{
  "id": "arbitrage",
  "title": "Arbitrage",
  "lang": "en",
  "source_lang": "en",
  "summary": "Arbitrage exploits inconsistent prices for equivalent assets or cash flows and provides a foundational principle for financial valuation and market integration.",
  "category": "economics",
  "subcategory": "finance",
  "type": "concept",
  "aliases": [
    "Financial arbitrage"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/arbitrage/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/arbitrage/",
      "json": "https://aiwiki.page/zh/arbitrage.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Arbitrage is the exploitation of inconsistent prices for the same asset, equivalent bundles of assets, or identical future cash flows. In its strict financial meaning, an arbitrage opportunity permits a trading strategy that requires no net investment, cannot produce a loss, and has a positive probability of producing a gain. A closely related case involves receiving money immediately while assuming no future net liability. Buying an asset in one market and simultaneously selling the identical asset at a higher price elsewhere is the simplest example.\n\nArbitrage connects prices across financial markets, trading locations, currencies, and dates. It also supplies a central principle for valuing securities: instruments that generate the same payments under the same conditions should have the same price, after allowing for relevant costs and restrictions. Actual trading is less straightforward than this principle suggests. Execution delays, financing requirements, delivery conditions, and legal differences can turn an apparent price discrepancy into compensation for risk. Market practitioners also use “arbitrage” more broadly for strategies that seek profits from relative mispricing but can incur substantial losses."
    },
    {
      "heading": "Definition, payoff conditions, and the law of one price",
      "text": "The essential feature of strict arbitrage is not merely buying cheaply and selling dearly, but constructing a position whose overall cash flows exclude an economic loss. The comparison must include acquisition costs, sale proceeds, financing, contractual obligations, and the eventual closing or delivery of every position. A trade with a high expected profit is not necessarily an arbitrage; a trade with a small guaranteed profit may be one.\n\nIn a simple one-period model, let \\(V_0\\) denote a portfolio’s initial cost and \\(V_T\\) its terminal payoff. A standard definition requires\n\n\\[V_0=0,\\qquad V_T\\geq 0\\ \\text{almost surely},\\qquad \\Pr(V_T>0)>0.\\]\n\n“Almost surely” means that the condition holds except on events assigned zero probability by the model. The strategy must also be self-financing: after its establishment, changes in its holdings are funded from the portfolio itself rather than by uncounted external contributions. A negative initial cost with a nonnegative terminal payoff is an equivalent form of opportunity when the initial proceeds can be invested safely.\n\nThe law of one price follows directly. Suppose two tradable portfolios deliver exactly the same payments at the same dates in every relevant state of the world. If one costs less, an arbitrageur can buy it and sell the expensive portfolio, retaining the price difference without assuming a net future payment obligation. This argument assumes that the expensive portfolio can actually be sold, including through short selling when it is not already owned, and that all contractual commitments can be fulfilled.\n\nFor example, an identical security offered for 100 at one venue and bid at 101 at another appears to yield one unit of profit per security. The relevant prices, however, are the executable offer and bid, not the last reported transaction prices. Commissions, the bid–ask spread, transfer charges, borrowing fees, and settlement costs must all be counted. If these consume the difference, there is no arbitrage.\n\nArbitrage also differs from ordinary speculation. A speculative purchase profits if an uncertain future price rises; a strict arbitrage links offsetting transactions so that such a forecast is unnecessary. Market making is different again: a dealer earns spreads by standing ready to trade but usually bears inventory and adverse-selection risks. These activities can overlap operationally, yet their economic sources of profit are distinct."
