Mark Perry updates his chart on divergent inflation trends.
Seven of those goods and services have increased more than average inflation, led by hospital services (+211%), college tuition (+183.8%), and college textbooks (+183.6%). Average wages have also increased more than average inflation since January 1998, by 80.2%, indicating an increase in real wages over the last several decades.
The other seven price series have declined since January 1998, led by TVs (-97%), toys (-74%), software (-68%) and cell phone service (-53%). The CPI series for new cars, household furnishings (furniture, appliances, window coverings, lamps, dishes, etc.) and clothing have remained relatively flat for the last 21 years while average prices have increased by 56% and wages increased 80.2%.
These cross-sectional differences in price movements are an order of magnitude greater than the time-series variation of “the” inflation rate. Indeed, I read these data as saying that “the” inflation rate is pretty close to a meaningless number
These services whose providers have benefited from hyperinflation are a prime target for the technology industries.
Often it is the set of rules that these services operate under that is the real culprit, some of them dating back decades and even centuries. Automate them and sell the relevant appliance from somewhere not affected by inflation enabling legislation, and watch people vote with their cash.
Pricing takes time.
Me and the IRS are still pricing my past taxes from 20 years ago. we get price revisions going back ten years once in a while, and none of the current indices are any good unless prices have had six months of averaging. The practice of price targeting has led to a very destruction inflation adjustment clauses in all our government obligations. Inevitable cyclic crisis.
What is sad about these trends is that inflation is in necessary goods (we need medical care, we need college degrees to get good jobs) while the deflation is in sectors were we could take or leave it without much fuss (we don’t need a flat screen TV, we don’t need a bunch of toys for our kids).
So I would say the effect this has is to reduce financial security or flexibility.
The decline in prices of the latter is the cause of the rise in prices of the former.
This is the concept of Ricardian rent or scarcity rent.
If certain goods and services are inelastic in supply do to natural scarcity or positional ranking, and one needs these goods or services to access lifestyles and levels of social status which are significantly greater than the next best alternative, then these sectors become Neo-Ricardian and are collectively able to extract excess rents up to the XNPV of that gap. (As I’ve pointed out in comments here before, there really is no other good explanation for the obscene college textbooks price curve, especially since many courses have no need for new content).
So technological and economic progress that makes certain items cheaper (that is, it requires a smaller portion of real income to satisfy certain preferences) merely frees up the funds for people to bid up the prices in the Neo-Ricardian sectors. This is even true for the other Neo-Ricardian sectors, so long as there is at least one left (which there always is, in the government).
By this logic, it is always the things “we really need” (inelastic demand and supply) that will increase in relative price when the prices of things “we don’t really need” (elastic demand and supply) decrease.
Random Anecdote:
I was in the lobby the other day with a young man in his 20s who was living in the condo and whose father was a professor at JHU. Apparently this professor was working on the !18th! edition of his textbook and the revenue from this edition was enough that he could employ his son “full time” on aiding him with that process while paying him a salary that allowed him to live independently.
I must say that upon hearing that he was workin on the 18th edition I actually reacted with a kind of amazed disgust, raising my voice in surprise a bit and commenting that pumping out an 18th addition seemed a something of a scam.
First off, there was no mention of land.
Health care and education are highly elastic on the supply side. There is also a substantial amount of elasticity in demand, even in healthcare.
College textbooks? Practically unlimited elasticity of supply. There is only one reason they are expensive. A professor usually has absolute authority over assignment, and exit is destructive to the more important goal of a obtaining a degree.
None of this has anything to do with Ricardian rent theory.
The sectors with the highest price increases are the one’s where demand is most heavily subsidized (college education, healthcare).
I don’t see any evidence of the ACA bending the cost curve on health care in these numbers, either. Not even the slightest inflection point.
How right this is about “necessary goods”. I call it “blackmail pricing”.
A similar phenomenon is when someone has bought an expensive but good value item such as a camcorder for say $400 and, say, a switch fails out of guarantee and they are charged $100 to replace it.
Or if something is required to obey a regulation – it is priced at a large multiple of what something of similar complexity would otherwise cost.
They are less necessary than other goods like food and water. Also today college degrees and medical care are cheaper the poorer you are. Still I think more competition should be allowed in those areas.
