Universal Insurance is an US-based P&C Insurance company which has been on my extended watch list for some time now. Why ? Well, the company always traded cheap (single digit P/E), was very profitable (~31% ROE for the last 9 years on average) and growing strongly year by year (400% over 9 years). So from the outside this looked like a cheap but highly profitable growth stock.
The main reason why I didn’t analyze the stock further is that Universal is a specialized homeowner insurance company which almost exclusively operates in Florida.
The company has a market cap of currently ~620 mn USD.
Not surprisingly Universal now is in a tough spot as “Irma” is creating havoc on Florida as I write this post. The stock price has dropped by around -30% by Friday:
Florida Insurance market
According to the latest 10K of Universal, the Insurance market in Florida is very special:
- it is dominated by “local” players with Universal being the largest player (10% market share)
- Following strong hurricane seasons in the early 90ies and then again in 2004/2005, national insurance companies reduced their exposures
- There is a state-owned insurer called “Citizen’s” which acts as “insurer of last resort
- there is mandatory Reinsurance offered by the Government:
All residential insurance companies that write business in Florida, including us, are required to obtain a form of reinsurance through the Florida Hurricane
Catastrophe Fund (the “FHCF”), a state-sponsored entity that provides a layer of reinsurance protection at a price that is typically lower than what would otherwise
be available in the general market. The purpose of the FHCF is to protect and advance the state’s interest in maintaining insurance capacity in Florida by providing
reimbursements to insurers for a portion of their catastrophe hurricane losses. Currently, the FHCF provides $17 billion of capacity annually to its participating
insurers, which may be adjusted by statute from time to time
So it is quite interesting that only local players serve this market. Normally, local players should have no chance as the bigger players have much better diversification and should be able to offer the same policies cheaper, but for some reason (maybe because of the Government activities ?) this doesn’t seem to work here.
The Universal Reinsurance program
Due to the nature of its exposure, it is no surprise that Universal spends several pages on explaining their reinsurance program. I try to summarize this in my own words as good as I can:
– Universal has a net retention of only +35 mn USD for the first Cat event they encounter in a season in Florida plus 5 mn for all other states. The cover reaches until a total loss of around 2,5 bn for the first event
– they have additional protection for the next 3 storms as well, however net retention increases to 55 mn USD and the maximum coverage for those events is around 800 mn USD
All in all they seem pretty well protected even under adverse scenarios.
We have used the model results noted above to stress test the completeness of the program by simulating a recurrence of the 2004 calendar year, in which four large
catastrophic hurricanes made landfall in Florida. This season is considered to be the worst catastrophic year in Florida’s recorded history. Assuming the
reoccurrence of the 2004 calendar year events, including the same geographic path of each such hurricane, the modeled estimated net loss to us in 2016 with the
reinsurance coverage described herein, would be approximately $84 million (after tax, net of all reinsurance recoveries), the same as it would have been in 2015.
So losing 84 mn doesn’t sound too bad. But not so fast. at the end of the Reinsurance section we find this sentence:
The third-party reinsurance we purchase for UPCIC is therefore net of FHCF recovery. When our FHCF and third-party reinsurance coverages are taken together,
UPCIC has reinsurance coverage of up to $2,487 million for the first event, as illustrated by the graphic below. Should a catastrophic event occur, we would retain
$35 million pre-tax for each catastrophic event, and would also be responsible for any additional losses that exceed our top layer of coverage.
This last sentence is important: If the total loss is above 2,5 bn USD, they have to pay every single dollar themselves.
How large could a loss be ?
To get a feeling for what is at stake, one can check page 11. Universal has written coverage for a total of 135 bn USD in Florida alone. So the big question is. How much will be the final damage as the reinsurance program covers as we discussed only 2,5 bn Usd.
There may yet be a Florida insurance market on Tuesday,” said Charles C. Watson Jr. of Enki Holdings L.L.C., who had been projecting losses of $172 billion on Sunday morning but just $49 billion by evening.
According to the 10k, Universal’s share of totally insured property in Florida is 5,5% (Page 9). Without knowing the exact distribution and just applying this percentage, Universal’s share of the losses could be between (5,5% *49) = 2,7 bn in the lower case and 9,5 bn in the upper one (assuming the estimate are only homeowner property losses).
I guess they would still survive the 2,7 bn hit but a 200 mn extra loss (pre tax) would wipe out close to 50% of their equity and might require them to raise capital.
However for me it is not clear what percentage in the above quoted damage is property damage or not. So the loss could also be a lot lower.
Interestingly they also quote Citizen’s projections for a large cat event:
Today, Citizens is about one-third its former size. With fewer policyholders, it estimated that it would get only $6.6 billion in claims in a 100-year storm.
Citizen’s has insured around 30% more in USD value than Universal and most likely also worse risks but still, it looks like that for Universal a loss above 2,5 bn is not totally unrealistic if Irma was really an extraordinary event.
