Posts this month
A blog on financial markets and their regulation
I have a perspectives piece in the current issue of Vikalpa about Blockchain in Finance. I have been teaching an elective course on the Blockchain for over three years now, and my approach has been to treat both mainstream finance and crypto finance with equal dollops of scepticism, cynicism and openness. That is what I do in this piece as well:
Blockchain – the decentralized replicated ledger technology that underlies Bitcoin and other cryptocurrencies – provides a potentially attractive alternative way to organize modern finance. Currently, the financial system depends on a number of centralized trusted intermediaries: central counter parties (CCPs) guarantee trades in exchanges; central securities depositories (CSDs) provide securities settlement; the Society for Worldwide Interbank Financial Telecommunication (SWIFT) intermediates global transfer of money; CLS Bank handles the settlement of foreign exchange transactions, a handful of banks dominate correspondent banking, and an even smaller number provide custodial services to large investment institutions. Until a decade ago, it was commonly assumed that the financial strength and sound management of these central hubs ensured that they were extremely unlikely to fail. More importantly, it was assumed that they were too big to fail (TBTF), so that the government would step in and bail them out if they did fail. The Global Financial Crisis of 2007–2008 shattered these assumptions as many large banks in the most advanced economies of the world either failed or were very reluctantly bailed out. The Eurozone Crisis of 2010–2012 stoked the fear that even rich country sovereigns could potentially default on their obligations. Finally, repeated instances of hacking of the computers of large financial institutions is another factor that has destroyed trust. When trust in the central hubs of finance is being increasingly questioned, decentralized systems like the blockchain that reduce the need for such trust become attractive.
However, even a decade after the launch of Bitcoin, we have seen only a few pilot applications of blockchains to other parts of finance. This is because cryptocurrencies (while being extremely challenging technologically) encountered very few legal/commercial barriers, and could therefore make quick progress after Bitcoin solved the engineering problem. The blockchain has many other potential finance applications – mainstream payment and settlement, securities issuance, clearing and settlement, derivatives and other financial instruments, trade repositories, credit bureaus, corporate governance, and many others. Blockchain applications in many of these domains are already technologically feasible, and the challenges are primarily legal, regulatory, institutional, and commercial. It could take many years to overcome these legal/commercial barriers, and mainstream financial intermediaries could use this time window to rebuild their lost trust quickly enough to stave off the blockchain challenge. However, whether they are successful in rebuilding the trust, or whether they will be disrupted by the new technology remains to be seen.
Blockchain is still an evolving and therefore immature technology; it is hard to predict how successful it would be outside its only proven use domain of cryptocurrencies. History teaches us that radically new technologies take many decades to realize their full potential. Thus it is perfectly possible that blockchain would prove revolutionary in the years to come despite its patchy success so far. What is certain is that businesses should be looking at this technology and understanding it because its underlying ideas are powerful and likely to be influential.
Last week, Anwer S. Ahmed, Brant E. Christensen, Adam J. Olson and Christopher G. Yust posted a summary of their research on how banks with leaders experienced in past crises fared in global financial crisis (GFC). Their conclusion is:
We find that banks led by executives and directors with past crisis experience had significantly higher ROA before and during the GFC, fewer failures during the GFC, lower risk-weighted assets in the GFC, less exposure to real estate loans both before and during the GFC, timelier loan loss provisions in the GFC, and more persistent earnings before and during the GFC.
There are two ways of looking at this result. At the micro level, organizations should try to recruit managers with such experience. More important in my view is the macro level implication: it is good for society to have a large pool of managers with past crisis experience. That would ensure that the entire financial system copes better with new crises. But for that to happen, we need crises (at least mild crises) to happen with some degree of regularity.
Already, a decade after the GFC, I think a whole generation of traders and bankers have entered the financial system who have no first hand knowledge of dealing with a crisis. All that they have seen is a financial market numbed by ultra loose monetary policy and policy-puts. Their experience so far is that large economic and geo-political shocks (Brexit or the US-China trade war) have very mild and transient effects on market prices and volatility. The complacency of this generation is probably balanced by the battle scarred veterans who dominate the senior ranks of most banks. But over a period of time, many of these crisis-experienced leaders will retire or leave. It is quite likely that when the next big crisis comes along, there will be a shortage of crisis experience in the trenches.
