A reader sent me an email asking me why the company shown in my blind stock valuation contest has free cash flow that's so much lower than reported earnings.
Letās look at something Warren Buffett said in his 2009 letter to shareholders:
āOur BNSF operation, it should be noted, has certain important economic characteristics that resemble those of our electric utilities. In both cases we provide fundamental services that are, and will remain, essential to the economic well-being of our customers, the communities we serve, and indeed the nation. Both will require heavy investment that greatly exceeds depreciation allowances for decades to comeā¦.ā
Without giving too much away, the mystery company in the blind stock valuation exercise fits that description.
But why does any company fit that description?
Why would a company require capital spending āthat greatly exceeds depreciation allowancesā not just for one year or two but decade after decade?
The answer is inflation.
Now, before I get into illustrating why railroads all vastly overstate their true economic earnings, I want to stress that this overstatement is not caused by fraud or managementās desire to mislead. If investors are misled, itās only because they look at the P/E ratio without looking at free cash flow.
It is well known that railroads canāt pay out all their earnings in dividends. The big reason for this is that some of those stated earnings are completely illusory.
Itās an earnings mirage.
This railroad earnings mirage come from a continual gap between capital spending and depreciation. Year after year, railroads spend more on new property than they count as a replacement expense. In other words, the āupkeepā expense at railroads is vastly understated. This causes vastly overstated earnings.
Railroads depreciate against the original cost of their old assets. But they replace their old assets with new assets. Because of inflation, replacing any asset bought or built in 1991 with an identical one in 2011 will cost more money.
Depreciation doesnāt adjust for cost inflation.
Obviously, inflation isnāt specific to railroads. But an inflation rate of say 2% a year makes little difference when youāre replacing an asset after 1 or 2 or 3 years.
Inflation makes a huge difference when youāre replacing an asset after 10 or 15 or 20 years.
A movie ticket cost $4.20 in 1991. Try getting a replacement ticket for $4.20 today.
Railroads tend to have free cash flow that is no greater than about 50% of their reported earnings. That's true of American railroads over the last 15 years or so (when inflation ran 2% a year using the GDP deflator). A couple railroads had free cash flow - the actual amount they could have paid in dividends each year - running around half of earnings. Most actually did even worse than that. This means their earnings don't all come in the form of cash.
Some of the 50% difference was caused by capital spending meant to grow the business. But a lot of that 50% difference between net income and free cash flow was caused by a gap between depreciation expense and the actual cost required to replace the railroadās assets.
Depreciation charges were too low relative to the actual cash upkeep needed to maintain the same assets and do the same volume of business.
Thatās bound to happen if railroads are replacing equipment that's 15 or 20 years old. Obviously, inflation makes the replacement cost higher than the original cost.
Yet depreciation isnāt adjusted for inflation.
Companies depreciate against the original cost of the asset they use up over those 15 or 20 years. So if they buy something for $1,000 that will last 20 years they charge themselves $50 a year ($1,000 / 20 years = $50/year). But even if inflation is just 3% a year over the next 20 years, the actual replacement cost when the thing breaks down in 2031 will be $1,800. It'll be $2,200 if inflation is 4%.
Let's assume inflation will be 3% over the assetās 20 year life. That means the replacement cost of something we buy in 2011 and replace in 2031 will be $1,800 in 2031 for every $1,000 we pay today.
If instead of depreciating against the original $1,000 cost we make up a non-GAAP fantasy account I'll call "Provision for Future Equipment Replacement" we might use a straight-line method where we take the actual replacement cost ($1,800) in 2031 and space it out evenly over 20 years. $1,800 divided by 20 years equals $90 a year.
So, we'd report depreciation of $50 a year under the original cost method - assuming 3% inflation - even though we know full well that we have to set aside $90 a year for the actual replacement cost of the equipment in 2031 if we live in a world of 3% inflation. That means our owner earnings should be reduced by multiplying depreciation by 1.8 ($90/$50 = 1.8). So, economic depreciation is really 1.8 times the depreciation we put on the books.
If you look at a group of 5 or 10 railroads, power companies, etc. you'll get a feel for how long the useful life of their entire business is. A big clue is property, plant, and equipment / capital spending. If a no-growth business has $32,000 of PP&E and $2,000 in capital spending, it's shedding and regrowing its PP&E skin at a rate of once every 16 years (since it's replacing 1/16th this year). That's how long the business's life span is. It has to regenerate all its tangible assets every 16 years.
This is only meaningful for an asset heavy business like a railroad. Itās not a relevant calculation for businesses that depend primarily ā or entirely ā on their current assets to produce earnings.
