The Effect of COVID on Ridership, Was It Real?

0
814

Abstract

The COVID-19 pandemic’s effect on a large number of people is undeniable; however, one aspect that has not received much coverage or had conclusive evidence is the effect of COVID cases on public subway ridership. Subway ridership is the primary source of transportation in New York City and controls the entire cities economy as workers need to travel to get to their work. This article attempts to answer how New York City and the five borough’s subway participants changed based on the COVID-19 rates, specifically through three sets of regressions: Ridership Entries on COVID cases in New York City, Ridership Entries on COVID cases in the state of New York, and Ridership Entries on COVID cases in the United States during 2021 and 2022, from NYC Gov, by viewing the amount of cases per day and the amount of ridership on the same day to find any significance. The original, ridership and COVID cases fixed simultaneous data showed a relationship between ridership at all three levels; however, after accounting for fixed effects (time in month and location in borough) for New York City and the State, it was proven to have no statistical significance (if p < 0.05 there is a statistical correlation). The final regression accounting for the whole United States only used fixed effects accounting for time in month, and there was no statistically significance within the regression as well. Thus, there was no statistical significance on Subway Ridership in New York City from COVID because the subway is an inelastic good for most workers as the transportation is necessary for all their work. Future research would account for later years and 2020 COVID ridership data, as well as view other obstacles that should affect subway ridership counts. If the amount of riders doesn’t change, it would prove the good continues to be inelastic to most consumers. While the study effectively accounted for COVID cases on ridership, it did not account for potential spreading habbits and other measures of the severity of COVID that could have impacted riders. The effect of new variants has not been explored, also public transportion has not been evaluated in depth through studies which this study attempts to bridge the gap.

Introduction

The COVID-19 pandemic has affected nearly every aspect of life throughout the past two years, most in a negative way. One transformation that is debated heavily by many economists at The Economist, World Economic Forum, and countless other organizations is the effect of COVID on the transportation industry, and more specifically, the public transportation industry. This study attempts to answer the question of how COVID cases effected daily ridership on the Subway during 2021, attempting to find any correlation between the two would indicate the elasticity of the subway for citizens of the 5 boroughs. As The Economist states, “What has collapsed is rush-hour commuting, particularly among well-paid workers in the knowledge economy. Rich countries should accept this new reality.” Travel patterns have changed for good. Transport systems should. Working from home gives employees less work, more sleep, and countless other benefits. Even if COVID ever ends, the chance of a portion of workers moving back to the office is not very high. This surprise is not unique to New York City. According to the same Economist article, Paris shopping records in 2021 were the highest they had ever been, while lifestyle is different between the two cities, the simple fact is individuals are less concerned about catching the virus and will prioritize activities they enjoy.

This paper attempts to determine the impact of the coronavirus on public ridership in New York City, and within that, to what extent ridership has changed. A hypothetical interpretation before conducting the study is that as COVID data became more public to a larger geography of people, subway riders would react significantly — as more citizens tested positive, the subway ridership would decrease and vise versa. An example is during the Omnicron outbreak, since it was a countrywide spike in positive cases, it is assumed ridership would fall greatly; however, small scale COVID outbreaks just in a neighborhood or even in the borough would not be taken as seriously because it does not have the scale to get to the public and also still does not necessarily mean the virus is spreading any faster than normal. However, after gathering data, the assumptions were proven incorrect as ridership based on city COVID cases had no more effect than ridership based on borough wide cases accounting for time per month and location per borough, thus proving an unconditional correlation. The next set of regressions run was ridership vs COVID data by viewing a sample of each positive case in the 365-day span from January 1st, 2021 to December 31st, 2021 from each of the 5 boroughs, from the NYC governmental page (for city and borough cases) and the CDC COVID cases (for national level cases), with a focus on positive cases, date, and location as the primary variables. In the state of New York with the same fixed effects, the results should have no statistical correlation when adjusting for time and location as well. The final relationship ran was ridership vs COVID data in the United States, in this set the only fixed effect evaluated was borough, the results showed no correlation after the regression. While each regression did not show statistical significance for the relationship between ridership and COVID cases, the study was able to gain a grasp on how COVID affected the industry and an answer to whether subway ridership is stable or not. The two most important variables during the study were Subway Entries as shown in Figure 1, and COVID case count as shown in Figure 2, both during 2021, to represent the raw variable, each value is averaged then sorted by month. In Figure 1, Subway Ridership increased each month until December where it took a steep downturn. While in figure two, the COVID rate decreased at a steady rate of roughly 150 cases each month until November and December where there was a significant spike. Subway Entries is characterized as the amount of people riding the subway each day from that specific station. COVID case count is the total amount of postive COVID cases at each individual borough location.

