Concurrent Seasonal Adjustment With Census X-11

Much investigative effort is being expended worldwide toward improving seasonally adjusted estimates from the Census X-11 program. Several recent investigations have resulted in a recommendation to obtain the most recent month's adjustment by applying X-11 to data through that month (concurrent adjustment), rather than relying on projected factors obtained from an earlier run. Although there are theoretical arguments supporting the applicaton of X-11 concurrently, a comprehensive study documenting the results of concurrent adjustment with X-11 on U.S. economic series has not previously been undertaken. This study evaluates the effect of applying X-11 concurrently to a set of selected Census Bureau economic time series. The accuracy of concurrent estimates, in terms of mean absolute deviations from historical estimates, is examined. The results obtained are generally favorable to concurrent adjustment.