The purpose is to find differences in return on capital for Japanese pulp and paper companies. The hypothesis is that return on capital can be explained by mill size, productivity, production line, financial data and use of wood resources. Return on capital is measured as income before tax related to stockholders' equity. This variable can be split into profit margin and asset utilization rate. The dataset from 1991-2001 consists of 13 larger Japanese pulp and paper companies. The best model fit is found when using asset utilization rate as dependent variable. Significant variables are asset utilization rate lagged one time period, labor productivity, capital productivity, paper production as a share of total paper and board production, total value of assets on the books and solidity. Among these variables, lagged dependent variable, capital productivity, and total value of assets on the books are negatively correlated with asset utilization rate; the others are positively correlated.