Tracking Liftoff: Is September Out of the Picture?

The Chinese yuan devaluation and global market volatility has increased the uncertainty surrounding the timing of liftoff.  Several important economic reports, such as employment, will be released before the next two-day Federal Open Market Committee (FOMC) meeting on September 16 and 17; and may provide more clarity on the timing of liftoff.

The graphic below allows you to track how FOMC members have adjusted their estimates of liftoff. Click on the photo of an FOMC member to see that person’s view about the timing of liftoff, how the view may have evolved since last December, and key quotes that are hyperlinked to full speeches.

The photos of voting members are shown in circles with nonvoting members in squares. Key economic indicators are presented on the right (where the data shown represent the original estimates that were available at the time of the meeting rather than more recent revisions).

Subscribe to our RSS feed to receive notifications when the graphic is updated.

As Employment Grows, When Will We See Wage Growth?

Wage growth remains relatively flat despite indicators of economic recovery. As the unemployment rate falls and employment grows, the increasingly smaller supply of workers is expected to lead to wage growth. As Loretta Mester, president of the Federal Reserve Bank of Cleveland, put it recently in the Wall Street Journal, “basic economics hasn’t gone out the window […]when employment grows, wages will start to grow.”

Over the past two years, however, most workers have not seen much wage growth. In fact, there is a somewhat weak but negative relationship between employment growth and wage growth for over 800 occupations from 2012 through 2014 (each occupation is weighted by the number of people employed in that occupation in 2014). There are some outliers, including occupations in the arts with an especially wide range of wages (such as models and makeup artists), but the majority of occupations are clumped approximately equally around low employment growth and low wage growth.

Employment and Wage Growth by Occupation

Using the Bureau of Labor Statistics’ occupation profiles and typical entry-level education requirements for each occupation, the relationship between employment growth and wage growth from 2012 through 2014 differs depending on the education typically required for an occupation.

Employment and Wage Growth by Typical Entry-Level Education for Occupation

Based on a review of the charts by education required, much of the negative relationship for all occupations is being driven by lower-skilled occupations, those that typically require a high school diploma or equivalent or less. Low-skill occupations with employment growth have seen a decline in real wages over this period, while many of the higher paying occupations have seen declining or stagnant employment. This is likely an indication that there is a surplus of available workers at the low-skill level. In fact, unemployment rates for workers without a college degree were well above the national average, as shown in the chart below from the Bureau of Labor Statistics.

Earnings and unemployment rates by educational attainment

For higher-skilled occupations such as those requiring at least a bachelor’s degree, the expected positive relationship between employment growth and wage growth is evident, indicating that labor market slack for these positions has been eliminated or nearly eliminated. Meanwhile, for middle-skill occupations (typically requiring some college or an associate’s degree) the relationship has been flat, which may suggest a tipping point in the near future as the remaining slack diminishes. Even so, the wage disparity between high and low-skilled jobs has been increasing for decades.

These charts align well with other reports indicating that employment is growing for jobs with higher wages and benefits packages, and most of these jobs are going to people with a bachelor’s degree or higher. This is good news for college graduates, but only part of the story—occupations requiring at least a bachelor’s degree made up less than a quarter of total employment in 2014, while  66% of employment in 2014 was in low-skill occupations.

Until the negative or flat relationship between wage growth and employment growth in lower- and middle-skill occupations reverses, we will likely continue to see little real wage growth in the economy at large.

Research support was provided by Patrick Clapp.

Estimating Spectators for Economic Impacts is Tricky

The UCI World Championships will soon be held in Richmond, Virginia from September 19th through the 27th. In 2011, when the city was preparing the bid for hosting the cycling championship, Chmura was asked to estimate the event’s potential economic impact. One of the most formidable challenges in that process was to estimate the number of spectators.

We estimated the number of spectators would be 450,000.  As we clarified in a previous blog, 450,000 spectators is not the same as 450,000 visitors.  Despite having thought we explained this difference, we continue to field questions like “how did that number come about?”  The purpose of this blog is to de-mystify the process of estimating this figure and to provide guidance for others who are estimating the economic impact of events in their region.