    },
    {
      "heading": "Historical development and the emergence of modern theory",
      "text": "Arbitrage predates modern securities exchanges. Merchants and foreign-exchange dealers compared the prices of goods, bills of exchange, and metallic currencies across commercial centers. A bill payable in another city could be valued through several currency conversions or alternative payment routes. Discrepancies encouraged offsetting purchases and sales, although transport, communications, and the reliability of counterparties limited how closely quotations could be aligned.\n\nUnder the gold standard, exchange-rate deviations were constrained by the cost of moving gold. When purchasing foreign currency became sufficiently expensive relative to the official metallic relationship, it could become cheaper to acquire gold, ship it, insure it, and exchange it abroad. The resulting “gold points” were therefore bands rather than a single exact exchange rate. Their width depended on freight, insurance, interest during transit, and operational conditions. This is an early illustration of the difference between a theoretical parity and an economically enforceable price bound.\n\nThe development of telegraphic communication and organized securities markets accelerated the comparison of prices between locations. As information arrived faster, some opportunities that had depended on slow communication narrowed. Arbitrage increasingly involved not only identical instruments traded in different places but also different securities linked by contractual conversion, delivery, or redemption provisions.\n\nModern financial economics gave these practices a general theoretical role. Franco Modigliani and Merton Miller’s 1958 analysis of corporate capital structure used arbitrage reasoning to show that, under specified frictionless assumptions, changing the division between debt and equity does not by itself change a firm’s total value. The argument depended on investors being able to reproduce corporate leverage through their own borrowing and lending. Taxes, bankruptcy costs, and financing differences can alter that conclusion.\n\nIn 1973, Fischer Black and Myron Scholes, together with related work by Robert C. Merton, established a landmark approach to option valuation through replication and dynamic hedging. Stephen Ross’s 1976 arbitrage pricing theory used the absence of arbitrage to connect expected asset returns with common risk factors. Later mathematical work, including that of J. Michael Harrison and David Kreps in 1979, clarified the relationship between arbitrage-free prices and probability measures suitable for valuation. Black’s contribution was recognized in the scientific background to the 1997 Nobel Memorial Prize in Economic Sciences, awarded to Merton and Scholes; Black had died in 1995.\n\nThese developments transformed arbitrage from a particular trading technique into a method of economic reasoning: identify cash flows that can be reproduced, compare their costs, and determine which prices can coexist without permitting a free gain."
    },
    {
      "heading": "Spatial arbitrage and foreign-exchange relationships",
      "text": "Spatial arbitrage links markets located in different places or operated on different trading systems. For a standardized security, the basic operation is to purchase where the executable price is lower and sell where it is higher. For physical goods, the transaction also requires transport, storage, inspection, and delivery. The relevant price relationship consequently includes the cost of moving the good, not simply the difference between two displayed quotations.\n\nIn commodity markets, apparently similar products may not be interchangeable. Crude oils differ in sulfur content and density; grain contracts specify grade, moisture, and delivery location. A price difference between two such goods can reflect genuine differences in production use or delivery obligations rather than arbitrage. Even identical material can command different prices when shipping capacity is scarce. Spatial price gaps therefore provide information about both trading opportunities and physical bottlenecks.\n\nThe foreign-exchange market supplies a particularly clear example because currencies can be exchanged through multiple routes. Suppose one euro buys 1.20 US dollars, one British pound buys 1.50 US dollars, and one pound buys 1.30 euros. Starting with 1.50 dollars, a trader could buy one pound, exchange it for 1.30 euros, and convert those euros into 1.56 dollars. Before costs, the cycle produces 0.06 dollars.\n\nThis triangular arbitrage would place pressure on all three quotations. Buying pounds with dollars tends to make pounds more expensive in dollars; selling pounds for euros tends to reduce the pound’s euro price; selling euros for dollars tends to reduce the euro’s dollar price. Competitive trading brings the direct and indirect exchange routes into alignment, subject to trading costs and available quantities.\n\nActual calculations must use the appropriate side of each currency quotation. If \\(S_{A/B}\\) is measured in units of currency B per unit of currency A, frictionless consistency requires\n\n\\[S_{A/C}=S_{A/B}S_{B/C}.\\]\n\nWith separate bid and offer prices, this equality becomes a set of inequalities: neither direction around the conversion cycle should yield a positive net return. A discrepancy calculated using midpoint quotations may disappear entirely at executable prices.\n\nCurrency convertibility, transfer restrictions, and the legal location of deposits also matter. An offshore deposit and an onshore deposit denominated in the same currency need not be freely interchangeable. Capital controls can prevent the purchases, transfers, or borrowing needed to enforce parity. In such cases, persistent price differences may indicate segmented markets rather than an opportunity available to every trader."
