It's helpful to keep in mind that in BRK valuations, we’re trying to arrive at a reasonable
estimate of Berkshire’s “discounted value of the cash that can be taken out of the business during its remaining life,” as Buffett puts it.
As we close the hurricane season, we might consider whether a catastrophic event that hypothetically would result in major losses for Berkshire’s insurance business would also result in a dollar-for-dollar reduction in Berkshire’s intrinsic value. The answer depends on how we defined Berkshire’s intrinsic value in the first place.
--Subscribers to ‘float is an asset’ type valuation methodologies – valuation methods fairly unique to Berkshire (versus how we’d value other insurers) including ‘two column’, the ‘intrinsivaluator’, and other treatments of ‘float’ as immediate additions to value – reasonably saw insurance losses as direct hits to intrinsic value.
--Those who instead calculated Berkshire’s IV using direct valuation of the insurance companies as businesses (reserving premiums received against future losses rather than treating them immediately as unencumbered, equity-like funding) concluded, also correctly from their stand-point, that insurance losses did not directly flow to IV.
estimate of Berkshire’s “discounted value of the cash that can be taken out of the business during its remaining life,” as Buffett puts it.
As we close the hurricane season, we might consider whether a catastrophic event that hypothetically would result in major losses for Berkshire’s insurance business would also result in a dollar-for-dollar reduction in Berkshire’s intrinsic value. The answer depends on how we defined Berkshire’s intrinsic value in the first place.
--Subscribers to ‘float is an asset’ type valuation methodologies – valuation methods fairly unique to Berkshire (versus how we’d value other insurers) including ‘two column’, the ‘intrinsivaluator’, and other treatments of ‘float’ as immediate additions to value – reasonably saw insurance losses as direct hits to intrinsic value.
--Those who instead calculated Berkshire’s IV using direct valuation of the insurance companies as businesses (reserving premiums received against future losses rather than treating them immediately as unencumbered, equity-like funding) concluded, also correctly from their stand-point, that insurance losses did not directly flow to IV.
'Float-is-as good-as-equity' has been part of many Berkshire shareholders’ conventional thinking for so long that we sometimes don’t stop to consider otherwise. The alternative valuation – valuing the insurance business with reserves for inevitable and anticipated losses – results in a lower valuation, and we can take a shot at calculating that differential.
Using some hopefully reasonable assumptions, we’ll see that the IV differential between the two perspectives might be in the range of $25k to $30k per share. For example, if we arrived at a valuation of $145k using a float-based methodology (eg, two-column), the equivalent IV on a ‘reserved’ basis would be something like $115k to $120k.
Let’s look at this more closely:
We can step away from ‘two-column’, etc, and carve out the insurance subs from that ‘cash and securities’ column and treat them as separately valued businesses. We do this for the other operating businesses, so why not give it a shot with the insurance subs? What would our build-up of IV look like then, vs two-column? Intuitively, we’d think that IV would have to be lower – after all, we’re recognizing the existence of insurance reserves that we gloss over under two-column and the intrinsivaluator -- but how much lower is that valuation?
This is a reasonable question. After all, the markets value most insurance businesses as ‘straight-up’ insurance companies – with investors considering the companies’ cash and securities portfolios as being encumbered (to say the least) by actuarial estimates of potential claims.
Float-as-equity is not an attribute that markets recognize for insurers in general – even very well managed ones. The ‘pass’ we give Berkshire is largely a function of Buffet’s long-running ability to grow the float-leveraged portfolio beyond Berkshire’s corresponding reserve requirements.
In distinctly segregating the insurance subs, we’d have to also segregate the portion of cash and securities necessary to support the insurance businesses as stand-alone entities.
When we look at the balance sheets of major insurance competitors we see that for the most part they are pretty straight-forward, with cash & securities and other funding assets comprising most of the assets, and insurance-related liabilities - reserves – comprising the liabilities. For the other major reinsurers, these two elements pretty much comprise book value. Their ‘other’ assets and liabilities related to the business – operating payables, receivables, PPE and such - are pretty much a wash, and in any event not particularly material.
