Special articles
Neoclassical Finance and the Fully Convertible Rupee
Recent discussions about full capital account convertibility have addressed the risk of such a move in terms of concepts such as crisis, contagion, fragility and crashes. Such terms seem as amenable to a sociological analysis as an economic one. Arguments about convertibility should be grounded in theories about financial markets from both these disciplines.
SIVAKUMAR ARUMUGAM
T
The group argues forcefully in their letter that the problem with implementing full convertibility is that it “would expose the Indian economy to extreme volatility” and, further, it “should be noted that the Indian economy could insulate itself from the contagion of the south-east Asian financial crisis in the late 1990s only because the capital controls in place at the time prevented destabilising movements of capital” [Krishnaswamy et al 2006:1222]. Although they go on to assert that the “gains to the economy from this move are nil”, it seems that the core point is one of uncertainty: “According to the government’s own pronouncements, the economy is currently doing very well. In such a context, exposing the country to unpredictable movements in capital flows creates a potential for fragility and crisis that is completely avoidable. This is of particular concern because the stock market is already experiencing a speculative boom” [Krishnaswamy et al 2006:1222].
Such concepts as uncertainty, crisis, contagion and fragility are also deployed in arguing against (or against the current pace of) liberalisation more generally in India. But, these concepts appear to be eminently sociological1 ones. What role then do these concepts play in economic theory about markets in money?
And what might an sociological alternative look like? These questions strike at the contemporary viability of theoretical justifications for liberalisation.
Financial Liberalisation in India
Rather fortuitously, the Economic and Political Weekly carried an issue devoted to ‘Money, Banking and Finance’ just a week before Manmohan Singh’s announcement. And, in addition, the May 13-19, 2006, issue carries more articles devoted specifically to capital account convertibility.
An article by C P Chandrasekhar and Parthapratim Pal in the ‘Money, Banking and Finance’ issue made some of same points that the group letter subsequently did. They argue that “liberalisation of controls on inflows and outflows of capital...have resulted in an increase in financial fragility in developing countries, making them prone to periodic financial and currency crisis” [Chandrasekhar and Pal 2006: 975]. With respect to the Indian case, they date the current period of liberalisation to the balance of payments crisis in 1990-91: “the crisis triggered a major reform of the financial sector that has been unfolding since, but has yet to culminate in full convertibility on the capital account” [Chandrasekhar and Pal 2006:975].
Chandrasekhar and Pal first attempt to establish that inflows of foreign institutional investors (FII) money and general upward movement of the stock market are closely related. In addition, they suggest that “it is well recognised that stock market buoyancy and volatility have been a phenomenon typical of the liberalisation years”2 [Chandrasekhar and Pal 2006:977]. They also point to a change in the tax regime for equity investment in 2003: investment in equities for longer than one year would no longer be subject to a capital gains tax.3 They give three reasons as to why markets in India tend to be shallow or thin: firstly, that “only stocks of a few companies are actively traded in the market”, secondly, that “only a small proportion [of stocks are] routinely available for trading, with the rest being held by promoters, the financial institutions and others interested in corporate control or influence”, and, thirdly, that “the number of players trading these stocks is small” [Chandrasekhar and Pal 2006:979]. This is held to be important because “[i]n as much as an increase in investment by FIIs triggers a sharp price increase, it would provide additional incentives for FII investment and in the first instance encourage further purchases...And when the correction begins, it would have to be led by an FII pull-out and can take the form of an extremely sharp decline in prices”. Consequently, “as and when FIIs are attracted to the market by expectations of price increases that tend to be automatically realised, the inflow of foreign capital can result in an appreciation of the rupee vis-à-vis the dollar”, for example, which would increase the “return earned in foreign exchange, when rupee assets are sold and the revenue converted into dollars”. Also, “the growing realisation by the FIIs of the power they wield in what are shallow markets, encourages speculative investment aimed at pushing the market up and choosing an appropriate moment to exit” [Chandrasekhar and Pal 2006:979]. And, finally, the behaviour of FIIs will at least in part be determined by “events outside the country, whether it be the performance of other equity markets or developments in non-equity markets elsewhere in the world”, such that “when they make their portfolio adjustments, which may imply small shifts in favour of or against a country like India, the effects it has on host markets are substantial” [Chandrasekhar and Pal 2006:979].
Chandrasekhar and Pal go on to examine the consequences of financial liberalisation for the banking sector and the real economy. They argue that liberalisation, and therefore the increasingly influential behaviour of investors (principally, FIIs), has forced the state to adopt a “deflationary macroeconomic stance, which adversely affects public capital formation and the objectives of promoting employment and reducing poverty” as well as resulting in a “credit squeeze for the commodity producing sectors and a decline in credit delivery to rural India and smallscale industry”, and, lastly, that the “belief that the financial deepening that results from liberalisation would in myriad ways neutralise these effects has not been realised” [Chandrasekhar and Pal 2006:988].
