20 March 2014
“Beans, bacon, whisky and lard” are the old sailor’s 4 basic food groups.
When it comes to measuring inflation, exactly how much pure price increase is in a can of beans?
As the FOMC muses about inflation, consider the complications in measuring generalized price increases. A few include:
- What question are we answering? A measure of inflation depends on how the answer will be used. Thus, measures have been constructed differently, including:
- Consumers by:
- Occupation, primarily Consumer Price Index for Urban Wage Earners and Clerical Workers (CPI-W) used mostly to calculate Social Security cost of living adjustments (COLAs) as reported by the U.S. Bureau of Labor Statistics (BLS).
- Geographic area, such as the Consumer Price Index for all urban consumers (CPI-U), and regional and local measures. 89% of consumers are covered in CPI-U; “urban” areas include smaller cities. Also from BLS.
- Producers by type of good or stage of production as reflected in the Producer Price Index. Again from BLS.
- Subsets of CPI such as “core” CPI (excluding food and energy) to reduce influences of weather and geopolitics.
- Individual products and aggregates reported as Personal Consumption Expenditure (PCE) price indices. Government and business buyers are included in GDP price indices. Both estimates are from the U.S. Bureau of Economic Analysis (BEA).
- Country source of product, important in import/export calculations
- Quality of product:
- If a can of beans costs more because it has a new sauce, is that pure price increase or not because the “quality” is better?
- Quality to whom? If the government mandates a change in automobiles that people won’t pay for voluntarily, is that better quality or only price increase?
- Is the change in “quality” so different that a product becomes a new product? How fast does this change occur? Measuring pure price increase in “Computers, Peripherals and Parts” is particularly troublesome as described by the BEA in the note on Table 4.2.6 of the National Income and Product Accounts.
- If beef prices increase and people buy more pork to make BBQ sandwiches with the same protein, is that pure price increase?
- If I buy store-brand extra zippy sauce beans instead of name-brand regular beans, is net quality better or worse?
- Timing – in adjusting for all the above, when are changes made?
- Data: As in anything, error creeps in from source data completeness and accuracy, sampling design and implementation, processing and more.
- Calculation: Nuances matter in calculating month-to-month or month-to-prior-year-month changes to create the indices. Then, are averages or medians more helpful?
For investors, range of implications:
- Forecasting company financials is different based on nature of input products. Commodity good prices driven by scarcity are different from service prices driven by growth in global educated workforce and downward price pressures (such as in outsourced professional services).
- Forecasting capital goods replacement is different. A 10 year old product today might retain more value than a 10 year old product 10 years ago. How many Windows upgrades will Microsoft sell after XP support expires in April?
- If trying to outguess the market in a CPI-adjusted financial product, don’t apply the wrong methods or data.
Know what matters to you in a specific investment decision and adjust appropriately
This comment was suggested by a helpful article from the BLS, Beyond the Numbers: Why does BLS provide both the CPI-W and CPI-U? by Stephen B. Reed and Kenneth J. Stewart, February 2014 http://www.bls.gov/opub/btn/volume-3/why-does-bls-provide-both-the-cpi-w-and-cpi-u.htm