Warren Buffett likes to say Dexter Shoe was his worst acquisition.
He not only bought the dying shoe business for $400 million in 1993, but he paid for it in his most precious currency: Berkshire Hathaway stock. As he wrote in the 2007 Berkshire annual report:
What I had assessed as a durable competitive advantage vanished with a few years. But that’s just the beginning: By using Berkshire stock, I compounded this error hugely. That move made the cost to Berkshire shareholders not $400 million, but rather $3.5 billion. In essence, I gave away 1.6% of a wonderful business – one now valued at $220 billion – to buy a worthless business.
Yet General Re, the huge reinsurance business Buffett acquired in 1998 for $22 billion—reportedly without input from his partner, Charlie Munger—looks worse, at least on a dollar basis. While General Re, unlike Dexter, is not going away any time soon, it has caused Buffett the kind of aggravation he despises from companies in the Berkshire “family.”
Like, government investigation-type aggravation, which resulted in the conviction of four former General Re executives earlier this year for allegedly cooking up an insurance transaction to make AIG’s business look better.
Also, General Re had a derivatives problem—23,000 derivatives, to be more precise—that Berkshire spent years, and more than $400 million, unwinding back when derivatives could be unwound.
But more than that, Buffett paid for General Re in stock. $22 billion in stock. At the time, people said Buffett was selling high.
In the last decade, however, that stock as almost doubled. The current value of the General Re deal is now closer to $40 billion.
It is hard to imagine that for $40 billion, Warren Buffett could not have re-created General Re, with money to spare. Not being insurance types, however, we put the question to our readers:
Was General Re worth $40 billion? Informed opinions are welcome.
Jeff Matthews
I Am Not Making This Up
© 2008 NotMakingThisUp, LLC
The content contained in this blog represents the opinions of Mr. Matthews. Mr. Matthews also acts as an advisor and clients advised by Mr. Matthews may hold either long or short positions in securities of various companies discussed in the blog based upon Mr. Matthews’ recommendations. This commentary in no way constitutes investment advice, nor is it a solicitation of business in any way. It is intended solely for the entertainment of the reader, and the author.
26 replies on “How Much is General Re Worth, Anyway?”
It may be a good idea to check the contribution of GenRe to Berkshire, to make sure there is no circular dependency, i.e. Berkshire’s stock doubled, but it may be *because* it now has GenRe…
The core franchise of Gen Re – the bit that Buffett could not reproduce – was the old investment banking type franchise.
Businesses would insure with Gen Re because they always had. The file was passed from generation to generation at Gen Re and at the home company.
Problems were sorted out CEO to CEO. Remember offering reinsurance is almost equivalent to offering capital to an insurance company. Reinsurance is capital.
A reinsurer that is always there for you is as good to an insurance company as an investment bank that is always there for you is to a big industrial.
Ford ALWAYS dealt with Goldman Sachs. AIG always dealt with Gen Re.
The CEO to CEO thing that Hank G (AIG) had going – and which caused Gen Re all these problems – was precisely part of this tradition.
The biggest problem with Gen Re is that the proportion of broker business to direct business is changing – and not in nice ways.
I am not so sure that Ford is quite as wedded to Goldies as before either…
Its less good a business. The moat is thinner.
—
Now I will have a go at answering the question of what it is worth – another day…on my blog…
John Hempton
http://www.brontecapital.blogspot.com
A fair few legal problems in Australia relating to the insurers FAI and HIH which General Re was involved in “re-insuring” (actually a fiddle of the books disguised as re-insurance with no claim possible). http://www.smh.com.au/news/Business/APRA-asks-Buffett-to-show-cause/2005/03/16/1110913666663.html
Not an insurance specialist here but I really think your (and by inference, Mr Buffet’s) thinking is wrong on this.
Let’s look at what Buffet said because his example is a simpler one.
The value destruction (or possibly creation, I don’t know) should be the $400m less any cash generated (or spent) to date, less the value of that business if sold today. No more, no less.
