Description

­I am recommending HCI Group (HCI) – a Florida-only homeowners insurance company – as a short.  The company is priced at 2.6x book value, which pricing accrues from high recent ROEs and earnings, leaving the company priced at only 7.7x LTM earnings.  My view is that earnings are unsustainable, the company is a commodity business with a dry moat, state regulatory changes will erode HCI’s ability to generate premium, and valuation should therefore be tethered to book value, not earnings.   

Background & Context

HCI Group is a Tampa-based homeowners insurance carrier that only holds Florida policies.  The company was incepted in 2006 following the exodus of several large national carriers following the disastrous 2004-2005 hurricane seasons.  In the wake of this decision, a group of smaller, “homegrown” Florida-based insurance companies filled the void.  HCI was one of these small local carriers.  Here’s an article that covers the topic, including HCI.  

Like several of these small carriers, HCI does not generate policies via agents, the internet or like distribution.  Instead, HCI generates premiums by assuming policies from Citizens Property Insurance Corporation – the state government-owned “insurer of last resort.”  Citizens was formed in 2002 and is intended to provide insurance coverage for homeowners who cannot obtain coverage through the private markets.  Citizens’ mandate allows the company to insure policies for which there is no private market rate within 15% of Citizens’ rate.  Not surprisingly, this government intervention in the market did not evolve as intended, and Citizens consistently holds over 20% market share – much larger than originally conceived.

To jettison policies and shrink its policy base, Citizens “depopulates” its policy book by allowing private insurance companies to assume (“take out”) policies from Citizens.  A company like HCI approaches Citizens, indicates a number of policies it would like to assume, HCI culls through the Citizens’ policies and the two parties agree on the specific policies to be assumed.  HCI then notifies the selected policyholders who then have the right to opt out of the assumption and remain with Citizens.

Citizens attempts to shrink its policy base via depopulation, but the company cannot stay ahead of the tide of new incoming policies that offsets Citizens ceded policies.  As evidence, here are Citizens’ policy counts over the past five years along with policies depopulated in the same year:

2013 – Policies 1.2m (Jul 31), depopulations 145k YTD

2012 – Policies 1.3m, depopulations 277k

2011 – Policies 1.5m, depopulations 54k

2010 – Policies 1.3m, depopulations 60k

2009 – Policies 1.0m, depopulations 150k

The beast just can’t seem to shrink.

So this is the background.  HCI is an insurance carrier that insures a single product type in a single state with no distribution channel other than policy assumptions from a state-run insurance company.  One could not conceive of a more commoditized, less defensible business model.  

What would you pay for such for such a business?  2.6x book value, seriously?  To put that valuation in context strong specialty insurers with defensible business models like Markel and WR Berkeley are priced ~ 1.2 to 1.5x book.  For further context, in 1996, Buffett paid a premium to buy the remaining 100% of GEICO (he already owned 50%).  GEICO is the arguably the best PNC company in the world with a sustainable cost advantage that has enabled it to grow market share from 2.5% at the time of the deal to 9.7% now, all the while generating an underwriting profit.  Over the past 13 years, GEICO has grown materially while achieving a 93.6 average combined ratio.  That’s simply unfathomable and without peer.  The price Buffett paid, control premium included?  2.4x TBV – less than HCI now trades.

In short, HCI is materially overvalued.

HCI is presently priced at $35.60 against book value of $13.44 (2.6x) and LTM earnings of $4.60 (7.7x).  Those earnings equate to an ROE of ~43%. 

That’s a very high ROE.  How can a commoditized insurance company earn such high returns?  There are a couple of points here.  First, hurricane and tropical storm activity has been placid since HCI incepted.  Second, and more importantly, HCI executes a hurricane season-based arbitrage in terms of when it assumes policies from Citizens.  Over the previous four years HCI has executed four rounds of Citizens policy assumptions – all four occurred in November or December.  By waiting until the end of hurricane season to assume policies, HCI is able to pull a large premium flow through its income statement without having to reinsure against the premium until the following June (when hurricane and reinsurance season rolls back around).  This strategy produces seven months of very high margin revenue.

