Lancashire Holdings: at a premium?
A few weeks ago I participated in an interesting discussion here about the pitfalls of accruals, and how companies with accruals-based earnings are net underperformers, over time. This creates a big problem if you want to invest in financial firms, who are the "poster boys" of accruals-based accounting. In this context I thought it would be interesting to take a look at the one financial stock I hold, Lancashire Holdings (LON:LRE).
Just who are Lancashire, then?
EPIC: LRE
Market Cap: £1.13bn (approx)
Current share price: 700.5p
Website: www.lancashiregroup.com
Lancashire Holdings (LON:LRE) listed on AIM in 2005, raising capital to take advantage of perceived opportunities in speciality insurance arising after Hurricane Katrina. It transferred to to the main market in 2009 after an astonishingly succesful four years and is now firmly established in the FTSE 250.
Lancashire are a Bermuda-based P&C (property and casualty) insurer and reinsurer, with Lloyds- and non-Lloyds based operations. In other words, they insure and reinsure risks such as offshore/onshore oil and gas operations, hurricanes and storm damage, other energy (e.g. power plants, nuclear), shipping, aviation, anything really except direct consumer facing stuff (such as say Admiral or Churchill) and life/life-related business. Their business is largely done through brokers so they are very different from the consumer facing insurers and more comparable with the likes of Amlin, Hiscox, Brit, other Bermuda-based insurers such as Catlin, and the many US P&Cs.
Like most insurers (and all P&C insurers) they make money in two ways: firstly, by paying out less in claims and expenses than they take in in premiums (the ratio of claims and expenses being known as the combined ratio, one of the most important metrics in insurance investment) and secondly by investing premiums and profiting on the investment yield on that premium before claims arise for payment.
So simple it's complicated
This sounds very simple, and, well, it is! All P&Cs operate within this basic model and therefore should, in theory, be highly comparable. The problem that arises is that while the idea is simple, the practice is not. This is for exactly the reason highlighted in the accruals article I linked to - namely that, at the point of sale (i.e. when a policy is concluded) the insurer does not know his cost of sales - and will not do so for many years. He still has to estimate it, though, and book it in the accounts (giving rise to the numbers that allow us to calculate the combined ratio). At the same time, he has to set aside appropriate reserves to meet claims as they are made. As claims sometimes arise many years after the policy was made, this is a process that is uncertain at best and as you can imagine, the quality of reserving and estimation is, ahem, variable ;-) across the industry.
Add to this a tendency to overdeploy capital when prices are low, in order to keep market share, hoping to make the difference back by relying on investment returns, and you have a recipe for investing disaster. So how do Lancashire Holdings (LON:LRE) stack up against this seemingly unpromising background?
Corporate principles
Lancashire Holdings (LON:LRE) state that they have three basic guiding principles, which are usually plastered at the front of their results releases:
- Underwriting comes first
- Effectively balance risk and return
- Operate nimbly through the cycle
The aim of these is to avoid the pitfalls of overdeploying capital and reaching for market share at the wrong time, relying on investment returns to make up the difference.
That's all very well, but how do they stack up in practice?
Combined ratio
Lancashire Holdings (LON:LRE) 's combined ratio since inception is 57% (i.e., for every £1 in premiums they write, they have paid out £0.57 in claims). This is an astonishing record that far outstrips normal industry performance. The nature of the business is that the combined ratio fluctuates considerably but the table below (taken from the 3Q11 investor presentation) sums up their expertise neatly:
| 2006 |
2007 |
2008 |
2009 |
2010 |
2011 YTD |
|
| LRE combined ratio (%) |
44.3 |
46.3 |
86.3 |
44.6 |
54.4 |
60.7 |
| Sector avg combined ratio (%) |
76.1 |
76.6 |
88.0 |
78.1 |
87.9 |
118.2 |
| LRE outperformance (%) |
31.8 |
30.3 |
1.7 |
33.5 |
33.5 |
57.5 |
As the table eloquently shows, Lancashire Holdings (LON:LRE) have outperformed their peers in each year, and in four out of five years substantially so. Nor have they (as yet) suffered a net loss from insurance operations in any year so far - this in a period when the market has generally been soft, and there have been numerous extremely large loss events. It is an exceptional performance by any measure.
Reserving
Another key area to look at is reserving. Complex losses of the type LRE and their peers insure take a long time to develop and assess accurately. Reserving history (i.e. the development of reserves set aside for future payments on policies written in the past) is therefore a key factor in judging management competence (and honesty...;-)). £LRE's relatively recent vintage means they only have five years of reserving history, but there have been net releases from reserves every year for each accident year so far - in other words £LRE's reserving appears to be prudent and cautious.