Figure 1. Subway Ridership Entries per Month in 2021
Figure 2. COVID-19 cases per Month in 2021

Literature Review

The main driving force that inspired this research paper was the underground nature of the topic, while many economics topics have been examined in detail with thousands of studies and rigorous data analysis done to find outcomes, COVID ridership effects have not gained nearly as much traction because it is very new, but by running through the few studies and data analysis; there seems to be a direction in the literature. Most believe, after COVID-19 took storm in March of 2020, the ridership data directly decreased because of an increase in working from home, and that this effect would maintain for a long period of time. Important to know before engaging in the study is twofold, COVID and subway/Manhattan lifestyle. First, COVID, beginning in March of 2020 when the pandemic forced global isolation and when COVID data began tracking. Within this, the biggest case breakout, the Omicron wave, is assumed to start on December 1st, 2021 and end January 31st, 2022 according to the Guardian and Nature News, respectively. While Guardian and Nature News accurately gives recent trends on COVID and new mutations and varations, it fails to account for the impact on public transportation and how new COVID cases effect the subway which this study attempts to bridge. Next is the subway lifestyle. According to nycgo.com, (the subway is the most popular method of transportation in New York, and is especially popular for traveling within the boroughs1

While addressing subway ridership importance throughout New York City, it fails to account for varations and changes to the system that a global pandemic like COVID would affect, this study attempts to bridge the two gaps and find the correlation of the two. Knowing how people view the subway and COVID is important in evaluating the reasoning behind the experiment. One common experiment attempted was the Time-Lag effect of COVID on Seattle and New York City. Similarly, they used ridership data and COVID numbers, in addition to travel patterns, to find out that COVID has made ridership significantly more difficult in terms of finding rides2. Another germane publication is COVID-19 effect on public and private transportation, a study conducted at International Journal of Transportation Science and Technology found that private ridership such as cars will go up by 143% as pre-pandemic, while public transportation will go down to 73%3. The main reason to back up their claim is that people have a fear of COVID and as a response adapt to more of a public transportation model. Travel patterns have been studied and the effect of a global pandemic like COVID have had an effect on ridership in other parts of the world, specifically with private transportation like cars. This study views the effect on daily COVID cases while other studies focus on cases during a month or shorter sample, none exam daily for a year so the accuracy of specific riders and other riding habbits like work time and location are significantly limited.

Method

Throughout this process, the study used data exclusively available to the public from nyc.gov4; the data gathered was ridership data from 2019, 2020, and 2021, and COVID-19 case count and hospitalizations from the same time period. While gathering the data, each day was checked to make sure only one entry per borough was processed and no subway location was double counted in each borough to make sure there were no recounts. After gathering the data and transporting it over to a statistical software, Stata, the study ran multiple regressions on the relationship between COVID case counts to subway entries and found the relationship between COVID hospitalization and subway entries. The choice to use Stata allowed for a diverse set of datasets to be tested and allowing over 500,000 variables to be tested, allowing for a diverse amount of information to be tested. In addition, the sorting and specialization is crucial to gather data, especially in such a wide sample size like this study. Initially, simple regressions were run, after merging the two variables, to see if there was any sign of significance, and keeping each other variable controlled during the process, after each regression viewing the significance. A significance value above .05 would mean there is a correlation. Starting with two variables is necessary because if there was no correlation, there would be no significance and a conclusion could be made. The next step was to sort data specific to the five boroughs: Manhattan, Queens, Brooklyn, Staten Island, and the Bronx to view if any individual borough reacted differently to a COVID outbreaks, whether it was less or more serious. After each specific COVID range, the study ran a similar regression but from a different level of COVID geographical aggregation, while using the same case count variable for these studies. For example, after completing the New York City COVID data, the study then moved to the New York state data, until reaching the national level of COVID. While this should control the majority of endogenity issues, accidental omitted variables or reverse casual chance that other effects besides COVID had an effect on ridership could still exist.

