In Berkshire circles there's often discussion as to the PE that might be applied to Berkshire’s
non-insurance business, the group that makes up about $10B of the current normalized earnings of $17B that Buffett mentioned. Whether we use the two-column approach or are developing an overall total-company PE, it’s helpful to at least have an approach that has its basis in the market, and is relative to Berkshire’s actual businesses (vs say, the S&P 500).
Technically, if we were doing this work-up with an eye towards supportability (eg, if we expected we’d have to explain it in a deposition with an astute analyst sitting on the other side) we’d likely look at this more specifically to the business units- at least the major ones, and we’d draw our assumptions from actual, appropriate market data. As an aside, we’d also do this on an EV/EBIT basis rather than using a PE, so we’re apples to apples on leverage and ‘excess’ cash & securities – ie, we’d do our two-column analysis. While there’s danger in precision, this process at least forces us to look at the Berkshire operating companies realistically.
For these purposes, we’ll take some short-cuts and jump right to PE’s. Purists can take the same approach at an EV/EBIT level, which again is more appropriate. This is an illustrative example; use your own examples and data. The main thing to keep in mind is that consistent application is the key.
These are PE’s below are 2010 actual/estimates, as available, corresponding to components of the $17B ‘current normalized’ pre-tax Buffett mentioned (and detailed out, to some extent) in the annual report.
To the extent forward earnings are used in an analysis like this, forward-looking PE’s (not the 2010e PE’s shown here) should also be used. These are generally lower. To the extent I did not get the ‘eliminations’ completely right on the pre-tax breakout, I'll note that for my purposes here they are probably close enough-- the purpose here is simply for weighting the various business segment PE's.
Again, I’m using the less-accurate ‘PE’ here just for convenience, using ‘market value’ rather than ‘enterprise value’ which is more appropriate. And again, the data is illustrative to the methodology. Pick your own public comparisons, and use your own assessment of those vs the Berkshire units. Where a Berkshire company is stronger than its peers, we can certainly apply a premium. This way though, any premium we apply will be more targeted, rather than broad-brush.
PreTax ($MM) %Tot %NonIns Comp.PE Wtd.PE
Underwriting 2,013 12%
Invest. Income 5,145 30%
Tot Insurance 7,158 41%
BNSF 4,080 23% 40% 18 7.2
MidAmerican@90% 1,385 8% 14% 14 1.9
Finance 813 5% 8% 17 1.4
Marmon @ 60% 488 3% 5% 17 0.8
McLane 369 2% 4% 15 0.5
Other Mfg 1,911 11% 19% 17 3.2
Other Svce 984 6% 10% 16 1.5
Retail 197 1% 2% 20 0.4
Tot Non-insurance 10,227 59% 100% - 16.9
Total 17,385 100%
Those wishing to work out a total-company PE can take a stab at the Insurance numbers and recalculate, using an all-inclusive weighting. I don’t do it here simply because if we were going to do that, I’d use EV instead of PE’s, stripping out cash/securities and all debt (using unlevered values through-out) then bring the debt and cash back into the valuation estimate at the end .
Notes:
BNSF – 2010 pre-tax earnings ‘annualized’ from 46 weeks reported, to 52 weeks ‘normalized’.
The BNSF comparisons are to Union Pacific (2010est PE: 17), Norfolk Southern (16), and CSX (18). For this illustration I used 18 for BNSF. We can update the comparisons as better data is available.
MidAmerican—Berkshire’s 90% interest shown. Comps might include Duke (13); Dte (13); Xcel (14)
Finance—Berkshire units in this group, per that annual report: Clayto, XTRA, CORT, BH Finance. Potential comps include Aaaron’s (16), Tal (15), and McGraths (18). For illustration, used 17 for Berkshire.
Marmon – currently (as of year-end) 60% of earnings applicable to Berkshire. Public companies might include Tyco, Actuant, and Carlisle companies.
