Department of Labor Logo United States Department of Labor
Dot gov

The .gov means it's official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you're on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Article
June 2024

Two hours to the office, two minutes to the kitchen table: trends in local public-transportation expenditures from 2018 to 2021

The COVID-19 pandemic altered consumer spending on public transportation, including both intracity (mass transit, taxi, and limousine) and intercity (air, ship, bus, and train). According to the Consumer Expenditure Surveys, spending on public and other transportation fell 66.3 percent in 2020. With shifts toward virtual work and school attendance, many commuters’ 2 hours to the office became 2 minutes to the kitchen table. This article examines changes in dollars spent for public transportation, mainly intracity mass-transit spending from January 2018 to December 2021, among various unique demographic groups: urban and rural residency, occupation type, educational attainment, and selected metropolitan statistical areas. Essential workers generally were unable to telework. Simultaneously, private-transportation costs declined in 2020 for all, while affordability rose for many. Gasoline prices fell 46 cents (17 percent) per gallon in 2020, according to the Consumer Price Index. Furthermore, average annual income fell for certain occupational groups and rose for others (income rose 2 percent for all consumer units). A model of indifference curves and budget constraints is used to show which members of specific education groups are likely to substitute private for public transportation. Results in 2020 show that income growth and lower private-transportation costs compared with those of public transportation resulted in increased private-transportation spending of up to 5 percent, with accompanying public-transportation spending reductions of 40 to 50 percent. Although intracity mass-transit spending rebounded in 2021, it did not reach prepandemic levels.

Position yourself on the corner of 7th Avenue and 32nd Street in Midtown Manhattan on February 20, 2020, gazing on Amtrak’s Pennsylvania Station (known to locals simply as Penn Station), when COVID-19 was an “elsewhere” event, and the thoughts of a crippling virus “happening here” was not on the radar. What would you see? Most likely, you would see nearly 600,000 commuters flooding out of the catacombs from Long Island, New Jersey, and New York City’s outer boroughs, either disembarking from intracity commuter trains and subways, transferring to buses or taxi cabs, or driving by the intersection in their personal cars to reach their cubicles or corner office desks. In just 1 month, the COVID-19 pandemic not only broke out but also accelerated, prompting a mandatory stay-at-home order from the Governor of New York on March 20th.1 The order forced private firms to adopt mandatory work-from-home arrangements. Penn Station became considerably more deserted, with few commuters.

Two hours by rail to the east of New York City sits the small town of Mattituck, NY. Served by one morning and one evening rush-hour train, with a platform barely long enough for one railcar, the town sees roughly 20 riders a day for the over 2-hour trip to the city. The other few thousand Mattituck residents either own local businesses on Love Lane (Mattituck’s main street), are employed in service sector positions in surrounding towns, or work the farms and vineyards dotting Long Island’s North Fork. Then on March 20, 2020, stay-at-home orders constrained local businesses. Of those 20 daily commuters, most stayed home. The vineyard workers, cooks, farmers, and the like, however, still drove to their jobs. Between 2019 and 2020, transportation expenditures fell 8.5 percent on the aggregate, but this story runs far deeper than this first townscape view.

A July 2021 release of the U.S. Bureau of Labor Statistics (BLS) American Time Use Survey (ATUS) found that the average round-trip commuting time fell from 72 to 47 minutes between May 2019 and December 2020, a decline of 36.1 percent.2 The share of individuals who commuted fell by 17 percentage points from 84 to 67 percent in 2020, with many of those taking the 2-minute venture to the kitchen table, laptop in hand. This downward trend in commuting was present across virtually all demographic groups. Because of these declines in commuting time, the effects of the COVID-19 pandemic created a downward bias on spending for public transportation, organizations held business trips virtually, and families put vacations on hold. According to data from the U.S. Travel Association, 2020 domestic business travel was just 32 percent of prepandemic levels of 2019. International business travel eroded more steeply, declining to just 22 percent of its 2019 value.3 As shown in chart 1, airfare spending declined almost 69 percent from 2019 levels, while intercity rail and bus expenditures declined by 74 percent and 85 percent, respectively. Changes in commuting patterns were reflected in the transportation estimates of the Consumer Expenditure Surveys (CE) and were observed across a range of demographic groups.

Urban consumer units (CUs) rely more on public transportation to commute to work compared with those in rural areas who have restricted choices in using public transportation.4 In certain rural areas, public transit may be in limited supply or not an option compared with urban areas where more choices exist between using a car and riding public transportation. Focusing on whether a CU resides in an urban or rural area as a demographic of interest can help determine how much an individual’s area of classification affects transportation consumption.

In addition to the differences by area, intracity mass-transit spending also varied by occupation. Retail and service workers were deemed “essential workers,” while managers and professionals were more likely permitted to telework.5 Did this trend exist before the pandemic or did a new phenomenon occur in which workers in entry-level jobs became more reliant on local public transportation? Highlighting spending by occupation of reference person allows a view into how strong the tie is between a job position and reliance on local public transportation. A similar analysis applies to education: how much could educational attainment have swayed a given CU’s transportation consumption, and could positive income shocks have swayed a particular group?

The BLS CE program’s metropolitan statistical area (MSA) tables break down data by selected metropolitan area according to the four regions of the country.6 This detailed breakdown allows data users to get a microlevel view of how CUs in various cities of similar transportation infrastructure behaved and to determine if their behavior, based on workforce characteristics and other key variables, was homogenous in reducing spending.

After nearly a year of the COVID-19 pandemic, had the tides begun to turn in 2021? Were commuters returning to the office in observable numbers? Had domestic and international travel demand picked back up? Did prepandemic-era transportation spending truly rebound in 2021? This article highlights trends of local public-transportation spending for various demographic groups, such as place of residence and education, and rationalizes these trends at a more granular level. The article presents specific characteristics often not highlighted in CE data, such as area type, occupation, and highest level of education tables. In addition, internal microdata for annual MSA estimates are used. Finally, this article arrives at the 2021 “terminal” to determine how consumers have adjusted, or not, to the ever-changing pandemic landscape. Thank you for boarding this journey into the local transportation expenditure trends of the 4 years used for this article; stand clear of the closing doors.

Data, methods, and sample exclusions

CE detailed tabular data were extensively researched to capture a diverse sample of consumers expenditure patterns on local public transportation. Data on area type (i.e., urban or rural residence), educational attainment, and occupation are derived from the CE data tables of detailed-level integrated calendar-year means. These detailed-level tables provide robust and extensive expenditure data that the published calendar year or MSA tables do not offer.7 These expenditure data include the primary focus of this article: intracity mass-transit expenditures. Integrated tables pull data from both components of the CE.8 That is, they pull from the quarterly Interview Survey, which focuses on more costly items, such as a monthly MetroCard for New York City. Integrated tables also pull from the Diary Survey, which is designed to capture small-ticket items and more frequent purchases over a 2-week period, such as specific food items purchased from a grocery store (e.g., bread or rice). The data for the MSA tables provide annual averages that span a 2-year-collection period, thus tables are labeled, for example, as “2019–20” to denote the full scope of the sample gathered. One can examine commuting patterns down to the metropolitan-area level with these tables, which provide a level of context with a narrower focus on public-transportation spending by selected MSA. Public transportation itself is critical to the efficient flow of labor into and out of cities. The 2-year collection period is a constraint because it is difficult to isolate the factors that influence annual spending trends in either year of the 2-year window. This article addresses this issue by using special tabulations developed with BLS CE internal microdata to perform research with 1-year estimates. These tabulations were used as a part of this research effort for selected MSAs. This approach allowed for a more thorough and meaningful analysis. In addition, several outside sources were used to provide relevant data regarding employment distribution by education type, commuting length, workforce dynamics, population flows, and other critical information to support the analysis. Lastly, all reported expenditures analyzed strictly capture out-of-pocket CU spending and do not reflect business-related expenses or transit-related reimbursements.

This analysis is completed in a “descriptive” format. A collection of charts is generated from expenditure data from the integrated detailed-level CE tables.9 Expenditure values are then converted into a percent-change format to represent shifts in spending from the preceding year.

