Cherry Blossom Season and The District’s Sales Tax

Peak blossoms for the cherry trees in Washington DC were expected earlier this year because of unusually warm weather for many days in February and early March. However, starting March 10, there were ten consecutive days of freezing weather with a wintry mix of precipitation on March 14, putting the official date of peak bloom in jeopardy. (The National Park Service defines peak blooms as the day when 70 percent of Yoshino cherry blossoms are open).

We wondered if the change in weather would damage the cherry blossoms, and consequently if this would have an impact on sales tax collections. Since it’s too early to tell the impact for 2017, we looked at data on cherry blossom peak dates over recent years, to see if there is a visible pattern on the corresponding sales tax collections.

Because the Cherry Blossom Festival spans both March and April, we looked at data for those months individually and combined for fiscal years between 2005 and 2016. We compared past sales tax collections with the historical peak bloom dates. We also analyzed past sales tax collection data to see if the strength of March and April activity ties to an overall better sales tax performance for the fiscal year.sharain1

As figure 1 shows, over the period FY 2005 to FY 2016, the month of the peak blossoms (March or April) was also the month of higher sales tax activity in ten out of twelve years. In 2008 and 2009 the peak month differed from the month of higher sales activity.

In fiscal years with strong collections in springtime (March and April), the total sales tax collections for the entire year were also strong.  In the years 2006 and 2009, where sales tax collections during the spring were not as strong as the previous year, total collections for the entire fiscal year also were not as strong as the previous fiscal year.

As figure 2 shows, the contribution of the cherry blossom season to the District’s sum of sales tax is clear, sales tax collections from these two months, for the period FY 2005 to FY 2016, average about 17.7 percent of total sales tax collections during the year.


What exactly is this data?

Our data on sales tax is from the Office of Revenue Analysis monthly cash collections reports. Information on Peak Bloom dates were obtained from the National Park Service also available at

Seble Tibebu and Bob Zuraski contributed to this post.

Revised data show more jobs located in DC in 2016, a slower pace of growth at year end, and a different view of recent trends

As it does each year at this time the US Bureau of Labor Statistics (BLS) revised its labor market data for all of the states and the District of Columbia based on additional information that has become available. For DC, this year’s revisions showed that at the end of last year—the December 2016 quarter—there were 2,267 (0.3%) more wage and salary jobs located in DC, but 4,052 (1.1%) fewer employed DC residents than had been previously estimated. (See below for BLS’s explanation of the basis for the revisions.)



These revisions to DC’s final quarter of 2016 seem relatively modest, but there is more to the story. The revisions over the past two years changed the pattern of growth not only for DC but for the Washington metropolitan area as well. These revisions result in a changed picture about how the recent dynamics of DC’s labor market compare to those in the metropolitan area and the US. We look at five such changes.

1. DC job growth at the end of the year was slowing down, not speeding up. The revisions increased job growth over the last half of 2015 and the first part of 2016, but reduced it in the last half of 2016. In the 2015.4 quarter , for example, job growth over the prior year was revised upward from 8,700 to 18,633—more than double. Even though 2,267 jobs were added to the 2016.4 quarter, the year ended with job growth slowing rather than speeding up.

graph 1

2.At the end of the year DC’s rate of job growth was below the US average, not above it. Previously, it appeared that an increasing rate of growth in jobs brought DC to the point where its rate of increase in jobs exceed the national rate of 1.6% in the 2016.4 quarter. The revision boosted DC’s rate of growth above the US for most of 2015 and the first half of 2016, but it slowed DC’s rate to well below the US average by the end of the year.

graph 2

graph 3


3. At the end of the year DC private sector jobs were growing at a faster rate than public sector ones, not at a slower rate. The upward revision of 2,267 jobs for the 2016.4 quarter was a net number, resulting from a 4,867 cut in the public sector and an increase of 7,133 in the private sector. The decrease in the public sector was mostly in federal government jobs (down 3,733), but local government ones were also reduced by 1,133. In the private sector there was modest increase in professional and business services (367), but most (6,767) was a 1.8% net increase in all other parts of the private sector.

