The financialization of tech

China is racing ahead in semiconductors and drones while we're building better ways to separate retail investors from their savings.

Fifteen years ago, if you wanted to make it in America — if you wanted to get in on the ground floor of that smooth upward escalator — you went into the finance industry. Lots of people were upset about this. Our best and brightest, they said, were dedicating their effort and their intelligence to finding new ways to sell each other complex financial products instead of building the technologies of tomorrow. When the financial house of cards collapsed in 2008, everyone who complained naturally felt vindicated.

In the decade that followed that crash, it seemed like the problem was fixing itself. Wall Street crashed, and the Dodd-Frank legislation tamed the high-flying sector. Talent streamed out of investment banking and headed west, to Silicon Valley and the burgeoning tech industry. Google and Amazon and Facebook were the new Goldman and Morgan Stanley and Merrill Lynch, and startups were the new hedge funds.

And it seemed to work to society’s benefit. Sure, some folks groused about Facebook ads and the occasional goofy startup. But the second tech boom gave us electric cars that actually sold, smartphone cameras good enough to make professional movies, e-commerce that sustained us through the pandemic and helped create a small business boom, video conferencing and workflow apps that made remote work a reality, convenient payment apps, cheap rockets that can take off and land, better batteries, mRNA vaccines, and much more. We had successfully redirected national resources from finance to tech, and we were getting tangible benefits. And most Americans appreciated those benefits.

During the later years of the second tech boom, however, I started to notice another trend. Tech companies were increasingly getting in on the finance game. Not so much in the lending space, but in payments and trading. Payments are a utility — the plumbing of the financial world — and here the tech industry was producing a lot of real value (a prime example being Stripe, which lets me receive payments for this blog). But trading — which along with lending drove much of the finance industry’s expansion from 1980 to 2008 — is another matter entirely.

In recent years, trading has taken an increasingly prominent place in the fintech world. Platforms like Robinhood have onboarded a whole ecosystem of retail investors (that’s finance-ese for “regular folks”) into the markets. And cryptocurrency and web3 are hard at work creating new assets for people to trade, and new markets in which to trade them. Both fintech and crypto/web3 saw enormous, record inflows of VC funding in 2021. Here’s TechCrunch on the fintech boom from January 2022:

In 2021, global fintech funding jumped to a new record of $131.5 billion across 4,969 deals. That compares to $49 billion across 3,491 deals in 2020. As you can see, the pace at which capital was invested into fintech startups in 2021 grew much more rapidly than total deal count, leading to larger rounds on average. The CB Insights data that we’re citing here also helps put the pace of fintech investing into context compared to its peer startup groups, with financial technology companies raising one in every five venture capital dollars last year, or some 21%.

And here’s Blockworks on the crypto boom, also from January 2022:

In 2021, venture capitalists invested over $33 billion into crypto and blockchain startups, according to a report by Galaxy Digital…

About $22 billion, or 67%, went into rounds with deals over $100 million. Companies that focus on digital asset trading or building in Web3 raised the most capital overall, the report found. 

And of course these numbers are proxies for the real resources being redirected toward trading tech — smart engineers, bold entrepreneurs, high-end computer chips, and so on.

Of course these numbers will come down — along with everything else — in the wake of this year’s tech crash. But when I look at the redirection of the technology industry’s resources toward trading, I see an uncomfortable echo of the financialization boom of the 2000s. Finance — especially behavioral finance — gives us plenty of reasons to question whether trading, past a certain point, is simply a way to separate foolish retail investors from their hard-earned savings. And it’s disturbing to think that while China is leaping ahead in semiconductors and leading the world in drone technology, America’s brightest minds are spending their time and energy thinking of new ways to trade tokens back and forth.

Bilking retail traders — the dark side of behavioral finance

Economists generally give market activity the benefit of the doubt — if they see a bunch of people buying and selling stuff, they start off with the presumption that there’s probably a good reason for it. When it comes to trading in financial markets, some economists assume that it’s all part of a natural and healthy process of “price discovery” — that traders have to trade assets back and forth a lot as a way of figuring out what they’re really worth. Here’s Kenneth French from 2008:

In aggregate, active investors almost certainly improve the accuracy of financial prices. This, in turn, improves society’s allocation of resources. Thus, my estimate of the cost of active investing also measures society’s cost of price discovery.

But when it comes to the specifics of how this process works, economists are generally at a loss. There’s a very famous result in financial economics called the “no-trade theorem”, which says that adverse selection — the knowledge that if you’re buying an asset because you think it’s undervalued, someone else is probably selling it to you because they think it’s overvalued — should prevent most trades from happening. People should have a natural wariness about assuming they’re smarter than the market, and this should make them trade only rarely.

