From 2008 peak to now: a dozen ways the District’s economy has changed

The official start of the Great Recession was December 2007; but the District’s employment continued to grow until September 2008. The years that followed included several shocks to DC’s economy, most notably the effects of the recession and federal cutbacks (shutdown and sequester). Since then, our demographics and workforce changed as we chronicled in many places in this blog.

Here are 12 notable economic trends that have shaped the District the September 2008 employment peak to January 2015 (or as close to the January 2015 date as data permit).

  1. More people, more households: population grew by 81,200 (14 percent), households by 29,000 (11 percent).
  2. More employed residents working in D.C. or elsewhere. Resident employment, including self-employed and residents who work in the suburbs—increased at a pace twice that of wage and salary jobs located in DC (15 percent v. 7 percent). We added 50,300 wage and salaried jobs and 45,500 resident employees.image004
  3. More of the earnings stay in the city. Total wages and salaries received by DC resident grew faster than wages earned in DC. (33% v 19%)—not surprising, given the strong growth in resident employment. Residents seem to be getting higher paid jobs, including ones formerly held by suburbanites who have retired or moved on.image006
  4. Not all District residents benefit from the boom. Total number of unemployed residents rose 36 percent (7,800 more)—but unemployment compensation payments received by DC residents decreased 31 percent. (This data do not explain how these two fit together, but there is a story here.)image008
  5. Personal Income of all D.C. residents combined grew 25 percent. Adjusted for inflation, this is a 14 percent growth. Incomes of households, again adjusted for inflation, grew 2.3 percent, and per capita real income did not grow at all.image010
  6. Almost all job growth was in the private sector. Federal civilian employment first grew with the stimulus, and then declined, netting additional 3,300 jobs for the period. The decline in state and local jobs (3,067) offset almost all of the federal gain. image012
  7. Employment became more diversified. The federal government and professional services together accounted for 42 percent of all jobs in September 2008. Both sectors continued to add jobs, but this amounted to only 18 percent of DC’s net job growth. By contrast, the four fastest growing industries—retail, education, health, and hospitality—accounted for 73 percent of the growth.image014
  8. The fastest growing sectors are not the highest paying. Wages and salaries in the four fast growing industries noted above accounted for only 23 percent of all DC wage growth. Despite their 18 percent share of DC’s job growth, federal government, and professional services accounted for almost half of the increase in wage and salary earnings. This is not too far from their 54 percent share of the 2008.3 total.image016
  9. DC added many more new households than new housing units. The increase in households (29,000) appears to be much greater than the increase in housing units whether one measures it by new housing permits issued by the DC government (12,483) or by the combined number of new condo units sold, permits for small projects of one to four units, and increased occupancy in larger market rate rental apartment buildings (14,372). Population increase therefore seems likely to have been accompanied by significant adjustment to and renovation of the DC housing stock beyond what is captured in the housing permit data. (The Census Bureau defines a household as an occupied housing unit.)
  10. Housing prices increased faster than incomes. The increase in value of single-family housing units in DC as measured by the Federal Housing Finance Agency outpaced the increase in DC Personal Income (37 percent v. 25 percentimage019).
  11. Occupied commercial office space increases matched the increases in jobs in DC (8 percent v 7 percent). Vacancy rate went up a little (6.2 percent to 6.9 percent) as the total inventory rose by 8 percent. image021
  12. DC government tax collections grew at rates similar to personal income. Tax collections, measured by the 12-month moving total went up by 22 percent while personal income went up by 25 percent. The increase in collections reflect policy changes as well as changes to the economy.image023

You can find more on these trends in the District’s monthly Trend Reports.

What exactly is this data? Population data are from Moody’s Analytics. Quarterly data, seasonally adjusted. Moody’s derives quarterly population estimates from annual Census Bureau population; it also provides estimates of the number of households. Labor market data are from Bureau of Labor Statistics, Seasonally adjusted. September is the average for the quarter. All measures of employment include full time and part time. Resident employment includes persons working outside of DC and self-employed or contract workers. Personal income data are from US Bureau of Economic Analysis and ORA. ORA estimates wages of DC residents based on the assumption that benefits as a percent of wages are the same for wages earned by DC residents as for wages earned by all persons working in DC. Property income is rent, dividends and interest and does not include capital gains.