And all of this is in nominal terms. In reality, the companyās property, plant, and equipment isnāt being replaced every 17 years in a business where annual capital spending is 6% of PP&E (1/0.06 = 16.67). Itās not like youāre replacing the same 100 miles of track every 17 years. Itās more like youāre replacing the same $100 worth of track every 17 years. How much track $100 buys changes over the years.
Why should you care how long it takes a business to shed and regrow its balance sheet skin?
It gives you a clue as to how inflation is disguising the economic reality behind the reported earnings. Only businesses that are very asset heavy will be affected by inflation in this way.
But ā even among asset heavy businesses ā thereās a big difference between a business that replaces those tangible assets every 6 years and a business that replaces those tangible assets every 16 years.
A business that replaces its tangible assets every 6 years will be depreciating some of those assets in 2005 dollars instead of 2011 dollars. But a business that replaces its tangible assets every 16 years will be depreciating some of those assets in 1995 dollars instead of 2011 dollars.
Railroads are an odd business in that their assets donāt have to be replaced very often. Itās just that they have to use unfathomable amounts of assets to move stuff around. Some other asset heavy businesses ā especially very dense, very heavily used networks ā have high capital spending but short asset lives.
Their depreciation expense is more accurate even though theyāre also big capital spenders, because the original cost of their assets is always closer to the replacement cost of those same assets.
None of this would be true if we lived in a non-inflationary world. But since the 1940s, we in the United States have tended to have more or less constant inflation. Prices never really fall for very long. And they rise ever higher over time.
The result is that our companies understate their depreciation expense in proportion to the age of the assets chewed up by their day-to-day operations.
Ravi Nagarajan of Rational Walk sent me an email where he mentioned that some industries do a good job of breaking down capital spending and explaining depreciation. Thatās true. The two examples he gave were railroads and offshore oil drillers.
If youāre wondering about how railroads calculate depreciation, thereās a note in each companyās 10-K that explains how it calculates depreciation.
Here is note #10 from the Union Pacific (UNP, Financial) 10-K:
āā¦Properties and equipment are carried at cost and are depreciated on a straight-line basis over their estimated service lives, which are measured in years, except for rail in high-density traffic corridors (i.e., all rail lines except for those subject to abandonment, yard and switching tracks, and electronic yards), which are measured in millions of gross tons per mile of trackā¦ā
Basically, Union Pacific is saying that they donāt adjust their depreciation charges for inflation. They just take the original cost of the property and depreciate it using one of two methods.
The easiest way to think of the two depreciation methods is to imagine a hammer and an anvil. The property is the hammer. You can either set aside money to pay for a new hammer in equal installments every day or in equal installments every time the hammer strikes the anvil. For something like a hammer, the economically accurate way to depreciate the asset is based on the number of anvil strikes, because the anvil will eventually destroy the hammer. Time doesnāt destroy hammers. Anvils destroy hammers.
Always ask what forces a company to replace its assets.
Is it time? Is it use? Is it technology? Or is it customer taste?
The best asset is one that can survive all those changes.
The worst asset is one that canāt survive any of those changes.
Twenty years ago ā back in 1991 ā this is what Warren Buffett told a group of Notre Dame Students:
āThe telephone company (AT&T), with the patents, the MBAs, the stock options, and everything else, had one problem, and that problem is illustrated by those figures on that lower left hand column. And those figures show the plant investment in the telephone business. Thatās $47 billion, starting off with, growing to $99 billion over an eight or nine year period. More and more and more money had to be tossed in, in order to make these increased earnings, going from $2.2 billion to $5.6 billion. So, they got more money, but you can get more money from a savings account if you keep adding money to it every year. The progress in earnings that the telephone company made was only achievable because they kept on shoving more money into the savings account and the truth was, under the conditions of the ā70s, they were not getting paid commensurate with the amount of money that they had to shove into the pot, whereas Lord Thompson, once he bought the paper in Council Bluffs, never put another dime in. They just mailed money every year. And as they got more money, he bought more newspapers. And, in fact, he said it was going to say on his tombstone that he bought newspapers in order to make more money in order to buy more newspapers. The idea was that, essentially, he raised prices and raised earnings there every year without having to put more capital into the business. One is a marvelous, absolutely sensational business, the other one is a terrible business.ā
So why did Warren Buffett buy Burlington Northern?
Thatās a great question. Buffett has given some clues. His best explanation is probably what he told Charlie Rose.
Since I donāt agree with Buffettās purchase, I canāt do his argument justice.
I just have no interest in owning a railroad. Itās not a business I want to be in. That may be close minded. But thatās the truth.
What I wanted to do here is just make sure that the folks who are interested in buying railroads arenāt overly fixated on the P/E ratio.
A railroadās reported earnings are not comparable to reported earnings in other industries.