Four years ago, faced with the challenge of estimating the number of spectators for a major event, the prudent approach was to look at past, similar events.  This mirrors the typical process of any economic projection— utilize data from the past to provide valuable information that helps predict the future.  The number of spectators did not come out of a magic “black box.” Rather, the process was guided by academic research in the tourism industry.

Tourism literature consistently indicates that a region’s population base is one of the key determining factors for the number of visitors/attendants to tourism attractions such as historic sites, festivals, concerts, parks, and museums. Other key factors are the population’s interests, economic conditions such as travel costs, and the existence of a tourism infrastructure such as roads and airports. 

Since we know the population base of Richmond, its surrounding counties, and other major cities within a few hours’ drive, the missing piece is how many of the nearby residents are interested enough to attend the race.  For that information, we examined the number of spectators who attended past UCI World Championship racing events relative to the population base of the host region. 

Right away, we faced challenges. The majority of past races have been held in Europe, which has a long history of public support for cycling. This being said, the public interest in these races in Europe would be higher than can be reasonably expected in the United States, therefore using European races as examples would likely over-estimate the attendance in Richmond.  Over the past decade, the only two non-European championship races were held in 2003 in Hamilton, Canada, and in 2010 in Melbourne, Australia. The 2003 Hamilton race reported 230,000 spectators while the 2010 Melbourne race reported 300,000 spectators.

Among those two races, Australia is very far from cycling centers in Europe or North America while Hamilton has more similarity to Richmond. For that reason, we used a survey from the Hamilton race to derive our estimate. 

Hamilton is a mid-sized city (over 700,000 persons in the metro area) not too far from Toronto with a gateway international airport. Richmond is also a mid-sized city (over 1,000,000 population in the metro area) and not too far from several major U.S. cities.  Both Hamilton and Richmond are on the eastern part of the North American continent and are relatively easy to reach.

Hamilton’s survey of over 1,000 race attendants identified the spectators based on the distance of their home to Hamilton. Using that information as a proxy, we calculated the percentage of the regional population base that would travel to see the race based on their distance to Richmond.  The Hamilton survey also contained information on the number of races each visitor attended. Adjusting for the fact that the Richmond race is longer (9 days as opposed to 6 days for Hamilton), we estimated that the total spectators to the Richmond race would be 450,000.

Four years ago, we used the past events to make a future projection, just like any economist would do.  Like any economic projections, there are many unforeseeable events that can affect the actual number of spectators. For example, the global economy and exchange rates can play a role in attendance. With the European economy struggling and the high value of the dollar, European visitors may find it too expensive to travel to America.  Marketing and outreach efforts will also affect the number of spectators. Locally, traffic and parking can affect the number of spectators from the region, and even weather can play a role in attendance.

With the race just about to begin in 22 days, we look forward to measuring the true number of spectators.


How Many Visitors for Richmond 2015?

The residents of the city of Richmond and surrounding communities are eagerly awaiting the Union Cycliste Internationale (UCI) Road World Championships to be held in Richmond in September - referred to locally as “Richmond 2015”.  Likewise, cycling enthusiasts around the country are excited about the championships returning to the United States for the first time in almost 30 years.

Businesses in the region, including hotels, restaurants, shops, and taxi drivers are hoping that the influx of athletes, race officials, and visitors to Richmond 2015 will mean a boon for businesses.  According to a study completed by Chmura Economics & Analytics, “The World Championships is a 9-day event from which Richmond can expect to draw more than 450,000 on-site spectators over the course of the nine-day event from around the United States and the world.”[1]

The number “450,000” has been quoted frequently by many media outlets. While some quoted it properly as 450,000 spectators, unfortunately, there are some mischaracterizations that warrant clarification.

First, the number 450,000 refers to the total number of spectators, not visitors. Some social media and general public comments have included phrases such as “450,000 visitors are expected in Richmond,”[2] giving the impression that there will be 450,000 individuals visiting the Richmond region during the event.