What is happening to middle-wage jobs in the District?

Middle-wage jobs lost a lot of ground during the great recession, but they are making a comeback across the nation.  Yet high-wage and low-wage jobs are growing faster than the middle-wage jobs, suggesting that over time, the job market will be more polarized.

Is the job market in the District becoming more polarized?  To see this, we looked at employment and wage data by occupation for years 2005 through 2014 (BLS just released the 2014 data on March 25).  We then grouped the data by wage level (low, middle, or high).  Middle-wage occupations have median salaries that fall within the 25th and 75th percentiles of the annual wages paid in the District.  We define occupations with median wages below the 25th percentile as low-wage, and occupations with median wages above the 75th percentile as high-wage.

This grouping exercise reveals interesting things about our city. First, a low-wage job in the District pays a lot (but might not buy much): In 2014, the District’s 25th percentile of the annual wages, at $37,440, was greater than the U.S. median wage of $35,540.


Second, middle-wage jobs–the bulk of positions in professional and business service areas, health, and education–have grown through the recession, but they are no longer the engine of District’s economy.  These jobs never went away, but their growth declined with the recession, and never really boomed.  Occupations that pay low-wages—these include positions in retail, food services, grounds maintenance, and personal services—have more than recovered from the losses during the recession.  These jobs have been the biggest source of increase in employment since the official end of the recession.  Between 2011 and 2014, the District added about 21,000 low-wage jobs, easily reversing the loss of 7,500 jobs in this category during the recession.  Finally, high-paying jobs increased rapidly during the recession, first due to federal hiringand then with population growth, but the growth in these occupations have also slowed down.


The composition of middle-wage occupations could change over time.  Middle-wage occupations are changing either because these occupations are no longer needed in the District economy, or because these jobs are no longer middle-wage.  Between 2013 and 2014, middle-wage jobs declined by 9,700, partly because there were fewer accountants, property, real estate, and community association managers, fundraisers, counselors, and legal secretaries.  Another part of the decline was that some occupations typically considered to pay middle-level wages transitioned into low-level wage occupations. For example, office and administrative support workers (about 2000 of them in the District) made the permanent transition to low-wage workers around 2012.  Of course, some office and administrative support workers still get paid a middle-level wage, but if one were to look for a job in this occupation, chances are his or her salary would be less than $37,400.

The interactive graphs show the number of jobs paying middle-wages and the number of occupations that could be considered middle-wage under each major occupation category.


What exactly is this data? We used BLS Occupation Employment Statistics to construct a series of low-, middle-, and high-wage occupations.  Low-wage occupations are defined as occupations with annual median salaries below the 25th percentile of wages for the entire city in the same year.  High-wage occupations are defined as occupations with annual median salaries above the 75th percentile.  Middle-wage occupations lie in between.

12 years of change in D.C. economy, yet some residents remain excluded from employment booms

A recent blog post noted that in the five years after the Great Recession, DC resident employment increased rapidly, but unemployment went down only slowly.

The loose connection in D.C. between resident employment and unemployment is not new, but remains cause for concern. From 2009 to 2014, DC unemployment went down by only one for every increase of ten in employed residents.


In the 2002 to 2007 pre-recession period, however, the comparable impact of resident employment gain on unemployment was even less. Then the decline was only one-half of an unemployed person for every ten additional employed residents (that is, reduction in the number of unemployed DC residents, 1,041, was only 5 percent as much as the 23,652 increase in resident employment). The strong growth in labor force and resident employment relative to the population overpowers the relationship between unemployment and resident employment in the chart below:


The relationship between resident employment and unemployment in D.C. was also quite loose in the recession years. From 2007 to 2009, unemployment rose by 15,829, but there was only a decline of 1,920 in employed residents. That is, unemployment rose 8 times faster than the resident employment. In the U.S. the relation was closer to one-to-one, and in the suburbs, less than 1 to 2.image005

