Plus: Curse of the IPO, Reputational Value, Index Providers
|Marc Rubinstein||May 21||16||2|
Welcome to the one year anniversary issue of Net Interest. No paywall (yet) so it remains your destination for free expert insight on the world of finance. Over 25 years as an analyst and investor in the space and there’s still lots to say! If you haven’t done so yet, please sign up for the weekly email, and if you’re looking at an interesting investment opportunity in fintech or financials broadly, do reach out.
To understand the financial services industry at the time of the financial crisis, you were well placed if you’d read Hyman Minsky. Minsky himself was ten years in his grave by then but his theories linking financial markets and the broader economy served a useful model of what was going on.
Today, that economist is Ronald Coase. If you want to know what’s going on in financial services today, he’s the man to study.
Coase was active from the 1930s all the way up to his death, aged 102, in 2013. Like Minsky, he had a practical streak – both men challenged the idea of efficient markets (enhancing their credibility among investors). Through his macroeconomic lens, Minsky attacked the notion that markets tend towards equilibrium. From a different perspective, Coase wondered why, if markets and prices are so efficient, corporate entities exist. In 1937, Coase wrote an essay, The Nature of the Firm, where he attempts to explain why some business activities sit within the control of corporate management while others are exchanged through arm’s length transactions in the marketplace.
Coase’s insight, which may seem obvious but has far-reaching implications, is that transacting is not a costless activity. In his essay he showed that the size and structure of firms, and the location of the border between internal exchange within the firm and external exchange through markets, stem from the costs of transactions. In short, firms will expand or shrink until the cost of “making” equals the cost of “buying”.
This insight explains a lot of the trends underpinning fintech.
Take Marqeta, which filed for IPO this week. Founded in 2010, the company provides a platform for customers to issue their own payment cards. But the company – at least for now – only really has one customer: Square. In the first quarter of this year, Square accounted for 73% of Marqeta’s net revenue. On top of that, the company also has significant concentration around one key supplier: Sutton Bank. Over the same period, 94% of total payment volume was settled through Sutton Bank. So Marqeta acts as a kind of conduit between Square and Sutton Bank. Which begs the question: why does Square bother to contract to Marqeta rather than just issue payment cards itself?
It’s a question we’ll come back to. First, let’s zoom out and look at the overall financial services landscape.
High Transaction Costs
When you look at the financial services sector, you find one of the most vertically integrated industries in the economy. Stroll down any commercial high street and you’ll (still) see bank branches nestled among retail stores and restaurants. But while most retail stores don’t sell products they’ve manufactured or restaurants food they’ve harvested, banks continue to control much of the financial services value chain from product design right through to distribution and servicing.
Other industries used to be like that. Andrew Carnegie once produced the coke to fuel his steel mills. And Henry Ford built a giant plant at River Rouge near Detroit to take iron ore and rubber all the way through into cars, with 100,000 people working there at peak. The reason they did this was captured by Coase: it was cheaper to make than to buy.
Yet in each case, the trade-off between make versus buy shifted over time, prompting these companies to focus on a narrower set of activities. In Carnegie’s case, it became evident by the 1880s that his coke ovens were costing more to operate than the equivalent cost of buying coke independently, so Carnegie sold off the ovens and bought coke on the open market. In Ford’s case, car parts became more standardised and the cost of finding them declined with better communications technology.
The two firms shrank because of a reduction in transaction costs.
One reason banks have been loath to shrink is that historically they have derived a benefit from size. In particular, they enjoyed a state subsidy if they were “too big to fail”, reflected in lower funding costs. According to the Bank of England, in the five years prior to the financial crisis, the implied annual subsidy to the world’s largest banks averaged $70 billion a year, equivalent to half of their profits over that period. The bigger they were, the bigger the absolute contribution of the subsidy. That subsidy – which couldn’t be bought on the open market – made it cheaper for banks to keep transactions in-house, on their own balance sheet, rather than look beyond their perimeter. In line with Coase, banks chose the size to maximise their private value.