The articles published in the more recent May 13-19, 2006, issue of Economic and Political Weekly similarly list the dangers of capital account convertibility in India. Some emphasise that the pace of the liberalisation process ought at least be slowed down. John Williamson, for example, places other liberalisations ahead of finance: “there are many other liberalising reforms – from electricity pricing to making courts work expeditiously to pruning the fiscal deficit – that deserve to be priorities over complete capital account liberalisation...” [Williamson 2006:1850]. Similarly, but with respect to finance only, Partha Sen finds that there are three conditions that ought first to be met: “(a) financial sector reforms, (b) fiscal balance; and (c) properly designed monetary (and exchange rate) policy” [Sen 2006:1855]. Amitava Krishna Dutt lists some of the benefits that convertibility can give to the real economy: “The economic logic of the liberalisation of the capital account for less developed countries...is related to the goals of increasing capital inflows in order to increase GDP and growth (and hence contribute to overall economic development) and of smoothing consumption through international borrowing” [Dutt 2006:1850]. Dutt finds, however, that the practical result is likely to be one of boombust capital flow cycles where the “instability caused by capital flows further increases uncertainty and reduces investment” [Dutt 2006:1852].
All the accounts surveyed above leave the theory of financial markets themselves untouched. The writers cleave only to arguments about the possibility of added risk in the form of increased variation of prices and to suggestions of the benefits and costs of financial liberalisation to the wider economy as such. What remains is an account not only of the standard theories of financial markets (and implicitly how to reform them), but also the possibility that there might be some dispute within the sub-speciality of financial economics on the topic at hand. This becomes doubly important given the wide overlap between theorists of finance and practitioners of finance. Very often, the very same people propose theories about financial market behaviour and, at the same time, act in those markets and/or advise institutions like the IMF and the US government about how to form and reform those markets.
Neoclassical Finance and the Efficient Markets Hypothesis
Textbooks on economics invariably concentrate on discussions of the efficient markets hypothesis in covering financial markets. A very popular economics textbook in the US by Paul A Samuelson and William H Nordhaus that was first published in 1948 (by Samuelson alone) and that has been continually revised since, begins with an account of the relationship between risk and reward: “Individuals generally prefer higher return, but they also prefer lower risk because they are risk-averse” [Samuelson and Nordhaus 2005:521, emphasis in original]. The result is that such individuals will require a higher return for agreeing to purchase riskier assets, where “[e]conomists often measure risk as the standard deviation of returns; this is a measure of dispersion whose range encompasses about two-thirds of the variation” [Samuelson and Nordhaus 2005:521]. They go from that explanation of the relationship between reward and risk to a section on “bubbles and crashes” that contains some of the language used by opponents of financial liberalisation: “History is marked by bubbles in which speculative prices were driven up far beyond the intrinsic value of the asset...Speculative bubbles always produce crashes and sometimes lead to economic panics” [Samuelson and Nordhaus 2005:523].
This dire warning gives way to a section on ‘Efficient Markets and the Random Walk’ that takes the reader directly onto the central terrain of neoclassical economic thought as it relates to finance: “An efficient financial market is one where all new information is quickly understood by market participants and becomes immediately incorporated into market prices”. This immediately gives markets the following property, argue Samuelson and Nordhaus: “It is not possible to make profits by acting on old information or at patterns of past price changes. Returns on stocks [or any other financial asset] will be primarily determined by their riskiness relative to the market” [Samuelson and Nordhaus 2005:525]. As news is by definition “new” and therefore unpredictable, it also seems to follow that prices of assets over time will move in an entirely random manner: “[b]ecause stock prices move in response to erratic events, stock prices themselves move erratically, like a random walk” [Samuelson and Nordhaus 2005:525].
Samuelson and Nordhaus link together the section on speculative bubbles and crashes with the section on efficient markets by noting, first, that “[e]conomists who look at the historical record ask whether it is plausible that sharp movements in stock prices could actually reflect new information” such as with the “30 per cent drop in stock prices that occurred from October 15-19, 1987” and, second, that “the efficient-market view applies to individual stocks but not necessarily to the market as a whole”. On the latter point, they suggest that large-scale stock market price movements might “reflect changes in the general mood of the financial community”. Under such conditions, particular participants “could not individually buy or sell enough stocks to overcome the entire national mood” [Samuelson and Nordhaus 2005:525]. In the case of financial liberalisation in less developed countries, however, this last point is usually taken not to hold, hence the concern over FII flows of money and its effect on overall markets.
Samuelson and Nordhaus here implicitly connect together the efficient markets hypothesis and something that has come to be called the “fundamental theorem of finance”. This theorem was first stated explicitly by Stephen A Ross in 1973. Ross has written recently that “the basic theorem and its attendant results have unified our understanding of asset pricing and the theory of derivatives, and have generated an enormous literature that has had a significant impact on the world of financial practice”. The theorem provides a framework in which to derive the value of any given asset in terms of the constitute parts of that asset: “[t]he basic intuition that underlies valuation is the absence of arbitrage” where an “arbitrage opportunity is an investment strategy that guarantees a positive pay-off in some contingency with no possibility of a negative pay-off and with no initial net investment” [Ross 2005:1].
Ross finds this no arbitrage condition “compelling because it appeals to the most basic beliefs about human behaviour, namely that there is someone who prefers having more wealth to having less” and, given that “save for some anthropologically interesting societies, a preference for wealth appears to be a ubiquitous human characteristic, it is certainly a minimalist requirement” [Ross 2005:2]. The no arbitrage condition allows market participants to accurately derive the value of any given asset. This in turn would seem to allow such investors to recognise when any such asset is mispriced in the market: “[t]he “smart” money ferrets out arbitrages and eliminates them...” [Ross 2005:63]. This is usually deemed to work both with individual assets and at the level of whole markets, where, for example, emerging market funds raise money by issuing equity or debt in a developed country and invest those funds in some mix of equities and debt in a less developed country.
