Image: 123RF/ lucadp
Image: 123RF/ lucadp

Why banks?

Here is a hilarious paper from Juliane Begenau of Stanford and Erik Stafford of Harvard called "Do Banks Have an Edge?" No, is their answer. You'd be better off just passively buying Treasuries than buying the complicated mix of stuff that banks actually own:

We construct a simple buy-and-hold portfolio that each month purchases a six-year maturity UST and holds it to maturity, such that the three-year average maturity of this portfolio matches that of the aggregate banking sector. This is a highly conservative benchmark as it does not charge for the credit and illiquidity risk that banks have added to their asset portfolios, both of which realized positive risk premia in capital markets over this period. Remarkably, we find that the unlevered bank assets, inclusive of their share of operating expenses, underperform passive maturity-matched investments in US Treasury (UST) bond portfolios over the period 1960 to 2015. This suggests that the specialized asset-based activities of banks have contributed negatively to bank performance over this period.

And you'd be better off borrowing in the market than relying on the deposits that are supposed to give banks their funding advantage:

A second analysis compares the average cost of bank deposits, inclusive of their share of operating expenses, to the average cost of non-deposit bank debt issued in the capital market. We find that the cost of deposits exceeds the cost of debt, suggesting that banks have a funding disadvantage associated with deposits relative to capital market debt. 

If you combine the two it's even worse: "passive maturity transformation portfolios have CAPM betas near zero, while portfolios of bank equities have CAPM betas exceeding one," suggesting "that the active components of the bank business model appear to be the source of the systematic risk in banks and to have realized negative risk-adjusted returns offsetting the strong tailwinds associated with the underlying business model of maturity transformation over this sample."

What is going on here? One story you could tell is: Hahaha banks are stupid. They have taken a simple and uncorrelated business model -- borrowing short to lend long -- and turned it into a systemically risky mess that also manages to underperform the simple model. "Since 1960, banks appear inefficient in that they have not covered their opportunity cost of capital."

Another story you could tell is that banks are massive charitable enterprises. Begenau and Stafford find that bank assets have a lower return than Treasuries, sure, but intuitively it is good that someone is making mortgage loans instead of just putting their money in Treasuries. They find that bank deposits are not cheaper than capital-market funding, sure, but intuitively it is good that checking accounts exist. Sure a passive leveraged Treasury portfolio might outperform a big bank, but it might outperform my local hardware store too. I am still glad to have the hardware store; not every business model needs to be optimized to outperform a capital-markets portfolio. Banks oil the gears of commerce even though it's not profitable for them.

If you buy this story, you don't have to necessarily believe that bankers, or bank investors, are the ones funding the charity. Perhaps the story is that society wants mortgages and checking accounts, and so the government provides subsidies to banks in the form of deposit insurance and mortgage guarantees and too-big-to-fail backstops, and banks eat some of those subsidies themselves and pass the rest on to us in the form of underpriced mortgages and free checking. It is not an obviously efficient system. Yet it has the ring of truth to it.

But there's a third story you can tell that has nothing really to do with banking. This story is just that modernity is coming for all of us, and it comes first not in the form of automation but in the form of statistical analysis. Everyone's work product can be decomposed into active and passive share. If you take a rigorous inventory of yourself you will probably find that the passive share is much bigger than the active share, and that there's a decent chance that the active share is negative. Mostly you just do the stuff that everyone else does, which could be done by anyone else, or a robot. Generating uncorrelated alpha probably makes up a pretty small part of your day. Most of what I do is sit in my chair staring blankly into space with an unfocused sense of terror; the time that I spend coming up with brilliant witticisms is like three minutes a day, tops. But no one ever shows up at my desk to run a regression.

Because they are too busy showing up at the desks of investment managers, where you see this sort of active-versus-passive analysis all the time. Hedge-fund managers come to work and do stuff all day, but statistical analysis reveals that they could be replaced by a set of linear exposures to the seven-factor Fung and Hsieh model. Warren Buffett drinks Cherry Coke and makes folksy sex jokes, but statistical analysis reveals that he could be replaced by leveraged exposure to the betting-against-beta and quality-minus-junk factors.

These analyses are sometimes interpreted to mean that"most hedge funds don't appear to be doing any hedging or active management at all," that they "are a scam," that really hedge-fund managers are chucking their investors' money into index funds and going to the beach. But that is not what the analyses mean at all. The hedge-fund managers do stuff; they work long hours and are constantly coming up with ideas and testing them against reality and trying to pursue their edge. It just turns out, upon analysis, that the results of all that effort look like a (particular, arguably cherry-picked) passive strategy. It's not that the hedge-fund managers aren't doing things. It's that they shouldn't.

Similarly, if you work at a bank you are making loans or trading bonds or managing branch networks or doing hundreds of other particular business things. The lived experience of banking is not a passive leveraged long-Treasuries strategy. But statistically that's what it turns out to be. Or actually the passive strategy turns out to be better.

Hugh Kenner writes about pre-electronic computers:

There were automata long before Vaucanson – histories of the subject commence with Hero of Alexandria (first century A.D.). There were mechanical aids to computation before Babbage – Pascal designed a digital adding machine. But Hero and Pascal would not have called their artifacts simulators, but rather toys or tools, utilized by men who were metaphysically something other. The eighteenth and nineteenth centuries were less sure that man was other. To trace, in their automata, an advanced technology derived from looms and watches, enlightens us less than does consideration of their novel uncertainties about where, if indeed it existed, the boundary between man and simulacrum lay. If a man does nothing with his life but spin threads, then just how is a thread-spinning machine not a purified man? And indeed it can replace him.

Once, everyone assumed that the people who managed investments were metaphysically something other than the average market return. The great and terrible innovation of the passive-investing revolution was to make people less sure that the investment managers were something other than the market. If a banker does nothing with her bank but maturity transformation, then just how is a leveraged long-Treasury strategy not a purified bank? And indeed it can replace her.

- Bloomberg