For instance, in November 2012, HCI assumed 60,000 policies from Citizens, increasing its policy count by an estimated 60% in one fell swoop.  These large policy assumptions are critical as HCI has a high policy churn rate (I estimate ~30%) and no natural way to otherwise sustain premium.  Thus, without these large periodic policy assumptions HCI’s earning have a material “decay effect.”  Specifically, HCI earned ~$1.78/share in 1Q13 on the heels of the Nov 2012 policy assumption.  I estimate (using a combined ratio of 80% - 45% loss ratio and 34% expense ratio) that without the large, high-margin premium inflow from the assumption, HCI would have earned ~$0.37/share, or ~20% of reported earnings.  These premium inflows are inherently high margin for a couple of reasons: (1) the annual premium rate is premised upon hurricane risk but by assuming the policies in November, HCI gets paid for this risk for seven months without actually shouldering the risk and does not have to reinsure against it; and (2) there is no policy acquisition cost associated with the Citizens assumptions so expense ratios are unnaturally low (low to mid-20s versus ~35% at typical insurers).  These effects combined with the calm weather have produced combined ratios between 46% and 60% over the past three Qs. 

This is obviously an attractive arbitrage HCI is executing.  The issue remains how sustainable is the model?  With no barriers to entry and a changing regulatory dynamic vis-à-vis Ctizens, I’d suggest the strategy and earnings are not defensible.  To understand the lack of sustainability, more background on Citizens and its current evolution is in order.

Current Citizens Take-Out Process

As summarized above, a take-out insurer (e.g., HCI) approaches Citizens and indicates it would like to assume 20,000 policies with certain characteristics – risk grade, location, property value, etc.  Citizens and HCI sort through various policies to arrive at 20,000 policies that meet HCI’s needs.

There is no real pricing of the policies per se as Citizens’ only goal is to get the policies off its books and HCI is not buying the policies but rather is assuming the future liabilities and future premiums going forward. The policies are simply transferred from Citizens to HCI.

This process has enabled HCI and other take-out insurers to assume the choicest / best-priced policies from Citizens. Citizens has historically held policies that were priced above market rates relative to the risks assumed, and take-out insurers can target these policies. Citizens can end up holding policies for which it was not the lowest bidder because the current system does not ensure a competitive bidding process.  Example: a captive agent can only write policies for a single or a few carriers; that agent’s client only receives policy quotes from one or two carriers and then Citizens; if the captive carrier is not within 15% of Citizens’ price, then the policy goes to Citizens even if there are five other Florida insurers that would have offered a lower price.  In this way, HCI is able to assume policies from Citizens that are priced at above market rates and Citizens is not truly functioning as the insurer of last resort.

Policies can also currently stay within Citizens because insureds have the ability to opt out of having their policies assumed by HCI and other take-out insurers. HCI has to notify an insured that it has chosen to assume his policy. The policy holder can then opt to stay with Citizens.

As a result, a large portion of the policies on Citizens books are priced above “market” rates – i.e., a more robust bidding process would have received quotes from private insurers below Citizens’ rates.  A recent analysis of 500,000 Citizens policies indicated 40% received lower quotes from private companies.  Generally, estimates I’ve seen suggest that 20% to 40% of Citizens’ policies are above market rate.  This obvioulsy creates an attractive dynamic for HCI and other take-out insurers who can assume above market rate policies without having to pay acquisition costs.  HCI has certainly generated impressive profits with its strategy.  However, as starting point, this strategy offers no competitive advantage – any glutton with capital can assume policies from Citizens.  High ROEs will attract capital, competition will increase and returns / the ability to attractively assume policies from Citizens will decrease.  This is already being seen in Florida: (a) Heritage Insurance was formed in Aug 2012 to assume policies from Citizens and was approved to assume 43,000 policies in Dec 2012 and another 20,000 policies in Jan 2013 and (b) Weston Insurance was formed in Dec 2012 and was approved to assume 6,000 policies in Feb 2013 and another 18,000 policies in Mar 2013.  Subsequently, Heritage has been involved in a controversial 60,000 policy assumption that is triggering political pushback given perceived lobbying influences that may have facilitated the deal.  Two points are made clear here: (1) HCI has no moat – these competing firms were formed, capitalized and then assumed policies within a matter of months; and (2) capital has seen the high recent returns generated by HCI and like carriers and is flooding to the opportunity.

