Although horse trading over the House and Senate versions of regulatory reform is limping to an end, I thought I’d bring up some fundamental observations the Bank of England’s Andrew Haldane made earlier this year on systemic risk and cost, bank size and diversification. Andy, as some Internet sources refer to him, is Executive Director, Financial Stability, at the BOE.

These are points culled from empirical research, the sort of fundamentals one would always wish policy makers would take as their starting point, instead of the kind of sloganeering that seems to pass for legislative activity these days. I also thought they would be fitting contemplations for the titans of finance running the Godzilla banks (including one, whom I know personally to be an exemplary centurion of banking, name checked above). Or, if not for the CEOs, then certainly food for thought for the past, present and future bank shareholders whose investments they shepherd.

Count the Cost of Systemic Crisis

Speaking last March on “The $100 Billion Question,” Haldane introduced his topic of how big a bank should be by offering some estimates of the damage wreaked by the recent disasters. The media was all over this piece of Haldane’s exegesis, like as they say white on rice.  But it’s worth recapping, if only because lots of reporters tend to leave the best thinking behind, eviscerated and shredded, when they grab that money shot.

Haldane defined systemic risk as a pollutant, a noxious by-product of the banking industry. Just like other industries, banking can be seen as producing private benefits for investors, customers and bank employees and social costs via banking crises (which have been occurring with distressing regularity over the recorded history of banking). Once the situation is drawn in this way, reckoning the cost of the latest crisis “helps calibrate the intervention necessary to tack system risk, whether through regulation of restrictions.”

He then works through some hefty academic work in banking, finance theory, operations research, even the behavioral and physical sciences to conclude that the optimal size for banking enterprises might be under US$100bn. (His bibliography cites more than 40 works!) Grounding his argument in the actual historical progression, in US law, from the Depression to the present, he does so with considerable wit and irony and mostly without too heavily taxing non-financial types’ intellectual resources.

In terms of fiscal transfer from the government to the banks, the cost of the crisis to US taxpayers is estimated around $100bn. Not too big a bill to present to the banks. Haldane estimated that taxing the banks $5bn a year could recoup it, assuming a systemic crisis occurs every 20 years.

The true social cost is better captured by foregone output — estimated at about $4trn worldwide. Moreover, some of the GDP losses may persist, in which case the present value cost of the crisis would significantly exceed the immediate cost. This is a bill too big to put to the banks. Making that same assumption of a systemic banking crisis every twenty years, the estimated annual levy on banks would exceed the current market capitalization of the largest global banks by several hundred billion dollars.

Banks of course do not bear the full blame. That is, “For every reckless lender there is likely to be a feckless borrower.” Haldane tries another measure, the implicit fiscal subsidy provided by government support, which can be estimated by the difference between banks’ stand-alone ratings and their supported ratings (worth 1.5 to 4 rating notches in Haldane’s data, rising from an average 1 notch in 2007 to 3 in 2009). Worth about $250bn in 2009 lower interest payments on bank liabilities (ex deposits) to a sample of global banks.

More to the point, the bigger the banks, the greater the implicit subsidy provided by being too big to fail. Looking at UK banks and building societies, Haldane found the large institutions sopped up 90% of the implicit subsidy. Another study found that the annual subsidy in funding costs for the 18 US banks with over $100bn in assets was $34bn per year.

The Al'Qaeda Metaphor

Aside from putting pricing out the social pain, Haldane’s speech got a lot of attention — if not much profound consideration by policy makers — because he used the structure of Al'Qaeda to promote the idea that smaller could be better. His point was that "modularity" dissipates some systemic risk. In a nutshell, because Al'Qaeda’s exists as a bunch of decentralized cells, it is difficult to infiltrate and destroy.

Haldane cites other kinds of systems and networks where structural modularity translates into systemic resilience, but the most intriguing – and much more persuasive than Al'Qaeda – comes via attempts on the world domino-toppling record. For what is our global financial system, if not a long line of dominos!

In the 1980s, an attempt involving 8000 dominos failed when a member of the TV film crew dropped his pen, spilling the majority of the dominos. Twenty years later, a sparrow topped 23,000 dominos, but "750 built-in gaps averted a systemic disaster" and allowed a new record of over 4m dominos was set.

Haldane’s point? If all these big, full service conglomerate banks are fully diversified, in effect they hold the same portfolio. The system as a whole lacks diversity and is more prone to generalized systemic collapse. "Homogeneity breeds fragility."

That’s not all. Despite the "intuition" of conglomerating banks that big is more cost-efficient, Haldane can demonstrate empirically that the size or diversity of 24 global banks does not translate into more stable income. In fact, Haldane’s data indicate that size and diversity may increase income variability. Worse, he can show that during the crisis the larger, more diversified banks suffered proportionally greater losses.

Academic research — "the literature" — indicates that economies of scale are exhausted at much lower levels than might be presumed — perhaps at $5bn – $10bn in assets. Studies indicate that economies of scope are similarly limited. Analysis of US bank holding companies suggest that gains achieved by diversifying business lines may well be offset by exposures to businesses such as trading with more volatile income streams. (UBS shareholders, who revolted in 2008 over the firms US mortgage trading losses, might agree.)