    },
    {
      "heading": "Intertemporal arbitrage, financing, and storage",
      "text": "Intertemporal arbitrage connects a price today with a contract for delivery at a later date. The simplest example is cash-and-carry arbitrage: buy an asset now, finance the purchase, hold it, and sell it forward. If the forward selling price exceeds the total financed holding cost, the combined position generates a gain, assuming delivery and financing are secured.\n\nLet \\(S_0\\) be the current price of a non-income-producing asset, \\(r\\) a continuously compounded financing rate, and \\(T\\) the time to delivery. Under frictionless conditions, its forward price is\n\n\\[F_{0,T}=S_0e^{rT}.\\]\n\nIf the forward price is higher, the trader can borrow \\(S_0\\), purchase the asset, and contract to deliver it for \\(F_{0,T}\\). Delivery produces enough cash to repay \\(S_0e^{rT}\\), leaving the difference. If the forward is too cheap, a reverse cash-and-carry trade may be possible: short the asset, invest the proceeds, and acquire it later through the forward.\n\nIncome received while holding the asset changes the calculation. For a stock index with a known continuous dividend yield \\(q\\), the idealized relationship is \\(F_{0,T}=S_0e^{(r-q)T}\\). For assets with specified cash payments, the present value of those payments is deducted from the spot acquisition cost before financing it forward. Actual borrowing and lending rates may differ, and stock borrowing may be unavailable or subject to recall.\n\nPhysical commodities introduce storage expenses and the benefits of immediate possession. A commonly used cost-of-carry expression is\n\n\\[F_{0,T}=S_0e^{(r+u-y)T},\\]\n\nwhere \\(u\\) represents proportional storage costs and \\(y\\) the convenience yield. Convenience yield captures the operational benefit of available inventory, such as keeping a refinery running during a supply interruption. Unlike a contractual dividend, it is not necessarily a cash flow that every investor can collect. Commodity parity relations can therefore be less tightly enforceable than those for readily borrowed financial securities.\n\nThe corresponding currency relationship is covered interest parity. If \\(S\\) and \\(F\\) are spot and forward rates quoted as domestic currency per unit of foreign currency, and \\(r_d\\) and \\(r_f\\) are comparable domestic and foreign interest rates, then, using simple interest,\n\n\\[F=S\\frac{1+r_dT}{1+r_fT}.\\]\n\nBorrowing one currency, converting it, investing in the other, and fixing the return conversion through a forward should not outperform equivalent domestic financing without risk. Persistent deviations, often discussed as a cross-currency basis, can nevertheless arise when financing, balance-sheet capacity, and counterparty access differ across institutions."
    },
    {
      "heading": "Derivative replication and relative-value identities",
      "text": "A financial derivative derives its payments from an underlying asset, rate, index, or event. Arbitrage-free valuation compares those payments with a portfolio that reproduces them. If the derivative and its replicating portfolio have identical future cash flows, a price difference permits buying the cheaper position and selling the more expensive one. The strength of the conclusion depends on how precisely and practically replication can be performed.\n\nA central example is put–call parity. Consider European call and put options on a non-dividend-paying share, with the same strike \\(K\\) and maturity \\(T\\). A call gives the holder the right to buy the share for \\(K\\); a put gives the right to sell it for \\(K\\). Buying the call and holding cash sufficient to provide \\(K\\) at maturity produces the same terminal value as buying the share and the put:\n\n\\[C+Ke^{-rT}=P+S_0.\\]\n\nBoth portfolios are worth the larger of the terminal share price and \\(K\\).\n\nIf the two sides trade at materially different executable prices, an arbitrageur can buy one portfolio and sell the other. The relationship is unusually robust because it does not require a forecast of the share price or an estimate of its volatility. It does require matching contract terms. Dividends, early exercise rights, different settlement conventions, or restrictions on short sales alter the relevant comparison. American-style options, which may be exercised before maturity, generally require bounds or additional adjustments rather than unrestricted use of the European identity.\n\nReplication can also be dynamic. In a one-step binomial model, suppose a share worth \\(S_0\\) can end at either \\(S_u\\) or \\(S_d\\), while a derivative pays \\(V_u\\) or \\(V_d\\). Holding\n\n\\[\\Delta=\\frac{V_u-V_d}{S_u-S_d}\\]\n\nshares and an appropriate cash position reproduces the derivative in both states. Its current price is then determined by the cost of that portfolio. Repeated over many trading periods, this reasoning leads toward continuous-time option models.\n\nIn practice, dynamic hedging is an approximation. Prices may jump, trading is discrete, volatility changes, and rebalancing incurs costs. A model-consistent price difference is consequently not always a strict arbitrage. Traders must distinguish a contractual identity, such as a matching terminal payoff, from a hedge whose success depends on modeling assumptions.\n\nRelated mechanisms operate in exchange-traded funds. Authorized participants can create or redeem fund shares against specified baskets of assets or cash. Buying an underpriced basket and delivering it for higher-priced fund shares, or reversing the process, helps connect the fund’s market price with its underlying holdings. Access restrictions, underlying-market closures, and basket trading costs determine how tightly this mechanism operates."
