I too "feel" that I see mostly newer vehicles on the road; however, the data does not seem to support my observations. My source, however, is Wikipedia, so take it with a grain of salt (
https://en.wikipedia.org/wiki/Automotive_industry#World_motor_vehicle_production), but the data there suggests that, excluding China, Korea, and India, which have experienced massive recent growth, the amount of vehicles produced per year over the past 20 years has remained relatively steady. If cars were failing sooner, then consumption would have to increase, unless people were surviving with fewer cars, which is, of course, a possibility. I remember in the early 2000's (based on observations of my friends' families) it was not uncommon for families to have more vehicles than drivers in the household. I don't see this happening as much any more, but of course, my observations are not quantified, and they are limited to my locale, so they are heavily biased. Another thing to consider is that where I live, we have very strict state inspection regulations. I used to live in MD, which only required an initial state inspection upon vehicle purchase (no recurring state inspections). I saw many older vehicles driving around (usually with parts falling off, mind you). I don't see that here in New England.
There are many things to consider. Perhaps cars are failing sooner and people are buying fewer cars (which is one possibility that could support the argument that cars are failing sooner while the world production rate remains relatively steady). Perhaps cars are not failing sooner and people are buying cars at a rate similar to that of 20 years ago. The one important thing to consider is that in addition to vehicle operation, vehicle production also contributes to pollution. Factories require electricity, raw materials (mining), recycled materials (processing). We, as a civilization, need to ensure that we don't get tunnel vision. Creating more efficient vehicles that fail more quickly may not necessarily reduce pollution; the pollution generated by the manufacture of vehicles used to replace failing vehicles may very well offset the reduced emissions from the vehicles themselves. Of course, I don't have any concrete data to support one side or the other; I'm just spitballing here.