There are three elements to this distinct valuation of the insurance subs:
a) Our assessment of reserves
b) Any book value we assign above those reserves
c) Any premium that we estimate exists above (or discount below) book value
Some attributes of these three components:
-- In comparison to ‘two-column’ and similar methodologies, items (a) and (b) are reclassifications out of the ‘cash and securities’ column, to the operating sub column.
-- Item (a) is a reduction to intrinsic value versus ‘two-column’, while item (c) is a plus to IV.
-- Any amount we assign to item (b) is a ‘wash’ in intrinsic value versus two-column – simply a transfer with no IV impact.
Most large insurers, and notably reinsurers, are selling for less than book. SwissRe and MunichRe are selling for a bit under 70% of book.
If we assume that in Berkshire’s case, the insurance reserves on the books are conservatively presented, and that as an operating company the Berkshire subs are worth substantially more than others in its insurance peer group -- book value or more -- then the amount of cash & securities we’d have to include as we carve out these businesses would be approximately the sum of (a) and (b) above -- our assessment of reserves, plus any ‘book value’ we assign above those reserves.
Using some hopefully reasonable assumptions, we’ll see that the IV differential between the two perspectives might be in the range of $25k to $30k per share. For example, if we arrived at a valuation of $145k using a float-based methodology (eg, two-column), the equivalent IV on a ‘reserved’ basis would be something like $115k to $120k.
Let’s look at this more closely:
We can step away from ‘two-column’, etc, and carve out the insurance subs from that ‘cash and securities’ column and treat them as separately valued businesses. We do this for the other operating businesses, so why not give it a shot with the insurance subs? What would our build-up of IV look like then, vs two-column? Intuitively, we’d think that IV would have to be lower – after all, we’re recognizing the existence of insurance reserves that we gloss over under two-column and the intrinsivaluator -- but how much lower is that valuation?
This is a reasonable question. After all, the markets value most insurance businesses as ‘straight-up’ insurance companies – with investors considering the companies’ cash and securities portfolios as being encumbered (to say the least) by actuarial estimates of potential claims.
Float-as-equity is not an attribute that markets recognize for insurers in general – even very well managed ones. The ‘pass’ we give Berkshire is largely a function of Buffet’s long-running ability to grow the float-leveraged portfolio beyond Berkshire’s corresponding reserve requirements.
In distinctly segregating the insurance subs, we’d have to also segregate the portion of cash and securities necessary to support the insurance businesses as stand-alone entities.
When we look at the balance sheets of major insurance competitors we see that for the most part they are pretty straight-forward, with cash & securities and other funding assets comprising most of the assets, and insurance-related liabilities - reserves – comprising the liabilities. For the other major reinsurers, these two elements pretty much comprise book value. Their ‘other’ assets and liabilities related to the business – operating payables, receivables, PPE and such - are pretty much a wash, and in any event not particularly material.
There are three elements to this distinct valuation of the insurance subs:
a) Our assessment of reserves
b) Any book value we assign above those reserves
c) Any premium that we estimate exists above (or discount below) book value
Some attributes of these three components:
-- In comparison to ‘two-column’ and similar methodologies, items (a) and (b) are reclassifications out of the ‘cash and securities’ column, to the operating sub column.
-- Item (a) is a reduction to intrinsic value versus ‘two-column’, while item (c) is a plus to IV.
-- Any amount we assign to item (b) is a ‘wash’ in intrinsic value versus two-column – simply a transfer with no IV impact.
Most large insurers, and notably reinsurers, are selling for less than book. SwissRe and MunichRe are selling for a bit under 70% of book.
If we assume that in Berkshire’s case, the insurance reserves on the books are conservatively presented, and that as an operating company the Berkshire subs are worth substantially more than others in its insurance peer group -- book value or more -- then the amount of cash & securities we’d have to include as we carve out these businesses would be approximately the sum of (a) and (b) above -- our assessment of reserves, plus any ‘book value’ we assign above those reserves.
Again, in two-column terms, that’s what we’d have to pull out of the ‘cash and securities’ column to show the insurance subs as stand-alone.