Follow Geoff at Gannon On Investing
Letās look at something Warren Buffett said in his 2009 letter to shareholders:
āOur BNSF operation, it should be noted, has certain important economic characteristics that resemble those of our electric utilities. In both cases we provide fundamental services that are, and will remain, essential to the economic well-being of our customers, the communities we serve, and indeed the nation. Both will require heavy investment that greatly exceeds depreciation allowances for decades to comeā¦.ā
Without giving too much away, the mystery company in the blind stock valuation exercise fits that description.
But why does any company fit that description?
Why would a company require capital spending āthat greatly exceeds depreciation allowancesā not just for one year or two but decade after decade?
The answer is inflation.
Now, before I get into illustrating why railroads all vastly overstate their true economic earnings, I want to stress that this overstatement is not caused by fraud or managementās desire to mislead. If investors are misled, itās only because they look at the P/E ratio without looking at free cash flow.
It is well known that railroads canāt pay out all their earnings in dividends. The big reason for this is that some of those stated earnings are completely illusory.
Itās an earnings mirage.
This railroad earnings mirage come from a continual gap between capital spending and depreciation. Year after year, railroads spend more on new property than they count as a replacement expense. In other words, the āupkeepā expense at railroads is vastly understated. This causes vastly overstated earnings.
Railroads depreciate against the original cost of their old assets. But they replace their old assets with new assets. Because of inflation, replacing any asset bought or built in 1991 with an identical one in 2011 will cost more money.
Depreciation doesnāt adjust for cost inflation.
Obviously, inflation isnāt specific to railroads. But an inflation rate of say 2% a year makes little difference when youāre replacing an asset after 1 or 2 or 3 years.
Inflation makes a huge difference when youāre replacing an asset after 10 or 15 or 20 years.
A movie ticket cost $4.20 in 1991. Try getting a replacement ticket for $4.20 today.
Railroads tend to have free cash flow that is no greater than about 50% of their reported earnings. That's true of American railroads over the last 15 years or so (when inflation ran 2% a year using the GDP deflator). A couple railroads had free cash flow - the actual amount they could have paid in dividends each year - running around half of earnings. Most actually did even worse than that. This means their earnings don't all come in the form of cash.
Some of the 50% difference was caused by capital spending meant to grow the business. But a lot of that 50% difference between net income and free cash flow was caused by a gap between depreciation expense and the actual cost required to replace the railroadās assets.
Depreciation charges were too low relative to the actual cash upkeep needed to maintain the same assets and do the same volume of business.
Thatās bound to happen if railroads are replacing equipment that's 15 or 20 years old. Obviously, inflation makes the replacement cost higher than the original cost.
Yet depreciation isnāt adjusted for inflation.
Companies depreciate against the original cost of the asset they use up over those 15 or 20 years. So if they buy something for $1,000 that will last 20 years they charge themselves $50 a year ($1,000 / 20 years = $50/year). But even if inflation is just 3% a year over the next 20 years, the actual replacement cost when the thing breaks down in 2031 will be $1,800. It'll be $2,200 if inflation is 4%.
Let's assume inflation will be 3% over the assetās 20 year life. That means the replacement cost of something we buy in 2011 and replace in 2031 will be $1,800 in 2031 for every $1,000 we pay today.
If instead of depreciating against the original $1,000 cost we make up a non-GAAP fantasy account I'll call "Provision for Future Equipment Replacement" we might use a straight-line method where we take the actual replacement cost ($1,800) in 2031 and space it out evenly over 20 years. $1,800 divided by 20 years equals $90 a year.
So, we'd report depreciation of $50 a year under the original cost method - assuming 3% inflation - even though we know full well that we have to set aside $90 a year for the actual replacement cost of the equipment in 2031 if we live in a world of 3% inflation. That means our owner earnings should be reduced by multiplying depreciation by 1.8 ($90/$50 = 1.8). So, economic depreciation is really 1.8 times the depreciation we put on the books.
If you look at a group of 5 or 10 railroads, power companies, etc. you'll get a feel for how long the useful life of their entire business is. A big clue is property, plant, and equipment / capital spending. If a no-growth business has $32,000 of PP&E and $2,000 in capital spending, it's shedding and regrowing its PP&E skin at a rate of once every 16 years (since it's replacing 1/16th this year). That's how long the business's life span is. It has to regenerate all its tangible assets every 16 years.
This is only meaningful for an asset heavy business like a railroad. Itās not a relevant calculation for businesses that depend primarily ā or entirely ā on their current assets to produce earnings.
