A Theory of Grift

Plus! Buy The Change You Wish to See in the World; Schmuck Insurance; Uncertainty; Block Trading; Ads

In this issue:

A Theory of Grift

Google Ngram Viewer, the world's canonical source for all slow-moving zeitgeist shifts, indicates that there's been a bull market in calling people "grifters" since the mid-1980s. Google Trends shows a completely different year-to-year trend trend, but also indicates a rise in the number of people searching for the term, perhaps originally catalyzed by the memorable tale of the Hipster Grifter (this coincides perfectly with the April 2009 spike), and further ebbs and flows as either specific grifters become famous of specific famous people get accused of some kind of grift.

What do we even mean by grift? Why is it so abundant? What, if anything, should we do about it?

The heart of a grift is that it's a promise that could be true: the fitness influencer who says you'll look like a model in four weeks is obviously lying, but the one who offers exercise guides and meal plans that put you on pace for steady weight loss, and who has the before-and-after pictures to prove it? They could easily be telling the truth. Products that fit that description exist; some of them work! But often in this category, they're simply not very good, or not much of an improvement on some diligent Googling and prompt engineering.[1]

So the heart of a grift is that you get something that, in a technical sense, is what you paid for, but that is also not worth what you paid for it.

In politics, an increasingly grifty arena, the pitch is once again plausible, just not going to happen. One notorious example of this is fundraising appeals that say that donations will be matched, doubled, quadrupled, etc. despite campaign finance limitations on actually doing this. (Is there really a cohort of four people who are willing to donate the maximum, but only if the latest fundraising email pulls in at least one pay-the-max donor? That's the implicit claim behind matching donations!) There are grifty policy ideas, too, usually the ones that are appealing to a party base but unworkable nationally.[2]

Finance has its own ecosystem of grifts. Leaving aside things like ponzi schemes, which are outright scams, the broad categories of finance grift are:

  1. A fancy-sounding high-fee wrapper on something that is trivial to implement manually, and
  2. A strategy that looks good in backtests because it has some kind of nonlinear blow-up risk.

There's a lot of both. In the aggregate, one of the biggest historical finance grifts was running a strategy that implicitly replicated the S&P 500, but charged much higher fees than an index fund. As index fund fees have declined—to as low as zero for some categories!—it's gotten harder for firms to get away with this, at least for simple products. But for more complex ones, there's still room to take a straightforward options strategy and overcharge for it.

The options tie into the second kind of grift, pitching a product with good historical returns while declining to highlight its risks. One notorious instance of this was inverse volatility exchange-traded notes, which would systematically bet that volatility would decline. This is a positive-carry trade, i.e. most of the time, you can bet against volatility at greater than fair value, and make money even if nothing changes. If volatility declines, returns are great—you sold something at 17 when its fair value was 14, and then fair value went down to 13, so you made some extra money. But there's a good reason it pays to sell volatility on average: sometimes, it really, really doesn't: in early February 2018, volatility more than tripled in a few days, partly because of a feedback loop with the products themselves, and the inverse-volatility trade blew up.

One reason finance can be prone to grifts is that it's entirely possible to drift into one. Plenty of reasonable professionals treated betting against volatility as part of their toolkit, and an exchange-traded product made this easier. But retail investors who looked at the chart of these products wouldn't necessarily know that 1) they were selling insurance over time, 2) the size of the inverse-volatility funds was actually affecting the price of volatility, adding to their returns in the short term but lowering the insurance premium, and 3) that a sufficiently large move could wipe investors out.

What are the mechanics of a grift? One way to think of it is that it's a particularly pathological example of a company with high churn. Of course, churn rates vary by industry; people stick with checking accounts for a long time—far longer than they do for app-based dating sites. But within the same industry, some companies optimize for keeping customers forever, while others focus on acquiring them quickly, running through them profitably, and repeating that cycle.

So grifts tend to target the middle of whatever the relevant bell curve is. There are a lot more average people than non-average people, so the market is bigger. And their averageness makes it easier to reason about their motivations.[3] Targeting the average also enables the plausibly-deniable part of the grift: the fitness influencer really did sell you a product that meets the specifications, the inverse-vol note did disclose in its prospectus that it could blow up.

One possible reason grifts seem to have proliferated is elite overproduction, specifically elite overproduction of extroverts. There's an ongoing debate over whether or not people skills are undervalued, and perhaps for many people they are, but it's hard to deny that there are many, many more ways for someone who doesn't like social interaction much to get rich. If ads and sales are on the same continuum, then the world's best salespeople are engineers, data scientists, and product managers.

Meanwhile, the growth of software corresponds to a drop in the number of required human touchpoints for a given transaction. When travel was booked by phone or in person, there was a direct financial reward to being a good conversationalist in the travel business; rhapsodizing about the beauty of a beach was a good way to upsell customers on a nicer hotel a bit closer to it, and sussing out whether someone was more interested in beaches, landmarks, or bars meant figuring out exactly what to pitch them. All this work is now silently and efficiently happening on the backend of the big online travel agencies, with no human interaction required.