Ridership Data 

While this study primarily focuses on 2021 ridership data, an interesting observation to look into is 2020 ridership because of a huge drop around March because of the initial Coronavirus, as shown in Figure 3. This proves, for many causes that we will later discuss, people traveling due to COVID did decrease. Whether there was an actual correlation between entries and months is still to be determined, but there is no doubt less people traveled at peak COVID.

Figure 3. Entries vs Monthly Ridership 2020

First, comparing the surface level ridership data from pre-Coronavirus to post-Coronavirus determines if there was any substantial difference due to the deadly virus. According to Metro Ridership, a site that evaluates yearly data surrounding ridership on any public transportation, reports that in 2019, weekday ridership estimated around 1,174,751 passengers from Monday through Friday. (Figure 4) Metro Ridership calculates the total number of riders at each station by calculating the total tickets sold, while not perfect because just because a ticket is sold, does not mean they rode the subway, it is net accurate, In addition, a lower value on the weekend with roughly 648,909 riders. (Figure 5) On the other hand, in 2021, the data not only decreased by a significant margin for both weekdays and weekends, but ridership from Monday to Friday plummeted. Importantly, consumers view the subway during the weekend much more elasitically because they do not need to go to their jobs while on the weekdays, they have no choice but to travel using the subway. Importantly, noticed by NYC government, during peak hour traffic, most decide to use the subway rather than find a taxi or walk themselves, this leads to times around 5 P.M. to have extremely high ridership rates but only on the weekdays5. Besides a change in lifestyle, a few notable reasons for the continued low ridership numbers in 2021 is at the start of the year, a COVID vaccine was at the very beginning of development and not many individuals had the opportunity to be vaccinated. Especially in a liberal and more COVID concerned city like New York City, vaccination definitely influences the use of public transportation at the start of the year. Not only was the vaccine development at a starting stage at the beginning of 2021, but the vaccine for children between the ages of 12-15 did not gain access until May. In turn, COVID-concerned families would be much less likely to travel with children. Each month from February saw more and more citizens get vaccinated and return to the subway, similarly as more got vaccinated, more people continued to use the subway because COVID was less of a concern, according to the COVID data available on the NYC government website6.


Figure 4. Weekday Ridership by Month 2021
Figure 5. Weekday Ridership by Month 2021

The second major factor that obviously has an effect on transportation in 2021 is the spike with Omicron in December of 2021 that would spread into 2022. From January of 2021 to November of 2021, ridership increased slightly each month; however, weekday ridership decreased from 844,930 riders in November to 740,746 in December. The main explanation for such a significant decrease in ridership in a single month is Omicron. The spike of COVID cases was the highest seen since the very beginning of the pandemic. An interesting distinction is the weekend to weekday, while weekdays face a heavy spike decrease in ridership, weekends barely decrease by around 7,000 passengers from November to December, and the December weekend data is still higher than many other months. Without evaluating both graphs, multiple conclusions can be drawn for the decrease such as: more COVID means people want less exposure, less activities were held during this time because of the spike, and the virus hurting public transportation, forcing people into private transportation or, even, other public transportation like buses or the railroad. However, the subway system offers a unique opportunity for citizens to receive a ride at nearly any time of the day and the flexibility of stopping at any place they want in the city. While railroads or a bus might be effecive for a short term solution, unless it gets workers directly to their target location, it is less effective. In addition, buses have to deal with traffic that a subway does not, this gives workers an advantage to choice the subway because of its extremely affordable price along with quick times.

While all these potential claims definitely do have some effect on Omicron’s effect on COVID, none of them has as much affect as the simple fact that individuals forced to work in public conditions away from their home, received in the virus at a higher rate, and, in turn, less people traveled on the weekday. However, the small increase on the weekend can be associated with more people that work from home spending their weekends away from home and traveling despite the serge. Also, weekends give many more possibilities of opportunities citizens can have such as bars, clubs, shopping, restaurants which are all more populated during Saturday and Sunday. Especially given most local shops cannot close because of their high fixed cost and need to stay in business, they will continue to stay opened despite the pandemic making the weekend very populated with more people available to travel. The amount of citizens working from home during the pandemic spiked as well—according to the US Census, working from home nearly tripled from 2019 to 2021 and a majority of workers were eligible to work from home.7 While there is no one cause to the situation, the significant difference between weekdays and weekends indicate which effects are the most significant. Another thing that could play a role in the data that could be accounting for the difference is mindset. Individuals at this point had endured over a year and a half of a virus; many individuals prioritized enjoyment and traveling around the city over the protection possibly gained by not traveling on weekends. On the surface, this spike decrease and difference between weekend and weekday seems like a coincidence or a one time effect because of Omicron, but working from home is not going anywhere. Workers with the ability to work from home have a multitude of advantages ranging from flexibility in their schedule to saving on driving costs. As technology advances, it becomes easier for workers to complete all their tasks without an office workspace. 