Now let’s compare this to ‘two-column’ valuation, on an overall basis. Looking at the impact this reclassification of the insurance subs would have on ‘two-column’ intrinsic value, we’d see Berkshire’s IV:
(a) reduced by the liabilities - those reserves that we are assigning to the insurance businesses - but
(b) we’ll also get an increase for the amount that we value the insurance businesses above ‘book value’ (regardless of what we actually assign for book value).
Again, thinking in two-column terms, whatever number we assign as insurance book value is an IV ‘wash’ vs two-column -- a ‘plus’ in the operating sub column, directly offset by a ‘minus’ on the cash & securities column.
Buffett refers to the premium we assign over book value, in his latest annual report discussion of Geico, as ‘goodwill’. Any goodwill we attribute to those carved-out insurers becomes a straight ‘plus’ to IV. Funding of reserves, on the other hand, is a reduction to IV. Again, from this perspective, catastrophic losses and other claims are now funded.
We need to determine both the reserves and the ’goodwill, the value we estimate to be in excess of book value.
Now let’s compare this to ‘two-column’ valuation, on an overall basis. Looking at the impact this reclassification of the insurance subs would have on ‘two-column’ intrinsic value, we’d see Berkshire’s IV:
(a) reduced by the liabilities - those reserves that we are assigning to the insurance businesses - but
(b) we’ll also get an increase for the amount that we value the insurance businesses above ‘book value’ (regardless of what we actually assign for book value).
Again, thinking in two-column terms, whatever number we assign as insurance book value is an IV ‘wash’ vs two-column -- a ‘plus’ in the operating sub column, directly offset by a ‘minus’ on the cash & securities column.
Buffett refers to the premium we assign over book value, in his latest annual report discussion of Geico, as ‘goodwill’. Any goodwill we attribute to those carved-out insurers becomes a straight ‘plus’ to IV. Funding of reserves, on the other hand, is a reduction to IV. Again, from this perspective, catastrophic losses and other claims are now funded.
We need to determine both the reserves and the ’goodwill, the value we estimate to be in excess of book value.
--We can get the reserves from the financial statements. They total $66B as of 2011 Q2, however that’s ‘gross’, and for our purposes we’ll use ‘net’ of $60B and count on $6B of relief we might anticipate from outside reinsurers (per the AR).
--The valuations of the subs versus book include some judgment, but we can give them a stab.
Geico and Progressive are reasonably similar in size and performance. Progressive currently enjoys an industry-leading $6B market mark-up vs book (aka ‘goodwill’) which seems reasonable, so we can start with that as a base level for Geico.
For the top of the Geico goodwill range we might look to Berkshire’s 2010 annual report, where Buffett commented that when Berkshire completed the purchase of Geico, his assessment of Geico’s goodwill was just under 1x annual premiums. Using that multiple now, we get a hefty $14B markup over book value (whether the market would agree with that if Geico were still publicly traded could be debatable, but let’s not be conservative in this exercise, and go ahead and use that $14B as a top-range estimate for goodwill, for now).
Looking at GenRe’s major global re-insurance competitors, Swiss Re and Munich Re, as we noted earlier we see that they are trading at less than 70% of book (discounts of $8B and $9B, respectively). Let’s step up and give GenRe a hefty 40% ‘goodwill’ premium versus SwissRe and MunichRe, and value GenRe at full book value.
We might also apply the same sort of 40% premium for BHRG, assessing that up at BV. On the one hand, BHRG may be fairly ‘key man dependent’ compared to other reinsurers – but on the other -- since BHRG is particularly efficient and no doubt conservatively reserved, let’s even give it an additional valuation premium even over book value, say an additional $3B goodwill.
Depending in the premium we assign to Geico, $6B-$14B, we have a combined range of $9B to $17B of insurance related goodwill in excess of book value for the Berkshire insurers.
In that recent two-column exercise, using the multiples of publicly traded peers by business line, I had come up with a two-column ballpark value of $238B, or $144k per share.
Looking at the Intrinsivalator on its ‘more conservative’ setting, we currently see $227B, or $138k per share.