Next, microeconomic concepts and theory are applied to draw conclusions on the unique motivations behind intracity mass-transit spending decisions among various demographic and geographic groups. This article revisits one of the core tenants of consumer theory in microeconomics: the concept of achieving utility-maximizing consumption bundles is shown with an indifference-curve and budget-constraint map. The analysis in the article shows how both a change in consumer income and a change in the price of one of two modes of available transportation will affect spending through a breakdown of the income and substitution effects.

The title and framework of this article revolves around the phrase “local public-transportation” expenditures. This phrase refers to the CE expenditure category “intracity mass transit,” which records spending on local commuter rail, intracity bus usage, subway ridership, cross-river ferry fares, and so forth. Among the specific demographic groups, intracity mass-transit expenditures are compared for all analyses by using data from the detailed-level integrated tables. MSA tables are constrained in that they do not have the same robust detail and depth as that of their detailed counterparts. Thus, as stated previously, BLS CE internal microdata were used instead as a substitute for MSA tables.

Data were endogenously adjusted and excluded to streamline the analysis and minimize “white noise” to the most feasible degree when analyzing certain demographic groups. For example, school-bus fares were excluded when analyzing the components of “public and other transportation.” On average, a fraction of a dollar is spent on school-bus fares annually, and they comprises a minimal portion of the broader metric in the aggregate totals. Most states provide free transportation via school buses, apart from a small number of counties that charge fixed annual fees. Even if the amount or structure of school-bus fees changed, the amount would be so small that it would not meaningfully affect public and other transportation expenditures. Therefore, school-bus fees were not included in this analysis. (See the appendix of this article for a complete list of components of public and other transportation that are included or excluded in the analysis.)

For those expenditures analyzed by occupation, CUs in which the reference person was identified as retired or self-employed were excluded. Those CUs who are retired are defined as not actively working and thus engage in public-transportation spending only for leisure consumption, for local needs, or on out-of-town trips for travel and vacation. Thus, this group is excluded from the analysis.

Self-employed people have their own set of constraints. Unlike standard salaried workers, self-employed people do not have the same certainty of pay structure. While CUs in every other occupation category are paid a fixed salary or salary plus commission, self-employed workers might take their pay from their clients or from other means. This group makes up one of the smaller sample sizes, although larger than construction workers and mechanics and operators, fabricators, and laborers. A small sample, in general, can lead to variability in data year over year, but this reason alone would not be a compelling reason to exclude self-employed workers from the analysis. In addition to the issue of a small sample size, the self-employed would not be entitled to fringe benefits, such as transit subsidies or other programs that encourage the use of public transit. Thus, both retirees and self-employed workers did not board the train to its final destination of analysis.

Initially, eight metropolitan regions were selected for the MSA analysis: New York, Boston, Philadelphia, Chicago, Atlanta, Dallas, San Francisco, and Washington, DC. The goal was to include multiple MSAs in each region that had relative homogeneity in intracity mass-transit layout, infrastructure, and ridership. Data concerns arose with certain MSAs when an extensive analysis of the expenditure patterns of these eight metro regions was conducted. A few MSAs had an insufficient sample size, sometimes with a sample size as low as 15–20 CUs. This quantity is well below any conventional lower bound for an acceptable sample size to obtain meaningful results and can lead to extreme variability and volatility in the data for metropolitan regions from year to year.10 Therefore, the scope of this analysis was limited to four MSAs: New York, Chicago, San Francisco, and Washington, DC. These MSAs have a robust network of subways, local buses, and intracity rail for those commuting for work or leisure. The most used transportation agencies of these MSAs are the following: New York’s Metropolitan Transit Authority (MTA); the Washington Metro Area Transit Authority (WMATA) in Washington, DC; the Chicago Transit Authority (CTA) of its namesake Chicago; and San Francisco’s Bay Area Rapid Transit (BART) system, along with the separate San Francisco Municipal Railway (Muni), which manages the subway and bus networks.

As mentioned previously, the goal was to select the most used transportation agencies available within the sample, but a couple of perceived constraints of the CE MSA data prevented the use of four of the eight MSAs. Ridership figures are not comparable relative to the population of the metro area. For example, as of 2019 (the last year of data in the pre-COVID-19 era), subway trips per day averaged 5.5 million for the MTA, 626,000 for WMATA, 600,000 for the CTA, and just over 400,000 for the BART. When adjusted for population and usage, these four metro regions and transit agencies were the most comparable with what was available in the constraints of CE geographic data. This adjusted analysis also narrows the focus to one MSA for each region of the country as defined by the CE. The analysis does not explicitly compare each MSA region with the other three regions. However, the analysis does draw conclusions at the most narrow and detailed level possible, relative to making sweeping analyses by region.

A focus on dynamic adjustments in expenditure shares and annual measures

This section of the article examines public and other transportation, a broader category within the transportation measure of the CE. After this section, the focus shifts back to intracity mass transit. Chart 2 depicts the relative shares for selected expenditure subcategories that make up public and other transportation. Intracity mass transit is housed in this broader station (public and other transportation) and accounts for about 10 to 14 percent of the total expenditure for public and other transportation. Although intracity mass transit accounts for about 2 to 5 times the spending of other expenditures under the public and other transportation category, as of 2021, it remains almost one-sixth of airline fares. Airlines fares take up at least a 60-percent share of public and other transportation, and in 2021, when CUs flocked back to the airlines, the share reached 71 percent.

When compared with transportation in the aggregate, public and other transportation only accounts for a small quantity of total transportation spending, never surpassing 9.0 percent and falling to just 2.7 percent in 2020, as shown in chart 3. The data for 2021 show that public and other transportation makes up just over 4 percent of total transportation spending, 3 percent lower than prepandemic levels. Intracity mass transit was less than 1 percent of total transportation spending from 2018 to 2021. Transportation demand was slow to rebound, and the share of intracity mass transit represented by public and other transportation remained flat in 2021. These two categories (public and other transportation and intracity mass transit) fell by 66.3 percent and 53.7 percent, respectively, in 2020, as shown in chart 1. This 2020 decline follows a nearly 5-percent reduction in public and other transportation in 2019, when intracity mass transit saw a modest 1.1-percent gain. Now that a frame of reference as to the current trends in broad-based total transportation spending has been established, there will be an examination of the various demographic groups to explore in detail how intracity mass-transit spending shifted over the 4 years from 2018 to 2021.

From urban to rural residency, spending by area type

Traveling back to 2020 to the Mattituck, NY, train station, suppose that most of the 20 daily commuters traded in their monthly train tickets, reducing ridership to a couple of passengers a day. Chart 4 presents the percent change in intracity mass-transit expenditures by area type from 2019 to 2020. Intracity mass-transit expenditures for rural CUs plummeted by nearly 63 percent from 2019 to 2020. Other urban CUs followed in a similar way, while those in central-city areas reduced spending by almost 48.9 percent.11 One variable that can explain the differential between those residing in central-city areas and those in more rural regions is the substitutability between personal automobiles and intracity mass transit. As of 2020, 95 percent of rural CUs reported owning or leasing at least one automobile. In contrast, only 84 percent of CUs in designated central-city areas reported auto ownership or having a lease payment on a car or truck. CUs may purchase automobiles that are not exclusively used by the CU. For example, automobiles may be purchased for a child who lives at a different residence, such as a child who is away at college, or an urban CU may purchase a car that is kept at a relative’s home or at a second home in the suburbs because of the expenses and limitations associated with city parking. When intracity mass-transit service was reduced as the COVID-19 pandemic began, those living in rural areas had the ability to drive their personal automobile to work in lieu of taking the train, bus, subway, or combination of the three. Their central-city counterparts did not have a choice if their jobs required in-person work. Residual demand for intracity mass transit was present for those CUs in central-city areas without cars who still had to complete basic tasks such as grocery shopping. Although not commuting related, residual demand for intracity mass transit ties into the lack of substitutability these CUs face. In the aggregate, this lack of choice led to a 15-percent lower reduction in intracity mass-transit demand in central cities compared with that of rural and other urban areas.