While comparing the relationship between pre-COVID and post-COVID to each other and ridership from two different years demonstrates a significant idea of what the effect of COVID on ridership is; it unfortunately does not tell the full story about the interaction between COVID and public transportation in New York City. One crucial task the first comparison does is narrow down the pool of possible primary causes – again, there are hundreds of little factors that influence the ridership, but not all of them have nearly the same magnitude. The next set of data comparison is a regression, comparing ridership to COVID cases. Despite significantly less attention in 2021, the 2020 March spike drop does indicate that a massive spike in COVID cases or a requirement from the government to social distance and quarantine will affect the number of riders. As the effect of COVID begans to become more sustainable and easier to address, more workers will continue to ride on the subway whether they view it as a normal good or a necessity. In the long term, ridership should continue back to an extremely useful option for all citizens of New York City to use as it gives endless possibilities of locations to travel and there is no longer a barrier that contrains many citizens’ health. Most of the information is covered with this data; however, some days COVID cases in specific counties were not accounted for and thus they are missing on the graphs and charts, but because the sample had so many days, it is unlikely it would impact the results of the study. In addition, some ridership data did overlap through boroughs which makes it hard to directly connect the data and might cause some repeats.

Results and Discussion

The first study conducted through Stata involved New York City ridership against COVID cases per each individual borough. Initially, the hypothesis was there would be significant correlation between the entries in a particular subway station and the case count in that borough. There were two main reasons behind the hypothesis: fear and lower riding population. First, as individuals get the virus, it is expected the number of riders decreases because they cannot ride with COVID because of testing requirements to protect all passengers. While the data is significant enough to make a general conclusion since it accounts for 365 days, since it only takes a year sample size and doesn’t account for the very peak of COVID in 2020, it still does not fully answer how COVID has affected ridership. In addition, high COVID numbers should, in theory, increase the fear of other people in the borough to not ride on the subway for fear of catching the virus. For example, if one person’s neighborhood had 5 cases, it is much more likely they would test and take safety precautions to not spread the virus. (Table 1) Thus the study ran this regression, comparing the number of entries onto a subway against the number of cases in each borough. The regression concludes that the p value is greater than 0.05 indicating no significant correlation which means it is impossible to conclude that there is a significant correlation between the data. (Table 1) It is clear a multitude of reasons could account for the correlation to not exist—it is highly likely that the data overlapped boroughs and counted them mulitple times increasing the amount of riders making the rider data saturated compared to the COVID cases.

Table 1
12
entriesentries
New York Borough Case Count-0.0615**
(-7.49)
New York State Case Count-0.00164
-(0.80)
cons4016.0***3977.6***
(280.96)(276.68)
N154361154361
t statistics in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001

To see if the data is statistically significant, P>|t| should be less than 0.05, from this data there is significant statistical significance, per the two star rating. However, the data on its own is not enough to actually prove the data has a causal relationship. To find out if it has a real impact, it must have some fixed effects that act as controls for omitted variables that might change the results. The first fixed effect used was location, the data used 4 different sets of locations since the data did not include Staten Island, instead only using the other four boroughs: Queens, Manhattan, Brooklyn, and the Bronx. Each borough effect is represented in Table 2. From each of the boroughs, it is clear that COVID has a minor effect on ridership; however, based on the case count alone, it cannot be fully assumed they are directly correlated and it could possibly be other obstacles such as family members attracting COVID or the threat of new cases on the news, or, even, a non-related situation such as workers being laid off and no longer needing the subway to get to work.