HCI, however, does not simply suffer from a low-moat business model that is attracting competition – the entire system that has enabled take-out insurers to assume policies from Citizens is changing . . .

Legislative Reform of Citizens and the Clearinghouse

Given Citizens’ large market share and the fact the company has held a material number of policies for which it is not truly the insurer of last resort there have been mounting calls for the State to reform Citizens.  Governor Rick Scott has made Citizens reform a priority.  To this end, in June 2012, Gov. Scott appointed Barry Gilway to head Citizens.  Gilway is 66 years of age, has four decades of private market insurance experience, and has a mandate to shrink Citizens.  Importantly, Gilway has had a successful career and is heading toward retirement – not the typical dynamic of a government regulatory head who has a primary goal of further entrenching his fiefdom. 

Beyond appointing Gilway to head Citizens – and more importantly from the perspective of fundamentally altering Citizens’ role within the Florida homeowners market – three months ago Gov. Scott signed legislation intended to prevent policies from entering or staying within Citizens when there are private carriers who will insure the policies at comparable rates.  This goal is accomplished by the creation of a clearinghouse which will cover all homeowners policies within the State and allow all participating carriers to assess and bid on each and every policy, ensuring a more transparent and robust pricing process before policies are allowed to enter Citizens. 

All new policies in Florida will be entered into the clearinghouse for a period of 48 hours. This will allow all participating insurance carriers to assess and bid on the risk (versus the current dynamic where agents only receive pricing from carriers with whom they have a relationship). If a private bid is less than 15% above Citizens’ rate, the policy will go to a private insurance company. If there is a private bid cheaper than Citizens, it will be up to the agent as to which private insurance company receives the policy (agent will consider the price versus the financial stability of the company, etc.).  Likewise, renewing Citizens policies will enter the clearinghouse. If a private insurer offers a rate at or below Citizens, the policy will be transferred to the private market.

The legislation also reduces in stairstep fashion the maximum home value that can be insured by Citizens – $2m currently to $1m Jan 2014 to $900k 2015 to $800k 2016 to $700k 2017.

Through these mechanisms, Citizens should be left only holding high-risk policies for which there is no private market bid competitive with Citizens’ rates. This should ultimately end the process of private companies assuming advantageous policies from Citizens.

The clearinghouse should materially reduce the inflow of policies to Citizens and thus reduce HCI’s pipeline of policies that can be assumed.  This reduced supply of Citizens policies in conjunction with the increased demand to assume Citizens policies (i.e., Weston, Heritage and other recently-capitalized take-out insurers) should have a two-pronged negative effect on HCI.  Moreover, if Citizens is currently writing policies for which it did not offer the lowest price, the clearinghouse should remove this source of “premium priced” Citizens policies, leaving HCI to assume fewer and relatively less profitable policies.

Further, the clearinghouse should also reduce rates / pricing by increasing transparency, increasing the number of companies competing for policies and – importantly – funneling policies to the lowest bidder rather than to Citizens when it’s not the lowest bidder.

The conversations I’ve had with Florida agents and related parties confirm the view that the clearinghouse will increase pricing efficiency, remove premium-priced policies from Citizens and eliminate the ability of take-out insurers to assume well-priced risk from Citizens.

So what does the future look like for a company like HCI whose business is 100% premised on assuming policies from Citizens? The new system will likely necessitate a transition to bidding on policies from the clearinghouse. This should have several effects:

• HCI will now only receive premium through a transparent competitive bidding process. Pricing will be lower for each unit of risk assumed. Again, this is a commodity product, so underwriting profitability will suffer as risk-adjusted prices fall.
• The ability to arbitrage the take-out process by assuming policies in November – following the hurricane season – and not reinsuring as would otherwise be the case will end. This will substantially hurt profitability as margins should shrink.
• HCI will have to develop a network of agency relationships in order to assume and service policies via the clearinghouse. This should increase expenses – including policy acquisition costs – for HCI will have to morph into more of a traditional operating entity.

