Accurate Methods for Approximate Bayesian Computation Filtering


Journal of Financial Econometrics

automne 2015, vol. 13, n°4, pp.798-838

Départements : Finance, Economie et Sciences de la décision

Mots clés : Bandwidth, Kernel density estimation, Likelihood estimation, Model selection, Particle filter, State-space model, Value-at-risk forecasts

The Approximate Bayesian Computation (ABC) filter extends the particle filtering methodology to general state-space models in which the density of the observation conditional on the state is intractable. We provide an exact upper bound for the mean squared error of the ABC filter, and derive sufficient conditions on the bandwidth and kernel under which the ABC filter converges to the target distribution as the number of particles goes to infinity. The optimal convergence rate decreases with the dimension of the observation space but is invariant to the complexity of the state space. We show that the adaptive bandwidth commonly used in the ABC literature can lead to an inconsistent filter. We develop a plug-in bandwidth guaranteeing convergence at the optimal rate, and demonstrate the powerful estimation, model selection, and forecasting performance of the resulting filter in a variety of examples

Central clearing and collateral demand


Journal of Financial Economics

mai 2015, vol. 116, n°2, pp.237-256

Départements : Finance, GREGHEC (CNRS)

Mots clés : Central clearing party; Margin; Credit default swap; Collateral; Client clearing

We use an extensive data set of bilateral credit default swap (CDS) positions to estimatethe impact on collateral demand of new clearing and margin regulations. The estimatedcollateral demands include initial margin and the frictional demands associated with themovement of variation margin through the network of market participants. We estimatethe impact on total collateral demand of more widespread initial margin requirements,increased novation of CDS to central clearing parties (CCPs), an increase in the number ofclearing members, the proliferation of CCPs of both specialized and non-specialized types,collateral rehypothecation practices, and client clearing. System-wide collateral demand isincreased significantly by the application of initial margin requirements for dealers,whether or not the CDS are cleared. Given these dealer-to-dealer initial margin requirements,mandatory central clearing is shown to lower, not raise, system-wide collateraldemand, provided there is no significant proliferation of CCPs. Central clearing does,however, have significant distributional consequences for collateral requirements acrossmarket participants

Dynamics of Innovation and Risk


Review of Financial Studies

mai 2015, vol. 28, n°5, pp.1353-1380

Départements : Finance

We study the dynamics of an innovative industry in which agents learn about the likelihood of negative shocks. Managers can exert risk prevention effort to mitigate the consequences of shocks. If no shock occurs, confidence improves, attracting managers to the innovative sector. But, when confidence becomes high, inefficient managers exerting low riskprevention effort also enter. This stimulates growth, while reducing risk prevention. The longer the boom, the larger the losses if a shock occurs. Although these dynamics arise in the first-best, asymmetric information generates excessive entry of inefficient managers, earning informational rents, inflating the innovative sector, and increasing its vulnerability

Equilibrium fast trading


Journal of Financial Economics

mai 2015, vol. 116, n°2, pp.292-313

Départements : Finance, GREGHEC (CNRS)

Mots clés : High frequency trading, Liquidity welfare, Adverse selection, Investment

High-speed market connections improve investors' ability to search for attractive quotes in fragmented markets, raising gains from trade. They also enable fast traders to observe market information before slow traders, generating adverse selection, and thus negative externalities. When investing in fast trading technologies, institutions do not internalize these externalities. Accordingly, they overinvest in equilibrium. Completely banning fast trading is dominated by offering two types of markets: one accepting fast traders, the other banning them. However, utilitarian welfare is maximized by having i) a single market type on which fast and slow traders coexist and ii) Pigovian taxes on investment in the fast trading technology

Estimating ambiguity preferences and perceptions in multiple prior models: Evidence from the field


Journal of Risk and Uncertainty

16 décembre 2015, vol. 51, n°3, pp.219-244

Départements : Finance, GREGHEC (CNRS)

Mots clés : Ambiguity. Decision-making under uncertainty. Multiple prior models. Alpha-MaxMin model

We develop a tractable method to estimate multiple prior models of decisionmakingunder ambiguity. In a representative sample of the U.S. population, we measureambiguity attitudes in the gain and loss domains. We find that ambiguity aversion iscommon for uncertain events of moderate to high likelihood involving gains, butambiguity seeking prevails for low likelihoods and for losses. We show that choicesmade under ambiguity in the gain domain are best explained by the a-MaxMin model,with one parameter measuring ambiguity aversion (ambiguity preferences) and asecond parameter quantifying the perceived degree of ambiguity (perceptions aboutambiguity). The ambiguity aversion parameter a is constant and prior probability setsare asymmetric for low and high likelihood events. The data reject several othermodels, such as MaxMin and MaxMax, as well as symmetric probability intervals.Ambiguity aversion and the perceived degree of ambiguity are both higher for men andfor the college-educated. Ambiguity aversion (but not perceived ambiguity) is alsopositively related to risk aversion. In the loss domain, we find evidence of reflection,implying that ambiguity aversion for gains tends to reverse into ambiguity seeking forlosses. Our model’s estimates for preferences and perceptions about ambiguity can beused to analyze the economic and financial implications of such preferences