Another potential explanation for the larger shift in spending for rural and other urban CUs could be a shift in the proximity to their workplace. With the prospect of having to return to the office in the short or long run, more employees in rural or other urban areas than in central-city regions could have shopped around in the labor market for positions closer to home or that permitted full-time continuous telework. Firms might offer full-time telework positions but with reduced wages relative to the office-facing role of an identical position. This scenario presents a question of opportunity cost for those employees. Does the prospect of spending less on intracity mass transit, having more family time, and reducing other expenditures outweigh the reduction in wages? If this is the case, then demand for intracity transit would be reduced considerably, and the 2-hour commute to the city would be replaced with a 2-minute commute to the dining room. Both components clarify the differential between the near two-thirds reduction in spending by those outside the city and the near 50-percent reduction by central-city residents.

Intracity mass-transit spending the following year in 2021 saw heterogeneous results when broken down by area type. In 2021, the general trend that developed was a positive relationship between population density of area type and the resurgence of intracity mass transit. Rural CUs began substituting private transportation for public transportation; this finding is reaffirmed by the fact that 94 percent of CUs reported owning at least one vehicle, according to the latest detailed-level table by area type.12 These results are in accordance with the substitution theory discussed above. Thus, spending fell a further 20.1 percent from 2020 levels. Other urban CUs began to resume commuting to the office at a slow rate as 2021 progressed, a modest gain in spending of 5.8 percent, whereas those in central cities returned more quickly, at a 22.3-percent rate. (See chart 4.) Detailed-level tables for 2020 and 2021 were used to find that growth in the number of CUs in other urban areas exceeded 4.5 percent, mainly drawing from rural areas.13 As selected firms shifted toward hybrid work arrangements or away from full-time telework, consumers may have sought to minimize commute time to maximize leisure time when not working. This reason may have prompted them to move toward the city and out of rural areas, leaving rural locations for those who do not rely on intracity mass transit or who drive locally to work. The number of CUs in central cities grew as well, but by less than 1 percent.14 The data for central-city CUs might be capturing more than just the number who commute to work. For these CUs, taking the bus or subway for errands or other local tasks might be common, especially among the 17 percent without cars as of 2021.15

From construction workers to managers, analyzing intracity mass-transit spending by occupation of reference person

This section breaks down intracity mass-transit spending by occupation of reference person. The occupation categories analyzed include the following: managers and professionals; technical, sales, and clerical workers; service workers; construction workers; and operators, fabricators, and laborers.

Managers and professionals

For the top earners in the firm, the demand for public transportation decreased well before the onset of the COVID-19 pandemic. In 2019, a modest decline in intracity mass-transit expenditures of 7 percent was seen from the previous year, potential evidence that hybrid telework policies were implemented. The following year, managers and professionals reduced mass-transit spending by 64.9 percent in 2020, as shown in chart 5.

The 2021 ATUS release of 2019 and 2020 data records that 22 percent of the domestic labor force was already engaged in some telework in 2019. Those with jobs classified as management, business, and financial operations and as professional and related (the two titles that best capture managers and professionals) had telework rates in 2019 that were 34.4 and 34.6 percent, respectively. The following year, the rates rose to 68.1 and 59.9 percent, respectively.16 CUs with reference persons who occupations is managerial and professional had the second-highest average age of the five occupation categories analyzed. Only CUs whose reference persons are fabricators and laborers, on average, were older.17 The manager and professional class also tends to skew toward the upper end of the age distribution, placing its members in the highest risk category for COVID-19 complications according to the Centers for Disease Control and Prevention. Public-transportation expenditures most likely declined as people chose not to return to crowded subways, trains, and buses once their cities and municipalities lifted restrictions on intracity mass transit. This occupation category could have also substituted taxi and limousine services in place of intracity mass transit in their commute to work. With either of these modes of transportation, the group’s worries of crowds would have been alleviated, and protective glass in many taxi and limousine vehicles would have ensured separation between the driver and passenger. However, this theoretical substitution was not proven in the data, potentially because the crumbling of taxi demand for leisure travel. Between 2019 and 2020, taxi and limousine spending declined 41 percent, further corroborating the broad-based telework shift.18

If a spending rebound of any noticeable quantity were to have occurred from 2020 to 2021, then a rollback of telework policies and a return to in-person work and commuting could have facilitated this since higher transportation demand means higher spending. However, as shown in the 2021 ATUS from June 2022, 59.0 percent of workers in the management, business, and financial operations and 56.5 percent in the professional and related fields still reported working at home on an average workday.19 ATUS explicitly states that this measure includes both hybrid and full-time telework arrangements or any setup in which an individual worked at home at least a portion of the workweek. The reluctance to end telework and hybrid work arrangements is corroborated by Michael Dalton and Jeffrey A. Groen in their 2022 Monthly Labor Review article, “Telework during the COVID-19 pandemic: estimates using the 2021 Business Response Survey.”20 Dalton and Groen find that the professional business services and financial activities industries had the highest concentrations of telework and hybrid work arrangements aside from information systems, which fits more closely into technical, sales, and clerical. This reluctance to transition back to full-time, in-person work arrangements could have resulted in a downward bias in spending returning to prepandemic levels. But intracity mass-transit expenditures for managers and professionals in 2021 increased 26.2 percent, as shown in chart 6.

Technical, sales, and clerical and service workers

Technical, sales, and clerical workers, along with service workers, underwent spending changes in intracity mass-transit expenditure consistent with forced lockdown measures in the second quarter of 2020. The former occupation category (technical, sales, and clerical workers) experienced a 58-percent decline in intracity mass-transit spending after an increase of 20 percent from 2018 to 2019, as described in chart 5. Those in the service sector reduced their spending by 36 percent, following a 15-percent increase in spending over the same period from 2018 to 2019. The job activities done by the former occupation category are more conducive to the telework transition, and they are not classified as essential workers in the same way as those in the service industry.21 Grocery stores, big-box retail outlets, and restaurants, whether for takeout or sit-down dining, remained open during the pandemic, all of which can be classified as service-worker-intensive institutions. These firms experienced layoffs as demand fell, but those workers who remained employed were relied on for in-person labor hours. This result led the service workers’ demand curve for intracity mass transit to contract, but it contracted less than the curve for transit for those workers in technical, sales, and clerical positions.

A further negative shift in the demand curve of intracity mass transit for technical, sales, and clerical workers might be attributed to the breakdown of the occupation by sex. Of the five occupation categories (managers and professionals; technical, sales, and clerical workers; service workers; construction workers; and operators, fabricators, and laborers) highlighted in this analysis, the technical, sales, and clerical group has the largest sample of females compared with males, 56 to 44 percent, a 12-percentage-point differential, according to the integrated detailed-level tables. Lockdown measures kept children out of school, forcing at least one parent to remain at home to care for the children, often the mother of the household. Research by Enrica Croda and Shoshana Grossbard, published in the Review of Economics of the Household, shows that mothers took on a higher childcare burden compared with fathers.22 In 2020, Croda and Grossbard found that 48 percent of mothers reported being solely responsible for childcare responsibilities compared with 14 percent of fathers. This finding held true for working mothers as well: approximately 33 percent of working mothers reported taking on full childcare responsibility compared with 11 percent of working fathers. The high cost of childcare could also prevent both parents from returning to work.23 A 2020 study from the U.S. Chamber of Commerce Foundation found that 58 percent of working parents left their current jobs because of insufficient childcare solutions to meet their current needs.24 This situation might also lead to reduced hours worked, in which an individual might commute part time or engage in some alternative arrangement to meet household needs, reducing transportation demand. High childcare cost and insufficient solutions might partially explain the slower transit spending growth of technical, sales, and clerical workers of 47.6 percent compared with the growth of other occupations, as seen in chart 6. Service workers, on the other hand, spent less on intracity mass transit by 27.2 percent in 2021, the only group to do so. (See chart 6.) Coincidently, service workers were the only occupation category to observe a decline in percent reporting.25 In addition, they were the only group to report growth in car ownership; all other occupation categories reported declines in car ownership.26 Service workers could have substituted public transportation with private transportation and car ownership in 2021.