The clearinghouse is statutorily scheduled to begin operating January 2014.  Two weeks ago the state hired Bolt Solutions under a five-year $22m contract (extendable to ten-years and $45m) to provide the software platform necessary to operate the clearinghouse.  Gilway estimates that between the clearinghouse and incremental depopulations, there will be a 550,000 policy count reduction from Citizens.  550,000 policies represents ~45% of Citizens total policy count, supporting the view that Citizens holds many policies for which it is not the insurer of last resort.

Recently-Announced Citizens Depopulation

On the heels of hiring Bolt Solutions, Gilway did not wait very long to execute the incremental depopulations noted in the last paragraph.  On August 29, 2013, Citizens announced an enormous 400,000 policy depopulation spread amongst 10 carriers.  HCI was approved to assume 50,000 of these policies.  There are a few interesting takeaways from the recent announcement.

First, during 4Q12 Citizens ceded 203,000 policies to 7 insurers.  This year Citizens is ceding 400,000 policies to 10 insurers.  This supports the view that the assumption process has gotten more competitive with more insurers involved.  Notably, Heritage – one of the newly-formed take-out companies – is approved to assume 50,000 policies, confirming the complete commodity nature of this business model. 

Further, given that Citizens has 1.2 million policies, this 400k cession will reduce policy count by 33%.  This is a large reduction.  There's a real possibility this cession eliminated the vast majority of policies Citizens has on its books that private insurers are willing to assume.  If so, it makes sense that this will be the last take out prior to the clearinghouse taking effect in Jan 2014.  This view is also supported by Gilway’s above-cited estimate that 550,000 policies would be eliminated via depopulation and the clearinghouse.  With 400,000 of the 550,000 policy reduction target already approved, the remainder could certainly be eliminated via the clearinghouse as Citizens’ policies renew and enter the clearinghouse.

In sum, the size of this recently-announced depopulation certainly appears to be a large "one last bite at the apple" take-out with carriers looking to assume Citizens policies while they still can. With Citizens potentially now down to 800k policies and the clearinghouse coming online in January, this could have been a last hoorah take-out as Citizens should see reduced policy inflows and can jettison policies on its books as they renew and are forced to enter the clearinghouse.

Financials & Valuation, Part II

In the wake of the announced Citizens policy depopulation and with the clearinghouse scheduled to come online January 2014, HCI faces a very real question surrounding its business model and sustainability of earnings.  The company may very well have benefited from its last large-scale Citizens policy assumption, and the question now arises as to how the company will generate new policies and premium.  If the answer lies in the clearinghouse, profitability should be materially reduced as (a) price competition should be fierce (b) reinsurance costs should rise relative to gross premiums (no more late-fall policy arbitrages) and (c) operating expenses should rise (policy acquisition costs).

Current book value is $13.44/share.  If the November 2013 takeout proves as profitable as HCI’s 2011 and 2012 fall assumptions and the company is unable to execute incremental take-outs, I model YE2013 book value of $15.70/share and YE2014 book value of $20.60/share.  Keep in mind this assumes combined ratios across that time of 46% to 70%, which are conservative assumptions with no catastrophe losses and would be the envy of other insurance companies.  With no incremental policy generation, earnings would peak at 1Q14 with $1.85/share and decline to $0.80/share in 4Q14.  This decline is attributable to the policy churn (I’m using 20% churn which is below my estimate of historic churn) inherent in HCI’s business model.  To demonstrate, since HCI’s recently announced November 2013 Citizens assumption will not roll through the income statement until 4Q13, I model 3Q13 EPS of less than $1.00/share versus $1.42 in 2Q13 (some of the decrease accrues from increased reinsurance costs).  Of course, modeling combined ratios on a quarterly basis is inherently imprecise but the directional trend is clear. 

