The impact of forest offset credits under a stochastic carbon price on agriculture using a rational expectations and real options framework

Abstract

With climate change becoming an increasingly pressing issue and a world population expecting to reach seven billion people in 2011, policies to mitigate greenhouse gas emissions are likely to be enacted domestically as well as internationally. The possible interference of those policies with commodity supply, and hence food security, are the subject of this dissertation.In 2009, a bill to reduce U.S. greenhouse gas emissions passed the House of Representatives but did not pass in the Senate. The bill would have established an emission trading system to reduce emissions from the energy, industrial, and transportation sectors. The bill also included an amendment which would have allowed the agricultural sector to provide the market with carbon offset credits to lower compliance costs for capped sectors and to compensate farmers for an expected increase in energy prices. Soon after the announcement of the offset provisions, concerns of higher commodity prices surfaced because the amendment allowed for credits from afforestation activities on cropland. This dissertation quantifies the effects of those offsets in terms of commodity prices, land allocation, landowner's welfare, and carbon sequestration.The basic model involves a landowner whose plot of land can be in either of two regimes: agriculture or forestry. Revenues in both regimes are uncertain due to price and yield fluctuations while in agriculture and allowance price volatility while in forestry. The sunk cost associated with switching as well as the uncertainty motivates the use of a real option switching model. It might be optimal for a landowner to delay afforestation in order to gain more information about the future carbon price or agricultural revenue. Furthermore, the investment in planting a forest is difficult to reverse. Besides the high costs of forest clearing, the legislation requires a plot of land to be in forestry for several years in order to earn carbon credits.In our model, the landowner observes each period's net revenue in both activities and forms expectations about the future evolution of prices and then decides whether switching to a different regime is optimal or not. A key aspect of our model is the presence of competitive markets. Real option models usually assume an exogenous stochastic process. In our case, revenues are influenced by the switching of landowners from one regime to the other and thus, are endogenous.The model is calibrated to the contiguous United States and includes nine crops plus pasture while in agriculture. For forestry, we impose the type of trees to be planted and show when and where land conversion between agriculture and forests occurs under domestic forestry offsets. The analysis is done at the county level in the United States to take spatial heterogeneity and biophysical constraints such as sequestration rates and yields into account. The value of the wood is included in our analysis but is assumed to be non-stochastic which facilitates the computational analysis.We show that in the presence of uncertainty, significantly less land gets converted from cropland to forestry over the projection period of 40 years. Pasture area is reduced because of low opportunity costs and because it serves as a land pool in the case of cropland expansion in counties which do not switch to forestry but increase crop area because of higher prices. In general, switching from agriculture to forestry starts occurring after a period of 25 years and leads to rising commodity prices thereafter. Ultimately, net revenue from agriculture and forestry start rising with the allowance price. Also, almost no afforestation takes place in the Corn Belt.From a policy perspective, less afforestation leads to smaller welfare effects for farmers than previously estimated and to a higher carbon price because domestic offsets are not supplied in quantities that allows for a significant allowance price reduction.