No man can serve two masters

Rogue Trading at Lloyds Bank International, 1974: Operational Risk in Volatile Markets

By Catherine Schenk (Glasgow)

Abstract Rogue trading has been a persistent feature of international financial markets over the past thirty years, but there is remarkably little historical treatment of this phenomenon. To begin to fill this gap, evidence from company and official archives is used to expose the anatomy of a rogue trading scandal at Lloyds Bank International in 1974. The rush to internationalize, the conflict between rules and norms, and the failure of internal and external checks all contributed to the largest single loss of any British bank to that time. The analysis highlights the dangers of inconsistent norms and rules even when personal financial gain is not the main motive for fraud, and shows the important links between operational and market risk. This scandal had an important role in alerting the Bank of England and U.K. Treasury to gaps in prudential supervision at the end of the Bretton Woods pegged exchange-rate system.

Business History Review, Volume 91 (1 – April 2017): 105-128.

DOI: https://doi.org/10.1017/S0007680517000381

Review by Adrian E. Tschoegl (The Wharton School of the University of Pennsylvania)

Since the 1974 rogue trading scandal at Lloyds’s Lugano branch we have seen more spectacular sums lost in rogue trading scandals. What Dr Catherine Schenk brings to our understanding of these recurrent events is the insight that only drawing on archives, both at Lloyds and at the Bank of England, can bring. In particular, the archives illuminate the decision processes at both institutions as the crisis unfolded. I have little to add to her thorough exposition of the detail so below I will limit myself to imprecise generalities.

Marc Colombo, the rogue trader at Lloyds Lugano, was a peripheral individual in a peripheral product line, in a peripheral location. As Schenk finds, this peripherality has two consequences, the rogue trader’s quest for respect, and the problem of supervision. Lloyds Lugano is not an anomaly. An examination of several other cases (e.g. Allied Irish, Barings, Daiwa, and Sumitomo Trading), finds the same thing (Tschoegl 2004).

In firms, respect and power come from being a revenue center. Being a cost center is the worst position, but being a profit center with a mandate to do very little is not much better. The rogue traders that have garnered the most attention, in large part because of the scale of their losses were not malevolent. They wanted to be valued. They were able to get away with their trading for long enough to do serious damage because of a lack of supervision, a lack that existed because of the traders’ peripherality.

In several cases, Colombo’s amongst them, the trader was head of essentially a one-person operation that was independent of the rest of the local organization. That meant that the trader’s immediate local supervisor had little or no experience with trading. Heads of branches in a commercial bank come from commercial banking, especially commercial lending. Commercial lending is a slow feedback environment (it may take a long time for a bad decision to manifest itself), and so uses a system of multiple approvals. Trading is a fast feedback environment. The two environments draw different personality types and have quite different procedures, with the trading environment giving traders a great deal of autonomy within set parameters, an issue Schenk addresses and that we will discuss shortly.

Commonly, traders will report to a remote head of trading and to the local branch manager, with the primary line being to the head of trading, and the secondary line being to the local branch manager. This matrix management developed to address the problem of the need to manage and coordinate centrally but also respond locally, but matrix management has its limitations too. As Mathew points out in the New Testament, “No man can serve two masters, for either he will hate the one, and love the other; or else he will hold to the one, and despise the other” (Matthew (6:24). Even short of this, the issue that can arise, as it did at Lloyds Luggano, is that the trader is remote from both managers, one because of distance (and often time zone), and the other because of unfamiliarity with the product line. A number of software developments have improved the situation since 1974, but as some recent scandals have shown, they are fallible. Furthermore, the issue still remains that at some point the heads of many product lines will report to someone who rose in a different product line, which brings up the spectre of “too complex to manage”.

