good summary but i want to comment on the accounting part. generally accountants aim to create financial statements that are useful to most readers under normal circumstances. fuzzy concepts like "useful", "readers" and "normal" mean this goal is complicated and involves trade offs. accountants recognize these trade offs and financial statements include notes to try and explain how they were created and include numbers that help users create different statements for different circumstances. a few related points:

- there are several financial statements that are interconnected. changing the balance sheet automatically changes the income statement and the cash flow statement. making the balance sheet "more useful" can have the unwanted effect of making the income statement and cash flow statement "less useful"
- for example, "mark to market" on the balance sheet means creating "unrealized gains and losses" on the other statements. most users (and accountants) refer to these as "fake gains and loses" and hate them. most of the time they are not useful, but once in a while they become very important
- creating financial statements that are "more useful" during rare events like a bank run could mean having "less useful" financial statements the rest of the time

accounting rules evolve over time mostly in response to changes in what readers think is useful. bank runs are not a new phenomenon so i'd be surprised if it suddenly becomes more useful to show mark to market assets on the balance sheet. accounting rules can also be changed by regulation, but this is mostly bad, since it means implementing rules that are not useful. i think fears that accountants create rules to favor specific industries or companies at the expense of readers is unfounded. however, it's true that most readers of financial statements are in the financial industry, and it's plausible that "normal people" would prefer different rules (if they ever read financial statements).

i'm not sure if you're disagreeing with me, i too would like technology to be defined properly. i think the common understanding of technology would be confined to things like gadgets, software, automation. economists seem to use more of a civ version of the word, which includes things like writing, buddhism, plastic. if technology is not framed as "everything else", i'd be curious to know what kinds of things are used in the production function but don't fall in one of the three buckets. however, i'm more curious to know which bucket ai ends up in. i think most people would call it a technology, but i could be persuaded to stick it in capital (tool) or labor (brain).  

this is an interesting topic, thanks for the great post. i find it frustrating that economists use some non-standard definitions for common words (i'm not looking for precise definitions but consistent definitons would be helpful). in the context of the solow-swan model, i think the definitions/buckets would be something like:

input: raw materials, energy, time
capital: tools, infrastructure, liquid assets
labor: people, effort, skills
technology: everything else (including things like research, entrepreneurship, political institutions)
output: consumption goods, services

is this right? i think "technology" is the term that's farthest from its common usage.

I'll add that for competitive games, you want to encourage your opponent to use more luck and less skill, for example by praising his superior dice rolling abilities, and reframe your skillful moves as totally lucky, no way that could happen again!

In poker, you'll sometimes see a player get upset and criticize his opponent if he loses a hand because the opponent made a lucky draw or call (how could you call that, only one card could help you!) Rational players understand that having the opponent call with a bad hand is the only way they make money, so they'll praise the lucky draw (great call, I just can't seem to beat you!)

Thanks, I loved this post. Its obviously hard to distinguish luck from skill when it comes to investment returns, I don't think focusing on money and track record would be enough to convince me of skill. I think I would need to also rationally evaluate the investment strategy to determine if it can be reasonably expected to beat the market. Common strategies like "we do what everyone else does but better", or "we invest in value/growth" would probably not convince me (risk premiums, liquidity premiums, animal spirits probably affect market prices but probably not in a way that can only be reliably predicted by a "chosen one"). Strategies like "we have a data connection that's 1 millisecond faster" or "we hide microphones in every lawyer's office" seem like they have a chance to beat the market, but they're hard/expensive to pull off. With respect to covid, it wouldn't be enough to say "I knew stock prices would go down" without explaining how you knew, what your confidence was, how you determined your confidence and your prediction accuracy/confidence calibration (the lack of confidence estimates on predictions is a huge red flag for me). My sense is that when someone says "my intuition tells me stock prices will go down due to covid", what they mean is something like "the market price currently implies the risk of a global pandemic is 20%, but my intuition tells me it's closer to 50%, therefore stock prices should go down half the time and my returns should be positive after 2 or 3 global pandemics". However, even with hindsight, it seems hard to determine if the actual risk of a global covid pandemic in Feb 2020 was 20% or 50% or some other amount, so focusing on outcomes doesn't look reliable. It seems like evaluating the prediction process rather than the outcome would be an easier approach ("we hide microphones in all WHO offices" seems like it would work).