Financial Crises of Emerging Markets

Emerging markets have witnessed major financial crises in recent decades, including the Mexican crisis of 1994-5, and the East Asian crisis of 1997 which spread to Russia and Latin America. Explanations about their causes abound, and may be aggregated into two meta-interpretations that reflect basic beliefs about markets and rationality. The first emphasizes self-adjusting markets and rationality, where market failure arises from government-led distortions and crises are caused by rational calculations in situations of government intervention and institutional weakness. The second emphasizes non-rationality or other types of rationality than neoclassical theory assumes, where market failure of well-functioning markets are routine, intervention is needed in order to modify outcomes, and crises result from non-rational calculations in situations of inadequately regulated financial markets.

Indeed, the theoretical debate is better characterized as paradigms talking past each other. Empirical testing is desperately needed, but the problem is one of generalizability – depending on the crisis or country, multiple explanations may contain elements of truth. Economic orthodoxy has become the ideology of the market in addition to being a theory and approach, with many believing that the big debates in economics were settled long ago. As a result, specific research programs in mainstream economics, and indeed in “American” IPE which is also resolutely methodologically individualist, have focused on an increasingly narrower set of questions. This creates blind spots hiding structural developments in the political economy, and preclude serious debates about finance as a system or as a term capturing fundamental trends. Indeed, as “the infrastructure of the infrastructure” of the global economy, finance is not an actor whose behavior can be modelled. This has significant implications for policy prescriptions.

This paper does not aim at hypothesizing and formally testing, but rather at off-setting the consensus on a “normal science” based on rationalist assumptions in accordance with neoclassical theory by giving prominence to non-rational factors contributing to emerging market crises. To that end, the first section elaborates on the complexity of financial crises and the two overarching views of how they occur. Since the latter reflect differences in basic beliefs of rationality and markets, an explanation of the premise of how financial markets is provided, drawing from sociological insights. Particularly, the critical distinction between risk and uncertainty is clarified to highlight the fundamental role of social conventions in financial market processes, as policy-makers and market actors make decisions in uncertain environments. The second section provides an overview of the different generations of rationalist explanations that dominate the debate to gain insights into the causal factors that are apparently deemed most important in generating financial crises. Some scholars have provided non-rational accounts of emerging market crises acknowledging the inherent instability of markets, and in order to illustrate how these more plausible explanations look the two next sections use a case study of the Asian financial crisis of 1997-99. The third section introduces the political economic context of the crisis, followed by an endogenous account of the crisis in the fourth section. I will draw from arguments of scholars who have provided endogenous accounts of the crisis, and add depth to our understanding of the crisis by analyzing the importance of social conventions. The final section concludes.

2 Financial Crises: Nature, Views, and Underlying Assumptions

Financial crises are typically multidimensional events which are difficult to capture with a single indicator. Different types of crises may overlap and as a result do not always take places independently. One type of crisis may lead to another, or two types of crises may occur simultaneously. Classifying crises as one type may then be misleading if one event derives from another. For example, Kaminsky and Reinhart (1999) finds that currency crises often overlap with banking crises. Emerging market crises have typically been combinations of currency and banking crises that are associated with sudden stops in capital flows, often transforming into a sovereign debt crisis. Indeed, of the 431 crises studies by Laeven and Valencia (2013), 68 are twin crises and 8 classified as triple crises. 20 out of 67 sovereign crises are also banking crises, and 42 of them also currency crises.Despite this complexity, they have common elements. They are often associated with significant changes in credit volume and prices of assets; disruption in supply of external financing and in financial intermediation; balance sheet mismatches that are large in scale (of households, firms, sovereigns, and financial intermediaries); and excessive government support in terms of liquidity and recapitalization (Claessens and Kose 2013). However, despite clarifications and identification of fundamental factors driving crises – macroeconomic imbalances, external and internal factors – identifying deeper causes have met little success. This is related to the way crises are understood, to which we will now turn.

2.1 Exogenous vs Endogenous explanations of Crises

Broadly speaking there are two different ways of understanding financial crises (Hall 2009). The first stream of thought is the exogenous approach, which has dominated economic thought in recent decades. The modern founders of this tradition includes Hayek and Friedman who believe in the efficient market hypothesis – left to their own devices, markets efficiently allocate resources. Financial markets correctly price securities since they always fully reflect all available information (Fama 1970:383). Here, financial crises are deviations from the market’s normal state, and hence, the approach focuses on ‘external’ causes – government failures in particular. An exogenous account assume constant adjustment in the behavior of market participants as new information enters the market. Market prices always reflect what others are willing to pay, and therefore prices are never false or inflated – the idea of bubbles, in which assets likes stocks or houses deviates from the underlying fundamental values, is simply rejected.

By contrast, an endogenous account holds that crises arise primarily from within finance. For Marx and Polanyi, “laws of motion” internal to capitalist modes of production generate change and upheaval between supply and demand rather than equilibrium. For Keynes, passions in speculation or “animal spirits” generates risky gambits. The endogenous view generally disagrees with the idea that traders merely integrate information coming from outside the market; rather, they focus on the behavior of other traders to anticipate their activities and do better than average, at least. Finance is subject to the pathologies of social life and thus give rise to manias, panics, and crashes periodically and normally (Hall 2009).

