[Week 44 of 2025] Access, Algorithms, and Anxiety

Welcome back to Price and Prejudice with a few musings from Week 44 of 2025.

Democratizing Private Markets

This Bloomberg piece discusses the latest push by large private equity firms — Blackstone, Apollo, Carlyle, KKR — to sell their funds to ordinary investors. What began as an institutional asset class, structured around university endowments and pension funds, is now being repackaged for 401(k)s and retail advisory channels. The campaign imagery is familiar: talk of “democratization,” slick marketing, and sports sponsorships designed to make private markets feel as approachable as index funds. Beneath the branding is also a regulatory shift: the U.S. Department of Labor has opened defined-contribution retirement plans to private equity allocations, creating a $13 trillion opportunity that the industry is rushing to capture.

The economic question is what, exactly, democratization delivers. In general, when studying a market that does not yet exist, there are two options. One is to build a model — write down some preferences, frictions, and budget constraints — and see what equilibrium pops out. The other is to look at the few cases where it’s already happening and extrapolate. We don’t have much of the first, but we do have a bit of the second: in a recent paper, my coauthors and I look at 65,000 PE investments made by individual investors. On average, they performed about as well as institutions, and even beat public markets a bit. The catch is that the distribution is steep: the richest investors outperform the least wealthy, mostly because they worked with better advisers. Once you subtract the extra advisory and platform fees, the outperformance for smaller investors mostly disappears.)

So the early data suggest that democratization mostly means intermediation. The new entrants don’t get better deals but instead they get access to more layers of fees. But this is only the beginning. The "general equilibrium" effects – meaning, roughly, what happens once everyone reacts to everyone else – could be more interesting than the direct returns. If private capital floods in from retail channels, more firms may choose to stay private longer, knowing they can raise money without public disclosure. At the same time, a new layer of financial plumbing may emerge: platforms that pool small investors, standardize reporting, and compress fees the way ETFs once did for mutual funds. In theory, those intermediaries could make private equity cheaper and more transparent.

Corporate Compliance as a Turing Test

This FT article discusses how the oldest workplace crime in the book – faking receipts – has discovered AI. Expense software firms now say roughly 14 percent of fraudulent documents are AI-generated, up from zero last year. What used to require Photoshop and patience now takes a few words typed into ChatGPT: “make a dinner receipt from Nobu, total $186.72.” Corporate compliance has become a Turing test, and the question is how to curb this behavior.

Economists have a simple framework for this, thanks to Gary Becker’s 1968 classic, “Crime and Punishment: An Economic Approach.” Becker’s insight was that crime is just another labor-leisure choice under uncertainty: you commit the act if the expected utility from cheating exceeds that from behaving. That expected utility depends on three things — the gain from fraud, the probability of getting caught, and the punishment if you are. Society’s job is to minimize the total cost of crime: not just the losses, but the resources spent on enforcement and the pain of punishment itself. Hence, Becker’s famous trade-off — you can deter crime by raising the likelihood of detection, or by increasing the penalty, whichever is cheaper. Historically, catching fraud was expensive (think auditors poring over receipts), so firms opted for mild penalties but frequent checks.

AI reverses the cost structure. Generating fake receipts is now nearly free, which increases the temptation dramatically. But detecting them has also become cheaper — machines can scan metadata, flag patterns, and identify inconsistencies faster than any compliance team. We’ve moved from “probability versus severity” to “algorithm versus algorithm.” If the expected cost of cheating depends mostly on the model’s precision, then deterrence itself becomes a software feature. In practice, that may mean companies shift toward automatic penalties: if the system flags your receipt, reimbursement just doesn’t go through. No trial, no HR drama, just code.

There’s a broader equilibrium question here. If AI reduces the marginal cost of both fraud and detection, we might end up in a steady state of constant low-level dishonesty — small lies caught by small machines. Some firms will tolerate a background rate of AI-generated noise, just as retailers price in shoplifting. Others will overcorrect, treating every reimbursement as a potential forgery. Either way, the moral dimension collapses into calibration: how many false positives can you afford? Becker reimagined crime as an optimization problem; AI has simply made the optimizer autonomous.

The Doom Spenders

Every generation invents a moral vocabulary for its debts. For the boomers it was “prudence,” for millennials “hustle.” For Gen Z, it’s “YOLO”—a term that began as a joke and somehow became a balance-sheet strategy. A recent piece profiles Canada’s “doom spenders”: young adults who, facing rents that swallow their paychecks and home prices that mock their savings, are borrowing with abandon. Consumer debt among Gen Z jumped 30 percent last year, much of it through buy-now-pay-later apps that let you finance groceries or concert tickets in installments. The story reads like sociology but doubles as a macro forecast: when the future feels unaffordable, the present becomes the only investment horizon.

A rational economist might sympathize. If you expect your long-run income prospects to stagnate and real asset (e.g. housing) prices to keep rising, then the marginal utility of future consumption falls. In plainer English: if tomorrow looks worse than today, spending now is a coherent, even optimal, response. Debt becomes a hedge against hopelessness. In that sense, doom spending isn’t irrational; it’s intertemporal realism — a consumption-smoothing strategy under collapsing expectations. The problem isn’t that people miscalculate; it’s that the calculation is correct.