Discrepancy in Average Age of Automobile on the Road?

[dorsiacapital]

Hello,

 

I think I'm missing something obvious, but I've been struggling to reconcile the following figures. Any help would be much appreciated.

 

Average Age of Automobiles on the Road 

(http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_statistics/html/table_01_26.html_mfd)

 

2010: 10.6

2011: 10.9

2012: 11.2

2013: 11.4

2014: 11.4

 

Total Highway Registered Vehicles

http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_statistics/html/table_01_11.html

2009:  254,212,610

2010: 250,070,048

2011: 253,215,681

2012: 253,639,386

2013: 255,876,822

2014: (?)

 

New Automobile Sales

https://research.stlouisfed.org/fred2/series/ALTSALES

2009: 10.4

2010: 11.55

2011: 12.735

2012: 14.43

2013: 15.53

2014: 16.435

 

So, here’s my question.  Between 2009 and 2013 (no 2014 data), the totalhigh way registered vehicles stayed basically flat at 255 million.  From 2009 to 2013, a total of 81 million new vehicles were sold (average blended age of about 2.5 years), and yet, the average age of the automobiles on the road actually increased by a year. 

 

This doesn’t seem to be possible. All of the new vehicles being added to a vehicle population that stays the same size should drive down the average age of the automobile on the road significantly. What am I missing?  (I’m very curious about this because the continuing increase in average age of automobile on the road has been repeatedly cited as a source of strength/opportunity for the auto industry…).

 

A few minor points

 

Total highway vehicles contains a mix of things including motorcycles and heavy duty trucks – I checked and each category appears to stay basically constant over the relevant period.

 

More anecdotally, vehicle miles driven and vehicles per capita has remained flat for the last few years (and declining a bit from car usage in the early 2000s).  These anecdotes could be wrong, however, and American car ownership could be increasing per capita.

 

http://usa.streetsblog.org/2013/05/14/millennials-will-drive-more-as-they-age-but-still-less-than-their-parents/

 

http://usa.streetsblog.org/2013/06/21/has-america-already-hit-peak-car/

 

[LongHaul]

Might be from the 17.5m rate of vehicle sales from ~2000-2007.  As those cars age each year they add to the average age.  Then there was a steep drop off in 09 and 10 which is more recent. 

 

It depends on the mix of volume and age by year.  That is my guess anyhow.

 

[TwoCitiesCapital]

Hello,

 

I think I'm missing something obvious, but I've been struggling to reconcile the following figures. Any help would be much appreciated.

 

Average Age of Automobiles on the Road 

(http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_statistics/html/table_01_26.html_mfd)

 

2010: 10.6

2011: 10.9

2012: 11.2

2013: 11.4

2014: 11.4

 

Total Highway Registered Vehicles

http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_statistics/html/table_01_11.html

2009:  254,212,610

2010: 250,070,048

2011: 253,215,681

2012: 253,639,386

2013: 255,876,822

2014: (?)

 

New Automobile Sales

https://research.stlouisfed.org/fred2/series/ALTSALES

2009: 10.4

2010: 11.55

2011: 12.735

2012: 14.43

2013: 15.53

2014: 16.435

 

So, here’s my question.  Between 2009 and 2013 (no 2014 data), the totalhigh way registered vehicles stayed basically flat at 255 million.  From 2009 to 2013, a total of 81 million new vehicles were sold (average blended age of about 2.5 years), and yet, the average age of the automobiles on the road actually increased by a year. 

 

This doesn’t seem to be possible. All of the new vehicles being added to a vehicle population that stays the same size should drive down the average age of the automobile on the road significantly. What am I missing?  (I’m very curious about this because the continuing increase in average age of automobile on the road has been repeatedly cited as a source of strength/opportunity for the auto industry…).

 

A few minor points

 

Total highway vehicles contains a mix of things including motorcycles and heavy duty trucks – I checked and each category appears to stay basically constant over the relevant period.

 

More anecdotally, vehicle miles driven and vehicles per capita has remained flat for the last few years (and declining a bit from car usage in the early 2000s).  These anecdotes could be wrong, however, and American car ownership could be increasing per capita.