2.3 Financial Markets and Insights from Sociology

It is clear that the key difference between exogenous and endogenous approaches to the study of financial crises relates to their respective understandings of the processes of market rationality. Exogenous explanations are grounded in rationalism, whereas endogenous ones draw heavily on sociological and constructivist insights. The latter is important in understanding why we pay dearly for viewing crises almost entirely through the lens of rationalism and for prescribing policy solutions accordingly (Hall 2009): first, there is a prevailing belief amongst market actors and central bankers overseeing the financial system that financial markets operate in accordance with the Efficient Market Hypothesis of neoclassical economics, when, in fact, prices do not reflect their fundamental value. Frequently, demand does not stimulate supply, but rather a lack of supply stimulates demand (Polanyi 1994, Cooper 2008). Hence, market participants buy financial assets with the collective expectation that they will increase in value in the future, which generates destructive asset bubbles (Minsky 1986). This observation resulted in Minsky’s Financial Instability Hypothesis, stating that internal market dynamics generate waves of credit expansion and asset inflation, followed by rapid contraction of credit and asset deflation. It points to a second misconception of market rationality causing crises, namely that markets’ self-discipline precludes excessive financial risk accumulation. However, central banks are keen to intervene to supply liquidity – consistent with Keynes’ demand management (Keynes 1964) – especially in the prospect of a recession, which generates liquidity expectations in times of downturn in credit cycles. As authorities loosen monetary policy, credit is further extended to borrowers endogenously (through the commercial banking system), which means that additional credit is injected at a time where “demand” for credit should be falling in an effort to reduce debt burdens of the real economy. Risky/bad lending practices may arise due to the resulting incentives over time, forcing out good lending practices (Minsky 1986).

This process results in mispricing of risk, a phenomenon that originates from intersubjective expectations amongst market actors – they vigilantly await signals for how to react to events – rather than from utility calculations as per “rational expectations”. This process generates stable market transactions until these intersubjective expectations are shaken or frozen by disappointment. In that case, market volatility ensues as actors search for a new “equilibrium” (Keynes 1964). Further, intersubjective expectations have both causal and constitutive effects (Wendt 1999:88). The causal effect on markets is behavior unrelated to rational analysis of economic “fundamentals” since actors rely on confidence in the intersubjective understandings about the present and the future. In terms of constitutive effects there is mounting evidence of financial markets being driven by the products of financial economics instead of the processes reflecting a pre-existing reality. As Donald MacKenzie (2006), an economic sociologist, argues, it is an engine actively transforming markets. This is the idea of performativity (or constitutive social processes in constructivist IR). Using economics does not simply have an effect on its processes, but alter those processes such that their “conformity to the aspect of economics in question changes” (MacKenzie 2006: 18-22).

These key sociological insights suggest that “manias or booms or asset bubbles (..) quickly become panics, crashes, or busts” (Hall 2009), as irrational expectations can become intersubjective shared amongst market participants, and when they change the result may be massive market reversals.

2.4 Risk and Uncertainty: the Role of Social Conventions

These insights are far from new, but rather forgotten in mainstream political economic thought. Indeed, it follows from Knight’s (1921) and Keynes’ (1948) conceptual distinction between risk from uncertainty, to which we will now turn.

In a world characterized solely by risk, sophisticated decision-makers have no trouble sorting events into different classes. The future is unknown, but known probability distributions with fixed means and variances means that unforeseen events are accounted for (Meltzer 1982). This makes plausible the assumption that agents follow rational decision rules, consistent with exogenous understandings of financial market crises. However, when parameters become too unstable for quantification of the prospects of events that could happen, such assumptions become untenable (Keynes, 1937). Drastic changes in the underlying economic structure shifts the mean of the distribution and thus affects realms of uncertainty (Meltzer 1982). In such an environment decision makers (actors) cannot determine the objective probability distribution. In accordance with an endogenous view, market actors and policy-makers must often rely on intersubjective expectations and social conventions to confidently make decisions. These are shared templates and understanding that guide economic interaction (Wagner 1994, Langley 2008) by enabling classification schemas about events, thus delineating the situations in which the convention applies. This, in turn, makes clear what decisions are reasonable despite the lack of precise decision rules (Herrigel 2010). At best then, conventions can stabilize uncertainty contingent on the individual or collective actors’ interpretative capabilities, but it cannot eliminate uncertainty.

Scholars disagree as to whether social conventions arise as self-conscious solutions to coordination problems made possible by their intersubjective character of common knowledge, which facilitates stable patterns of cooperation (Lewis 1969 and Schelling 1960) or at random, in which case common knowledge is taken for granted to the extent that social conventions arise as practices and patterns of imitation and conformability. As the sociological insights suggest, the latter view is more plausible – conventions are collectively shared social templates employed to deal with epistemic uncertainty rather than solutions to strategic coordination games (Nelson and Katzenstein 2014). Crises are usually not exogenous shocks followed by struggles over distribution amongst rational actors which eventually yields an equilibrium. Rather, they illustrate the quintessential feature of social conventions: actors adopt them in order to cope with uncertainty, which sows the seed of systemic crisis endogenously. Hence, the rationalist assumption that we live in a world of quantifiable risk is too simple. Social and cultural contexts within which rational actors face uncertainties inherent in financial markets needs paying attention to (Best and Paterson 2010).

