Description

Willis Group is the third largest insurance broker in the world with a 7% marketshare. Willis globally provides risk management and insurance brokerage services for middle market and larger Fortune 1000 clients. The company helps clients determine the best way to address risk management needs and acts as an intermediary between clients and insurance carriers. Willis generates revenues based on commissions and fees from insurance placements and consulting arrangements (70% of revenues in 2004 thru commissions). The company has significant marketshare in the US, UK and several other countries through its various subsidiaries. Willis, Aon Corp, and Marsh McLennan are the three global providers.

For more detail about the background and other aspects of the industry, Abrams705 wrote an excellent piece on the company in 2003. What has changed since then, or more appropriately, what has gone wrong is widely known. In April 2004, NY State Attorney General Eliot Spitzer turned his attention to the insurance brokerage industry and noticed what now seems obvious; when brokers are paid by clients to obtain the optimal insurance coverage and are given kick backs by the selected insurance carrier, there is a huge conflict of interest. In many cases, the commissions paid by insurance carriers (“contingent commissions”) weren’t gratuitous, but the line clearly was crossed in many instances.

The primary culprit was Marsh McLennan, which generated 12% of its insurance brokerage revenues through contingent commissions. The company went so far as to set up phony bids from insurance carriers and fix prices for coverage. Eventually, Marsh settled with Spitzer, agreeing to pay an $845 million fine. While Willis only generated 3% of its revenues through contingent commissions in 2004 and settled for a $51 million fine, the company has suffered in the aftermath due to legitimate issues adjusting their business model and problems involving perception.

Willis announced in October 2004 that they would stop use of contingent commissions and would provide full transparency to clients involving the commissions they are generating. Other insurance brokers followed suit and it has been a somewhat painful process for the industry as clients reassessed their risk management / insurance needs. Additionally, when Willis announced the decision to terminate use of contingent commissions, the company differentiated between problematic contingent commissions and market service agreements (“MSAs”). Market service agreements involve legitimate services to insurance companies that don’t involve volume or profit targets for coverage. Examples of MSAs are claims processing services, risk modeling, benchmarking, and market research. While these services are justifiable, insurance carriers have shrewdly balked at paying these fees in light of the issues with contingent commissions.

Combine these issues with a declining price environment / soft market for insurance rates, and it appears on the surface that Willis could be in real trouble for the first time since going public. First quarter 2005 revenue was basically flat from a year ago and adjusted operating margins were down nearly 500 bps. The company thus trades at $32.31, near its 52 week low (range of $30.40 - $42.53).

I believe that concerns are overblown. While the near-term earnings power of the company may be impaired, Willis’ long-term viability is intact and the growth potential, in spite of falling premium rates, is significant. Even with very modest growth assumptions, Willis is trading at an extremely compelling valuation.


VALUATION

In an extremely conservative situation, Willis is currently trading at 16.1x 2005 earnings and 13.6x 2006 earnings.

Scenario A

$ in millions
2003A 2004A 2005E 2006E 2007E
Revenue 2,076 2,275 2,174 2,174 2,283
EBIT 620 666 512 576 642
Net Income 414 427 332 366 410
Diluted EPS 2.45 2.54 2.01 2.38 2.73
FCF/Share 2.34 2.53 2.00 2.36 2.68
PE Ratio -- 12.7x 16.1x 13.6x 11.8x
FCF Yield -- 7.8% 6.2% 7.3% 8.3%
ROE 31.3% 30.0% 23.0% 30.3% 31.0%
*Excludes restructuring and Spitzer settlement fee
**Incremental free cash flow used for share repurchases

These estimates represent the very low end of profits Willis is capable of generating. I’ve assumed that Willis’ core business experiences 0% revenue growth for the remainder of 2005 and for all of 2006, after stripping out contingent commissions and potential market service agreements earned in 2004.

The only factor driving earnings growth in 2005 and 2006 is the assumption that margins will gradually drift closer to levels achieved in 2003 and 2004. In 2003 and 2004, the company achieved general and administrative margins of 71.3%; I’ve assumed 75% and 73% margins for 2005 and 2006, respectively (which translates into mid 20% operating margins). I’ll discuss the rationale as to why this is reasonable later in the writeup.

