Quick Links
Key Findings
- Job recovery nationwide has slowed down significantly since August ‘20. The cold weather expected to hit starting in November poses a significant threat to an already fragile economy, and may potentially erase the past two months of job gains, an estimated $190 billion in economic activity.
- Many businesses across two highly affected industries—Retail Trade and Leisure & Hospitality— made adjustments to the way they operated during summer months in order to open back up (e.g. outdoor dining). However, these adjustments will be significantly less feasible during winter.
- Using a conversative estimate, the 1.4 million jobs that have been recovered since April could be lost from the impact of winter weather.
- Job losses could number 2.8 million under a more severe scenario, in which half of the gains made across all private industries are lost due to cold-weather induced spike in coronavirus cases, for instance.
- These losses are most acute in America’s mid-size and smaller cities: 14 of the 20 cities facing the largest projected losses have populations below 200,000 people.
- Black and Latinx workers bore a disproportionate share of job losses at the onset of COVID. In cold weather cities, Black workers are projected to lose 140% of recovered jobs across Retail Trade and Leisure & Hospitality. Latinx workers are projected to lose 51% of recovered jobs in these industries, and women are projected to lose 76%.
- In cold-weather cities, small businesses account for 54% of employment within these two industries. If small business failure rates approach those seen in the spring, an additional 20,000 firms could be forced to shutter.
Introduction
In September, employers added 661,000 jobs across the United States, a steep drop in pace from the 1.4 million jobs added in August and a considerable slowdown in the economic recovery experienced since the onset of the pandemic-induced recession in April. The modest recovery over the summer has been fueled by the adjustments that businesses made to cope with restrictions in indoor service, by providing retail and dining services outside. The arrival of cold winter weather has loomed over businesses across the U.S., creating significant concerns about the inability to continue current strategies that worked in warmer, summer months. As a result, there should be significant concern that winter will eliminate the ability to use outdoor spaces for physical distancing and result in a broader economic contraction, which could be made worse with a spike in COVID cases.
This report provides a prediction for the potential impact to economic activity as a result of colder weather, and to estimate the number of jobs at risk if no further steps are taken to prevent a rise in positive COVID cases or to aid businesses in adapting service to the drop in temperatures. Cities facing this looming crisis are spread not only across the northern U.S., from New England to the Pacific Northwest, but extend into the south and southwest and represent 45% of total private employment nationwide. If Retail and Leisure businesses are unable to adjust to the coming drop in temperature, 1.4 million jobs that have been recovered in these sectors since April could be lost. As many as 2.8 million of recovered jobs are at risk of being lost if a spike in cases (which is expected in winter months) causes a broader contraction in economic activity across all sectors of the economy. This does not account for additional losses as a result of businesses that may need to furlough workers as a result in decreased foot traffic or businesses are unable to maintain headcount as their initial PPP lifelines run out.
The arrival of colder weather threatens to wipe out the jobs gained in August and September, or roughly $190 billion in economic activity. These losses are most acute in America’s mid-size and smaller cities: eight of the ten cities facing the largest projected losses have populations below 200,000 people. Black and Latinx workers bore a disproportionate share of job losses at the onset of this pandemic, and these further losses threaten to exacerbate the economic burden placed on those already struggling.
Where–and How Big–Is the Risk?
The first step to assessing the risk that winter weather poses is to identify the areas that typically become cold enough that prolonged outdoor time is not comfortable or enjoyable. We do this by classifying cities as “at risk” if the average monthly air temperature drops below 50 degrees Fahrenheit by November.
Figure 1 plots the mean temperature in November for each Micropolitan or Metropolitan Statistical Area–our measure of a city in this analysis. The risk posed by colder temperatures is concentrated in the northern portion of the country–from New England, through the Midwest, and out to the Pacific Northwest–but extends considerably far south as well. These “at-risk” cities represent a significant portion of national economic activity, accounting for 45% of total private employment in March 2020, before the pandemic-induced recession began.
We’ve focused on the Retail Trade and Leisure & Hospitality industries throughout this analysis because they are likely to be most affected by the drop in temperature–that is, they are most affected by fears of virus transmission and businesses in those industries have made adjustments that rely to the greatest extent on outdoor spaces. Importantly, the risks posed by cold weather extend beyond outdoor dining, although that may be most visible. Retailers have moved services outdoors to sidewalks and parking lots, and socially-distant indoor shopping often relies on consumers waiting in lines outdoors. Concert venues and movie theaters have also relied to differing extents on transitions to outdoor service over this summer.