2.4.1 Rationalist view

In the past few decades many economists has refuted the idea that decision making may not happen according to rational axioms in the special case of uncertainty (see, for instance, Hirshleifer and Riley, 1992). Their theories draw from Subjective Expected Utility Theory (SEUT) where decision makers behave according to a probability distribution. Yet under uncertainty they may not have objective probabilities. Choices that deviate from SEUT’s axioms have been considered distant from valid inference because “inconsistency is costly” and thus financial market agents invest in information to avoid systematic mistakes. Subjective probability estimates should thus converge toward objective probabilities, giving rise to the rational expectations idea where agents’ subjective probabilities and the probabilities arising from the economic model in which those agents function are equated (Hansen and Sargent 2010). Thus, if market participants share a model of the economy that is correct, and information is properly distributed, then expectations about possible state of the economy in the future will converge and promote a self-enforcing, stable equilibrium (Blyth 2003). As such, rational investors who have “common knowledge” about the true underlying economic structure will not drive prices too far from equilibrium.

Reducing the world to risk is not only a progress relevant to economic theorists. As previously mentioned many specialists within IR and IPE followed suit in embracing the idea that the distinction between risk and uncertainty was a “sterile one” (Hirshleifer and Riley 1992) – uncertainty is neglected or dismissed and defined simply as risk, an assumption that is an integral part of social science today. The problem is that the set of choices they face is more likely characterized in part or solely by uncertainty (Nelson and Katzenstein 2014). Anomalous behavior deviating from SEUT is widespread (Camerer et al 1992, Zeckhauser 2010), and this uncertainty disqualifies predicting future returns based on past events (Leamer 2010). Crises are amongst such events, and their causes remain either uncertain or are difficult to diagnose. Hence, we look to sociological insights next.

2.4.2 Sociological view

Constructivist and sociological approaches recognize that actors and those overseeing financial markets adopt social conventions in order to make decisions, thus allowing the market to operate according to “expectations which define efficiency” (Storper & Salais 1997: 16). It is important to note that social conventions have varying degrees of materiality, taking form both as public discourse – so called ‘new era stories’ (Akerlof and Shiller 2009) – and material forms such as new technologies. They make possible different degrees of efficiency in financial markets. Moreover, they may be influenced by the theories with which we seek to explain the social reality (MacKenzie, 2005)

The sociological approach counters the rationalists’ portrayal of markets as unaffected by social relations. Rather, it sees them as uncertain environments stabilized by social conventions. Actors may thus classify events, refine expectations about future events, and act accordingly. Coordinating behaviors with the goal of creating mutual expectations to reduce risk is possible, but, often, conventions are followed due to their “prescriptive element” that allows judgment of appropriateness of behavior, thus reduces epistemic uncertainty (Biggart & Beamish 2003:444). Conventions are thus internalized by market participants, informing their expectations. Hence it is not the underlying economic fundamentals, but rather social conventions, that shape the market. The views are usually developed in the context of a social environment where the understanding of finance is affected by “rumours, norms, and other features of social life” (Sinclair 2009).

Crises remind us that our world is characterized not only by risk but also uncertainty (Nelson and Katzenstein, 2014). Hence, the composition of journal articles in prominent IR journals – almost exclusively quantitative and largely grounded in the liberal paradigm (Wade 2009, Palan 2009) – is not a signal of the discipline’s health. Rather, the failure of neoclassical theory to capture how people act under such uncertainty demonstrates the costs of resolute adherence to the tenets of rational choice theory, which preempts sociological explanations (Akerlof and Shiller 2009). As we shall see in the following section, the dominant explanation for emerging market crises are no different.

3 Emerging Market Crisis Theory: Three Generations of Models

Emerging markets crises have been the subject of extensive theoretical and empirical economic literature, giving rise to three generations of explanations evolving as the nature of those crises has changed throughout the past decades. To this day they are employed to prescribe solutions to financial stability, which means it is crucial to understand which view they take on financial market functioning.

3.1 First Generation Models: Speculative Attacks

The first generation of theory was motivated by the collapse in the price of gold which was the most important nominal anchor before the 70s where exchange rates were floated, and is often used in explaining currency devaluations in developing countries (Claessens 1991). It is based upon seminal papers by Paul Krugman (1979) and studies by Flood and Garber (1984) to show that speculative attacks on fixed or pegged currencies can be explained by rational investors correctly foreseeing that excessive deficits have been mounting, financed by central bank money/credit. This happens when governments take an overly expansionary stance on monetary policy causing domestic absorption to exceed production, which then spills over into deteriorating balance-of-payments. The central bank will start losing their liquid assets or their foreign exchange reserves supporting the exchange rate level until a point where these fall below a certain threshold. Here, a speculative attack gets launched because investors do not expect the peg to hold (Krugman 1979). This quickly eliminates remaining foreign assets and the peg is abandoned, and will subsequently secularly depreciate to reflect the relatively looser domestic policy stance vis-à-vis other countries. Hence, the model seeks to show how inconsistencies between domestic economic conditions and the chosen exchange rate regime will lead to a collapse of the peg.

