Stock prices and inflation
admin | June 7, 2019
This document is intended for institutional investors and is not subject to all of the independence and disclosure standards applicable to debt research reports prepared for retail investors.
Fed Chairman Powell surprised most market participants at his May 1 post-FOMC press conference by downplaying early 2019 cooling in core inflation. The move through March well below 2% in year-over-year core PCE was largely “transitory,” he said. One of the factors that Powell cited was portfolio management services. The cost of these services is closely correlated to trailing movement in stock prices. The swoon in stock valuations in late-2018 contributed heavily to slower inflation around the turn of the year and the subsequent rebound has already helped tip core inflation up in recent months, validating the Chairman’s contention. It’s a case study in following the inflation numbers.
Financial services prices
One of numerous key differences between the CPI and PCE deflator is the treatment of financial services. The CPI tends to use an out-of-pocket costs concept, so certain budget items that are paid for indirectly have relatively small weights in the CPI, most notably medical care. The financial services category is another example. Financial services cumulatively account for less than 0.25% weighting in the CPI, because most costs for financial services are paid for through implicit fees rather than out-of-pocket expenditures. In contrast, the PCE deflator accounts for all consumer spending, even if the actual outlays are paid for indirectly, in the case of implicit financial services charges, or by a third party, in the case of medical care covered by private insurance or through Medicare/Medicaid. In the PCE deflator, financial services account for just over 5% of the total. This is one key source of differences between the two major measures of consumer prices.
Portfolio management services
Within the financial services category, one large component in the PCE data is “portfolio management services.” This line item accounts for roughly 1.5% of overall spending and about 1.7% of the core PCE deflator. Measuring the cost of portfolio management services is problematic. What is a unit of output for this category of spending? Is it one months’ worth of taking care of someone’s money? The return that the investment manager earns? Once the BEA has determined the nominal amount of spending in this category, much of it implicit, statisticians have to determine what is real output and what is price inflation. Such determinations can be difficult and potentially subjective.
The BEA’s methodology appears to reflect the incidence of management fees. The structure of those fees makes the price data correlated to the performance of the underlying assets. Consider a hedge fund which charges a management component and a performance component: 2% of assets and 20% of returns, respectively. When the market plummets, performance fees drop and vice versa.
There is clear linkage between the portfolio management services component of the PCE deflator and the stock market. Examining monthly changes in the PCE deflator component vs the S&P 500 index for one of the most volatile periods for stocks in recent memory, it is evident that the two series move together, both falling substantially in 2008 and rebounding sharply in 2009 (Exhibit 1).
Exhibit 1: S&P 500 vs portfolio management services prices 2008-2009
The left and right scales suggest that the proportions during this time period were almost 3-to-1, that is, a 3% move in the S&P 500 index tended to produce about a 1% move in the portfolio management services component of the PCE deflator. The moves were also virtually contemporaneous as there is no discernible lag.
The relationship over the past five years has remained pretty tight (Exhibit 2) but has changed in two important ways. First, the relative scales show that the ratio is now less than 2-to-1. Now a 2% move in the S&P 500 index is associated with a swing of less than 1% in the portfolio management services component of the PCE deflator. Perhaps this reflects cuts in performance fees over the last decade. Second, for many of the swings in the S&P 500 index, the corresponding moves in the price deflator appear to be lagged by a month or so. In particular, the swings in the S&P 500 in 2015 and early 2016, as well as the steep drops late last year, are reflected in the portfolio services deflator with a lag.
Exhibit 2: S&P 500 vs portfolio management services prices 2015-2019
Replotting the same S&P data against a one-month lag for the price changes makes the correlation more visually obvious (Exhibit 3).
Exhibit 3: S&P 500 vs portfolio management services prices 2015-2019
The recent monthly swings have largely conformed to the rules of thumb laid out above. In October, November, and December, the S&P 500 index slid by 4.0%, 2.2%, and 5.7% respectively, a cumulative total of 11.9%. Incorporating a one-month lag, the portfolio management services component of the PCE deflator sank by 2.8%, 0.1%, and 4.4%, from November through January, for a cumulative drop of 7.3%. This relationship is broadly consistent with the scale of the 2015-2019 graph, indicating a somewhat less than a 2-to-1 relationship.
Subsequently, the S&P 500 index rebounded by 1.6% in January, 5.7% in February, and 1.8% in March, a cumulative 9.1% gain. The portfolio management services index increased by 2.6%, 1.5%, and 4.1% in the February-to-April period, an 8.2% rise. Thus, if anything, the PCE deflator gauge may have increased faster than would normally be warranted (perhaps already capturing some of the additional 3.6% jump in the S&P 500 index in April).
Given the significant 1.7% weighting of this category within the core PCE deflator, the recent volatility in the S&P 500 has had a noticeable impact on the core readings. For example, the 4.4% slide in January subtracted 7½ bp from the January core PCE reading. A flat reading for the portfolio management category, all else equal, would have pushed the January core inflation monthly increase to 0.11% (the actual reading was up 0.03%). Conversely, the 4.1% jump in April added 7 bp to the core PCE, whereas a flat change in the portfolio services component would have limited the core gain (which was 0.25%) to 0.18%.
Chairman Powell’s declaration that the dip in core inflation as measured by the core PCE deflator early this year was transitory was borne out to a degree by the steep monthly advance in April, which almost rounded up to +0.3%. The portfolio management services component contributed to the recent swings, which he alluded to in his press conference. Another key contributor to the soft start to the year, apparel, has yet to reverse and was almost certainly down sharply in March and April in large part due to change in the BLS’s data collection methodology, an entirely different dynamic.
One might reasonably ask whether swings in stock prices should be driving noticeable swings in the critical core inflation figures, since they have little to do with the type of cyclical inflation pressures that monetary policy should be systematically responding to. Although market participants generally accept the inflation data at face value, inflation figures often swing around on a high-frequency basis due to anomalies that should not have implications for the Fed. Understanding why inflation is accelerating or moderating can be important in gauging what the Fed’s response may be.