The decision to issue stock rather than pay cash is a secondary mistake that is more relevant to capital structure. Unfortunately, there’s no way to calculate this. Should it be the value of the stock issued in excess of the cash amount paid for the stock? Or should we calculate a theoretical value for the Berkshire shares if the transaction had not taken place. What if paying in cash would have put Berkshire in a precarious financial position? If so, then it would not have even been a mistake as the value of the shares would fall to liquidation value.
Certainly one could consider this to be a more costly decision than the one to buy the shoe company in the first place. Like pornography, you can’t define it with precision but you know it when you see it.
as well as selling high, Buffett was also diluting an illiquid (due to the size of the stakes) all-equity portfolio
Inscrutable Chicken seems to be makng a simple calculation very complex, but maybe we’re missing the point.
First, paying $400 million in cash for Dexter would not have made a dent or even a paint-scratch in Berkshire’s AAA rating, so the capital structure issue is not relevant.
Second, recall that Dexter is “worthless” in Buffett’s words; and assume Dexter didn’t contribute meaningfully to Berkshire’s cash flow during its few years of life in the Berkshire family. Value received: pretty much zero.
So the issue is, what did he pay? And he paid in 25,203 shares of Class A Berkshire stock.
Current value, $3.1 billion.
Put another way: he could have bought back $400 million in Berkshire stock instead of buying Dexter, and he’d have $3.1 billion worth of Berkshire stock in the treasuary, instead of $0 worth of Dexter shoe company stock.
Let’s focus on General Re. That’s a deal that certainly did affect Berkshire’s capital structure.
JM
A further follow up on the issue of distribution:
If you go back to the 1999 annual report Buffett said this:
Even though a reinsurer may have a tightly focused and rational compensation system, it cannot count on every year coming up roses. Reinsurance is a highly volatile business, and neither General Re nor Ajit’s operation is immune to bad pricing behavior in the industry. But General Re has the DISTRIBUTION , the underwriting skills, the culture, and — with Berkshire’s backing — the financial clout to become the world’s most profitable reinsurance company. Getting there will take time, energy and discipline, but we have no doubt that Ron Ferguson and his crew can make it happen.”
I have emphasized DISTRIBUTION – but it is quite clear that Buffett listed distribution first because that was his reason for buying. He thought Ajit was a brilliant underwriter. But he couldn’t seem to bring in the business like the grandest old house (Gen Re). He thought if he tacked Berkshire’s financial strenght to Gen Re’s distribution he would come up trumps…
So he purchased the old house. For its distribution.
Does anybody seriously believe now that the distribution advantage is intact?
Buffett stuffed that one.
John Hempton
http://www.brontecapital.blogspot.com
Jeff,
Framing perhaps peter’s response with some numbers….
WB had an overvalued stock due in part to overvalued stocks like KO.
By buying GenRe he obtained $15B in float to invest – gotta assume some rate of return on that. He did get hit by the impact of hidden bombs in the insurance portfolio and his float cost did climb for the next few years. However, he also said in the 99 annual report that he preferes a lumpy 15% growth rate to a 12% smooth pace. Since the end of 2001 he has had a negative cost of float btw.
I suspect you may have run some of the numbers – therefore I have not gone further. Perhaps if you’ve not then someone else could throw in their estimates.
The dexter deal was a dud – WB has generally been upfront about it. He has been upfront about his other mistakes as well – see USAir, etc. He’s generally been very successful at going “all in” at great times. No investment manager has a 100% batting average, the best like WB though don’t stike out like Casey.
NJ
I’m a little rusty on the details, but have in the past looked into the timeline of Berkshire’s purchase of GenRe.
The agreement was struck when Berkshire’s stock price (and GenRe’s) was quite a bit lower then when the deal actually was voted on (and Goodwill gets booked). Memory says it ran up 50% to 60%. Then came down quite a bit.
So, was it a $13 billion deal, a $22 billion deal, or a $17 billion deal? That all depends on the day in 1998 you use.
GAAP calls for $22 billion. Buffett probably had his own number, the ratio of Berkshire stock to GenRe (though, he’s admitted the deal wasn’t his best).
Berkshire has never written off any of the Goodwill it booked. Which has to pass a DCF test.
