Have you ever noticed that virtually everything about traffic planning is counter-intuitive? Mostly, that is because the real world works opposite the way traffic engineering operates.
Here is a pretty fascinating study by the HSIS (Highway Safety Information System) looking at road diets from Iowa specifically (15 sites) as well as around the country (30 sites in addition to the 15 in Iowa). Predictably, "accidents" (shall we say collisions) dropped on average from 23.74 to 12.19 yearly accidents per mile averaged from the Iowa sites and 28.57 to 24.07 on the nationwide HSIS sites. Though the information isn't provided regarding average speeds before or after, we can hypothesize that the severity of the accidents diminished as well.
More interestingly however, is that average daily traffic counts actually INCREASED when reducing the amount of lanes on a road. On the dieted Iowa roads, traffic (vehicular only) increased from 7,987 to 9,212 cars per day. On the 30 HSIS dieted roads, traffic increased from 11,928 to 12,790. More traffic, slower roads, safer.
The question is why? The key to answering that question is a simple understanding of human psychology and how it influences urbanism. Lots of pavement and fast moving cars (potential danger - not to mention exhaust fumes, noise, etc.) are repulsive. They are sociofugal spaces. People want to get away from them as quickly as possible. However, safer, more attractive places attract more people, even if they have to drive to get there.
More traffic = more value*. (*Disclaimer: provided that said traffic is not fast moving and entirely automobile-based.)
Investors look at traffic counts. Nobody wants to put their business at place nobody sees or wants to go. The dirty little secret of traffic planning and engineering is that roads that get an A grade are well below capacity. In the traffic engineering world, that is a good thing. Cars are free to drive as fast as they want without pesky other people getting in the way. Because in the traffic engineering world, roads are designed for the optimal condition of one singular person on the road. In the real world, we give these roads Fs.
Because of dendritic nature of modern traffic planning, funneling traffic towards larger and larger roads, we end up with a condition where certain roads are terribly congested. But for every one congested road there are 100 roads that are nearly abandoned. Six lane roads with a capacity of 44,000 vehicles a day carrying 5- or 10,000 cars per day (I'm looking at you Hatcher St in South Dallas).
Nearby Metropolitan Ave has a capacity of 31,000 but carries 3,000 cars per day. MLK Blvd is a six lane road, but at its current traffic counts could handle being a 2-lane road. Between Fair Park and the Trinity River, that's 12 acres of unnecessary pavement (And we wonder why cities are broke. Thank a traffic planner). But, they get an A grade. Counter-intuitive indeed.
The sacrosanct formulae we mindlessly obey from traffic engineering leads to two and only two types of environments, invaded places and abandoned ones. Neither is attractive. To people nor investment. Hence, the slow, inevitable decline of areas where widenings occur. Disintegration = disinvestment. This hasn't been deemed a Law of Urban Dynamics, but you might as well chalk it up as one. Nay, write it in ink. Because it is fact.
We're by nature a social species, searching for other people. Where are people going to meet? At safe, attractive places. Sociopetal places. Convergence points in the human network. Therefore, highly integrated, interconnected places should ideally have the highest level of design and cultivation as anthropocentric places.
If Mayor Rawlings wants to grow South Dallas, he'll start by reducing the size of all the roads and increasing the interconnectivity of the disintegrated grid.
This concludes your daily evisceration of an entire profession.