Month: March 2022

In an interview with Ezra Klein, economist Larry Summers says there are two things that he thinks are likely to produce a long-term inflation crisis. The first is that demand is too high. The second is that if wages go up too much (hooray!) they will eventually produce even higher inflation and before long the purchasing power of workers starts to decline (boo!).

So let's take a look. On the demand side, consumers have two main sources of income: wages and savings. Here's what that looks like:

The spikes coincide with the two big COVID rescue bills, which filled up household savings accounts. But as you can see, that money has largely been spent and there's no more on the way. If we follow the current trend, we're likely to revert to our pre-pandemic trendline in a few more months.

Another possible source of money for households is consumer loans. Even if wages and savings are falling, a frenzy of loan activity could produce higher household spending and goose demand dangerously. But that's not happening:

Finally, on the wage side, I'm a little puzzled about what the problem is supposed to be:

A '70s-style wage-price spiral happens when wage demands outpace inflation, which then forces companies to raise their prices above the inflation rate, rinse and repeat. But that's not what's happening today. Wage increases have been below the inflation rate for nearly a year.

So we have declining real wages, declining household resources, and a restrained consumer loan environment. Beyond that, there's no special reason (is there?) to think that our supply chain shortages will last much longer now that everyone is back to work.

IANAE, and I understand that this is a simplistic, 100,000 foot view of things. But if it's wrong, I'd sure like to see a more detailed exposition of where inflationary pressures are supposed to come from.

The Washington Post reports that Donald Trump's phone log on January 6, 2021, goes suddenly silent for 7 hours and 37 minutes starting when Trump addressed the rally to overturn the vote and continuing through the attack on the Capitol. Shortly after the rioters were finally cleared, the phone logs once again list calls to and from Trump:

Aside from this gap being pretty unlikely under any circumstances, we know for a fact that Trump made and accepted numerous calls during this period. Obviously he's not trying to hide those from investigators since they're already public record, attested to under oath by various of Trump's friends and aides. One of them is even listed in the presidential daily diary. So what is he trying to hide?

According to the Bureau of Labor Statistics, here's the inflation rate for new cars after it's been adjusted for the overall rate of inflation:

This probably looks very wrong. A Toyota Corolla cost about $5,000 back in 1980 and costs about $20,000 today. That's an increase greater than the overall rate of inflation, and it's typical of all cars. So why does the BLS say the inflation-adjusted price of cars has gone down by nearly half?

The answer is hedonic adjustments. The BLS measures not just prices, but also changes in quality. The price of that Corolla may have gone up a little faster than overall inflation, but the modern version has 17 airbags, antilock brakes, a navi-tainment system, a more reliable engine, and so forth. Essentially the BLS is saying that a Corolla costs roughly the same as it did in 1980 but you're getting twice as much car for your money.

Hedonic adjustments are applied to everything. In some cases, like loaves of bread, they don't make much difference. In others, like personal computers, they make a huge difference.

So how about health care? Everyone agrees that you have to make hedonic adjustments for health care, but the amount varies considerably depending on the treatment. An aspirin is an aspirin and requires no adjustment. Conversely, a friend of mine had a stroke a few years ago, but thanks to quick treatment with the latest technology he recovered completely by the next day. How do you measure quality improvements like this?

Tyler Cowen points to a new paper that uses a pile of cost-effectiveness data to do exactly this. It suggests that over the period 2000-17 the BLS has way underestimated quality improvements in health care and therefore way overestimated health care inflation. The authors illustrate this with a price index that's a little obscure and hard to replicate, so instead I'm going to roughly mimic their results using the ordinary PCE price index:

Very roughly speaking, official figures suggest that annual health care inflation has averaged about 0.8% above the overall inflation rate. But if you recalculate that number using the bias adjustment in the paper, prices have gone down by about 2% per year compared to overall inflation.

For the record, the numbers in the paper, which are based on a sectoral output price index and a complicated rebiasing formula, are 0.53% and -1.33%. However, there are a couple of reasons not to take the precise numbers too seriously:

• They are based on a database of quality improvements for individual procedures, along with an estimate of how widely these procedures are used and how much their increased efficiency has diffused throughout the health care industry. This obviously has a huge scope for error.
• The quality improvements are converted into QALYs, or the number of additional years of life that patients get from better procedures. However, the authors say their results are highly sensitive to estimates of the value of a single year of life. They use $100,000, but if instead they used $50,000 it would eliminate their main result.