I tend to agree with the Inscrutable Chicken on the measurement question, which puts me in the unfortunate position of disagreeing with BOTH Mr. Matthews and Mr. Buffett. (And for the second time, as well – this appears to be a DQ redux). Simply comparing the return that Berkshire actually received from a particular investment to the return on Berkshire’s stock over the life of the investment is not appropriate. Taken to its conclusion, that logic would dictate that Berkshire should have done nothing but repurchase its shares over the past 20 years or so – given its outsized returns. But that’s obviously non-sensical, as Berkshire’s stock has appreciated primarily because of the investments its made. Unless you can isolate the portion of Berkshire’s overall return that is attributable to each portfolio company (which I think is impossible) I don’t think you can make a fair comparison. In a nutshell, when you do so, you assume way too much.
Hold on a sec – all I’m saying is that there are two calculations here. It’s not over-complication, it’s breaking it down into two distinct components. First one is the loss from choosing to buy Dexter (if worthless and no positive or negative cashflows during ownership then it’s $400m), which we’re both agreed on.
Secondly, is the decision to issue shares rather than pay cash and this one is less clear cut. Yes, the shares have seen massive appreciation since the deal was done but that is value dilution, rather than value destruction. What if the Berkshire shares were flat over the period? Then we wouldn’t be having this debate (or at least it wouldn’t be Warren Buffett’s company we are talking about).
What if the shares fell over the same period? You would then say it was a good move to issue shares rather than pay cash? I raise this because I used to specialize in telecoms, where there were loads of deals that were done at the top of the bubble (so the decision to buy at that price was the wrong one) but many were done with shares (and because those shares have fallen, the decision to pay in shares was the right one).
Now, applying this to General Re. 1) Was it a good purchase? $22bn +/- cashflows generated since acquisition – current market value. If this is positive then it was a value creating, if negative then destructive. 2) Was it paid for in the right way? The shares have approximately doubled since the deal so it was not (unless paying with cash would have jeopardized its credit rating).
It’s important to make the distinction because it applies to the build v buy question. If Berkshire spent $22bn in cash in 98, would the value creation be higher than (1) for General Re? If yes then should have built, if no then should have bought. Decision 2 is irrelevant to the discussion because, even if Berkshire chose to build it would still have to decide whether to fund it through cash or issuing new shares.
Look, now you’ve made me get all serious and stuff which is really not what I am about….
Ps I disagree with you but I still love your blog.
Didn’t buying General Re give Berkshire access to a huge float? Buffet would sometimes brag about negative rates–getting paid to borrow from the float.
There’s an unknown opportunity cost to the cash that’s not being counted in the analysis. If Berkshire had paid cash for Dexter or General Re, perhaps there’s something else he didn’t buy that would have grown faster than BRK stock.
NJ: Right–the issue is not Dexter or whether Buffett hasn’t been up-front about Dexter as a mistake. It’s a whopper, and he’s the first to admit it. It drives him nuts.
So back to Gen Re: could he have not issued $15 billion of BRK’s ‘over-valued’ stock and given it to Ajit Jain to invest? Without the derivatives and the headaches?
JM
It’s not that simple for Gen Re.
From a strictly economic point of view, it wasn’t a bad deal. If you include the regulatory issues, forget about it.
The economic argument is that Gen Re now has $22 billion of float and is making an underwriting profit, so Gen Re is far from worthless today.
BRK picked up $15 billion in float in 1998. They also got the capital. BRK was able to lighten up on its stock portfolio when some of the stocks were at their peak.
The way that Berkshire is structured, it is impossible to strictly allocate assets to business, since they are effectively pooled. If the insurance subs had only required capital, then it would be easier, but they have excess capital, so any division would be arbitrary.
Overall, I would be surprised if the deal didn’t work out as a positive, but there is no way to definitively demonstrate it.
You have to ask if Buffett got over a 6% return on the Gen Re assets. If so, then anything much over the 6% is due to the deal.
Gen Re today is worth at least what BRK paid 10 years ago. I think it is worth more. In addition, the investment gains on the assets (however you want to compute them) have to be compared to the current stock price to evaluate the deal.
It’ no Dexter.
