When Is It a Depression & Not a Rolling Sectoral Readjustment Rotation?
Noah Smith has views, views I disagree with, on what are the macroeconomic lessons from the recent tech bust:
Noah Smith: The US economy shrugged off the tech bust: ‘Jason thinks there’s a “rolling recession”… [because] much of the tech industry, especially the venture-funded startup industry, has taken a beating since the end of 2021. It’s easy and natural to extrapolate the performance of our own little corner of the economy to the whole thing.
But in fact, it’s worth asking why the tech bust didn’t drag the U.S. economy into a general recession…. It’s pretty widely accepted that the Great Recession of 2008 was prompted by the implosion of the finance industry, and most people think the bursting of the dotcom bubble had something to do with the mild recession of 2001. So it’s worth asking why the tech bust of 2022-23 didn’t have similarly negative results….
Forward interest rates… started to rise in November 2021… [and] earnings plateaued…. These… sparked a big collapse in tech stocks that lasted through the end of 2022….
Slowing earnings, falling stock prices, and the prospect of rate hikes sparked a wave of layoffs by big tech companies, which crested in early 2023.…
The carnage in medium-big tech companies also changed the game for the startup world as well…. Meanwhile, all this time, another type of tech bust was happening: the crypto crash…. The final piece of the tech bust was the collapse of regional mid-sized banks that served a lot of tech industry people and companies in California. Silicon Valley Bank and First Republic….
[But] though the tech bust moderated expectations for future growth… it didn’t really hurt actual output that much. Of course, the generative AI boom might have something to do with this…. The tech bust might have been severe enough to impact the national economy if ChatGPT hadn’t suddenly ridden to the rescue….
A final possibility, though, is that… macro models of sectoral recessions depend on network effects…. If software is a relatively self-contained production network, those linkages will be weaker…. Software might act as sort of an “equity tranche” for the U.S. economy—a high-risk, high-reward sort of icing on the cake. When it does well, it creates a ton of value, and when it does badly, it doesn’t crash the rest of the economy…
The polar opposite of the disembedded software industry is the heavily embedded housing industry. Is there an industry that has more backward links into manufacturing, mining, and so forth than housing construction. And yet the housing bust of the 2000s did not bring down the economy. Between January 2006 and November 2007 the number of privately-owned housing units started fell from a 2.27 million annual rate to 1.20 million. And yet the economy as a whole cruised along: unemployment was 4.7 percent in both months.
It was only when finance started to crash in the fall of 2007—and people started to demand to hold lots more “cash”—that the economy as a whole got into trouble.
Remember the course of events in 2007 and early 2008:
January 2007: several subprime lenders declare bankruptcy.
February 2007: Freddie Mac announces it will no longer buy the riskiest subprime mortgages and mortgage-related securities.
April 2007: The SEC loosens the net capital rule, which had limited broker-dealers and investment banks to a 12-to-1 leverage2.
July 2007: Bear Stearns liquidates two hedge funds that invested in various types of mortgage-backed securities.
August 2007: Federal Reserve lowers the discount rate to help ease concerns in the credit markets.
September 2007: The Federal Reserve lowers the federal funds rate by 50 basis points to 4.75%.
October 2007: Bank of America announces a $2 billion investment in Countrywide Financial.
November 2007: Citigroup, JPMorgan Chase, and Bank of America announce a pool of funds to purchase high-risk mortgage-backed securities.
December 2007: The Federal Reserve lowers the federal funds rate to 4.25%.
January 2008: Bank of America announces it will buy Countrywide Financial for $4 billion in stock. Fitch Ratings downgrades Ambac Financial Group’s insurance financial strength rating to AA, Credit Watch Negative.
February 2008: The Federal Reserve lowers the federal funds rate to 3.00%.
Up until February 2008 the Federal Reserve thought not only that the ongoing housing bust was not going to bring on an economy-wide recession, but also that it was successfully managing the rough patches in finance. It was only in February 2008 that the Federal Reserve shifted to emergency-policy mode. It was only in February 2008 that the Federal Reserve, at least, believed that recession would come and was in fact here.
So I reject Noah Smith’s third possibility: that the software industry is self-contained without a great many backward linkages does not, in my opinion, matter much.
So what does determine what kinds of sectoral busts are followed by economy-wide downturns—by the coming of a “general glut”—and what are not?
I do not think that looking at sectoral linkages actually gets us very far. Oh, it gets us some of the way: after all, a fall in the final demand for a sector of 50% will mean that twice as many people will have to find new jobs as work in the sector if backward linkages are threefold, as opposed to just half as many people as work in the sector having to find new jobs if there are now backward linkages at all. The market dance of reknitting a new division of labor in response to a shock does depend crucially on the extent and sophistication of the division of labor.
But doing this market dance is one thing that competitive markets are good at, as Jean-Baptiste Say (1767-1832) was perhaps the most eloquent at noting.
Back up to 1803. Earlier, Jean-Baptiste Say was the first person to write down his thoughts on this issue. Say had been special assistant to Tom Paine’s friend and France’s Girondist Secretary of the Treasury Etienne Claviere—who was then purged, arrested, imprisoned, and executed by Maximilien Robespierre’s Mountain Party.