Ross implicitly agrees in part with Samuelson and Nordhaus in that he finds “some bothersome facts that are difficult to reconcile with our intuition about market efficiency”. He is “particularly troubled that contemporaneous news seems to explain so little of the contemporaneous motion of prices” [Ross 2005:64]. But, importantly for Ross, the magnitude of such inefficiencies appears to be very small. He estimates that hedge funds and other participants that explicitly market themselves to investors as being able to uncover market inefficiencies make annual excess profits of (generously estimated) approximately $40 billion dollars: “[t]his is a tidy sum, but not when measured against a rough estimate of $50 trillion of marketed assets”. In such a case, “prices are “off” or inefficient to about $40 billion/$50 trillion, which is less than 0.1 per cent” [Ross 2005:64].
Arbitrage Inefficiencies
One crucial consequence of the no arbitrage theorem is that it appears to solve neatly a problem that has been posed to neoclassical economics since its first formulations in the late 19th century, namely its insistence on treating all consumers as rational self-interested actors with well-defined preferences. In a general market for goods and services, many of the important results in neoclassical economics depends upon this assumption. In financial markets, the question of efficiency is posed differently: here, all that is required is that at least some participants seek to make rather than lose money (hence Ross’s claim about “anthropologically interesting societies”).
A group of behavioural economists have explored some of the limits of this assumption. Owen A Lamont and Richard H Thaler suggest that the “first law of economics is clearly the law of supply and demand”, but they “nominate as the second law ‘the law of one price’, hereafter simply the law”, where the “law states that identical goods must have identical prices” [Lamont and Thaler 2003:191]. They note that the “absence of arbitrage opportunities is the basis of almost all modern financial theory, including option pricing and corporate capital structure”. In particular, “the law says that identical securities (that is, securities with identical state-specific payoffs) must have identical prices; otherwise, smart investors could make unlimited profits by buying the cheap one and selling the expensive one”, and, further, that “it does not require that all investors be rational or sophisticated, only that enough investors (dollar weighted) are able to recognise arbitrage opportunities” [Lamont and Thaler 2003:192]. However, the authors argue, “it turns out that the application of the Law in financial markets is not as uncontroversial as was originally thought” [Lamont and Thaler 2003:192].
They examine a number of possible violations of the no arbitrage theorem. In one kind of example, arbitrage might not be possible for legal reasons. Infosys, the Indian technology company, listed American Depositary Receipts on the Nasdaq stock exchange.4 By the March 7, 2000, Infosys was trading at $ 335 on the Nasdaq: “The ADR was trading at a 136 per cent premium to the Bombay shares” [Lamont and Thaler 2003:194]. As Lamont and Thaler quickly point out, however, “official barriers prevented Americans from buying the shares trading in Bombay, and so there was no way for American arbitrageurs to create new ADRs and thus instantly profit from this relative valuation”. Moreover, in “segmented markets, it can be rational for the same asset to have different prices in different markets, reflecting differences in supply and demand” [Lamont and Thaler 2003:195].
In a related situation, arbitrage might be legally possible but extremely difficult to put into practice. Lamont and Thaler examine a US technology company, 3Com, whose subsidiary, Palm, was the maker of then popular handheld computers. 3Com first sold a small amount of Palm in an initial public offering (IPO) and proposed granting the remainder of Palm to all 3Com shareholders, at a rate of 1.5 shares of Palm for each share of 3Com held. Lamont and Thaler note that “[o]n the day before the IPO, 3Com was selling for $104. The Palm shares were sold to the public at $38 a share, but ended the day selling for $95...During this same day, the stock price of 3Com actually fell 21 per cent during the day to $82” [Lamont and Thaler 2003, 198]. Although there was some (very limited) uncertainty about whether 3Com would in fact be able to distribute 1.5 shares of Palm for each of its own shares, the market in this case appeared to be valuing Palm shares highly but not the Palm shares still embedded within 3Com. Lamont and Thaler consider the “so-called ‘stub value’ of 3Com, which is the implied value of 3Com’s non-Palm assets and businesses...one just has to multiply the Palm share price by 1.5 [approximately, because of the small amount already sold in the IPO] to get $145 and then subtract this from the value of 3Com, obtaining the novel result of a negative $63 per share...” [Lamont and Thaler 2003:198]. Further, this “mispricing was not in an obscure corner of capital markets, but rather took place in a widely publicised initial public offering that attracted frenzied attention...On the day after the issue, the mispricing was widely discussed, including in two articles in the Wall Street Journal, one in the New York Times, and it even made USAToday” [Lamont and Thaler 2003:198]. The problem was that arbitrageurs could not find holders of Palm shares who were willing to lend them their shares, thus enabling the arbitrageurs to sell Palm short and complete the arbitrage by buying 3Com: “In the case of Palm, retail investors rather than institutional investors held most of the shares, thus making Palm hard to borrow” [Lamont and Thaler 2003:199].
Lamont and Thaler recognise explicitly that the examples they have given are not frequently occurring ones, and that there were specific reasons, legal or otherwise, that had prevented arbitrageurs from driving divergent prices together, but the problem is that “these special cases are interesting because they should be situations in which it is particularly easy for the market to get things right” even in the absence of arbitrage possibilities: “If the market is flunking these no-brainers, what else is it getting wrong?“ [Lamont and Thaler 2003:201].