    },
    {
      "heading": "No-arbitrage valuation and the fundamental theorems",
      "text": "The fundamental theorem of asset pricing formalizes the link between the absence of arbitrage and consistent valuation. In a finite, frictionless market model, absence of arbitrage is equivalent to the existence of a probability measure, equivalent to the model’s original measure, under which discounted asset prices are martingales. Equivalence means that the measures agree about which events have zero probability.\n\nA martingale has the property that its conditional expected future value equals its current value. If a safely accumulating bank account is used as the unit of valuation, discounted tradable prices must satisfy this property under the valuation measure. This does not imply that investors actually expect all assets to earn the safe rate. The probability measure changes the weights assigned to future outcomes so that risk-adjusted prices become internally consistent.\n\nFor a replicable terminal payment \\(X_T\\), with a constant safe rate \\(r\\), the valuation formula is\n\n\\[V_0=e^{-rT}\\mathbb{E}^{Q}[X_T],\\]\n\nwhere \\(Q\\) is a suitable risk-neutral measure and \\(\\mathbb{E}^{Q}\\) denotes expectation under that measure. The expression resembles ordinary discounted expected cash flow, but the probabilities are valuation probabilities rather than necessarily real-world frequencies. Confusing the two can produce incorrect conclusions about both fair values and expected investment returns.\n\nThe distinction between complete and incomplete markets is equally important. In a complete market, every contingent payment allowed by the model can be reproduced by trading the available assets. Under the standard finite-model assumptions, the equivalent martingale measure is then unique. In an incomplete market, some payments cannot be replicated, and more than one valuation measure can be consistent with no arbitrage. The principle may determine a range of admissible prices rather than one definitive price.\n\nFor a one-period claim with payments \\(X_s\\) in states \\(s\\), the same logic can be written using state prices:\n\n\\[V_0=\\sum_s \\pi_sX_s,\\]\n\nwhere \\(\\pi_s\\) is the current price of receiving one unit of money in state \\(s\\). Strictly positive state prices prevent a nonnegative payoff with some positive outcome from having zero or negative cost. This representation links arbitrage theory with contingent-claim approaches to general equilibrium.\n\nContinuous-time models require additional care. Strategies that repeatedly double a losing bet, or rely on unbounded borrowing, can generate misleading mathematical opportunities. Admissibility restrictions exclude such constructions. In general semimartingale settings, the relevant condition is commonly formulated as “no free lunch with vanishing risk,” a stronger technical requirement than merely ruling out an individual arbitrage strategy. No-arbitrage valuation is therefore a theorem about a specified model and its permitted strategies, not an unconditional claim about every observable market price."
    },
    {
      "heading": "Statistical arbitrage, merger arbitrage, and other risky strategies",
      "text": "Market usage extends “arbitrage” well beyond its strict definition. Many strategies buy relatively cheap instruments and sell relatively expensive ones without guaranteeing that their payoffs will converge. Their returns depend on uncertain events, estimated relationships, and the timing of market adjustment. The term “relative-value strategy” often describes their economic character more accurately.\n\nStatistical arbitrage uses historical or real-time data to identify patterns in relative prices or returns. In pairs trading, a trader might buy one company’s shares and short another’s after their prices diverge from an estimated relationship. A position can be designed to reduce exposure to the broad equity market, but market neutrality does not eliminate firm-specific or model risk. Two companies in the same industry can diverge permanently because of financing problems, technology, regulation, or changes in competitive position.\n\nMore elaborate strategies combine hundreds of securities and use forecasts of short-horizon returns, factor exposures, and transaction costs. Machine learning may be used to estimate signals, but profitability remains sensitive to data quality, changing relationships, and implementation. A statistically unusual price is not necessarily incorrect. Apparent opportunities in historical tests can also result from survivorship bias, look-ahead information, or repeatedly searching many candidate rules.\n\nMerger arbitrage, also called risk arbitrage, centers on announced acquisitions. In a cash acquisition offering 50 per share, the target might trade at 48 because completion is uncertain and payment is delayed. Purchasing at 48 can earn the difference if the deal closes, but the share price may fall substantially if it fails. In a stock-for-stock transaction, traders may buy the target and short the acquirer in proportions related to the exchange terms. Regulatory approval, financing, shareholder votes, and contractual termination rights all affect the outcome.\n\nConvertible-bond arbitrage typically combines a long position in a bond convertible into shares with a short position in the issuer’s equity. The hedge can reduce some equity-price exposure while leaving exposure to volatility, credit deterioration, interest rates, dividends, and stock-borrowing conditions. Fixed-income relative-value trades similarly compare securities with related cash flows but may retain differences in liquidity, financing, or credit quality.\n\nThese strategies can have positive expected returns without containing any strict arbitrage opportunity. Their losses are often concentrated in unusual market conditions, precisely when positions are difficult to close. Evaluating them requires attention to the full probability distribution of outcomes, not merely a high historical success rate. Frequent small gains can coexist with infrequent losses large enough to dominate long-run performance."
