BRK and leverage

[Graham Osborn]

I was just looking at BRK's growth both in revs and tbook these past 20 years and they were good - almost too good.  BRK may be run by a slightly ossified genius, but the businesses themselves are solid at best - we're not talking about an industrial steamroller like Google, Microsoft, or Standard Oil here.  So how is he doing it?  I've asked myself the same question for one too many a business since 2009, so one first looks to the balance sheet trend.  Since 2000 BRK's tbook had grown about 380% while LT debt has grown around 4000%.  D/E is off 2011 highs but still over 0.25.  The industrialist in me knows this is probably prudent use of cash given the cost of capital.  But it still leaves BRK "atypically" levered until they tighten their belt a bit.  More significantly, I wonder whether BRK can keep up the pace of rev and tbook growth once their cost of capital goes up.  That would be a continuation of a secular reversal in a 15-year D/E trend.  I could see us down in EV/ Rev 1.3 territory.  Of course with the leverage PSR < 1 possible.  The interesting thing is if they deployed the cash at PBR < 1.2 and borrowed more the D/E would be back up with is a cyclical predictor of a depressed MC - somewhat counterintuitive.

[shalab]

I think you are adding up the debt in BNSF and BH Energy and Financial products - if you look at the principal payment schedule in page 59, it looks very manageable.

 

Furthermore, S&P looked at this recently and re-affirmed the ratings ( even when not paid to do so )

 

[John Hjorth]

I was just looking at BRK's growth both in revs and tbook these past 20 years and they were good - almost too good.  BRK may be run by a slightly ossified genius, but the businesses themselves are solid at best - we're not talking about an industrial steamroller like Google, Microsoft, or Standard Oil here.  So how is he doing it?  I've asked myself the same question for one too many a business since 2009, so one first looks to the balance sheet trend.  Since 2000 BRK's tbook had grown about 380% while LT debt has grown around 4000%.  D/E is off 2011 highs but still over 0.25.  The industrialist in me knows this is probably prudent use of cash given the cost of capital.  But it still leaves BRK "atypically" levered until they tighten their belt a bit.  More significantly, I wonder whether BRK can keep up the pace of rev and tbook growth once their cost of capital goes up.  That would be a continuation of a secular reversal in a 15-year D/E trend.  I could see us down in EV/ Rev 1.3 territory.  Of course with the leverage PSR < 1 possible.  The interesting thing is if they deployed the cash at PBR < 1.2 and borrowed more the D/E would be back up with is a cyclical predictor of a depressed MC - somewhat counterintuitive.

 

I think of this in a totally different way. It's not about long term debt [interest bearing] on the group balance sheet. Yes, the debt is a large number, but it does not matter that much in the total picture.

 

It's about :

 

Insurance float, now at USD B 88 [and still counting], basicly free borrowed money from the policy holders

Deferred taxes, now at approx. USD B 63 [and still counting], basicly free borrowed money from Uncle Sam.

 

To me, by making calculations over a 20 year time span for BRK for evaluation of performance, you are making a serious mistake. BRK has over the last 20 years morphed into a totally different company - by acquisitions - than 20 years ago.

 

To me, BRK is in an increasing degree over time becoming a Nestle, Unilever or AB InBev [just not only doing business in food or beer, but a lot of things], plus an insurance operation, and there is no sign that will change going forward.

 

And the CAGR in BV pr. share has been seriously hurt by that change with in the last 10 years or so. In January I actually made some calculations on that, and the picture of it is striking.

 

Please see attached file, which is quite rough in formatting, based on the first column figures on first page in Annual Report 2014.

 

You might say, that the file is so rough in it's calculations, that it is flawed, because I'm using simple mathematical averages in stead of the IRR function in Excel. My point is, that the outcome in the diagram won't change materially by doing the calculations totally correctly, because the development in the tendencies is so striking.

 

To me, people investing in BRK now perhaps expecting [or trying to get] 10% - 15% CAGR going forward are destined to get disappointed and should consider taking on other risks than investing in BRK.

John_Hjorth_-_BRK_-_Book_value_per_share_-_v1_-_20160113.xlsx

[Graham Osborn]

 

I think of this in a totally different way. It's not about long term debt [interest bearing] on the group balance sheet. Yes, the debt is a large number, but it does not matter that much in the total picture.