I’d also highlight I believe my modeling assumptions are generous as they assume a continuation of very low loss rates which have been historically low.  There will be increased loss rates – even absent hurricanes, current loss rates are understated.  The LTM loss rate is 32.0% versus HCI’s five-year average of ~48%.  Florida PNC life has been very placid over the past several years, but LTM was even more so.  This won’t last – it won’t even take a hurricane to materially hurt HCI earnings – just “normalization” relative to a period of time that had no major adverse events.

HCI group is materially overvalued.  The company faces a fundamental business issue of how it is going to continue to generate policies and premium amidst the Florida clearinghouse and increased competition for whatever depopulations (if any) occur in the future.  Should HCI morph into a more traditional insurer that generates policies via the clearinghouse, the company’s profit margins should erode as pricing will decrease (with the competition facilitated by the clearinghouse) and expenses will increase (policy acquisition costs and reinsurance).  The days of 40% ROEs are likely over.  As such, at 2.6x book value, HCI is too richly valued.  With an estimated $20.60/share of YE2014 book value – which assumes HCI’s current book of business generates historically rich profitability – and declining profitability from 1Q14 forward, current book value of $13.44/share or a slight premium thereto appear to be more realistic valuations given the existential issues facing the company.  This is especially true in light of the catastrophe risk facing a company with geographic concentration in the state of Florida.

Other Considerations

The above analysis is the primary, fundamental driver of my negative view on HCI.  However, there are a few collateral considerations. 

• Odd capital allocation – if the beauty of insurance companies is creating low- or no-cost investable float via breakeven underwriting, HCI turns this strategy on its head.  With 82% of its investable assets held in cash, HCI foregoes investment income and drives profitability with underwriting.  Other odd components of capital allocation include:
• Company acquired two Pinellas County marinas and related retail space for $13.7m (~8% of equity).  This is strange on its face for an insurance holding company of this size.  Beyond that, these properties were acquired in Apr 2011 and Apr 2012 and are still losing money in aggregate.  However, the CEO isn’t concerned as operating profitability wasn’t part of the analysis; it’s all about real estate speculation: ”Having said that, as you also recall, all of these properties are less about cash flow income as they are more about capital appreciation. So I mean, we carry them on our books at purchase cost, but the real value is somewhere down the road, you’re going to have a capital appreciation out of them somewhere, yes.” (3Q12 call).  I’d also add it is not likely good hedging practice for a Florida property insurer to acquire Florida coastal property for investment purposes. 
•  In Feb 2013, HCI paid $2m for a 16k sqft office building in Ocala, FL.  The office building previously housed a failed bank.  The office is vacant and but HCI intends to utilize as a redundant facility in case of catastrophes affecting the Tampa HQ.
• Company paid $7.1mm for its HQ – should not have been a priority at this point in the company’s life cycle – amounted to 15% of equity at the time.
• Related-party transactions with captive reinsurance company
• In 2012, HCI’s CEO and an SVP formed an entity (Moksha Re) to personally assume risk from HCI’s captive reinsurance company.  The CEO contributed $700k, the CEO’s family contributed $943k, the SVP contributed $200k with the remaining $8mm coming from “parties unrelated to the company.”  In 2012, Moksha assumed $13.8m of exposure for $4.0m of premium.  With no 2012 reinsurance losses the CEO and partners essentially extracted – via Moksha – $4.0m of potential profit from HCI shareholders last year.  If the pricing of this risk was so attractive, HCI’s captive reinsurance company could have retained the risk for the benefit of shareholders.  Perhaps instead of buying $13m worth of money-losing marinas, HCI could have used this capital to retain the Moksha exposure and earn $4m of profit?  On the heels of the 2012 transaction, HCI just announced a second related-party entity (Oxbridge Re) is set to assume 2013 premium from HCI’s captive reinsurer.  In 2013, Moksha and Oxbridge will garner an aggregate of $9.2m of premium in return for assumption of $25.5m of risk (company does not appear to disclose via SEC filing where in the reinsurance waterfall these risks reside).  In addition to the CEO and SVP (who recently resigned but is still on the board) from the 2012 transactions, two more HCI board members are participating in the related-party deal this year.  
• Insider selling between $20 to $25/share last fall and this spring

Catalyst