McLane – compares to Loblaw (15), and perhaps SYSCO (15)
Other Mfg – Acme, Ben Moore, JohnsManville, Shaw, Fruit/Loom, Forest River, etc. The group had a pre-tax margin of about 11%. Conducted a search of similar size-range cyclical manufacturing companies with 2010 pre-tax margins of 8% to 20%--resultant median PE: 17
Other services—Berkshire's grouping in this category includes: NetJets, TTI, BusWire, DairyQueen. The group had pre-tax margin of about 13% in 2010. Used a similar search of 8%-20% pre-tax service companies. Median PE: 15
Retail: Berkshire’s category for the Home Furnishing stores, Jewelers, and Sees had pre-tax margins averaging 7% in 2010. Mid-size retailers in the 5% to 15% pre-tax range (40 companies) had a median PE of 18.5. Used 20 for the Berkshire group.
_______________
Using these examples – again, use your own selection – we get an earnings-weighted PE of about 17 for the group. And saying it again, if we are looking to apply this multiple to forward earnings, we need to re-look at the comparison companies – their forward multiples on estimated earnings for 2011 or beyond are likely lower also.
Lastly, if we wanted to be particularly thorough, we’d look at transaction multiples for recent acquisitions – public or private – of companies similar to the Berkshire units. We all see plenty examples of this, of course, but very often M&A transactions suggest higher valuations than day-to-day markets would indicate – often the result of a strategic vs financial M&A motive. But for us, acquisitions might not be as helpful as they otherwise would be, since we are unlikely to ever realize 'M&A' value.
We know that Berkshire doesn’t ‘conglomerate’ companies, and generally doesn’t realize synergies (as say, a Johnson & Johnson might). Most of Berkshire’s units would have greater ‘strategic’ value to some other company. In the instances where Berkshire does divest – Gillette; CapCities/ABC — it’s almost always to a strategic buyer. We always have that latent 'value reserve'. (My stock example: FlightSafety's value to Boeing). As noted though, we are highly unlikely to ever realize that last dosage of value – the reason Berkshire was able to make these acquisitions in the first place.
Anyway, plug in the PE's weight them, and get a composite 'value' number.
non-insurance business, the group that makes up about $10B of the current normalized earnings of $17B that Buffett mentioned. Whether we use the two-column approach or are developing an overall total-company PE, it’s helpful to at least have an approach that has its basis in the market, and is relative to Berkshire’s actual businesses (vs say, the S&P 500).
Technically, if we were doing this work-up with an eye towards supportability (eg, if we expected we’d have to explain it in a deposition with an astute analyst sitting on the other side) we’d likely look at this more specifically to the business units- at least the major ones, and we’d draw our assumptions from actual, appropriate market data. As an aside, we’d also do this on an EV/EBIT basis rather than using a PE, so we’re apples to apples on leverage and ‘excess’ cash & securities – ie, we’d do our two-column analysis. While there’s danger in precision, this process at least forces us to look at the Berkshire operating companies realistically.
For these purposes, we’ll take some short-cuts and jump right to PE’s. Purists can take the same approach at an EV/EBIT level, which again is more appropriate. This is an illustrative example; use your own examples and data. The main thing to keep in mind is that consistent application is the key.
These are PE’s below are 2010 actual/estimates, as available, corresponding to components of the $17B ‘current normalized’ pre-tax Buffett mentioned (and detailed out, to some extent) in the annual report.
To the extent forward earnings are used in an analysis like this, forward-looking PE’s (not the 2010e PE’s shown here) should also be used. These are generally lower. To the extent I did not get the ‘eliminations’ completely right on the pre-tax breakout, I'll note that for my purposes here they are probably close enough-- the purpose here is simply for weighting the various business segment PE's.
Again, I’m using the less-accurate ‘PE’ here just for convenience, using ‘market value’ rather than ‘enterprise value’ which is more appropriate. And again, the data is illustrative to the methodology. Pick your own public comparisons, and use your own assessment of those vs the Berkshire units. Where a Berkshire company is stronger than its peers, we can certainly apply a premium. This way though, any premium we apply will be more targeted, rather than broad-brush.