 

It's about :

 

Insurance float, now at USD B 88 [and still counting], basicly free borrowed money from the policy holders

 

 

This seems like an interesting point.  Berkshire's float certainly would seem to be more optimally structured than typical debt in the sense that coupons are low or potentially even negative depending on how well the insurance businesses are run (less advantage than usual with rates this low).  And the principal is repaid in increments rather than as a lump sum (the reinsurance businesses may be an exception here since the insured events may be correlated and the claims large).  But even if it is well-structured debt it should be still considered leverage no?  If you can't pay out claims your insurer becomes insolvent or needs a bailout.  I would definitely think is should still be used in the D/E and levered beta calculation.

 

 

To me, by making calculations over a 20 year time span for BRK for evaluation of performance, you are making a serious mistake. BRK has over the last 20 years morphed into a totally different company - by acquisitions - than 20 years ago.

 

 

Assuming BRK's can be considered just a advantageous form of debt, I don't see any reason to give BRK preferential treatment relative to any other conglomerate.  Berkshire was a conglomerate capitalized partly through insurers 20 years ago and it still is.  If the relative dependence on float vs FCF from the non-insurance businesses has changed, that should be no different than a "typical" conglomerate making investments with debt vs FCF at different times.

 

I'm quite prepared to acknowledge I probably understand the intricacies of the insurance empire less well than many here.  But I just have a hard time putting BRK on a pedestal as this unicorn that can magically defy the laws of cost of capital by using float.  There has to be a catch somewhere.  Maybe the reinsurance businesses have more claims during times of financial tightening?

[thepupil]

I would definitely think is should still be used in the D/E and levered beta calculation.

 

For better or worse, I think you'll have a tough time finding a Berkshire shareholder who makes a "levered beta calculation". I'm in the "for better" camp.

[Graham Osborn]

I would definitely think is should still be used in the D/E and levered beta calculation.

 

For better or worse, I think you'll have a tough time finding a Berkshire shareholder who makes a "levered beta calculation". I'm in the "for better" camp.

 

The trouble is it does matter.  D/E is a good predictor of how much a stock will "crump" when the general market tanks.  This has nothing to do with cost of capital per say, it just reflects the fact that for a given change in EV the MC takes a bigger hit.  If BRK is more levered than it was in 08-09, it will crump harder.  If so, that will give some of us better buy prices.  And the stock will be spring loaded for heftier gains afterward.

[John Hjorth]

 

I think of this in a totally different way. It's not about long term debt [interest bearing] on the group balance sheet. Yes, the debt is a large number, but it does not matter that much in the total picture.

 

It's about :

 

Insurance float, now at USD B 88 [and still counting], basicly free borrowed money from the policy holders

 

 

This seems like an interesting point.  ...

 

 

To the members of this board invested in BRK, this is just not an interesting point, but a fact, on which [among others] investment in BRK is based.

[thepupil]

I would definitely think is should still be used in the D/E and levered beta calculation.

 

For better or worse, I think you'll have a tough time finding a Berkshire shareholder who makes a "levered beta calculation". I'm in the "for better" camp.

 

The trouble is it does matter.  D/E is a good predictor of how much a stock will "crump" when the general market tanks.  This has nothing to do with cost of capital per say, it just reflects the fact that for a given change in EV the MC takes a bigger hit.  If BRK is more levered than it was in 08-09, it will crump harder.  If so, that will give some of us better buy prices.  And the stock will be spring loaded for heftier gains afterward.

 

- Berkshire has never been less levered.

 

- Look at the balance sheet of consolidated subs BNSF and BHE

https://www.bamsec.com/filing/93461215000028?cik=934612

https://www.bamsec.com/filing/108131615000018?cik=1081316

 

These two have about $160B of assets and $56B of equity and have $57B / $82B (~70%)  of Berkshire's debt. They also have $30B of the $60B DTL (50%). so they've got 70% of the financial debt and 50% of the DTL despite being ~29% of the assets. They are where the leverage resides. The two have assets/equity of 2.85 versus Berkshire's total of 2.1, so they are more levered than the rest of Berkshire.