Noise and Arbitrage Failures
Lamont and Thaler mention briefly the peculiar case of Long-Term Capital Management (LTCM).5 They suggest here that when prices of associated assets diverge, it is rarely the case that an arbitrage trade can be put in place without using at least some amount of capital initially. But that immediately implies that any unexpected divergence of prices, even in cases where the arbitrage guarantees a profit over some fixed period of time, can jeopardise the entire trade: “In extreme cases, this widening spread can cause the arbitrageur to approach bankruptcy, as his net worth becomes negative and he no longer has the collateral to hold his positions” [Lamont and Thaler 2003:200].
Sanford J Grossman and Joseph E Stiglitz have constructed a model that suggests that informationally efficient markets are impossible because then “those who arbitrage make no (private) return from their (privately) costly activity”. They suggest an “equilibrium degree of disequilibrium” such that “prices reflect the information of informed individuals (arbitrageurs) but only partially, so that those who expend resources to obtain information do receive compensation” [Grossman and Stiglitz 1980:393]. The model treats all market participants as homogeneous agents, excepting that each one can decide whether to spend some defined sum of money to become informed: “When informed individuals observe information that the return to a security is going to be high, they bid its price up, and conversely when they observe information that the return is going to be low...Thus the price system makes publicly available the information obtained by informed individuals to the uninformed” [Grossman and Stiglitz 1980:393]. But if the efficient markets hypothesis does in fact hold, informed individuals have no advantage over uninformed individuals. Such a situation would not form an equilibrium. Conversely, “[h]aving no one informed is also not an equilibrium, because then each trader, taking the price as given, feels that there are profits to be made from becoming informed” [Grossman and Stiglitz 1980:404]. Thus, insofar as information is costly, an efficient market is impossible. However, Grossman and Stiglitz do demonstrate that “when information is very inexpensive, or when informed traders get very precise information, then equilibrium6 exists and the market price will reveal most of the informed traders’ information”. Unfortunately, because their model treats the individuals as identical to one another, “such markets are likely to be thin because traders have almost homogeneous beliefs” [Grossman and Stiglitz 1980:404]. In general, “prices cannot perfectly reflect the information which is available, since if it did, those who spent resources to obtain it would receive no compensation” [Grossman and Stiglitz 1980:405].
An important paper by J Bradford De Long, Andrei Shleifer, Lawrence H Summers and Robert J Waldmann examines the case of persistent noise in market prices by modelling two kinds of market agents: irrational noise traders with randomly erroneous beliefs and rational arbitrageurs. They suggest that there is “an important source of risk borne by short-horizon investors engaged in arbitrage against noise traders: the risk that noise traders’ beliefs will not revert to their mean for a long time and might in the meantime become even more extreme”. In particular, they observe that “arbitrage does not eliminate the effects ofnoisebecause noise itself creates risk” [De Long et al 1990:705]. The core feature of the model they construct involves a continual uncertainty introduced by the (possible) changes in noise traders’ future beliefs about asset values: “[t]his uncertainty about the price for which (assets) can be sold afflicts all investors, no matter what their beliefs about expected returns, and so limits the extent to which they are willing to bet against each other”. Otherwise, if they held certain, but different, beliefs about future prices, they would “try to take infinite bets against each other” [De Long et al 1990:710]. A further feature is the stipulation that the arbitrageurs have a (relative to the noise traders) short investment horizon. In general, “as the horizon of the (arbitrageur) agents becomes longer, arbitrage becomes less risky and prices approach fundamental values” [De Long et al 1990:713].
The authors find that noise traders in their model need not earn lower returns as a group than the arbitrage traders. The noise traders suffer two problems: as they become (erroneously) more bullish about some asset, they drive up its price, tending thereby to reduce their returns from holding that asset; in addition, because they behave stochastically they may tend to buy the asset just when other noise traders are buying it. As “the variability of noise traders’ beliefs increases, the price risk increases”, but as the arbitrageurs are risk averse, this mean that the arbitrageurs will “reduce the extent to which they bet against noise traders in response to this increased risk”. The net effect is such that if the noise traders are overly bullish their returns will suffer in relation to the arbitrageurs’ returns, and similarly, if they are bearish. However, for “intermediate degrees of average bullishness, noise traders earn higher expected returns” where these returns “come at the cost of holding portfolios with sufficiently higher variance (giving) noise traders lower expected utility” than those of the arbitrageurs [De Long et al 1990:715].
Andrei Shleifer and Robert W Vishny argue (in a later article) that arbitrageurs are likely to be even less aggressive than in the model above. They note that arbitrage price theory assumes the presence of “a very large number of tiny arbitrageurs, each taking an infinitesimal position against the mispricing in a variety of markets” such that “[b]ecause their positions are so small, capital constraints are not binding and arbitrageurs are effectively risk neutral toward each trade”. But, they assert, “millions of little traders are typically not the ones who have the knowledge and information to engage in arbitrage” [Shleifer and Vishny 1997:36]. Rather – this is the “agency relationship” they explore in the paper
– the money placed in the hands of sophisticated arbitrageurs “comes from wealthy individuals, banks, endowments, and other investors with only a limited knowledge of individual markets” [Shleifer and Vishny 1997:36-37]. This relationship is founded, of course, on the claims that arbitrageurs make to their putative and current investors about their knowledge of the markets. Shleifer and Vishny produce a performance based arbitrage model in which such claims are legitimated (solely) by past performance records. In such a situation, their “model shows that the times when they lose money are precisely the times when prices are far away from fundamentals, and in those times the trading by arbitrageurs has the weakest stabilising effect” [Shleifer and Vishny 1997:46]. Further, there may be an internal performance based limitation: “If the boss of the organisation [i e, the arbitrage fund] is unsure of the ability of the subordinate taking a position, and the position loses money, the boss may force a liquidation of the position before the uncertainty works itself out” [Shleifer and Vishny 1997:48]. Even worse, if the arbitrageurs base their estimations of future returns on what they would have earned in the past, that is, if they are “Bayesians with an imprecise posterior about the true distribution of returns”, then “a sequence of poor returns may cause an arbitrageur to update his posterior and abandon his original strategy”. The net result is that, in this model, arbitrageurs “care about total risk, not just systematic risk” [Shleifer and Vishny 1997:52].