Outside of finance, it is well understood that preventing small crises is a bad idea: frequent small earthquakes are better than an occasional big one; periodic restricted forest fires are preferred to one rare but big conflagration, and so on. In finance, there is a reluctance to permit even small failures. Regulators and policy makers are rewarded for moving swiftly to “solve” mini-crises. The tragedy is that this leaves institutions, individuals (and even regulators) ill equipped to cope with the big crises when they come.
In the last few years, the intermediary theory of asset pricing has emerged as a single factor model of asset pricing that does as well as the standard four factor model and thus subsumes the size, value and momentum factors (Adrian, T., Etula, E., & Muir, T. (2014). Financial intermediaries and the cross‐section of asset returns. The Journal of Finance, 69(6), 2557-2596). The theoretical justification for this model is that since financial intermediaries are the marginal buyers of many assets, their marginal value of wealth is a more relevant stochastic discount factor than that of a representative consumer. Though the idea that leverage is a good proxy for marginal value of wealth strains credulity, the empirical results seem quite strong, and there is some case to be made that the shadow price of a leverage constraint is related to the marginal value of wealth.
I see two problems with this. First of all, the major risk factors (like Momentum, Value, Carry and BAB) have been demonstrated in two centuries of data (1799-2016) from across all major world markets (Baltussen, Guido and Swinkels, Laurens and van Vliet, Pim, Global Factor Premiums (January 31, 2019). Available at SSRN: https://ssrn.com/abstract=3325720). It is evident that the structure of financial intermediation has changed beyond recognition over the last two centuries; for example, 19th Century giants like the Rothschilds operated with far lower levels of leverage than modern security dealers, and were in fact more principals than intermediaries. If the risk factors are solely due to intermediary leverage constraints, I would not expect to see such strong Sharpe ratios for the risk factors in the 19th Century data.
Second, there is a vertical split within the intermediary theory itself. He, Kelly and Manela presented a competing theory (Intermediary asset pricing: New evidence from many asset classes. Journal of Financial Economics, 2017, 126(1), 1-35) with drastically different results. I sometimes joke that Adrian, Etula & Muir (AEM) and He, Kelly & Manela (HKM) refute each other and so there is nothing more to be said. The first direct contradiction is that AEM find a positive price of risk for leverage, while HKM find a positive price of risk for the capital ratio (which is the reciprocal of leverage). Second, HKM get their nice results when they measure capital of the primary dealers at the holding company level unlike AEM who measure security dealer leverage at the unit level. Finally, AEM find book leverage to be more important, but for HKM, it is the market value capital ratio that is relevant.
I am veering around to the view that risk factors are not priced because of intermediary leverage constraints, but it is the other way around. Factor risk premiums have very long and deep drawdowns (for India, the drawdown plots are available at https://faculty.iima.ac.in/~iffm/Indian-Fama-French-Momentum/drawdown.php). As Cliff Asness put it,
I say “This strategy works.” I mean “in the cowardly statistician fashion.” It works two out of three years for a hundred years. We get small p-values, large t-statistics, if anyone likes those kind of numbers out there. We’re reasonably sure the average return is positive. It has horrible streaks within that of not working. If your car worked like this, you’d fire your mechanic, if it worked like I use that word.
So it is easier to harvest factor premiums if you are gambling with other people’s money especially with a taxpayer backstop for extreme tail events. Since Too Big to Fail (TBTF) banks are ideal candidates for doing this, you could well see significant correlations between the factors and the capital/leverage of these banks, but these correlations might be very sensitive to the measurement procedures that you use. In short, perhaps, we need to invert the intermediary theory of asset pricing.
The traditional recipe for reducing the leverage of an over indebted business conglomerate is to (a) sell non core peripheral unviable businesses, and (b) focus on improving the cash flows of the core profitable businesses. Most companies tend to do this, at least after they have gone past the stage of denial and business as usual.
But there is an alternative view expressed most forcefully two decades ago by a senior Korean government official in response to a restructuring proposal submitted by the Daewoo group: you do not reduce debt by selling unviable business, you do it by selling profitable businesses. (This statement most probably came from the Korean Financial Supervisory Commission (FSC) then led by the no-nonsense Lee Hun Jai, but I am not now able to trace this quote though the tussle between Daewoo and the government was well covered in the international press.)