The issue of precautionary or governance rules, and their non-enforcement, is a clear theme in Schenk’s paper. Like the problem of supervision, this too is an issue where one can only do better or worse, but not solve. All rules have their cost. The largest may be an opportunity cost. Governance rules exist to reduce variance, but that means the price of reducing bad outcomes is the lower occurrence of good outcomes. While it is true, as one of Schenk’s interviewees points out, that one does not hear of successful rogue traders being fired, that does not mean that firms do not respond negatively to success. I happened to be working for SBCI, an investment banking arm of Swiss Bank Corporation (SBC), at the time of SBC’s acquisition in 1992 of O’Connor Partners, a Chicago-based derivatives trading house. I had the opportunity to speak with O’Conner’s head of training when O’Connor stationed a team of traders at SBCI in Tokyo. He said that the firm examined too large wins as intently as they examined too large losses: in either case an unexpectedly large outcome meant that either the firm had mis-modelled the trade, or the trader had gone outside their limits. Furthermore, what they looked for in traders was the ability to walk away from a losing bet.

But even small costs can be a problem for a small operation. When I started to work for Security Pacific National Bank in 1976, my supervisor explained my employment benefits to me. I was authorized two weeks of paid leave per annum. When I asked if I could split up the time he replied that Federal Reserve regulations required that the two weeks be continuous so that someone would have to fill in for the absent employee. Even though most of the major rogue trading scandals arose and collapsed within a calendar year, the shadow of the future might well have discouraged the traders, or led them to reveal the problem earlier. Still, for a one-person operation, management might (and in some rogue trading scandals did), take the position that finding someone to fill in and bring them in on temporary duty was unnecessarily cumbersome and expensive. After all, the trader to be replaced was a dedicated, conscientious employee, witness his willingness to forego any vacation.

Lastly, there is the issue of Chesterton’s Paradox (Chesterton 1929). When a rule has been in place for some time, there may be no one who remembers why it is there. Reformers will point out that the rule or practice is inconvenient or costly, and that it has never in living memory had any visible effect. But as Chesterton puts it, “This paradox rests on the most elementary common sense. The gate or fence did not grow there. It was not set up by somnambulists who built it in their sleep. It is highly improbable that it was put there by escaped lunatics who were for some reason loose in the street. Some person had some reason for thinking it would be a good thing for somebody. And until we know what the reason was, we really cannot judge whether the reason was reasonable.”

Finally, an issue one needs to keep in mind in deciding how much to expend on prevention is that speculative trading is a zero-sum activity. A well-diversified shareholder who owns both the employer of the rogue trader and the employers of their counterparties suffers little loss. The losses to Lloyds Lugano were gains to, inter alia, Crédit Lyonnais.

There is leakage. Some of the gainers are privately held hedge funds and the like. Traders at the counterparties receive bonuses not for skill but merely for taking the opposite side of the incompetent rogue trader’s orders. Lastly, shareholders of the rogue traders firm suffer deadweight losses of bankruptcy when the firm, such as Barings, goes bankrupt. Still, as Krawiec (2000) points out, for regulators the social benefit of preventing losses to rogue traders may not exceed the cost. To the degree that costs matter to managers, but not shareholders, managers should bear the costs via reduced salaries.

References

Chesterton, G. K. (1929) ‘’The Thing: Why I Am A Catholic’’, Ch. IV: “The Drift From Domesticity”.

Krawiec, K.D. (2000): “Accounting for Greed: Unraveling the Rogue Trader Mystery”, Oregon Law Review 79 (2):301-339.

Tschoegl, A.E. (2004) “The Key to Risk Management: Management”. In Michael Frenkel, Ulrich Hommel and Markus Rudolf, eds. Risk Management: Challenge and Opportunity (Springer-Verlag), 2nd Edition;

Keynes and Actual Investment Decisions in Practice

Keynes and Wall Street

By David Chambers (Judge Business School, Cambridge University) and Ali Kabiri (University of Buckingham)

Abstract: This article examines in detail how John Maynard Keynes approached investing in the U.S. stock market on behalf of his Cambridge College after the 1929 Wall Street Crash. We exploit the considerable archival material documenting his portfolio holdings, his correspondence with investment advisors, and his two visits to the United States in the 1930s. While he displayed an enthusiasm for investing in common stocks, he was equally attracted to preferred stocks. His U.S. stock picks reflected his detailed analysis of company fundamentals and a pronounced value approach. Already in this period, therefore, it is possible to see the origins of some of the investment techniques adopted by professional investors in the latter half of the twentieth century.