Table 1 presents the largest of these cold-weather cities, sorted by the portion of private employment in each city that consists of jobs in these two industries. These metro areas are spread across the northern part of the United States, and all have significant portions of their economy in these affected industries, ranging from roughly 30-40% of employment. These cities most at risk are also not large metropolitan areas, but smaller cities reliant on bars, restaurants, and retail outlets for economic activity. Estimates of exposure among the largest cities are available in Table A1.
Table 1: Cold-Weather Cities, by Concentration in Retail, Leisure, and Hospitality
| Metropolitan Area | Fraction of Employment in Retail andLeisure and Hospitality | Population (Thousands) |
| East Stroudsburg, PA | 41.8% | 170.3 |
| Santa Fe, NM | 39.9% | 150.4 |
| Kingston, NY | 35.9% | 177.6 |
| Wenatchee, WA | 35.4% | 120.6 |
| Iowa City, IA | 34.3% | 173.1 |
| Lawrence, KS | 33.6% | 122.3 |
| Bremerton-Silverdale, WA | 33.3% | 271.5 |
| Bend-Redmond, OR | 32.5% | 197.7 |
| Asheville, NC | 32.4% | 462.7 |
| Coeur d’Alene, ID | 32.3% | 165.7 |
| St. George, UT | 32.3% | 177.6 |
It is also possible to identify cities threatened by the change in weather in other ways–for instance, certain areas not identified here may experience significant increases in precipitation, making outdoor activity less likely. In this way, the methodology presented represents a conservative estimate of areas in which businesses must make further service adjustments.
Assessing the Possible Drop in Employment
With the scope of the risk now in mind, we assess the magnitude of job loss faced by economies in these cold weather areas. The approach we use benchmarks the drop in employment to gains made since the trough of the recession in April. In short, we estimate the number of jobs at risk by asking “what would the employment picture look like within these cold-weather cities if businesses in the Retail Trade and Leisure & Hospitality industries lose half of the jobs that have been gained back since April?”
This exercise is thus underpinned by the assumption that half of the retail and leisure jobs recovered over the summer are due to employer adjustments that are not sustainable in the colder weather. Surely there are other factors that contributed to job gains over this time, but recent economic research has found that these forces, including the lifting of state and local shutdown orders contributed very little to the resumption in economic activity (Goolsbee 2020). Economic activity was much more responsive to business efforts to decrease crowding within the establishment.
Furthermore, we scale the possible employment losses to the gains made over the summer for two reasons. First, the drop in temperatures poses a threat to the slow but ongoing economic recovery–so a conservative approach at quantifying this threat is to scale it to the gains made since the depth of the recession. Second, this approach is not dependent on assumptions about the mechanisms through which this job loss occurs. Economic gains have been made because businesses have been able to adjust to outdoor dining, virus cases have slowed in some areas, and some localities have lifted state-mandated shutdowns. Instead of attempting to predict when states may shutdown or at what point consumers will stop going out–and then modelling the economic threat posed by each channel–we simply examine each city’s subsequent recovery to get an estimate of how many jobs are at risk of disappearing again.
This methodology also has the advantage that it is scaled to the individual experience of each metropolitan area. Cities across the county have had very different economic and public health experiences during the spring and summer, but using industry-level employment data within each city allows us to account for different local recovery trends.
It is reasonable to expect that cold weather cities will shed half of the jobs they gained back in Retail Trade and Leisure & Hospitality since April. That would equal a loss of 1.4 million jobs. Figure 2 plots the possible impact to employment, separating out the Retail Trade and Leisure & Hospitality industries, relative to a baseline scenario in which employment within these industries continues its current trend. In August 2020, the Retail Trade and Leisure & Hospitality Industries employed 11.4 million workers in cities where the temperature will soon cool down. This figure represents a marked improvement from the combined 8.6 million employed at the trough of the recession in April. If even half of these workers gained back were due to business adjustments that will become infeasible once the weather cools, we can expect an estimated drop of 1.4 million jobs.
Retail businesses were hit less hard by the pandemic than those in Leisure & Hospitality, likely because many retail shops were able to transition services to delivery or e-commerce. Over the summer, however, retailers were able to recover lost jobs by opening vending on sidewalks, parking lots, and closed streets and by limiting indoor capacity in physical storefronts through queuing customers outside.
Given September’s monthly employment gain of 677,000 jobs, a loss of 1.4 million jobs represents roughly two months of recovery efforts. To assign a monetary value to these job losses, 1.4 million jobs is roughly one percent of total employment. Scaled to Gross Domestic Product, a one percent contraction in economic activity represents $190 billion in economic activity that would be lost as a result.