Construction workers

Trends of intracity mass-transit expenditures for construction workers underwent the largest decline in 2019. (See chart 5.) Intracity mass-transit spending for those workers building your next house or building the next office complex declined 76 percent from 2018 to 2019 and 56 percent from 2019 to 2020. Annual expenditures fell from $83 to just below $9––an 89-percent decline in spending over just a 2-year period. Operators, fabricators, and laborers, or those biased toward the manufacturing sector, followed a similar trend with annual percent declines of 33 and 45 percent, respectively. According to a 2021 Pew Research study, U.S. suburbs underwent a net gain in population of 2.2 million individuals in 2020, while cities observed a decline of 2.0 million individuals, displaying a clear shift in preference of area type.27 This finding follows a trend of “exodus” from the city to the suburbs that stretches back to 2017. According to Macrotrends, cities such as San Francisco, CA, and Buffalo, NY, lost 0.33 percent and 2 percent, respectively, of their populations in that time frame.28 This population outflow would raise construction demand in rural or other urban areas compared with that in cities.

To reduce spending on intracity mass transit, construction workers and contractors might have chosen to use private transportation, such as company vehicles, to travel from house to house or to work at an office-park complex, where bus, train, and subway options are less plentiful and efficient. Yet, in 2021, this theory was refuted by available data. Intracity mass-transit expenditures increased by 372 percent to an average annual level of $42, double the prepandemic value in 2019 but half of the high watermark of 2018.29 One potential explanation of this change might be that new housing construction rose. A greater number of housing starts raises demand for construction workers; this demand for construction workers then increases their use of intracity mass transit for work travel. Beginning in late 2020 and continuing throughout 2021, new housing construction grew for both single and multifamily residencies compared with construction growth before the pandemic. Fluctuations in the time series were present, as is the case with any volatile dataset. The overall trend after the initial phase of the pandemic showed an increase from 50,000 to over 300,000 additional starts a month compared with prepandemic highs.30 Operators, fabricators, and laborers followed the trend of construction workers, but, in 2021, operators, fabricators, and laborer increased their spending on transportation by 235.1 percent. (See chart 6.) Because of the many job titles housed under the general umbrella of operators, fabricators, and laborers, from bus and machine operators to apparel stitchers, identifying movement from any unique group is difficult.

Expenditures across various education levels

Unsurprisingly, intracity mass-transit spending fell in the aggregate for all levels of education in 2020, but the trend line from 2018 into the full year of 2019 was less predictable. As shown in chart 7, only one of the six education cohorts reduced spending from 2018 to 2019: those CUs with master’s, professional, or doctoral degrees. This finding could partially be explained by people moving to the suburbs, as cited in the 2021 Pew Research study mentioned earlier.31 Demand for intracity mass transit appeared to have already been weakening before the COVID-19 pandemic began. If a constant marginal cost for an intracity mass-transit fare is assumed, which is relatively “sticky,” then any reduction in demand, in this case, because of changing tastes and preferences (such as a change in workplace structure), will cause a corresponding reduction in quantity of trips and expenditures. “Sticky,” in this case, describes price changes to intracity mass-transit fares not set by supply and demand market forces. Prices, or fares, in the transit world are generally set by a governing board of the transit agency and can remain the same for years. The remaining five cohorts saw expenditures rise from 3 percent for high school graduates to 67 percent for those holding associate’s degrees. (See chart 7.) The estimates in this section are partitioned based on the highest level of education obtained by any member of the household.

Associate’s degree holders

By 2020, those with a completed 4-year degree in higher education (bachelor’s and above) spent over 60 percent less, on average, than the prior year for transit. (See chart 7.) According to data from the Education Data Initiative, as of 2020, at least 82 percent of those holding a bachelor’s degree or higher were employed in either professional or management positions.32 These positions adapted to extended telework policies after the initial COVID-19 lockdowns, compounding on declining demand from 2019. People holding associate’s degrees were dispersed widely but clustered in office, sales, service, management, and professional positions, leading to falling expenditures but at a heterogeneous rate. According to 2012–16 data from a 2021 Pew Research study, over one-fourth of associate’s degree holders are concentrated more in urban settings, keeping mass-transit reliance more stable.33

High school graduates

High school graduates had similar job clustering, albeit skewed slightly more toward the service sector, but high school graduates reduced their transportation spending at twice the rate of those holding an associate’s degree. (See chart 7.) In contrast to associate’s degree holders, roughly 4 in 10 high school graduates reside in rural areas, while just over 2 in 10 reside in urban areas.34 The two additional education cohorts without a completed college degree of any type (those who are classified as “less than a high school graduate” or “high school graduate with some college”) experienced similar spending reductions to those with completed high school diplomas. In 2020, the average pretax income for a CU, whose highest level of educational attainment was a high school graduate, was $44,994. This household’s income growth continued a steady upward trend from $42,599 and $41,259 over the previous 2 years, bolstering their budget constraint.35 High school graduates can be seen as the education cohort most on the intensive margin between public-transportation and private-transportation use. A positive income shock for high school graduates could affect their behavior between spending on public versus private transportation more than other education cohorts. A reduction in the price of gasoline coupled with a positive income shock could show a change in behavior. Gasoline prices trended downward from 2018 to 2020. According to the CPI Database for Average Price Data, U.S. City Average, the national average for a gallon of gasoline fell 46 cents, from $2.70 in 2019 to $2.24 in 2020.36 All other things being equal or held constant, what would the impact be on the consumption of public transit relative to automobiles?

Indifference-curve analysis and application

Thus far, tabular expenditure data and supplemental sources have been used to analyze consumer spending on intracity mass transit across various demographics. Consumers care about changes in income and prices because both affect spending behavior. A budget-constraint and indifference-curve model can be used to understand the cumulative effects of income and price changes on consumer spending. Budget constraints show the possible combinations of two goods, x and y, that a given CU can purchase for a set dollar amount. For example, if x costs $1, y costs $2, and the CU has $10 to spend, then the CU can purchase 10 units of x, and none of y; no units of x and 5 of y; or some combination that costs $10 (e.g., 2 units of x and 4 units of y).

An indifference curve represents a combination of goods or a connecting set of consumption bundles that yields a constant level of utility (often described as “satisfaction”) for the CU. The indifference curves in this application are convex in shape, implying the conventional assumption of a diminishing marginal rate of substitution (MRS). The MRS is the slope of the indifference curve at any given point. It equates to a CU’s willingness to substitute one unit of x for another unit of y while keeping the same level of utility. Real-world expenditure data for goods x (gasoline) and y (public transportation), average gasoline prices, and the percentage increase in income, which functions as the income shock to the budget constraint, are used to show the theory in action.

While income and prices play a key role in the decision making of any CU, those two forces do not operate in a vacuum; additional factors can influence consumer behavior as well. Specifically, the balance of supply and demand in both the intracity mass-transit and gasoline markets can affect transportation decisions by CUs. As mentioned previously, demand for intracity mass transit and gasoline was negatively impacted by the COVID-19 pandemic in 2020. CUs, regardless of education group, used fewer gallons of gasoline in 2020 than in 2019 and took fewer trips on any mode of intracity mass transit. Tastes and preferences were altered by changing personal health assessments and risk. In addition, new travel needs certainly had an impact on utility maximization. On the supply side, while oil companies did not reduce oil extraction and gasoline refining, mass-transit agencies did reduce service in 2020 in response to falling demand and ridership because of lockdowns and maximum telework arrangements. It has been shown above that other factors are clearly at play as possible explanations for the changing consumption of mass transit and gasoline by CUs. Nevertheless, income and prices are the most easily modeled using CE and CPI data to explain how an average CU maximized its utility in this two-good scenario.