Krugman’s model is a balance-of-payments crisis theory that has produced four classes of insights (Eichengreen et al 1997). First, it helps identify relevant macroeconomic fundamentals that determine the exchange rate and balance of payments. Second, it demonstrates how crises may occur before reserves has fully depleted. Speculators wish to liquidate the holdings of domestic currency while the central bank can still absorb sales volumes. The attack occurs such that the sales exactly eliminates the central bank’s stock of foreign assets. Third, in order to maintain an exchange rate peg the central bank needs adequate foreign exchange reserves. An attack which depletes its reserves results in the peg being abandoned, and the exchange rate regime transitions to floating. Hence, reserves must be rebuilt before the exchange rate can be re-pegged. Further, in the standard model the sales volume causing crisis is small, corresponding to the decrease in money base that is needed to ratify the lower money demand associated with the higher interest rates prevailing after the transition to a depreciating exchange rate (Eichengreen et al 1997). Domestic actors need to hold a share of the base after the peg has collapsed in order to transact. However, in economies heavily reliant on the dollar that transaction demand may be quite small (Edwards 1989), in which case the share of the monetary base liquidated in a crisis can be very large. The problem exacerbates if the currency crisis provokes a banking crisis, in which case the central bank will be required to purchase other domestic liabilities too (Goldfajn & Valdés 1995). Fourth and finally, the model implies that authorities stand little chance in defending against an attack. Only if the authorities choose not use sterilized intervention the attack can be repelled, yet at the cost of allowing the interest rates to rise and the base to shrink. In a situation where a sizable share of the base is involved, interest rate rises may increase to the point where the currency crisis precipitates a banking crisis (Reinhart & Rogoff 2011). Resuming sterilization will be necessary but weakens the central bank’s ability to defend the peg.

3.2 Second Generation Models: Multiple Equilibria and Contagion

The non-uniqueness property of theoretical macroeconomic models based on rational expectations later became the foundation for literature on speculative bubbles (Obstfeld 1986, Flood & Garber 1984) as examples of multiple equilibria and self-fulfilling attacks in foreign exchange markets were documented. This “second generation” of models offers a different perspective on crisis causes. When market participants anticipate that a successful attack will change policy it becomes the expected future fundamentals post-attack that are incompatible with the peg, rather than the present or future expected fundamentals absent an attack. This gives rise to two equilibria. In the first, where neither an attack nor change in fundamentals take place, the peg is maintained indefinitely. In the second a speculative attack happens and the expected exchange-rate change precipitates a change in fundamentals ex-post (Flood & Garber 1984). An obvious limitation to the model is that policy (before and after crisis) is set arbitrarily – should an attack occur, an expansionary policy response is assumed. The literature on multiple equilibria and self-fulfilling crises is not generalizable theory but rather a collection special cases. Paradoxically, Krugman (1996) argues that only when fundamentals are “good” we might end up in a situation of multiple equilibria. When they are clearly inconsistent with the exchange rate peg the model will converge on the equilibrium where attack and devaluation occurs. Further, unaware of authorities’ preferences there might be market ‘testing’ where unsuccessful attacks are launched to gain more information of these. Second generation models imply that merely having good fundamentals cannot preclude currency crisis. To prevent an attack in such a situation what matters is the credibility of the central bank in not relaxing policy once the peg is abandoned.

The issue of contagion has been added to this generation of models. Gerlach and Smets (1995) considered two countries trading in merchandise and financial assets, and finds that a successful attack and the following real currency depreciation enhances the competitiveness of its exports and thus produces a trade deficit in the second country, a gradual decline in its central banks foreign assets, and a currency attack. Contagion may also occur in response to the impact of crisis and currency depreciations in the first country on the overall price level in the second. Real depreciation reduces import prices in the other country, which reduces the CPI and demand for money by domestic residents. Efforts to gain more foreign exchange depletes the foreign assets of the central bank. The effect is a shift in the second economy from one equilibrium (no-attack) to another where investors have an incentive to launch it. Goldfajn and Valdés (1995) emphasizes the role of illiquidity and introduces financial intermediaries in their model. They show how small anomalies may provoke large-scale currency runs because of intermediaries – they supply liquidity to foreigners that are not making long-term investments. Hence, their services transform maturity structures and augments capital inflows. When foreign investors withdraw deposits it can be costly for the intermediates to liquidate those assets and therefore they risk failure. In that case a bank run can induce self-fulfilling banking crisis. As foreign investors convert their deposits into foreign exchange the bank run can spill over on the currency. When liquidity difficulties exist due to a banking crisis in one country, international investors may respond by liquidation of positions in other domestic markets, which means that these crises can spread contagiously.

Efforts to systematically test the predictions of these theories have faced difficulties given their biased samples. Eichengreen, Rose and Wyplosz (1994, 1995, 1997) seeks to measure crises as a weighted average of (a) exchange rate changes, (b) changes in foreign reserves available to respond to speculative pressures, and (c) changes in differences in interest, as these can be increased to defend against attacks. They test two dozen OECD economies since 1959. Their findings of causes and results of changes in exchange rate are consistent with predictions of mainstream models in the sense that countries face external balance problems prior to devaluations due to their relatively loose monetary policies, as well as internal balance problems (high unemployment). Indeed, the former may be a response to the latter. This leads to a generalization that revaluations mirror devaluations. Eichengreen et al (1997) finds no other events in foreign exchange markets generalizable, such as transitions of exchange rate regimes.

Kaminsky and Reinhart (1996) analysed the connection between speculative currency attacks and banking crises in 20 countries in Asia, Europe, the Middle east and Latin American from 1970 to 1995, using the same index as Eichengreen et al (1994, 1995) excluding interest rate differentials due to lack of data. Their findings suggest that declining economic activity, deteriorating export sectors, declining stock markets, and high real interest rates are good indicator currency crises. Moreover, crises are characterized by accelerating growth in money and growing liabilities in the banking system, and banking crises are leading indicators in predicting currency crises but only in few cases the reverse is true.