In many ways, this represents a conservative estimate of the number of jobs at risk due to the onset of colder weather. Here we have restricted the potential losses to the two most at-risk industries. Under a more severe scenario, the losses in these industries or a cold-weather induced spike in COVID cases could create additional pullback in economic activity. If half of the jobs gained back across all private sectors are shed, the total number of jobs lost would increase to 2.8 million. This number does not take into account (1) businesses that have furloughed workers and for whom the drop in foot traffic over the winter may be a final straw and (2) businesses that have been able to maintain headcount through the Paycheck Protection Program, who may be forced to lay off workers as these lifelines run out.
Table 2 lists the cities which would be hit hardest by the more conservative scenario (in which losses are confined to Retail and Leisure & Hospitality). The largest losses, as a portion of the local economy, are concentrated in the Midwest and Mid-Atlantic Regions, which could lose up to six percent of employment as a result of colder weather. As in Table 1, these cities facing the largest potential losses are not largely big metropolitan areas, but rather smaller cities reliant on retail, restaurants, and bars. No cities in the top 10 besides Asheville, NC and Providence, RI have a population above 200,000 people.
Table 2: “Cold-Weather” Cities That Face Greatest Job Loss, as a Fraction of Total Private Employment
| Metropolitan Area | Losses as a Fraction of Total Private Employment | Projected Employment Losses (Thousands) | Population (Thousands) |
| Glens Falls, NY | 6.27% | 2.5 | 125.1 |
| East Stroudsburg, PA | 6.04% | 2.55 | 170.3 |
| Portsmouth, OH | 5.49% | 3.95 | 75.3 |
| Kingston, NY | 5.25% | 2.4 | 177.6 |
| Rapid City, SD | 5.17% | 2.95 | 142.1 |
| Jackson, MI | 5.13% | 2.3 | 158.5 |
| Asheville, NC | 4.74% | 7.35 | 462.7 |
| Niles-Benton Harbor, MI | 4.64% | 2.35 | 153.4 |
| Muskegon, MI | 4.55% | 2.3 | 173.6 |
| Providence-Warwick, RI-MA | 4.45% | 21.75 | 1,624.6 |
| Iowa City, IA | 4.45% | 2.65 | 173.1 |
| Lancaster, PA | 4.40% | 10.35 | 545.7 |
| Pocatello, ID | 4.39% | 1.25 | 95.5 |
| Pueblo, CO | 4.37% | 2.2 | 168.4 |
| Medford, OR | 4.31% | 3.2 | 220.9 |
| Elmira, NY | 4.24% | 1.25 | 83.5 |
| Carson City, NV | 4.23% | 0.85 | 55.9 |
| Utica-Rome, NY | 4.22% | 3.7 | 290.0 |
| Bay City, MI | 4.21% | 1.1 | 103.1 |
| Bellingham, WA | 4.17% | 3.2 | 229.2 |
Table 3 presents estimates of these employment losses for the 10 largest cold-weather metropolitan areas, now sorted by population. In absolute numbers, the New York-Newark metropolitan area faces the largest employment drop, over 200,000 jobs, and all cities face losses of roughly two to three percent of private employment this coming winter.
Table 3: Employment Losses in Largest “Cold-Weather” Cities, by 2019 Population
| Metropolitan Area | Population (Thousands) | Projected Employment Losses (Thousands) | Losses as a Fraction of Total Private Employment |
| New York-Newark-Jersey City | 19,216.2 | 214.6 | 2.8% |
| Chicago-Naperville-Elgin | 9,458.5 | 96.7 | 2.5% |
| Washington-Arlington-Alexandria | 6,280.5 | 41.8 | 1.7% |
| Philadelphia-Camden-Wilmington | 6,102.4 | 57.9 | 2.4% |
| Boston-Cambridge-Newton | 4,873.0 | 41.2 | 2.7% |
| Detroit-Warren-Dearborn | 4,319.6 | 49.9 | 3.0% |
| Seattle-Tacoma-Bellevue | 3,979.8 | 41.2 | 2.4% |
| Minneapolis-St. Paul-Bloomington | 3,640.0 | 49.7 | 3.0% |
| Denver-Aurora-Lakewood | 2,967.2 | 33.5 | 2.6% |
| St. Louis | 2,803.2 | 29.3 | 2.5% |
Winter Could Erase the Gains Made by the Hardest-Hit Communities
An important aspect of this recession is the disproportionate burden borne by historically underserved communities. This recession has hit the Retail Trade and Leisure & Hospitality industries particularly hard, and both rely on part-time and low-wage employment. The concentration of female, Black, and Latinx workers in these industries has meant that they faced the brunt of employment losses as the pandemic unfolded.