However, “spectators” are defined as people watching races on-site (and it includes some double counting). According to the Richmond 2015 website, there are a dozen races over 9 days. If an individual attends multiple races, he or she would be counted multiple times.  In fact, the Chmura report assumed the average individual will watch over 4 races during the 9-day event. Obviously, visitors travelling to Richmond for the race may attend more races, while casual fans from the region may only watch a couple of them.

Another misconception by the general public is to assume that all spectators are visitors from outside the region. For example, a report from in Richmond quoted one local real estate agent as saying, “450,000 visitors are expected in Richmond. There’s not enough hotels to house those people.”[3]  If hotel operators expect that there will be 450,000 people needing lodging, they will be disappointed. That is because the 450,000 spectators include local residents in the Richmond region, as well as a large number of people living outside Richmond that will make a day trip to the region. The overnight visitors are estimated to total over 50,000. Adding athletes and their support staff, journalists, and officials, the overnight visitors are estimated to be more than 60,000.

With about 18,000 hotel rooms in the Richmond region,[4] even the estimated fifty to sixty thousand overnight visitors seem to be exceeding the region’s capacity. Is it really? The answer is no. That is because many visitors will share rooms, reducing the demand for hotel rooms. Furthermore, many visitors will not stay for all nine days. Some will come for the entire event and some will come for a few days while many may only come for a weekend.  Additionally, some overnight visitors may choose to stay with families and friends in the region or rent rooms or houses directly from area residents. In the end, the regional hotels rooms may be sufficient after all.  


[1] Source:

[2] For an example, please see

[3] For an example, please see

[4] Source:

Federal Spending Back on the Increase?

Federal spending is back on the upswing. Or, at least it is according to the President’s budget which shows an increase of 7% percent in Fiscal Year (FY) 2015 (October 2014 through September 30, 2015) compared with the prior year.[1] 

The President’s Budget has federal spending growing on average 5.4 percent a year from FY2015 through FY2020 with national defense increasing essentially 0 percent.[2] Although slower than the 6.6 percent annual average growth from FY2000 through the peak in FY2011, the return to some growth in spending is good news for states and metropolitan statistical areas (MSAs) that are dependent on federal contract spending.

We’re not out of the woods yet. A January 2015 Congressional Budget Office report indicates that defense and nondefense funding are equal to or below the FY 2015 budget caps,[3] but the President’s Budget for FY 2016  ignores the caps put in place by the Budget Control Act of 2011 and modified by the Bipartisan Budget Act of 2013.  If the caps are exceeded, a sequestration will reduce federal discretionary spending by approximately $139 billion in FY2016 below the President’s Budget request.[4] About 64 percent of the reduction will occur in defense spending based on current laws.

So which metro area economies are most dependent on revenues derived from contract work for the federal government? Which ones are most at-risk if we see another round of sequestration? Chmura Economics & Analytics took a look at federal contract spending data and assigned it a metro geography based on where the awarded firm performed the work and adjusted it for time of performance since some contracts are awarded for work that is performed over a number of years.

Out of 381 MSAs in the country, the Washington-Arlington-Alexandria, DC-VA-MD-WV MSA topped the list of federal spending with $71 billion in FY 2014. Dallas-Fort Worth-Arlington, TX was a far second with $19.3 billion, and Los Angeles-Long Beach-Anaheim, CA was third at $14.8 billion in FY 2014.

A better way to assess the risk of a region to potential cuts in federal spending is to consider the concentration relative to employment. From that perspective, Idaho Falls, ID ranked the highest ($47,691 per employee); followed by California-Lexington Park, Maryland ($39,940); Amarillo, Texas ($31,407); and Huntsville, Alabama ($30,799).

The interactive map and table below show the dependence of all MSAs in the nation on federal spending.

Research support was provided by Patrick Clapp.


[1] President’s Budget FY 2016, Table S-1

[2] President’s Budget FY 2016, Table 28-1 Net Outlays by Function, Category, and Program


[4] President’s Budget FY 2016, Table 5.6—Budget Authority for Discretionary Programs: 1976–2020