I do recall one company that sold its best business without any prodding from creditors or government: RJR Nabisco under the private equity group KKR. Way back in 1995, with the tobacco business in the doldrums (as a result of Marlboro_Friday and tobacco litigation), RJR sold a part of the more attractive food business in a public issue, and used the proceeds to pay off some of its humongous debt. Apparently, the reason for not selling off the entire food business was legal advice that this could expose the board members to liability for fraudulent conveyance. (Baker & Smith discuss this episode in some detail in Chapter 4 of their book on KKR – The new financial capitalists: Kohlberg Kravis Roberts and the creation of corporate value. Cambridge University Press, 1998).
There are two arguments in favour of the radical approach of selling your best businesses to reduce debt. The first is that deleveraging is often carried out under acute time pressure and it is the good businesses that can be sold quickly and easily. Dilly dallying over deleveraging can quickly take things out of the control of management, and potentially lead to the complete dismantling and liquidation of the group as happened to Daewoo. The second argument is that financial stress at the conglomerate level acts as a drag on the good businesses that might need capital to grow or might need strong balance sheets to retain customer confidence and loyalty. In times of financial stringency, the functioning of the internal capital markets within the conglomerate becomes impaired and the good businesses tend to suffer the most. When internal capital markets start prioritizing survival over growth, good businesses should be rapidly migrated to stronger balance sheets that can both preserve value and support growth.
Many business groups in India are today trying to deleverage in response to changes in the legal regime that empower creditors, but they are still focused on selling their bad businesses. The risk is that this may prove too little, too late. At least some of them should consider the heretical idea of selling their crown jewels.
Globally, perhaps the largest conglomerate that needs to evaluate the strategy of selling its best business is GE. The aviation business is the crown jewel that is at risk from the troubles in the conglomerate. A year ago, John Hempton explained why this business needs a pristine balance sheet: whoever buys a plane powered by a GE engine needs to be confident that GE will be around and solvent in 40 years to actually maintain that engine. Moreover, the business needs massive investment in research and development, and the ability of a struggling GE to do this might be questionable. John Hempton proposed an equity raising as the solution, but the window for that might be slipping away as the share price continues to slide.
In times of stress, companies need level headed managers who can take rational decisions without being swayed by a maudlin attachment to their crown jewels.
Operational risk has always been less glamorous compared to market risk, interest rate risk and credit risk which are all now dominated by sophisticated mathematical models and apparent analytical rigour. Regulators too are uncomfortable dealing with operational risk because of its judgemental nature. Yesterday, for example, the US Federal Reserve Board announced that the largest US banks would no longer be subject to the “qualitative objection” which was the rubric under which it dealt with operational risk (see pages 13-14 of the summary instructions).
The reality however is that in big financial institutions with large well diversified portfolios, most risk management failures involve operational risk. This was true for example of JP Morgan’s London Whale, of the Nirav Modi scam at Punjab National Bank, of Nick Leeson, and many other cases. Even in the Global Financial Crisis, many of the largest losses were due as much to operational risk as to systemic events (which is why some banks had much larger losses than others).
Chernobai, Ozdagli and Wang have a paper showing that operational risk is aggravated for large and complex institutions (Business Complexity and Risk Management: Evidence from Operational Risk Events in U.S. Bank Holding Companies (December 18, 2018). Available at SSRN). They show that operational risk increased significantly when the business complexity of banks increased and provide evidence that this results from managerial failure rather than strategic risk taking. A year ago, I wrote on this blog that
banks are so opaque that even insiders cannot see through the opacity when bad things happen … Even a very competent chief executive can be clueless about some activities in a corner of the bank that have the potential to bring down the bank or at least cause severe losses.
Ignoring operational risks for the largest and most complex banks because it is too qualitative and judgemental does not appear to me to be a very good idea.
Gresham’s law states that if good money and bad money are circulating simultaneously, everybody would hoard the good money and spend the bad money thereby driving the good money out of circulation. Essentially, the good money becomes a store of value, and the bad money becomes the medium of exchange. I am beginning to think that an even more perverse outcome is possible – the good money having ceased to be money can suddenly become nearly worthless (because the store of value function of the previously good money depended on its being money). This strong form of Gresham’s law came to my mind after reading Wiegand’s recent paper presenting a “prisoners’ dilemma” model of Germany’s adoption of the gold standard in the 1870s.
