3L Finance Workshop

09:00-09:40 — The Impact of Clearing Fees on Market Quality — Hans Degryse, Jose Mendoza, and Gunther Wuyts (all KU Leuven etc) — Discussant: Nikolaos Karagianis, KU Leuven

This paper aims to provide a more integrated view of market quality by linking the whole value chain in securities transactions: trading and post-trading. We empirically study the impact of post-trade fees, particularly clearing fees, on various dimensions of market quality. As a natural experiment, we use a 25% average reduction by LCH.Clearnet’s clearing fees on April 1, 2008. After this fee reduction, we find an improvement of various spread measures (quoted, realized, effective). But while this dimension of liquidity improves, depth decreases by around 14% with respect to average pre-treatment levels, and resiliency (the recovery of liquidity after a shock) deteriorates as well. These results imply that small trades could benefit from the post-trade fee reduction, while larger trades may face higher trading costs. Finally, we find no effect on trading activity and volatility, nor on price impact (adverse selection).

09:40-10:20 — Once Upon a Broker Time: Order Preferencing and Market Quality — Hans Degryse and Nikolaos Karagiannis (all KU Leuven etc) – Discussant: José Mendoza, KU Leuven

We develop a one-tick dynamic microstructure model to study how priority rules determine market quality and investor welfare. We compare order preferencing, modeled as price-broker-time priority (PBT), to price-time priority (PT). Priority rules impact investors’ choice between limit and market orders. When the tick is tight, trading rates are higher with PBT whereas investor welfare is higher with PT. The opposite holds for a wide tick. PBT endogeneously results when brokers individually choose between PT or PBT. Our model has testable implications regarding systematic patterns in order flow, market depth, trade composition, and market fragmentation.

10:50-11:30 — MIDAS-PRO-LASSO: Mixed Frequency Data Regression Models with Parameter Profiling and LASSO — E. Ghysels (UNC); L. Iania and Jonas Striaukas (both LSM) — Discussant: Luca Barbaglia, KU Leuven and NBB

We study MIDAS regressions with potentially large set of predictor variables. We propose a two-step efficient estimation procedure that can be used with both linear and nonlinear MIDAS regression models which we call MIDAS-PRO-LASSO. We adopt a MIDAS framework with polynomial parameter profiling as suggested in Ghysels and Qian (2016). A typical MIDAS regression involves two types of parameters: (1) intercept and slope, and (2) MIDAS distributed lag parameters. In the MIDASPRO-MIDAS procedure, the MIDAS polynomial parameters are estimated separate from the intercept and slope coefficients. The latter are subject to LASSO-type penalties whereas the former will be profiled and not subject to penalty. The two-step procedure allows us to apply fast and efficient algorithms for estimating penalized slope coefficients using for example the coordinate descent approach put forward by Friedman et al. (2007). We also provide results for MIDAS Ridge regression estimators using the variable transformation considered by Hastie and Zou (2005). Consequently, we relax the zero restriction on the penalty term for MIDAS parameters which enables us to consider cases with more covariates than observations (henceforth denoted as p ≫ n). In this case, we can also use the Elastic Net as a penalty function to perform variable selection. Lastly, we also discuss the relevance of equivalence results between Adaptive Ridge and LASSO in estimating MIDAS-PRO-LASSO (see Grandvalet (1998) for further details).

11:30-12:10 — Asymmetric Term Structure Modelling — H. Dewachter (NBB); E. Ghysels (UNC); L. Iania and Jean-Charles Wijnants (both LSM) — Discussant: Hamza Hanbali. KU Leuven

We develop a new class of term structure models in which the pricing factors can exhibit time-varying conditional skewness. A recent literature. on empirical asset pricing focuses on the predictive role of macroeconomic variable skewness on equity excess returns. Colacito et al. (2016) document that the cross-sectional skewness of the distribution of professional forecasters’ expected US GDP growth, which is time-varying, has predictive power on future equity excess returns over and above the standard predictors in the equity return forecasting literature. The workhorse Gaussian affine term structure model, assuming symmetry, is silent about this salient feature of macro-financial data. To fill this gap in the literature on reduced-form asset pricing of interest rates, the conditional distribution of the risk factors in our term structure model has skew-normal distributions and time-varying parameters, which also impact the conditional mean, conditional variance and conditional kurtosis. Estimation pertains to the literature on non-Gaussian and non-linear state space models and will require the use of Sequential Monte Carlo simulation methods. Our preliminary results tend to highlight the important changes in the conditional asymmetry of interest rates over the business cycles. The explanation for these variations is to be found both in the cyclical behaviour of monetary policy and the reaction of macroeconomic fundamentals to business cycles. In addition, interest rates’ conditional skewness seems to be an important predictor of future bond excess returns in addition to common predictors of bond risk premia identified in the literature.

13:00-13:40 — Googlization and retail investment decisions — Christophe Desagre and Catherine D’hondt (both LSM) — Discussant: your name could be here

Several papers have highlighted a positive link between Google queries, measured by the Search Volume Index (SVI), and financial markets. These studies focused mostly on trading volume, a publicly disclosed measure encompassing all trading activities. In the present paper, we further investigate this relationship on a proprietary database that covers over two millions retail investors’ transactions that were conducted on a Belgian brokerage online platform between January 2003 and March 2012. Our objective is twofold. First, we analyze whether a relation exists between retail volume and Google SVI. Then, we construct portfolios according to that index and test whether that strategy generates abnormal excess returns. Our results confirm that there exists a positive and significant relation between SVI and retail investors’ transactions, implying that retail investors are likely to search for information using the stock ticker on Google before trading.

13:40–14:20 — Tone Management and Labor Unions — Özgür Arslan-Ayaydin, James Thewissen and Wouter Torsin (all KU Leuven) — Discussant: your name could be here

This study examines whether managers deflate the tone of earnings press releases to convey a pessimistic picture about the firm’s future prospects, with the objective to bolster the firm’s bargaining position with labor unions. We find that the tone of the qualitative information in earnings press releases is significantly more pessimistic as the degree of unionization increases. Our results also show that this effect is more pronounced in states without right-to-work laws, for profitable and financially flexible firms, and firms in states strongly influenced by the Democratic Party. Results from quasi-natural experiments suggest that labor unions causally affect the use of tone deflation and that tone deflation is stronger during the labor negotiation period. Our findings also indicate that labor unions lead to a significant weakening of the signaling value of the tone of earnings press releases to predict future performance.

14:20–15:00 — TBA (Lille2) — Discussant: your name could be here

Adaptive Markets: Financial Evolution at the Speed of Thought. One of the most important economic debates of our time is whether investors and markets are rational and efficient, as modern financial theory assumes, or irrational and inefficient, as behavioral economists believe. In this talk, Andrew Lo cuts through this debate with a new framework, the Adaptive Markets Hypothesis, in which rationality and irrationality coexist. Drawing on psychology, evolutionary biology, neuroscience, artificial intelligence, and other fields, Adaptive Markets shows that the theory of market efficiency isn't wrong but merely incomplete. When markets are unstable, investors react instinctively, creating inefficiencies for others to exploit. Lo's new paradigm explains how evolution shapes behavior and markets at the speed of thought—a fact revealed by swings between stability and crisis, profit and loss, and innovation and regulation.