Source: Business History Review (2016), 90(2,Summer), pp. 301-328 (Free access from October 4 to 18, 2016).

Reviewed by Janette Rutterford (Open University)

This short article looks at Keynes’ purchases of US securities in the period from after the Wall Street Crash until World War II. The investments the authors discuss are not Keynes’ personal investments but are those relating to the discretionary fund (the ‘Fund’) which formed part of the King’s College, Cambridge endowment fund and which was managed by Keynes. The authors rely for their analysis on previously unused archival material: the annual portfolio holdings of the endowment fund; the annual report on discretionary fund performance provided by Keynes to the endowment fund trustees; correspondence between Keynes and investment experts; and details of two visits by Keynes to the US in 1931 and 1934.

The authors look at various aspects of the investments in US securities made by Keynes. They first note the high proportion of equities in the endowment fund as a whole. They then focus in detail on the US holdings which averaged 33% by value of the Fund during the 1930s. They find that Keynes invested heavily in preferred stock, which he believed had suffered relatively more than ordinary shares in the Wall Street Crash and, in particular, where the preference dividends were in arrears. He concentrated on particular sectors – investment trusts, utilities and gold mining – which were all trading at discounts to underlying value, either to do with the amount of leverage or with the price of gold. He also made some limited attempts at timing the market with purchases and sales, though the available archival data for this is limited. The remainder of the paper explores the type of investment advice Keynes sought from brokers, and from those finance specialists and politicians he met on his US visits. The authors conclude that he used outside advice to supplement his own views and that, for the Fund, as far as investment in US securities was concerned, he acted as a long-term investor, making targeted, value investments rather than ‘following the herd’.

This paper adds a small element to an area of research which is as yet in its infancy: the analysis of actual investment decision making in practice, and the evolution of investment strategies over time. In terms of strategies, Keynes used both value investing and, to a lesser extent, market timing for the Fund. Keynes was influenced by Lawrence Smith’s 1925 book which recommended equity investment over bond investment on the basis of total returns (dividends plus retained earnings) rather than just dividend yield, the then common equity valuation method. Keynes appears not to have known Benjamin Graham but came to the same conclusion – namely that, post Wall Street Crash, value investing would lead to outperformance. He experimented with market timing in his own personal portfolio but only to a limited extent in the Fund. He was thus an active investor tilting his portfolio away from the market, by ignoring both US and UK railway and banks securities. Another fascinating aspect which is only touched on in this paper is the quality of investment advice at the time. How does it stack up compared to current broker research?

The paper highlights the fact that issues which are still not settled today were already a concern before WWII. Should you buy the market or try to outperform? What is the appropriate benchmark portfolio against which to judge an active strategy? How should performance be reported to the client (in this case the trustees) and how often? How can one decide how much outperformance comes from the asset allocation choice of shares over bonds, from the choice of a particular sector, at a particular time, whilst making allowance for forced cash outflows or sales such as occurred during WWII? More research on how these issues were addressed in the past will better inform the current debate.