Eichengreen et al (1997) asks whether currency crisis in one country is correlated with a currency crisis in another country at the same time after accounting for lagged and current domestic political economic influences. They find a strong positive correlation, but this could also be due to common shocks to economic fundamentals rather than definitive evidence of contagion. Thus, the study goes further to explore alternative transmission channels for contagion and their respective explanatory power and finds that currency crises spread contagiously primarily due to trade links rather than macroeconomic similarities. The debate over contagion has taken a focus on emerging markets in particularly in Asia and Latin America. The data used by Eichengreen et al (1997) pertains to developed economies and thus makes it difficult to systemically analyse cross-country effects in emerging markets. They call for future research in this regard.

3.3 Third Generation Models: Moral Hazard

Largely motived by the Asian financial crisis of the late 90s and the difficulties of prior models to account for some key factors associated with it (Krugman 1980), the third generation explores how financial crises can arise from rapid balance sheet deteriorations following asset price fluctuations, including exchange rates changes. These models (e.g. Chang and Velasco 1998) show how balance sheet mismatches in financial and corporate sectors can lead to currency crises since they generate large vulnerabilities in situations where macroeconomic imbalances appear to be small, for instance where there is surplus in fiscal positions and current account deficits can be managed. This may trigger currency crisis, which, if domestic banks have sufficiently high debt in foreign currency, leads to a banking-cum-currency crisis. This generation of models emphasizes also the role of banks and self-fulfilling crises.

It is argued that international and domestic institutional deficiencies leads to moral hazard and thus inefficient investments as well as excessive risk-taking. Of the varieties of moral hazard especially crony capitalism is popular, suggesting that connections and patronage determines the terms of and access to credit (Frankel 1998, Krugman 1998). The role of industrial policy is important here, as investment projects are provided guarantees that in turn encourages excessive risk-taking on behalf of their managers (Economist 1997). Implicit guarantees such as deposit insurance are also considered and claimed to be a major source (Krugman 1998, McKinnon and Pill 1998). The too-big-to-fail logic is also given attention as domestic firms similarly take on excessive risks (Pyo 1998). The general policy prescriptions to domestic and international institutions influenced by these arguments are to speed up the liberalisation process and enhance transparency. The former would abolish market-defying industrial policies and reduce cronyism, and the latter will further limit and expose cronyism. This should be done by breaking up large business groups by for instance restricting cross-subsidisation and loan guarantees; improve stock market resilience; and enforce minority shareholder rights (Krugman 1998).

3.1.4 Predictive Performance

It is immediately obvious that the three generations of crises take an exogenous view of financial crises in emerging markets. Given the weakness of resolute adherence to rational choice theory associated with such an approach it is not surprising that no empirical research can differentiate which of the three generations of theories explaining and predicting emerging market crises is the best characterization (Claessens and Kose 2013). There is clearly a need for sociological insights to improve our understanding of the causes. As in an ecological system, the health of disciplines in the business of predicting crises is critically dependent on a certain degree of diversity, and will degenerate below a certain threshold.

Fortunately, a few authors have provided endogenous accounts of emerging market crises, constituting what might be thought of as a fourth generation of explanations. Yet, as we shall see in the following section, these accounts have been incomplete as they ignore the role of social conventions. In order to emphasise this point in the clearest way possible, a widely discussed case study is used for illustration: the Asian financial crisis of 1997-99. Before juxtaposing an endogenous account with the three generations of EM crisis explanations, the following section will summarize the political economic context in which the crisis unfolded.

4. The Political Economy of the Asian Financial Crisis: A Brief History

The East Asian financial crisis of 97-99 equaled the Great Depression in terms of scale of output and consumption declines, and ensuing poverty and insecurity increases. It pushed the Asian economies back 10 years in terms of the world income hierarchy, whilst international lenders escaped with small or no losses (Wade 2000). It generated debates amongst scholars and authorities as to fundamental issues within economics, finance, and economic policy-making. How did the East Asian countries get into so much trouble and how did they return to growth?

4.1 Trajectory of Events

The crisis unfolded in multiple overlapping phases, starting in Thailand and subsequently spreading to other South East Asian countries. The dating of the crisis can conveniently be traced to the  2^nd of July 1997 when the Thai baht was floated. Soon after, all the South East Asian exchange rate regimes transitioned to floating currencies, which subsequently depreciated sharply. The ensuing economic meltdown meant that most of the countries (Indonesia, the Philippines, Thailand, Indonesia) had to seek assistance from IMF and other multilateral institutions and donors. Despite this, confidence failed to be restored quickly and loan packages required revisions in 1998 and 1999. Malaysia was also forced to give up its currency peg in July 1997 yet avoided having to seek help from the IMF. In September 1998, the Malaysian government fixed its exchange rate and imposed capital controls.

Taiwan’s decision to float in October 97 marked the onset of a second phase of the crisis. Speculation turned to the Hong Kong dollar which was still pegged to the USD. Large reserves and a currency board system allowed authorities to fend off the attack, but the necessary increase in the interest rate led to substantial and rapid stock market decline. Interest-rate sensitive property development companies were badly hurt, and overall, emerging markets witnessed rapidly widening spreads.