In this base case scenario, Willis still generates a free cash flow yield (net income plus depreciation minus capex) of 6.2% and 7.3% in 2005 and 2006, respectively. Insurance brokerage is inherently a low capital expenditures business and the company is able to throw off tons of cash even in a difficult environment. Combine this with a 23% return on equity in 2005 and Willis achieves an attractive return on investment even in a very conservative scenario.

That being said, in this scenario, Willis certainly isn’t dirt cheap and isn’t at a meaningfully better valuation than levels at which it has historically traded. The company’s average forward P/E since going public has been roughly 15x earnings. So why is this so compelling?

The real value arrives if one starts to evaluate reasonable upside scenarios as the story post the Spitzer saga unfolds. If one considers Willis’ true earnings potential if the company is able to achieve some of (certainly not all) the goals it accomplished pre-Spitzer, the situation becomes far more enticing.

Below are the estimates if Willis can achieve a modest 5% top line growth in its core business, revert G&A margins to 2003 and 2004-like levels, and generate half the level of MSAs achieved in 2003 and 2004.

Scenario B

$ in millions
2003A 2004A 2005E 2006E 2007E
Revenue 2,076 2,275 2,295 2,450 2,572
EBIT 620 666 541 695 730
Net Income 414 427 352 446 470
Diluted EPS 2.45 2.54 2.13 2.90 3.13
FCF/Share 2.34 2.53 2.12 2.88 3.08
PE Ratio -- 12.7x 15.2x 11.2x 10.3x
FCF Yield -- 7.8% 6.6% 8.9% 9.5%
ROE 31.3% 30.0% 24.1% 34.1% 31.7%
*Excludes restructuring and Spitzer settlement fee
**Incremental free cash flow used for share repurchases

The ability to buy Willis at 11.2x 2006 earnings when only assuming 5% core revenue growth (with some pickup in MSAs) in my view is extremely compelling. Traditionally, Willis’ valuation relied on significant revenue growth; at current levels, it really only hinges on the company’s ability to grow modestly and thoughtfully re-adjust its cost structure, as it has in the past.

Based on these assumptions, I believe that a mid-$40s price for Willis is extremely reasonable. I’ll add that the company’s ability to generate higher than mid single digits growth is very plausible and could certainly drive significantly higher earnings than I’ve forecasted in 2006 and beyond.

From A TO B

So how do we get from the fairly conservative (well below street estimates) Scenario A to the slightly more aggressive Scenario B? The main drivers are the top line growth, margins, and restoration of MSAs.

Top Line Growth

The main engines for Willis’ top line revenue growth are hiring, insurance rates, and acquisitions.

Willis is currently expanding its revenue generating work force by approximately 5-6% annually. More insurance brokers and risk management consultants generally translate into more revenues; within the insurance brokerage industry, hiring more brokers is roughly analogous with a retailer opening more stores (with the insurance carrier itself functioning as the manufacturer of goods). Like increasing number of stores / square footage, hiring more brokers generally increases coverage/revenues. However, in sticking with this parlance, what really matters is the same store sales equivalent, which is increasing revenue per revenue generating employee. Willis doesn’t break this statistic out cleanly, but from comments made by Plumeri, one can roughly back into the figures; revenue per broker was $740,000, $830,000, and $870,000 in 2002, 2003, and 2004, respectively. In other words, Willis’ revenue generating worker productivity increased by 12.5% and 4.4% in 2003 and 2004.

To provide an idea of how this fits with my base case conservative estimates, based on the company’s plan to increase revenue generating employees by 5%, Willis’ productivity per worker would have to fall by 9.1% and 4.8% in 2005 and 2006. While this certainly is possible considering the presence of falling insurance rates and the troubles Willis is experiencing, it provides an idea of how drastically trends would have to change for the company to only be able to achieve those targets. Moreover, it is important to note that in the first quarter of 2005, revenue per worker fell by only 4% year over year in what probably will prove to be the most tumultuous quarter considering the changes in business structure that occurred (first complete quarter of full transparency and contingent commissions’ elimination).

For Scenario B to occur, it would only require that productivity per employee fall by 4.0% in 2005 and rise by 1.7% in 2006. While the 2006 estimate assumes some improvement in productivity from 2005 levels, I’m comfortable with the 2006 estimate because on an absolute level it still is below historic levels ($840,000 versus $870,000 in 2004) and because of the types of employees Willis has been adding this year. Willis has absorbed a number of relatively senior defects from Marsh McLennan the past two quarters. While I can’t personally attest to their quality, the level of their seniority provides some indication that their 2005 hires could be of a unique caliber. As we wait to determine the quality, it is important to note that the impact of hires from Marsh likely won’t be felt until late 2005 / early 2006 as there obviously is a lag between hiring and eventual revenue generation / RFP victory.

