Looking into BNSF's capital expenditures...The ATSF merger was finalized at the end of 1996, but
the consolidation wasn’t immediate – post-merger BNSF initially reported combined routes of over 50,000 miles of track. By 1998 they were reporting 34,000 route-miles of track, so it might be appropriate to look from that time period, up through the end of 2009, and the Berkshire acquisition.
Principal assets, 1998 & 2009.
Land & track comprised 85%-90% of balance sheet fixed assets, both initially and at the BRK re-valuation, with remaining few $B for the locomotives, freight cars and such. In terms of miles of track and numbers of locomotives of cars:
. 1998 - track: 34,000mi total -- 25,000mi owned – 5k locomotives – 97k freight cars
. 2009 - track: 32,000mi total -- 23,000mi owned – 7k locomotives – 79k freight cars
The 10K’s give us a break-out of BNSF revenue as follows:
. 1998 - $9B – 7,884k cars/units - $1,132 rev/car
. 2009 - $14B – 8,418k cars/units - $1,614 rev/car
Over the span of the ’98 through ’09 period, the total earnings, ‘owners’ cash flow’, debt issued, and use of that cash -
1998 through 2009 -
. Reported earnings: $15.5B
. Cash flow: $7.8B
. Debt issued: $5B
Principal uses of the cash flow plus additional debt, 1998-2009:
- Dividends: $3.4B
- Share buy-backs: $7.6B
In the process, capital expenditures over the 1998-2009 totaled $23B, including money essentially funded from that $8B reported earnings vs cash flow gap. No doubt there were improvements in various areas, but it would appear that a large portion of that spending was probably needed to keep things going.
Considering Berkshire assigned about 90% of the value of BNSF’s overall assets (in the $34B purchase) to the land, track, roads, and such – vs rolling equipment - we might suppose that the $23B spent in the prior decade was not the major driver in the purchase (though it was of course necessary to maintain the physical property over the ’98 to ’09 period). The company’s value couldn’t grow if at least the assets weren’t continually maintained and improved.
But the question still remains, when we look at our returns from operating the railroad, do we look at reported earnings, knowing that we are going to put half of them – and likely half of all reported earnings well into the future, back into keeping that property usable? I would look at the history of reported earnings and actual cash flow, and recognize costs of doing business for what they are- costs, and not count that continued spending as part of my own ROI.
I’m not saying we shouldn’t spend those capex dollars. We need to in order to at least realize those inflation-driven - and volume driven – returns. But that spending isn’t really our current period ROI. Not yet. And again, in all likelihood, we can look forward to putting a good share of even future reported earnings back into the business to keep it going. Half of our 'owner' or shareholder returns seem to go to 'feed the beast'. At least that's the history.
What sets BNSF apart from other less-desirable capital-intensive businesses is pricing dynamics (at least so far, and in modern times). But we shouldn't forget the 'spending' characteristics of capital-intensiveness, where there's always a need for redeployment of a portion of our (accountant-defined) earnings in order to stay up-to-date and operating.
We also have to question how much of capex is simply maintenance versus supporting 'new business', where we'd expect an incremental return. For example, BNSF spends far more on track maintenance than new track, but it does build almost 1,000 miles of new track each year. But system-wide, it shows no overall growth, year after year, of track in use. As many miles are abandoned each year as added. So what's the rational? If our accounting conventions were real-world in this case, actual (cash) earnings would be half those being reported.
______________________________
From a different time and place, when the all-important pricing aspect of the railroad business was much different and less attractive than today:
It is usually unsound to make blanket recommendations of
whole classes of securities, and there are equal objections to broad
condemnations. The record of railroad share prices in Table 14-6
shows that the group as a whole has often offered chances for a
large profit. (But in our view the great advances were in themselves
largely unwarranted.) Let us confine our suggestion to this:
There is no compelling reason for the investor to own railroad
shares; before he buys any he should make sure that he is getting
so much value for his money that it would be unreasonable to look
for something else instead.*
That's from Graham, of course.
the consolidation wasn’t immediate – post-merger BNSF initially reported combined routes of over 50,000 miles of track. By 1998 they were reporting 34,000 route-miles of track, so it might be appropriate to look from that time period, up through the end of 2009, and the Berkshire acquisition.
