Description

Background
If you like cheap financials trading at a deep discount to book value and have a two-year investment horizon, then you will love PMA Capital (“PMA”, Ticker: PMACA). Based on its current price of approximately $9.75, I believe there is 60%+ upside from current levels. Also, with the stock is trading at ~75% of stated GAAP book value (ex FAS115 gains/losses) you have a very sizable margin of safety at current levels.

Similar in story to the successful turnaround at Allmerica (now called “Hanover Insurance Group”), PMA is a sum-of-the-parts story with two good businesses (a well-run and increasingly profitable Workers Compensation insurer and a fast-growing and high ROIC Third Party Administration business) and a bad business (a reinsurance business now in runoff). PMA almost went under in 2003 because of massive reserve problems at the reinsurance business. Specifically, the reinsurance business blew up because of general liability, E&S lines (namely professional liability coverage on nursing homes and builders’ defects coverage for construction claims) and commercial auto business written from 1997 to 2000. Unfortunately for the Workers Comp business, it was structurally subordinated to the reinsurer (they were “stacked” with the workers comp business under the reinsurance business), so when the reinsurer went into runoff, the ratings agencies rightfully became concerned about the stability of the capital at the Workers Comp business. A.M. Best slashed not only the ratings on the reinsurance business, but also on the Workers Comp business to below A- which is the kiss of death in this line of insurance. As a result of the downgrade, LTM net written premium volumes declined significantly from $604mm in 2003 to $348mm in mid 2005. Without this volume, the combined ratio spiked to ~108% due to the negative operating leverage.

Following the adverse development at the reinsurance company and the resulting stock price collapse, the Board promoted Vincent Donnelly to the CEO role. Vincent is an actuary by training and was already the head of the Workers Comp business. Immediately he began a plan to (1) stabilize the Runoff Business and salvage its surplus and (2) extract the Workers Comp business from underneath the Runoff Business. Working with the Pennsylvania Department of Insurance, PMA was able to “de-stack” the Workers Comp business from the Runoff Business in exchange for agreeing to a prohibition from paying dividends to the holding company until such time that the regulators became comfortable that the business was stabilized. During this time, Vincent and his team did a fantastic job of communicating with the brokers and key employees about its plans for the turnaround. Based on my channel checks, I learned that PMA didn’t lose any key employees (most importantly, they retained the client-facing claims adjusters) during this entire period when they didn’t have their ratings. Their open and honest communication with the brokers set the stage for significant premium growth once ratings were restored (which we will get to later).

In order to wind-down the runoff business, Vincent and Bill Hitselberger (the CFO) aggressively approached clients of the runoff business about commuting their business with PMA. In a commutation transaction, the insured agrees to settle their claims in advance against PMA in exchange for an upfront payment. PMA’s argument is fairly compelling in the negotiation. They argue that the insured customers are similar to impaired creditors and these customers have a choice to either (1) accept an upfront payment for an amount less than the estimated future claim (which creates a gain for PMA) or (2) roll the dice and hope that the Runoff Business has the claims-paying resources to pay the bill when the time comes. Considering the company has performed over $400 million of commutations (approximately ~$250mm excluding contractual commutations), it is clear that PMA has been highly successful in running down the reserves on favorable terms. While it’s impossible to calculate based on the information, I suspect there is hidden value based on these commutation gains. Specifically, management told me that, on average, they were able to negotiate a discount rate of ~8% on claims that had a ~4 yr weighted average life remaining, which implies the company was settling claims at ~73% of par. On $250mm of negotiated commutations, this equates to ~$66mm of implied commutation gains. The company has been doing the right thing and adding 100% of these “gains” to IBNR reserves over the past three years in order to build a margin of safety in the reserves. While it’s impossible to know exactly the movement between IBNR to case base reserves including the impact of these commutation benefits, I think it’s reasonable to assume that a material portion of the ~$40mm of remaining IBNR reserves today relate to the commutation gains achieved by the company. As long as the case based and IBNR reserves don’t develop adversely, there should be significant redundant reserves that at some point could be released back into earnings.

Earlier this year, PMA petitioned regulators for an extraordinary dividend from the runoff business. The Company announced it had received approval on May 3rd for a $73.5mm special dividend from the Runoff Business to the Holding Company. The capital will be used primarily to pay off holding company indebtedness (notably the converts currently trading at ~113% of par). Management alluded to future special dividends due to the excess capital in the business (Risk Based Capital ratios is still ~350% pro forma for the dividend). Future dividends will likely be used to buyback stock given the reasonable amount of pro forma debt to capital on PMA after the recent dividend.