Insurance rates are an entirely different animal. I certainly won’t pretend that I know where premium levels are heading. On the surface, because Willis’ commissions are based on a percent of insurance premiums, it might appear that there would be a one-to-one correlation between rates and insurance brokerage revenue. This isn’t necessarily the case because of the tendency for clients to purchase more insurance coverage when rates are weak and vice versa. Insurance brokers enjoyed a significant tailwind from insurance rates post 9/11, but it is unclear how much additional insurance clients might have purchased had the market been softer (i.e. clients’ elasticity is the driving factor).

In 2004, insurance rates fell significantly over the course of the year but Willis held up fairly well. According to reports provided by the Council of Insurance Agents & Brokers, the average property/casualty premium for medium and large accounts fell by 4% and 7%, respectively. Willis’ organic revenue, however, only fell by 2% as a result of insurance rates. Additionally, because 6% more insurance was purchased by Willis’ clients, net organic growth actually increased by 4% year over year (6% - 2% = 4%). The first quarter of 2005 provides further evidence of Willis’ resiliency; insurance rates fell by 11% for medium and large accounts while Willis’ organic revenue growth attributable to falling insurance rates was only negative 2%.

These results are over a relatively short time frame so obviously are inconclusive at best. Perhaps a more effective exercise is to examine insurance broker net organic revenue growth during the 1990s, which was a fairly weak environment for insurance rates. The best way to conduct this analysis is by evaluating Marsh. From 1993-2000, Marsh’s organic revenue (excluding acquisitions and foreign exchange) on average increased by 4.6% annually. Organic revenue growth wasn’t a result of excessive costs, as operating costs only increased by 3.3% on average from 1993-2000. Although Brown & Brown is a far smaller regional player, the company is a top notch provider so I’ll note that their average organic revenue growth was similar, increasing by 3.9% annually over the same time period.

My hope isn’t to establish that Willis is impervious to falling insurance rates, but rather to realistically suggest that the notion that Willis won’t be able to grow its top line in a weak rate environment is fallacious. They did it in 2004 and Marsh and Brown & Brown have done it for years. Insurance rates certainly can function as more of a headwind in the upcoming future, but in no way does that imply that the company can’t grow its revenues by mid single digits. Additionally, the fact that 30% of revenues are fee based and not generated through commissions further insulates Willis during a soft market environment.

It is difficult for me to comment on how acquisitions will affect the top line growth. Plumeri seems to handle acquisitions and their integrations very prudently from a financial and operational perspective. He understands that what he is truly paying for is manpower; if he can hire brokers without having to pay a multiple of revenues, that is far better than acquiring an entire company. Willis has recently made some interesting moves acquiring subsidiaries internationally – notably in China – and I would expect that to continue going forward. International revenue has increased by 15% and 10% over the past two years.

Margins

Over the past few years, Willis has enjoyed best in class margins among the global insurance brokers. Operating margins were 29% in 2003 and 2004. This is largely a function of Joe Plumeri’s ability to squeeze costs out of every segment of the business, a fanaticism he undoubtedly developed through his years with Sandy Weill. While margins in the first quarter of 2005 declined by approximately 5% on an adjusted basis (excludes special charges), it seems reasonable to believe that Willis can gradually revert to its historical levels over time as Scenario B dictates. A big contributor will be cost cuts; Plumeri has indicated that they plan on cutting approximately $50 million of annual costs effective Q4 2005. Based on his past ability to fastidiously drive through cost cutting measures, this assumption seems reasonable.

MSAs

Willis’ ability to partially regain lost market service agreements revenue should also take some pressure of margins and provide an incremental kick to the top line. There is no guarantee that they will be able to regain lost revenue ($63 mm and $77 mm of total revenue in 2003 and 2004) but I find it difficult to believe that this revenue stream will disappear in entirety. Insurance companies obviously are thrilled with the current arrangement in which they basically receive services for free, but as clients become more comfortable with the full transparency provided by insurance brokers, it seems reasonable that the notion of brokers collecting revenue from insurance carriers for services unrelated to volume or profits will be accepted. In scenario B, I’ve assumed only a partial recapture of MSAs and don’t believe that is particularly aggressive.