 

Now imagine if instead of wholly owning these, Berkshire owned $90B of stock in them ($60B of BNSF and $30B of BHE), they would just be an asset on the balance sheet and you would say Berkshire is not really all that levered. Also, consider the beta of utility equity. It is low. I'm sure railroads' beta is decently high though.

 

I'm not going to put a to of effort into the calculations of seeing what it would look like if they held them as stocks instead of consolidated subs, but I think you get my point. Berkshire isn't any more levered than it was in the past and has only gotten stodgier and less levered with time.

 

Also, if you care about beta and all that stuff, wouldn't the beta calculated from Berkshire stock movement be more useful? I believe it's about 0.8.

 

 

 

[thepupil]

as a follow up and using the 10-K instead of the old Q's, if you peeled out the BNSF and BE, you'd take away roughly $165B of assets and $102B of liabilities (1/2 the DTL is related to those)

 

Assets would be      $552-$165  = $387B

Liabilities would be  $293-$102  = $191B

 

And you'd have $63B of equity in the utility and railroad. If, hypothetically those traded @ $90B (not heroic valuations for either of those, $60B = 0.9UNP since BNSF earning = 0.9UNP Earnings and $30B = 15X earnings for BHE), you'd have $477B of assets and $191B of liabilities and $286B of equity and Berkshire would look less levered than ever.

 

So it may appear to have fast growing debt because it happens to have two capital intensive regulated companies on the balance sheet but if it owned these companies in stock form you wouldn't notice the leverage at all. Hope that helps. 

 

EDIT: So ex-BNSF and BHE, you'd have $191B of liabilities. $30B of that is DTL related to stock price appreciation (0% interest loan from government) and only $25B is interest bearing debt (the rest being float and operating liabilities).

 

Pre-PCP, the insurance co holds $61B of cash and $27B of fixed income, $88B total of liquid, high quality paper. After PCP, the debt figures will obviously go up (by about $14B) and the cash will go down (by $20B), but the earnings power will go up and they'll re-load the gun within a year or two.

 

In short: it isn't that levered when you disaggregate the balance sheet.

 

 

[jay21]

BRK is crazy overcapitalized / under levered

[Palantir]

But why look at leverage without BNSF and BE? I don't see why comparing it to owning in stock form is more accurate. Rather, owning it in stock form would simply understated true leverage...

[thepupil]

But why look at leverage without BNSF and BE? I don't see why comparing it to owning in stock form is more accurate. Rather, owning it in stock form would simply understated true leverage...

 

- The debt is not guaranteed by Berkshire, so that's one reason

 

- OP was asking why D/E is rising/ is that sustainable, and the answer is "they own two large businesses that have 70% of the debt". Decide for yourself if those businesses can handle that debt.

 

- It isn't more or less accurate, but when looking at "how levered is Berkshire" or just trying to value Berkshire, I find it much more useful to pull out its two largest businesses and value the equity stakes in those businesses. To take it to an extreme, Would it help or hurt you if you consolidated 9.8% of Wells Fargo's balance sheet or 15% of Amex's into Berkshire's financials? Wells fargo has $1.8 trillion of assets and $1.6 trillion of liabilities and Berkshire owns 9.8% of it. Obviously it would seriously muck up the analysis. Rather than do that, you'd simply value the stake in the equity and stress that for risk assessment purposes. You might say "BNSF is worth $30-$90B or BHE is worth $25-$50B" or something like that, and incorporating that into your bear/base/bull case. In my opinion that would be the better way to view things. Disaggregation versus consolidation is generally more enlightening.

 

- When you only look at a single consolidated quantitative metric (like debt / equity), it doesn't really tell you anything. A consolidated Berkshire balance sheet is a much more un-analyzable. But BNSF and BHE have separate public financials and publish K's and Q's and are independent debt issuers. We should use that to help us analyze the company.

 

- with any holdco, I'd look at the holdco debt and put it in the context of the equity stakes it owns. Does LUK have $1B of debt or $30B? Well it owns JEF which has $30B of liabilities and the equity stake in JEF could be wiped out by that debt, but it isn't guaranteed by LUK. So in my opinion, LUK owns a $3.8B tangible equity stake in  a bank (JEF), which is 10X levered. That to me is a more accurate way of portraying things than saying "LUK has $46B of assets and $35B of liabilities, wow debt / equity has increased DRAMATICALLY over the past 10 years".