A Sociology of Noise
Donald MacKenzie has investigated the strange, but exemplary, case of LTCM. LTCM was a very large arbitrage firm that operated across a variety of markets (equities, bonds, currencies) around the world. MacKenzie notes that typical sociological understandings of economics (of, in particular, Talcott Parsons) has it that the “technical core, so to speak, of the workings of market economics was the business of economists, not of sociologists” [MacKenzie 2003:350]. Contrawise, MacKenzie argues that one can hold arbitrage firms such as LTCM to be “the border guards, in economic practice, of the Parsonian boundary between economics and sociology” [MacKenzie 2003:350]. In other words, MacKenzie holds the activity of arbitrageurs to be responsible for the disciplinary division between economics and sociology. This requires a sociological analysis of the economics derived practice engaged in by arbitrageurs. Another consequence is that arguments within the discipline of economics about noise traders – which by definition are not amenable to sociological analysis because entirely random – work as arguments only in cases where the arbitrageurs do indeed act so as to make market prices efficient.
LTCM, headed by the well known Salomon Brothers bond trader John Meriwether, was “hugely successful: at its peak, it deployed what is almost certainly the largest single concentration of arbitrage positions ever” [MacKenzie 2003:352]. LTCM was founded in 1994 and its gross returns far outmatched even its own initial claims to investors: returns were 28.1 per cent in 1994, and then, 59 per cent and 61.5 per cent in 1995 and 1996, respectively [MacKenzie 2003:355]. However, in late 1998, “in one of the defining moments of the economic history of the 1990s, adverse price movements drove LTCM to the brink of bankruptcy” [MacKenzie 2003:352].
MacKenzie interviewed Meriwether and other partners in LTCM. He notes that “[r]isks were carefully calculated and controlled using the “value-at-risk” approach standard in the world’s leading banks’ and “[u]sing those estimates, it was then possible to work out the relationship between the magnitude of possible losses and their probabilities”. Further, the historical correlations between positions in different markets were monitored: “LTCM was aware that its own and other arbitrageurs’ involvement in these diverse positions would induce some correlation”, so the “standard deviations and correlations what went into LTCM’s aggregate risk model were...not simply the empirically observed figures but deliberately conservative estimates of their future values” [MacKenzie 2003:358]. Furthermore, “LTCM also ‘stress tested’ its portfolio, investigating the consequences of hypothetical events too extreme to be captured by statistical value-at-risk models, events such as a huge stock market crash, bond default by the Italian government, devaluation by China, or (particularly salient given its European involvement) failure of EMU”. But, according to MacKenzie, the traders were not blind to the difficulties of mathematical models as such either: “all those involved knew that models were approximations to reality and a guide to strategy rather than determinate of it” [MacKenzie 2003:359].7
MacKenzie argues that LTCM’s near-bankruptcy8 was a result not of excessive leverage or over reliance on mathematical models but rather the imitation of LTCM by other arbitrageurs acting in the same markets. The result was that in late 1998, the otherwise relatively unimportant “flight-to-quality [in bond markets] triggered by the Russian default” [MacKenzie 2003: 363] meant that as “arbitrageurs began to incur losses, they almost all seem to have reacted by seeking to reduce their positions, and in so doing they intensified the price pressure that had caused them to make the reductions” [MacKenzie 2003:363]. MacKenzie suggests, importantly, that the “effects of imitation run deep: it can, for example, affect the statistical distributions of price changes, causing distributions to become dangerously ‘fat-tailed’ (that is, the probability of extreme events is far higher than implied by standard normal or log-normal distributional assumptions)” [MacKenzie 2003:372]. This imitation effect thus constitutes “a limit on the performativity of economics: under some circumstances, arbitrage may be unable to eliminate what economic theory regards as pricing discrepancies”. This, for MacKenzie, suggests that “financial markets are not an imperfectly insulated sphere of economic rationality, but a sphere in which the ‘economic’ and the ‘social’ interweave seamlessly” [MacKenzie 2003:373].
Periodising Financial Economics
In a recent paper,9 MacKenzie examines the question of performativity more closely. In particular, he traces the effects of arbitrage pricing theory in its most important field – derivatives markets – from the early 1970s to the present. He suggests that the 1973 paper by Fisher Black and Merton Scholes on options pricing, along with Robert C Merton’s paper in the same year, was “a defining – perhaps, the defining – achievement of modern financial economics...” [MacKenzie 2005:8]. Previous models of options contracts, MacKenzie notes, “involved parameters the values of which were extremely hard to determine empirically, notably investors’ expectations of returns on the stock [or other asset] in question and the degree of investors’ risk-aversion...” [MacKenzie 2005:6].