The pre-Internet white collar economy was basically a universal job guarantee for personable people. That's increasingly going away, to be replaced with a more neoliberal attention economy with a more extreme distribution of outcomes. And one result of that is that it's possible to craft an online persona as the top of funnel for some product.

This is common: individual creators have converged with media brands (some of them are literally A/B testing facial expressions). But the difference comes back to extroversion (as well as moral flexibility): a grift achieves a healthy ratio of customer lifetime value to customer acquisition cost by making sure customers pay as much as possible upfront and are cheap to replace. So an extrovert who thrives on having a new audience all the time will also prosper from a business model where everyone who is burned out by working with them is quickly replaced by someone new. This shows up in the high-fee finance grift in a very specific way: a high cost structure for a fund with a commoditized offering can fund a sales force that persuasively argues it's not commoditized after all.[4]

In this model, grift will only get worse over time, as it's a function of media distribution that increasingly allows people to build a following—80th percentile knowledge in some domain rounds up to world's leading expert if you're far enough below that percentile—and a lack of more legitimate opportunities for charismatic people. But two forces push back against this:

  1. Automation is coming, even for the grifters. Real-world influencers are more charismatic than digital ones for now, but that won't stay true for very long. And digital influencers can be endlessly A/B tested both based on their core message and the variants for specific audiences. AI-generated video will, among other things, lead to a golden age of affinity fraud.
  2. Some software products replace human labor, some complement it, and some create entirely new categories. There might be some limited sense in which the Call of Duty franchise competes with bowling, but it's mostly a brand new thing. As the number of products proliferates, and as AI makes it faster to produce a customized one for different domains, the labor market's demand for extroverts will rise. Cheaper-to-produce software creates more demand for proof-of-stake in the form of a human being who pitches the product and offers support. Some of this can happen digitally, and can be replaced by avatars, but business travel is still recovering, and on their earnings call last week Delta specifically called out tech companies as former travel laggards who had finally started spending again.

So, for the moment, we're probably stuck with some baseline level of grift. No one is immune, because everyone is part of the mass middle of the bell curve for something. So for now we'll still see weird stuff like Goop and Infowars selling identical supplements with different labels. And, of course, once grifters learn to specialize, they'll stick around for a while: last June, the Hipster Grifter was sued for six months of unpaid rent.

  1. One useful feature for ChatGPT is giving serviceable workout recommendations subject to constraints. So something like "what's a good routine for arms and chest while I recover from a bench press injury?" you will get decent answers. This is especially helpful because there's little middle ground in workout advice from "what advice would you give a sedentary beginner if you were mostly worried about getting sued?" and "what's a good way for an experienced athlete to eke out a 0.5% performance improvement?" ↩︎

  2. One meta problem here is that in theory, someone who gets elected because they mobilize the party base to vote for them, dominating the primaries and eking out a national victory because of turnout, could enact such policies. But making a big unpopular change involves the other kind of politics, where you make lots of compromises and cut lots of deals. And that sort is the comparative advantage of party insiders who've been in office for a long time, i.e. exactly the sorts of people who don't win on some kind of insurgent platform. ↩︎

  3. Outright scams tend to target the tails: at one end, they're going after small customers who aren't sophisticated enough to know that 30% a month is not a realistic rate of return, but at the other end, they're targeting buyers who don't want to feel embarrassed asking the obvious questions. Bernie Madoff was always a few diligence questions away from seeing his whole scheme unravel, but by far the most awkward question you can ask during a fundraising pitch is: "By the way, is everything you just told me a complete lie?" ↩︎

  4. So asset management, like credit cards, has two fee equilibria, one in which companies compete on cost, and another in which they compete on distribution funded by a much higher cut of the transaction. ↩︎

Diff Jobs

Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:

Even if you don't see an exact match for your skills and interests right now, we're happy to talk early so we can let you know if a good opportunity comes up.

If you’re at a company that's looking for talent, we should talk! Diff Jobs works with companies across fintech, hard tech, consumer software, enterprise software, and other areas—any company where finding unusually effective people is a top priority.


Buy The Change You Wish to See in the World

On Friday morning, along with its earnings release, BlackRock agreed to buy infrastructure private equity firm Global Infrastructure Partners for $12.5bn. BlackRock has had some solid acquisitions in the past, but their previous big acquisition, Barclays Global Investors, was partly a way to achieve scale in the intrinsically scale-driven game of managing indexed assets. GIP is a very different model, funding the acquisition or construction of ports, airports, and energy assets.

What's notable about this deal is that the big source of assets and potential upside comes from trends that BlackRock itself has pushed in other areas. From their earnings call:

If we are going to decarbonize the world, the amount of capital and infrastructure is going to be very necessary. If we are going to be more and more reliant on interconnectivity worldwide, the need for the upgrading of ports is vital. As more and more human beings grow into a middle-class lifestyle, the demand for air travel grows dramatically, the need for high-quality airports grows dramatically.

