The role of trading intensity in duration modelling and price discovery : evidence from the European carbon market

Abstract

In this study, trading intensity is employed to investigate the role of information and liquidity in duration modelling and price discovery in the two largest exchanges of the European Carbon market, namely European Climate Exchange (ECX) and Nord Pool (NP). First, duration modelling is examined for the first time in this market, and existing ACD models are empirically extended to explore the impact of stylized facts, such as non-linear effects of trading intensity and OTC transactions. Second, the “time dimension” of information is investigated focusing on the informational content of trading intensity. A Smooth-Transition-Mixture of Weibull Distributions ACD (STM-ACD) model that distinguishes between three types of trades is proposed. Time, volume and OTC transactions measure how related related a trade is to information. Third, the price impact of the “time dimension” of information is examined. A new dynamic expectations, structural pricing model is proposed in order to account for the learning process of traders and their expectations. Trading intensity is used to measure the sensitivity of market participants to information and liquidity. The main findings indicate that empirical adjustments significantly improve duration modelling. In consistence with Bauwens et al. (2004), the specification of the conditional mean contributes more to model performance. Trading intensity appears to create a momentum, especially in ECX, whereas OTC transactions seem to slow down the trading process, probably due to information inflow, especially in NP. Furthermore, similar to Easley and O’Hara (1992) higher trading intensity is associated with increased presence of information. Trading intensity is found to be able to distinguish among three different types of trades, according to their informational content. The timing of acquiring information can make it further exploitable. A significant proportion of uninformed traders in the Carbon market is found to observe the market trying to extract price unresolved information. Consequently, informed traders are found to act strategically, according to Kyle (1985), but they are less efficient in covering their actions as market gains complexity, mainly because of higher liquidity levels and improved learning process. In addition, large transactions appear to increase the information price component, while the liquidity component seems to asymmetrically decrease, probably due to economies of scale. Consequently, trading intensity appears to have a dual impact on price, spread and price change volatility, which is determined by current market conditions and dealers’ exposure to risk. Finally, market making in this market seems to be profitable only when expected trading intensity is low.