This post includes topics covered in the (virtual) Timber Market Analysis course on October 11^th and 12th. Also, for those interested in this type of research, please visit Forisk’s forest economist job opening.

Each quarter, when publishing the Forisk Research Quarterly, our team reviews the performance of prior projections to learn and adjust. On an annual basis, we back-test these models to rebuild and improve them. While helpful and educational, the exercise of back-testing and refining models raises a question: are we forecasting the future or are we explaining the past?

Forecasting serves many roles. In addition to helping set budgets and estimate values, it can eliminate options, allocate resources, and identify key drivers. If forecasting was simply the projection of prices, why not simply draw a trend line and move on?  (Maybe we should.) However, in our respective fields, we have at our disposal hard facts and accumulated experience, and the process of developing forecasts and scenarios levers these to increase our understanding and make decisions.

In this way, building a model to forecast (timber) prices includes identifying and understanding relationships in historical data to infer future conditions. So, first, we do use back-testing to help explain what happened in the past. Our understanding of how, for example, natural disasters like hurricanes affect local timber markets and production is largely a retrospective one. We use what we’ve learned across markets over time to better explain the effects of unique events. Then, second, we take our new learnings and apply them to scenarios and other events in the future.

In short, the act of developing forecasts is both explanatory and prospective. For evaluating markets for timber and forest products, our process is explicitly applied. We take the variables we understand, such as wood-using mills and forest supplies, and evaluate a range of scenarios to provide a set of plausible future outcomes.

To increase the relevance and usefulness of this work, we want to structure the forecasting process around scenarios designed to help make decisions. Developing scenarios that answer client questions and support decision-making is a current focus and priority. If the goal of a timber forecast is to plug a hole or check a box, then we could use the current price or a trend line. But if the goal is to develop a nuanced view of the future that supports investment decisions, then we want to confirm the forecast process helps us do this.