Dynamics of Pinus wood prices for different timber assortments: comparison of stochastic processes

Abstract

Understanding the dynamics of market prices for Pinus wood is a prerequisite for strategic decisions concerning forest investment plans since, in terms of the market, the exogenous risk to a project depends on timber assortments. The stochastic process that represents the best way of pricing the underlying asset therefore needs to be known. This study set out to compare Fractional Brownian Motion and Geometric Brownian Motion, through econometric tests, to understand the stochastic model that best represents the price behaviour of Pinus wood from planted forests in the state of Santa Catarina, Brazil, for the pricing of the underlying asset and valuation of the real options intrinsic to forest investment projects. The time series of prices, for the period from June 2017 to July 2019, relate to three assortments of Pinus wood used for multiple products. The recommended econometric tests to analyse the time series were for normality of the data, trend, autocorrelation, stationarity and fractional differential estimation. The time series were then modelled by means of stochastic processes in line with the econometric tests. The time series showed normal behaviour and indicated the presence of a positive trend and non-stationarity in the data. In addition, a true long memory was found in all series. Fractional Brownian Motion proved to be the most suitable stochastic process for modelling the prices of three forest timber assortments, given the non-stationary characteristics and true long memory of the time series for Pinus wood prices.