Quarterly Report on Traffic Developments

Overview

Recent Developments

The Overview section of this report highlights recent developments in data describing changes in traffic patterns, insights, and overall road safety as of December 31, 2020. We highlight changes in the known main safety drivers: traffic levels, accidents volumes and severity, driver behavior, and economic activity. The reduction in traffic due to the COVID-19 pandemic presents an interesting ‘natural experiment’ to analyze traffic and accident relationships.

Traffic Decreased 40% YoY in April 2020, Rebounding to ~10% Drop by December

• 2020 Traffic levels were similar during January and February to 2019, dropping in March, and plunging dramatically in April year-over-year (YoY) as initial Shelter In Place (SIP) orders were instituted. By vehicle miles traveled (VMT) on tracked road segments, all systems together saw an average -40.1% decrease during that month. Subsequent months saw a gradual lessening of the decrease, from -25.5% in May to -10.3% by December 2020. The least severe decrease was in September (-8.4%)[2]. 

• There were some variations by drive environment. The most severe nationwide drops were seen on urban interstates in April 2020 (-44.8% YoY) with rural non-arterial non-interstate roads seeing the least severe drop at -33.3%. This trend continued through most of the year, with rural road driving (-4.7%) propelling the rebound in September overall traffic to -8.4%. October to December saw levels drop across the board and the divergence lessen, with December showing an overall -12.0% drop in Urban travel and -9.6% in Rural travel[2]. 

• State monthly traffic largely followed national trends, with deviations related to local environment (see Supplemental Table 1 for state breakdowns). Highly urbanized, populated states followed similar trends, while there were stark differences in predominantly rural states. For example, tourist-driven economies (eg Hawaii, VMT -13.9% in November) contrasted sharply with states with less tourist dependence (eg Idaho, VMT -0.5%)[2].

•  2020 cumulative VMT was 2.8 trillion miles, -13.2% (430 billion miles) from 2019, the lowest VMT level seen since 2002, when the US population was 14% smaller (Fig 1)[2].

Figure 1: Yearly Cumulative Vehicle Miles Traveled (Trillions in miles)

Stay At Home Population Surged to 25% of Population, 81.2M People on Average Day

• Bureau of Transportation Statistics (BTS) collects and estimates detailed trips data (number and length of vehicle trips) from various sources like traffic counters, cell phone data, and others [3]. Prior to COVID-19, BTS estimated that ~18-20% of the population was homebound on any given day (zero vehicle trips), with the homebound population dropping in August and peaking in December [3]. COVID-19 quickly changed this trend, as the March shelter-in-place (SIP) population rose to 24% and peaked at 28% in April. May through October saw the SIP population oscillate between ~24% and 26%, dropping to ~29% in November and December. For the year, roughly 25% of the population was at home on any average day (Table 1)[3].

• While a relatively moderate relative increase in SIP, BTS estimated the total observed population to be 327.1M, thus generating rather large absolute population changes. 2019 daily average population at home came out to 63.4M people, while 2020 averaged 81.2M, peaking at 94.9M in November[3].

• Trips data from BTS also indicates the magnitude of trip volume changes. In 2019, Americans averaged 1.35B trips a day, while in 2020 that dropped to 1.00B trips a day. Five months in 2020 saw less than 1.00B trips a day, with April plummeting below 900M trips a day, a reduction of over half a billion trips per day less than on the average 2019 April day (Table 2, following page)[3].  

Total Accidents Down Sharply

• Accident data collection and distribution were hit hard by complications from COVID-19 and are being backfilled.  Preliminary data from 11 states is available but caution must be used, as these numbers could be conservative and adjusted as additional data comes in. Nonetheless a common trend in the data showed reduced accident volume year over year in December 2020 compared to 2019 (Table 3)[various state sources, 4-14]. Interestingly, in 8 of 11 states, accident volume drops outpaced traffic volume drops (i.e., modest YoY drops in VMT of ~ -4-16% led to accidents volumes down ~ -11-55%).