Regarding distribution system, it is important, but I don’t think that Buffett thought that what is now BHRG would be part of Gen Re or that there would be significant synergies.
The key to Gen Re’s distribution system is that it is primarily direct, like GEICO. That is, no agents/brokers.
I think you can argue that Gen RE was and has been the premier reinsurer over the last decade — only it hasn’t been the greatest decade. Compare their performance with Munich Re and Swiss Re. The others at the time are mostly gone. Both Munich and Swiss had to raise capital and diluted equity.
I won’t argue that the reinsurance business is as good as it was (or seemed to be) 10 years ago. However, Gen Re, right now, is better then the competition not because of what it has, but what it doesn’t have.
That is, it no longer has esoteric stuff. Their derivative business is gone. They have permission to shrink — no need to chase business in a softening market. They don’t have to allocate capital. They don’t have to do a lot of retrocessions to keep earnings stable. They don’t have to think about the asset side of the balance sheet.
Let’s look at the numbers.
With Berkshire’s insurance businesses, I think the best way to determine valuation is to think about float growth. There are two components to float growth, #1 underwriting profit and #2 realized/unrealized investment gains. I am not going to parse the respective contributions of float growth by division except to note that the writedowns and losses at Gen Re have stunted its float growth over time.
Per Berkshire annual reports, since the Gen Re acquisition, Gen Re has grown float from about 4.5% annually versus GEICO’s 8.7%, Other Reinsurance at 14.2% and overall float at 8.8%.
Clearly, Gen Re has been a laggard. The only questions you can ask about its value outside of what is presented in the annual reports are #1 did Gen Re’s presence somehow contribute to BRK’s other reinsurance business growth and #2 would one expect Gen Re’s float to grow more in line with the rest of the insurance division’s float going forward.
If one assumed at the time of purchase that float was going to grow 4.4% in perpetuity, using a 12% discount rate one would come up with a value of ~$8.7B. This is the result of 4.4% yield on $15B divided by (k-g) or (12%-4.4%) If one assumed 8.8% float growth the value would be $41.25B.
Throwing some more numbers around…
WB paid $22b in stock for $17b in fixed
income investments. Assuming a 25%
overvalued stock price translates to
a roughly equal swap. Unfortunately he
then lost $6.5b from poor underwriting
and 9/11. Thru early 2003 the deal was
a dud and BRK was at best flat with its
own Brk RE having a hard time. I am
leaving out some gains from the float to
offset the losses but that needs to be
included.
Today GenRE represents $23b in float
and generates an underwriting profit of
$.5 billion. Someone has to make these
two components equal $40b – an exercise
left for the student perhaps? WB stated
in the 2007 annual report that the $3.5b
for Dexter was his worst deal. That
suggests his own GenRE valuation
approaches at least $37b.
If WB had raised $22b 10 years ago, he
would have netted ~21b after fees (or
not if the spread was different). That
$21b could probably have generated a
nice return in bonds and sold later
with the proceeds applied to buy one of
the troubled RE’s for a better price
than he got for GenRE. That’s 20/20
hindsight.
Would it have been better? Yes.
Would he have been able to raise that
much? I dont know.
Why would anyone have invested in an
insurance business when Internet stocks
were around?
NJ
cadamyale:
“I think the best way to determine valuation is to think about float growth. There are two components to float growth, #1 underwriting profit and #2 realized/unrealized investment gains.”
It sounds like you are talking about total assets and not float. Float is the assets related to insurance liabilities primarily for loss and loss adjustment expenses and unearned premiums. The only way to grow float is to write new business. Underwriting profits and investment returns are income statement items and they would be booked as net worth (shareholders equity) on the liability side with corresponding assets.
The point of float is to earn investment income on the assets, which flows into income, but pay the liabilities at their booked value which — using typical insurance accounting — are booked at ultimate nominal values.
Any attempt to model float independently of the insurance processes that generate it doesn’t make any sense to me.
GEICO float is growing rapidly only because they are writing large numbers of new auto policies.
BHRG frequently buys float via retroactive reinsurance agreements. The costs of this show up in the asset DCRA — which you could read about in his annual reports. The short version being that all retroactive insurance is booked to net $0 profit/loss when it is written.