And yet we see a ton of trading in financial markets. What could overcome adverse selection enough to cause all this trading? One very natural answer is “behavioral biases”. Lots of investors are not hyper-rational “homo economicus” types — they’re overconfident, or their attention is limited, or they follow the herd, etc. Behavioral finance researchers have done decades of yeoman’s work documenting these biases and showing how they lead retain investors to consistently underperform the market. The best survey of this research I know of is Barber and Odean’s “The Behavior of Individual Investors”, from 2013. From the introduction:

A large body of empirical research indicates that real individual investors behave differently from investors in [fully rational] models. Many individual investors hold under-diversified portfolios. Many apparently uninformed investors trade actively, speculatively, and to their detriment. And, as a group, individual investors make systematic, not random, buying and selling decisions…[M]any individual investors seem to have a desire to trade actively coupled with perverse security selection ability.

In other words, most regular folks (and yes, this almost certainly means you) are not very good at picking stocks, or tokens, or whatever. They tend to lose money. And who gets the money? Two groups of people — sophisticated investors, who know how to defeat the retail folks in the trading game and take their money, and market makers, who sell retail folks the rope they use to hang themselves. Market makers, of course, include the trading platforms where buyers and sellers come together; the house always takes its cut.

Economists have documented plenty of instances of retail traders getting burned by sophisticated investors. My favorite is a 2004 paper by Brunnermeier and Nagel that documents how hedge funds were generally able to sell out at the top of the 2000 tech stock bubble, leaving other traders holding the bag. Remember all the people who got into day trading in the late 90s? Almost all of those guys went bust, and the wise guys walked off with their money.

Some economists used to argue that retail traders who suffer from irrationality and biases will just lose their money and exit the market. And so in fact it went, with the day traders of the 90s. But the thing is, there are always new suckers being born every minute, and by the time 2020 rolled around, there was a whole new crop of eager day traders who had no recollection of the disaster that befell their predecessors two decades earlier. And so the cycle began again, with a new generation of retail muppets downloading Robinhood and FTX and betting their hard-earned life’s savings on the latest meme stock or shitcoin.

And of course the exact same thing happened as last time. Here’s Barber et al. (2021):

Using data from Robinhood, we find that Robinhood investors engage in more attention-induced trading than other retail investors. For example, Robinhood outages disproportionately reduce trading in high-attention stocks. While this evidence is consistent with Robinhood attracting relatively inexperienced investors, we show that it can also be partially driven by the app’s unique features. Consistent with models of attention-induced trading, intense buying by Robinhood users forecast negative returns. Average 20-day abnormal returns are -4.7% for the top stocks purchased each day.

It would be a strange world indeed in which these biased, losing retail traders were necessary for some economically efficient process of “price discovery”. It’s much more likely that day trading simply facilitates the redistribution of dollars from the pockets of regular shlubbs into the pockets of smart people.

And frankly, I heavily doubt that that redistribution constitutes a socially efficient deployment of the technology industry’s human, financial, and physical resources.

Web3 and retail extraction

Day trading is bad enough when it’s just stocks. Now, however, crypto has given technologists and financiers a way to create infinite new tradable assets, in the form of tokens. And it has given them new markets in which to trade those assets — not just exchanges like FTX and Binance and Coinbase and NFT marketplaces like OpenSea, but also web3 applications like Axie Infinity with internal marketplaces.

I am not an inveterate web3 skeptic or detractor (like, say, Liron Shapira). I believe that web3 creators will probably find useful applications for blockchains (unique online identities and DAOs are two ideas I think are promising, for example). But the vast majority of the web3 stuff that’s being created and envisioned right now seems to rely heavily on trading as part of its business model.

Here are some examples. In a recent thread, Tascha Che of Tascha Labs listed a bunch of hypothesized web3 applications:

Tascha @TaschaLabs

35 business ideas of how web3 could improve or disrupt the most successful companies:

Essentially all of the ideas she lists — social media platforms, search engines, warehouses, etc. — are things that work perfectly well without blockchains. A social media platform can easily pay you to post without a blockchain. A company can easily pay people to warehouse goods in their garages without a blockchain. The only thing that Tascha’s proposals add to these already-existing businesses is tokenization.

Tokenization means that each of the economic activities Tascha proposes would be associated with a token, which is an asset whose value traders can bet on in online financial markets. Her proposed web3 search engine would have its own token that would allow owners to access the search data. Her proposed web3 social network would have its own token that would allow owners to buy ads on the platform, and so on.

Without trading, this is just a very inefficient way to do everything. You can pay for search data in dollars. You can pay for social media ads in dollars. It’s very easy, and companies like Stripe are making it easier every day. There’s a good reason we don’t already pay for Facebook ads in FacebookBux or Google data in GoogleBux — it’s easier to just pay in dollars. There’s no reason for every business to have its own internal corporate scrip.

Unless that scrip also trades. If the scrip trades, then you have a second way of monetizing your service — not through selling actual valuable products, but through retail traders overpaying for your tokens. The retail muppets overpay because (as we saw in the last section) they’re overconfident and less-than-competent. And you, the creator of the token, get to dump that token on the retail muppets and transfer some of their hard-earned life’s savings into your own pocket.