MacKenzie, following Michel Callon, suggests that performing the discipline of economics itself can affect markets: however, “[f]or a claim of performativity to be interesting – for the use of economics to constitute what I call ‘effective’ performativity
– an aspect of economics must be used in a way that has effects on the economic processes in question...economic processes incorporating the aspect of economics must differ from their analogues in which economics is not incorporated” [MacKenzie 2005:11]. MacKenzie suggests that the Black-Scholes-Merton pricing theory “disentangled options from the moral framework in which they were dangerously close to gambling, and framed them by showing how they could be priced and hedged as part of the normal operations of mature, efficient capital markets” [MacKenzie 2005:18]. This was presumably an enabling condition of the rapid spreading of options markets: geographically, across asset classes, and, most importantly, in the terms of the volume of trading in options markets. MacKenzie notes that “[b]road features of the Black-Scholes-Merton model were indeed already present in the patterns of prices in markets prior to the formulation of the model”, however, “there were also significant discrepancies between the model and the pre-existing price patterns” [MacKenzie 2005:19]. But prices did later conform very closely to that predicted by the model. This MacKenzie terms “Barnesian” performativity, which is the “claim that the market prices informed by the model altered economic processes towards conformity with the model” [MacKenzie 2005:23].
MacKenzie suggests that the success of the model was due to “[f]inancial economists quickly [coming to] see the Black-Scholes-Merton model as superior to its predecessors...it involved no nonobservable parameters except for volatility, and it had a clear theoretical basis, one closely linked to the field’s dominant viewpoint: efficient market theory” [MacKenzie 2005:27]. Most importantly, MacKenzie argues, for options traders the “underlying mathematics might be complicated, but the model could be talked about and thought about relatively straightforwardly: its one free parameter – volatility – was easily grasped, discussed, and reasoned about” [MacKenzie 2005:27]. Further, given that such traders tended to sell options to institutional buyers of options, the “availability of the Black-Scholes formula, and its associated hedging and risk-measurement techniques, gave participants the confidence to write [sell] options at lower prices, helping options exchanges to grow and to prosper” [MacKenzie 2005:36].
MacKenzie argues, however, that the US stock market crash in October 1987 forms an important break in the performativity of the Black-Scholes-Merton model: “The fall was a grotesquely unlikely event on the assumption of log-normality”. This had an important consequence, for the “subsequent systematic departure from Black-Scholes option pricing – the so-called “volatility smile” or “volatility skew” – is more than a mathematical adjustment to empirical departures from log-normality: it is too large fully to be accounted for in that way” [MacKenzie 2005:38]. Thus, for MacKenzie, the history of options pricing has three phases. The first is of little options trading activity and a loose fit between the Black-Scholes-Merton model and actual traded prices; the second demonstrates a Barnesian performativity fit of model to prices; and in the third “from 1987 to the time of writing – option pricing theory is still performed in the generic and effective senses (it is used, and its use makes a difference), but it has lost its Barnesian powers” [MacKenzie 2005:39].
Regarding the 1987 crash, MacKenzie raises the “intriguing possibility that amongst the factors exacerbating the 1987 crash was an application of options pricing theory” in the form of portfolio insurance. This, MacKenzie suggests, “is Barnesian performativity’s opposite: the use of an aspect of economics altering economic processes so that they conform less well to their depiction by economics” [MacKenzie 2005:39]. This possibility and subsequent changes in options market prices reinforces MacKenzie’s claim of the historical contingency of the 1973-87 period.
Emanuel Derman, who contributed to the well known BlackDerman-Toy options pricing model, has argued that, following the 1987 crash, “[o]ver the next 15 years the volatility smile [has] spread to most other options markets, but in each market it took its own idiosyncratic form”. Derman suggests that the rapid growth of options markets themselves – some options markets are now larger than the markets on which they are based – implies that the problems Black and Scholes faced are no longer pertinent. Whereas Black and Scholes attempted to derive the price of an options contract from the price movement of the underlying stock, a “better model of the smile should be capable of calibration to liquid stock, bond and options prices”. But, Derman concludes, the problem is in fact much deeper: “All financial models are wrong, or at least hold only for a little while until people change their behaviour” [Derman 2003]. MacKenzie’s emphasis on the importance of the 1987 stock market crash has, then, been confirmed by at least some leading financial theorist practitioners. However, MacKenzie may have underestimated the importance of LTCM’s near-bankruptcy. It is possible that LTCM was such an important exemplar of large-scale arbitrage trading across the world that its downfall itself may form an important break in any adequate periodisation of financial economics.
Trend Following and Skewness of Arbitrage
Nassim Nicholas Taleb has argued that market participants such as LTCM heavily prefer owning assets that give asymmetric pay-offs, where, for example, a “considerably negatively skewed bet can present more than 99 per cent probability of making G and less [than] 1 per cent probability of losing L”. Taleb finds “strong evidence for the predilection for negative skewness on the part of investors” [Taleb 2004:2] and argues that a number of financial instruments seem designed to have negatively skewed pay-offs: loans and credit-related instruments give a high probability of small payments with a small probability of default on the entire amount; derivatives such as simple call and put options have negatively skewed pay-offs if the strike price is, respectively, much above or below the current underlying asset price – a “seller of an out-of-the-money option can make a profit as frequently as he wishes, possibly 99 per cent of the time, by, say, selling on a monthly basis options estimated by the market to expire worthless 99 per cent of the time”; and various classes of arbitrage trades, such as merger arbitrage (where one bets that the merger will in fact take place, thus forcing a convergence of the two stocks in question) [Taleb 2004:2-3]. LTCM effectively engaged in every one of these types of trades.