Marking the next phase was South Korea. Large business groups failed in the beginning of 1997, and in late October a liquidity crisis ensued as a result of the situation in Hong Kong. The won was floated, and in December, South Korea accepted a massive IMF program backed by many other sources. This proved inadequate, and new resources and conditions had to be introduced. This did not mark the end of the currency crises as the effects were felt in Russia, effectively defaulting in August 1998, and in Brazil. However, the focus of this account shall remain on the region which felt the crisis the most: South East Asia.

4.2 Economic factors

Regardless of the arguments inherent in exogenous and endogenous accounts of the crisis it is generally agreed that several factors created permissive conditions, or themselves constituted pressures. The Chinese monetary authorities devaluing the Yuan by 35% in 1994 and the country increasingly dominated export markets (Taylor 1998). Japans poor economic situation also had dramatic effects – unexpectedly, the dollar rose 50% against the Yen after Japanese devaluations in response to sluggish growth. This strengthened exchange rates of South East Asian currencies given their dollar peg, and had implications especially for South Korea which competed with Japan in several sectors (Haggard 2000). Terms of trade shocks also resulted: semiconductor prices collapsed and impacted South Korea and Malaysia immensely.

These factors alone could not have generated crises of the scale witnessed. The financial liberalization and integration process was a necessary condition (Wolf 1998, Palma 1998, 2000). In the early 90s Asia (excluding Japan) received sharply increasing capital flows both from foreign banks and hedge funds and speculators after recently having either opened fully their capital accounts or taken measures towards that end (Kregel 1998, Wade 2000). This created problems for South Korea, for instance, where external debt maturity profiles were a significant source of vulnerability given their increasingly short-term nature. Several countries failed to understand policy constraints of capital account openness; recognizing overheating of the economies, governments used monetary policy instruments that attracted more inflows, thus leading to further real appreciation. The fixed rate regimes generated risk-taking because it was perceived as a guarantee by investors, which justified unhedged borrowing.

Finally, there was deep distress in the corporate and financial sector which constituted important domestic vulnerabilities. The financial sectors saw rapid lending growth, high corporate leveraging, and excessive risk-taking. Malaysia, Thailand and South Korea had undergone bank-financed booms in investment prior to the crisis. As a result, lending increased rapidly (in spite of lowering returns on capital) and banks’ balance sheets worsened. When the bubble burst, liquidity and solvency problems of financial institutions became apparent since collateral values fell alongside asset prices. Banks had to decrease or stop lending altogether, furthering the process of asset deflation, which lends empirical support to arguments that foreign exchange crises may lead to financial crises and vice versa (Kaminsky and Reinhart 1998), and that financial vulnerability increases vulnerability to foreign exchange crises (Goldstein et al 2000). The weak domestic financial situation was further exacerbated by the market signals sent by fiscal and monetary policy tightening as required by the IMF programs. However, there is inconclusive evidence on the effects on the exchange rate. In any case, since investors respond to the actions of governments there is a need to account for how government (in)action affected markets, which highlights the central role of politics.

3. Political factors

Close relationships and interactions between business and government is a feature of governance in the region that in the past has contributed to its high investment levels and growth rates. This comes with risk, especially after countries have liberalized their capital accounts in a world of global capital mobility (Krugman 1998). Private public sector relations influenced government intervention, distorted incentives and politicized lending behavior and the losses that followed. Yet, more important were the risk arising from poorly regulated liberalisation and privatisation (Haggard 2000). Rather than being an antidote to corrupt practices these processes were often “captured” by the private sector, which shifted risks back to the government and increase their fragility by weakening regulatory processes. Underlying these risks in the region were political and institutional features – popularly viewed as cronyism in the West – characterized by political bias towards certain constituents in the private sector and an absence of monitoring mechanisms to detect these biases or business influence. In general, there was a lack of transparency.

Regardless of long-run vulnerability sources, political uncertainty at the onset of the crisis and in the following adjustment process impacted government responses. For instance, the public-private sector relationship hindered coherent and timely action to the arising difficulties (Pepinsky 2008). Broader political uncertainties associated with type of political regime were also relevant (McIntyre 2001). Authoritarian rule has the advantage of decisive action which allows for swift, correct responses that in turn portrays a stable policy stance to investors, but often risks arbitrary action and lack of transparency in business-government relation (Haggard 2000). In particular, the crises exposed certain weaknesses in Indonesia and Malaysia associated with their authoritarian regimes – these were more acute for Indonesia as it lacked any form of checks and balances in the political system. Confidence in the Indonesian government feigned as challenges mounted. Indeed, countries going through the deepest crises also experienced the most radical political changes.

By contrast, In South Korea and Thailand the nature of government policy processes and the checks and balances provided by veto players delayed initial efforts to respond to the crisis, but he Phillippines was able to avoid both the rigidity of democratic regimes and the volatility of authoritarian regimes (McIntyre 2001). Once the crisis was unfolding, the nature of political opposition was similar across all the countries – governments faced pressures from social groups and business who focused on specific reforms not necessarily consistent with those prescribed by the IMF. Hence, policy was not solely a response to external pressures from international institutions and the US (Wade & Veneroso 1998).