Principal assets, 1998 & 2009.
Land & track comprised 85%-90% of balance sheet fixed assets, both initially and at the BRK re-valuation, with remaining few $B for the locomotives, freight cars and such. In terms of miles of track and numbers of locomotives of cars:
. 1998 - track: 34,000mi total -- 25,000mi owned – 5k locomotives – 97k freight cars
. 2009 - track: 32,000mi total -- 23,000mi owned – 7k locomotives – 79k freight cars
The 10K’s give us a break-out of BNSF revenue as follows:
. 1998 - $9B – 7,884k cars/units - $1,132 rev/car
. 2009 - $14B – 8,418k cars/units - $1,614 rev/car
Over the span of the ’98 through ’09 period, the total earnings, ‘owners’ cash flow’, debt issued, and use of that cash -
1998 through 2009 -
. Reported earnings: $15.5B
. Cash flow: $7.8B
. Debt issued: $5B
Principal uses of the cash flow plus additional debt, 1998-2009:
- Dividends: $3.4B
- Share buy-backs: $7.6B
In the process, capital expenditures over the 1998-2009 totaled $23B, including money essentially funded from that $8B reported earnings vs cash flow gap. No doubt there were improvements in various areas, but it would appear that a large portion of that spending was probably needed to keep things going.
Considering Berkshire assigned about 90% of the value of BNSF’s overall assets (in the $34B purchase) to the land, track, roads, and such – vs rolling equipment - we might suppose that the $23B spent in the prior decade was not the major driver in the purchase (though it was of course necessary to maintain the physical property over the ’98 to ’09 period). The company’s value couldn’t grow if at least the assets weren’t continually maintained and improved.
But the question still remains, when we look at our returns from operating the railroad, do we look at reported earnings, knowing that we are going to put half of them – and likely half of all reported earnings well into the future, back into keeping that property usable? I would look at the history of reported earnings and actual cash flow, and recognize costs of doing business for what they are- costs, and not count that continued spending as part of my own ROI.
I’m not saying we shouldn’t spend those capex dollars. We need to in order to at least realize those inflation-driven - and volume driven – returns. But that spending isn’t really our current period ROI. Not yet. And again, in all likelihood, we can look forward to putting a good share of even future reported earnings back into the business to keep it going. Half of our 'owner' or shareholder returns seem to go to 'feed the beast'. At least that's the history.
What sets BNSF apart from other less-desirable capital-intensive businesses is pricing dynamics (at least so far, and in modern times). But we shouldn't forget the 'spending' characteristics of capital-intensiveness, where there's always a need for redeployment of a portion of our (accountant-defined) earnings in order to stay up-to-date and operating.
We also have to question how much of capex is simply maintenance versus supporting 'new business', where we'd expect an incremental return. For example, BNSF spends far more on track maintenance than new track, but it does build almost 1,000 miles of new track each year. But system-wide, it shows no overall growth, year after year, of track in use. As many miles are abandoned each year as added. So what's the rational? If our accounting conventions were real-world in this case, actual (cash) earnings would be half those being reported.
______________________________
From a different time and place, when the all-important pricing aspect of the railroad business was much different and less attractive than today:
It is usually unsound to make blanket recommendations of
whole classes of securities, and there are equal objections to broad
condemnations. The record of railroad share prices in Table 14-6
shows that the group as a whole has often offered chances for a
large profit. (But in our view the great advances were in themselves
largely unwarranted.) Let us confine our suggestion to this:
There is no compelling reason for the investor to own railroad
shares; before he buys any he should make sure that he is getting
so much value for his money that it would be unreasonable to look
for something else instead.*
That's from Graham, of course.
