| dc.description.abstract | This report gives a Bayesian analysis of operating cost escalation in the Norwegian Armed Forces. The amount of
relevant data is limited to nine data points from the time period 1994-2002. We present a statistical model of an
underlying cost escalation, with random noise both in the actual cost and the observation of it. Through discussions with
experts in the field, we have constructed prior distributions for these three parameters (cost escalation, process noise and
observation noise). We have then computed the posterior distribution by methods of Markov chain Monte Carlo
(MCMC) simulation and Kalman filtering. The model has been evaluated by the use of posterior predictive p-values,
and found to fit the data sufficiently well. The analysis gives a 95% credibility interval of (0.2% - 4.5%) for the
operating cost escalation parameter. | en_GB |