Investment performance
Investment is the other side of the coin from insurance results - having collected premiums, how they are invested can be a key driver (positive or negative) of long-term returns. LRE have a very conservative investment policy which seeks to conserve capital, rather than aggressively to grow it (unlike insurers such as, say, Berkshire, Markel and Fairfax, for whom investment returns have been almost as important as insurance results in growing value over time). Five year annualised investment returns have been 4.7%, for a sector outperformance of 1.2%. However, the markets of the last five years have been unkind to the more aggressive approaches of some of £LRE's peers and in the event of a lengthy 'risk-on' phase LRE would almost certainly underperform its peers' portfolios. Its bearish positioning in (relatively!) safe debt securities, with short average maturities (and no equity exposure whatsoever) should help conserve capital come hell or high water, however.
Capital allocation
Another key indicator of a well-run insurer is what they do when they think there is excess capital and insurance opportunities are scarce. Do they take money off the table or do they chase market share? All of them talk the talk of not chasing market share, but who walks the walk? LRE's record here is quite simply superb. In addition to the combined ratio above, LRE has returned $1,298m to shareholders, rather than write unpromising business. This has taken the form of:
- share repurchases: $311.5m
- special dividends: $918.1m
- ordinary dividends: $60.1m.
To put it in context, 132.7% of the capital raised at IPO in 2005 has been returned to shareholders (without counting the capital gain on the shares which effectively doubles the total return).
Summary
As this post is getting rather long, it seems appropriate to try to sum up here. There's no insurance company that offers the same exposure to overnight step changes in value that one can benefit from in oil and gas E&P, and it's crucial to be very careful to look at management's ability to allocate capital sensibly when buying into any financial. Having said that, even though the current investment environment is problematic for insurers, with low yields and increasingly scarce risk free assets, if you find talented management in an insurance company, they should be able to increase book value per share year in, year out - reliably hitting singles rather than swinging hard for occasional home runs. Lancashire Holdings (LON:LRE) are the best run non-life insure in my book. They should be able to maintain an RoE of 15-20% year in year out( barring exceptional accident years) and I am comfortable that at any price below, say, 1.3x book value, LRE will provide consistently decent returns, at relatively low risk.
Any thoughts and comments are very much appreciated.
Disclaimer
This is not financial advice. Our content is intended to be used and must be used for information and education purposes only. Please read our disclaimer and terms and conditions to understand our obligations.
7 comments
Quite a lot has happened to Lancashire Holdings (LON:LRE) since I last posted on this thread. Firstly, Lancashire's share price has recovered (and then some) since its decline on the news of the increased loss estimate I referred to in my post 4. More importantly, however, there have been four key events:
1) Resignation of Neil McConachie as an executive director;
2) Preliminary estimate of costa concordia related losses of approx $35mm net;
3) Announcement of preliminary results for 2011 (along with a slew of preliminary results from competitors);
4) Expansion of Accordion sidecar.
The preliminary results can be found here and the headline figures are as follows:
These results in miniature demonstrate neatly the problems that have faced the industry over the course of 2011: a very bad loss year (which may well turn out to be the worst, and is certainly one of the worst, in industry history) combined with anaemic investment returns as maturing investments have to be rolled into new, lower-yielding issues. Set against that, however, are suggestions that in light of the capital destroyed by this year's losses, rates are turning upwards. Though it would be premature to call this the start of a much needed hard market, it's clearly a positive for the year to come.
Obviously it's disappointing that LRE has lower investment income and a worse underwriting performance than last year, but that was not surprising given the state of the general market (and also the fact that it was clear in the third quarter results that the full year figures would look something like the above). Also, given that reporting season for LRE and its peers has now concluded, we have a set of comparators against which to judge its performance, and these (set out in the table below) give some interesting context to the results:
13.4%
0.9%
As the above neatly shows, Lancashire has had yet another industry leading year. Amlin's results are interesting, as Amlin has historically been a very well managed P&C, and one that I have considered taking a position in in the past. Its bad 2011 certainly tempted me to invest but a closer inspection put me off somewhat: firstly, it's having trouble integrating a recent acquisition, and secondly, its asset position concerned me. Amlin has a large position in French government bonds that unsettled me slightly and I would prefer to get paid a bit more to take that risk on than is implied in the current SP - if Amlin were available at or close to book, it would be very interesting, but at 1.4x there's quite a lot more risk IMHO.
On valuations, LRE is (unsurprisingly given the above performance) the most richly valued: it trades at 1.6x book (Amlin is at 1.4x, Catlin at 1.05x, Hiscox at 1.29x and Beazley at 1.1x). Even at that relatively toppy price, a float-based valuation would almost certainly suggest LRE is in, or close to, 'buy' territory. I am a little more cautious, and am not prepared to add at current prices, but will certainly look to increase my holding if LRE drops to 1.2-1.3x book. One reason is that part of the explanation of LRE's superior profitability is its discipline in turning down unprofitable business means that it can only grow if the market hardens substantially, so the key to getting a good return is to make sure your entry price isn't too high (LRE writes roughly the same amount of business as it did five years ago, so it isn't going to increase value by growing substantially).
