• Health spending Q&A

    After I posted the following chart from  Charles Roehrig, readers emailed me with questions. Some I can answer. Some I cannot. The Q&A is below the chart.

    Question: Does heath spending as a share of GDP only go up during and around recessions?

    Answer: Not exactly, but it’s a good rule of thumb anyway. The figure above only goes back to 1990. If you go back farther, you see that health spending as a share of GDP grows between recessions too. The following figure is also from Charles Roehrig. Ignore the black line (follow the link for details). See how there is growth between the early 1980s and early 1990s recessions? However, what is true is that health spending as a fraction of GDP tends to have some flat spots between recessions and steep growth during and around them. That’s largely due to changes in the denominator, GDP. When GDP growth is lower, the ratio of spending to it is higher and vice versa.

    Question: Why was health spending as a fraction of GDP flat during much of the 1990s?

    Answer: Managed care kept health spending growth down and GDP growth was substantial. The combination of the two just happened to keep the ratio nearly constant.

    Question: Why was health spending as a fraction of GDP flat in part of the 2000s?

    Answer: GDP growth is part of the answer. But health spending growth was modest too. Nobody I’ve asked has a good explanation for it.

    Question: What about other flat spots, prior to 1990, such as the one in the late 1970s?

    Answer: I don’t really know, but it is worth keeping in mind that our health system was very different before 1980. Prior to that year (or thereabouts), we had similar spending growth as other wealthy countries. Afterwards, the US pulled away from the pack. See:

    Question: What are the implications for capping spending at GDP + 1 percentage point (or 0.5 percentage points or some other GDP-keyed target)?

    Answer: This could be problematic during recessions. Health spending is relatively sticky. It’s rate of growth doesn’t necessarily fall when GDP falls, or not as much anyway. It’s clear from the charts above that during boom times, holding health spending growth to GDP + 1 (or something like that) is very doable. But, what happens when economic growth sags? Is it a good idea just to slash health spending?

    What would be better is to permit health spending to go above the target (GDP + 1 or whatever) during recessions and insist that we make up for it when the economy recovers. This would spread the pain out more evenly, demanding that we tame health spending relative to GDP when it seems more feasible to do so.

    Also, as health care becomes a larger fraction of the economy, its growth is more tightly correlated with overall GDP. After all, if the entire economy was health care then achieving a target health care growth rate no higher (or lower) than GDP would be trivial. We’d be living in caves, though.

    • Thanks for this. I had no idea about that 1980 break away. Is there a consensus about what caused that or is there any interesting writing on the subject you’d suggest? Off hand I can’t think of a big policy change then – on the scale of Medicare / Medicaid or tax deductions for employeer sponsored care – the big policy changes are all earlier in the century. What am I forgetting?

    • The change in the 80’s was the introduction of third-party payment and low or no deductibles in early managed care plans. Prior to tgrowing market penetration, patients paid out of pocket for most services and were reimbursed by insurers. Managed care plan penetrations peaked in the 90’s, when they are often attributed with controlling growth, but their introductory period was a decade earlier. Third party payment in Medicare also started in the early 80’s, freeing Medicare beneficiaries from up-front spending and claims filing.

      Third-party payment decreased the out-of-pocket cost for patients, decreasing the percieved price of care and driving demand and utilization up. Third-party payment increased the operating cost of providers and payers because of the claims handling costs on both sides of the transaction and also because the payers had to add new procedures and programs design to control utilization growth. Those utilization control programs amight have limited cost growth during the 90’s but they had also been responsible for some of the cost growth during the ’80s as both payers and providers added new administrative burden.

      As an illustration of this for physician practices, prior to 1980 the averge physician practice hired two non-physicians per practicing physician. Starting in the 80’s that number grew continuously. It now averages 5 per physician. Those new employees are handling claims filing, documentation and follow up plus eligibility determinations, referral documentation, contract negotiation, health plan credentialing, and more functions that were non-existent before the early 80’s.

      • This is the best explanation I’ve seen of the 1980s divergence of US health spending from other nations. Can you suggest a publication that documents it? Do other nations use third-party payment? If so, when was it introduced elsewhere and is an effect evident in data?

        • Sorry it has taken me so long to reply. I don’t know of any study that reaches that far back in time to see this. The changes in physician practice operating cost can be very well documented from the AMA’s Socioeconomic Survey series. it goes back at least to 1980 but stopped in 2000, so there is no current data, but it clearly shows the practice cost changes starting in the early 80’s. If we had the data on insurers’ administrative cost, it would surely be clear there, too. It certainly takes more administrative cost to run an HMO or other managed care plan than it took to run the old-style major medical plans, where claims were only filed once the deductible was reached. Furthermore, zero deductibles and low co-pays almost certainly increased utilization. Even emergency department utilization was probably affected, but I’m not at all sure that there was any other material effect on hospitals.

    • A quick comment on capping spending at GDP + .5 or GDP + 1. Austin suggests the following in his comment above:

      “What would be better is to permit health spending to go above the target (GDP + 1 or whatever) during recessions and insist that we make up for it when the economy recovers.”

      I think the way to do this is to use “potential” GDP (PGDP — what GDP would be at full employment) in the denominator. Suppose you peg health spending to PGDP + 1. During recessions, this would allow spending to exceed GDP +1 and during recoveries, when GDP grows faster than PGDP, it would hold spending below GDP + 1. In summary, capping health spending in terms of PGDP rather than GDP accomplishes what Austin describes above. As long as GDP continues to return to PGDP every now and then, holding to PGDP +1 will also ultimately equate to GDP +1 over long stretches of time.

      In response to Austin’s other comment about how the growing health share of GDP affects the ability of their growth rates to differ — I think it is only a matter of time before we start comparing health spending growth rates to that of the non-health portion of GDP rather than all of GDP. Just makes a lot more sense.