Coinucopia: Dealing with Multiple Currencies in the Medieval Low Countries

Enter the ghost: cashless payments in the Early Modern Low Countries, 1500-1800

by Oscar Gelderblom and Joost Jonker (both at Utrecht University)

Abstract: We analyze the evolution of payments in the Low Countries during the period 1500-1800 to argue for the historical importance of money of account or ghost money. Aided by the adoption of new bookkeeping practices such as ledgers with current accounts, this convention spread throughout the entire area from the 14th century onwards. Ghost money eliminated most of the problems associated with paying cash by enabling people to settle transactions in a fictional currency accepted by everyone. As a result two functions of money, standard of value and means of settlement, penetrated easily, leaving the third one, store of wealth, to whatever gold and silver coins available. When merchants used ghost money to record credit granted to counterparts, they in effect created a form of money which in modern terms might count as M1. Since this happened on a very large scale, we should reconsider our notions about the volume of money in circulation during the Early Modern Era.

URL: https://ideas.repec.org/p/ucg/wpaper/0074.html

Distributed by NEP-HIS on: 2015-11-21

Review by Bernardo Batiz-Lazo

In a recent contribution to the Payments Journal, Mira Howard noted:

It’s no secret that the payments industry has been undergoing a period of enormous growth and innovation. Payments has transformed from a steadfast, predictable industry to one with solutions so advanced they sound futuristic. Inventions such as selfie-pay, contactless payments, crypto currency, and biotechnology are just examples of the incredible solutions coming out of the payments industry. However, many payments companies are so anxious to deliver “the future” to merchants and consumers that they overlook merchants that are still stuck using outdated technologies.

The paper by Gelderblom and Jonker is timely and talks to the contemporary concerns of Mira Howard by reminding us of the long history of innovation in retail payments. Specifically, the past and (in their view) under appreciated use of ledger technology (you may want to read its current application behind Bitcoin inThe Economist Insights).

Gelderblom and Jonker set out to explain high economic growth in the Low Countries during the 17th and 18th centuries in a context of scarce media to pay by cash given low coinage, recurrent debasements and devaluations. Their argument is that scarcity of cash did not force people to use credit. Instead silver and gold coins were used as a store of value while daily transactions were recorded in ledgers while translated into a “fictional” currency (“a fictive currency, money of account or ghost money”, p. 7). This provided a common denominator in the use of different types of coin. For instance they cite a merchant house in Leiden transacting in 28 different coin types.

Gelderblom and Jonker build their argument using different sources including a re-examination of relevant literature, probates and merchant accounts. Together they build a fascinating and thought provoking mosaic of the financial aspects everyday life in the Early Modern age. One can only praise Gelderblom and Jonker for their detail treatment of these sources, including a balanced discussion on the potential limitations and bias they could introduce to their study (notably their discussion on probate data).

Comment

The use of a unit of account in a ledger to deal with multiple currencies was by no means unique to the Low Countries nor to the Medieval period. For instance, early Medieval accounting records of the Cathedral of Seville followed the standard practice of keeping track of donations using “maravadies” while 19th century Kuwaiti merchant arithmetic of trade across the Indian Ocean and the Persian Gulf was expressed in Indian rupees [1]. Gelderblom and Jonker, however, go a step beyond using trends in probate data to explore whether there was widespread use of credit and also, extant literature to determine the scarcity of different coins and precious metals.

As part of their arguments Gelderblom and Jonker also question the “efficiency” of the so called “stage theory of money”. This echoes calls that for some time economic anthropologist have made, as they have provided empirical support questioning notion of the barter economy prior to the emergence of money and thus pointing to the illusion of the “coincidence of and wants” (for a quick read see The Atlantic on The Myth of the Barter Economy and for an in depth discussion see Bell, 2001). The same sources agree that the Middle Ages was a second period of demonetization. Moreover, systems of weight and measures, both being per-conditions for barter, were in place by the Early Modern period in Europe then a barter or credit economy rather than the gift economy that characterized pre-monetary societies was a possible response to the scarcity of cash. Gelderblom and Jonker provide evidence to reject the idea of a credit economy while conclude that “barter was probably already monetized” (p. 18) and therefore

“we need to abandon the stage theory of monetization progressing from barter via chas to credit because it simply does not work. … we need to pus the arguments of Muldrew, Vickers, and Kuroda further and start appreciating the social dimensions of payments”.(pp. 18-19)

I could not agree more and so would, I presume, Georg Simmel, Bill Maurer, Viviana Zelizer, Yuval Millo and many others currently working around the sociology of finance and the anthropology of money.