So that's that. The authors conclude that the quality of health care has increased much more than official estimates, but even with painstaking effort this is always going to be a very difficult measurement to make. Even small changes in model estimates can balloon into large differences over the course of a couple of decades.

Still, it's an intriguing study. If the average treatment for a heart attack cost $100,000 in the year 2000 and provided five extra years of life, how does that compare to a modern treatment that costs $115,000 (adjusted for inflation) but provides seven years of additional life? The cost of the treatment is 15% more, but the cost per QALY has gone down 18%. So what's the right inflation measurement for that?

This is Interstate 15 where it crosses California Highway 76. The view is looking south. San Diego is shrouded in clouds at the southern end of I15; Oceanside is off the frame at the western end of Highway 76; and Camp Pendleton is off the frame to the lower right.

It seems like the whole Will Smith-Chris Rock thing needs some added juice to keep it going. But what? They're both Black, so there's no racial angle. They're both cis men, so there's no gender or sexual orientation angle. They're both the same age. They were both raised sort of vaguely working class. They both grew up in the northeast (Philadelphia, New York). Neither went to college.

These guys could practically be twins. There's just nothing to mine. However, Wikipedia informs me that Rock has a learning disorder. This means we can all redirect our outrage to the fact that Will Smith assaulted a guy with a learning disorder. That will have to do unless someone can come up with something better.

Exciting news, genealogy nerds! Census records are kept private for 72 years, which means the 1950 census records will be available soon. On Friday, to be exact.

What's even more exciting is that the 1950 records have already been digitized and will be fully searchable immediately. So if you want to find out where your mysterious uncle Joe was really living in 1950, this is your chance.

Ultraviolet light kills pathogens such as COVID-19 but is also a danger to human beings. But what if there was a version of UV light that killed pathogens without affecting people?

It turns out there is. Krypton chloride (KrCl) excimer lamps radiate light at a frequency of 222 nm, also known as Far-UVC, which is safe for people but still just fine for killing airborne pathogens similar to COVID-19:

This comes from a recent study in Nature, which didn't use actual COVID-19 pathogens but did use something similar enough that the researchers believe their results apply to COVID-19. Further research is still necessary to confirm this.

It's worth noting that this isn't really anything new. For example, Far UV Technologies would be delighted to sell you one of their broad range of krypton excimer lamps starting at the low, low price of $2,499. Amazon will sell you one for $1,699 with free shipping. There are others out there too, all of which promise to sanitize your environment safely and quickly.

Personally, I'd like a small, battery operated unit for $19.95 that I could put in a pocket then use wherever I happened to be. However, this appears not to be technologically feasible just yet. That's a shame.

For the record, I have no problem with President Biden's ad-libbed comment about Putin yesterday. Blunt talk about autocrats is sometimes welcome, and wartime seems like an especially auspicious occasion for it. What's more, I very much doubt that it will cause Putin to blow his top or anything like that. Quite the opposite: the fact that Biden is obviously very sincere in his loathing of Putin makes it clear that the US and NATO are unlikely to back down in Ukraine. This in turn should motivate the Russian leadership to look for ways to retreat from an obvious quagmire.

From Chris Wallace, longtime Fox News host, on why he finally left the network:

I’m fine with opinion: conservative opinion, liberal opinion. But when people start to question the truth — Who won the 2020 election? Was Jan. 6 an insurrection? — I found that unsustainable.

....Some people might have drawn the line earlier, or at a different point. I think Fox has changed over the course of the last year and a half. But I can certainly understand where somebody would say, "Gee, you were a slow learner, Chris."

Better late than never. Welcome back to the reality-based world, Chris.

The crypto community brags that there are thousands of cryptocurrencies these days. But doesn't that, almost by definition, demolish the idea that cryptocurrencies have real value?

When Bitcoin was unique, that was one thing. But over the past few years it's became clear that nearly any random person can create a cryptocurrency. All it take is some technical knowhow, a little bit of marketing skill, and a highly artificial mechanism for guaranteeing scarcity. Every cryptocurrency is nearly identical to every other one.

Put aside the complicated arguments and long conversations about why crypto is for real. Just answer one question. It makes no sense that just about anyone can create a cryptocurrency. So why can they?