The Bureau of Economic Analysis is moving ahead on the second phase of a three-pronged plan to refine its estimates of gross domestic product by reducing potential sources of “residual” seasonality. That’s when seasonal patterns remain in data even after they are adjusted for seasonal variations.
As part of the second phase, BEA has conducted a component-by-component review of data feeding into quarterly estimates of GDP and Gross Domestic Income to see if there is residual seasonality in the numbers and to identify causes. The results of that review are reported in a new BEA paper called “Residual Seasonality in GDP and GDI: Findings and Next Steps.”
“In the case of GDP, the evidence for residual seasonality is somewhat ambiguous with tests finding evidence for residual seasonality over some time spans but not others,” the review found.
The findings will be used to identify and implement improvements to seasonal adjustment in the annual update to GDP and its major parts in late July. For instance, BEA is likely to start seasonally adjusting some data series, which feed into GDP, that show residual seasonality at a quarterly frequency, but not at a monthly frequency.
This year’s annual GDP update will also include the regular updating of seasonal factors for the period that’s open to the update. That’s typically a three-year period – in this case, from 2013 through 2015 and the first quarter of 2016.
BEA will continue to work with the Census Bureau, the largest provider of data for the calculation of GDP, as well as other agencies that supply data for GDP, to implement strategies for removing residual seasonality in source data.
In addition, BEA will review and modify its own policies governing its annual update to GDP estimates. BEA will consider adopting a longer time period (more than the current three years) in which the numbers are subject to change. A longer time period would allow BEA to better incorporate updated seasonal adjustments in published estimates.
In 2018, BEA will update its historical time series for GDP as far back as necessary to remove any remaining residual seasonality. That will occur as part of BEA’s “comprehensive” update to GDP and its major components, which happens roughly every five years.
As part of the third phase of BEA’s plan, the agency will develop methods and procedures for compiling estimates for GDP and GDI that aren’t seasonally adjusted. This would help data users identify changes in seasonal trends over time and better understand why GDP numbers may change during the update process. BEA plans to publish these unadjusted numbers alongside its adjusted GDP estimates, beginning in July 2018.
The first phase of the plan came in July 2015 when BEA began seasonally adjusting several series of data used to calculate GDP, including some from Census’ quarterly services survey, that previously hadn’t been seasonally adjusted.