In Japan, banks pushed their boundaries in a different direction – by integrating with networks of companies via keiretsu. The high cost of bankruptcy in Japan incentivised companies to stay small. But rather than access the open market to transact, companies formed tight-knit relationships with other firms and with banks. This third approach, neither open market nor vertical integration, took root in Japan thanks to low levels of moral hazard. Compared with other countries, suppliers had a tendency to behave in a less opportunistic fashion. Coase is apparent here, too: the structure of Japanese banking evolved as a consequence of the trade-off companies made between growing too large (high transaction costs) and dealing with a small circle of associated companies (low transaction costs).
Although his immediate focus was on transaction costs, over his career Coase built on top a detailed analysis of the economics of contracts and the economics of property rights. His conclusion: that markets trade principally not in goods and services but in legal rights.
Economists commonly assume that what is traded on the market is a physical entity, an ounce of gold, a ton of coal. But, as lawyers know, what are traded on the market are bundles of rights, rights to perform certain actions. Trade, the dominant activity in the economic system, its amount and character, consequently depend on what rights and duties individuals and organizations are deemed to possess – and these are established by the legal system.
Legal rights are a useful way to think about the nature of distribution in financial services. The difference between operating an internal salesforce and using external agents rests on who has legal rights over the customer list. Insurance companies utilise both channels and can pay salespeople inside each on a similar basis through commissions. The advantage of the agency model is that agents bear their own costs, but tensions can develop if the insurance company tweaks product features or changes pricing between renewal dates. Indeed, if the agent owns the customer list, the company is not incentivised to invest in it and may be more inclined to tweak product features on renewal. For shorter duration products like auto, where renewals are sensitive to sales effort, it makes sense to use an agent network. From the agent’s perspective, if the insurance company shirks, they can sell a competitor’s products. But for longer duration products like life insurance, where renewals are more certain and less sensitive to the agent’s actions, it makes sense for the insurance company to own the list.
This analysis of vertical integration in insurance was developed by Sanford Grossman and Oliver Hart, economists heavily influenced by Coase (Nobel prizewinners, all). It applies in banking too. Deposit persistence is typically quite high, which is why banks own their distribution channel.
But underpinning all of this is transaction costs. Ownership emerges when it becomes too costly for a company to specify everything it wants from its supplier. There is a theoretical contract that would resolve the tension between the insurance company and the agent network, but it would have to cover so many eventualities it would be unwieldy.
Low Transaction Costs
A major feature of the past decade has been the steady reduction in transaction costs. Michael Munger, author of the book, Tomorrow 3.0: Transaction Costs and the Sharing Economy, regards this period as the third great human entrepreneurial revolution, where the key value proposition in the economy isn’t selling products but selling reductions in transaction costs.
This is where fintech comes in. New and specialised fintech companies are emerging to ease the movement of financial operations out of rigid corporate structures and into the marketplace. Marqeta is one of them. Of course Square could have developed its own card issuing operation, but Marqeta makes it cheaper (evidenced by the $120 million losses it has incurred over the past 27 months).
Indeed, there are a whole host of fintech businesses active in this infrastructure layer of financial services. Within card and payment processing there’s Galileo (owned by SoFi), which counts Chime and Monzo among its clients, and Modulr, which has Revolut as a client. There are also companies powering open banking initiatives – companies like Plaid and Truelayer offer solutions that allow providers of financial services to share data across the marketplace in a Coasian sense. In a world where there are APIs All The Way Down, financial companies can more readily transact with each other.
Purists will point to marketplace lending as an endpoint in the shift from the firm to the marketplace (and DeFi, but that’s for another time). Pioneers in the marketplace lending space like Zopa and Lending Club separated the borrowing side of banking from the funding side and put both out into the marketplace. But liquidity was not in equilibrium across each side and so transaction costs did not come down to a sufficiently attractive level. It may have been a development that came too early; the immediate outlook is for smaller institutions rather than a population of individual bankers.
The trend has regulatory implications since the marketplace encompasses both regulated and unregulated institutions. New fintech entrants have incentives to tap into the regulated sector sparingly; meanwhile, regulated entities have incentives to push out new fintech services into firms outside the regulatory perimeter. Square/Marqeta provides an example. They need to contract with a regulated bank to provide payment processing but chose Sutton Bank, a small bank whose asset size allows it to charge merchants higher interchange fees (as per the Durbin Amendment to the Dodd-Frank Act, which places a cap on interchange fees on larger banks). Square/Marqeta is able to earn ~120 basis points of gross interchange which is much higher than bigger banks can earn.