Passive investing is implicitly a bet on globalization, which leads to economic convergence. And BlackRock has been pressuring portfolio companies to adopt ESG standards for years. So this fits in with their broader strategy: reduce the world's aggregate capital expenditures for carbon-intensive assets to create more demand for carbon-mitigating assets, most of which have a longer duration than dirtier energy sources; create global portfoliosto slowly apply pressure on the rest of the world to conform to American standards for corporate governance and property rights. BlackRock is not the world's most influential organization, but at their scale it's impossible for them not to have an impact—and that impact is big enough that they can do deals to take advantage of it.

Schmuck Insurance

Sports broadcasting is obviously worth a lot, and at present it's very non-obvious where that value accrues: dumb pipes, legacy cable networks, direct-to-consumer sports apps, or league- or team-specific subscriptions. The uncertainty means people are looking for deals, but it also means they're in the market for schmuck insurance. So, the NFL is considering a deal in which it will sell some of its online media assets to ESPN, in exchange for ESPN equity. When company growth within an industry is a function of pricing power instead of rising unit volume, reaching the limit of that pricing power creates immense uncertainty—sometimes, one part of the value chain will grow and eat most of the margin, while in other cases aggregate profits will shrink fast as everyone fights for share. A merger like this mitigates some of that uncertainty without eliminating it, but in the end everyone involved is in the market for a bit less uncertainty.


Snippet Finance has a nice chart of analysts' uncertainty about company earnings, as captured by the dispersion in estimates. The short-term spikes in the chart have a simple driver: when there's an industry downturn, some analysts cut estimates fast and others take their time, so average earnings estimates going from $1 to $0.50 can represent half of the analysts cutting their estimates to $0.00 while the other half haven't gotten around to it yet. (That's one reason for the big increase in early 2015: everyone knew oil companies would take a giant earnings hit from the collapse in oil prices, but not everyone published this promptly.) What the chart highlights is that uncertainty has gone up post-Covid, and mostly stayed there. There's been a lively debate since about which companies' economics permanently reset higher (like sporting goods retailers seeing their net margins rise by ~50%) while others seem to have reset lower (Delta airlines' headcount is 10% higher than it was pre-pandemic, at a similar business size). Every quarter is either evidence that things are returning to normal or evidence that we're in a new normal.

Block Trading

Morgan Stanley has settled a complaint with the SEC over block trading. The way block trades are supposed to work is that a seller wants to move a large amount of stock with minimal market impact, and instead of dumping it on the market, they sell it to a counterparty (often Morgan Stanley) who then promptly sells to their customers. The bank's value-add is that they know more about what market depth looks like conditional on a large sale, and is willing to take some risk that their estimate is wrong. Suppose there's a stock trading at $42, and a seller wants to get rid of 10m shares. Perhaps they estimate that the price impact of selling would knock the stock down to $40, but Morgan Stanley figures they could sell carefully enough to only push the stock to $41. Morgan Stanley offers $40.50, makes a $0.50/share profit, and saves their customer $0.50/share, too.

One trader at Morgan Stanley discovered a fun variant on this trade: leaking information about it to potential buyers so they can profit from that initial drop (and, presumably, this customer will owe Morgan Stanley a favor and buy into future deals that are harder to sell). This is a straightforward wealth transfer from one client (the seller) to another client (the prospective buyer), and increases the market impact of the trade—once the hedge fund knows an offer is coming, their incentive is to short the stock aggressively and then cover by buying into the offering. Once the offering is committed, it's hard to back out even if there are fluctuations, and Morgan Stanley's customers probably don't audit every block trade Morgan Stanley ever did to see if there's some kind of pre-trade price impact indicating a leak.

There is an obvious temptation here, because the block trader has basically been told "there's a specific time at which this stock will move by a predictable amount, and while you yourself can't trade on it, somebody else would clearly owe you a favor if they did." The people doing this trade were pretty aggressive about it: one part of the SEC document references a case where one fund was 89% of the trading volume in a stock the afternoon before an offer. And Morgan Stanley even pitched customers on giving them exclusive rights to do the deal—by citing the price impact of a leak they themselves were responsible for! (It was, in fact, the same transaction where one fund tipped off by Morgan Stanley did almost all the trading volume in the stock.) The real trouble with this, from a white collar crime optimization standpoint, is that once someone goes looking, it's very easy to assemble evidence of exactly what happened: a straightforward illicit trading strategy is straightforwardly easy to spot after the fact.


The Diff wrote last year about how common it is for companies to either have an ad-based revenue model or to include ads in an existing model. The trend continues: here's an interview with the Chief Digital Officer of Planet Fitness about their new ad network. (Even someone writing in an ad trade magazine is surprised that PF is getting so much into ads.) Every in-person business is, from an advertiser's perspective, an audience that has been pre-selected based on some traits that predict other kinds of spending, and that is somewhat captive because the venue can control the local media environment. Naturally, this model works even better with a membership/loyalty program that makes it possible to track who saw which ad, and to retarget them later through an app. Planet Fitness is not one of the companies that will end up switching to an entirely ad-supported business, of course. But the gym business is already built on paying fixed costs in order to achieve high-incremental margin revenue. In fact, this is a sort of usage-based hedge: the customer who signs up because of their New Years Resolution in January and stops showing up in March is pure margin until they inevitably cancel; now, the customer who insists on continuing to work out throughout the year can contribute revenue, too.