In Table 3, we calculate the share of jobs held by women, Black, and Latinx workers as of March 2020 in Retail Trade and Leisure & Hospitality, and then estimate the shares of job losses from March-April accounted for by each group. Tables 4a and 4b presents estimates of the share of employment losses in Retail Trade and Leisure & Hospitality accounted for by each group and then calculates how much of the recovery in each industry could be reversed come winter, if the 1.4 million losses are distributed in the same way as in March. Within these cities, the employment drop projected this Winter may more than completely undo the recovery Black workers in these two industries have experienced since April. The drop in temperature also threatens to destroy half the progress made by Asian and Latinx workers, and three-quarters of the recovery seen among women. The losses this winter thus threaten to exacerbate the trends in inequality that have become a hallmark of this recession.
Table 4a: Portion of Job Gains in Cold-Weather Cities that Could be Lost, by Race/Ethnicity
| Race/Ethnicity | Share of March-April Job Losses in Retail Trade and Leisure & Hospitality | Share of April-August Gains that Could be Lost During Winter |
| % Asian/Pacific-Islander | 9.0% | 51.5% |
| % Black | 12.4% | 140% |
| % Latinx | 16.1% | 51.2% |
| % White | 58.3% | 56.7% |
Table 4b: Portion of Job Gains in Cold-Weather Cities that Could be Lost, by Gender
| Gender | Share of March-April Job Losses in Retail Trade and Leisure & Hospitality | Share of April-August Gains that Could be Lost During Winter |
| % Female | 49.3% | 75.7% |
| % Male | 50.7% | 53.1% |
The Impact of Winter on Small Businesses
Finally, this pandemic has been particularly hard on small businesses, who often don’t have the cash buffer to survive extended business closures or temporarily reduced business capacity. Survey data has shown that almost 110,000 small businesses had permanently closed by June 2020—almost two percent of small businesses nationwide, with rates as high as three percent in the Retail Trade and Leisure & Hospitality industries (Bartik et al. 2020).
Within cold-weather cities, businesses employing fewer than 100 employees employed 54% of workers in these two industries, greater than the 50% industry share nationwide (Census SUSB 2020). There are an estimated 944,000 small businesses operating in these two industries in cold-weather cities, or 925,000 taking into account the roughly two percent that have already failed. If the drop in temperature causes a small business failure rate in Retail Trade and Leisure & Hospitality firms near that in the spring (2%) an additional 20,000 firms could be forced to shutter. Additionally, under a more severe scenario in which business failures extend to all private firms, this number would increase to 98,000 small employers.
Policy Prescriptions: Winter with Coronavirus
Absent a broad-based public health effort, small businesses (which are already strapped for cash) need aid adjusting services for winter service. Grants for simple equipment such as heat lamps and tents could be relatively cheap and effective tools at making outdoor consumption more attractive during colder months. Specific policies that could help small businesses adjust during the winter include:
- Permit Paycheck Protection Program (PPP) funds to be used on additional non-payroll expenses, such as outdoor heating equipment and personal protective equipment (PPE)
- Provide specific grants and tax incentives that make it easier for businesses to get the necessary equipment to operate safely during a COVID winter
- Access to free COVID testing by mail, with or without insurance, so people do not have to travel to testing locations in adverse conditions
- Expanded small business access to capital (including a second-draw PPP, expanded Economic Injury Disaster Loan advances and loans) to help those with continued negative revenue impacts
Local governments in cities like Washington, D.C. have initiated grant programs to help businesses make outdoor areas “winter-ready,” but the federal government has a large role to play as state and local governments are already facing huge projected budget deficits.
References
Bartik, Alexander, Marianne Bertrand, Zoe Cullen, Edward Glaeser, Michael Luca and Christopher Stanton. 2020. The impact of COVID-19 on small business outcomes and expectations. Proceedings of National Academies of Sciences. 117(30). 17656-17666.
Bureau of Labor Statistics. 2020. Current Employment Statistics. https://www.bls.gov/ces/. Accessed 9 October 2020.
Bureau of Labor Statistics. September 2020. The Employment Situation – September 2020. https://www.bls.gov/news.release/pdf/empsit.pdf. Accessed 9 October 2020.
District of Columbia Mayor’s Office. 21 September 2020. “Mayor Bowser Invests $4 Million to Help Small Businesses Continue Outdoor Dining Through Winter Months” https://mayor.dc.gov/release/mayor-bowser-invests-4-million-help-small-businesses-continue-outdoor-dining-through-winter. Accessed 9 October 2020.
