This could easily be an accounting anomaly. Productivity is defined as real GDP / hours worked. If you shrink the denominator - say, by causing 25% unemployme

This could easily be an accounting anomaly. Productivity is defined as real GDP / hours worked. If you shrink the denominator - say, by causing 25% unemployment with lockdowns - while also boosting the numerator with government spending, you get high productivity, which we had during the pandemic. If you then shrink the numerator (say, with high inflation, which translates a given nominal GDP to a smaller real GDP) while increasing the denominator (through record low unemployment), productivity will drop. IMHO traditional economic metrics are not well adjusted to an economy where output is largely independent of effort or hours worked, particularly not when measured on a quarterly basis. I work for a (remarkably slow and bloated) tech company. If I choose to do nothing other than post on Hacker News, that will not become apparent in company performance for ~2 years, because that's the median time for a project to make it out to market and start having an effect on consumer behavior. The company could lay me off and it wouldn't affect the bottom line at all, but it'd boost productivity. Conversely, if I hire a new person, I don't see significant gains in output for ~2 years, but employment and hours worked has gone up, and so productivity is down. The limiting case for this is algorithmic cryptocurrency trading with mark-to-market accounting (of which there was plenty in 2021). Here, you have computers trading virtual assets back and forth at ever increasing prices. Because prices are going up, the value of everyone's assets increases, and with mark-to-market accounting you'd show a profit. And yet nobody is employed and no real work is being done. Productivity is effectively infinite, but it means nothing.