Institutional Investors, Spillovers, and Optimal Investment into Innovation

Firms internalize, and this really matters.

i. Cry Me a River

Imagine there’s a river which floods. 100 farmers live along it. They would all prefer if a dam were built to control its floods, and the total amount of money they would be willing to put up would be sufficient to build the dam, if you could elicit their true preferences. Of course, once the dam is built the benefits go to everyone, even those who don’t pay. It is a public good, which is to say that it is not diminished by one person consuming it, nor is it feasible to exclude people from receiving its benefits. It is costly to negotiate, and everyone has an incentive to not pay and get the dam for free. Let us say that the contributions of 60 farmers is sufficient to build it. There will still be quite plausible circumstances where negotiations break down, and no dam is built. 

Now suppose that one farmer is much more successful than the rest. He invents a secret technique for growing crops, and with the proceeds slowly buys out the rest. If he has the land of 60 farmers, it is now worthwhile for him to buy the dam all by himself. He need not consult or negotiate with anyone before building — he can privately justify providing a public good. The other 40 farmers will benefit as a byproduct of the self-interested investments of the big farmer.

ii. Sold my Soul to the Company Town

We should expect that as firms internalize more of the benefits from providing public goods, they will indeed provide more of them. Company towns, for example, would tend to provide wages more in the form of schools, infrastructure, and other such goods. Mendez and Van Patten 2022 found that areas owned by the United Fruit Company received more public goods, such as education or roads, because the firm could offer these amenities as a more efficient form of compensation for employees. Because of this, the areas granted to the United Fruit Company are persistently better off, to this very day. In coal mining towns in West Virginia, sanitation was provided by the company, and was of similar quality to similar cities. While they did not provide education, as the state did constitutionally, they did pressure local school boards to provide less segregated education. Coal companies would subsidize the education system in their district, in an effort to attract workers.

Contrary to legend, company towns paid well compared to the outside, and even if they were a local monopoly, worker welfare would be improved by companies owning the stores. (Imagine there is one employer and one supplier in town. Both are monopolies. To maximize profit, both of them will under-employ and under-produce, to the detriment of both. Combining ownership of the two reduces the total monopoly distortion). As Price Fishback wrote in 1986, “the company store’s monopoly power in nonunion districts was limited because store prices were part of an employment package offered to geographically mobile miners in a labor market with hundreds of mines.” Besides, as Braverman and Stiglitz proved, interlinking of markets is not necessary for the extraction of surplus. Rather, interlinking markets allows the landlord (in the context of agriculture) to induce greater investment in productive improvements. In a world of poor insurance markets, for example, the landlord can use their control over input prices to effectively provide insurance. (Obligatory reference to the study in Ghana quantifying losses due to lack of insurance here. I would be much obliged if anyone could point me to additional sources, preferably by dunking on me in quote tweets for not knowing the source already). :)

iii. Spillovers or Insurance?

The most important public goods are ideas. The case of the farmers should apply to index funds too, which have greatly increased in size. From 1975 to 2016, mutual funds increased from $45 billion to $13.2 trillion. This is brought to mind by a recent paper (Acharya, Johnson, Sundaresan, and Zhang 2024) measuring the impact of vaccine news on the stock market. Ending the pandemic was a really big deal [citation needed], with a large positive impact on the stock market. As holding companies get larger, it should become worthwhile to invest in innovations which you do not expect to capture directly, but which instead improve enough of the other companies you partly own as to be worthwhile. This is borne out by what evidence we have — ownership by institutional investors increases both R&D spending and productivity. 

Aghion, Van Reenen, and Zingales 2013 is the go-to study on this. They find that being purchased by index funds increases R&D spending by a bit, and R&D productivity by a lot, as measured by citation-weighted patents. Of course, index funds do not buy stocks at random, so simple observations aren’t enough. There are a number of large index funds which simply buy all the firms in the S&P 500 — inclusion in the S&P 500 causes a great surge in demand for the stock. (This method, I believe, was pioneered by Andrei Shleifer in “Do Demand Curves for Stocks Slope Downwards?” The answer, by the way, is “yes”). This is a sudden discontinuity — if research spending and patenting activity abruptly changes afterward, it is plausible that the institutional investors are responsible for it. The paper examines two theories as to why there’s an increase. First, the institutional investors may well be simply more competent owners — they press harder for the firm to be productive. More relevant to our concerns, though, is that they are less likely to let managers go during downturns which are correlated with other firms, or are simply bad luck. Investing more in research will increase the volatility of earnings, and earnings might go down through no fault of the manager. The evidence accords more with the second hypothesis — when competition is already high, managers are already disciplined by the market. People overestimate the degree to which poor outcomes are anticipatable. Institutional ownership might be like individual investors binding themselves to the mast. The correlation between institutional ownership and innovation is stronger in firms facing high levels of competition, though. The other big paper on this is Bushee 1998, which, being older and less causally convincing can be skipped over, but his findings accord with Aghion, van Reenen, and Zingales. Institutional investors are less likely to fire a manager for bad luck, and thus they encourage risk taking.

We aren’t able, from these papers, to sort out why they are more tolerant of risk. It could be because of the spillovers, but it could also be because they are simply larger, and better insured. Dating back to Arrow 1962 (and possibly beyond that, but I don’t care enough to track down the antecedents — Borges I am not), authors have pointed out the impossibility of buying insurance for your attempts to innovate. You could simply expend no effort, and be paid anyway. As the savings of firms grow larger, they are better able to self-insure. Related to this, perhaps some of the shares which index funds buy were previously owned by the executives themselves; if the executives were more risk-averse than the new owners (highly plausible) then this would also cause the increase in risk-taking. Unfortunately, no one’s data is sufficient to answer this. 

iv. What Is To Be Done?

So what should we, as a society, do about this? In either case, we should be very skeptical of concerns against bigness for the sake of bigness in firms. We do not, however, need to subsidize bigness in firms, which is different from most results about innovation. The benefits to agglomeration are captured by the investors. The concerns about index funds leading to less accurate capital allocation are likely overblown. I think they are overblown anyway, for the prosaic reason that Goldman Sachs et al is still quite capable of outbidding foolishness, but even if it were leading to inaccurate stock prices, this must be balanced against the inducement to innovation index funds provide. 

What would happen if institutional investors grow to such an extent that the entire economy becomes one firm? I am doubtful that this will happen, because people will have ideas for how to better manage companies — buying and selling equity is the right way to incentivize this. I just had a blog post on this — to elicit investments, the residual should go to the one most capable of making them. Even if it did, one must remember that the mutual fund would largely internalize the allocative inefficiency too. It could conceivably identify mutual fund members, like that of a Costco membership, but then everyone would buy into the mutual fund. If it cannot forbid resale, it would be unable to maintain this price discrimination in a lot of markets. Besides that, the usual conditions of free entry still apply! Were they to actually monopolistically price, in most markets competitors would be able to enter. I hope it is not panglossian to say that if an index fund were to own all firms, it would produce efficiently, or else not maintain its ownership of all firms!

The possibility of entry is why I’m not much concerned with the empirical evidence showing an inverted u-shaped curve in innovation. Firms become monopolies for reasons which can often hinder innovation — they’re given government protection, for example. Besides, index firms will surely not own the whole economy all the time. I predict that, in an ideal world, most firms would be owned by large institutional investors, and would cycle in and out of the control of private equity investors. They identify poorly performing firms, and the mutual funds would be happy to sell off the firms, not being able to improve the firms themselves.