Testing prevailing hypotheses, in this sense, is a treacherous game. Number of business starts, IPOs, patents and most attempted refinements of such measures ar

Testing prevailing hypotheses, in this sense, is a treacherous game. Number of business starts, IPOs, patents and most attempted refinements of such measures are very crude measures. For example, Cowen later softened some of his positions on learning how much of the data he commented on was retail driven. Retail is relevant, but he was accidentally modeling the retail market where he intended to model the economy. Besides crude metrics, there are always implied assumptions connecting new business starts, IPOs or unicorn milestones to the bigger picture. In short, this is rhetoric. Hypotheses testing is fine figuratively, but let's remember where the pitfalls are. IMO, it's worth noting where a lot of this is coming from: market caps. These are very much monopoly-related. The bull case expectation for a 2020s Great Company is 25-40% profit margins, double digit growth or potential for such. This is possible because Alphabet, FB, Apple & such do not need to borrow or raise to grow. This fuels a fire of speculative startup investing at the lower tiers, the only place where capital is required. The problem is, IMO, that there's almost no relationship between FB's size and gross economic output, productivity, economic utility, etc. Online communication is great, but online communication isn't better because FB is 10X bigger. Google's freeware is great, but we were getting search, gmail, etc when Google was 10X smaller. Where's the relationship between aggregate revenue (nevermind profit) at JPMorgan, Alphabet or FB and economic utility? IMO Cowen, The Economist and much of this cohort/generation of economists have an economic worldview that totally lacks an adequate model for market structure/maturity and such. This has led to a lot foolishness. If you are Ford in 1921, the world is young. You get more efficient each year. Cost per car decreases, but you can outrun this by making more cars. The market is growing in both dollars and number of cars. If you are intel in 2021, same game. The world is young. Moore's law only tells half the tale. Transistors halve in price and size every two years, but the number of transistors produced increases by more than than. The more they make, the more we want. Gotta mine that crypto, train GPT-N, etc. We need more chips every year, and the cheaper they get the more we need. If productivity keeps goes up, but we don't need more and more cars... the market contracts and this is now a different game. If $Nbn worth of grocery shopping that once happened in supermarkets now happens on amazon, the markets value this is a net capital gain, because $Nbn at amazon is presumably more profit generative in the long run. That doesn't society bought more groceries, or received much more value. This is not that hard. Look to the largest companies. What are they outputting? How much of it? How good is it? What changed? I get that it's harder to add up than the number of cars coming out of factories but, it's still less abstract than viewing the world through a purely financial lens.