Finally, the Asian financial crisis is characterized by systemic distress, which means many banks and firms face liquidity and/or solvency issues simultaneously. This makes it difficult to distinguish which bank or firm is viable, and for that reason governments engaged in forbearance resulting in substantial public losses. Crisis precipitates reforms with long-term consequences: changes in laws and regulation, and liberalization measures. The adoption of stronger financial regulation and corporate governance rules, better bankruptcy procedures, and FDI liberalisation gradually change the nature of private-public sector relations, however governments still faced obstacles to financial and corporate restructuring, as these run counter to the “Asian model”.

The political economic context in which the crisis unfolded provides a basis for analysing the market dynamics leading to the crisis. Rather than blindly accepting accounts based on false assumptions of rational calculation simply because they are in accordance with neoclassical theory, the following section will give an endogenous account with prominence to non-rational factors. Key insights from scholars who have taken this view of the Asian crisis are outlined, and depth will be added by analysis of social conventions, which no existing account has given attention to.

5. Endogenous account of the AFC

5.1 Mania, panic and crash

Endogenous accounts of emerging market crises (Kregel 1998, Palma 1998, Wade 2000) generally draw from the work of Charles Kindleberger’s work Manias, Panics, and Crashes (1978), which argues that excess borrowing and lending are endogenous to free and competitive, yet under-regulated financial markets. This precipitates cycles of mania, panic, and crash. In the context of the Asian financial crisis, Palma (1998) identifies four elements contributing to the emergence of Kindleberger’s mania phase. The first two are the propensity to over-borrow and over-lend, respectively. These are exacerbated by two other elements, namely distorted incentives at the domestic level and inadequate regulation. Panic is caused when lending stops abruptly, and crisis ensues. Despite adherence to some of the key sociological insights described in the first section, one crucial element in the operation of financial market processes is missing from the analysis and thus makes these endogenous accounts incomplete: the role of social conventions in financial markets. This section will present the endogenous accounts given, but add explanatory value through an evaluation of the role of social conventions in Kindleberger’s mania, panic and crash cycle.

5.1.1 Over-lending and over-borrowing

Emerging markets’ demand for funds are insatiable, and they can always borrow (at least short term) (Palma 1998). This means that resources are allocated inefficiently and, as a consequence, excessive risk is accumulated. This is neither privately nor socially efficient, and necessitates an effective regulatory system. This argument is consistent with the observation of rapidly increasing lending to East Asia by bankers and financiers in a time where spreads were declining (Kregel 1998), the macroeconomic fundamentals were deteriorating (Krugman 1998), unhedged private debt short-term in nature was building up (Chang et al 2000), gearing ratios were going up, profit margins in the real economy was decreasing, “cronyism” was becoming more prominent, and the investment levels were reaching incredible heights (Haggard 2000, Palma 1998). A key issue here is that when the amount of investment that needs ‘clearing’ grows larger, financial market players first exaggerates positive news and subsequently misjudge and exacerbate bad news. As documented in the Asian crisis, inflows – increasingly of short term maturity – into Korea, Malaysia, Thailand, Indonesia and the Philippines reached $220bn from 1994-96, whilst these countries’ foreign debt doubled in the same time period (Kregel 1998, Palma 1998). In the crisis, the economic situations of the countries that borrowed were “misjudged”. This biasing heuristic counters the idea of rational calculation assumed in the three generations of explanations which suggest that capital inflows were solely a function of capital account opening, fixed exchange rates, lax bank supervision that is inadequate for an internationalised financial system, depreciations of domestic currencies against the Yen due to their dollar-peg, and higher returns in Asia than in the US and in Europe.

This propensity to over-lend made “animal spirits” run amok, precipitating Kindleberger’s (1978) ‘mania’ phase. Indeed, as markets of ‘last resort’, countries in East Asia were particularly prone to this effect. Consistent with Minsky’s inefficient market hypothesis, willingness to lend created demand and consequently borrowing through overly optimistic expectations about the economies’ prospects. This was because the increased access to lending prior to the crisis generated initial improvement in economic performance due to the extra expenditure from more borrowing and foreign exchange availability. This became a self-fulfilling prophecy that was further bolstered by the deregulation and economic liberalisation as advanced by Washington consensus institutions and populist politicians (Wade 2000, Palma 1998). Further, exhilarating expectations meant that the need for hedging was overlooked, despite deteriorating current accounts in the region.

A critical shortcoming of this analysis of inflows is the lack of focus given to the social convention arising form the idea of “miracle Asia”. This had causal effects on the “misjudgement” of international lenders and to the acceptance of those funds by the borrowers. The remarkable developments by historical standards in the “newly industrialised countries” of South East Asia – particularly their success in adjustments and growth levels during turbulent time period from 1970-1990 – had drawn much attention. Emphasis was placed on accommodative market-friendly policies that got the “fundamentals” right in response to significant changes in the international economic environment. The result was generation of new resources and their efficient allocation. Thus, investors relied on imaginations of future events and a social convention in the form of a ‘new era story’ (Akerlof and Shiller 2005) in order to reduce complexity in the decision process. Despite being an efficient cognitive instrument, this distorted the perception of reality, turning this perception into self-fulfilling prophecies due to dynamics of performativity (MacKenzie 2006) and reflexivity (Soros 1987).