And all of this is in nominal terms. In reality, the companyās property, plant, and equipment isnāt being replaced every 17 years in a business where annual capital spending is 6% of PP&E (1/0.06 = 16.67). Itās not like youāre replacing the same 100 miles of track every 17 years. Itās more like youāre replacing the same $100 worth of track every 17 years. How much track $100 buys changes over the years.
Why should you care how long it takes a business to shed and regrow its balance sheet skin?
It gives you a clue as to how inflation is disguising the economic reality behind the reported earnings. Only businesses that are very asset heavy will be affected by inflation in this way.
But ā even among asset heavy businesses ā thereās a big difference between a business that replaces those tangible assets every 6 years and a business that replaces those tangible assets every 16 years.
A business that replaces its tangible assets every 6 years will be depreciating some of those assets in 2005 dollars instead of 2011 dollars. But a business that replaces its tangible assets every 16 years will be depreciating some of those assets in 1995 dollars instead of 2011 dollars.
Railroads are an odd business in that their assets donāt have to be replaced very often. Itās just that they have to use unfathomable amounts of assets to move stuff around. Some other asset heavy businesses ā especially very dense, very heavily used networks ā have high capital spending but short asset lives.
Their depreciation expense is more accurate even though theyāre also big capital spenders, because the original cost of their assets is always closer to the replacement cost of those same assets.
None of this would be true if we lived in a non-inflationary world. But since the 1940s, we in the United States have tended to have more or less constant inflation. Prices never really fall for very long. And they rise ever higher over time.
The result is that our companies understate their depreciation expense in proportion to the age of the assets chewed up by their day-to-day operations.
Ravi Nagarajan of Rational Walk sent me an email where he mentioned that some industries do a good job of breaking down capital spending and explaining depreciation. Thatās true. The two examples he gave were railroads and offshore oil drillers.
If youāre wondering about how railroads calculate depreciation, thereās a note in each companyās 10-K that explains how it calculates depreciation.
Here is note #10 from the Union Pacific (UNP, Financial) 10-K:
āā¦Properties and equipment are carried at cost and are depreciated on a straight-line basis over their estimated service lives, which are measured in years, except for rail in high-density traffic corridors (i.e., all rail lines except for those subject to abandonment, yard and switching tracks, and electronic yards), which are measured in millions of gross tons per mile of trackā¦ā
Basically, Union Pacific is saying that they donāt adjust their depreciation charges for inflation. They just take the original cost of the property and depreciate it using one of two methods.
The easiest way to think of the two depreciation methods is to imagine a hammer and an anvil. The property is the hammer. You can either set aside money to pay for a new hammer in equal installments every day or in equal installments every time the hammer strikes the anvil. For something like a hammer, the economically accurate way to depreciate the asset is based on the number of anvil strikes, because the anvil will eventually destroy the hammer. Time doesnāt destroy hammers. Anvils destroy hammers.
Always ask what forces a company to replace its assets.
Is it time? Is it use? Is it technology? Or is it customer taste?
The best asset is one that can survive all those changes.
The worst asset is one that canāt survive any of those changes.
Twenty years ago ā back in 1991 ā this is what Warren Buffett told a group of Notre Dame Students:
āThe telephone company (AT&T), with the patents, the MBAs, the stock options, and everything else, had one problem, and that problem is illustrated by those figures on that lower left hand column. And those figures show the plant investment in the telephone business. Thatās $47 billion, starting off with, growing to $99 billion over an eight or nine year period. More and more and more money had to be tossed in, in order to make these increased earnings, going from $2.2 billion to $5.6 billion. So, they got more money, but you can get more money from a savings account if you keep adding money to it every year. The progress in earnings that the telephone company made was only achievable because they kept on shoving more money into the savings account and the truth was, under the conditions of the ā70s, they were not getting paid commensurate with the amount of money that they had to shove into the pot, whereas Lord Thompson, once he bought the paper in Council Bluffs, never put another dime in. They just mailed money every year. And as they got more money, he bought more newspapers. And, in fact, he said it was going to say on his tombstone that he bought newspapers in order to make more money in order to buy more newspapers. The idea was that, essentially, he raised prices and raised earnings there every year without having to put more capital into the business. One is a marvelous, absolutely sensational business, the other one is a terrible business.ā
So why did Warren Buffett buy Burlington Northern?
Thatās a great question. Buffett has given some clues. His best explanation is probably what he told Charlie Rose.
Since I donāt agree with Buffettās purchase, I canāt do his argument justice.
I just have no interest in owning a railroad. Itās not a business I want to be in. That may be close minded. But thatās the truth.
What I wanted to do here is just make sure that the folks who are interested in buying railroads arenāt overly fixated on the P/E ratio.
A railroadās reported earnings are not comparable to reported earnings in other industries.
Follow Geoff at Gannon On Investing
