I've uploaded a set of teaching notes I use in my Research Methods class, which is basically a class on statistics for policy work. jwmason.org/wp-content/upl…
They're very much a work in progress, but if you teach econometrics or statistics (or are trying to learn them) please take a look - I'd be very interested if you find them helpful.
I'm self consciously trying to follow a somewhat different approach than most econometrics texts, by consistently taking a sample-first rather than a population-first perspective - starting from data as you actually encounter it rather than a hypothetical true distribution.
I'm also trying to think about whatis useful for someone who wants to data work in a thank tank or similar place. That isn't the audience for econometric textbooks, while books that are aimed at more of a policy audience are in my experience too basic and uncritical.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
You often hear that higher inflation is associated with higher real interest rates - that's supposed to be one of its major costs. I don't know where this bit of folk wisdom came from, but the reality is exactly the opposite. Higher inflation almost always means lower real rates.
The US has experience many episodes of inflation (and some of deflation) over the past century, and the early 1980s is the *only* case in which higher inflation has been associated with higher real interest rates.
To the extent that interest rates are linked to inflation, this is entirely mediated by the central bank. Higher inflation may cause the central bank to tighten, and bond yields do (slowly and incompletely) move with the policy rate.
There seem to be a significant number of economists and economics-adjaent people who generally support strong demand and increased public spending, but are deeply worried about inflation. I'm genuinely curious how they'd answer the following questions.
First, do you think that some (not necessarily all) of the slow growth in employment and GDP after the Great Recession compared to before it was due to the ongoing effects of weak demand (hysteresis)?
Second, if weak demand caused a lasting fall in GDP and employment below the earlier trend, could a sustained period of strong demand (with GDP above current estimates of potential) reverse that damage?
Here is one way to think about the inflation we are seeing now. (a thread)
Over much of the past two years, large parts of the economy were unable to operate as normal. Schools and daycares were shut down. Airports were operating at 5 percent capacity. No one was going to the gym or eating in restaurants. You probably remember this! 1/
Though you might have forgotten the scale of it. For example: In March of 2020, fewer than 100,000 people passed through TSA checkpoints each day, compared with 2.5 million a day in 2019. 2/
A funny thing about "labor shortages" is that with the end of pandemic UI we just had a very powerful experiment on the effect of changes in work incentives on employment, and yet no one seems to be drawing any broader lessons from it.
If ending pandemic UI did not raise employment, that is not just a fact about pandemic UI. It is very informative about how important labor supply is to employment in general.
A nice example of the refusal to learn from this is the beltway journalist who admits that non-effect of expiring UI on employment is surprising, but then just goes on with his but-this-one-goes-to-11 insistence that employment problems are all about labor supply, not demand.
The only way you could see today's rising prices as the beginning of a 1970s-style inflation is if you knew literally nothing about either period except that measured inflation was high.
A sustained period of strong demand-led growth has nothing in common with a sudden, exogenous interruption in production. To see them as equivalent requires a deliberate effort to ignore everything we know about what was happening in the world.
Between 2019 and 2020, worldwide auto production fell from 92 to 78 million vehicles. What is the scenario in which this *doesn't* cause an increase in auto prices? And why should we think this is at all informative about production capacity in 2022 or 2023?
This piece on inflation by @BuddyYakov gets it exactly right: the problem we need to be addressing - in both the short term and long term - is not excessive growth in demand, but sluggish growth in supply. noemamag.com/how-to-put-the…
In retrospect it's strange that "supply side economics" has been associated with conservatives - in general, raising the economy's productive capacity is going to require a bigger public sector, not a smaller one.
The idea that there is a direct link between labor supply and employment needs full debunking elsewhere. But I'm just following textbook economics when I say: Demand determines employment, then labor supply conditions determine the wage at that employment level.