References and Notes

Bell, Stephanie. 2001. “The Role of the State in the Hierarchy of Money.” Cambridge Journal of Economics 25 (149-163).

[1] Many thanks to Julian Borreguero (Seville) and Madihah Alfadhli (Bangor) for their comments.

By failing to prepare, you are preparing to fail

The European Crisis in the Context of the History of Previous Financial Crisis

by Michael Bordo & Harold James

Abstract – There are some striking similarities between the pre 1914 gold standard and EMU today. Both arrangements are based on fixed exchange rates, monetary and fiscal orthodoxy. Each regime gave easy access by financially underdeveloped peripheral countries to capital from the core countries. But the gold standard was a contingent rule—in the case of an emergency like a major war or a serious financial crisis –a country could temporarily devalue its currency. The EMU has no such safety valve. Capital flows in both regimes fuelled asset price booms via the banking system ending in major crises in the peripheral countries. But not having the escape clause has meant that present day Greece and other peripheral European countries have suffered much greater economic harm than did Argentina in the Baring Crisis of 1890.

URL: http://EconPapers.repec.org/RePEc:bog:spaper:18

Circulated by NEP-HIS on: 2015-01-26

Reviewed by: Stephen Billington (Queen’s University of Belfast)

Summary

In this paper Bordo and James seek to analyse the impact of the financial crisis of 2007-8 in the context of previous crisis. Specifically by comparing the experience of periphery countries of the Eurozone with those of the “classic” Gold Standard.

In their paper Bordo and James give a synopsis of the similarities which emerged between both monetary regimes. By adhering to a gold parity there was an expansion in the banking system, through large capital inflows, which was underpinned by a strong effective state to allow for greater borrowing. A nation was effective if it held an international diplomatic commitment, which in turn required them to sign into international systems, all the while this played into the hands of radical political parties who played on civilian nationalism[1]; these events combined lead to great inflows of capital into peripheral countries which inevitably led to fiscal instability and a resulting crisis. Similar dilemmas occurred within the EMU, but much more intensely.

This brings me to the main point that the authors emphasize, that of the contingency rule of the classic gold standard. The latter allowed member countries a “safety valve for fiscal policy”. Essentially this was an escape clause that permitted a country to temporarily devalue its currency in an emergency, such as the outbreak of war or a financial crisis, but would return to normalcy soon after, that is, they would return to previous levels. Bordo and James’ argument is that this lack of a contingency within the EMU allowed for a more severe financial crisis to afflict the periphery countries (Greece, Ireland and Portugal) than had affected gold standard peripheries (Argentina, Italy and Australia) as modern day EMU countries did (and do) not have to option to devalue their currency.

Bordo and James point out that crisis during the gold standard were very sharp, but did not last as long as the 2007-8 crisis. This because the exclusion clause during the gold standard enabled a “breathing space” and as a result most countries were back to growth within a few short years. The EMU story is quite different, say Bordo and James. Mundell (1961) argued that a successful monetary union requires the existence of a well-functioning mechanism for adjustment, what we see in the EMU are a case of worse dilemmas due primarily to this absence of an escape clause.

“Gold outflows, and, with money and credit growth tied to gold, lower money and credit growth. The lower money and credit growth would cause prices and wages to fall (or would lead to reductions in the growth rates of prices and wages), helping to restore competitiveness, thus eliminating the external deficits”

The above quote provided by Gibson, Palivos and Tavlas (2014) highlights how the gold standard allowed a country to adjust to a deficit. This point reinforces how Bordo & James argue that due to the constricting nature of the EMU there is no “safety valve” to allowed EU countries to release the steam from increasing debt levels. With respect to the Argentine Baring Crisis of 1890, while the crisis was very sharp in terms of real GDP, pre-crisis levels of GDP were again reached by 1893 – clearly a contrast with the Euro as some countries are still in recession with very little progress having been made as suggested by the following headline: “Greece’s current GDP is stuck in ancient Greece” – Business Insider (2013).