"KISS"

So why isn’t cost efficiency a linear function of size? Here’s where the breadth of Haldane’s scholarship kicks in. The answer — like the scrapped prescribed modularity of Glass-Steagall — has been around since the Thirties. In a 1934 paper, "The Problem of Management and the Size of Firms," Austin Robinson asserted the human mind and memory are the limiting factor. Quotes Haldane, "…Every increase in size beyond a point must involve a lengthening of the chain of authority…at some point the increasing costs of co-ordination must exceed the declining economies." (Those are Haldane’s ellipses.)

For a concrete example, Haldane looks outside banking and finance to military history. "“In Roman times, the optimal size of a military unit was 100 — hence the Roman centurion. This was the maximum number of men a general felt able to know well enough to lead and control." Two millennia and massive advances in telecommunications later, the optimal unit size in the US army is still held to be just under 100.

In fact, there is a law — Dunbar’s Law — that says the number of relationships humans appear able to control is less than 150 (this finding is drawn from neurological behavioral science). For most people, suggests Haldane, the number is probably single digits, despite LinkedIn or Facebook, iPhones or Blackberries. (Facebook contacts, Haldane quotes another saying, are not really friends.) By contrast, a Godzilla bank can be comprised of several thousand legal entities.

The recent crisis held many examples of failures rooted in what Haldane calls "an exaggerated sense of knowledge and control….When Lehman Brothers failed, it had almost one million open derivatives contracts — the financial equivalent of Facebook friends. Whatever the technology budget, it is questionable whether any man’s mind or memory could cope with such complexity." (Those are my ellipses.)

Not Even Jamie Dimon

I cannot read this wonderful passage of Haldane’s without thinking of Jamie Dimon, CEO and Chairman of JP Morgan Chase. Dimon is the most impressive manager I’ve ever seen in the flesh. (Disclosure: He is the reason I bought some shares of JPM in the dark days of the financial crisis and continue, sentimentally, to hold them.)

Here’s how. I was briefly at Smith Barney after Drexel Burnham, where I got my introduction to mortgage research, failed. Sandy Weill, in the person of Primerica, had recently acquired Smith Barney and one of Dimon's jobs was to oversee the trading businesses. The thing that impressed me so was that Dimon was regularly out on the trading floors, in his shirt sleeves, asking questions. He knew our names and what we knew. I could have been six reports below him, but more than once he stopped me to ask what I thought about IO strips because elsewhere in that conglomerate someone was building a portfolio of these leveraged prepayment bets on the premise they were cheap. Were they cheap, and what were the risks they wouldn’t pay off? Alas, I was too green an analyst to frame a response that could put me on the wrong side of my traders.

I’ve heard the same kind of stories from people who worked in the trenches at Bank One when JP Morgan bought it and Dimon in 2004. And Bank One was a big bank — $290bn in assets. Today, JP Morgan is over $2trn in assets. Dimon must have a number of able centurions below him, because the Wikipedia notes that President Obama cited JPM as one of "a lot of banks that are actually pretty well managed." The President conceded, "Jamie Dimon, the CEO there, I don't think should be punished for doing a pretty good job managing an enormous portfolio."

Still, I wonder how many of Dimon’s centurions walk among the troops and know who on the line does what?  And recent headlines say JPM is looking to expand in Europe. As great a general as Dimon is, I wonder if he might be spreading my contribution to his capital too thin.

A Recipe for Catastrophe

If Haldane’s remarks are relevant for the big bank managers charged with protecting their shareholders’ interests, they are doubly so for the policy and lawmakers who will presume to repair the financial system.

The financial system operated by mega universal banks is like other complex dynamic systems: "the distribution of risk may be lumpy and non-linear, subject to tipping points and discontinuities." Unlike the tidy world of finance theory, the distribution of outcomes in this system is unknown. This is a wholly different order of uncertainty than the kind of uncertainty (garden variety investment risk) that is traded for return. The technical term for this minefield uncertainty is “Knightian” uncertainty.

Haldane is conversant with the literature on how systems should be regulated in the face of Knightian uncertainty. Rule one, keep it simple. “Complex control of a complex system is a recipe for confusion at best, catastrophe at worst.” Complex control adds rather than subtracts uncertainty.

His example is the US constitution, four pages long. By contrast, the first draft of Christopher Dodd’s Senate financial reform bill was 1,336 pages long. (And the conference committee is not editing it!) Which, wonders Haldane, “will have a more lasting impact on behavior”?

Second, minimize the likelihood of the worst outcomes. Often the simplest way to do this is what non-economists call “structural reform”. That is, regulate structure, not behavior. Case in point: Glass-Steagall, simple “red-line regulation” and only 17 pages long, it separated commercial and securities brokerage banking businesses. (I guess the Volcker rule, by comparison, would be a pink line.)

Haldane’s counter example is yet more egregious than the US reform bill. Basel II “was anything but simple”. It comprises thousands of pages, took fifteen years to deliver, and was calibrated in the main “from data drawn from the Great Moderation, a period characterized by an absence of tail events”. Rigged with a complex menu of capital risk weights, it epitomizes “fine-line, not red-line, regulation”.

Amen.

NOTE: Linda Lowell writes a regular column, called Kitchen Sink, for HousingWire magazine.

Editor’s note: Linda Lowell is a 20-year-plus veteran of MBS and ABS research at a handful of Wall Street firms. She is currently principal of OffStreet Research LLC.