Yes, in this instance, perception matches reality.

This is interesting data and relevant to a Bayless household discussion topic.  My wife has insisted for some months that over the past eighteen months of Covid-19, while traffic volume is down, the crazy driver ratio is up.  

I am in agreement that the crazy driver ratio is probably up but being perhaps overly alert to recency bias, confirmation bias, etc. I have been more moderate in my assessment.  Perhaps drivers aren't quite as bad as they seem.

Discussion has also centered on the why.  If driving is worse, why?  I hear a lot of suggestions that it is a rusty skills issue - people haven't been driving as much and they have lost their edge.  Then there is the open roads claim - as the roads lightened up from normal volumes, people adjusted their speeds to take advantage.  Boredom is another popular mainstream media claim.

All likely contributors.  My suspicion has also been that the George Floyd protests and the reigning in of police departments everywhere (both explicitly and even more frequently indirectly) has been detrimental to traffic safety.  If police are not stopping speeding cars, you get more speeding.

Now we have some answers.  From Deadly Crashes in The US Surged During The Pandemic. This Could Be The Explanation by Peter Dockrill.  Yes, the increase in dangerous driving is both real and substantial and yes, there are some strong candidates for why.  Not among those listed above.

Despite Americans driving hundreds of billions of miles less in 2020 than they did in 2019, deadly traffic accidents actually surged in the year of lockdown, with crash fatalities during the pandemic hitting levels not seen in over a decade.

According to the US Department of Transportation's National Highway Traffic Safety Administration (NHTSA), traffic fatality projections for 2020 suggest an estimated 38,680 people died in motor vehicle traffic crashes last year, representing an increase of about 7.2 percent on 2019's fatalities.

That staggering rise – taking traffic fatalities to levels not seen since 2007 – came during a year in which Americans actually drove their cars much less than usual: an estimated 430.2 billion miles less traveled overall, representing a 13.2 percent decrease from the year before.

The new figures show that traveling on the nation's roads last year was a significantly riskier proposition overall compared to recent years, with the fatality rate per 100 million vehicle miles traveled (VMT) surging to 1.37 in 2020, up from 1.11 in 2019, and ending a string of yearly reductions in fatality rates that began in 2016, suggesting safer driving conditions.

A 7% increase in fatalities in a single year?  Yikes.

"Of the drivers who remained on the roads, some engaged in riskier behavior, including speeding, failing to wear seat belts, and driving under the influence of alcohol or other drugs."

[snip]

While it's difficult to authoritatively pin down what made people drive in more dangerous ways, the NHTSA acknowledges a "remarkable trend" of several worsening factors contributed to the surge in deadly crashes last year.

These included significant increases in fatalities on rural, urban, and interstate roads, fatalities at night-time and on weekends, crashes involving older vehicles, rollover crashes, speeding-related crashes, and many more.

The greatest increases over 2019 figures were fatalities involving non-Hispanic Black people (up 23 percent), occupant ejection (up 20 percent), and unrestrained occupants of passenger vehicles (up 15 percent, and tied with fatalities on urban interstate roads). 

I am still guessing that a reduction in roadway policing is a material factor but the data below looks like an obvious culprit.

"NHTSA's study of seriously or fatally injured road users at five participating trauma centers found that almost two-thirds of drivers tested positive for at least one active drug, including alcohol, marijuana, or opioids between mid-March and mid-July," a new report explains.

"The proportion of drivers testing positive for opioids nearly doubled after mid-March, compared to the previous six months, while marijuana prevalence increased by about 50 percent." 

50-100% increase in drug use among drivers?  Woof!  That seems relevant to the discussion of causation.

Interestingly, if you got to the original NHTSA data, alcohol involved accidents are up only 9%.  Yes, an increase, but nothing like that of drugs.  

I am struck by how little attention this report is getting in the mainstream media.  Or, at least, I am not seeing it.  

Even more astonishing is that statistic - an increase in African-American driving fatalities of 23% as part of the overall 7.2% increase in fatal crashes.  7.2% increase in a socio-econometric measure in a single year is pretty unusual.  A 23% increase in fatalities in a single year?  Astonishing.  

Like any good empirical, this moves you on to the next line of questioning.  My two biggest are:

Why the far greater increase in drug associated accidents (50-100% increase) versus the much small increase in alcohol associated accidents (9%)?

What is happening among African-Americans to cause an increase of 23% in fatalities?  Is it an age demographic factor (younger drivers)?  Are African-Americans more dependent on driving than others and therefore became a larger share of drivers?  Both these seem unlikely.  I don't have a good answer.