# Affordability

## Background

I've already written about how, contrary to popular belief, the median income has risen quite steadily since the 1970s. This page considers the related criticism that even ignoring income, the fact that specific expenditures are increasing makes it harder and harder to live a "decent" life.

It's worth pointing out that since the CPI and GDP deflator include all price increases already, this argument is effectively already accounted for in my earlier analysis; however, I wanted to investigate these particular claims anyways.

In particular, I look into the cost of houses, the cost of college tuition, and the cost of health care. I end up finding that housing is actually very cheap relative to historical levels and that college tuition is either flat or has risen by only $2k over the last two decades. I'm less confident about health care. It does look like that prices are rising faster there than for other services, but due to imperfections in CPI, its difficult to tell for certain. ## Housing While people have complained about the cost of housing since time immemorial, in truth, their thinking is typically muddled. For instance, while almost all new housing is expensive, this has always been the case historically, with lower-income people moving into older homes and wealthier people moving into new ones - not dissimilar to cars. Instead of comparing the prices of new homes to old homes, we need to see how the price of a specific home changes over time to get an apples-to-apples comparison. Of course an individual home isn't super illuminating, so economists do this for lots of homes and average the price changes together to create the Case-Shiller index. After adjusting for inflation (CPI), the graph for US homes looks like The two exceptions to this rule are that housing became noticeably more expensive post WWII and throughout the 21st century. The first exception can largely be blamed on government policy. During WWII the government banned new domestic housing construction except for defense purposes, while after WWII the Veterans Administration guaranteed millions of home loans HUD's Historical Timeline. Both these moves dramatically and artificially increased the demand for housing. I used to think the boom in the early 2000s was just a temporary surge due to issues leading up to the 2007 subprime mortgage crisis since prices eventually fell to late-90s levels. However, the significant rise in prices since then have started to change my mind. All that being said, there is to more to the cost of housing than a home's price. Most noticeably, since most homeowners borrow to buy their home, the monthly cost of a home is just as much determined by the interest rate on the mortgage. In particular, the formula for a monthly mortgage payment is the value of the mortgage multiplied by $$i \frac{(1+i)^n}{(1+i)^n-1}$$ where$i$is the monthly interest rate and$n$is the number of months. After adjusting for this using the average 30-year mortgage MORTGAGE30US, we achieve a monthly-payment version of the Case-Shiller index. Unfortunately, I only found data going back to 1971: From this it is pretty clear that the monthly cost of housing has fluctuated quite a bit, it's actually at a historical low now: if you wanted to buy the home your parents have, then (on average) you'll be paying less per month for it. ## College Unlike housing, the quality of college instruction probably hasn't changed much over the last several decades. It was and remains largely (a) a lecturer teaching students (b) students doing homework and getting feedback (c) some smaller discussion- or lab-based classes. The main exception to this is room and board, which (at least anecdotally) has gotten much nicer. For this reason, we'll be focusing on tuition to make sure we're doing an apples-to-apples comparison. Nevertheless, average published tuition has 2.1% faster than inflation over the last 55 years. This is bad enough, but that statistic actually hides the true increase in tuition prices since people are shifting from expensive to cheaper options. For instance, among 4-year and 2-year public schools, the increase has been 2.8% and 2.6%, respectively, while among 2- and 4-year private institutions tuition prices have increased by 2.5% and 2.0%, respectively Table 330.10. The obvious explanation is that most of this increase is due to price discrimination: colleges raise their sticker price to milk their wealthy students for more money while providing greater financial aid so that typical students don't see much of a change. This explanation is at least plausible since the percent of students receiving financial aid has increased over the last couple decades Sources of Financial Aid and the average amount of financial aid they receive has grown as well Table 331.20. For instance, between 2000 and 2017, the percent receiving financial aid rose from 70% to 84% while the average grant size increased by 51% - and that's all ignoring the billions of dollars in guaranteed and subsidized loans. Research by the College Board suggests that financial aid has generally reduced the true growth of the cost of college Average net price by sector over time. For instance, the net price of tuititon at 4-year nonprofit college has remained flat over the last 20 years even whilethe sticker price has increased ~50%. The evidence is more mixed for 4-year public colleges. Over the last 20 years, the cost of tuition has doubled even after accounting for inflation and financial aid. That being said, financial aid still makes attendance much more affordable, with average tuition paid shrinking to around$4k per year.

The story for 2-year public schools is even stronger, a small drop in the net price of tuition adjusted for inflation.

In short, after adjusting for inflation and financial aid, the cost of tuition has remained flat at nonprofit and 2-year public schools, while increasing from ~$2k/yr to around$4k /yr at public ones.

Finally, I should point out that since college is a service, most of its cost is labor. As mentioned, mainstream college hasn't really seen much productivity-enhancing innovation. These two facts strongly suggest the price of college should rise faster than inflation Baumol's cost disease. I consider a mild miracle that this has only actually happened at 4-year public schools.

Finally, I'd be remiss if I didn't at least mention that it would be probably be good if the cost of college rose.

## Health Care

Health care costs are by far the trickiest to examine since (a) it's the most heterogeneous; (b) its abilities have expanded the most, making apples-to-apples comparisons difficult; and (c) the price structure of our health care system is wildly complicated causing the same procedure to have multiple different prices.

For instance, controlling for inflation, health care has grown by nearly 4% per year for the last 4 decades DHLCRC1Q027SBEA CPIAUCSL. The question is how much of this is due to (a) newly invented procedures (b) more procedures (c) improved quality of procedures and (d) actual increase in price.

As far as I can tell, the only way out is to choose a very standardized procedure that hasn't changed much and look at its price over time. In principal this is exactly what the CPI does, though it is generally agreed it is imperfect and therefore overestimates the true increase in prices for three reasons United States Academic Decathlon:

1. substitution bias - CPI is a Laspeyres index
2. unmeasured quality bias - researchers attempt to adjust for quality improvements in goods over time, but this is hard to do perfectly
3. new goods & services - new goods often dramatically fall in price their first couple years of existence, when they're usually not in the CPI basket of goods

That being said, I don't see a better alternative.

The CPI for medical care has been rising faster than the overall CPI since at least 1950, averaging 1.7% faster per year CPIAUCSL CPIMEDSL. Services in general grew only 0.7% faster than overall prices over that time Table 1.1.4, which leaves a 1% unexplained gap. It's possible the three biases I mentioned above are more prominent in health care than other services, which makes me uncertain whether that 1% gap is real or due to methodological error. If that seems unreasonable, consider that a government report estimated CPI as a whole was biased upward by over 1% The Boskin Commission Report, so expecting twice that kind of discrepancy among health care seems plausible to me.

[ As an aside, the fact that services grew 0.7% faster than overall CPI is nearly equals the 0.7% growth in median income. This makes some sense due to Baumol's cost disease Baumol's cost disease. ]

Finally, I'm just going to leave this here.