The story as Wiegand describes it is as follows. In the mid 19th century, a large bloc of countries led by France was on a bimetallic standard with both gold and silver being used as money at a fixed exchange rate. New discoveries in California and Australia brought new supplies of gold in the 1850s, leading to relative shortage of silver whose output grew slowly. While in 1849, annual production (by value) of gold was less than that of silver, in the 1850s and 1860s, gold output was 2-3 times that of silver. Gresham’s law operated as expected to cause hoarding of silver in the bimetallic world: the share of gold in the French currency in circulation rose from below 30% in 1849 to over 80% in the 1860s. As the proportion of gold in France approached 100%, the possibility emerged of silver simply ceasing to be money. But if silver was no longer money, its price would decline to its value in cutlery or jewellery (the first photographic rolls using silver halide came only in the 1880s). We know from 40-year old first generation currency crisis models (Krugman, P. (1979). A model of balance-of-payments crises. Journal of money, credit and banking, 11(3), 311-325.), that the transition from France being 90% on gold to 100% on gold would not be smooth, but would happen in a sudden speculative attack that demonitizes silver. My reading of Wiegand is that Germany acted like a mega George Soros in executing this speculative attack by shifting to a gold standard and dumping all its silver on world markets; soon everybody abandoned silver and its price collapsed. (In the French bimetallic standard, it took only 15.5 ounces of silver to buy an ounce of gold; currently it takes more than five times that many ounces of silver to buy an ounce of gold.) Wiegand’s “prisoners’ dilemma” model is that Germany was forced to act pre-emptively to prevent France from launching a similar speculative attack on Germany’s silver standard.
This is what I am calling the strong Gresham’s Law: in a world of competing monies, the good money would be destroyed by a sudden speculative attack if it undergoes excessive deflation. All successful moneys have been mildly inflationary over sufficiently long periods (Triffin’s dilemma also leads to the same insight).
On the other hand, it is well known that the bad (inflationary) money could also become worthless if inflation accelerates beyond a point (Bernolz has labelled this reverse of Gresham’s law as Thiers’ law). The two laws together imply that all moneys are likely to die over multi-century time frames because of the low probability of staying on the razor’s edge between being demonitized by (a) deflation (the strong Gresham’s Law) and (b) inflation (Thiers’ law) for such long periods of time. This is consistent with the historical evidence: the ultimate fate of every fiat money in human history beginning with 11th Century China seems to be to become worthless. Near worthlessness has also been the ultimate fate of every commodity money except gold (and who knows how long gold’s luck will last?).
This has implications for crypto currency money supply rules as well. Seared by an abundance of hyper inflationary episodes in the 20th Century, crypto currencies have been designed with a deflationary bias. Many of them have inbuilt rules that freeze the money supply after an initial period of gradual monetary emission. In the wake of the collapse of crypto currency prices in recent months, some are making their systems more deflationary. Commentators are interpreting the reduced rate of monetary emission under tomorrow’s Constantinople Upgrade in Ethereum as a move to increase its market price. The weak and strong Gresham’s Laws suggests that all this might be misguided. It appears to me that after the rapid appreciation of crypto currencies in 2017, the weak Gresham’s Law kicked in and crypto currencies ceased to be medium of exchange; they became mere stores of value as exemplified by the hodl meme. It remains to be seen whether the 2018 price collapse in crypto currencies is the beginning of the effect of the strong Gresham’s Law that could destroy these currencies. Counter intuitively, an increased rate of monetary emission might actually be the way to salvage these currencies. Models with multiple equilibria are indeed quite messy.
During the global financial crisis, it became fashionable to say that a CDS (Credit Default Swap) is insurance in disguise and should be regulated as such. My response used to be that (a) a lot of insurance is derivatives in disguise, (b) an LC (Letter of Credit) issued by a bank is a CDS in disguise, and (c) it might be better for both them to be regulated as derivatives with mark to market discipline and some pre/post trade transparency. Reinsurance for example is best thought of as put options on a portfolio of non traded or illiquid assets as I wrote in a blog post nearly 11 years ago.