Sure, but I’m not talking food and water (which largely became cheaper a long time ago in the developed world, and outside of the time period mentioned in this article).
Yes, Medicaid is free and you can get financial aid if you are poor. The trick though is that all of these things have very high effective marginal tax rates. You start to earn a little money and medicaid kicks out. You earn some more and your kid doesn’t qualify for financial aid.
Sometimes I wonder if the goal is to make all income between $25,000 and $100,000 have a marginal tax rate close to 100%.
I would look at these divergent trends a different way.
Most of the high price increases are in segments where there is essentially no foreign competition.
While the segments with flat and/or falling prices are areas where imports have largely displaced domestic production.
This suggest a very different competitive environment at work.
I’ve seen numbers on average change in wages over the last few decades ranging from a decrease of 50%, to flat, to an increase of 50%, and now to an increase of 80%. This does not give me a lot of confidence in our ability to micromanage the economy.
How have clothing and household furnishings not declined more? Food and drink, also.
Arnold said CPI is pretty much meaningless. The comments so far seem to be crying out: we need better meaningless numbers!
As far as medical care is that really inflation or is care getting better?
I guess Robin Hanson might say it’s not getting much better and so it is inflation.
I’m undecided.
The article refers to an index of hospital services, so I think it is safe to assume it is tracking a bundle of standard services, like a hospital bed for a day. I don’t think we can infer anything about any improvements. Sounds like just inflation, or runaway cost bundling.
I take Arnold’s point about the limits of averages (RE: “‘the’ inflation index”), but, for my money, the natural reference point for the long run evolution of prices should be nominal per capita incomes. Something needs to happen for prices to fall substantially faster than incomes in the long run (e.g., rising productivity, benefits of trade, etc.) and some categories are naturally more subject to these sorts of influences than others. While I’m naturally sympathetic to the libertarian perspective, most of the prices Mark shows in his chart have declined relative to incomes. Prices for items that depend on relatively more on domestic labor inputs (e.g., healthcare, education, babysitting, etc.) or that are substantially weighted towards fixed/non-produced assets (e.g, land) are at a fundamental disadvantage vis-a-vis measured inflation when compared to manufactured goods, tradable items, etc, so we should probably expect them to exceed “the” inflation index by a large margin.
Some might object to referencing mean income because the consumer doesn’t necessarily have mean income growth, but a good number of these oft-discussed categories (especially health and education) the median or other distributional income measures are probably not more helpful given that the labor inputs aren’t necessarily the same (median etc.) and, more importantly, these are mostly being paid for by society writ large. Education is probably the most significant exception to this rule, RE: inflation-vs-income growth, but that is found at all levels (pre-K through tertiary), and I’d bet that’s mostly a price measurement issue (unlike healthcare these services typically aren’t disaggregated and don’t have prices…. makes it harder for BLS, BEA, etc. to do their jobs).
The CPI for primary and secondary education measures prices at private schools, not public schools.
Yeah, that’s about what I expect given that the gist of their mission is to measure price inflation for costs born directly by consumers. The BEA’s mission is much broader with the various NIPA components (PCE and Government consumption expenditures & investment) and their price indices show similar trends for primary, secondary, and tertiary education within public and private sectors. The government consumption side is priced at gross output (cost of inputs), which makes inflation in the public sector inflation appear a bit more modest. However, even the government indices show substantial excess price inflation relative to broad indices like PCEPI and CPI*, and given the parallel trends in per-pupil spending, declining student teacher-ratios, etc, I think it’s pretty safe to assume comparable inflation trends if they were measured comparably. (Keep in mind that relatively few private school students are attending the more expensive independent private schools you might be familiar near major urban centers & NE….)
I’m considering a simplified model where supply of any good is either growing or constrained and demand is either elastic or inelastic. Production efficiencies cause supply to grow; Baumol’s cost disease or inherent scarcity constrain supply. This gives us four types of goods:
* Elastic growing: These goods are getting cheaper, so people will consume more of them. Plasma TVs, cell phones, video streaming services, etc. Total spending on these items will still fall as some spending shifts to other quadrants.
* Elastic constrained: These goods are getting more expensive relative to growing goods, so most people use less (but the rich go in for conspicuous consumption). Broadway tickets, beachfront property.