The following graph highlights the issue that in Europe most countries are still lagging behind the pre-crisis levels of GDP.

Bordo and James clearly support this argument. Delles and Tavlas (2013) also argue that the adjustment mechanism of core and periphery countries limited the size and persistence of external deficits. They put forward that the durability of the gold standard relied on this mechanism. This is reinforced by Bloomfield (1959) who states it “facilitated adjustments to balance of payments disequilibrium”.

Vinals (1996) further supports the authors’ sentiments by arguing that the Treaty of Maastricht restricts an individual member’s room to manoeuvre as the Treaty requires sound fiscal policies, with debt limited to 60% of GDP and annual deficits no greater than 3% of GDP – meaning a member cannot smooth over these imbalances through spending or taxation.

Gibson, Palivos and Tavlas (2014) state “a major cost of monetary unions is the reduced flexibility to adjust to asymmetric shocks”. They argue that internal devaluations must occur to adjust to fiscal imbalances, but go on to argue that these are much harder to implement than in theory, again supported by Vinals (1996).

Comment

Bordo and James focus primarily on three EU periphery countries which are doing badly, namely Greece, Ireland and Portugal. However they neglect the remaining countries within the EU which can also be classed as a periphery. According the Wallerstein (1974) the periphery can be seen as the less developed countries, these could include further countries such as those from eastern Europe[2]. By looking at a more expansive view of peripheral countries we can see that these other peripherals had quick recoveries with sharp decreases in GDP growth, as in the case of the Gold standard countries, but swiftly recovered to high levels of growth again while the main peripheral countries the authors analyse do lag behind.

See note 3

Bordo and James do provide a strong insight into the relationship between an adjustment mechanism to combating fiscal imbalances as a means of explaining the poor recovery of certain peripheral countries (i.e. Greece, Ireland, Portugal) and highlight the implications of this in the future of the EMU. If the EMU cannot find a contingency rule as the gold standard then recessions may leave them as vulnerable in the future as they are now.

Notes

1) This process can be thought of as a trilemma, Obstfeld, Taylor and Shambaugh (2004) give a better explanation. In the EU the problem was intensified as governments could back higher levels of debt, and there was no provision for European banking supervision, the commitment to EU integration let markets believe that there were no limits to debt levels. This led to inflows in periphery countries where banks could become too big to be rescued.

2) Latvia, Lithuania, Slovakia, Slovenia, and even Cyprus can be included based on low GDP per capita which is equivalent to Greece.

3) Data taken from Eurostat comparing real GDP growth levels of lesser developed countries within the Eurozone who all use the euro and would be locked into the same system of no adjustment.

References

Bloomfield, A. (1959) Monetary Policy under the International Gold Standard. New York: Federal Reserve Bank of New York.

Business Insider (2013). Every Country in Europe Should be Glad it’s Not Greece. http://www.businessinsider.com/european-gdp-since-pre-crisis-chart-2013-8?IR=T [Accessed 19/03/2015].

Eurostat, Real GDP Growth Rates http://ec.europa.eu/eurostat/tgm/graph.do?tab=graph&plugin=1&pcode=tec00115&language=en&toolbox=data [Accessed 21/03/2015].

Dellas, Harris; Tavlas, George S. (2013). The Gold Standard, The Euro, and The Origins of the Greek Sovereign Debt Crisis. Cato Journal 33(3): 491-520.

Gibson, Heather D; Palivos, Theodore; Tavlas, George S. (2014). The Crisis in the Euro Area: An Analytic Overview.Journal of Macroeconomics 39: 233-239.