Taleb suggests that negatively skewed pay-offs have three important properties. First, that the true mean and variance of the pay-offs are camouflaged, thus a “typical return will, say, be higher than the expected return”, so it will be “easier for the observer of the process to be fooled by the true mean particularly if he observes the returns without a clear idea about the nature of the underlying (probability) generator”. The same holds for estimations of the variance of returns. Second, it will take “a considerably longer sample to (properly) observe the properties under a skewed process than otherwise”. And, third, “the more skewness, the more the process will generate steady returns with smooth ride attributes”, thus “an investor has, without a decrease in risk, a more comfortable ride most of the time, with an occasional crash” [Taleb 2004:3].
Although Taleb goes on to locate the preference for negatively skewed payoffs in prospect theory and hedonic psychology, i e, in recent research areas of behavioural finance, the important point here is his suggestion that arbitrage trading as such suffers from negatively skewed payoffs.10 But the large number of hedge funds that do not specialise in arbitrage generally engage in a wide variety of different trading strategies. William Fung and David A Hsieh have examined mutual fund and hedge fund returns. They performed statistical regressions of such returns in an attempt to ascertain the “returns from assets in the managers” portfolios, their trading strategies, and their use of leverage” [Fung and Hsieh 1997:276]. Unsurprisingly, they suggest that “where they invest, much less how they invest, is the key determinate of performance in mutual funds” [Fung and Hsieh 1997:279, emphasis in original] whereas “[h]edge fund returns have low and sometimes negative correlation with asset class returns” [Fung and Hsieh 1997:282]. They suggest that hedge funds seem to employ come combination of five different trading styles: “ ‘Systems/Opportunistic’, ‘Global/Macro’, ‘Value’, ‘Systems/Trend Following’, and ‘Distressed’ ”11 [Fung and Hsieh 1997:285].12 They define “Systems/Trend Following” hedge fund traders as “traders who use technical trading rules and are mostly trend followers...” [Fung and Hsieh 1997:285] and find that this style of trading “is most profitable during rallies in non-US equities and bonds, and during declines in the US dollar”. Furthermore, the style “has a return profile similar to a straddle (i e, long put and a call [option]) on US equities” [Fung and Hsieh 1997:288].13
Counter Arbitrage
It would appear that trend-following strategies,14 being equivalent to holding options positions, gives positively skewed returns: consistent small losses, with an occasional period of high returns. As such, this trading strategy should often lead to positions that run counter to those put on by arbitrage traders. Consider the following simple example. Suppose an arbitrage trader, using the no arbitrage condition, derives a model relating two asset prices. For example, the model may assert that some asset A should trade at a price of $10 higher than some given asset B.15 The arbitrage trader may observe that the price difference is usually about $10 but that it has recently risen steadily from $10 to $13. But, this does not amount to sufficient information with which to put on the appropriate arbitrage trade (here, shorting A and buying B), because the model does not give any information on how high the difference may go in the future. The model only posits that it is rational to expect the difference to lessen over time. Theorists of no arbitrage models suggest noise traders as the “reason” for the increase in price difference and MacKenzie, for example, gives sociological reasons instead (like the imitation problem that nearly bankrupted LTCM). The usual – LTCM style – solution is to examine the historical price changes of asset A and asset
B. If it has been observed historically that the maximum observed price difference has been, say, $16, then the arbitrage trader may “safely” put in place a much larger position than if the observed maximum difference had been $20 dollars. From the systemstrend following trader’s perspective, everything is reversed. There exists no no-arbitrage derived model, merely observations of varying price differences. Such a trader, having observed the steady rise from the usual $10 to a recent difference of $13, may decide to buy A and sell B, knowing in advance that it would be easy to get out of the trade if the difference moved back to $10 because that would tend to happen under normal market trading conditions. If, on the other hand, the difference continued to increase, this would give such a trader the opportunity to close the trend-following trade out at a profit. Clearly, the trend-following trader’s returns will be positively skewed: most of the time, their trades will be unprofitable; occasionally, they will be closed at a (hopefully larger) profit. The opposite will be the case for the arbitrage trader. Note, however, that the arbitrage trader will close out losing trades at the worst possible moment – when the price difference has increased to some extreme, entirely abnormal, level. Worse, the arbitrage trader will have no (perceived) reason to close out the trade other than pressing capital inadequacies. Conversely, this will be the moment when trend following traders are looking to close out their profitable trades. In other words, arbitrage traders will act as to exacerbate the moment of historical crisis, and trend followers will be acting as to ameliorate it.
Paradox of Efficiency?
Three important events mark the history of neoclassical finance: the development of no arbitrage condition models in the late 1960s and early 1970s; the breakdown of the one-factor Black-Scholes-Merton options pricing model in the aftermath of the 1987 stock market crash; and the near-bankruptcy of the arbitrage hedge fund, LTCM. The initial event marked the proliferation of markets, especially derivatives markets, in the US and then around the world. The second event, itself produced by an over reliance on the Black-Scholes-Merton model, marked the breakdown of consensus on how to derive models from the no arbitrage condition. This breakdown, paradoxically, made arbitrage trading seemingly more attractive for there were now many apparent infractions of the no arbitrage condition in markets everywhere. And the third event has marked the beginning of a current period in which there is no consensus amongst economists and finance theorists and practitioners on how best to analyse financial markets and think about the efficient markets hypothesis.