Figure 1 depicts a standard indifference-curve and budget-constraint map in which the price of gasoline is the x good and intracity mass transit is the y good. The diagram shows two distinct events: an increase in income and a reduction in the price of gasoline while the price of public transportation is fixed. The budget constraint is derived from summing the average annual expenditures of gasoline and intracity mass-transit expenditures for the hypothetical “average” CU; this is a CU in which graduating high school was their greatest level of educational attainment. Thus, the budget constraint sums to $1,621 ($1,558, the average expenditure for gasoline for this group, plus $63, the average for intracity mass-transit expenditures).37

An increase in income and reduction in the price of gasoline on high school graduates’ intracity mass transit consumption using true expenditure data 2019–20

The two expenditures ($1,558 and $63) create the point of tangency of indifference curve I1 to budget constraint BC1 at point Ei. The increase in income, which was 5.6 percent in 2020, shifts out the budget constraint from BC1 to BC2. Thus, the value of BC2 is $1,712. This value represents the amount a given member of this cohort would allocate to gasoline and public transportation if the increase in allocation in 2020 was proportional to the rise in income from 2020 (e.g., $1,712 is 5.6 percent larger than $1,621). The mean CU in this group will be able to consume a higher quantity of both goods with the higher income, affording the ability to move to a higher indifference curve from I1 to I2, and purchase the utility maximizing bundle of goods—that is, the bundle of goods that would bring the consumer the most satisfaction, given the budget constraint. That bundle is the combination of gasoline and public transportation at the new tangency point of the budget constraint and indifference curve.

The reduction of the price of gasoline, the x good, pivots the budget constraint. This pivoting of the budget constraint graphically represents the ability of consumers to purchase additional gasoline at the lower price, while holding the price of intracity mass transit constant. In general, one can safely assume that fares in major cities remained fixed. For example, the MTA did not raise their fares between 2015 and 2021; fares were raised in 2023, outside of the scope of this article.38 According to CBS news, the CTA raised its base rate in 2018 for the first time in 9 years; even with reductions in demand from COVID-19, fares would not have changed.39 The pivot is displayed in figure 1 by BC3 emanating from its intersection with the y-axis, which shows the quantity of intracity mass-transit trips that high school graduates could purchase if they allotted their full budget that way. The budgeted amount itself did not change, but the quantity of gasoline that can be purchased did.

This reduction in the price of gasoline results in two distinct effects: a substitution effect and an income effect. The substitution effect is the amount of increased consumption of good x that would result from a decline in the price of good x, while the price of good y is unchanged, if the consumer were to keep utility constant. Implicit in this computation is that the budget changes exactly enough to maintain this level of utility because either the consumer increases their budget by reducing spending on goods other than x and y or the price decrease of good x allows the consumer to maintain this level of utility while spending less on the new x/y combination. This finding is represented by the distance from QG1 to QG2 in figure 1. As the price of gasoline fell, high school graduates substituted away from intracity mass transit to hold utility constant.

The income effect results from the fact that the reduction in price of good x raises consumers’ purchasing power, therefore granting them the ability to make new choices regarding the consumption of both x and y. That is, they can still purchase the same amount of y as they did before the price change for x and buy more x. Also, they can purchase the same amount of x as before the price change and use the money saved to purchase more y. Furthermore, consumers could increase purchases of both x and y while still spending the same amount that they used to spend before the fall in the price of x. The difference between quantity of gasoline purchased after the price decline and the substitution effect (mentioned earlier) is the income effect. This income effect is represented by the distance from QG2 to QG3. The income and substitution effects shown by the changes from 2019 to 2020 moved in a parallel and not an opposing direction. This result indicates gasoline is a “normal good.” A normal good is defined as any good that CUs will purchase more of when their income levels rise or purchase less of in terms of quantity when their income levels fall. Income and substitution effects moving in opposing directions would show gasoline to be an inferior good. If gasoline was an inferior good, the substitution effect would cause the CU to purchase additional units of x, and the income effect would trigger an opposing reaction. However, if the income effect outweighed the substitution effect, only then would the CU purchase less gasoline, despite the drop in price. CUs purchase more of an inferior good when their income falls and vice versa.

Figure 1 shows that, although the price reduction allows the average CU within the group of interest to consume the same bundle of gasoline and intracity mass transit as before the price declined, the average CU will purchase considerably more gasoline and somewhat less public transportation than the CU did before the price decline. This behavior is only because of the way the indifference curves are drawn in the figure. The curves can possibly be redrawn so that quantities of both intracity mass transit and gasoline would increase after the price reduction in gasoline. To see an example of such a point on BC3, at any level of gasoline purchase between QG1 and QG2, the corresponding purchase of intracity mass transit is larger than it is at E1, the amount of mass transit purchased when gasoline is at its original price. This result is also true at some points greater than QG2, but this result is not explicitly shown in figure 1. However, with a horizontal line drawn from E1 to BC3, a vertical line is dropped at that point of intersection, and the represented quantity of gasoline purchased is called QG2'. The amount of mass transit that could be purchased after gasoline prices fall is higher than at E1, at all points greater than QG1 and less than QG2'.

Then in 2021, if another representative or average CU were randomly chosen in which the most highly educated member of the CU had a high school diploma, there would be stark differences in its new (2021) utility maximizing behavior when compared with the differences in 2020. With no further shocks in the model aside from the unanticipated income and price shocks previously mentioned, BC3 in figure 1 is transposed to figure 2 as BC1. Thus, the initial endowment point is the point of tangency between BC1 of $1,712 and the indifference curve I1 at point Ei. Then with the distribution of the budget constraint in figure 2 identical to that in figure 1, the first point of tangency would be ($1,645, $67). In 2021, income growth was 4.2 percent for any given CU whose highest educational attainment was a high school diploma.40 This finding is shown in the outward shift of BC1 to the new budget constraint BC2, demonstrating the potential for this CU to obtain a preferable bundle of goods on the higher indifference curve I2, with a budget constraint of $1,784. For this simple model, one can assume that the full amount of additional income was spent on these two goods. In this model, no leakages exist as to where this extra income is being invested or spent on any alternative goods. That is, it has been assumed that all extra income is spent on a combination of mass transit and gasoline.

A small increase in income and increase in the price of gasoline on high school graduates’ intracity mass transit consumption using true expenditure data 2020–21

According to CPI average price data from 2021, the same source cited above (CPI Database for Average Price Data, U.S. City Average), the average price faced between January and December 2021 was $3.13 a gallon, an 89-cent-a-gallon or 39.7-percent difference from the previous year.41 This increase in the price of the x good will cause a pivot inward compared with a kink outward of the budget constraint. In figure 2, this pivoted-in budget constraint is represented by BC3, a considerably steeper line, emanating from the point on BC2, in which a given CU would only purchase intracity mass transit. This type of purchase shows the CU’s inability to purchase the same quantity of the x good gasoline or, in this case, 28 percent fewer gallons if the entire budget went toward the higher price gasoline. Once again, price stickiness is assumed for intracity mass transit, so no movement occurs in price of the y good. This price stickiness refers to the lack of price changes to intracity mass-transit fares by supply and demand market forces. An executive board of the transit agency generally set the fares, which can go many years without change.

With no price change for intracity mass transit and higher gasoline prices affecting a CU’s ability to purchase the same bundle of goods before the price change, any given CU will move to the lower indifference curve I3 in figure 2. This curve represents a suboptimal bundle of goods compared with the bundle that was obtained the previous year in 2020. As was the case in figure 1, the aforementioned income and price shocks yield the income and substitution effects in figure 2. The substitution effect represents the move along the indifference curve I2 from point E1 to E2 after the positive income shock to show a movement away from gasoline, which has become more expensive than it was before the price change relative to intracity mass transit. Thus, the substitution effect is quantified by the distance from QG1 to QG2. The income effect, quantified by the distance from QG2 to QG3 shows the movement to a lower indifference curve because of the reduction in purchasing power from the upward movement in gasoline prices. As stated, this substitution represents that CUs are driving less, but not necessarily that they are substituting 1 to 1 a drive in the car for a trip on the local railroad, bus, or subway. In figure 1, gasoline exhibits characteristics of a normal good. Figure 2 corroborates this finding by showing the income and substitution effects working in a parallel direction.

Generally, as shown in chart 7, 2021 saw recoveries in expenditures for two-thirds of the education groups (i.e., 4 out of 6) analyzed. The clustering of groups that increased spending is important to note. Two distinct peaks occurred in spending gains: one at the lower end of the education distribution among high school graduates and lower and the other at the high end of the distribution with bachelor’s degree earners and above. On the lower education peak, those with less than a high school diploma and high school diploma earners saw gains of 23.3 and 44.8 percent, respectively. At the opposing end of the distribution, bachelor’s degree holders and those with at least a master’s degree saw similar patterns of intracity mass-transit spending in which master’s degree holders and above saw gains of 13.7 percent and bachelor’s degree holders 49.4 percent. The outlier in the dataset was associate’s degree holders.