In late 2004, A.M. Best upgraded the workers compensation business to A- and business at PMA has been growing ever since. The brokers I spoke with said they were waiting for ratings to be restored and, once they were, they felt highly-comfortable putting their customers back with PMA. Vincent’s constant dialogue with the broker force and his ability to retain the employees were key factors. According to brokers, PMA is regarded as having (1) superior service levels and (2) good upfront loss mitigation practices which save customers money in the long-run. PMA is able to charge a premium for its services compared to other players (I’ve heard between 5 – 15% depending on the program) as customers recognize the long-term cost savings from having PMA on board. Following the upgrade, premiums YoY have been growing 15-20%. As Q1 ’06 LTM net written premiums were $384mm versus $604mm in 2003, it is clear that PMA still has significant growth left in the Workers Comp business.

Valuation
I believe the proper way to value PMA is on a sum-of-the-parts basis. First, we have to disaggregate GAAP book value (ex FAS115 gains / losses) between the Workers Comp and Runoff businesses (pay attention here because this is critical to the story). According to the Company, there is ~$378mm of equity in the Workers Comp business, $221mm in the Runoff Business and ($185mm) of negative equity at the Holdco (essentially double leverage). Let’s use these numbers as the starting point for the valuation.

We now have to adjust the equity in the Runoff Business for the special dividend up to the holding company of $73.5mm. This leaves us with $148mm of estimated book value at the runoff business ($221mm less $73.5mm). This special dividend is an intra-company reallocation of capital, not a dividend back to shareholders. Therefore, GAAP equity will not decline as a result of this dividend. What will happen is the equity in the Runoff Business and the negative equity at the Holdco will each offset each other by $73.5mm. After the debt is reduced, we are left with $378mm at Workers Comp (unchanged), $148mm at the Runoff Business and $112mm of debt (the negative equity at the Holdco). Let’s assume for the sake of conservatism that 100% of the debt is allocated to the Workers Comp business and the Runoff Business is unencumbered. This implies adjusted GAAP book value at Workers Comp is equal to $266mm ($378mm of “Investment” in this subsidiary less $112mm of pro forma debt).

Valuing the Runoff Business first, I believe that 60-65% of book is a fair price for this pool of capital. Assuming a 4 year weighted average life remaining on the business, this translates into a ~12.5% IRR. Considering the likelihood of hidden value at this business due to the commutation gains currently hidden in IBNR reserves, I think that’s a reasonable valuation for this business. This yields a valuation of ~$92mm.

For the Workers’ Comp business, I think ~90% of book is a reasonable valuation today considering the growth prospects and its competitive positioning in the market, offset by the current low ROE (4-5%). 90% of book value yields a valuation of $239mm for the Workers Comp business (90% * $266mm).

Finally, PMA has a fast-growing Third Party Administrator business which is increasingly valuable. In the TPA business, PMA administers the workers comp programs for clients and handles the administrative aspects of the claims management process. Revenues have grown in this business recently at 20-25% and LTM operating profit is approximately ~$6mm (estimated operating margins of ~20-23%). My channel checks have come back very positive about the TPA management and PMA believes there’s significant upside in this business. Also, the TPA business requires no capital as it’s a processing-like business, so it’s a highly free-cash flow generative business. I think a 6x multiple on LTM operating profit is a conservative multiple. This yields an additional $36mm of value.

So summing this up, I estimate the total business is worth $368mm today ($92mm for Runoff, $239mm for Workers Comp, and $36mm for the TPA business), which equates to ~$11.40 per share or ~17% higher than current levels.

As a sanity check, let’s see where we “create” the Runoff Business at the current price of $9.75/share holding the valuation for the Workers Comp and TPA businesses constant. The current market value is ~$314mm, so this implies a value of $39mm for the Runoff Business ($314mm total less $239mm for Workers Comp and $36mm for TPA), or roughly ~25% of pro forma book value in this segment. I think paying 25% of book for what is effectively a pool of cash is a good risk to take, especially considering that the regulators have just scrubbed the books and allowed PMA to extract 1/3 of the statutory surplus from this business.