Proponents of financial liberalisation rest their arguments on an usually implicit (because taken as obvious) efficient market hypothesis basis. But, given the ambivalence of finance theorists and practitioners over the efficacy of the no arbitrage condition, such a position appears untenable. Opponents of financial liberalisation, conversely, argue that liberalisation of financial markets results in an added risk of crisis, either in the form of asset price crashes or in the form of asset price speculative bubbles. However, after the 1987 crash and LTCM’s near-bankruptcy that risk may even have lessened. No arbitrage condition derived models are no longer employed as straightforwardly as before in financial markets. It is thus possible that markets are less prone to moments of crisis because they are less efficient now. The only proper conclusion seems to be that proposing or opposing financial liberalisation without accounts, critical or otherwise, of the work being done by finance theoristpractitioners is unlikely to be convincing.
EPWEmail: sva2003@columbia.edu
Notes
1 Following the sociologist of science, Donald MacKenzie, I consistently use the term sociology in this paper. One might easily replace it with history, anthropology, political science, etc, insofar as the discipline of economics perforce takes its domain – the economy – to be outside the scope of all the other social sciences.
2 To substantiate this argument they rather unfortunately provide a graph of the BSE Sensex from 1975 to 2005. This does indeed seem to show a lack of movement in the Sensex until 1985 or so, but this is because the graph is not plotted log-normally: one would expect stock prices to grow geometrically, not arithmetically. Changes in the gradient of a lognormal graph would indicate substantial changes in the nature of stock asset returns. Such appears to be the case from 2003 onwards.
3 The authors attempt to estimate the loss to the government from that waiver with data on stock prices after 2003. As they point out, it is surely the case that the gains to investors after 2003 are due in part precisely to the waiving of capital gains tax. This means that estimates based on yearon-year prices changes after 2003 will neither provide a useful measure of losses to the government, nor gains to investors.
4 Infosys ADRs are claims on Infosys shares held by US based financial
institutions; the claims are traded on the Nasdaq stock exchange. 5 The LTCM case is discussed in detail later. 6 An equilibrium level of presumably slight disequilibrium. 7 But see Lowenstein 2000 for a more critical assessment, one that emphasises
the changing nature of LTCM’s trading as safer arbitrage opportunities dried up in 1996, itself partly as a result of LTCM’s own initial successes.
8 The Federal Reserve engineered a buy-out of LTCM by its major counterparties, the international investment banks operating on Wall Street. 9 The paper was delivered to the 2005 meeting of the History of Economics
Society, Tacoma, WA. 10 He notes regarding LTCM that the “fund derived steady returns over a dozen quarters then lost all of them in addition to almost all its capital
in a single observation...only for the main principals to restart a new, albeit milder, version of the strategy” [Taleb 2004:2].
11 The nomenclature comes from an analysis of investment community marketing literature.
12 Fung and Hsieh do not consider arbitrage trading styles because the style
– as used by specific hedge funds set up for that purpose – was rather new. There was insufficient returns data to work with. 13 They develop this last point in greater depth in a later article [Fung and Hsieh 2001]. 14 In the news media, this style is often termed “momentum trading” and is spoken of as akin to noise trading.
15 For concreteness, suppose that asset B is some developing country asset such as a high-yielding government bond, and that asset A is a US government bond. Many arbitrage trades are of this form.
References
Chandrasekhar, C and P Pal (2006): ‘Financial Liberalisation in India: An Assessment of Its Nature and Outcomes’, Economic and Political Weekly, XLI(11), pp 975-88.
De Long, J B, A Shleifer, L H Summers and R J Waldmann (1990): ‘Noise Trader Risk in Financial Markets’, The Journal of Political Economy, 98(4), pp 703-38
Derman, E (2003): Laughter in the Dark: The Problem of the Volatility Smile, unnumbered pages, Euronext Options Conference, Amsterdam.
Dutt, A K (2006): ‘Flawed Logic of Capital Account Liberalisation’, Economic and Political Weekly, XLI(19), pp 1850-53.
Fung, W and D A Hsieh (1997): ‘Empirical Characteristics of Dynamic Trading Strategies: The Case of Hedge Funds’, Review of Financial Studies, 10(2), pp 275-302.
– (2001): ‘The Risk in Hedge Fund Strategies: Theory and Evidence from Trend Followers’, The Review of Financial Studies, 14(2), pp 313-41. Grossman, S J and J E Stiglitz (1980): ‘On the Impossibility of Informationally Efficient Markets’, The American Economic Review, 70(3), pp 393-408.
Krishnaswamy, K, A Mitra, S Shukla, S Shetty, A Bagchi and N Chandra (2006): ‘A Dangerous Measure’,Economic and Political Weekly, XLI(13), p 1222.
Lamont, O A and R H Thaler (2003): ‘Anomalies: The Law of One Price in Financial Markets’, Journal of Economic Perspectives, 17(4), pp 191-202. Lowenstein, R (2000): When Genius Failed: The Rise and Fall of Long-Term Capital Management, Random House, New York. MacKenzie, D (2003): ‘Long-term Capital Management and the Sociology of Arbitrage’, Economy and Society, 32(3), pp 349-80.
Williamson, J (2006): ‘Why Capital Account Convertibility in India Is Premature’, Economic and Political Weekly, XLI(19), pp 1848-50.
Economic and Political Weekly
available at
Sundar News Agents
(Magazine Market) 158 D N Road, Shop No 6 Near Kelkar Hotel
Fort, Mumbai 400 001