While 4 of the 6 education groups saw spending gains and high school graduates with some college remained flat, expenditures by associate’s degree holders declined a further 59.2 percent from 2020 levels. One explanation could be an increase in car ownership rates and a subsequent move away from public transportation as was seen with technical, sales, and clerical workers. That theory did not apply to those holding associate’s degrees because their car ownership rates remained steady at 93 percent every year from 2019 through 2021. Population distribution by area type can shed light as well. If associate’s degree holders in urban centers relocated to suburban or rural areas, demand for public transportation would be reduced. A 2022 release of 2019 data from the National Center for Education Statistics shows no material difference from the 2012–16 data in the 2018 Pew Research study of associate’s degree holders dispersion by area type.42

Trends in intracity mass-transit spending by selected MSA

Intracity mass-transit and public-transportation usage is critical for the efficient flow of workers in major cities in the U.S. and abroad. It shuttles commuters every workday to metro areas, such as New York, Chicago, San Francisco, or Washington, DC, from their respective suburbs. Even as far back as 2018, trends in intracity mass-transit spending by metro region presented a mixed picture. Possibly, most CUs who use intracity mass transit are relatively insensitive to price changes, with cars and maybe bicycling as the only potential substitutes if they live close to their employer. Spending in both the Washington, DC, and Chicago metro areas experienced visible declines well before the pandemic began. Expenditures fell by 40.9 and 24.8 percent in the two respective metro areas from 2018 to 2019, following fare increases the previous year; persons on the margin between transportation modes might have substituted toward an alternative mode. The New York metro area opposed this trend with 5.5 percent higher reported intracity mass-transit expenditures in 2019. (See chart 8.) The MTA did report higher “ridership figures,” indicating more frequent purchases and higher annual spending, but determining why New York opposed the trend on either the demand or supply side is difficult. San Francisco’s transit agency BART raised fares on subways in both January 2018 and 2020 but not 2019.43 If fares had risen in 2019, the 25.1-percent increase in expenditures could have explained why ridership in 2019 fell as well. From a function of total expenditures, this is unclear because the general price instrument remained the same in 2019 while the quantity instrument fell.44 Further exploration found that San Francisco’s Muni, the city’s bus network, raised its fares in 2019, which could be the missing link in the transmission.45 The case could also be that those CUs who reported expenditures on intracity mass transit in the San Francisco metro area simply spent more than the previous year whether for leisure or other auxiliary tasks, thus pushing up the aggregate expenditures figure for 2019, leading to the 25.1-percent positive spending shock mentioned earlier.

In 2020, New York metro area residents reduced spending on public and other transportation by 55 percent. (See chart 8.) Lockdown restrictions were at the forefront, with workers classified as essential making up most public-transportation demand in the early months of the pandemic. One reason for the sharp drop in demand could be because of the population outflow that the New York metro area had been enduring. According to an analysis from the Brookings Institution, the population of the New York metro area has been declining at an increasing rate since 2016, averaging three quarters of a 1-percent decline in population each year or a 3-percent population loss by 2020.46 The 2020 loss alone was larger in magnitude than years prior, exceeding 1 percent, or more than 100,000 residents. The trend in population reduction of 2020 was mirrored closely by the Chicago metro area. The Chicago metro area also observed a 1 percent decline in 2020, evenly split between the primary city and suburbs. With market size for public transportation reduced, prices firm, and quantity and demand falling, expenditures declined by association.

Washington, DC, had a slightly sharper contraction in public-transportation expenditures, declining just over 55 percent. Could this be a function of regional occupation? According to a 2019 Business Insider report, Washington, DC, had one of the highest concentrations of government workers, at 25 percent of its workforce; the same rank as Alaska.47 When the government transitioned to mandatory telework for most employees in March 2020, the city with the highest proportion of employees in this sector was bound to see a profound effect on public-transportation usage. This result extends beyond just government employees. Since government employees were not populating the agencies’ buildings, foot traffic was reduced for local city outlets, such as restaurants, reducing the agencies’ labor demand and thus public-transportation demand for their employees. Many government employees work in New York and Chicago, but Washington, DC, exceeds them in levels, thus exacerbating this effect. This decline was further compounded by an extended period of zero fare collection. From March 2020 to January 2021, DC Metrobus fares were waived; thus, although CUs did not pay, some might have still ridden the bus.48 Although spending by CUs of the Washington, DC, metro area declined just over 55 percent, spending by CUs of the San Francisco metro area fell farther, plummeting 63.4 percent in the initial year of the COVID-19 pandemic.49 (See chart 8.) Could industry or labor force composition be a contributing factor? The San Francisco Chamber of Commerce lists IT and software as the city’s primary industry; one that is conducive to maximum telework arrangements.50 Furthermore, the city is seen by some as the financial center of the west coast, as shown in a 2013 report by the Federal Reserve Bank of San Francisco.51 According to the previously mentioned ATUS 2019–2020 report from July 2021, 69.7 percent of those employed in the financial services sector reported working from home in 2020 compared with 36.7 percent in 2019.52

In 2021, 1 year into the COVID-19 pandemic, clear heterogeneity occurred in the speed of recovery to prepandemic levels. New York metro area residents charged ahead of their Washington, DC, Chicago, and San Francisco counterparts. New York metro area CUs spent 34.6 percent more on intracity mass-transit expenditures in 2021 compared with the amount spent in 2020. In contrast, the Washington, DC, metro area residents spent 10.1 percent more. Chicago metro area residents, however, spent 6.5 percent less, while San Francisco metro area residents spent 77.2 percent less. (See chart 8.) What could explain such divergence?

As was the case in 2020, Washington, DC, fell the most because of the makeup of its workforce. Workforce composition could have prompted the rebound in New York the following year. The New York metro area labor market is synonymous with the financial services sector. According to a 2019 Monthly Labor Review article on employment and wage trends in the NYC financial services sector, this sector of the labor market accounted for greater than 10 percent of total nonfarm payroll for the city and more for the private sector labor force.53 In 2021, prominent financial firms pressured their employees to return to the office. The pressure to return to in-person work would lead to higher demand for intracity mass-transit use in the New York metro area and, in turn, higher expenditures for it. At the same time, because of the high concentration of government workers, the Washington, DC, metro area remained on a maximum telework posture in 2021, stifling the same comeback observed in New York. A CTA’s 2021 ridership report on Chicago metro area data revealed that total system ridership was down 0.8 percent, with bus ridership falling 3.4 percent and rail ridership rising 3.4 percent.54 Bus and subway ridership declines were observed during the weekday (prime commuting days), while ridership gains were seen on the weekends, although not enough to counteract the reduction in total weekday ridership. San Francisco metro area CUs reduced spending by a further 77.2 percent on top of the 63.4 percent decline from the previous year.55 (See chart 8.) The sluggish return to in-office work, particularly in locations such as San Francisco, might have downward bias on ridership and thus total expenditures for intracity mass transit.

Intracity mass-transit spending highlights over the last 4 years

As the train approaches its final stop, it is time to reflect upon key findings, discuss caveats, and chart the next leg of this trip. In this analysis, current events and microeconomic theory are used to examine trends in intracity mass-transit expenditures between 2019 and 2021. There is also an investigation of unique demographic breakdowns of interest in order to rationalize how and why intracity mass-transit expenditures have changed. Although the future of workplace locations remains uncertain and telework postures will likely contract noticeably, it is doubtful that intracity mass-transit expenditures will return to 2018 and 2019 levels. Data from the New York MTA show that as of October 2021, ridership remains between one-half and one-third of 2019 levels.56 While it is possible that central-city dwellers could gradually return to prepandemic spending patterns, what about those who live outside the central city (i.e., other urban and rural consumers), construction workers, or high school graduates? Will these groups develop a preference to use private transportation instead of public transportation? There is only one way to find out: to continue this journey with future analyses on private transportation.