As a wonkish side note, it’s much easier for the original seller of a token or asset to extract value from retail traders if the asset is in a thin, illiquid market where it’s hard to short-sell. In liquid markets with easy short-selling, the short-sellers can come in and trade against the over-optimistic retail traders and take their money. But when shorting is hard, the price of the asset is set by the most optimistic traders (actually it’s even higher, because of the option value of being able to re-sell it to optimists over and over!), so the original seller of the token can make out like a bandit. That’s why these thin, illiquid crypto token markets are such an attractive way for entrepreneurs to extract value from retail traders.

In other words, all the web3 stuff that’s just “X to earn” or “X, but tokenized” is actually just a play to dump tokens onto retail traders — to use normal shlubbs’ behavioral biases to extract money from them above and beyond the value of the actual web3 service.

Of course, this source of revenue eventually runs out, just as the day trading boom of the 90s ran out. The aforementioned web3 skeptic Liron Shapira recently noted how Helium — a company that lets people set up their own wi-fi hotspots and charge for their use — has become much less lucrative for users ever since retail interest in the company’s token started to dry up:

Liron Shapira @liron

These folks maintain false hope of positive ROI. They still don’t realize their share of data-usage revenue isn’t actually $20/month; it’s $0.01/month. The other $19.99 is a temporary subsidy from investment in growing the network, and speculation on the value of the $HNT token.

This kind of thing often gets called a “Ponzi scheme”, but Stripe’s Patrick McKenzie notes that it’s actually a bit different. Yes, some of these web3 applications get inflows of money from people paying up-front fees to buy into the system, and yes this can enable Ponzis. But a lot of the money just comes from retail traders overpaying for the companies’ tokens.

And note that all this tokenization is, on some level, a regulatory arbitrage. There is no technological reason you couldn’t tokenization without blockchains, but there is some degree of existing regulation that the U.S. government has set up to try to protect retail investors from being able to trade too much in traditional financial markets. Financial regulators have so far had trouble extending these regulations to the crypto/web3 world, meaning that the rules applied to stocks aren’t all applied to tokens, even though a token is just a slightly different type of security. So web3 enables retail trading — and therefore the extraction of value from overconfident, less-than-competent retail traders — that traditional financial markets restrict.

As with trading platforms, I question whether this activity constitutes a socially efficient deployment of the technology industry’s human, financial, and physical resources.

Fiddling while China builds

I’ve talked a lot so far about the benefits of trading technology, which I think are few. But what about the costs? In an age of cheap capital — which 2020-21 definitely was — what are we really losing by throwing money at these technologies?

As always in economics, the true cost is opportunity cost. We didn’t just throw money at trading technology, we threw real resources at it as well. Tons of startups hired tons of talented engineers to work on these things. They absorbed the entrepreneurial energies of a lot of bold smart risk-takers. They also used a bunch of high-end computer chips and electricity, but the human resources are really the scarce, precious thing in the innovation economy.

And while we were directing our smart and bold people to find new ways of taking money from retail traders, China — our great national rival — was directing its own smart and bold people toward industries like semiconductors, drones, and electric vehicles. Recently, China’s SMIC announced a 7nm chip, created despite heavy U.S. export controls on that company, that moves China closer to parity with the U.S., South Korea, and Taiwan. As for drones — which the Ukraine war has showed are a core platform of modern warfare — China is leading the pack. And Chinese state-backed EV companies are Tesla’s stiffest competition.

Now, China’s methods of directing resources toward these industries are not efficient and not something we should copy. In addition to showering the favored sectors with money and support, China’s leaders cracked down hard on other tech industries, and on finance and private education, in the hope of diverting human resources toward the industries it likes. This is likely to simply chill ambition and risk-taking throughout Chinese tech.

But even without copying China’s methods of resource allocation, I still think we’re not powerless in deciding which tasks our engineers and entrepreneurs direct their lives toward. VCs like Katherine Boyle have been fighting to divert resources toward companies that benefit national security and physical infrastructure. During the Obama administration, the government supported green technology companies and got some huge wins, most notably Tesla. Both the private sector and the public sector have levers they can pull here.

Now I don’t want to get too melodramatic about this. The numbers quoted above suggest that no more than 20-25% of VC money is going into trading tech, and probably less than that. Nor is VC money close to the only source of funding for innovation — there’s also big companies, which do a surprising amount of innovating, and of course there’s the government too. We are not seeing a return to the days of 2007, when it seemed like every smart young college grad and math geek in the country was headed to Wall Street.

Still, though, it gets me down to see the U.S. technology industry start to turn toward the sort of extractive financialization that it seemed to replace a decade ago. Not just because I’m sad that a lot of retail traders will lose their nest eggs — keeping overconfident, less-than-competent people from losing their money will always be an uphill battle — but because extracting money from retail traders simply doesn’t strike me as much of a technological achievement.

---