Last week, we discussed the landscape in the US mortgage lending market and I cited research showing that over half of the growth in non-bank mortgage lending derives from regulatory advantage with only around a third coming from superior technology. In fact, the two are intertwined. Developments in technology facilitate gaps in regulation. Without the technologies available today, Square/Marqeta wouldn’t have been able to contract with Sutton Bank to exploit Durbin at all.
Coase was wise to the issues of regulation. In his collection of essays, The Firm, the Market and the Law, he writes: “The interrelationships between the economic system and the legal system are extremely complex, and many of the effects of changes in law on the working of the economic system… are still hidden from us.” The classic approach to financial regulation is based on the regulation of entities. For example, some aspects of the US payments system are available only to insured depository institutions. But as the range of market-based transactions increases with technological developments, fewer activities are left within the management of regulated firms. This presents a challenge for regulators – something we discussed in The Policy Triangle. One option is to target a more activity based regulatory framework.
The boundaries of the firm are constantly in flux but they are being eroded more quickly than ever in the financial sector. As during the financial crisis it is hard to keep up but Ronald Coase provides a map.
More Net Interest
Curse of the IPO
The collapse in the price of Bitcoin and other crypto assets is interesting because it coincides with the IPO of Coinbase (and me writing about Galaxy Digital, the Goldman Sachs of Crypto, but let’s not go there). John Authers of Bloomberg highlights other market tops that have coincided with the IPO of a key participant in that market – Blackstone in 2007 (leveraged buyouts) and Glencore in 2011 (commodities); another would be Goldman Sachs in 1999.
One view is that the owners of these firms have deep market insight and are really good at market timing. I don’t buy it. This view cements an aura around these people who in most cases stay on to run their firms; there’s some short-term cynicism but the narrative enhances their reputation leaving them well-placed for the next cycle. Putting the small sample size to one side, I suspect any correlation is driven more by the demand side than the supply side. Global markets are still very siloed; few investors have a mandate to buy commodities and leveraged loans and crypto and equities. In most cases the pure equity investor will be the last one into the market and ends up with the most leveraged play on the market.
Interestingly, Klarna is happy not to feed the demand. Its CEO told the FT, “I preferably would not take a company public in a market that is that overheated. It would be nice to come into a little bit more sane environment.”
Scenes in Madrid this week, where a court hearing began between the head of Unicredit and the head of Santander over an employment contract. In case you missed it, Santander offered Andrea Orcel the job of CEO while he was at UBS in 2019, but revoked the offer after he’d resigned from UBS and announced it to the world. Orcel is now CEO of Unicredit, but is still put out by the incident. His original claim was for €112 million of lost compensation and “reputational damage” but that’s now been dropped to around €67 million. But it still includes €10 million of reputational damage.
Dan Davies (a Net Interest guest contributor) wonders what the number implies about the gross value of Orcel’s reputation. “He has, after all, now got a CEO post with a national champion bank, and his former colleagues still want to work for him. His Wikipedia page is glowing and he got Lunch With The FT. He had to miss Davos for two years, but one of them was cancelled anyway.” If €10 million is the impairment, the implied reputational value must be an order of magnitude greater than that, which seems high, but then European bank CEOs are in high demand.
We’ve discussed here before the amazing power that index providers hold (The Business of Benchmarking). Much attention centres on the asset management firms that run index money, but index providers run a very profitable toll on that activity. This week, S&P Global was fined by the US Securities and Exchange Commission for publishing stale values on the XIV exchange-traded note in 2018. Matt Levine has a very good write-up.
What went wrong was this: “On February 5, one of the two index managers responsible for monitoring the Index was out of the office. As a result, just one index manager (“Index Manager”) was left to monitor the Index, which was one of thousands of indices he was tasked to monitor that day.”
Although she voted against the action, SEC Commissioner Hester Pierce has raised concerns about the power of index providers: “This enforcement action may hint at a deeper, unspoken concern that index providers, whose products have become so integral to our securities markets, are not governed by a regulatory framework explicitly tailored to their activities.” Their rise to power has been stealthy but regulators are now taking note.