Fund managers under pressure: Rationale and determinants of secondary buyouts


Journal of Financial Economics

janvier 2015, vol. 115, n°1, pp.102-135

Départements : Finance

Mots clés : Leveraged buyouts, Secondary buyouts, Private equity, Limited investment horizon, Agency conflicts in fund management

The fastest growing segment of private equity (PE) deals is secondary buyouts (SBOs)—sales from one PE fund to another. Using a comprehensive sample of leveraged buyouts, we investigate whether SBOs are value-maximizing, or reflect opportunistic behavior. To proxy for adverse incentives, we develop buy and sell pressure indexes based on how close PE funds are to the end of their investment period or lifetime, their unused capital, reputation, deal activity, and fundraising frequency. We report that funds under pressure engage more in SBOs. Pressured buyers pay higher multiples, use less leverage, and syndicate less suggesting that their motive is to spend equity. Pressured sellers exit at lower multiples and have shorter holding periods. When pressured counterparties meet, deal multiples depend on differential bargaining power. Moreover, funds that invested under pressure underperform

Implied Risk Exposures


Review of Finance

octobre 2015, vol. 19, n°6, pp.2183-2222

Départements : Finance, GREGHEC (CNRS)

Mots clés : Herding, Risk Disclosure, (Stressed) Value-at-Risk, Regulatory Capital

We show how to reverse-engineer banks' risk disclosures, such as Value-at-Risk, to obtain an implied measure of their exposures to equity, interest rate, foreign exchange, and commodity risks. Factor Implied Risk Exposures (FIRE) are obtained by breaking down a change in risk disclosure into a market volatility component and a bank-specific risk exposure component. In a study of large US and international banks, we show that (1) changes in risk exposures are negatively correlated with market volatility and (2) changes in risk exposures are positively correlated across banks, which is consistent with banks exhibiting commonality in trading

Incentive pay and bank risk-taking: Evidence from Austrian, German, and Swiss banks


Journal of International Economics

juillet 2015, vol. 96, n°1, pp.123-140

Départements : Finance, GREGHEC (CNRS)

Mots clés : Trading income, Bank risk, Incentive pay, Bonus payments

We use payroll data in the Austrian, German, and Swiss banking sector to identify incentive pay in the critical banking segments of treasury/capital market management and investment banking for 67 banks. We document an economically significant correlation of incentive pay with both the level and volatility of bank trading income—particularly for the pre-crisis period 2003–2007, in which incentive pay was strongest. This result is robust if we instrument the bonus share in the capital market divisions with the strength of incentive pay in unrelated bank divisions like retail banking. Moreover, pre-crisis incentive pay appears too strong for an optimal tradeoff between trading income and risk, which maximizes the net present value of trading income. Further analyses indicate that the bonus moderation during the crisis has removed excessive pre-crisis incentive pay

Risk Attitude, Beliefs Updating, and the Information Content of Trades: An Experiment


Management Science

juin 2015, vol. 61, n°6, pp.1378-1397

Départements : Finance, GREGHEC (CNRS)

Mots clés : Behavior under uncertainty, Risk attitude, Belief updating, Financial market efficiency, Laboratory experiment

We conduct a series of experiments that simulate trading in financial markets. We find that the information content of the order flow varies with the strength of subjects' prior beliefs about fundamentals. The presence of intrinsic uncertainty about the asset's fundamentals reduces informational efficiency. This originates from subjects' risk attitudes and biases in the way some subjects update their beliefs. The behavior of approximately 63% of the subjects is consistent with the expected utility maximization. These subjects are either risk averse (52%) or risk loving (11%). About 22% of the subjects display non-Bayesian updating of beliefs: underconfidence emerges for weak prior beliefs, and confirmation bias occurs for strong prior beliefs. Non-Bayesian belief updating reduces market efficiency when subjects' prior beliefs are weak and increases it when the prior beliefs are strong

Robust Filtering


Journal of the American Statistical Association

décembre 2015, vol. 110, n°512, pp.1591-1606

Départements : Finance

Mots clés : Kalman filter, Particle filter, Robust statistics, State-space model weight degeneracy

Filtering methods are powerful tools to estimate the hidden state of a state-space model from observations available in real time. However, they are known to be highly sensitive to the presence of small misspecifications of the underlying model and to outliers in the observation process. In this paper, we show that the methodology of robust statistics can be adapted to sequential filtering. We define a filter as being robust if the relative error in the state distribution caused by misspecifications is uniformly bounded by a linear function of the perturbation size.Since standard filters are nonrobust even in the simplest cases, we propose robustified filters which provide accurate state inference in the presence of model misspecifications. The robust particle filter naturally mitigates the degeneracy problems that plague the bootstrap particle filter (Gordon, Salmond and Smith, 1993) and its many extensions. We illustrate the good properties of robust filters in linear and nonlinear state-space examples