Goolsbee, Austan, and Chad Syverson (2020) “Fear, Lockdown, and Diversion: Comparing Drivers of Pandemic Economic Decline 2020,” NBER Working Paper # 27432. https://www.nber.org/papers/w27432
Long, Heather, Andrew Van Dam, Alyssa Fowers, and Leslie Shapiro. “The covid-19 recession is the most unequal in modern U.S. history.” Washington Post. 30 September 2020. Available at: https://www.washingtonpost.com/graphics/2020/business/coronavirus-recession-equality/. Accessed 9 October 2020.
NASA North American Land Data Assimilation System. “North America Land Data Assimilation System (NLDAS) Daily Air Temperatures and Heat Index (1979-2011)” https://wonder.cdc.gov/nasa-nldas.html. Accessed 9 October 2020 through CDC WONDER Database.
Ritter, Emily. “‘I Can’t Keep Doing This:’ Small Business Owners are Giving Up.” New York Times. 13 July 2020. Available at: https://nyti.ms/3j8LgNc. Accessed 12 October 2020.
Ruggles, Steven Sarah Flood, Ronald Goeken, Josiah Grover, Erin Meyer, Jose Pacas and Matthew Sobek. IPUMS USA: Version 10.0 [dataset]. Minneapolis, MN: IPUMS, 2020. https://doi.org/10.18128/D010.V10.0
Shumsky, Robert and Laurens Debo. 2020. What Safe Shopping Looks Like During the Business Pandemic. Harvard Business Review. https://hbr.org/2020/07/what-safe-shopping-looks-like-during-the-pandemic. Accessed 11 October 2020.
US Census Bureau. 2020. Metropolitan and Micropolitan Statistical Areas Population Totals and Components of Change: 2010-2019. https://www.census.gov/data/tables/time-series/demo/popest/2010s-total-metro-and-micro-statistical-areas.html. Accessed 9 October 2020.
US Census Bureau. 2020. Statistics of U.S. Businesses (SUSB). https://www.census.gov/data/tables/2016/econ/susb/2016-susb-annual.html. Accessed 12 October 2020.
Appendix 1: Data and Methodology
This report combines city-level monthly temperature data with employment statistics at the city-industry-month level to estimate the possible drop in employment faced by cold-weather cities.
NASA’s North America Land Data Assimilation System (NASA 2020) calculates monthly air temperature for every county from 1895-2011. I aggregate November 2011 county-level monthly temperature data to the level of Core-Based Statistical Areas (CBSA’s, defined as Micro- and Metropolitan Statistical Areas), averaging county temperatures within each CBSA. CBSA’s are our measure of “cities” in this analysis. I count as “cold-weather” cities those CBSA’s whose November air temperature is below 50 degrees.
I merge CBSA-level temperature data with BLS State and Metro Area Employment estimates from the Current Employment Statistics from March-August 2020 (BLS 2020a). These data provide counts of employees in various industries, each month, for the 450 largest micropolitan and metropolitan areas. I restrict most analysis to the Retail Trade (BLS Industry Code 42) and Leisure and Hospitality (code 72) industries. In the wider scenario, I use the “Total Private Employment” Series.
For the above two industries within each city, I measure the number of jobs gained back since the trough of the recession by calculating the April-August gain in employment (April represents the trough of the recession in all cities). I estimate the number of jobs at risk of loss by calculating half of the gain in employment across these two industries (or across Total Private Employment) for all cold-weather cities.
Appendix 2: Table 1, for “Large” Cold-Weather Cities
Table A1: Large “Cold” Cities, by Concentration in Retail, Leisure, and Hospitality
| Metropolitan Area | Total Private Employment (Millions) | Fraction of Employment in Retail, Leisure, and Hospitality |
| Seattle-Tacoma-Bellevue, WA | 1.5092 | 23.6% |
| St. Louis, MO-IL | 1.0908 | 22.7% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 2.3425 | 22.5% |
| Denver-Aurora-Lakewood, CO | 1.1618 | 22.3% |
| Portland-Vancouver-Hillsboro, OR-WA | 0.9154 | 22.1% |
| Baltimore-Columbia-Towson, MD | 1.0128 | 21.7% |
| Chicago-Naperville-Elgin, IL-IN-WI | 3.6025 | 21.7% |
| Detroit-Warren-Dearborn, MI | 1.3572 | 21.6% |
| Pittsburgh, PA | 0.8663 | 21.3% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 2.1787 | 20.9% |
| Minneapolis-St. Paul-Bloomington, MN-WI | 1.5118 | 20.7% |
| New York-Newark-Jersey City, NY-NJ-PA | 6.7102 | 20.6% |
| Boston-Cambridge-Newton, MA-NH | 1.3805 | 18.9% |