The sector as a whole seems reasonably fully priced to me, so it's a case of maintaining a watching brief for the present.
Banking, not teaching....though I could have taught some bankers a thing or two ;-)
I'm more bothered about the liability profile, because that has the potential to blow a big hole in profits if their underwriting performance weakens from the apparent high level of the present. Re the assets, there won't be much matching required if they really do have a short-tailed liability book......the default asset of choice should be T Bills and short Treasuries (or equivalent in other liability currencies) - so why do they appear to own riskier stuff?
Some of that may be true of course. But some of it may just be down to luck. It is important to understand whether they have some sustainable advantage in their approach to underwriting - or whether they have just got lucky with the insurance equivalent of the drillbit....especially since this is an accruals business and changes in business profile can have multi-year impacts.
No rush though, as I said, LongbeardMinor ;-)
ee
Well I have finally got round to doing my homework! Well, the first part of it, anyway.
Taking the issues in turn:
1) Short-tail liabilities.
Lancashire underwrite and reinsure business in four sectors - property, energy, marine and aviation. The vast majority of that business is of the type where the existence of a loss would become apparent very quickly (probably during the accident year in question). Quantum of that loss is a different question, and can take many years to play out, so it's certainly not the case that "short tail" means that they will know during the year what their liability is going to be (this is apparent from the fact that they still have reserves for the 2006 accident year (their first) which, while shrinking, will still be on the balance sheet for some years yet). Having said that, most of the uncertainty revolves around the amount of a claim, not its existence, and 'incurred but not reported' (IBNR) is running at around 35% for the portfolio as a whole IIRC so two thirds of the reserving is 'just' about valuing reported claims. It's clearly better to know of a claim but be uncertain as to its final amount, than having to guess at the level of claims as well as quantum.
To put the 'short tail' in context, the risk disclosures in the 2010 AR state that, of $489m of insurance liabilities, $184m would fall due in less than a year, and $182m in 1-3 years. Obviously that can only be an estimate, but it gives some feel for the timespan.
They have avoided areas such as professional negligence, medical, and D&O, where claims can fester as IBNR for many, many years (asbestos is the notorious example). Not only is this easier to value, it is an area where arguably insureds understand the risk better than insurers - an unfortunate information asymmetry that obviously leads to poor underwriting.
Incidentally, Lancashire Holdings (LON:LRE) was down in trading today due to an increased loss estimate for the Japanese earthquake earlier this year...
2) Tax residence move to the UK.
I wouldn't want to second guess too much on the reasoning for this, but I wonder whether it's down to the reforms in the UK branch profits exemption and the impending implementation of Solvency II (the new EU directive regulating insurers' capital). Solvency II is going to push insurers to operating a branch (rather than a subsidiary) based corporate structure, and the branch profits reforms allow better use of overseas branch losses (as I understand it). That's just a guess though. The CFC reforms are another possible reason, so who knows?
Also, they're only moving the tax residency of the holding company - it would be impossible to move the tax residency of the Bermuda operations! So the Bermuda subsidiaries will still be subject to the Bermuda tax regime.
3) Accordion
Accordion is a sidecar facility. This is effectively an SPV, which is part owned by Lancashire (as to 20% in this case), with the remainder owned by passive outside investors. Lancashire's reinsurance subsidiary then cedes reinsurance to the sidecar. It is a way of risk sharing that allows LRE to write larger amounts of reinsurance which is then passed on to the sidecar and spread among the sidecar's investors in proportion to their holdings (it will be fully collateralised so the credit risk to LRE is minimal). It's basically a quick way of accessing capital to deploy to attractive reinsurance opportunities, and allows outside investors to benefit from Lancashire's reinsurance expertise (as it will be Lancashire that selects the risks which are reinsured and then passed on to Accordion) - while also ensuring that Lancashire is not on the hook for 100% of any loss (its capital commitment is for $50m out of a total Accordion capacity of $250m). Sidecars are not uncommon in reinsurance and tend to have a limited shelf life: Accordion has been formed in response to specific pricing opportunities and will probably be wound up in due course (LRE have used, and then wound up, at least one sidecar in the past). It's all part of what they refer to as 'operating nimbly through the cycle' - i.e. responding quickly and flexibly to allocate the appropriate amount of capital to the areas with the best risk:return characteristics.