 

- in a a crazy hypothetical, if Bernie sanders nationalized the railroads and utilities of the U.S. (I realize that isn't what he is saying he'd do, but just for illustrative purposes), would you a) deduct $165B from Berkshire's assets and adjust equity ($255-$165) or would you deduct the assets and the corresponding liabilities, in other words just the equity ($265-$63B). It would be more accurate to just deduct the equity because Berkshire would have $200B of equity after that event, not $90B

 

- the least an equity stake could be worth is $0, whereas if you consolidate, your stressing may give those business's negative value (which wouldn't happen, excepting some crazy scenario where debt not guarantee by Berkshire became guaranteed)

 

Just remembered we actually discussed this very topic already

 

Have members ruled out that BV might decline? I mean BRK is so leveraged and heavily into cyclical, industrial stocks which are getting pounded this year. Eg, NSC is down -30% YTD, and if we assume a similar valuation decline for BNSF, and multiply by 2x for leverage, the "true" BV might be lower than anticipated?

 

 

BNSF has a book value to Berkshire of $34B. Using a simplistic analysis of the decrease in market value of BNSF (simply saying the "right" multiple has gone from 20X to 14X), then BNSF went from being worth $90B to $56B, from 2.3X GAAP book to 1.6X GAAP Book.

 

It would be incorrect to say NSC is down 30%, Berkshire has 44% equity/ assets, ergo Berkshire's interest in BNSF is worth 60% less.

 

BNSF is simply "worth" a less giant premium to book than before. That premium moves up in down in an unlevered fashion because Berkshire's equity stakes in BNSF and BE are not themselves levered like that.

 

Interesting, thanks.

[Graham Osborn]

Also, if you care about beta and all that stuff, wouldn't the beta calculated from Berkshire stock movement be more useful? I believe it's about 0.8.

 

The interesting thing is the beta trend since 2004 mirrors the quoted D/E - up about 340% vs 250% (beta was around 0.2 then vs 0.8 now).  So BRK is in fact quite a bit more volatile relative to the market than it was historically - at record highs in fact. 

 

I haven't gone through all the subs for that whole period yet.  But from my personal investing perspective whether a sub's operations are consolidated or not is irrelevant to the fact that that piece of the business has a certain amount of debt associated.  The market isn't stupid - if I owned BRK and 50% of BRK was AXP, you can bet BRK would be tanking too.  Whether you prefer to say that is because BRK's portfolio notched losses or because BRK was levered through its minority owned subs is more about who you choose to blame :)

[thepupil]

Also, if you care about beta and all that stuff, wouldn't the beta calculated from Berkshire stock movement be more useful? I believe it's about 0.8.

 

The interesting thing is the beta trend since 2004 mirrors the quoted D/E - up about 340% vs 250% (beta was around 0.2 then vs 0.8 now).  So BRK is in fact quite a bit more volatile relative to the market than it was historically - at record highs in fact. 

 

I haven't gone through all the subs for that whole period yet.  But from my personal investing perspective whether a sub's operations are consolidated or not is irrelevant to the fact that that piece of the business has a certain amount of debt associated.  The market isn't stupid - if I owned BRK and 50% of BRK was AXP, you can bet BRK would be tanking too.  Whether you prefer to say that is because BRK's portfolio notched losses or because BRK was levered through its minority owned subs is more about who you choose to blame :)

 

beta doesn't equal vol.

 

Berkshire's beta to the market is indeed increasing (it's now an S&P 500 component, a large portion of XLF, an important megacap that will trade in line with markets).

 

It's vol relative to the market is random in a given year but I don't see an upward trend. I'll need to update my monthly returns but I've attached a rolling 12M standard deviation of berkshire thru August 2015. Don't really notice an upward trend. Downside vol relative to the S&P doesn't appear to be increasing either.

 

I don't think it's correct  to say Berkshire has become more volatile / levered. Many would say the decline in the crazy high growth of book value seen in the 90s / 2000s is because it is no longer a pretty levered portfolio of stocks.

 

[thepupil]

downside deviation relative to S&P seems to have peaked in the late 90's (when berkshire massively underperformed in the tech boom), overall volatility is decreasing.

