One of the sectors for which RISI makes forecasts is European graphic papers. The methodology for this model is detailed below and is representative of our modeling approach, although each model is unique to its sector.
Demand by End Use
Total consumption is estimated for each end use tracked. An end-use factor is forecast for each end use, which relates the tons of product consumed per unit of end-use activity. For example, the end-use factor for newspapers is tons of graphic paper consumed in newspapers per thousand dollars of real advertising expenditures in newspapers. The end-use factors are forecast as a function of time and the weighted real prices. Time is used to account for changes in consumer preference, technological change, and lack of complete data on activity levels in particular end uses. This last function is important in an end use such as magazines, where paper usage in directories is included, but the end-use indicator is only real advertising expenditures in magazines.
The weighted price is calculated by weighting the prices of each product used in an end use by the share of that grade within the consumption within the end use. Then, this weighted price is put into real terms by dividing by the producer price index for that category. The weighted real price is used to capture the demand elasticity with respect to price; i.e., the percentage change in demand for a one percent change in real prices. Once the end-use factors are forecast, the total in each end use is calculated by multiplying the end use factor by the level of activity.
Consumer Inventories by End Use and Apparent Consumption
After consumption by end use is forecast, the next step is to arrive at apparent consumption. Apparent consumption can be calculated historically by using the formula shipments+imports-exports. Total shipments of each grade are not always available, so we use production as a proxy for shipments in calculating historical apparent consumption. Apparent consumption tends to bounce around quite a bit on a quarter-to-quarter basis, mainly due to changes to consumer inventories. We have attempted to estimate these quarter-to-quarter changes in consumer inventories to allow the estimation of end-use factors, which should be calculated using actual consumption, and to allow explicit assumptions on consumer inventory behavior in the forecast.
The change in consumer inventories is forecast exogenously. That is, our economists make their best guess as to the movement of consumer inventories, given the estimated level of consumer inventories relative to consumption at the beginning of the forecast, forecast changes in prices, and market tightness. The inclusion of consumer inventory estimates can be quite important in near-term demand forecasting due to the large swings that can sometimes take place. We believe that it is important to show our assumptions on consumer inventories, even if the historical data has to be estimated.
Demand by Grade/Product in Each End Use
After apparent consumption has been estimated in each end use, the shares of grades/products are forecast. The shares are a function of time and the price of each grade relative to the weighted average prices within each end use. Time is used to account for changes in consumer preference; e.g., the movement toward higher quality, no matter what the price. Relative prices are used to capture the substitute price elasticity; i.e., the impact of price changes between grades on demand for each particular grade within the end use. Demand by grade in each end use is then calculated by multiplying the share by total apparent consumption in the end use.
Net Exports
Net exports are forecast for each grade as a function of the price of the grade in the region relative to the price in the U.S. and the amount of capacity available for export in that region. The relative price of paper in the region and the U.S. accounts for the profit incentive to ship out of the region. The U.S. price is used because of the large amount of product that is shipped by regional producers to the U.S. Capacity available for export accounts for the incentive to fill out excess capacity in a loose market by expanding into export markets. Capacity available for export is calculated for each grade as capacity minus apparent consumption. After net exports are forecast, production of each grade can be calculated as apparent consumption plus net exports.
Capacity
Capacity by grade/product is forecast exogenously, based on announced capacity changes over the near term and our estimates over the longer term. The long-term forecast for capacity is based on projected developments in profitability and demand growth for each grade. The capacity expansion forecast takes into account the time lag between changes in profitability and when capacity actually comes onstream. After capacity has been estimated, operating rates can be calculated as production divided by capacity.
Production Costs
The models are used to estimate historical production costs for major traditional producers. Therefore, cost models vary based on the products being forecast. The technological factors contained in the model are projected into the future, based on our assumptions about technological change in each of the components. Prices of the major inputs to the production process are then forecast, based on the general macroeconomic environment foreseen and assumptions about specific inputs within the individual countries. For example, we forecast wood prices in Finland and Sweden from demand/supply balance within the domestic timber markets, the general outlook for wood prices in major parts of the world and the pricing of products produced from the timber. RISI also forecasts exchange rates, which are used to translate the local currency costs into local currency terms.
Prices
Prices of selected major subgrades within each major grade are the last element in the industry forecast. Prices are forecast as a function of operating rates for each major grade and average variable costs for the grade. Average variable costs provide the cost floor for prices. Operating rates are used in forecasting prices to estimate how high prices will be off of the variable cost floor at any point in time. The ratio of shipments to capacity is used, where available. Where shipments are not available, we use the ratio of production to capacity; e.g., operating rates. Operating rates indicate the tightness of the market and the ability of producers to generate profit margins. Once prices are forecast, they feed back into the demand side of the model in the later time periods.