More recently, I am beginning to think that a convergence of derivatives and insurance could happen as “parametric insurance” moves from a fringe idea to a mainstream insurance product. The common description of parametric insurance reads almost like a definition of a weather derivative:
Parametric insurance, …, provides coverage monies automatically upon the existence of certain objective weather-related parameters based upon a set formula.
(Van Nostrand, J. M., & Nevius, J. G. (2011). Parametric insurance: using objective measures to address the impacts of natural disasters and climate change. Environmental Claims Journal, 23(3-4), 227-237.)
The parametric insurance literature talks a lot about “basis risk” which indicates convergence with derivatives not only in substance but also in terminology. More recently, proposals have emerged to move from digital call/put option payoffs (payout triggered by a variable such as rainfall amount, wind speed, or earthquake magnitude being observed to exceed a threshold) to more complex functional forms depending in non linear fashion on multiple indices (for example, Figueiredo, R., Martina, M. L., Stephenson, D. B., & Youngman, B. D. (2018). A Probabilistic Paradigm for the Parametric Insurance of Natural Hazards. Risk Analysis, 38(11), 2400-2414.) A traditional derivative structuring expert would be quite at home here.
Till now, parametric insurance has tended to be a niche product used for large transactions (often involving sovereigns or multilateral organizations). The derivatives analogy for this would be a transaction between two ISDA (International Swaps and Derivatives Association) counterparties. But that could change as well because FinTech (financial technology) players now see parametric insurance as an opportunity to break into the insurance space. They dream of using smart contracts and IOT (internet of things) to turn parametric insurance into a retail product. In some of these grandiose plans, a sensor in my home will inform the insurance company that it detected flood waters inside my home and the insurance company will automatically transfer the payout (or is it payoff?) to my bank account, and perhaps, all of this will happen on the blockchain. So we will have the equivalent of retail weather derivatives. I hope there will be a mark to market regulation somewhere.
The Insolvency and Bankruptcy Code (IBC) introduced in India a couple of years ago was from the very beginning a morality play with a thin veneer of economic theory. With the passage of time, the veneer of economic theory has eroded, the morality play has become stronger and the functioning of the code has become progressively more divorced from economic reality.
Until the IBC came along, Indian businesses were well protected from their lenders by mechanisms like the BIFR which were the morality play of an earlier era that had become perverted over time. Originally designed to protect the interests of workers of distressed companies, these mechanisms ended up entrenching incumbent management, and leaving lenders helpless. Since a lot of the lenders were public sector banks, there was a strong political pressure to redress the balance. After a couple of attempts to empower the financial sector (Debt Recovery Tribunals and SARFAESI) proved inadequate, the IBC was introduced to redress the balance decisively.
In this morality play, the corporate sector were the villains, and the banks were the saints. The obvious solution was to hand over insolvent companies to the financial creditors. Of course, companies have operational creditors, but since these tend to be businesses which were classified as villains, a decision was taken to exclude operational creditors from the decision making. Morality was also the path of expediency as the new system also served as a backdoor bailout of the beleaguered financial sector.
Decades of studying finance have taught me that the world of finance is full of villains, but there are hardly any saints. In my first blog post on the Indian bankruptcy reform, I wrote that in the real world bankruptcy was “very much like the familiar scene in the Savannah where cheetahs, lions, hyenas and vultures can be seen fighting over the carcass”. There is no fairness in the jungle, and victory belongs neither to the one that hunted down the prey nor to the one in greatest need of food; victory typically goes to the most wicked of the lot. The story is the same when it comes to distressed debt around the world. Last month, Jared Ellias and Robert Stark wrote a fascinating paper entitled “Bankruptcy Hardball” which documented several episodes of such wickedness in the United States.
The sidelining of operational creditors was initially the most egregious morality play in the IBC and I wrote more than one blog post on this issue (here and here). Moreover, even the morality of this exclusion became suspect when it was realized that home buyers who had paid an advance to an insolvent builder would be operational creditors. Politically, it was impossible to club home buyers with other villainous operational creditors, and exceptions were made for them.
But there was more to come. Very soon, instances arose where the incumbent managements of the insolvent companies were potentially the highest bidders in the bankruptcy auction of their companies. Under the original IBC, they would have prevailed, and this might have been the best outcome from the point of view of maximizing the economic value of the lenders. But since the IBC was from inception a morality play, this could not be permitted. So the law was hurriedly amended to prevent them from bidding.