 

1/1990 -12/2000, vol was 25.5% versus the S&P's 13.9%, downside deviation was 14.1% versus 8.8%

 

1/2001 - 12/2010, vol was 17% versus S&P's 16%, downside 11.5% versus 12.8%

 

1/2011 - 1/2016, vol was 13.2% versus 11.9%, downside 7.7% versus 7.5%.

 

I don't see really see a pickup in the vol of berkshire that reflects an increase in leverage. Downside deviation (and upside vol) are in a long term decline. This makes sense for an ever diversifying conglomerate megacap that is in fact not levering up and has become much more staid.

 

EDIT: I agree, the market isn't stupid, which is why volatility and downside deviation has decreased while the business has grown larger and more diversified, less levered and as regulated businesses have become a greater portion of assets and earnings.

 

And beta (which is not exactly voll, but rather a measure of systematic risk, ie how much a stock's moves can be explained by market moves) is increasing as the shareholder base has become more institutional, the stock has been included in indices, etc.

 

Attached Rolling 12M beta for reference.

[Graham Osborn]

Yes, pardon my misuse of terminology.  Beta is covariance of asset/ market return whereas volatility is an absolute quantity independent of the market.  Volatility is lower now than in years past because market volatility has been lower since 09 with a recent uptrend.  The relative volatility calculation above would be more useful I guess.

 

Anyway, BRK/ market beta has been on the rise and seems remarkably correlated with the D/E (I'm just plotting the two in Ycharts going back to 2000 or so).  Unfortunately the institutional ownership % in Ycharts doesn't run back that far so I can't say whether there is any pattern.  I wonder how big an effect that has on beta compared with the D/E?  I see some anecdotal comments online but am not familiar the research on this.  It does give me an idea for another parameter to add to my screens though so thanks.

[thepupil]

I cannot tell you why beta has increased. I would just offer that indexation/institutionalization is an alternative narrative to D/E.

 

Berkshire was added to the S&P in Feb 2010; you'll notice we've been in a consistent move towards beta of 1 since early 2011; maybe coincidence, maybe not.

 

I don't put a lot of faith/stock in this quantitative finance stuff and would encourage you to assess the business fundamentals in terms of risk and leverage, but I don't think there is much evidence to support increasing leverage or volatility.

 

[Guest longinvestor]

Great question for the agm.

 

Why is Beta up @ BRK? And what is management going to do about it? What we really want is Alpha, how do we plan to make that bigger?

[thepupil]

Great question for the agm.

 

Why is Beta up @ BRK? And what is management going to do about it? What we really want is Alpha, how do we plan to make that bigger?

 

While I know you are being sarcastic, alpha has been quite decent outside of the last 1 yr. Berkshire has a lot of characteristics that quanty folks love; typically lower peak to trough drawdowns than indexes over time, lower beta (though increasing as Graham points out), better gain to loss ratio than indices, better (more positive) skewness,  good upside capture versus downside capture, blah blah blah; all saying the same thing different ways really.

 

It likely does not matter to the vast majority of shareholders, of course.

 

Annualized alpha against S&P, Russell 2000, DJ Financials*

 

1 yr:  -9%, -6%, -8%

2 yr: +2%, +8%, 4%

3 yr: +2%, +8%,+3%

4yr:  +3%, +9%, +4%

5 yr: +1%, +6%, +4%

10 yr: +5%, +7%,+9%

 

* not equal to excess return http://markovprocesses.com/blog/2012/08/alpha-and-excess-return-not-synonymous/)

[Graham Osborn]

One fairly obvious point I missed - as BRK owns more of the market their beta will naturally converge to 1.  Warren wants that other 98% of Fortune 500 - if he does beta will be redefined as asset/ BRK ;)

[Palantir]

- with any holdco, I'd look at the holdco debt and put it in the context of the equity stakes it owns. Does LUK have $1B of debt or $30B? Well it owns JEF which has $30B of liabilities and the equity stake in JEF could be wiped out by that debt, but it isn't guaranteed by LUK. So in my opinion, LUK owns a $3.8B tangible equity stake in  a bank (JEF), which is 10X levered. That to me is a more accurate way of portraying things than saying "LUK has $46B of assets and $35B of liabilities, wow debt / equity has increased DRAMATICALLY over the past 10 years".

 

Good points