Abstract

The aim of this paper is to investigate the liquidity of non-government bonds fragmented across the three main Italian retail bond markets (DomesticMOT, ExtraMOT, and EuroTLX) from January 1, 2010 to June 30, 2013. In order to account for different aspects of liquidity, four measures are used: zero-trade, turnover ratio, price impact (Amihud indicator) and bid-ask spread (Roll indicator). The use of all these indicators is supported by the evidence of a principal component analysis, showing that liquidity of dual-listed bonds cannot be summarized by one single indicator over the sample period, since it results from the even contribution of the four measures. Fragmented bonds can be traded either on DomesticMOT and EuroTLX or on ExtraMOT and EuroTLX. As for bonds traded on DomesticMOT and EuroTLX, we find that liquidity is similar across the two venues when using zero-trade and turnover ratio, whereas it is higher on EuroTLX if we use price impact and bid-ask spread. As for the bonds traded across ExtraMOT and EuroTLX, liquidity is on average higher on EuroTLX. Moreover, irrespective of the trading venue, on average bank bonds turn out to be less liquid than non-financial bonds, especially during the sovereign debt crisis. We also show that securities’ characteristics (such as minimum trading size, coupon type, complexity, issuer sector and nationality) may impact differently on liquidity depending on the trading venue, thus suggesting that market microstructure plays a relevant role. The multivariate analysis confirms this evidence, by showing that, controlling for bond characteristics, liquidity changes across trading venues. Finally, the paper investigates the effect of fragmentation by comparing the liquidity of bank bonds fragmented across DomesticMOT and EuroTLX with otherwise similar bank bonds traded on DomesticMOT only. We show that bonds issued by Italian banks traded on DomesticMOT and EuroTLX have similar or higher liquidity (depending on the measure adopted) than otherwise similar Italian bank bonds traded on DomesticMOT only, whereas the opposite result holds true for foreign bank bonds. Therefore, we do not find a clear-cut evidence on the effect of fragmentation on bond liquidity, because it is probably intertwined with bonds’ attributes, such as the issue size (which, in our sample, is higher for the Italian bank bonds compared to the foreign ones). To our knowledge, this is the first paper to investigate the liquidity of dual-listed bonds and the impact of fragmentation on retail corporate bond markets, providing new empirical evidence on whether transparency and market microstructure rules may contribute to the development of an integrated secondary market. In this respect, the paper is relevant also on policy grounds, given that the recent MiFID review envisages greater transparency in non-equity markets.

The authors thank Giovanni Petrella, Andrea Resti, Juan Roboredo, Giovanni Siciliano and the participants to the Portuguese Financial Network 2014 Conference, held in Vilamoura, Portugal, on June 18-20, 2014, for useful comments to an earlier version of the paper. Of course, the authors are the only responsible for errors and imprecisions. The opinions expressed here are those of the authors and do not necessarily reflect those of Consob.