Table 2
1
Entries
Case Count0.0123***
(1.58)
Bronx2581.1***
(83.47)
Brooklyn2629.0***
(122.74)
Manhattan6671.5***
(284.55)
Queens3673.5***
(126.06)
N154361
t statistics in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001

This fixed effect essentially divided each data entry from each borough into individual subsections. For example, each station in Brooklyn joined the Brooklyn fixed effect (in this case location). This allowed for the data to test whether any of the boroughs had a specific effect on ridership. According to the hypothesis, all four boroughs should have a significant correlation based on case count because based on their COVID data, it should show a direct effect on ridership entries. Ideally, a borough with 1000 cases would have half as many riders as a city with 2000 cases because it would show the same reaction from subway riders if they responded to their borough’s neighborhood. From this data, it seems there is no correlation between the case count and entries relating to location, obviously individually the locations have significant correlation, but on the broader, more telling case count vs entries per each borough, there is zero significance as proven by no star value. This important conclusion proves that the data is not statistically significant enough when accounting location. 

According to the hypothesis, there should have been at least a marginal significance in the data; however, the data was not perfect. There were a few sources of error, most importantly: lack of sample size, incorrect date correlation, and lacking correct locations. First, lack of sample size which was the biggest error, because there were not enough entries to constantly count time inside each day, it is hard to determine when or why people are taking the subway which determines a lot of the ridership, instead it was simply a daily occurrence, which in theory should work; however, not every date entries has a corresponding COVID value which means each day is not actually accounted for which could change the data significantly. The next error could be incorrect date correlation, when merging the data sets, it was merged together using date and borough; however, since it was a single COVID case count data to multiple ridership accounts, there is a strong likelihood that some dates were not correctly matched leading to error because Stata can only match one ridership data point with one COVID data point. Finally, incorrect location data, this could happen again when merging the two sets because either a station is not sorted into a correct borough, or a station is not accounted for, either way causing significant error. 

More than likely, none of these errors were the main cause of the lack of correlation; probably most individuals riding the subway do not pay sufficient attention to their borough’s COVID cases, unless it is a major concern, and thus they do not react. Imagine a low-income worker must get to their job across the city, even if the case count in his borough is a high number, there is a very good chance that he ignores that protocol and attends work through the subway because he needs the work to cover expenses and the subway is one of the few affordable routes. The beauty and ugliness of public transportation is that no matter what the situation, people can still ride and ride for cheap. Borough representation in terms of location should have an effect, but after analyzing it and using empirical examples, it is clear that there is not much effect that can take place.

The next fixed effect accounted for was time, specifically in days. The new regression attempts to show that based on each day’s case count, ridership on the subway will vary. Similar to location, still entries and COVID cases are the main variables; however, instead of basing off living place, it is based on when the ridership happens. The thought process behind this is that on a day where COVID cases are high, or COVID cases are generally high during that time (ie Omicron) people should, in theory, be scared of the high case counts and the amount of people catching the virus. Thus limit their exposure and instead of traveling on public transportation like the subway, either stay home or take another route to their location. Surprisingly, the educated guess was not true. In fact quite the opposite was true, people do not care about the day’s COVID case count enough to prevent them from riding the subway. On the surface this is very surprising, but similar to location, the same logic applies. Some people do not have a choice; they have to take whatever method of transportation is offered to them to fulfill basic necessities. Not everyone has the luxury of having another source of transportation, most people use the subway for everything in New York City and whether the case count is high or not does not matter. A unique reason why time by day might not be the best factor is that people cannot account for it. For example, if a day has lots of new COVID cases spread, people will not find out until the next day more than likely, especially on such a small scale like a city. 

Figure 6. COVID Vaccination doses given per Month in 2021

One significant observation is the effect vaccines had on Ridership. As more people received vaccinations, more people would ride the subway with less concern. While COVID cases directly might not have effected ridership, it is clear sending a public signal that people are receiving treatment and that it is scientifically possible to make the disease less deadly, lead to many more citizens being comfortable traveling. As proven by the graph, very early in the year a large portion of people received the vaccine. Specifically, in March nearly 21,000 people per day would receive the vaccine, as shown in Figure 6. This made society significantly more safe, which in turn, led to more people willing to travel using public transportation like the Subway. This gave the Subway a lot of its original,  pre-COVID value, again. The excuse that COVID gave workers was no longer there, now workers were forced to either show up to their job, more than likely using the Subway, or stay at home for their work. Since nearly 3/4th of the study period had a very large portion of the New York City population vaccinated, it makes it very possible riders did not even account for COVID anymore and simply reverted back to their normal routine and lifestyle.

