The effect of self-driving lorries on BNSF

[LightWhale]

Attached is an article from today's FT about autonomous lorries, which should become widespread well before private cars.

This will reduce shipping costs dramatically, and I'm wondering how it may affect Burlington's pricing power. If I'm not mistaken, BNSF contributes about 15% of BRK's net earning and above 12% of book value, so any effect on BNSF can be dramatic for the holding-co.

Curious about your thoughts on this one, TIA.

Self-driving_lorries_to_be_tested_in_UK_next_year.pdf

[gfp]

I always wondered how one is supposed to merge into the interstate if self driving platoons of road trains are following closer than would be safe if humans were driving.  As for BRK, by the time this materially impacts BNSF, BNSF will be a much smaller percentage of BRK. It's just the nature of the compounding machine.

 

And we really don't know what these road trains will be asked to pay for their use of the public infrastructure, vs BNSF's privately maintained infrastructure.

 

What if 20 years from now autonomous lorries are a thing and $1.5 trillion market value Berkshire Hathaway (or a per-share equivalent gain in shareholder wealth through repurchases) is only reporting $400 million a quarter in profit from BNSF, instead of a billion. Is that a disaster?

[Jurgis]

OTOH, even if you have self-driving trucks, the likely transportation model will be something like https://en.wikipedia.org/wiki/RORO for long distance onto BNSF and then self-drive remaining shorter distance. Why not?

[rb]

Something tells me that the idea won't be so popular in the US for several reasons.

 

1. The US is the land of the concerned mothers of something or other. This is why I still have to drive at 50 mph on lot of freeways in the US. 40 ton behemoths rolling at 50 mph by themselves is surely going to concern a lot of mothers.

2. You have the political BS. Lots of truck drivers in the US.

3. The whole industry structure would have to change. Currently in the US the majority (90%?) of trucks on the road are owners/operators. Now a trucker may be able to buy/finance and operate a truck. But the independent trucker is surely not going to be able to buy/finance a convoy. The cost savings may not be big enough to trigger a complete redesign of the industry.

4. While the self driving stuff will decrease the price gap between truck and rail, rail will still have a cost a cost advantage.

 

All this stuff sounds cool. But really I think that the shift away from coal represents a much bigger and present risk for BNSF (all rails really) than self driving vehicles.

[LightWhale]

by the time this materially impacts BNSF, BNSF will be a much smaller percentage of BRK. It's just the nature of the compounding machine.

 

Good point. the WFC position illustrates it well.  WEB hasn't sold much since the scandals popped up, but through investing in other positions, WFC's weight in the portfolio halved.  However, implicit in my remark about private cars was that Geico could get battered at the same time. So these are two major revenue sources for BRK.

 

self-drive remaining shorter distance

 

How do you reach this conclusion? If it works for short lags, why should it not be extended? Why not assume the opposite, that long routes might be the first where driverless will be deployed. That's where we see a shortfall of human drivers. It's a financially unrewarding job for the driver (train crews earn double their salary), with little sleep, and far away from home.

 

 

2. You have the political BS. Lots of truck drivers in the US.

3. The whole industry structure would have to change. Currently in the US the majority (90%?) of trucks on the road are owners/operators. Now a trucker may be able to buy/finance and operate a truck. But the independent trucker is surely not going to be able to buy/finance a convoy. The cost savings may not be big enough to trigger a complete redesign of the industry.

 

 

points 2&3 equally applied to cab drivers until 3-4 years ago. Changes are slow to come, but quick to materialise when they do eventually come.  If big players find value opps, there could easily be a consolidation, just like in any other industry that develops technologically.

 

 

While the self driving stuff will decrease the price gap between truck and rail, rail will still have a cost a cost advantage. All this stuff sounds cool. But really I think that the shift away from coal represents a much bigger and present risk for BNSF (all rails really) than self driving vehicles.

 

I know little about the industry but read that for the same volume, train cost is 100$ vs. trucking 130$ (though trucks have the door-to-door advantage). If the driverless convoy eliminates both fuel and personnel costs, a 30%-50% reduction in expenses seems reasonable, which means cost advantage can flip. With coal, its volume comprises about 18% of BNSF's freight revenues, so that's quite substantial. Yet on the other end BRK compensates for that through its MidAmerican Energy sub and clean energy initiatives.

 

[dartmonkey]

I know little about the industry but read that for the same volume, train cost is 100$ vs. trucking 130$ (though trucks have the door-to-door advantage). If the driverless convoy eliminates both fuel and personnel costs, a 30%-50% reduction in expenses seems reasonable, which means cost advantage can flip.

 

The general rule of thumb, often cited by Buffett, is that it costs 3 times as much fuel to move a ton by truck compared to the costs of moving it by train. Treehugger says the same: https://www.treehugger.com/cars/rail-versus-trucking-whos-the-greenest-freight-carrier.html However, double-stacking trucks, or having them run in convoys, can make up for some of that difference.

 

Not all of it: trains have much less rolling resistance, and require less human resources too, in addition to the fact that trucks on the roads contribute to a lot of the congestion. So it is hard to see how the cost advantage could flip, except for relatively short distances. But if trucks can  narrow the gap, they could still take some business away from trains.