    },
    {
      "heading": "Limits to arbitrage and financial crises",
      "text": "An apparent mispricing can persist when exploiting it requires scarce capital, uncertain financing, or exposure to losses before convergence. These obstacles are known collectively as limits to arbitrage. They explain why the presence of sophisticated traders does not guarantee immediate elimination of every price discrepancy.\n\nShort selling is one important constraint. An arbitrageur must generally locate securities to borrow, provide collateral, and pay borrowing fees. A lender may recall the securities, forcing the short position to close before the expected price adjustment. Some securities are difficult or impossible to borrow, and legal restrictions may prevent particular institutions from shorting them. A seemingly overpriced asset can therefore remain expensive even when many investors recognize the discrepancy.\n\nFinancing creates a second constraint. Through margin trading and repurchase agreements, traders can fund positions with borrowed money, but lenders impose margins or haircuts. If market prices move adversely, additional collateral may be required. A position whose final payoff is tightly hedged can still fail operationally if the trader cannot meet intermediate cash demands. Contractual convergence at maturity does not guarantee survival until maturity.\n\nAndrei Shleifer and Robert Vishny’s 1997 analysis of the limits of arbitrage emphasized the organizational problem of managing outside capital. Investors may withdraw funds after losses, even when a manager believes the underlying opportunity has improved. Anticipating withdrawals, an arbitrageur may rationally avoid a large position in a visibly mispriced asset. The horizon and incentives of the capital provider become part of the pricing mechanism.\n\nThe 1998 crisis surrounding Long-Term Capital Management illustrated these dangers. The fund held highly leveraged relative-value positions whose expected convergence was disrupted by market turmoil following Russia’s debt default and ruble devaluation. Spreads moved against several positions, liquidity deteriorated, and financing pressure intensified. A private-sector recapitalization was arranged in September 1998 with the facilitation of the Federal Reserve Bank of New York. The episode demonstrated that correlated financing and liquidity shocks can overwhelm positions intended to offset ordinary market movements.\n\nDuring the global financial crisis of 2007–2009, differences between funding rates and persistent deviations from covered interest parity likewise highlighted the importance of intermediary balance sheets. The theoretical trade could require access to funding at rates unavailable to the institution attempting it. When many traders must reduce financial leverage simultaneously, sales depress prices, prompting further collateral demands and additional sales. Such fire-sale dynamics can widen discrepancies instead of correcting them, at least temporarily."