The next test accounted for was entries vs the United States COVID cases. The hypothesis was as the scope increased, more people would take the precautions to COVID more seriously and, thus, use public transportation less. Since the numbers are on a higher scale and broadcasted on the news to millions of people, it is much easier for people to adapt and change their lifestyle, or that is what was believed to be. Thus we ran a regression comparing ridership entries vs the national COVID cases, and here are the results:

Table 3
1
entries
Nation Cases0.00202***
(27.17)
cons3070.8***
(112.87) 
N20592
t statistics in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001

In addition, the national COVID data throughout the year is similar to New York Cities data throughout the year. Both data sets have a strong increase during the end of the year most likely due to Omicron, each data set has a drop around April to May when COVID tensions were more light. (Table 3) Due to this, the direct correlation of response should not be too different. During November and December, both studies expect the ridership totals to be lower because of the higher case count, and Figure 7 details the significant drop.

Figure 7. National Case Count vs Month

On the original regression, comparing New York City ridership entries to case count had an effect before accounting for fixed effects. However, after accounting for time in days, the only fixed effect of this study, it was proven to be insignificant. Since the data was similar to the New York City data, it is not a surprise that when accounting for the same variable, there is no effect. This regression did not include a location variable either because national COVID cases should have no distribution based on cases in individual boroughs. It would only have an effect on a national level, as each study becomes more broad, basing it on location becomes less and less important because the dependent variable is much more broad.  Despite a belief that expanding the range on COVID cases from city to state to nation would have an effect, it seems to have no effect at any of the three levels.

The original question asked when starting this experiment was, “To what degree did COVID-19 affect public transportation, specifically subway ridership in New York City?” According to the regressions, there was no effect on subway ridership in the year 2021 from monthly COVID cases on the city, state, or national level. This set of findings is quite significant for future ridership and COVID management, transportation is an inelastic good. No matter the situation, people need a method to get from one place to another. Public transportation is the option most people choose because of the affordability. Most people cannot afford their own vehicle and the added costs like gas, so even if COVID cases are high, they do not have a choice, they have to travel on the subway. While this is a logical conclusion and makes sense based on other core economic principles, the hypothesis disagreed with the results. Initially it was thought that each level would have some set of statistical significance because of previous studies and also logical conclusions drawn from the Economists and other news sources. Thus, this study is an outlier as most literature bases believe that the effect of COVID is large enough to have statistical significance. However, many of these papers are based on theories and do not use many studies to account for fixed effects and are not on a day-by-day basis like our study. In addition, the amount of articles and studies on the topic are, relatively and surprisingly, small. The regressions run between two unique factors like case count and ridership entries has not been done as a primary study through my findings. That makes it difficult to assess whether many flaws in the study existed or if the data just did not have a significant correlation. It is more than likely that the data simply did not have a correlation and is warranted by being one of the few affordable sources of transportation for people across New York City. However, there were a few limitations that potentially harmed results. Primarily is lack of sample size and merging complications. Since the study only focused on a one year basis, it is possible that 2021 is not enough information to give a long term assessment on whether or not COVID has a significant impact on ridership. In addition, when merging sets of data in a set, many entries do not merge with one another, which leads to exclusion between data points. Slight errors such as not 100% accurate data every single day might exist but are unlikely and are highly negligible considering the extremely high sample size. However, it is extremely important to account for this because it could potentially mismatch certain cases to boroughs which could have a spill over effect to other data entries; this is unlikely as after testing multiple data points without regressing, they were directly matched together. After this study, the next step would be to explore more than just the year of 2021, see how COVID trends changed over 2020 and if any other major factors had a huge impact on ridership. While it might not be COVID, I believe there is a reason for the ridership decrease we have seen in 2021 compared to 2019. However, another highly plausible cause is that workers have moved away from the office. According to Bloomberg news, over half of the workforce worked from home in 2021 due to the deadly virus. That accounts for a minority of the portion of the population, and the people working from home view public transportation as an elastic good, so there is not a significant distinction between their pre and post-COVID ridership because some people completely stopped using it, while some people started using it a lot more. Specifically, high tech workers, bankers, stock investors, and countless other positions that are mostly based on technology would not need to go to the office, in turn, it is likely many stayed at home to perform their job. Also, when the pandemic hit the impact had 3x as many people working from home from 2019 to 2021 which severly impact the amount of people riding the subway. That means a high majority of the people riding the subway before the pandemic and after view it as an inelastic good, so despite the virus spreading, people have no choice but to ride on the subway. Public transportation is directly tied to the labor force, and more specifically the distinction between in-person or online work.