Mundell, Robert A. (1961). A Theory of Optimum Currency Areas. The American Economic Review 51(4): 657-665.

Obstfeld, Maurice. Taylor, Alan. Shambaugh, Jay C. (2004). The Trilemma in History: Trade-Offs among Exchange Rates, Monetary Policies and Capital Mobility. National Bureau of Economic Research (NBER working paper 10396).

Vinals, Jose. (1996). European Monetary Integration: A Narrow or Wide EMU?. European Economic Review 40(3-5): 1103-1109.

Wallerstein, Immanuel (1974). The Modern World-System I: Capitalist Agriculture and the Origins of the European World-Economy in the Sixteenth Century. New York: Academic Press.

Technology and Financial Inclusion in North America

Did Railroads Make Antebellum U.S. Banks More Sound?

By Jeremy Atack (Vanderbilt), Matthew Steven Jaremski (Colgate), and Peter Rousseau (Vanderbilt).

Abstract: We investigate the relationships of bank failures and balance sheet conditions with measures of proximity to different forms of transportation in the United States over the period from 1830-1860. A series of hazard models and bank-level regressions indicate a systematic relationship between proximity to railroads (but not to other means of transportation) and “good” banking outcomes. Although railroads improved economic conditions along their routes, we offer evidence of another channel. Specifically, railroads facilitated better information flows about banks that led to modifications in bank asset composition consistent with reductions in the incidence of moral hazard.

URL: http://econpapers.repec.org/paper/nbrnberwo/20032.htm

Review by Bernardo Bátiz-Lazo

Executive briefing

This paper was distributed by NEP-HIS on 2014-04-18. Atack, Jaremski and Rousseau (henceforward AJR) deal with the otherwise thorny issue of causation in the relationship between financial intermediation and economic growth. They focus on bank issued notes rather deposits; and argue for and provide empirical evidence of bi-directional causation based on empirical estimates that combine geography (ie GIS) and financial data. The nature of their reported causation emerges from their approach to railroads as a transport technology that shapes markets while also shaped by its users.

Summary

In this paper AJR study the effect of improved means of communication on market integration and particularly whether banks in previously remote areas of pre-Civil War USA had an incentive to over extend their liabilities. AJR’s paper is an important contribution: first, because they focus on bank issued notes and bills rather than deposits to understand how banks financed themselves. Second, because of the dearth of systematic empirical testing whether the improvements in the means of communication affected the operation of banks.

In 19th century north America and in the absence of a central bank, notes from local banks were substitutes among themselves and between them and payment in species. Those in the most remote communities (ie with little or no oversight) had an opportunity to misbehave “in ways that compromised the positions of their liability holders” (behaviour which AJR label “quasi-wildcatting”). Railroads, canals and boats connected communities and enabled better trading opportunities. But ease of communication also meant greater potential for oversight.

ACJ test bank failure rates (banks that didn’t redeem notes at full value), closed banks (ceased operation but redeem at full value), new banks and balance sheet management for 1,818 banks in existence in the US in 5 year increments between 1830 and 1862. Measures of distance between forms of communication (i.e. railroads, canals, steam navegable river, navegable lake and maritime trade) and bank location emerged from overlapping contemporary maps with GIS data. Financial data was collected from annual editions of the “Merchants and Bankers’ Almanac”. They distinguish between states that passed “free banking laws” (from 1837 to the early 1850s) and those that did not. They also considered changes in failure rates and balance sheet variance (applying the so called CAMEL model – to the best of data availability) for locations that had issuing banks before new transport infrastructure and those where banks appear only after new means of communication were deployed:

Improvements in finance over the period also provided a means of payment that promoted increasingly impersonal trade. To the extent that the railroads drew new banks closer to the centers of economic activity and allowed existing banks to participate in the growth opportunities afforded by efficient connections.(p. 2)

Railroads were the only transport technology that returned statistically significant effects. It suggested that the advent of railroads did indeed pushed bankers to reduce the risk in their portfolios. But regardless of transport variables, “[l]arger banks with more reserves, loans, and deposits and fewer bank notes were less likely to fail.” (p.20). It is thus likely that railroads impact banks’ operation as they brought about greater economic diversity, urbanisation and other measures of economic development which translated in larger volume of deposits but also greater scrutiny and oversight. In this sense railroads (as exogenous variable) made banks less likely to fail.