PreTax ($MM) %Tot %NonIns Comp.PE Wtd.PE
Underwriting 2,013 12%
Invest. Income 5,145 30%
Tot Insurance 7,158 41%
BNSF 4,080 23% 40% 18 7.2
MidAmerican@90% 1,385 8% 14% 14 1.9
Finance 813 5% 8% 17 1.4
Marmon @ 60% 488 3% 5% 17 0.8
McLane 369 2% 4% 15 0.5
Other Mfg 1,911 11% 19% 17 3.2
Other Svce 984 6% 10% 16 1.5
Retail 197 1% 2% 20 0.4
Tot Non-insurance 10,227 59% 100% - 16.9
Total 17,385 100%
Those wishing to work out a total-company PE can take a stab at the Insurance numbers and recalculate, using an all-inclusive weighting. I don’t do it here simply because if we were going to do that, I’d use EV instead of PE’s, stripping out cash/securities and all debt (using unlevered values through-out) then bring the debt and cash back into the valuation estimate at the end .
Notes:
BNSF – 2010 pre-tax earnings ‘annualized’ from 46 weeks reported, to 52 weeks ‘normalized’.
The BNSF comparisons are to Union Pacific (2010est PE: 17), Norfolk Southern (16), and CSX (18). For this illustration I used 18 for BNSF. We can update the comparisons as better data is available.
MidAmerican—Berkshire’s 90% interest shown. Comps might include Duke (13); Dte (13); Xcel (14)
Finance—Berkshire units in this group, per that annual report: Clayto, XTRA, CORT, BH Finance. Potential comps include Aaaron’s (16), Tal (15), and McGraths (18). For illustration, used 17 for Berkshire.
Marmon – currently (as of year-end) 60% of earnings applicable to Berkshire. Public companies might include Tyco, Actuant, and Carlisle companies.
McLane – compares to Loblaw (15), and perhaps SYSCO (15)
Other Mfg – Acme, Ben Moore, JohnsManville, Shaw, Fruit/Loom, Forest River, etc. The group had a pre-tax margin of about 11%. Conducted a search of similar size-range cyclical manufacturing companies with 2010 pre-tax margins of 8% to 20%--resultant median PE: 17
Other services—Berkshire's grouping in this category includes: NetJets, TTI, BusWire, DairyQueen. The group had pre-tax margin of about 13% in 2010. Used a similar search of 8%-20% pre-tax service companies. Median PE: 15
Retail: Berkshire’s category for the Home Furnishing stores, Jewelers, and Sees had pre-tax margins averaging 7% in 2010. Mid-size retailers in the 5% to 15% pre-tax range (40 companies) had a median PE of 18.5. Used 20 for the Berkshire group.
_______________
Using these examples – again, use your own selection – we get an earnings-weighted PE of about 17 for the group. And saying it again, if we are looking to apply this multiple to forward earnings, we need to re-look at the comparison companies – their forward multiples on estimated earnings for 2011 or beyond are likely lower also.
Lastly, if we wanted to be particularly thorough, we’d look at transaction multiples for recent acquisitions – public or private – of companies similar to the Berkshire units. We all see plenty examples of this, of course, but very often M&A transactions suggest higher valuations than day-to-day markets would indicate – often the result of a strategic vs financial M&A motive. But for us, acquisitions might not be as helpful as they otherwise would be, since we are unlikely to ever realize 'M&A' value.
We know that Berkshire doesn’t ‘conglomerate’ companies, and generally doesn’t realize synergies (as say, a Johnson & Johnson might). Most of Berkshire’s units would have greater ‘strategic’ value to some other company. In the instances where Berkshire does divest – Gillette; CapCities/ABC — it’s almost always to a strategic buyer. We always have that latent 'value reserve'. (My stock example: FlightSafety's value to Boeing). As noted though, we are highly unlikely to ever realize that last dosage of value – the reason Berkshire was able to make these acquisitions in the first place.
Anyway, plug in the PE's weight them, and get a composite 'value' number.