But this creates another problem. Originally, creditors were put in charge of the decision making because it was supposed to be a purely business decision. As the Bankruptcy Law Reforms Committee wrote in its report:
The evaluation of these proposals come under matters of business. The selection of the best proposal is therefore left to the creditors committee …
However, with the exclusion of tainted bidders, the choice of the best proposal is no longer one of economics, but one of theology. Some of the feverish debates in the courts on which bidders are tainted enough to be excluded reminds me of medieval scholastic debates about “How many angels can dance on the head of a pin?“. From an economic point of view, these debates are ridiculous. As the Roman emperor Vespasian said while imposing a tax on urine, Pecunia non olet (money does not stink). Morality plays tend to forget this principle.
By elevating morality above economics, the IBC is failing to live up to its promise. Instead, we see confusion reign paramount. We see distressed companies boasting of a respectable market capitalization while their debt trades at less than half of book value. We see bankruptcy remote vehicles delaying payment on their obligations after the parent group filed for insolvency. The time has come for us to deemphasize the morality play. It is time to hold our noses like emperor Vespasian, and get on with the ugly business of economics.
I have blogged several times about how Covered Interest Parity (CIP) is not valid in the multi-curve discounting framework that is the standard in finance after the Global Financial Crisis. (My last post a couple of years ago argued that economists who still believe in CIP unreservedly are simply ignoring risk; earlier posts described the cross currency basis and the multi curve discounting framework).
Recently, I read a paper by Wong and Zhang that is perhaps the most lucid explanation that I have seen of the phenomenon of CIP violations and the emergence of a large cross currency basis. They are able to explain not only why the forward premium is not equal to the Libor differential, but also why the CIP violation persists when Libor is replaced by (near) risk free rates like OIS (Overnight Indexed Swaps) or repo.
Wong and Zhang point out that the Libor-OIS spread reflects two different things. First, Libor carries significant counterparty credit risk because it involves unsecured lending for a non trivial time period, while the overnight tenor of OIS reduces the credit risk to negligible levels. Second, Libor carries an exposure to funding liquidity risk because the lender has to fund the loan till maturity, while OIS involves only an exchange of interest cash flows without any principal funding.
The Cross Currency Basis Swap (CCBS) in its post-crisis form does not expose the counterparties to credit risk because of collateralization and variation margins. But it does involve funding liquidity risk (each party receives liquidity in one currency and gives up liquidity in another currency). Thus the CCBS spread reflects only one part of the Libor-OIS spread – the part that accounts for funding liquidity risk. The empirical results in the Wong and Zhang paper show that in some currencies, the Libor-OIS spread is dominated by credit risk while in other currencies (notably the US dollar) it is dominated by funding liquidity risk. As a result, a CIP violation is observed whether one measures the interest differential using Libor or OIS.
Of course, all this is consistent with the multi-curve discounting framework, but this analysis is probably a lot easier to understand.
(This was posted on my blog yesterday but due to an oversight was not copied to mirror sites until now.)
Adriana Robertson argues in a recent paper that index investing is not passive investing; it only delegates the active management to the index proviver. (Passive in Name Only: Delegated Management and ‘Index’ Investing (November 2018). Yale Journal on Regulation, Forthcoming. Available at SSRN). This is a problem because mutual funds are regulated, but index providers are not. The paper presents data showing that the vast majority of indices in the United States are used as a benchmark by only 1 or 2 mutual funds, and so it is hard to argue that these index providers are subject to strong market discipline.
She offers an ingenuous suggestion to solve this problem without new intrusive regulation.
While a mutual fund cannot deviate from its fundamental policies, as stated in its registration statement, without a shareholder vote, there is no restriction on an index’s ability to change its methodology.
Fortunately, there is a simple solution to this problem. Once we recognize that delegating to an index is no different from delegating to a fund manager, we can craft a solution based on the existing rules: Any time the underlying index makes a change that, if made by the fund manager in a comparable actively managed fund, would trigger a vote, the fund manager is required to hold a vote on retaining the index. This simple change would harmonize the protections offered to investors in the two types of funds.
I can think of at least two significant index changes that would qualify under this rule, and on both these, I think Adriana Robertson’s solution makes eminent sense: