Journalists in older cities like New York, Boston or San Francisco may see the role of rail transit as critical to a functioning modern city. In reality, rail transit has been a financial and policy failure outside of a handful of cities.

In 23 metropolitan areas that have built new rail systems since 1970, transit’s share of commuting — including all forms, such as buses and ferries — has actually slipped a bit, from an average of 5.0 percent before the rail systems opened to 4.6 percent in 2013. The ranks of those driving alone continue to grow, having increased 14.4 million daily one-way trips since 2000, nearly double transit’s overall daily total of 7.6 million, according to Census Bureau data.

Virtually all the actual increase in rail commuting has occurred in the “legacy cities”: New York, Boston, San Francisco, Washington, Chicago and Philadelphia. These are older cities built around well-defined cores that were developed mostly before the automobile. Together the core cities of these metro areas, excluding the suburbs, accounted for 55 percent of all transit work trips in the nation in 2014, according to the latest American Community Survey data. Overall, transit’s work trip market share in these six metropolitan areas rose from 17 percent to 20 percent between 2000 and 2014. In the entire balance of the country, where most of the new rail systems have been built, transit’s market share is only 2.2 percent, up a scant 0.2 percentage points since 2000, according to Census Bureau data.

Manhattan alone, in fact, accounts for more than 40 percent of all rail commuters in the nation. New York is the only U.S. city where more than 20 percent of workers labor in the central business district (downtown). In most cities, the percentage is less than half of that, and in many others, even smaller. In Los Angeles, less than 3 percent of employment is downtown. In Dallas only 2 percent of metropolitan employment is downtown. In Houston, where numerous large companies maintain headquarters, it’s still only 6.4 percent.

For transit to work effectively, employment needs to be concentrated. This explains why between 2013 and 2014, New York accounted for a remarkable 88 percent of the total increase in train commuting. But what works for Brooklynites headed to Union Square does not generally work so well for people living in our increasingly dispersed metropolitan areas. Indeed in most cities — Dallas-Fort Worth, Houston, San Diego, and even the new urbanist mecca of Portland, according to 2015 American Community Survey data, where new transit lines have been put in, it has failed to increase the share of commuters who take public transportation, and in some cases the actual ridership has dropped.

It has even failed where cities are booming and their downtowns flourishing. Houston’s light rail system opened in 2004, but has done little to change the car-dominated commuting pattern of America’s energy capital. Between 2003 and 2014, Harris County’s population grew 23 percent, but transit ridership decreased 12%, according to American Public Transportation Association data. This means that the average Houstonian took 30 percent fewer trips on the combined bus and light rail system in 2014 than on the bus-only system in 2003.

The Next Great Transit City

Nowhere is the transit mania more profound than in Los Angeles, a city progressive blogger Matt Yglesias describes as “the next great transit city.”

There seems to be a conscious strategy of making auto commuting in Los Angeles and the rest of California so unpleasant as to force people into transit. Mayor Eric Garcetti has made bold predictions that commuting times will drop in half, largely by people moving from cars to trains. Of course this is folly, since transit commuting generally takes considerably longer than commuting by car. The Governor’s Office of Planning and Research has called for putting all California on “a road diet,” meaning that traffic will simply continue to worsen. This in a state which has among the worst roads in the country – 68 percent of which are in poor or mediocre condition.

Can rail solve or mitigate congestion? L.A. has already spent over $15 billion on rail yet this has proven less than effective in either boosting transit ridership or lessening congestion.

Since 1980 before the rail expansion the percentage of Los Angeles County commuters who take transit has actually dropped from 7.0 to 6.9 percent while the transit share of the combined statistical area has dropped from 5.1 to 4.7 percent. Even the total numbers of riders is heading down. Recently the transit booster Los Angeles Times published statistics that showed that there were now 10 percent fewer boardings on the Los Angeles MTA system than in 2006, and that the decline was accelerating.

One reason for the poor performance is that much of the train ridership turns out to have been former bus travelers in the first place, which limits actual gains there. Taxpayers, however, should be screaming about this switchero; the subsidy for new L.A. new bus riders, who tend to be the poorest of the poor, cost taxpayers $1.40 while the cost for a new rail rider was $25.82 over the period of 1994 to 2007. If you believe in transit as public good, clearly building more trains makes less sense than expanding bus operations.

But it’s not just a cost issue. Los Angeles is a vast and dispersed metropolis in which only one in five residents even lives within the city limits, and even much of the city — notably the San Fernando Valley — is essentially suburban in form. Transit travel takes much more time to get to work than the car, even on the region’s miserable roads and overcrowded freeways. With downtown only a minor employment center, people increasingly travel there for cultural events, sports or even a restaurant, not for work.

Other factors also seem to be contributing to the decline. One is the trend toward working at home; in 2014, the number of Angelinos working at home surpassed the number taking transit. Although this saves more energy, and produces less carbon than transit ridership, there is virtually no government support for this innovative approach to traffic reduction from the climate-obsessed state government.

Finally, there are now other options such as Uber and Lyft, which provide reasonably priced door to service, always available, often on short notice. Down the road, the path for transit looks even bleaker with the development of self-driving cars, which will make even long suburban commutes easier. Looked at objectively, the drive for a traditional transit dominated Los Angeles is on a collision course with reality.

 Taking Stock and Changing our Approach

In the alternative world that dominates our transit planners and retro-urbanists, nothing succeeds like failure. Some urban experts still predict that the Sun Belt cities are ripe for a huge infusion of rail transit, despite all evidence to the contrary.

Given what we know about the share of commuters using transit in most cities, pumping money into this form of transportation seems doubly wrong while other needs such as roads, schools, sewers and parks are neglected. Rather than try to fit all cities, and all parts of metropolitan areas, into a 19thcentury technology, maybe we should look to encourage 21st century innovation.

Clearly some of this is already with us, notably in the rise of services like Uber and Lyft which, for many, seems a far more effective way of getting around with your own car. Ride-sharing and services like Zipcar also provide new alternatives. And other innovations could be developed, with expanding shuttle and dial-a-ride services. In many big cities dedicated commuter buses, connecting the dispersed employment centers, would make great sense in cities such as Houston, which has many large employment centers, notes my Center for Opportunity Urbanism fellow, Tory Gattis.

But it’s changing work patterns that may provide the most promising opportunities to reduce traffic and reduce greenhouse gases. In the U.S., working at home, not transit, was the principal commuting alternative to the automobile in 39 of the 53 major metropolitan areas with populations over 1 million as of 2014, according to Census Bureau data. The share of work access accounted for by home workers rose by more than a third between 2000 and 2014, from 3.3 percent to 4.5 percent

Many of the most striking work at home share gains are taking place in the country’s leading technology regions, including Austin, Raleigh, the San Francisco Bay Area, Denver, Portland and San Diego. Millennials in particular, notes a recent Ernst and Young study, embrace telecommuting and flexible schedules more than previous generations did, in large part due to concerns about finding balance between work and family life.

All this suggests we need to revamp our ideas of transit, particularly in the newer, fast-growing cities. Trains may elicit a nostalgic smile about the good old days, but most Americans, and the vast majority of our cities, need to live not in the past but in an increasingly dispersed, and choice-filled reality. Time to embrace that future.

Cross-posted at New Geography.