Appendix: Universal Classification Codes of public and other transportation

Appendix table A-1 lists all expenditure Universal Classification Codes (UCCs) included within the category of public and other transportation, a description of each, and whether the UCC was included in the analysis of chart 2 of the main article.

Table A-1. Public and other transportation, universal classification codes and descriptions
Universal classification codeDescriptionIncluded in chart 2

530110

Airline faresYes

530210

Intercity bus faresYes

530311

Intracity mass-transit faresYes

530312

Local transportation on out-of-town tripsYes

530411

Taxi fares and limousine services on tripsNo

530412

Taxi fares and limousine servicesYes

530510

Intercity train faresYes

530901

Ship faresYes

530902

School-bus faresNo

Source: U.S. Bureau of Labor Statistics.

As table A-1 shows, seven of the nine expenditures under the category public and other transportation were included in chart 2 to contribute to the expenditure-shares analysis. The two expenditures that were left on the platform include taxi fares and limousine services and school-bus fares. Both expenditures combined account for less than 2 percent of the total share of public and other transportation. Furthermore, as mentioned previously, school-bus fare collection is an anomaly in the United States outside of a small number of counties. Positive or negative shocks on spending for school-bus fares (as well as taxi fares and limousine services on trips), given their small contribution to the total spending category of interest, are not expected to meaningfully shift public and other transportation spending, thus bolstering the decision to exclude them from the section.

 

Suggested citation:

Shane Meyers, "Two hours to the office, two minutes to the kitchen table: trends in local public-transportation expenditures from 2018 to 2021," Monthly Labor Review, U.S. Bureau of Labor Statistics, June 2024, https://doi.org/10.21916/mlr.2024.9

Notes


1 For more information on the New York State Governor’s stay-at-home order, see “Governor Cuomo signs the ‘New York State on PAUSE’ executive order” (New York State, March 20, 2020), https://www.governor.ny.gov/news/governor-cuomo-signs-new-york-state-pause-executive-order.

2 The U.S. Bureau of Labor Statistics (BLS) American Time Use Survey released in 2021 strictly captures data from May to December of 2019 and from May to December of 2020 respectively, see American Time Use Survey News Release––May to December 2019 and 2020 Results, USDL-21-1359 (Department of Labor, July 22, 2021), https://www.bls.gov/news.release/archives/atus_07222021.htm.

3 Historical data from 2019 to 2021 is from the U.S. Travel Association, which shows the breakdown of what components of travel (such as airlines) led to declines in transportation spending. For more information, see “U.S. travel forecast,” U.S. Travel Association, Summer 2022, https://www.ustravel.org/sites/default/files/2022-06/us_travel-forecast_summer2022.pdf.

4 According to the BLS Consumer Expenditure Surveys (CE) program, a consumer unit (CU), which is loosely related to that of a household, consists of any of the following: (1) “All members of a particular household who are related by blood, marriage, adoption, or other binding legal arrangements.” (2) “A person living alone or sharing a household with others or living as a roomer in a private home or lodging house or in permanent living quarters in a hotel or motel, but who is financially independent.” (3) “Two or more persons living together who use their incomes to make joint expenditure decisions. Financial independence is determined by spending behavior with regard to the three major expense categories: housing, food, and other living expenses. To be considered financially independent, the respondent must provide at least two of the three major expenditure categories, either entirely or in part.” For more information, see “Questions and answers,” Consumer Expenditure Surveys (U.S. Bureau of Labor Statistics) https://www.bls.gov/cex/csxfaqs.htm.

5 For additional information on what constitutes an “essential worker,” see Elka Torpey, “Essential work: employment and outlook in occupations that protect and provide,” Career Outlook (U.S. Bureau of Labor Statistics, September 2020), https://www.bls.gov/careeroutlook/2020/article/essential-work.htm.

6 For CE definitions on urban and rural area type and metropolitan statistical area (MSA), see “MSA definitions for the U.S. Consumer Expenditure Survey, 2015–2016,” Consumer Expenditure Surveys (U.S. Bureau of Labor Statistics), https://www.bls.gov/cex/ce_msa_201516.htm.

7 For the BLS Consumer Expenditure Survey and MSA tables, see “Tables,” Consumer Expenditure Surveys (U.S. Bureau of Labor Statistics), https://www.bls.gov/cex/tables.htm#calendar and https://www.bls.gov/cex/tables.htm#geo.

8 Integrated tables contain data from both the quarterly Interview Survey and weekly Diary Survey to provide a more comprehensive and dynamic analysis of average annual expenditures. For additional information, see “Section 3. Table types and structure,” Consumer Expenditure Surveys Tables: Getting Stared Guide (U.S. Bureau of Labor Statistics), https://www.bls.gov/cex/tables-getting-started-guide.htm#section3.

9 You can contact Consumer Expenditure Surveys staff for more information regarding detailed-level integrated tables at https://www.bls.gov/cex/contact.htm.

10 For more information on the variability of transportation spending, see “Consumer spending on public and other transportation in metro areas before and during COVID-19,” The Economics Daily (U.S. Bureau of Labor Statistics, March 2, 2023), https://www.bls.gov/opub/ted/2023/consumer-spending-on-public-and-other-transportation-in-metro-areas-before-and-during-covid-19.htm .

11 CE defines central city as living within the bounds of an MSA. For CE definitions on central-city area type and MSA, see Consumer Expenditure Surveys Glossary (U.S. Bureau of Labor Statistics), https://www.bls.gov/cex/csxgloss.htm.

12 For the 2019 release, urban and rural expenditures in tabular form were adjusted in the CE. Up to 1984, urban and rural data had been available but as a component of the housing tenure table. For the 2019 release, urban and rural data were separated into a “new” table known as “area type.” CE recently adjusted the definitions for collection of data for this type of area table. For collection years 2019 and 2020, the CE program used its own definitions for “urban” and “rural.” For collection year 2021, CE switched to the most current definitions of what constitutes an urban and rural area as designated by the U.S. Census Bureau while simultaneously adopting the Office of Management and Budget’s definition of “principal city.” The CE program produced both table 1720 (old) and table 1721 (new) in 2021. Table 1720 is used here to ensure continuity in data for the time series of 2019–21. For more information on the impacts to CE data from these definition changes, see Scott Curtin, “Changing how BLS defines ‘geographic areas’ in the Consumer Expenditure Surveys,” Beyond the Numbers, vol. 12, no. 5 (U.S. Bureau of Labor Statistics, February 2023), https://www.bls.gov/opub/btn/volume-12/changing-how-bls-defines-geographic-areas-in-the-consumer-expenditure-surveys.htm.

13 The cited data for CUs are from CE’s detailed-level tables, which are available upon request.

14 The cited data for CUs are from CE’s detailed-level tables, which are available upon request.

15 Information on automobile ownership rates for 2021 or any year is from CE’s detailed-level tables, which are available upon request.

16 American Time Use Survey News Release––May to December 2019 and 2020 Results, USDL-21-1359.

17 The cited data for CUs are from CE’s detailed-level tables, which are available upon request.

18 Data referenced here are found in CE’s detailed-level tables by occupation of reference person for 2019 and 2020, which are available upon request.

19 For data on the division of in-person and remote-work dynamics in 2021, see American Time Use Survey–– 2021 Results, USDL-22-1261 (U.S. Department of Labor, June 23, 2022), https://www.bls.gov/news.release/archives/atus_06232022.htm.

20 Michael Dalton and Jeffrey A. Groen, “Telework during the COVID-19 pandemic: estimates using the 2021 Business Response Survey,” Monthly Labor Review, March 2022, https://doi.org/10.21916/mlr.2022.8.

21 Torpey, “Essential work.”

22 For this information and that on the allocation of childcare responsibilities during the pandemic, see Enrica Croda and Shoshana Grossbard, “Women pay the price of COVID-19 more than men,” Review of Economics of the Household, vol. 19, February 15, 2021, pp. 1–9, https://doi.org/10.1007/s11150-021-09549-8.