 

http://usa.streetsblog.org/2013/05/14/millennials-will-drive-more-as-they-age-but-still-less-than-their-parents/

 

http://usa.streetsblog.org/2013/06/21/has-america-already-hit-peak-car/

 

All things being equal - the average age of the entire fleet would increase by 1 year every year. Each year ~17M old cars will come out of the fleet as they are retired and 17M new cars will be added - but that's still only 34M vehicles acting towards bringing down the average while the remaining 221M act to take it up by a year.

 

It makes sense that the average age is still rising despite new car sales given this dynamic. You'd need to get to the point where we are seeing signifcantly more sales/retirements to start bringing that average down.

 

 

[Jurgis]

Dorsia,

 

You are thinking about this incorrectly I believe.

 

Take 255m cars. Let's say 16M cars enter the population, which means that 16M cars exit the population. The rest (255M-16M=239M) age 1 year. So you could do back of napkin computation (239M * (11.4+1) + 16M * 1) / 255M to get the new average. This is actually not correct in general, since the average will be affected by which cars exit the population. If really old cars exit, it reduces the average, if newer cars exit (through accidents I guess), this increases the average. However, we don't have that data, so we have to use this dirty method.

 

Now, for average age to go down, you have to solve for X in ((255M - X) * (11.4+1) + X * 1) / 255M < 11.4

This solves to X > 22M

So for average age to go down, you need at least 22M of sales in a year.

 

This is quick and dirty, I know caveats, so likely you need less than 22M of sales for average age to go down. But likely more than 16M. ;)

[rkbabang]

Dorsia,

 

You are thinking about this incorrectly I believe.

 

Take 255m cars. Let's say 16M cars enter the population, which means that 16M cars exit the population. The rest (255M-16M=239M) age 1 year. So you could do back of napkin computation (239M * (11.4+1) + 16M * 1) / 255M to get the new average. This is actually not correct in general, since the average will be affected by which cars exit the population. If really old cars exit, it reduces the average, if newer cars exit (through accidents I guess), this increases the average. However, we don't have that data, so we have to use this dirty method.

 

Now, for average age to go down, you have to solve for X in ((255M - X) * (11.4+1) + X * 1) / 255M < 11.4

This solves to X > 22M

So for average age to go down, you need at least 22M of sales in a year.

 

This is quick and dirty, I know caveats, so likely you need less than 22M of sales for average age to go down. But likely more than 16M. ;)

 

Yes this is the way to think about it.  I would think that with cars being totaled in accidents there would be a skew towards older vehicles (or at least cheaper ones), because newer vehicles (and more expensive vehicles) will tend to have better accident avoidance features. And maybe more important older/cheaper vehicles need considerably less damage to consider them totaled.  For instance if a vehicle is only worth $1500 it wouldn't take much of an accident to make it not worth fixing, where if it is worth $35K it would be fixed after all but the worst accidents.

 

[dorsiacapital]

Dorsia,

 

You are thinking about this incorrectly I believe.

 

Take 255m cars. Let's say 16M cars enter the population, which means that 16M cars exit the population. The rest (255M-16M=239M) age 1 year. So you could do back of napkin computation (239M * (11.4+1) + 16M * 1) / 255M to get the new average. This is actually not correct in general, since the average will be affected by which cars exit the population. If really old cars exit, it reduces the average, if newer cars exit (through accidents I guess), this increases the average. However, we don't have that data, so we have to use this dirty method.

 

Now, for average age to go down, you have to solve for X in ((255M - X) * (11.4+1) + X * 1) / 255M < 11.4

This solves to X > 22M

So for average age to go down, you need at least 22M of sales in a year.

 

This is quick and dirty, I know caveats, so likely you need less than 22M of sales for average age to go down. But likely more than 16M. ;)

 

Thanks all, I was looking for a quick and dirty formula but couldn't quite think of it. I still plan to play with the math a bit, but it makes sense. (I would imagine that most of the cars being replaced are the old ones, but certainly accidents could play a big factor too....).