But ACJ note that means of transportation were not necessarily exogenous to banks. Reasons for the endogeneity of transport infrastructure included bankers promoting and investing in railroads to bring them to their communities. Also railways could find advantages to expand into vigorously active locations (where new banks could establish to capture a growing volume of deposits and serve a growing demand for loans).

Other empirical results include banks decreased the amount of excess reserves, notes in circulation and bond holdings while also increased the volume of loans after the arrival of a railroad. In short, considering railroads an endogenous variable also results in transport technologies lowering bank failure rates by encouraging banks to operate more safely.

Comment

The work of AJR is part of a growing and increasingly fruitful trend which combines GPS data with other more “traditional” sources. But for me the paper could also inform contemporary debates on payments. Specifically their focus is on banks of issue, in itself a novelty in the history of payment systems. For AJR technological change improves means of payment when it reduces transaction costs by increasing trust on the issuer. But as noted above, there are a number of alternative technologies which have, in principle, equal opportunity to succeed. In this regard AJR state:

Here, we describe a mechanism by which railroads not only affected finance on the extensive margin, but also led to efficiency changes that enhanced the intensity of financial intermediation. And, of course, it is the interaction of the intensity of intermediation along with its quantity that seems most important for long-run growth (Rousseau and Wachtel 1998, 2011). This relationship proves to be one that does not generalize to all types of transportation; rather, railroads seem to have been the only transportation methods that affected banks in this way.(p4)

In other words, financial inclusion and improvements in the payment system interact and enhance economic growth when the former take place through specific forms of technological change. It is the interaction with users that which helps railroads to dominate and effectively change the payments system. Moreover, this process involves changes in the portfolio (and overall level of risk) of individual banks.

The idea that users shape technology is not new to those well versed in the social studies of technology. However, AJR’s argument is novel not only for the study of the economic history of Antibellum America but also when considering that in today’s complex payments ecosystem there are a number or alternatives for digital payments, many of which are based on mobile phones. Yet it would seem that there is greater competition between mobile phone apps than between mobile and other payment solutions (cash and coins, Visa/Mastercard issued credit cards, PayPal, Bitcoin and digital currencies, etc.). AJR results would then suggest that, ceteris paribus, the technology with greater chance to succeed is that which has great bi-directional causality (i.e. significant exogenous and endogenous features). So people’s love for smart phones would suggest mobile payments might have greater chance to change the payment ecosystem than digital currencies (such as Bitcoin), but is early days to decide which of the different mobile apps has greater chance to actually do so.

Wall Street (1867)

Another aspect in which AJR’s has a contemporary slant refers to security and trust. These are key issues in today’s digital payments debate, yet the possibility of fraud is absence from AJR’s narrative. For this I mean not “wildcatting” but ascertaining whether notes of a trust worthy bank could have been forged. I am not clear how to capture this phenomenon empirically. It is also unlikely that the volume of forged notes of any one trusted issuer was significant. But the point is, as Patrice Baubeau (IDHES-Nanterre) has noted, that in the 19th century the technological effort for fraud was rather simple: a small furnace or a printing press. Yet today that effort is n-times more complex.

AJR also make the point that changes in the payments ecosystem are linked to bank stability and the fragility of the financial system. This is an argument that often escapes those discussing the digital payments debate.

Overall it is a short but well put together paper. It does what it says on the can, and thus highly recommended reading.