Plus one should not forget that this is just the start of the Hurricane Season.
Other Florida only players:
There are two other Florida insurance plays: Heritage Property & Casualty Company and Federated National Holding Company. As the stock chart shows, all stocks have been hit equally hard in the last few days:
Without any deeper analysis I guess that those peers have the same issues than Universal.
So what to do now ?
Normally common wisdom in Insurance is to buy the stocks after a big Cat event happens. Usually, insurance market “harden” after such events and the higher premium then leads to nice profits going forward.
In order to make money from this as a shareholder, there is one big caveat: You have to survive.
And in my opinion this is currently the big issue with Universal and its peers: I am not 100% sure if they survive or at least need a significant capital infusion.
Insurers do have some instruments in these situations, such as booking claims slowly and play for time, but I am not sure if this works here when the damage is just too large.
A good example against buying after a big cat event is Thai Reinsurance, the dominating Thai reinsurance company after the disastrous flood in 2011. Prem Watsa’s Fairfax bought 25% of the company shortly after the cat event
. Looking at the stock chart it is easy to say that Prem lost significant amounts of money with this purchase as the company never really recovered:
I don’t really know what happened there but these days capital is plenty and markets “harden” only for very short periods and might not offer the chance for the old players to earn it back.
This is also one of the big perils of Nat Cat insurance: These businesses can look very good for long periods of time but then get knocked out quickly if a larger than modelled event appears. It will be very interesting to see where in the Reinsurance industry these losses will ultimately pop up.
So from my perspective, I will watch this one closely, but only to learn and not necessarily to invest.
what do you think of Trisura? Recently spun off from Brookfield Asset Management.
Never heard of them
Great answer. Here’s the link if you are interested:
I found an interesting comment in the yahoo board:
—UVE has ~573,000 homeowners policies in FLA, 220,000 of those are located in the higher-risk areas (Broward, Miami and parts south). In order for UVE to blow through their reinsurance, they would need to incur ~4,800 in losses for every single policy. Also, keep in mind that most of the policies written in FLA have substantial wind deductibles (2,000 or more per policy). So, windstorm damage would have to be at least 7,000 or more for every single policy on a ground-up basis before reinsurance limits are reached. I cannot see that happening, since wind speeds are diminishing. Also, flood is excluded on UVE policies and separately covered.–
hmm, 7000 USD doesn’t sound like much to me, I am not sure what a new roof alone costs in Florida.
And to play with numbers: The 573 k of policies cover ~99,3 bn insured value which then translates to ~173k USD insured value per policy. So the 7000 USD mentioned represent only 4% of insured value which then doesn’t sound like such a big number.
We will see what happens, but I think at the moment there is only speculation. And don’t forget: Hurricane season just started and for the next storm it will look VERY different as those numbers only apply for the first event…..
I’ve held GNW since January this year. I think the merger is a complete distraction – it will get barred by CFIUS. See what happened to Lattice Semiconductor who did their initial filing about 40 days before GNW – they’ve just announced they’re going to ask Trump direct – and he’ll be even more mad than on an average day to let it through. I personally think GNW is undervalued – yes they keep taking hits on the LTC business, but they still have like a $16bn discount to book equity value. At some point their models will stop deteriorating, and if this happens before 16bn of equity is consumed, you’ve bought in at an incredible valuation. So my thesis is that LTC will continue to deteriorate some more but not enough to account for the gigantic discount to book. But it’s definitely a long slow play, and there’s more risk than I fully understand.
MMI – another great article. I’ve looked at several insurers – including UVE (which I agree with you on), and the thing that I can’t get my head around is how their equity behaves in a rising rate environment, when their fixed income portfolio ought to be loss making from a book value perspective. Do you have any thoughts on this?
In short: (slowly) rising interest rates would be good for insurers because then they can earn more on their float. The mark-to-market effect is only temporary.
I get that at the high-level hand wavy explanation. But I decided to dig down and try to quantify exactly how this looks as I feel understanding what ought to happen when rates start to rise will be pertinent in the coming years.. So you’ll forgive me for using your blog comments section as my library… Here goes.
I’ve been looking at Employers Insurance Group. It’s a small US insurance business of which the only special thing for me is that I’ve looked at it because it had a low p/e. I never invested. In any event, from their 2016 10-k, we find the following:
“As of December 31, 2016, our investment portfolio, which is classified as available-for-sale, consisted of 91.9% fixed maturity securities. We strive to limit the interest rate risk associated with fixed maturity investments by managing the duration of these securities. Our fixed maturity securities (excluding cash and cash equivalents) had a duration of 4.3 years at December 31, 2016. To minimize interest rate risk, our portfolio is weighted toward short-term and intermediate-term bonds; however, our investment strategy balances consideration of duration, yield, and credit risk. Our investment guidelines require that the minimum weighted average quality of our fixed maturity securities portfolio be “AA-,” using ratings assigned by S&P. Our fixed maturity securities portfolio had a weighted average quality of “AA-” as of December 31, 2016, with 57.2% of the portfolio rated “AA” or better, based on market value.”
