Markov chain is more appropriate to develop and verify a deterioration predication model for the given LGA because continuous inspection data sets are not available. Also, the assumption of unit change of facility condition is not appropriate for building condition data because the deterioration of building components might undergo a multistate transition (i.e., jumping from condition 3 to condition 6 within one time step). Hence, the Markov chain is selected to derive the deterioration prediction model. Local government agency inspection data of two subsequent inspections were used. Out of the two data sets available, one data set is used for model parameter calibration (Data 1) and the other data set is used for model validation (Data 2). The building condition prediction model presumes only the condition degradation in a one-step time transition (i.e., deterioration). Hence, the component cannot be at a better condition in the subsequent time step. Under this assumption all repair, rehabilitation, or maintenance work that would bring the building component condition to a better state are eliminated. An exemple 10-state MC model used in deterioration prediction of building components is shown in Fig. 2.