* Inelastic growing: Getting much cheaper. Not a good sector to be in, since price drops aren’t offset by increased consumption.
* Inelastic constrained: Over time, the fraction of an average consumer’s spending in this sector should rise. In the limiting case (cost of “growing” goods goes to zero), nearly all spending is in this sector and economic growth effectively stops. Health care, education.
Healthcare may be highly inelastic at the individual level (especially in the developed world over the past several decades), but that doesn’t mean it’s not highly elastic at the national level. Rich countries engage in all manner of cutting edge healthcare that poor countries don’t (or do much, much less of), and they increasingly do them as they get richer.
Healthcare is also elastic at the individual level. One reason payers don’t want to reimburse telemedicine (which, in theory, would lower cost) is that it would increase net usage.
No argument from me there (also lots of seasonality in the industry tied to people meeting their deductibles etc also), but my point is on individual income is a very poor predictor of the sort of healthcare services one gets over the course of a year. If the coefficient on income isn’t exactly zero or even slightly negative (estimates vary some), it’s quite close to it. The national or even sub-national data show a very different pattern (though sub-national is distinctly intermediate to individual and national)
There has been much mention of the elasticity of demand for healthcare. To the extent that at the individual level, discretionary-out of pocket costs are so low (because of private, public insurance and all sorts of price opacity) and at the national level (though deficit financing seemingly without consequences), how could one ever measure its real elasticity.
Agree with this “individual income is a very poor predictor of the sort of healthcare services one gets over the course of a year”
and point to this para from the linked paper that jumped out at me:
“Within an insured group, the bulk of the health resources will be allocated to those individuals who are ill and stand to benefit from medical care. Individual budget constraints and ability to pay concerns are largely removed through pooled insurance financing. However, variation between groups which do not share access to the same insurance pool is largely determined by differences in aggregate group funding. The group mean will vary with the budget constraint, not with the amount of illness.”
With the proviso that in the US there effectively “commercial”, “Medicare”, and “Medicaid” pools…at least until ACA messed up the “commercial” one.
That said, not being a health economist, I’d echo MG’s question. From my coal-face-eye-view the market is so opaque and byzantine that even calling the provider portion a “market” seems optimistic.
A toy 2×2 matrix model is going to have serious limits. Elasticity is a continuum, not a binary. Elasticity can depend on details of the payment system. The price elasticity of healthcare depends on the exact treatment and on the patient’s prognosis without treatment. The model assumes all tech reduces cost; there’s no place in it for inventing new expensive stuff.
All economic models are false; some models are useful. I get the feeling this one points to some understanding, but I’ve been wrong before.
Ivy League schools are positional goods. Also-ran state schools not so much. Although I like Handle’s Neo-Ricardian analysis a lot I’m not sure it exclusively explains everything. I think there is “causal density” for a lot of phenomena.
For example, severe inflation (and much price volatility) in the housing sector is mostly caused by “urban growth boundaries” around cities. If you chart housing price changes for only cities without severe suburban-growth restrictions, you see relatively little inflation (because supply is elastic and competitive).
I think inflation in costs for non-positional education is chiefly driven by rent-seeking (for example, colleges and politicians conspired to make students borrow vast amounts of money and give it to colleges–and indirectly to politicians. Colleges and politicians conspired to enable colleges to practice nearly perfect price discrimination– by forcing students and parents to disclose all their assets, income, and tax returns to colleges’ misnamed “financial aid” offices).
A lot of demand for higher-ed is driven by rent-seeking. For example, prompted by colleges the Washington DC government has decided that each daycare worker must have a two-year college degree to change a diaper or supervise a playground. Now humble people seeking modest daycare jobs must first become college students. Some will become college dropouts, many will end up “paying back” their student loans forever.
Really, Mancur Olsonism explains a lot of inflation.
colleges and politicians conspired to make students borrow vast amounts of money and give it to colleges–and indirectly to politicians.
I don’t see community colleges making huge campaign contributions the way, say, the finance industry does. I think the inflation is caused by inelastic demand (plus Baumol’s cost disease) with a cause that is ultimately cultural. The American Dream is one of opportunity. American youth would rather pay ridiculous prices for an implausible promise of opportunity than accept that their economic circumstances will remain constrained.