Here is a Wall Street Journal story about how the Office of the Comptroller of the Currency and the Federal Reserve have told banks that they can't make really leveraged loans, and how that makes life harder for private equity firms, whose lifeblood is really leveraged loans. This is controversial, because private equity firms don't like to be deprived of their lifeblood, and because banks don't like it when their regulators tell them what to do.

What is nice about the controversy is that the sides are entirely predictable. If you don't think that regulators should be telling banks how to run their businesses, at the specific loan level, then this is an obvious bad. Here is a guy who is disturbed:

"Micromanaging lending in this manner could potentially shake the fundamental tenets of the leveraged-loan markets and lead to unintended consequences, such as distorted pricing or limited supply," said Alex Jackson, who invests in high-yield loans on behalf of Cutwater Asset Management, which oversees $26 billion. "It's a little disturbing."

It's particularly tenet-shaking because of the blunt arbitrariness of the rules: Lending to companies with a debt-to-earnings (no, debt to our good friend Ebitda, earnings before interest, taxes, depreciation and amortization) ratio of greater than 6 times is basically right out, but a 5.9x ratio is fine. Give or take.1 You can buy a company and lever it more than 6 times, but not with (any) bank debt. You gotta go get bonds, and "Since bonds are typically more expensive than loans, the revised structure can make the deals more costly."

If on the other hand you think banks are rapacious idiots who will lever up the system until it explodes, then this guidance is straightforwardly good, and those complaints should be ignored. Here is an unusually boisterous bank regulator:

"The impact on private equity, a significant driver of what we see as risky practices, is an intended consequence of our actions," Martin Pfinsgraff, the OCC's senior deputy comptroller for large-bank supervision, said in an interview. "As regulators, we certainly hope to change bad practices and remove the extraordinary froth that's experienced at the peak of a credit cycle. If we can mitigate that, it reduces the size of the valley to follow."

My knee-jerk assumption is that banks know more about the specific details of how to bank than bank regulators do, but that is obviously a debatable question.

One thing I did was build a very toy model of how much leverage is "allowed" in the typical buyout. This is meant as a toy model, not a real one, but it's maybe something to think about:

Source: Excel, Bloomberg data, arithmetic, my imagination, etc.
Source: Excel, Bloomberg data, arithmetic, my imagination, etc.

If you buy a company and put no debt on it, then you finance it 100 percent with equity, tautologically enough. If you buy a company with 50 percent equity and 50 percent bank loans, then your purchase is financed ... about 58 percent with equity. That's the 50 percent equity that you provide, plus the equity that the bank needs to have as capital against your loan. But if you're borrowing more than 6x Ebitda, you gotta borrow it from the bond markets, not from banks, and bond investors are probably less levered than banks (though who knows?). So, paradoxically, if you borrow 6.5x Ebitda you'll end up with more total equity funding than if you borrow 5.5x (on this simplistic model), because as you get more levered, your lenders will themselves be less levered characters.

How should one feel about that? Well, I mean, the shape of that graph is ugly. Jagged cutoffs are always ugly, as a matter of aesthetics but also as a matter of ... arbitrariness? Arbitrage? Those two arbi-things are sort of formally equivalent. Making 6.01x leverage much much more expensive than 5.99x leverage is arbitrary and blunt and unfair and all that bad stuff, so the less imaginative will complain bitterly. The more imaginative will find a way to exploit the arbitrary cutoff to achieve things the rules were designed to prevent. Smooth curves are both fairer and harder to game.

But the direction of the graph makes some sense, no? Most companies probably should get a good chunk of their funding from people who can afford to lose it -- either equity investors, or at least equity investors in their debt -- as opposed to from long chains of highly levered, highly sensitive risk-averse investors. It's not easy for regulators to ensure that result -- you can't just ban companies from taking out lots of loans,3 and bank capital regulation probably won't do it either4 -- though, arguably, the market sort of does it itself: It's a lot easier to borrow against a Treasury bond than a junk bond. But if you think that the market is sometimes insufficiently cautious about piling leverage on leverage -- and, y'know, there is some evidence in that direction -- then this blunt method of forcing it in the right direction may seem pretty reasonable.

1 "Loans with debt ratios above the guidance may be permitted if, say, regulators view the industry the borrower operates in especially favorably, or if the bank convinces regulators the loan has strong odds of repayment," says the Journal, though "exceptions will not be the norm." Here is the OCC guidance, which roughly defines leveraged lending as loans where senior debt (i.e. bank) exceeds 3x Ebitda or where total debt exceeds 4x Ebitda, and which says that "a leverage level after planned asset sales (that is, the amount of debt that must be serviced from operating cash flow) in excess of 6X Total Debt/EBITDA raises concerns for most industries."

2 What is going on here? First of all I assume that our buyout comes at an enterprise value to Ebitda ratio of about 10.2x, which I get from running Bloomberg MA on (1) all asset sales, company takeovers, going-privates, management buyouts, reverse mergers, tender offers and PE buyouts (the Bloomberg MA categories) (2) where target and acquirer are in the U.S. (3) completed in 2013. That gets 5,416 deals with an average size of $403 million and an average EV/Ebitda ratio of 10.18x. (Limiting it to just PE buyouts gets you a 9.67x multiple, so close enough.)

So, you know, a deal funded with 50 percent debt has 5.1x debt-to-Ebitda ratio (50 percent of 10.2x).

Then I assume that banks fund loans with 12 percent equity. That is pretty utterly arbitrary. It comes from the facts that:

  • under the Basel III standardized approach, corporate loans have a 100 percent risk weight; and
  • under Basel III, banks are required to have equity equal to ... some percentage of their risk-weighted assets. See page 21 here; if you just tot up the common equity tier 1 requirements (4.5 percent for common equity tier 1 core requirement, 2.5 percent capital conservation buffer, 2.5 percent maximum countercyclical buffer and 2.5 percent maximum G-SIB surcharge) you get 12 percent. Plus 3.5 percent required non-common equity capital. I don't know, 12 percent seems like a reasonable number.

Then I assume, most unjustifiably, that the average bond investor is levered 2x. Bond mutual funds can't be levered more than 1.33x. Bond hedge funds ... can be. Pages 17-18 here are basically the Treasury Office of Financial Research spinning its wheels about the impossibility of knowing how levered the average asset manager is. But the general assumption is that the answer is at least "less levered than the average bank," which is why that OFR report -- on maybe regulating asset managers under too-big-to-fail authority -- is so controversial.

Then I assume that the OCC doesn't allow banks to lend anything in deals with more than 6x total leverage, because that seems to be the import of these rules. Also just for giggles I assume that banks don't lend more than 4x Ebitda in bank debt, so a deal with between 4x and 6x leverage is funded 4x with bank loans and the excess with bonds. (And that anything below 4x leverage is funded entirely with bank debt.) Those last assumptions aren't especially rigorous but they make the chart look more interesting.

Anyway here is how the chart would look in the absence of the OCC rules, just assuming up to 4x in bank debt and the rest in bonds:

Obviously you can make this look worse by letting banks lend more than 4x, etc.
Obviously you can make this look worse by letting banks lend more than 4x, etc.

Equity decreases monotonically in this one. In the other one, equity never gets below 50 percent.

3 Can you? You shouldn't, anyway.

4 Because making capital regulation finely attuned to the riskiness of the underlying asset is sort of a dangerous game. Lots of people dislike the whole concept of risk-weighting, but finely tuning risk weights based on the leverage of the underlying companies seems even less popular. That is how advanced-approach banks do it, but the standardized approach is a blunt 100 percent risk-weight for AAA and junk-rated loans alike.