cost index. Labor compensation, of course, is a key cost component to many services industries, where inflation tends to be cyclical. However, while wages are firming, inflation rates in cyclical industries have actually disappointed in recent months.
To identify the sources of recent inflation weakness, and to provide a framework for thinking about the outlook, Steve Friedman of our Macro Research Team analysed the PCE price index by dividing its constituents into one of two categories:
- cyclically sensitive prices (which respond to economic slack), and
- a-cyclical prices (which tend to be driven by sector-specific factors and hence have little correlation with the economic cycle).
This technique is based on the methodology of Stock & Watson and, to a lesser degree, the Federal Reserve Bank of San Francisco. Importantly, the cyclically sensitive or “pro-cyclical” inflation categories are all services – housing, food services and accommodation, recreation, non-profits serving households, and other services, which include communications, education, professional, personal care, housing utilities ex. energy, and social services. These categories compromise just 44% of core PCE inflation. In other words, a bit over half of core inflation is driven by factors other than the economic cycle.
With inflation divided into these two categories, we can see that since 2012 pro-cyclical inflation has been rising – suggesting that the Phillips Curve isn’t flat, one just needs to know where to look for it. And despite recent softness, pro-cyclical inflation has risen this year, from 2.4% at end of 2017 to 2.7% currently, that it is contributing 1.2 percentage points to core inflation.
But the firming since 2017 is not particularly impressive – it largely represents base effects, namely, the roll-off of the early 2017 weakness in cell phone plan prices (within “other services”). And while housing services continued to make a strong contribution to both PCE and CPI inflation in 2018, it did not pick up further. In sum, even if pro-cyclical inflation has firmed as the expansion has continued, recent readings have been somewhat disappointing, especially given signs of little resource slack.
Shelter costs, meanwhile, have remained a steady contributor to CPI inflation, rising by 3.2% year-over-year. Rents, and thus OER, have paradoxically been supported by the falling affordability of owner-occupied housing (which forces some households to rent rather than buy).
Meanwhile, a solid labor market has maintained households’ demand and their ability to pay for rental units, even as supply has picked up, keeping vacancy rates low. Rent remains a major contributor to inflation, but it is nevertheless hard to see rental inflation picking up further.
US dollar and administered prices
A-cyclical price components, meanwhile, have contributed just 60bp to core PCE inflation over the past year. The cause of weak a-cyclical inflation is two-fold:
- persistent goods deflation, which has been aggravated by dollar strength and softer growth abroad, and
- a recent step-down in services inflation driven by two categories: healthcare and financial services.
Some have speculated that higher tariffs, and the associated substitution effects, could help support goods prices (if the trade dispute with China escalates and there is limited exchange rate offset). But with global growth moderating, lower commodities prices, ongoing technological advances and the renminbi (CNY) responding to US tariffs, we would not be overly optimistic about goods inflation.
On the healthcare front, however, there are grounds for optimism. This may sound surprising, given that healthcare is just as susceptible as other sectors to the deflationary impact of technological advances. But this observation misses the point that a large portion of medical services is subject to administered prices by the US government, and these administered prices also exert a strong pull on private medical service prices.
For 2019, the government has larger price increases planned for Medicare and Medicaid compared to 2017, suggesting a further step-up in healthcare services inflation. This is extremely important given the almost 20% weight of healthcare services in core PCE inflation. For 2019, PCE healthcare services prices should rise by over 2%, continuing the firming trend of recent years.
Source: BLS. PPI is the source data for the BEA’s health care services price indices
We maintain our view that core CPI and PCE inflation should gently firm in 2019, with the contribution from shelter being supplemented by a growing contribution from cyclically-driven services inflation. As can be seen by examining pro-cyclical inflation relative to labor market slack, the unemployment gap – the difference between the unemployment rate and the CBO’s NAIRU estimate – has moved into negative territory over the past two years.
As would be expected, pro-cyclical inflation has firmed over this period. But the firming in cyclical inflation components has been weak relatively to the long-term trend, as measured from 2002 to 2015. After falling in 2017, pro-cyclical inflation has picked up in 2018, suggesting that tighter levels of resource utilization should continue to drive up many services price components of core inflation.
The inflation environment is not without downside risks. The US dollar and oil prices have already been mentioned. An additional and arguably more important structural headwind is technology.
Recall that in 2017, Janet Yellen attributed the shortfall of inflation from the 2% objective to ad hoc/transitory factors such as the sharp decline in cell phone services prices. Another way to think about those price declines is through the lens of technological advancement – new technologies allowed cell phone service providers to offer unlimited data plans without raising prices, leading the BEA to incorporate this quality improvement into their inflation calculation for cellular services.
We will continue to see such quality adjustments to price indices that relate to technological advancement. In fact, a near-term risk is the BEA’s planned January 2019 quality adjustment to residential telecommunications services, which includes land lines, internet and cable television services. The details of the methodological changes have not been released, but the direction seems clear given the BEA’s reference to “the rapid technological changes in these services.”