When Buffett writes about float, he is always correct and precise, but not always thorough enough make it clear how it translates into accounting entries. When ever he talks about balance sheet items, the liability (unpaid losses, unearned premiums, for example) are only “good” in the sense that they necessarily have an associated asset (investments).
Sorry to get all technical regarding basic accounting, but it isn’t that obvious, obviously.
“Growth of float is important – but it’s cost is what’s vital.”
This comes from BKH’s 1999 Annual Report (on page 7). I think to intelligently answer Jeff’s question on whether Gen Re was worth $40 billion, one should look at whether the cost of Gen Re’s float BKH purchased has risen faster than their float.
For the record, I am not an accounting major nor an insurance analyst, so my knowledge of the reinsurance business is cursory at best.
After skimming through BKH’s annual reports , however, and doing some quick Google searches, I understand that underwriting losses, where losses incurred and underwriting expenses exceed premiums earned, impact the cost of float.
If you notice when reading BKH’s annual reports, the float cost rises/falls dramatically over time. For example, back in 2000, their cost of float was around 5%. In 2001, after the 9/11 attacks, their cost of float shot up to 12.8% (I am not making this up – readers can see this for themselves by perusing BKH’s 2001 annual report).
The cost of their float has dramatically declined since 2001, but so has the growth in their float. I think this is due to the sub-par investment results that have come from reinvesting GenRe’s premiums before paying losses which come due.
My thinking here is that you want float to grow over time while making sure the cost of the float stays low or declines over that same time period.
If float is less than what you paid for the business over a given period of time, then, to answer Jeff’s question, I’d argue, no, GenRe wasn’t worth the $40 billion paid.
I could be wrong though.
Final Comment:
You can actually look it up. http://www.sec.gov/Archives/edgar/data/109694/0000898430-98-002837.txt
BRK issued 268k shares of stock in 1998 to do the deal.
For the 268k shares, he got a company that had $15 billion in investable funds from insurance and a book value of $8 billion, which consisted of investable funds. He ended up with the same dollar amount of investable funds as he would have had he been able to sell the 268k shares for a near bubble price of $80,000 per share.
Prior to the purchase, BRK had $7 billion in investable funds from insurance (float) and a book value of of $37 billion. Given the 1.25 million A equivalent shares, that comes out to investable funds of $35k per A share.
I don’t know if Buffett could have issued 268k shares of stock in 1998 for $80,000/share, but even if he had, he would have ended up with roughly the same amount of money to invest.
Even today, if you combine float with Gen Re’s $8 billion 1998 book value you get $31 billion or $116,000 per A share in investable funds.
From a strictly financial perspective, it is hard to argue that this wasn’t a good deal.
It had to have been, given the subsequent problems. Gen Re had a couple of billion of liabilities that later emerged. The reinsurance business went to hell. September 11 cost a lot of money. 2004 and 2005 were the two worst years on record for hurricane losses.
So financially it is a Buffett type of situation. A deal with enough of a margin to survive some estimation errors and bad luck and still work out.
I would say that the regulatory problems were so bad that they make the deal negative, regardless of the financials.
This thread shows the danger of
posing an open question where
the original message requests
informed opinions. It also
shows the risk of not filling
in the blanks of a discussion
on a blog. I’ve spent more time
on this then I planned, but
this is my last response –
barring an informed question.
The original question posed by
JM: is GenRE worth $40b?
The answer is yes. I still
have not done the heavy lifting,
but a table borrowed from elsewhere
points to a current value above
$40b. That will allow the
discussion to focus on, perhaps,
just the numbers.
There is no doubt, however, that
WB would have done better by
doing nothing. The RE business was
heading for a downturn and 9/11
was just around the corner. It
took until 2003 for GenRe’s unit to
again report underwriting profits.
Its underwriting profits are still
minimal compared to Brk Re. Brk Re
experienced losses thru 2001.
I originally suggested WB was
basically exchanging overpriced
stock for bonds. It would have been
better to issue stock when it
was overvalued, but it would not
have been WB’s style.