Situations in which an increase in employed residents is associated with a relatively small decrease in the number of unemployed are not unique to D.C., but they seem to occur where job growth occurs and the unemployment rate is already quite low.  For example, in the U.S. from 2002 to 2007, the decrease in the number of unemployed was about 10 percent of the growth in resident employment, but the U.S. was starting from a 5.6 percent unemployment rate, close to what some thought at the time was a full employment number. In the DC suburbs, over the 2002 to 2007 period, the reduction in unemployment was about 7 percent of the growth in employed residents, but the suburban unemployment rate in 2002 was 3.4 percent. In North Dakota, as the energy sector expanded from 2009 to 2014, the decline in the number of unemployed was 5 percent of the gain in resident employment, but in 2009 the unemployment rate there was 3.6 percent. These are all instances where the booming economy and a tight labor market invites new working adults to join the labor force.

In D.C.’s case, the unemployment rate in 2009 was 10.0 percent and the bump up in resident employment over the next five years did not translate into more jobs for the existing unemployed.  Even when plenty of people were unemployed in the District, they did not benefit from the economic expansion or found jobs.

The BLS labor force and Census population numbers are summary data snapshots at two points in time. These snapshots, as important as they are, do not give a very good picture of the degree to which changes may be occurring in the composition of the labor force and population that have a bearing on the connection between resident employment and unemployment. But these snapshots do suggest some significant changes have occurred in the dynamics of the DC market over the past 12 years.

This can be seen most clearly in the pre-recession period from 2002 to 2007. DC’s labor force went up by 22,611 over those years while the population increased by only 2,575. Where did the additional labor force come from? Some no doubt came from the existing population as the falling unemployment rate encouraged more people to seek and obtain work. But to get such a large increase from the existing population seems questionable when the 1,041 decrease in unemployment that occurred was such a small percentage of the growth in resident jobs. The most likely explanation is that the growth in the labor force was largely the result of migration patterns, the churn in the city’s population. With little change in the overall level of population, the net result of who moved in and who moved out could easily have produced an increase in persons who were in the labor force and employed.


Summary labor market and population statistics do not give information on who the unemployed are and how the composition of unemployment may have changed over the last 12 years, but such details are important for understanding current labor market dynamics in D.C. and when considering policy remedies. From numerous studies that have been done in D.C. and elsewhere it would be expected that a large share of the unemployed in D.C. are people lacking job skills and work experience. Beyond that, however, other matters to consider are the degree of turnover among the unemployed, how important the discouraged worker phenomenon is in the composition of measured unemployment in D.C., whether migration patterns that have increased the labor force also contribute in some way to D.C. unemployment, how long unemployed persons have lived in D.C., and whether the operation of the District’s unemployment insurance system has any particular effect on measured unemployment. It would also be useful to know more about how the relationship between resident employment and unemployment described here for DC compares with that of other cities.

What exactly is this data? Labor market data is from the Bureau of Labor Statistics. BLS develops these statistics with the assistance of models, with inputs that include population numbers from Census, unemployment insurance statistics, the American Community Survey, and the Current Population Survey administered by the Department of Labor for and in cooperation with Census. BLS revises data as more information becomes available. The data used here reflect the comprehensive revisions released on March 4 2015 to take account of population and other changes; these revisions were not just to the most recent year but to the entire data series.

Labor market data is the seasonally unadjusted quarterly average for the December quarter for the years shown. The data for the DC suburbs is calculated by subtracting amounts for the District of Columbia from the totals for the Washington Metropolitan Area.

Population numbers for the December quarter are taken from Moody’s Analytics (, which derives quarterly estimates from annual Census Bureau population numbers.

DC labor market over the last 5 years: many more jobs for residents, modest cut in unemployment

According to the US Bureau of Labor Statistics, from December 2009 to December 2014, the period of recovery from the Great Recession and subsequent expansion, the number of employed D.C. residents increased by 49,166—that is 16.1 percent. The percentage increase is quite remarkable—two-and-a-half times the rate across the US as a whole (6.4 percent) and the D.C. suburbs (6.5 percent), and even greater than in the resident employment growth in North Dakota (15.6 percent), an energy boom state, with many new jobs attracting a lot of new residents.image002What accounts for the increase in resident employment? It is not reduction in unemployment. From 2009 to 2014, D.C. resident employment grew by 49,166 while unemployment declined by only 4,776. This is a modest decline: For every 10 additional D.C. residents who got jobs since 2009, the number of unemployed residents went down by only one. The relatively loose relationship between resident employment and unemployment contrasts with the suburbs and the entire US, where for every gain of 10 in employed residents, the declines in unemployment were an average of 2.6 and 6.8, respectively.

