Files in this item

FilesDescriptionFormat

application/pdf

application/pdfHUANG-THESIS-2016.pdf (854kB)Restricted Access
(no description provided)PDF

Description

Title:Are futures prices good price forecasts? Nonlinearities in efficiency and risk premiums in the soybean futures complex
Author(s):Huang, Joshua
Advisor(s):Serra, Teresa; Garcia, Philip
Department / Program:Agr & Consumer Economics
Discipline:Agricultural & Applied Econ
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:M.S.
Genre:Thesis
Subject(s):risk premium
market efficiency
nonlinearity
Abstract:Prior to 2005, the evidence suggested that futures markets were relatively efficient in the long run, but short-run inefficiencies existed in certain markets for particular periods. Recent research has pointed to a reduction in predictive content in several agricultural markets, but the specific sources of the decline are less well understood. We investigate short-run forecasting in the soybean futures market complex to more clearly identify predictive content and the sources of forecast errors. We concentrate on two-month forecast horizons using a nonlinear framework and 1973-2016 data that allows us to identify the effects of market exuberance/pessimism and risk premiums on prices. A non-parametric local linear regression framework (Fan and Gijbels, 1996) is first applied to investigate biasness, and to guide the specification of parametric regime-switching models in which we perform statistical testing. To identify effects of risk premiums, which have been difficult to estimate in a forecasting context (Frank and Garcia 2009), we use a realized GARCH model (Hansen, Huang and Shek 2012) that has been shown to improve conditional volatility modeling. We focus on the markets in the soybean complex because of their economic importance, and differences that exist in the nature of markets (e.g., storability). Also, beginning with Rausser and Carter (1983), the forecast accuracy of the soybean complex has been called into question, and use of our long sample period, permits us to gain perspective on the sources of forecast errors over time. Our preliminary data analysis for the entire period identified average absolute forecast errors in percentage terms to be 1.4% for soybeans, 1.0%; soybean meal, and 2.4% for soybean oil. Non-parametric and parametric findings indicate nonlinearities in efficiency and risk premiums are present. Depending on the level of futures prices, thresholds or regimes of predictive performance exist. Evidence of market exuberance/pessimism emerges in all three markets. When prices are high (low), markets tend over- (under-) forecast subsequent prices. Differences in the regimes and sources of forecast accuracy also emerge across the markets. In soybeans, a low price regime is inefficient with no evidence of a risk premium, a middle regime is primarily affected by a risk premium, and a high price regime is affected by inefficiency and risk premiums. In soybean meal, a low price regime is inefficient with some evidence of a risk premium, and a second regime covering both middle and high prices is unbiased. In soybean oil, a first regime covering low and middle prices is unbiased, and a high price regime is inefficient without a risk premium. While results differ in the three markets, they indicate that the recent market booms did affect forecasting performance through exuberance and changing risk premiums. However, they also identify that in periods of low prices predictive content is primarily affected by “pessimism”. In most cases, these non-linear findings differ substantially from those generated in the conventional linear framework. On balance, the research highlights the importance of investigating the ability of futures markets to forecast spot prices in a non-linear context. Most research to date has ignored these non-linear price linkages when testing for biasness and the presence of risk premiums. Our research demonstrates that failure to account for these non-linear relationships can distort our understanding of market effectiveness. Finally, we show that the use of higher frequency data can be useful in identifying the presence and magnitude of risk premiums. This finding may make uncovering risk premiums in agricultural commodity markets more tractable in the future.
Issue Date:2016-12-07
Type:Thesis
URI:http://hdl.handle.net/2142/95612
Rights Information:Copyright 2016 Joshua Huang
Date Available in IDEALS:2017-03-01
Date Deposited:2016-12


This item appears in the following Collection(s)

Item Statistics