23 Stephanie Ferguson, Understanding America’s Labor Shortage: The Scarce and Costly Childcare Issue (Washington, DC: U.S. Chamber of Commerce, April 27, 2022), https://www.uschamber.com/workforce/understanding-americas-labor-shortage-the-scarce-and-costly-childcare-issue.

24 Piecing Together Solutions: The Importance of Childcare to U.S. Families and Businesses, Chamber of Commerce Foundation report, vol. 6 (U.S. Chamber of Commerce Foundation, December 2020), https://www.uschamberfoundation.org/sites/default/files/EarlyEd_Minis_Report6_121420_Final.pdf.

25 Percent reporting refers to the quantity of CUs in a given wave of the Interview Survey or week of the Diary Survey who reported spending money on a certain good. Expenditures reported in CE tables are average annual expenditures and have data entered as zeros for those who did not purchase a specific good in a given timeframe.

26 Data referenced here are found in the detailed-level tables by occupation for 2020 and 2021, which are available upon request.

27 Richard Fry and D’vera Cohn, “In 2020, fewer Americans moved, exodus from cities slowed” (Pew Research Center, December 16, 2021), https://www.pewresearch.org/fact-tank/2021/12/16/in-2020-fewer-americans-moved-exodus-from-cities-slowed/#:~:text=Overall%2C%204.9%20million%20Americans%20left,3.1%20million%20for%202016%2D2018.

28 “Largest cities by population: 2024 metro area ranking” (Macrotrends, LLC), https://www.macrotrends.net/cities/largest-cities-by-population.

29 Dollar amounts can be found in CE’s detailed-level tables, which are available upon request.

30 “New privately-owned housing units started: total units (HOUST)” (Federal Reserve Bank of St. Louis, 2024), https://fred.stlouisfed.org/series/HOUST.

31 Fry and Cohn, “In 2020, fewer Americans moved, exodus from cities slowed.”

32 For these data and for data on likely job title by education cohort, see Melanie Hanson, “Education attainment statistics” (Education Data Initiative, October 15, 2023), https://educationdata.org/education-attainment-statistics.

33 For these data and for data on area type by education cohort, see Kim Parker, Juliana Menasce Horowitz, Anna Brown, Richard Fry, D’vera Cohn, and Ruth Igielnik, “Demographic and economic trends in urban, suburban and rural communities,” in What Unites and Divides Urban, Suburban and Rural Communities (Pew Research Center, May 22, 2018), https://www.pewresearch.org/social-trends/2018/05/22/demographic-and-economic-trends-in-urban-suburban-and-rural-communities/.

34 For data on area type by education cohort, see ibid.

35 “Calendar year means, shares across all items, and variances tables by demographic characteristics,” Consumer Expenditure Surveys (U.S. Bureau of Labor Statistics), https://www.bls.gov/cex/tables/calendar-year/mean-item-share-average-standard-error.htm#rf-add.

36 "Consumer Price Index (CPI) databases," Consumer Prince Index (U.S. Bureau of Labor Statistics), https://www.bls.gov/cpi/data.htm. To retrieve the gasoline data use, visit the CPI databases site and use the average price data multiscreen database.

37 “Calendar year means, shares across all items, and variances tables by demographic characteristics,” https://www.bls.gov/cex/tables/calendar-year/mean-item-share-average-standard-error.htm#rf-add.

38 Emma G. Fitzsimmons, "M.T.A is raising fares and tolls; one subway or bus ride will cost $2.75," The New York Times, January 22, 2015, https://www.nytimes.com/2015/01/23/nyregion/mta-raises-fares-subways-and-buses.html; “Changes to MTA fares and tolls in 2023” (Metropolitan Transportation Association), https://new.mta.info/transparency/mta-fares-tolls-2023; and Shaye Weaver, “NYC subway fare will increase for the first time in 8 years,” Timeout New York, July 19, 2023, https://www.timeout.com/newyork/news/nyc-subway-fare-will-increase-for-the-first-time-in-8-years-071923.

39 “First CTA fare increases In 9 years takes effect Sunday,” CBS News, January 7, 2018, https://www.cbsnews.com/chicago/news/first-cta-fare-increases-in-9-years-takes-effect-sunday/.

40 For more information, see “Calendar year means, shares across all items, and variances tables by demographic characteristics,” https://www.bls.gov/cex/tables/calendar-year/mean-item-share-average-standard-error.htm#rf-add.

41 Gasoline data are taken from BLS CPI’s average price data. For more information, see “Consumer Price Index (CPI) databases,” https://www.bls.gov/cpi/data.htm.

42 “Educational attainment in rural areas,” Condition of Education (Institute of Education Sciences, National Centers for Education Statistics, October 2022), https://nces.ed.gov/programs/coe/indicator/lbc/educational-attainment-rural?tid=1000.

43 “Get Clipper and save! New fares effective Jan. 1, 2018” (Bay Area Rapid Transit [BART], January 01, 2018), https://www.bart.gov/news/articles/2017/news20171201; and “Fare increase January 1, 2020” (Bay Area Rapid Transit [BART], January 2, 2020), https://www.bart.gov/news/articles/2020/news20200102.

44 “2019 BART facts” (Bay Area Rapid Transit [BART], 2019), https://www.bart.gov/sites/default/files/docs/2019%20BARTFacts2019%20FINAL.pdf; and “BART 2018 factsheet” (Bay Area Rapid Transit [BART], 2018), https://www.bart.gov/sites/default/files/docs/2018_BART%20Factsheet.pdf.

45 “Muni fare increase begins July 1, 2019” (San Francisco Municipal Transportation Agency [SFMTA], June 18, 2019), https://www.sfmta.com/reports/muni-fare-increase-begins-july-1-2019.

46 William H. Frey, “America’s largest cities saw the sharpest population losses during the pandemic, new census data shows” (Washington, DC: The Brookings Institution, June 8, 2021), https://www.brookings.edu/research/the-largest-cities-saw-the-sharpest-population-losses-during-the-pandemic-new-census-data-shows/.

47 Andy Kiersz, “Here’s how much of each U.S. state’s workforce is employed by the government,” Business Insider, January 25, 2019, https://www.businessinsider.com/percentage-workforce-employed-by-government-every-us-st ate-2019-1.

48 Justin George and Ian Duncan, “D.C. votes to eliminate Metrobus fares in movement toward free transit,” The Washington Post, December 6, 2022, https://www.washingtonpost.com/transportation/2022/12/06/dc-council-free-metrobus-vote/.

49 These annual estimates were obtained using internal Consumer Expenditure Surveys microdata.

50 “IT & Software,” Key Sectors (San Francisco Chamber of Commerce), https://sfchamber.com/resources/economic-development/key-sectors-2/.

51 “Federal Reserve Bank of San Francisco 2013 annual report” (San Francisco Federal Reserve Bank, 2013), https://www.frbsf.org/wp-content/uploads/2013-Annual-Report.pdf

52 American Time Use Survey––2019 Results, USDL-20-1275, (Department of Labor, June 25, 2020), https://www.bls.gov/news.release/archives/atus_06252020.htm; and American Time Use Survey News Release––May to December 2019 and 2020 Results, USDL-21-1359.

53 Mario A. González-Corzo and Vassilios N. Gargalas, “Recent trends in employment and wages in New York City’s finance and insurance sector,” Monthly Labor Review, April 2019, https://doi.org/10.21916/mlr.2019.7.

54 “Annual ridership report: calendar year 2021,” Ridership Analysis and Reporting (Chicago Transit Authority, January 24, 2022), https://www.transitchicago.com/assets/1/6/Ridership_Report_-_2021_Annual_Report.pdf.

55 These annual estimates were obtained using internal Consumer Expenditure Surveys microdata.

56 “Focus on ridership,” Ridership and Crossings Update—October 2021 (New York’s Metropolitan Transit Authority, October 2021), https://new.mta.info/document/59346. This document provides time series data on ridership levels for the Metropolitan Transit Authority New York City transit for all major components of intracity mass transit including subway, bus, and commuter rail to show current ridership dynamics.

article image
About the Author

Shane Meyers
Meyers.Shane@bls.gov

Shane Meyers is an economist in the Division of Consumer Expenditure Surveys, U.S. Bureau of Labor Statistics.

close or Esc Key