 

[rb]

 

2. You have the political BS. Lots of truck drivers in the US.

3. The whole industry structure would have to change. Currently in the US the majority (90%?) of trucks on the road are owners/operators. Now a trucker may be able to buy/finance and operate a truck. But the independent trucker is surely not going to be able to buy/finance a convoy. The cost savings may not be big enough to trigger a complete redesign of the industry.

 

 

points 2&3 equally applied to cab drivers until 3-4 years ago. Changes are slow to come, but quick to materialise when they do eventually come.  If big players find value opps, there could easily be a consolidation, just like in any other industry that develops technologically.

 

The taxi cab industry is nothing like the trucking industry. In cabs the companies own fleets of cabs and the drivers lease these cars from the company as individual contractors. The trucking industry is mostly owner operators where the driver owns the truck. The raxi industry is taxed pretty heavily through licenses. Trucking gets a big subsidy in the form of infrastructure. The cost structure is very different as well. You can get a Prius for about 30k. But a class 8 truck will set you back about 125k. On top of that parts along with maintenance are repairs are way more expensive for trucks than for cabs. This causes the labor component of price to be significantly lower for trucking compared to cabs.

 

 

While the self driving stuff will decrease the price gap between truck and rail, rail will still have a cost a cost advantage. All this stuff sounds cool. But really I think that the shift away from coal represents a much bigger and present risk for BNSF (all rails really) than self driving vehicles.

 

I know little about the industry but read that for the same volume, train cost is 100$ vs. trucking 130$ (though trucks have the door-to-door advantage). If the driverless convoy eliminates both fuel and personnel costs, a 30%-50% reduction in expenses seems reasonable, which means cost advantage can flip. With coal, its volume comprises about 18% of BNSF's freight revenues, so that's quite substantial. Yet on the other end BRK compensates for that through its MidAmerican Energy sub and clean energy initiatives.

Firstly, once you go the convoy route you loose the flexibility of trucking.

 

Secondly, the 100 vs 130 train vs truck is incorrect. Revenue per ton mile for rail is about 4 cents. Revenue per ton mile for trucking is 12-15 cents.

 

Thirdly, in terms of fuel efficiency the article was talking about maybe 10% increase in fuel efficiency with the convoys. Rails already use between 1/3 and 1/4 of fuel per ton mile compared to trucks and that gap is growing.

 

Fourthly, rails carry a lot of stuff. I think about 1/3 more ton miles than trucks. You shift a meaningful amount of that stuff from rail to trucks and you'll see bunch of negative issues. An increase in congestion, crashes, and spills. Environmental effects from burning more fuel. Also a deterioration in infrastructure from wear and tear caused by trucks. US infrastructure is already kinda shabby.  Who pays for that? Rail guys like Matt Rose and Hunter Harrison will make damn sure everyone knows why their roads are shit.

 

Lastly, I don't see how Berkshire benefits through the utility unit from the decline in coal shipments on BNSF, much less offset those.  BHE's utilities' compensation is determined by their rate base not their fuel mix. If I'm wrong and someone has a good understanding of this please let me know.

[LightWhale]

Lots of valid points. I do think certain elements of the subject have been ignored, but since both of you say that the cost differential of rail/truck is 1:3, then the threat of AV to BNSF in the visible future is small, and the threat to BRK even smaller. The option to have such shared discussions and quickly reach a conclusion is invaluable. Thanks for the inputs.

[DooDiligence]

Dynamic Positioning systems installed onboard offshore vessels & the operators (DPO's) who run them, are subjected to rigorous testing & periodic trials (getting a DPO unlimited is crazy difficult now.)

 

These vessels operate in an environment with significantly less traffic than interstate travel (driving much larger vehicles at lower speeds with a f*kton more inertia, similar to the lorry-train described.)

 

I don't see regulators allowing un-manned behemoths on interstates any time soon.

 

The last mile scenario sounds more likely with small tows driven on specified routes.

 

As the standards of certification, testing & practices get ironed out, longer tows & more routes would get approved.

 

Someone has to make up the tows & the ability to do this autonomously would be an incremental money saver.

 

Here's the worldwide standards & certification org for shipboard DPO's & systems

 

http://www.nautinst.org/en/dynamic-positioning/index.cfm

 

--- edit ---

 

Just found this

 

https://www.computerworld.com/article/3134879/car-tech/self-driving-18-wheeler-delivers-the-first-shipment-beer.html

 

-

 

I don't know how good this report is but

 

http://ieeexplore.ieee.org/document/5723343/?reload=true

 

-

 

Great article on truck platooning (I guess small tows on both long & short routes could work.)

 

http://www.truckinginfo.com/channel/fuel-smarts/article/story/2016/04/platooning-is-on-the-way.aspx

 

Quote - The “most challenging aspect,” of DATP, according to the report, comes down to who to platoon with. Most fleets would prefer platooning with their own trucks, while owner-operators would prefer platooning with other owner-operators, with only 7% of surveyed owner-operators willing to platoon with large fleets and 5% of large fleets willing to platoon with owner-operators. - end quote.

 

Seems like there'd need to be a full understanding of existing trucking companies since there's likely be some consolidations.

 

Also, how would 3PL operators like C.H. Robinson fit in?

 

http://www.inboundlogistics.com/cms/article/readers-choice-top-10-3pl-excellence-awards-2016/

 

(basically just a list of operators to start with.)