So we can simplify EIGs portfolio as 90% AA bonds of duration 4.3 years and 10% equities, for which I’ll just assume the S&P to keep things simple.
The key number that tells you the sensitivity of a portfolio’s price to interest rate changes is modified duration. Investorpedia has a calculator here: http://www.investopedia.com/calculator/mduration.aspx
The St Louis Fred tells us that the BofA Merrill Lynch US Corporate AA index is currently yielding 2.6% (https://fred.stlouisfed.org/series/BAMLC0A2CAAEY). Unfortunately I can’t find info on the index’s duration anywhere but given that 10 year t-bills are today yielding 2.24% I would guess the index duration is somewhere in the right ballpark of 4.3 years. So if we plug in 4.3 years and 2.6% in our modified duration calculator, we get an MD of 4.02.
For the S&P portfolio, the answer is a bit more hand wavy as I can’t find a good source which tells me how to find modified duration for an equity portfolio. But if we assume (very roughly) that the S&P generates owner earnings at 5% in perpetuity, which doesn’t seem mad given current valuations, this gives us a modified duration of 19.2.
Thus our blended portfolio modified duration for the EIG investment portfolio is 5.5. So if interest rates go up by 1%, the EIG portfolio ought to drop in value by 5.5%. Given that EIG as a business generated a return on book equity of 12.7%, this tells me that it will take roughly 0.4 years for the business to generate sufficient value to make up for what will be lost in the event of a 1% rise in interest rates. And then on top of that, there will be the added benefit of greater income from the portfolio going forward.
So net net, the reason rates are not a huge risk for EIG is that the bulk of the investments are in relatively short dated bonds.
As an interesting aside, there is a slightly scarier calculation that can be done. US banks hold $750bn of T-Bills and about $1.6bn of Mortgage Backed Securities (https://www.bloomberg.com/news/articles/2016-10-30/banks-amass-2-4-trillion-hoard-of-bonds-as-bofa-leads-stampede). The MD of a 10 year T-Bill is 9, while if we assume that the average MBS yields double the 10 year bond with a 15 year duration, we get an MD of 10.8. Weighted, banks portfolio duration then becomes 10.25 – let’s call it 10. This means that for every 1% increase in rates, the banks fixed income portfolio will lose $240bn of market value. When you compare this against US banks total equity of $1.8trn (https://fred.stlouisfed.org/series/USTEQC), that’s a 13% reduction in bank equity for every 1% the fed raises rates. You can see why they’re so reluctant to do it and why in reality rates cannot go up by that much anytime soon.
Anyways, thanks for stimulating me to do this info, and I hope you don’t mind me posting such long comments not really related to the original article.
I agree that the merger is a distraction, my assumption is that it keeps down the share price more than without merger talks. Every GNW investor just hoped to get freed from the agony with the merger. I believe that if this merger is finally from the table (somewhere this year) the stock might take another hit but will catch up once everyone starts focusing on the current business and due debt in 2018. Then the stock should be trading higher than today. Then again, that’s what people have been saying about GNW for as long as I can remember 🙂
Thanks for the write-up, mmi. What do you currently think of GNW? I know your stance on life insurers, and I agree that it’s bad business. But at current share prices, there seems upside whatever happens. Spread with the deal price with Oceanwide is huge because everyone thinks the merger won’t get approved and that GNW is dead if it doesn’t. But there is enough cash available to pay the 2018 debt, so I think a default scenario is very unlikely. In that case, and if GNW continues to reprice existing contracts and write profits as they did in the last 2 quarters, I think a share price around 4,5 USD for the no merger scenario is conservative. My odds:
15% merger goes through at 5,43 USD
10% renegotiated merger at 4 USD
70 % no merger, 2018 debt paid at 4,5 USD
5% no merger, default at 0 USD
That results in 4,36 USD in short term (<1 year) and I feel I'm not going out on a limb here. At current prices (quite volatile in the last weeks) that is around 20% upside. As I said: not the business for a lifetime investment, but an interesting special situation?
Relating back to your article, there is one more thing I wanted to write. As I’ve learned from you and other people who know more about insurance than I do: insurance business is great if you can underwrite well. I think the variables (catastrophe occurrence) and the impact (1 occurrence can wipe out several times a statistic average of claims) are just too unknown for nat cat to underwrite reliably. The difference is huge with e.g. car insurance where none of a lot of single, non-related events have a big impact.
Warren Buffett has shown that you can underwrite Natc Cat profitably for a long time. The trick is to have a large diversified book of business.
Thanks for your reply, you’re right!
hmm, I really just don’t like GNW.
plus a Chinese buyer….no thanky ou.
Update: the gap on my case got closed this week on some unspectacular news, so I’ll just forget about it.