Conclusion

While there was not a significant effect on ridership entries by COVID-19 on city, state, or national level after accounting for fixed effects, it is clear that the public transportation industry has changed long term. Currently, research has been highly limited to a small sample on COVID data and unknown about the rebound effects of the pandemic as it hasn’t been long enough since the threat of COVID actually decreased. In addition, there has not been any scholarly study or reggresional analysis regarding subway ridership in New York City. Other ridership studies have usually concluded the other way however, finding a significant correlation between public transportation (or private in the case of the Seattle study) and COVID cases. Many factors such as geographical significance, fear of COVID, and workers choice all factor into each of these studies and this study so it it is impossible conclude a clear brightline quite yet. Because of the few amount of articles, many might not view the data in-depth, instead relying on baseline interpretations and logically claims rather than empirical analysis and evidential claims which risks the legitimacy of the entire public ridership industry.

While the paper was unable to find a true answer to is the industry financially stable, that would be the next step. This idea would be less related to COVID and could depend on many other factors, but it would be interesting to see the results. If the results show that the industry is unsustainable, it would definitely indicate that there is some external factor that might actually have the effect we believed COVID would. Based on the study, a potential conclusion is that ridership in abstract is an inelastic good, and, within that, the subway is one of the few sources nearly everyone can afford. 

The specific COVID data was found on NYC.gov, which in terms of creditablity, is the best possible as it is the New York government recording on cases, it is possible some political bias could exist, but it is unlikely it would push them to creating fake statistics. As for the ridership data, it was found from NYC.go transit which is also extremely creditable; however, during a difficult time like COVID, it is possible to give the impression that the subway was thriving as to not get rid of funding, some data could not be accurate, however there is no proof of this and their methods of calculating ticket sales is very trustworthy.

The two main causes I believe are vaccination rates and the inelastic demand from public transportation, and they can even be applied to one idea. What I believe happened is as more and more people received the COVID vaccine, they began traveling more, for work, pleasure, or any sort of extra factor. Since this happened relatively early in 2021, it had a major effect on the correlation between case counts and ridership. People believed, even though case counts increased, they were produced from the virus and traveled freely. While for private transportation it might not be the case, since a large portion of the population must use the subway for their livelihood, it definitely influenced the ridership data the study accounted for. Future studies must evaluate the depths of the literature past 2021, specifically how 2022 and 2023 differ and if as more powerful vaccines and new.

The findings discussed can be generalized to the entire transportation industry. It seems like industries that previously took a big hit from the pandemic are now less so, and the world is starting to revert back to previous forms before COVID disrupted most parts of life. The world is moving away from the pain suffered in the last two years to a more lively world where people can live life without such a large concern regarding COVID. While public transportation does not show all aspects of life, it does represent a shift and change that the world has taken, it is now more free and available for everyone to revert to their previous life before the two year shutdown.

Acknowledgements

Thank you to Eric Donald for helping with every step of the process in writing the paper as my mentor. 

Thank you to Lumiere Publication Program for having detailed schedules and assistance when it came to writing my paper.

Thank you to Calvin Edwards for reviewing my paper and constantly helping me correct issues that existed.

References

  1. MTA: NYC Transit Authority’s annual ridership by division 2021, https://www.statista.com/statistics/1294015/new-york-city-mta-network-total-annual-ridership-by-division/ []
  2. Z. Bian, F. Zuo, J. Gao, Y. Chen, S. S. C. Pavuluri Venkata, S. Du-ran Bernardes, K. Ozbay, X. J. Ban and J. Wang, Transportation Research
    Part A: Policy and Practice, 2021, 145, 269–283. []
  3. D. Wang, B. Y. He, J. Gao, J. Y. Chow, K. Ozbay and S. Iyer, International Journal of Transportation Science and Technology, 2021, 10, 197–211. []
  4. COVID-19: Latest Data – NYC Health, https://www.nyc.gov/site/doh/covid/covid-19-data.page. []
  5. Beyond Rush Hour, https://comptroller.nyc.gov/reports/beyond-rush-hour/. []
  6. COVID-19: Data on Vaccines – NYC Health, https://www.nyc.gov/site/doh/covid/covid-19-data-vaccines.page. []
  7. U. C. Bureau, The Number of People Primarily Working From Home Tripled Between 2019 and 2021, https://www.census.gov/newsroom/press-releases/2022/people-working-fromhome.html. []

LEAVE A REPLY

Please enter your comment!
Please enter your name here