![IconGems-Print](images/IconGems-Print.png)
580 – Time Series and Factor Models
Relative and Trend Revision Methods for Benchmarking Two Sets of Time Series
Lynn Imel
U.S. Census Bureau
Data from the same target population collected at different spans of time can be inconsistent. For example, estimates collected monthly do not necessarily sum to measures from an annual survey. The adjustment process referred to as benchmarking solves inconsistency problems such as this. Economic programs at the Census Bureau use the relative and trend revision methods to benchmark quarterly and monthly time series. Fagan (1999) generalized the methods to benchmark a series broken out in two ways. This paper describes the methods presented by Fagan and use of them on economic data.