The revision was enough to change the relationship of DC’s public and private sector job growth over the past two years. Previously, the rate of increase in public sector employment was shown overtaking the private sector in 2016. The revision substantially cut the growth of public sector jobs in 2016, so that they grew more slowly than private sector ones—even though growth in the private sector was slowing.


table 2



graph 4agraph 5


4. DC’s rate of private sector growth over the past two years has been similar to that in the suburbs, not significantly different.   The revisions to Washington metropolitan area wage and salary employment cut 19,033 jobs from the area total in the 2016.4 quarter, a 0.6% reduction. The net reduction in the metro area total was entirely due to a 21,300 (0.9%) reduction in suburban jobs. Most of the suburban reduction, 16,167, was in the private sector—7,133 of which was shifted to DC and 9,033 was lost to the area. The suburban private sector loss was about equally divided between business and professional services and all other private sector jobs.

A consequence of the change to metropolitan area job growth over the past two years is that the pattern of change in DC’s private sector is now seen to track that of the suburbs fairly closely. Previously, the rate of change in DC’s private sector appeared to be much weaker than in the suburbs over most of the past two years. With the revision, DC’s private sector is now shown to have grown faster over most of that time, just falling below the suburbs at the end of 2016.


table 3.PNGtable 4.PNG


graph 6graph 7

5. DC resident employment did not end the year with a sharp increase. The 4,052 (1.1%) downward revision to resident employment in the 2016.4 quarter was notable because it reversed a sharp increase which previously had been reported. This revision mostly results from cuts to the labor force (a 1.2% cut of 4,832), not higher unemployment. (Unemployment was actually reduced by 779, resulting in a 0.1 percentage point reduction in the unemployment rate.) The reduction to the labor force is consistent with slowing population growth which occurred in 2016. DC ended the year with growth rates in the labor force and resident employment similar to those of the Washington area suburbs and the US average.


table 6graph 8



graph 9.PNGgraph 10.PNG


According to BLS, momentum in DC’s labor market seems now to be slowing at a time when federal spending policies under consideration may weaken the region’s economy. Should such policies materialize, the preceding discussion underscores the difficulty of keeping current with how well DC’s labor market is responding to the new environment. Data can be revised.

What is this data?  All  data is from the US Bureau of Labor Statistics (BLS).  One set  is wage and salary employment (which is determined from surveys of employers) and the other set is labor force and unemployment statistics (determined from surveys of population). Both sets of data are for the District of Columbia, the Washington DC  metropolitan area, and the total US economy, and cover the period from the fourth quarter of 2014 (2014.4) to the fourth quarter of 2016 (2016.4). The data referred to in the text as the “previous estimate” is the data issued in January 2017 for the period up to an including December 2016.  The data referred to in the text as the “revised estimate” was issued in March (February in the case of the US data) for the period up to an including January 2017.

The BLS web site explains the basis for the labor market data revisions as follows:

Nonfarm payroll estimates for states and metropolitan areas have been revised as a result of annual benchmark processing to reflect 2016 employment counts primarily from the BLS Quarterly Census of Employment and Wages (QCEW), as well as updated seasonal adjustment factors. Not seasonally adjusted data beginning with April 2015 and seasonally adjusted data beginning with January 2012 were subject to revision.

The civilian labor force and unemployment data for states, the District of Columbia, and modeled sub-state areas were revised to incorporate updated inputs, new population controls, re-estimation of models, and adjustment to new census division and national control totals. Both not seasonally adjusted and seasonally adjusted data were subject to revision from January 2012 forward.

Data for the DC suburbs is calculated by subtracting District of Columbia estimates from those for the entire Washington metropolitan area.

The information here was presented in the District of Columbia Economic and Revenue Trends: March 2017 prepared by the DC Office of Revenue Analysis.