Title: Endogenous and exogenous volatility in the foreign exchange market
Authors: Leonardo Bargigli - Università di Firenze (Italy) [presenting]
Abstract: Two sources of heteroskedasticity in high-frequency financial data are identified. The first source is the endogenous changing participation of heterogeneous speculators to the market, coupled with the time-varying behaviour of the market maker. The second source is the exogenous flow of market-relevant information. We model the first one using a Markov switching (MS) SVAR process and the second one utilizing a GARCH process for the MS-SVAR structural errors. Using transaction data of the EUR-USD market in 2016, we detect three regimes characterized by different levels of endogenous volatility. The impact of structural shocks on the market depends on both sources, but the exogenous information is channelled to the market mostly through price. This suggests that the market maker is better informed than the speculators, who act as momentum traders. The latter can profit from trade because, unlike noise traders, they respond immediately to price shocks.