Looking out two years, I think this business will be worth ~$15.50/ share. How do I get there? Well, I think PMA will continue to run-down reserves and get favorable commutations in the Runoff Business, so the value of this business should slowly accrete towards book value. Assuming the same ~12.5% IRR rate, I think this business will be worth ~$117mm in two years (roughly ~80% of adjusted book), which seems reasonable. After management extracts the remaining low-hanging fruit in the business via commutations (which is why management has been reluctant to sell the business as they think there’s more upside remaining), I think they will monetize the capital in runoff via a sale, similar to Hanover’s sale to Goldman Sachs. With respect to the Workers’ Comp business, I believe that (1) management will grow this business to ~$500mm in revenues (still far below 2003 levels) and (2) the combined ratio will drop to 100% due to increased operating leverage. I see no structural reason why this business can’t earn a solid 8-12% ROE over the cycle, especially considering the quality of management in this business. Assuming a reasonable ramp in the ROE towards these levels, I value the Workers Comp business at ~$332mm looking out two years, which assumes 100% of forward 2008 book value. Realistically this business would trade north of 100% of book if the ROE does ramp like I think it will, so that’s additional upside. Assuming the TPA business continues to grow at 15% on the top line and maintains margins, it should generate $8mm in operating profit in 2007. Assuming no multiple expansion, this business will be worth ~$51mm. So summing it all up, I believe this business will be worth ~$500mm looking out two years ($117mm in Runoff, $332mm in Workers Comp, and $51mm in TPA) or ~$15.50 per share. This equals ~60% upside from current levels.


Risks
So how do you get crushed on this investment? The key risk is future adverse development in either of the Runoff Business or Workers Comp business. So much of this story is predicated on management’s ability to extract capital from the Runoff Business and improve the ROE in the Workers Comp business. Any hiccup in either business would be bad for the turnaround story.

Regarding the adequacy of the reserves, I think it’s important that we step back for a minute and remember that in order to simply stand in front of the state regulators this year and make the request for a special dividend, the regulators required (1) the company get a third party review by an outside actuary of the reserves and (2) that PMA put up another good year of operating performance in both businesses (ongoing ops + runoff ops). After reviewing all the data, the regulators (who, remember, have absolutely zero incentive to allow a runoff entity to upstream capital given their mandate to protect policyholders), allowed the company to extract not a token dividend, but approximately one third of the statutory surplus in that entity. While I’m no actuary, it seems to me that the guys who (1) know the stat and probability tables and (2) have their jobs on the line felt pretty comfortable that these reserves are adequate.

Additionally, the company has ~$75mm of an adverse loss development cover remaining with Swiss Re that provides significant protection against an additional ~$105mm of development. If you look at the 2005 statutory filing for the runoff business, you’ll see there was only twenty six dollars of development in 2005. Furthermore, the company still has ~$40mm of IBNR reserves remaining in addition to the Swiss Re cover to protect against future claims that still might develop. We are now 6-9 years past the period of “bad” business (’97-’00) and we are getting towards the back end of the curve for the majority of the exposures. While adverse development can always flare up, the fact that the book has shrunk so much and that a significant amount of time has passed makes me comfortable that management and its actuaries have a pretty good handle on future loss payments. Also, with interest rates being higher the PV of the long-tail loss payouts becomes less and less and less, especially since the main reserve builds were in 2002 and 2003 and rates are up materially since then.

Finally, Workers comp insurance is subject to market swings in pricing just like other lines of insurance. Notably, Pennsylvania just reduced the estimated filed loss cost rate used in pricing business by 8.6%. PMA uses this loss cost rate as the floor pricing and then adds to the pricing depending upon the individual insured company’s loss history / projections. While the company expects to have flat-to-slightly up pricing overall in 2006, the trends in some of PMA’s states are flat to down. Offsetting this base rate decline is this fact that a substantial portion of PMA’s business in Pennsylvania has loss sharing agreements with customers, so they share in the upside and downside. Also, PMA has so much business still to recover that it should have plenty of years remaining of solid, profitable overall growth.

Catalyst

Near-term catalysts include (1) management finally taking the offensive and marketing the stock this month (June), (2) new Sell-Side research coverage (KeyBanc just initiated) and (3) continued positive earnings momentum in the Workers Comp business. Longer-term catalysts include (1) additional special dividends from the runoff business, (2) share buybacks (the fist special dividend had to be used to repay debt....future dividends can be used for buybacks), (3) a monetization of the runoff business and (4) a sale of the entire company in a few years.