4) Underwriting profitability
Why are LRE so much more profitable? This is the crucial question really. I don't think there's one neat answer to it. We can certainly rule out pricing power as a reason for better than average results, no insurer has pricing power! So then it comes down to luck, fraud or some sort of managerial or structural advantage. Let's take those in turn.
Could LRE be lucky? It's possible, but five years of industry leading performance - in four different areas of coverage - is enough of a sample size to make me consider this unlikely, especially when LRE has managed to maintain profitability in areas that have experienced significant losses which have caused underwriting losses for competitors. There have been some bad accident years for LRE's portfolio of risks in that time span, and LRE has inevitably had exposure to those losses; the fact they've contained the losses and remained profitable speaks volumes for their capital and risk management IMHO. It's hardly been a benign environment for P&Cs, in fact, in terms of the market conditions you might argue they've been unlucky.
So what about fraud/misstatement/misestimation of risk, as suggested by shipoffrogs? Well it's an accruals based business...you never know definitively until all the accruals have unwound. All the same, LRE have:
-experienced net releases from loss reserves every year for each past accident year
-kept a handle on their top line; they write roughly the same amount of business now than they did in 2006, despite having generated ~$1billion of surplus capital in that period
-returned the vast majority of surplus capital to investors
-maintained a disciplined balance sheet with very low gearing and substantial net cash
-had positive operating cash flows of $188m (HY11), $269m (2010), $278m (2009), $360m (2008), $521m (2007), and $314m (2006).
All of these say to me that misstatement is unlikely (though of course, not impossible) as an explanation for the level of the CR.
Ok...that's enough for now...part 2, hopefully tomorrow, will cover why LRE might actually be better at underwriting than everyone else, what they mean by BLAST, and also look a bit more at the asset side of the balance sheet.
Thanks for the detailed reply.
All very fair comment - though sometimes quantum can be crucial. For example, BP's problem in the Gulf of Mexico may attract four times the penalty per barrel if negligence is proven than if it isn't (ballpark $20bn vs $5bn...IIRC). Such things are likely to turn on a court interpretation of the evidence (due soon in that case, incidentally). If one was exposed to such an event, where would you pitch your reserves estimate? Would you aim to cover the worst case scenario or would you (effectively) hope for the best? There is no right answer, really, given the scale of uncertainty.
Accordingly, the approach to reserving can have a material impact on the reported numbers - as implicitly evidenced by your comment Incidentally, Lancashire Holdings (LON:LRE) was down in trading today due to an increased loss estimate for the Japanese earthquake earlier this year... where the market has obviously had a negative surprise (which one hopes is not part of a systemic bias towards under-reserving...which is a potential factor to add into your list of possible reasons for superior profitability?).
Thanks for the explanation re Accordion - all very clear.
rgds
ee
Your article doesn't explain why their combined ratio is so much better than the competition, who I presume are not dolts.
Northern Trust were lauded for producing much better metrics than their competition, there was an explanation for that, but for most people it only became apparent too late.
Do you have an understanding as to why their CR is so good?
Cheers
Excellent write-up!
Not normally the sort of thing that I would consider at all, given the mixed history of reinsurers and the wide variations in the way they are run - though, as you say, it's crucial to be very careful to look at management's ability to allocate capital sensibly.
I'm interested mainly in their potential liability profile. How do they control their risk profile? I notice that the website makes reference to "short-tail" liabilities - which is a good thing from a valuation standpoint and avoids potential issues such as long-term medical/catastrophe risks - but what exactly do they mean by that in practical terms?
I also see from their latest quarterly that they are about to move to become UK tax-resident (if I've correctly interpreted the slide).
What is "accordian"? I notice they make a big thing of it and have a 20% stake?
I'm interested to see that until recently they ran an equity portfolio and that they still have 11% of assets in BBB credits or lower and a further 21% in A rated assets. I'm not sure how that squares with their "number 1 rule" which is "don't lose money on investments".
What does their BLAST model purport to do?
How does their risk mix on slide 11 compare with the industry mix?
Just a few questions - no rush.
ee
Let me guess, EE - you were a teacher in a former life...you certainly know how to set homework, anyway!
All excellent points though, as is shipoffrogs' - I will endeavour to answer in due course. Liability profile is definitely a key area to focus on IMO, though cannot be completely divorced from the assets as the asset portfolio will be designed with some level of liability matching in mind. (Also, one has to be very confident that the assets will be money good when claimed upon, so they need as much watching as the liabilities in truth.)
The slide you highlight underlines something that I think you have touched on in relation to E&Ps(?). The value of a £ of premium in one company is different to a £ of premium earned by another - different parts of the market, different underwriting, reserving and investing strategies make comparisons that seem straightforward rather more complex than they appear at first blush, just as not all barrels of oil are equal.
Anyway, hopefully I will be able to turn my mind to my homework later this week ;-)