"Are early stage investors and venture capitalists better than an algorithm? This paper suggests that the answer to that question is ‘no’. Although venture capitalists outperform the broad stock market quite significantly, most of this overperformance is driven by only a subset of their investments. An algorithm can prune out the investments that are ‘predictably bad’. This increases returns by approximately 7–41%, creating an even larger magnitude of outperformance versus the stock market. The author believes the reason for the poor investment decisions is the over-weighing of the characteristics of the founders of startups. Interestingly, I wonder if there is a repeated interaction component that might be at play – if you back the founder during a failed investment, the founder might accept you as an investor in the future when they’ll have a more successful investment."
> if you back the founder during a failed investment, the founder might accept you as an investor in the future when they’ll have a more successful investment
Apparently this is why Sequoia is backing Musk with x.ai.
Hi! Author here. This was actually my second article on wages and inflation (the first one can be found on my site - Dec 4, 2022 - Wages and Inflation). In that one I tackled the generalizability of the result, especially to the US economy. I 100% agree that for a result to hold, we need similar assumptions and conditions (basically the issue of external validity of a study). Regarding wage growth and inflation and the causality between the two, it actually has been studied quite heavily. From my previous post:
"Due to the simplicity and logical appeal of this theory [wage-price spirals], it has been heavily tested empirically. Most empirical studies to date suggest, however, that wages do not cause1 inflation. Schwerzer and Hess (2000) from the Cleveland Federal Reserve did an overview of the economic research at the time and found very little evidence supporting the idea that wages cause inflation. Only one study showed a causal impact2, while three others, and Schwerzer’s and Hess’ own work were not able to find this causality. The reason for the ambiguity in results is because inflation and wages move so closely together that attempting to separate and isolate which one causes which is not straightforward to do. Their own work focused solely on establishing the direction of causality, using what is called in economics and statistics “Granger causality”, which is a test whether the future values of one time series3 (inflation in our case) can be predicted by past values of another time series (nominal wage growth) and vice-versa. The review and analysis conducted by Schwerzer and Hess suggests that increasing wages do not cause inflation. On the contrary, evidence likely points to inflation driving increased wages."
So I would say so far the preponderance of evidence suggests there is unlikely to be causality of wage growth on inflation [in the US]. Multiple methods have shown the causal link is unlikely.
Are there circumstances in which wage growth can be the trigger of runaway inflation - I suppose there will be such conditions (how feasible/realistic they are is also a question). Have they occurred it in the US - does not appear so.
However, regarding the main issue at hand (i.e. current inflationary surge), there is even more evidence that it has not been the case, and we are nowhere near wage growth that would make us wonder about the wage-price spiral. Real wages are growing well below productivity.
Summarizing - you are 100% right that it is important to keep an eye out on external validity issues. It is also important to state the underlying assumptions (whether it is regarding the conditions of the economy or the mechanisms). It is something I am trying to get across with all of my writing and maybe could have done a better job in this particular case (although I have written a more theoretical post on the same topic, so that influenced the writing here)
Hi! Author here! You are correct - that we have real wage growth. However, it is worth nothing that productivity has gone up 3-4% over the last several quarters (see FRED data). Wages should grow by this amount in real terms at least. That's where the standard assumption of 3.5% nominal wage growth comes from (2% inflation + 1.5% productivity growth). So the 1.8% real wage growth is, in my opinion, very low given the economy.
Labor productivity data can be hard to interpret over the short term because of its definition as real output / hours worked. If, for example, the government sends all non-essential workers home while continuing to pay them through PPP, productivity will rise because hours worked has gone down (as it did in 2020), even though no layperson would say that people were more productive during COVID. Similarly, if companies are laying off people without significant declines in output (as in the 2023 tech industry), this also shows up as increased productivity. If CPI is dropping but nominal output and hours worked remain constant, this also shows up as increased productivity (real output goes up, because of lower inflation), though this scenario should be pretty rare since actual prices are a major component of nominal output.
I'm reminded of a coworker who produced absolutely nothing of value for 18 months (his project was canceled at the end without ever launching), collected ~$1.5M in total comp, and worked 3-4 hour days. His productivity was sky-high: $1.5M in economic transactions for ~1000 hours worked. It's just that the economic transactions were basically lighting money on fire. For that matter, his department of ~3000 burned around $6B over ~5 years and is currently in the business of bricking otherwise-useful devices. It just had massive layoffs, but he transferred to another department, made his bosses look very good there, and survived the layoffs.
You really need a long enough time period to know whether the shitty quality products you put out will result in fewer transactions before you can interpret productivity data. Long-term (multi-year) productivity growth usually presages real wage gains. But short-term (couple quarters) productivity growth could just be an indicator of layoffs coming, which usually means real wage cuts.
Hi! Author here! - I think I agree with the general point you're making. To elaborate: I've actually written about the underlying inflation mechanism (May 10, 2023 - (Un)intuitive Inflation).
The main model used by central banks for modelling inflation is the New-Keynesian Model. This model explains that the rate of inflation is related to the level (sic) of real marginal costs, desired (Note: not realized) mark-up (profit margin) and inflation expectations. The main element through which wages feed through to inflation is the real marginal cost. However, if current wages are below pre-pandemic real wages (which for many still are), then real marginal cost is lower, thus pushing the inflation rate down below the 2% inflation target.
During the 2021-23 inflation surge, the supply shock and bottlenecks created significantly pushed up the real marginal cost of production (things like over-time, things like it taking to produce longer than before due to supply delays etc). This has reverted bringing back down the level of real marginal cost. Which is why we are close to the 2% inflation target.
The mechanism regarding price setting - between firms and workers - is actually really nicely put by a recent Werning and Lorenzoni paper - Inflation is Conflict. Interestingly, they are capable to generate a theoretical result where we have inflation with no money! Meaning there is no monetary policy.
Hi! Author here. Yes - this is a stereotype and it is unfortunate it exists (to some extent I think the blame should be placed on the profession that doesn't speak up). Good economics, whether theoretical or empirical models, should allow for multiple outcomes. Then when it is tested with data, conclusions can be drawn on what is likely or not. Many economics questions have these conclusions - it's just a lot of theories are mentioned over again, without actually referencing the fact that a lot of research has already been done and demonstrated a theory is no longer appropriate.
https://www.nominalnews.com/p/new-research-highlights-decemb...