University of Illinois Urbana-Champaign

Price volatility and liquidity cost in grain futures markets

Wang, Xiaoyang

Content Files
Xiaoyang_Wang.pdf

Permalink

https://hdl.handle.net/2142/50729

Description

Title
Price volatility and liquidity cost in grain futures markets
Author(s)
Wang, Xiaoyang
Issue Date
2014-09-16
Director of Research (if dissertation) or Advisor (if thesis)
Garcia, Philip
Doctoral Committee Chair(s)
Garcia, Philip
Committee Member(s)
  • Irwin, Scott H.
  • Mallory, Mindy L.
  • Peterson, Paul E.
  • Sanders, Dwight
Department of Study
Agr & Consumer Economics
Discipline
Agricultural & Applied Econ
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
Ph.D.
Degree Level
Dissertation
Date of Ingest
2014-09-16T17:25:56Z
Keyword(s)
  • Grain Futures Markets
  • Volatility
  • Liquidity Cost
  • High Frequency Trading
Abstract
Significant changes have taken place in grain futures markets. This dissertation consists of three essays investigating issues in the price volatility and liquidity cost in grain futures markets influenced by these changes. The first essay examines the sources of long memory in three major grain futures contracts, and assesses its usefulness to forecast price volatility in periods of moderate and heightened uncertainty. Using data from corn, soybeans and wheat futures contracts in 1989-2011, statistical tests and estimation results indicate that much of the long memory patterns arise from seasonality and structural breaks. After accounting for these factors, a less pronounced but still significant long memory effect exists in corn and wheat, but it disappears in soybeans. Directly modeling structural breaks through a semi-parametric method generally fails to improve forecast accuracy due to likely estimation errors that can arise in over-parameterized models. During recent heightened structural breaks, a simple long memory model provides the best forecasts especially at distant horizons, but the forecast performance of all models in this period is poor. Our findings suggest that though long memory models can be used as a parsimonious specification for structural breaks in forecasting, the reduction in forecast errors is limited. While long memory forecasts have slightly fewer rejections of unbiasness, their improvement relative to short memory forecasts is marginal. Modeling seasonality is important for better forecasting performance in these markets. The second essay is the first paper to analyze liquidity costs in agricultural futures markets based on the observed bid-ask spread (BAS) faced by market participants. Using the order book for electronically-traded corn futures contracts, this study reveals a highly liquid corn market, which with few exceptions offers order execution at minimum cost. BAS responds negatively to volume and positively to price volatility, but also affects volume traded and price volatility. While statistically significant, these responses on a cents/bushel or a percentage basis are generally small. Liquidity costs are also virtually impervious to short-term changes in demand for spreading and trend-following trader activity, as well as differences from day-of-the-week changes in market activity. Much larger cents/bushel and percentage changes in BAS occur during commodity index roll periods and on USDA report release days. The roll period findings point to a sunshine trading effect, while announcement effects identify the importance of unexpected information and adverse selection on order execution costs. Overall, the research demonstrates that the move to an electronic corn market has led to low and stable liquidity costs even in a recent period of market turbulence. The third essay pioneers research on the high frequency quoting noise in electronically traded agricultural futures markets. High frequency quoting – quickly canceling posted limit orders and replacing them with new ones – emerges as a strategy for liquidity-providing high frequency traders (HFTers) to cope with predatory trading algorithms. High frequency quoting can generate noise in price quotes which adds uncertainty to order execution and harms market quality. We measure high frequency quoting noise by the level of excess variance and discrepancies in bid/ask price co-movement at time scales as small as 250 milliseconds. Using the Best Bid Offer (BBO) dataset in 2008-2013, we simulate sub-second time stamps using a Bayesian framework. Excess variance and co-movement discrepancies are estimated using a wavelet-based short-term volatility model. We find excess high frequency quoting variance exists. It is highest at 250 milliseconds, which is 90% higher than normal. In terms of economic magnitude, net excess volatility – square root of variance – is negligibly small. At 250 milliseconds it ranges from 0.86% to 4.61% of one tick (0.025 cents/bushel), which is the minimum allowed price change. Bid/ask price co-movement shows a low degree of discrepancy with average correlation of 0.67 at 250 milliseconds. Both excess variance and bid/ask co-movement discrepancy indicate high frequency quoting noise has declined through the period.
Graduation Semester
2014-08
Permalink
http://hdl.handle.net/2142/50729
Copyright and License Information
Copyright 2014 Xiaoyang Wang

Owning Collections

Graduate Dissertations and Theses at Illinois PRIMARY
Graduate Theses and Dissertations at Illinois
Dissertations - Agricultural and Consumer Economics