5.1.3 Distorted incentives and regulatory inadequacy

Other central arguments in endogenous accounts relate to distorted incentive at the domestic level and inadequate regulation. As for the former, inflows undertaken in the belief that these will be immune from standard risks and protected by government guarantees reinforced the propensity to over-lend (Kregel 1998). Artificially high exchange and interest rates induced by domestic policies produced dramatic distortions as the resulting spreads attracted dramatic increases in arbitrage flows. This further overvalued exchange rates and precipitated interest rate rises, partly to counter the expansionary effect of increased reserves (Palma 1998). The implicit or explicit guarantees of bank bailouts provided by the government, and previous IMF-endorsed bail outs in Latin America reinforced the belief that East Asian banks would receive similar treatment (Krugman 1998). The resulting moral hazard was another inflow-attracting distortion.

The latter element takes starting point in the market failure in East Asian domestic banking systems. Pressure to take measures that reduce reserve requirements, facilitate foreign borrowing, and remove private debt financing requirements mounted. As a result, the region increasingly liberalised and deregulated their domestic financial systems, leading to inadequate financial regulation and monitoring, which in turn allowed banks to make more risky loans (Krugman 1998). Domestic bank lending in Thailand more than doubled from 1990-1996; the Philippines saw a 3.3-fold increase; and Indonesia a 9-fold increase since mid-1980s. As previously mentioned, the maturity structure became increasingly asymmetric too (Chang 1998). Excess liquidity and deregulation caused problems of transparency and information. For instance, growth in developments of financial instruments such as derivatives that allowed banks to shift commitments off balance-sheets (Neftci 1998; Kregel 1998) greatly reduced transparency. Further, lack of monitoring in the private sector created information issues by encouraged excessive risk-taking in debt financing and favouring managers rather than shareholders in equity financing (Stiglitz 1998). Hence, this represented dramatic distortions to capital markets. Another issue contributing to ‘mania’ was in the evaluation of available information, both by market players and domestic governments. The down-side risks were ignored in favour of up-side risks, and, in addition, lending was increased during the Tequila hangover, which caused an increase in bank lending to Asia of around US$70 bn.

Again, social conventions are precluded from the analysis. In terms of distorted incentives, it is true that lenders paid no attention to downside risks, but to a large extent it can be argued that this was critically carried along by the “gestalt” of miracle Asia. This also casts doubt on popular moral hazard arguments that implicit and explicit IMF guarantees caused these massive inflows. In fact, there is no evidence that lending occurred to those with stronger ex-ante bailout presumptions(Wade 2000). The same point may be applied to the argument that a lack of transparency about balance sheets, forex reserves, and foreign debts caused inflows to become that large. In fact, the dangers of this trend were made clear by rating agencies and by the Bank of International Settlements (BIS, 1996: 5). Investors ignored this due to focusing on the up-side and in the belief that the miracle of Asia would continue. This also explains the development of derivatives as a social convention taking material form, grounded in the commonly held belief that Asia would continue it’s growth path and accommodative monetary policies and liberalisation measures.

5.1.5 From ‘mania’ to ‘panic’

As Kindleberger (2000) points out, the only thing worse than lending excessive amounts to developing countries is stopping that lending abruptly. The collapse of banks in Thailand and corporations in Korea, and the subsequent speculative attack on the baht triggered the ‘panic’ stage (Palma 1998). Costly international bail-outs provided liquidity of foreign exchange in order to have capital account convertibility, which meant that foreign capital, and big international players, could leave unscathed after the crisis, and substantial losses were incurred by other participants. Hence, financial markets provided ‘only carrots and no sticks’ (Palma 1998: 798) regardless of how well risks were priced and the efficiency with which financial resources were allocated. This contributed to a too big to fail logic, a moral hazard that shortened the ‘collective memory’ of crises in financial markets (Kindleberger 1978). Financial markets accepted ex post intervention of bailouts, but were reluctant to ex ante interventions of effective regulations that could reduce vulnerabilities. In East Asia, the reversal of flows was rapid and large in volume: net external financing dropped more than US$70 bn. between 1996 and 1997, which is almost equivalent to their combined reserves, and constituted more than 10% of GDP. The majority of that share pertained to commercial bank lending (IIF 1998), which created a large supply shock. Consequently, exchange rate adjustment and internal policy revisions were insufficient to deal with the dramatic contractionary effects. What would have otherwise probably been a ‘local balance of payments crisis’ (Kregel 1998: 37) became a full-blown financial crisis.

A sociological optic on the sudden outflows suggests that a deeper causal factor was the abrupt shift in intersubjectively shared irrational expectations. The mania stage generated stable market transactions, but as soon as these expectations were shaken and changed, market volatility ensued as market participants searched for a new “equilibrium”. The result was massive reversals. Investors and speculators pulled out after starting to focus on micro indicators such as debt maturity structures, which they could have tracked in the mania stage. In addition, contrary to the IMFs belief that austerity and institutional reforms would restore confidence, the news of IMF conditionalities being negotiated with the East Asian countries contributed to the breakdown of the social convention, and thus aggravated the loss of confidence. The “equilibrium” settled on “crony” Asia rather than “miracle” Asia. Combined with signals that large institutional reforms shifting away from the Asian model and towards the tenets of the prescriptions of the Washington Consensus were required for growth, these news prompted an even bigger exit-rush. This helps explain the size of the outflows. As such, panic was not simply a trigger or messenger of the crisis; rather, it was a primary cause. Social conventions caused changes in behaviour much larger than warranted by ‘real’ factor changes. Hence, a sociological analysis suggests that the massive outflows in Thailand was necessary for the crisis to happen.