    },
    {
      "heading": "Execution, settlement, and market infrastructure",
      "text": "A price discrepancy becomes a realizable arbitrage only if all required transactions can be completed on compatible terms. Trading systems must identify prices, verify available quantities, submit orders, obtain financing, and manage delivery. An executable opportunity may last less time than it takes to transmit an order between venues. The difference between an observed quotation and an obtainable transaction is therefore fundamental.\n\nOne operational problem is leg risk: one part of a multi-part trade executes while another does not. A trader buying an asset on one exchange may discover that the higher bid elsewhere has disappeared. Some exchanges offer combination orders or spread contracts that execute linked transactions together, reducing this risk. They cannot necessarily coordinate every external market, currency conversion, or financing agreement involved in the strategy.\n\nAutomated systems use direct market data and rapid order routing to exploit short-lived differences. High-frequency trading includes forms of cross-market arbitrage, but not all high-frequency strategies are arbitrage and not all arbitrage requires high speed. Physical commodity transactions or merger-related positions may last months. Speed is most valuable when competing traders can readily identify and execute the same small discrepancy.\n\nClearing and settlement determine whether the contracted exchange of assets and money actually occurs. Delivery-versus-payment arrangements link securities delivery to cash payment, while payment-versus-payment mechanisms reduce the danger of paying away one currency without receiving the other. A central counterparty can substitute itself between buyers and sellers, requiring margin and managing defaults. These arrangements reduce some bilateral risks but create their own collateral, liquidity, and operational requirements.\n\nSettlement timing also matters. Two instruments may represent equivalent economic claims but settle on different dates, producing a temporary financing need. Securities may be held in different depositories, subject to transfer restrictions, or unavailable for immediate delivery. A security traded in two markets may not be freely transferable between them at the moment a price difference appears. Operational fungibility must therefore be verified rather than inferred from a similar name or identifier.\n\nTechnology introduces additional failure modes. Incorrect reference data, stale quotations, mismatched units, and faulty algorithms can transform intended arbitrage into an unhedged position. Trading controls commonly include position limits, order-size checks, credit limits, and mechanisms for disabling malfunctioning systems. These controls do not create the economic opportunity, but they determine whether it can be pursued without incurring unintended exposure. Execution quality, financing access, and reliable post-trade processing are often as decisive as the price calculation itself."
    },
    {
      "heading": "Economic functions, regulatory arbitrage, and policy distinctions",
      "text": "Arbitrage contributes to price coordination. Purchases in a cheaper market and sales in a dearer one transmit information about relative scarcity and reduce inconsistent valuations. This can make prices more useful for production, consumption, and investment decisions. In securities markets, connections among cash instruments, derivatives, and fund shares allow investors to obtain related exposures through different instruments without facing wholly unrelated prices.\n\nThese benefits do not imply that every arbitrage activity produces an equal social gain. Part of a trader’s return may reward information gathering, inventory provision, or the use of scarce financing capacity. Other expenditures, particularly those aimed at gaining a tiny speed advantage over competitors, can redistribute trading profits without proportionately improving the allocation of real resources. The social value of arbitrage consequently depends on the market structure and the particular discrepancy being corrected.\n\nArbitrage also differs from a claim that prices always reflect fundamental value. No-arbitrage conditions impose consistency among tradable claims. Several mutually consistent assets may still be valued using optimistic forecasts or a low required return. If no available strategy produces a nonnegative payoff in all relevant states, disagreement about an asset’s fundamental worth does not by itself establish an arbitrage. No-arbitrage reasoning is therefore narrower than a full theory of market efficiency.\n\nRegulatory arbitrage uses differences in legal rules, classifications, or jurisdictions to obtain a more favorable treatment of economically similar activities. An institution might alter the legal form or location of a transaction, or transfer an exposure to an entity subject to different requirements. Unlike strict financial arbitrage, this need not involve inconsistent market prices or a riskless trading profit. Its central mechanism is a mismatch between economic substance and regulatory categories.\n\nTax arbitrage similarly exploits differences in taxation across instruments, investors, or jurisdictions. The relevant outcomes depend on enforceable legal treatment, eligibility rules, and anti-avoidance provisions. A transaction described commercially as tax or regulatory arbitrage is not automatically unlawful; legality depends on its structure and the applicable law. Conversely, the label does not establish that an arrangement complies with legal obligations.\n\nFor financial authorities, the practical issues include reliable settlement, market access, funding resilience, and the concentration of leveraged positions. Requirements under the Basel accords, margin rules, and other prudential measures can affect the cost of intermediary activity and thus the width of observable price gaps. Relaxing constraints may permit tighter price alignment while increasing exposure to funding shocks. Policies directed at systemic risk therefore address not only whether arbitrageurs can correct discrepancies, but also how their positions are financed and how losses would be absorbed if expected relationships break down."
    }
  ],
  "links": [
    "financial-market",
    "probability",
    "law-of-one-price",
    "short-selling",
    "bid-ask-spread",
    "market-maker",
    "gold-standard",
    "insurance",
    "nobel-memorial-prize-in-economic-sciences",
    "commodity",
    "foreign-exchange-market",
    "balance-sheet",
    "financial-derivative",
    "exchange-traded-fund",
    "expected-value",
    "general-equilibrium",
    "machine-learning",
    "probability-distribution",
    "collateral",
    "repurchase-agreement",
    "global-financial-crisis",
    "financial-leverage",
    "fire-sale",
    "taxation",
    "basel-accords",
    "systemic-risk"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}