erc/metu
INTERNATIONAL CONFERENCE IN ECONOMICS  IV
September 13-16, 2000, Ankara

A General Framework for Forecasting the Futures Market Prices: Non-Parametric Regression Analysis

Morteza Haghiri (University of Saskatchewan, Canada)

Abstract

An old definition about market says: “a market is a place or situation that puts sellers and buyers in communication with one another where prices are discovered, and where ownership may be transferred.” This global definition applies to wide variety of markets from local market as well as to the futures markets. Futures markets exist to assist in the process of price discovery and provide a mechanism that allows producers, holders, and users of a commodity to transfer and reduce price risk. The transfer is accomplished by taking equal and opposite positions in the futures and cash markets. There is also no question that futures markets move quickly, although its motives in the markets and the accuracy of the prices it generates, is always being debated.

Four main functions have been recognized for the performance of the futures markets: (1) facilitate risk management, as they provide a hedging facility; (2) facilitate stock-holding, as the difference between the futures price and the spot price is a guide for inventory control and may be interpreted as a price of storage; (3) act as a centre for the collection and dissemination of information, and (4) perform a forward pricing function, as futures prices can be interpreted as an indication of what the market can anticipate to be the subsequent cash price.

Price analysis is the key to successful use of any futures or options market. Whether a trader is speculating or hedging, some expectation of prices is required for development of a trading strategy. These price expectations can be derived in a number of ways. As date, two general categories of price analysis have been used in futures and options markets. First, fundamental analysis that is a way of looking forward, gathering data concerning supply and demand conditions expected in the future. Secondly, technical analysis which is a way of looking backward, making forecasts based on historical data. Fundamental analysis is usually considered to be more difficult than technical analysis because there is no so much information to evaluate. It also requires knowledge of a specific market, whereas the same technical rules can be applied to a number of different markets.

This study tries to introduce a general framework for forecasting futures market prices using a non-parametric analysis. Non-parametric analysis, unlike the parametric one, has the advantage of not having a known priori functional form in advance. In this type of analysis we let data speak for them to have some particular form of distribution. There are 5 main sections in this paper. Section one introduces basic characteristics of futures markets. In section two, both methods of futures market analyses will be discussed in detail. Section three introduces the principle of forecasting. In section four, non-parametric regression analysis is discussed with an empirical analysis of the futures market prices. Section five concludes some remarkable points.

Economic Research Center
Middle East Technical University
06531 Ankara Turkey
Phone: + 90 312 210 3044, 210 2003
Fax: +90 312 210 1244
e-mail: metuerc@metu.edu.tr