[Digressing a moment, a principal difference between my ‘two-column’ and the Intrinsivaluator’s ‘more conservative’ is the calculator’s more restrained valuation of ‘deferred taxes'. This is probably at least partially appropriate these days, if not a tad heavy-handed - considering the evolution of Berkshire’s deferred taxes to a higher proportion of short-term accelerated depreciation – ‘non-float’(in the case of BNSF’s deferred taxes, assumption of a deferred tax bill that benefited prior owners) -- rather than the actual deferral of taxes that comprised almost all of this category in decades past.]
The result, treating the insurance subs as insurance businesses, with their own earmarked assets and reserves and as if they were stand-alone – awarding ourselves $17B in insurance goodwill -- is a net reduction of $43B of intrinsic value versus two-column or the intrinsivaluator, or $26k less per share. Using a Progressive-like premium on Geico, the reduction to IV would be $30k per share.
I’ve long been in the two-column camp – more specifically ‘enhanced two-column,’ recognizing the possibility of some premium to book for Berkshire’s insurance activities. Using two-column logic, catastrophic losses and such aren’t reserved. Additionally, my estimate of two-column value would be overstated in the event of a hypothetical break-up of Berkshire. For example if the insurers were spun off at generous but hopefully realistic market prices, looking at the insurers as funded businesses, we’d still be at a value that was substantially less than ‘two-column’ however, because in essence we wouldn’t be perpetually deferring unreserved liabilities, and instead attaching funded reserves.
Discussions of our overall IV exposure to catastrophic losses remind us that ‘worst case, Berkshire is still worth the sum of its businesses, plus that additional, unencumbered portion of its portfolio. in effect, excess reserves.
This approach at least gives us a benchmark base of the businesses value. As we noted above, if we think the value is $144k on a two-column basis, then total IV is likely in that $112k to $120k per share ballpark on an insurance-sub-valued-as-insurers basis, with reserves in place. But at that $112k intrinsic value, at least all actuarial anticipated insurance events are covered without further impacting IV with each new event.
..........................................
Shifting gears a bit: a foundation of Berkshire’s two-column premise –and similar float-based estimations - is that insurance-based float, even ‘free’, translates into at least some value if supported by expected returns.
If we can see (or incorporate) those 'additional' returns into an income-based discounted cash flow analysis in some manner, we can continue to rationalize a value. If we can’t, then perhaps we should re-examine our long-held assumptions. Going back to Buffett’s original definition of intrinsic value (“cash that can be taken out of the business’) we’re not going to ever see a realization of float principal – just the benefits.
If we no longer are experiencing, and can no longer project, Berkshire’s long-running historical levels of returns on float - "cash taken out" - then we might reasonably challenge its dollar-for-dollar addition to value. On the other hand, if present-value-of-return estimates are understated in some way, we’d want to recalibrate from that end.
Buffett seems to have alluded to this in his new addition to the annual report this year – that ‘third pillar’ reference in assessing two-column approaches to IV – the “efficacy with which retained earnings will be deployed in the future.”
That’s the question, of course, in comparing historical valuations that include float (more accurately, the present value of future expected returns from float) versus the historical record that we’ve extrapolated forward in the past. While it’s tempting to assign a unique value to one component and tack it on to other elements of value, we also really need to step back and look at our overall return expectations, and value those.
Geico and Progressive are reasonably similar in size and performance. Progressive currently enjoys an industry-leading $6B market mark-up vs book (aka ‘goodwill’) which seems reasonable, so we can start with that as a base level for Geico.
For the top of the Geico goodwill range we might look to Berkshire’s 2010 annual report, where Buffett commented that when Berkshire completed the purchase of Geico, his assessment of Geico’s goodwill was just under 1x annual premiums. Using that multiple now, we get a hefty $14B markup over book value (whether the market would agree with that if Geico were still publicly traded could be debatable, but let’s not be conservative in this exercise, and go ahead and use that $14B as a top-range estimate for goodwill, for now).
Looking at GenRe’s major global re-insurance competitors, Swiss Re and Munich Re, as we noted earlier we see that they are trading at less than 70% of book (discounts of $8B and $9B, respectively). Let’s step up and give GenRe a hefty 40% ‘goodwill’ premium versus SwissRe and MunichRe, and value GenRe at full book value.