Here are the numbers:
WB has said that in evaluating
insurance businesses, an investor
needs to know three things: amount
of float, its cost(amount of under-
writing profit or loss) and the
long-term outlook for these two.
What’s also needed though is an
assumption about the investment
returns earned on the float.
As of ye 2007, GenRe’s float was
$23b, its operating profit was $.5b
(hence its cost of float was a
negative $.5). Float at GenRe grew
at a 4% CAGR from $15b in 1998.
Over time I assume that GenRe’s cost
of float will be close to zero.
In the past BRk has had periods of
negative and high cost float. Since
1984 I come up with a total of $6b
in negative float cost. There have
been 12 years of float cost and 13
with negative float cost. Float
does tend to the negative, I’ll admit.
(I really didn’t want to do a DCF for
the blog, so I found a website where
the calculation was already done. The
link is below, its an old TMF page.)
I assume that float will continue to
compound at a 4% AGR, there will be
a zero (or negative) float cost, and
look at the three scenarios on the
website. In the case most like now
the value of the float is at least
double the amount of float. The
biggest assumption maybe that when
WB kicks the bucket then his
replacement will be able to still
exceed the risk free return. If
value of float = 2x float, then the
value of float currently is $46b.
General Points
Assets are valued as sum of DCF. In
switching back and forth I’m sorry
if I lost people. WB paid $23b for $15b in float. He got $17b in fixed income investments along
with other assets and liabilities.
-NJ
Flame Away!
TMF link:
http://www.fool.com/community/pod/2001/011002.htm?ref=foolwatch
First, Inscrutable Chicken made an excellent follow-up that might have gotten lost in the follow-up comments. If you’re just reading this one, please scroll back a few.
Second, NJ makes the premise that ‘float is worth 2x float.’
Why is float worth more than what Buffett paid? Where is a comparable reinsurer trading in the open market?
Cheers,
JM
I got props from JM! That’s made my week and it’s only Monday.
Jeff, can you clarify your initial question? your post seems to indicate a price paid at time of acquisition of both 22m and 40m. It seems to me the price paid was consideration worth 22m in 1998 that today is worth 40m. So the question is not was genre worth 40m but is genre worth 40m.
i think insurance companies should trade at a multiple of bv. the actual multiple paid would be reflective of the roe.
on a prospective basis, the argument would have been that synergies and superior management would have resulted in genre growing into the 22b price paid (at the time no one knew the consideration would later be worth 40b)
this is an experiment without a control as 911 etc. impacted actual results.
Regardless, genre has a value today , and i think your question is, ‘is that value 40b?’???
non life insurers today are trading around 10x pe; I don’t see any indication that genre earns 4b
non life insurers trade at around 1.4x bv (brk.a trades at 1.6x) Lets set aside the whole issue of the associated goodwill,
Assets attributed to genre are 32b, brk.a as a whole has an e/assets of .44; that means genre equity may be approx. 14m. On a Comp basis this would indicate a value of 20b, at brk.a’s own multiple a value of 22.4b
we dont really know what the equity attributable to the genre business is; non life insurers trade at .42x assets and brk.a trades at .7x assets. neither imply a 40b value.
my conclusion on a very rough basis therefore is that:
if genre were a standalone company today its market value would not be 40b
if wb were to sell genre today he would not receive 40b
if wb were to buy genre today he would not pay 40b
but if your question is was genre worth 40b in 1998 not sure how to address that…
Dan,
That was exactly the question: is Gen Re worth $40 billion today.
We know it was worth $22 billion at the time of sale, because Warren Buffett paid $22 billion for it.
What we’re trying to understand is, given that the value of the equity paid for Gen Re is now worth $40 billion, has it been a good deal? Could BRK sell Gen Re for $40 billion today?
So far, it appears not. But further informed opinions are welcome.
JM
Jeff:
I went ahead and took a crack at answering the question on my previously unused blog. http://capitalvandalism.blogspot.com/
The question of what it is worth today and whether it was a good deal are different, since Gen Re funds were used for other investments.
In addition, at today’s BRK stock price, $125k, the cost at today’s share price is $33.5 rather then $40 billion.