Here is another way to look at it: Over the past five years, the District reversed only about third of the increase in unemployment caused by the recession. The suburbs reversed nearly half of it and the US reversed nearly 80 percent.

image004The most likely explanation for why resident employment rose so rapidly and unemployment fell so slowly is the growth of DC population. From 2009 to 2014, DC’s population increased by 65,400. This 10.9 percent increase in population growth did not directly cause either the 13.1 percent increase in DC’s labor force or the 16.1 percent increase in DC’s resident employment.


Some of the increase may well have come from the existing population reentering the labor force; however, these two labor market indicators could not have grown as they did in the absence of more population. The opportunities for employment in both DC and the suburbs (reachable from DC by commute) appear to have attracted additional people to the city. The March trend report coming out this week has much more on this topic.

We will write more on the relationship between population, resident employment, or unemployment in another post.

What exactly is this data? Labor market data is from the Bureau of Labor Statistics. BLS develops these statistics with the assistance of models, with inputs that include population numbers from Census, unemployment insurance statistics, the American Community Survey, and the Current Population Survey administered by the Department of Labor for and in cooperation with Census. BLS revises data as more information becomes available. The data used here reflect the comprehensive revisions released on March 4, 2015 to take account of population and other changes; these revisions were not just to the most recent year, but also to the entire data series.

Labor market data is the seasonally unadjusted quarterly average for the December quarter for the years shown. The data for the DC suburbs is calculated by subtracting amounts for the District of Columbia from the totals for the Washington Metropolitan Area.

Reverse commuters now hold higher paying jobs

We generally talk about the commuter bite—income earned in the District by non-residents, which the District cannot tax—but rarely about the District residents who work in the suburbs and pay income taxes in the District. Let us call them reverse commuters.

Reverse commuters were thought to be residents who cannot find work in the District and therefore go to the suburbs, especially Maryland, to work jobs in retail, construction, or building and grounds maintenance.   We have been seeing strong income growth among District residents, while job growth and wage growth in the District has been stagnant. This made us suspect that the nature of reverse commuters is changing.

We looked at data from the American Communities Survey and found the following:

  • A quarter of the District residents who are employed work in the suburbs. This share has been relatively stable, varying between 24 percent and 27 percent since 2000.image002
  • Personal earnings of reverse commuters have been increasing. Until 2011, reverse commuters earned less than District residents with jobs  in the District. Since 2011, reverse commuters, on average, earn more. This change would have come earlier had it not been for the federal stimulus after the great recession, and the consequent ramp up in federal hiring between 2009 and 2011.image004
  • Personal earnings of reverse commuters is higher because they now hold higher paying jobs. In 2001, only one in five reverse commuters held jobs in the professional and business services area. In 2013, this share was almost one in three. The growth is coming from higher paying jobs in professional services such as lawyers, managers, scientists, or architects and not from lower paying jobs  in business services such as buildings maintenance, travel agencies, or administrative support.image006
  • A smaller share of reverse commuters work in retail, construction, and health sector jobs in Maryland or Virginia while the share of such jobs held in the District by D.C. residents did not change since 2001. Finance sector jobs declined everywhere, but even more so in the District, suggesting that some of the reverse commuters successfully held on to such jobs in the suburbs.


What exactly is this data? We used data from American Communities Survey between 2000 and 2001, looking at the place of work and personal earnings of DC residents who are employed.  We focused on jobs in DC, MD, and VA. Some DC residents hold jobs in other states but it is not easy to make a reliable estimate about them. Their size in the sample is too small.  We also looked at the industries DC residents hold jobs in, again focusing on DC, MD and VA. Detailed breakdowns of jobs are, once again, not reliable, given the small sample size.  So we reported on broader industry areas, sticking to what we think are reliable estimates.