We might also apply the same sort of 40% premium for BHRG, assessing that up at BV. On the one hand, BHRG may be fairly ‘key man dependent’ compared to other reinsurers – but on the other -- since BHRG is particularly efficient and no doubt conservatively reserved, let’s even give it an additional valuation premium even over book value, say an additional $3B goodwill.
Depending in the premium we assign to Geico, $6B-$14B, we have a combined range of $9B to $17B of insurance related goodwill in excess of book value for the Berkshire insurers.
In that recent two-column exercise, using the multiples of publicly traded peers by business line, I had come up with a two-column ballpark value of $238B, or $144k per share.
Looking at the Intrinsivalator on its ‘more conservative’ setting, we currently see $227B, or $138k per share.
[Digressing a moment, a principal difference between my ‘two-column’ and the Intrinsivaluator’s ‘more conservative’ is the calculator’s more restrained valuation of ‘deferred taxes'. This is probably at least partially appropriate these days, if not a tad heavy-handed - considering the evolution of Berkshire’s deferred taxes to a higher proportion of short-term accelerated depreciation – ‘non-float’(in the case of BNSF’s deferred taxes, assumption of a deferred tax bill that benefited prior owners) -- rather than the actual deferral of taxes that comprised almost all of this category in decades past.]
The result, treating the insurance subs as insurance businesses, with their own earmarked assets and reserves and as if they were stand-alone – awarding ourselves $17B in insurance goodwill -- is a net reduction of $43B of intrinsic value versus two-column or the intrinsivaluator, or $26k less per share. Using a Progressive-like premium on Geico, the reduction to IV would be $30k per share.
I’ve long been in the two-column camp – more specifically ‘enhanced two-column,’ recognizing the possibility of some premium to book for Berkshire’s insurance activities. Using two-column logic, catastrophic losses and such aren’t reserved. Additionally, my estimate of two-column value would be overstated in the event of a hypothetical break-up of Berkshire. For example if the insurers were spun off at generous but hopefully realistic market prices, looking at the insurers as funded businesses, we’d still be at a value that was substantially less than ‘two-column’ however, because in essence we wouldn’t be perpetually deferring unreserved liabilities, and instead attaching funded reserves.
Discussions of our overall IV exposure to catastrophic losses remind us that ‘worst case, Berkshire is still worth the sum of its businesses, plus that additional, unencumbered portion of its portfolio. in effect, excess reserves.
This approach at least gives us a benchmark base of the businesses value. As we noted above, if we think the value is $144k on a two-column basis, then total IV is likely in that $112k to $120k per share ballpark on an insurance-sub-valued-as-insurers basis, with reserves in place. But at that $112k intrinsic value, at least all actuarial anticipated insurance events are covered without further impacting IV with each new event.
..........................................
Shifting gears a bit: a foundation of Berkshire’s two-column premise –and similar float-based estimations - is that insurance-based float, even ‘free’, translates into at least some value if supported by expected returns.
If we can see (or incorporate) those 'additional' returns into an income-based discounted cash flow analysis in some manner, we can continue to rationalize a value. If we can’t, then perhaps we should re-examine our long-held assumptions. Going back to Buffett’s original definition of intrinsic value (“cash that can be taken out of the business’) we’re not going to ever see a realization of float principal – just the benefits.
If we no longer are experiencing, and can no longer project, Berkshire’s long-running historical levels of returns on float - "cash taken out" - then we might reasonably challenge its dollar-for-dollar addition to value. On the other hand, if present-value-of-return estimates are understated in some way, we’d want to recalibrate from that end.
Buffett seems to have alluded to this in his new addition to the annual report this year – that ‘third pillar’ reference in assessing two-column approaches to IV – the “efficacy with which retained earnings will be deployed in the future.”
That’s the question, of course, in comparing historical valuations that include float (more accurately, the present value of future expected returns from float) versus the historical record that we’ve extrapolated forward in the past. While it’s tempting to assign a unique value to one component and tack it on to other elements of value, we also really need to step back and look at our overall return expectations, and value those.