• 3 of the 11 states (Ohio, Utah, Vermont) did not follow the trend of accident volumes outpacing the drop in VMT in December 2020. Ohio’s VMT drop (~ -13%) was slightly below the concordant accident volume drop (~ -11%), while Utah (VMT -7.2%, accidents +1.7%) and Vermont (VMT -19.1%, accidents +5.8%) showed elevated accident volumes despite VMT drops. However, in November, all three states showed accident volume drops outpacing VMT, inline with the larger December trend (Table 4)[10,12,13]. Statistical noise in state data collection and/or weather-related effects may be behind the divergence, with further research on finalized data needed in the future.

Economic Indicators Dropped Consistent with Reduced Traffic and Accident Levels

• Detailed research from the International Transport Forum (ITF) arm of the Organization for Economic Cooperation and Development (OECD) following the Great Financial Crisis (GFC) indicated that hard economic times correlate with increased traffic safety across the board. Research suggests that economy-related drops in high risk driving populations (particularly young and/or intoxicated drivers) propel much of the rise in safety, with wider related effects seen from reduced trips and conservative driving styles to conserve gas and mitigate expenses [15].

• COVID-19 produced an economic shock far greater than the GFC, in both size and time of onset. Unemployment rose sharply across the US and remained elevated. In January 2020 state unemployment averaged 3.5% of the workforce and ranged from 2%-5.6% (see Supplemental Table 1 for state breakdown); by January 2021, average state unemployment was 5.7% and the range had grown from 3.2%-10.2%. Recovery continues unevenly, with some states seeing only a modest 7 percent relative rise in January YoY Unemployment Rate (South Dakota, from 2.9 to 3.1), while others have seen their rate double (eg North Dakota, 2.2 to 4.5) or quintuple (eg Hawaii, from 2.0 to 10.2). In total, over 10M jobs remain missing from the workforce over the period January 2020 to January 2021 [16]. Additional data in Supplemental Table 1.

• The general trend of traffic safety dropping in line with economic growth seems to be holding for COVID-19. All states have seen increased unemployment alongside lowered VMT and accident volume drops in states where preliminary counts are available. States with available accident data show the bulk of the drops coming from accidents involving property damage only (PDO accidents, commonly called ‘fender benders’) or involving minor injury. New York saw roughly 135k less accidents, from 419k to 283k, with PDO accounting for 101k of the drop. Florida likewise saw a reduction of 61k accidents, from 402k to 340k, with PDO accidents dropping by ~35k. Similarly, Ohio saw a reduction of 51k accidents, from 297k to 245k, with PDO accidents dropping by ~41k. Given the preliminary nature of these counts, it’s prudent to use caution in overinterpreting them. Nonetheless there is a clear trend of accident volume dropping, largely concentrated among PDO-related occurrences [5,7,9,10].

Fatality Levels Rise, Speeding and Seatbelt Misuse Likely Culprits

• One disturbing trend is the reported rise in fatalities and fatal crashes during 2020. Previous recessions like the GFC generated noticeable drops in fatalities (compare 2008 to subsequent years, figure 2 below; see also research in 15) alongside drops in VMT (see trough in earlier Fig 1, pg. 2). Preliminary 2020 data from BTS, National Highway Transportation Safety Administration (NHTSA), and National Safety Council (NSC) showed a notable elevation in fatalities, both in raw numbers (from 39,107 fatalities in 2019 to 42,060 in 2020) and especially in VMT-normed fatalities per 100M VMT (quarterly ranges of 1.2-1.6, up from 1-1.1 in 2019). In March (-186 people; 1.25 per 100M) and April (-559 people; 1.49 per 100M) absolute fatality rates dropped YoY compared to 2019, but normed fatality rates began climbing above 1.1 as VMT drops outpaced drops in fatalities. Subsequent months saw a surge in absolute fatalities compared to 2019, with October recording 694 more deaths than seen in 2019 [17-20].     

• Several possibilities may explain the rise in fatalities, among them a rise in intoxicated drivers or other reckless behaviors, like speeding and seatbelt misuse. COVID-19 differs from earlier recessions in that traffic volume dropped much more dramatically, suggesting that the rise in fatalities may be related to this altered ‘traffic normal’. Preliminary data analyzed by NHTSA indicates that driver speed increased markedly on monitored roads, particularly in urban areas. Risky behavior rose in other areas as well, including seatbelt misuse and driver intoxication, particularly among males [19]. Preliminary crash data (not shown) from several states does not indicate a direct causal rise in intoxicated operators causing the fatality increase; however, this may be an artifact of the data collection lag and awaits further study.