The optimal level of budgetary goal difficulty: an experimental study

Budgets are commonly used for multiple purposes. Whereas planning budgets reflect organisations' 'best guesses' of future events, control budgets are intended to motivate high levels of effort. Budgetary control mechanisms modelled in a principal-agent setting typically find that the use of such schemes is Pareto optimal only when the budgetary goals are set to very difficult levels. This project studies how individual effort levels vary with different levels of goal difficulty in an experimental setting. We use a between-subjects factorial design. Subjects are assigned to either linear or budget-based compensation schemes at levels of goal difficulty ranging from 10% to 50% probability of goal attainment after controlling for skill and learning effects. An ink-stamping task is used to measure subjects' effort levels. Results indicate that the optimal level of budgetary goal difficulty may be less than that prescribed by information economics models.