Projecting county pulpwood production with historical production and macro-economic variables

We explored forecasting of county roundwood pulpwood production with county-vector autoregressive (CVAR) and spatial panel vector autoregressive (SPVAR) methods. The analysis used timber products output data for the state of Florida, together with a set of macro-economic variables. Overall, we found the SPVAR specification produced forecasts with lower error rates compared to CVAR specifications. Nonetheless, high forecast errors across counties revealed the uncertainty associated with projecting volumes of county pulpwood production.