Open Enrollment and the CPI - Part Two

Last week, in Part One of this look into rising medical costs as they relate to the Consumer Price Index, a great disparity was identified - the difference between rising health care costs as detailed in the annual Kaiser Family Foundation study and rising health care costs as reported in the Bureau of Labor Statistics CPI.

Near ten percent versus four percent, respectively - as it turns out, that was only half of the story.

Even more significant than the rise of medical care costs are the differing views of how much medical care spending contributes to total consumer spending. Not surprisingly, the amount in the CPI is low compared to others measures.

As shown in the table below, the current make up of the CPI includes medical care costs as 6.1 percent of total consumer spending.



While the Bureau of Labor Statistics goes to great lengths to determine what percent of consumption goes into which categories in their Consumption Expenditure Survey, and while the survey is due for an update in just another couple months, the fact remains that according to current calculations in the CPI, health care costs contribute only 6.1 percent to total consumer spending.

There is mind-numbing detail about how data is collected and how it is interpreted, but again, the bottom line is 6.1 percent.

The Consumer Expenditure Survey (CE) collects annual consumer spending for each CPI category; this provides the basis for the item category weights. BLS replaces these weights every two years with ones based on more recent Consumer Expenditure Surveys. To obtain the category weights, BLS combines expenditures from the CE's for two years and updates them for price change to the December before their first use in the CPI. For example, the expenditures reported on CE's for 2001 and 2002 updated to December 2003 became the basic weights for use in the CPI from January 2004 through December 2005. Every month, to compute that month's index, BLS updates the base weights for price change from the previous month. Every year, the BLS publishes relative importances for the previous December; these are base weights updated for price change and expressed as a percentage of total weight. Weights for components with greater than average price change will increase more than those with smaller than average price change. As a result, the change in a component's relative importance from one December to the next reflects its price change relative to that of all other categories as well as every two years, the biennial weight replacement.

It goes on like this for pages and pages - it sounds impressive.

A Different View

Over at the Bureau of Economic Analysis, they do an alternative consumption expenditure survey, based on "an internationally accepted set of guidelines". This survey results in a markedly different conclusion regarding the importance of medical care costs to one's budget.

Over there, they get a number that is more than three times higher - 18.7 percent.



Both the BLS and the BEA data are from 2003, the most recent year for which data is available. It is widely acknowledged that as a percent of GDP, spending on medical care is much closer to what the BEA reports, than that which is detailed by the BLS, as evidenced by this New York Times article from earlier this year.

The torrid pace of growth in national health spending cooled a bit in 2003, but the spending, at $1.7 trillion, topped 15 percent of the gross domestic product for the first time, the government said on Monday.

But to really understand what medical care costs might feel like to the common man, some simple math should make help make the point.

Some Simple Math

From White House data it is learned that current median household income is roughly $45,000, from which, for this example 25 percent will be subtracted for taxes, retirement savings, and other non-consumption items, leaving about $34,000.

Now, using the 6.1 percent share for medical care per the BLS, this works out to just shy of $2100, meaning that the median family with employee sponsored health care is still about $600 shy of meeting their share of the average annual insurance premium of $2713.

That is, before anyone steps foot into a doctors office or a pharmacy.

So, the 6.1 percent number really seems kind of dumb - despite how many pages the BLS provides about how it is calculated. What would be a better number?

Continuing with this example, in addition to the $2713 per year in medical insurance, add say another $300 in deductibles, drug prescriptions totaling $900, various other co-payments totaling $300, maybe some dental and optical expenses for another $600, and miscellaneous other medical costs of $500.

This all adds up to $5313, which as a share of $34,000 is about 16 percent.

Here's a real-world example from the USA Today article a couple months ago:

As co-payments, premiums and deductibles rise, those facing financial difficulties are increasingly moderate-income families, such as Jeffrey Herchenroder's, who have health insurance.

The 45-year-old high school teacher lives just outside Albany, N.Y., with his wife, Cindy, and two of his three children on an annual income of about $65,000. His oldest child just graduated from college. His job provides insurance, for which Herchenroder pays about $3,000 annually toward the premium; his employer pays the rest. He has rheumatoid arthritis, and his 11-year-old son has asthma and diabetes. They recently paid $1,000 for their share of the cost of an insulin pump. Everyone in the family takes at least one prescription drug. Each month, he estimates, the family spends $280 on drugs alone.

"We dip into credit cards, and it starts to become a big cloud," he says. "We have good insurance but still some amount of (financial) trouble. Medical costs affect our lifestyle. We don't vacation. We don't eat out. I buy cars with blown engines and put new engines in them so we have something to drive. I've never owned a car newer than 6 years old."

It is easy to see why medical expenses are the leading contributor to bankruptcy filings - with budgets tight and getting tighter, any major medical expense can become a tipping point.

Back to the CPI

So, let's say that the medical care cost component of the total consumer spending should really be 18.7 percent, as indictated in the BEA survey, and that the increase in medical care costs was really 9.2 percent in the last year, as cited in the Kaiser study for health insurance.

The Consumer Price Index as reported by the Bureau of Labor Statistics for the month of September was 4.7 percent on a year-over-year basis. What would the new calculation be?

Some simple algebra will do the trick - first solve for inflation exclusive of medical care:

(93.9% other CPI components * x percent) +
(6.1% medical care component * 3.9 percent) =
4.7 percent inflation
x = 4.75 percent

Now, using the 9.2 percent increase in medical costs as 18.7 percent of the total CPI:

(81.3% other CPI components * 4.75 percent) +
(18.7% medical care component * 9.2 percent) =
5.6 percent inflation

Almost a full point higher than the official headline number of 4.7 percent.

To go a step further, if the 9.2 / 18.7 percent medical care numbers are combined with real housing costs in place of owner's equivalent rent (as discussed here last month), the new headline inflation number would be over 8 percent with a core rate over 6 percent.

Obviously, real GDP would be adversely affected by the combined 3.5 percent understatement of inflation, as calculated here.

In fact, this 3.5 percent difference would make real GDP about zero.

Notes:

1. This does not consider quality adjustments - surely health care quality is improving, however not nearly as rapidly as costs are rising.
2. This does not consider either pre-tax payment of employee medical insurance costs or health savings accounts, both of which would reduce employee costs.
3. The 9.2 percent number may in fact be too low. There has been a drop in the number of companies offering health care plans and a rise in high deductible plans which may have influenced this figure.
4. There is a wealth of interesting survey data from the Kaiser study here.