COMPETITION

Marsh

Marsh McLennan is the largest player in the industry and controls roughly 25% of the market. Marsh historically has been the strongest player in the industry but now faces significant issues as the company had the largest exposure to contingent commissions (12% of revenues). The fact that some of the bid rigging tactics taken by the company’s brokers became public certainly doesn’t help either. The company faces significant issues regaining trust from clients and it is unclear how badly its franchise has been impaired. Marsh is also losing employees to Willis and Aon in flocks.

Another issue Marsh faces is the necessity to re-examine profitability. Previously, contingent commissions allowed certain smaller accounts to be sufficiently profitable. Marsh was able to accept cheaper commission rates in exchange for contingent commissions down the road as eventual form of compensation. Without contingent commissions, Marsh will either have to stop writing some business (which management has indicated they will do) or raise their commissions if they hope to make comparable returns on capital. Contingent commissions made scale an even bigger advantage.

I won’t delve into Marsh’s valuation because I believe that there is too much uncertainty surrounding the franchise’s impairment, the value of departing employees, and Cherkasky’s ability to restore order. Assessing what percent of assets the problematic Putnam should trade for also seems to involve sufficient guesswork. There is significant potential upside but the risk reward profile is of an entirely different nature and not one upon which I am willing to speculate.

Aon

Aon is the other major global player in the industry with 23% marketshare. The company had far less exposure to contingent commissions than Marsh (approximately 4% of revenues) and has also been poaching brokers. However, believing in Aon is believing in the turnaround story. The opportunity is that margins are significantly below Willis’ (and Marsh’s historically). If management could ever get their act together, Aon could be a great investment. Their reinsurance brokerage business is a dominant player and the company seems more focused than it ever has before. However, in my view, management has to prove that they can achieve consistency. If they operated their business with inferior margins and return on capital before the insurance brokerage industry faced meaningful issues, who is to say that management can do so when there is a strong regulatory eye and full transparency required for clients?

The company currently trades for an attractive 12.2x 2005 earnings and 11.5x 2006 earnings, but I consider Willis to be the superior investment. Not only is the valuation more or less the same, but Willis is a smaller, more nimble company with stronger growth prospects, superior management, and a cleaner balance sheet (I’ve never been a big fan of under-funded pensions…)

Regionals

Like Willis, the regional brokers (Brown & Brown, AJ Gallagher, and Hilb Rogal & Hobbs to name a few) have a significant opportunity to steal marketshare as clients look to diversify their insurance brokerage providers. Plumeri indicated recently that Willis has seen 3x the number of RFPs (request for proposals) and that, in 70% of the situations, the client changed to a new broker for at least one line of business. The elimination of contingent commissions also helps smaller players’ pricing become relatively more competitive.

What the regional brokers lack, however, is the ability to serve clients on a global scale. Clients looking for an insurance broker with extensive reach (that is not Marsh or Aon) will in many cases find the smaller players insufficient. Willis has the ability to not only steal marketshare, but also actually serve a major client’s needs. That being said, the regional players are likely attractive investments. They trade at comparable valuations (with the exception of the highly regarded Brown & Brown) and have solid growth prospects.

In this context, it is important to note the incremental ability Willis has to steal marketshare. The entire industry is approximately $30 billion; every 0.5% of marketshare represents $150 million of revenue. For Willis, which currently generates $2.2 billion, the opportunity to grow by stealing marketshare is huge.

MANAGEMENT

I’ve met Joe Plumeri on a couple occasions and found him very impressive. While I typically have ample disdain for the “go-go,” cheer the company on type of CEO, Plumeri is unique. He is an exceptional motivator and has reinvigorated Willis. He was handpicked by Henry Kravis and spent years learning from Sandy Weill. If you listen to Plumeri discuss business operations, it is clear that his ultimate goal is to increase shareholder value, either through high returns on reinvested capital or share repurchases (previously debt repurchases). Additionally, like Weill, he has an exceptional knack for cutting costs. Willis enjoys some of the best margins in the industry and it is clear that Plumeri is in large part responsible for this success. Tom Colraine also seems like a thoughtful, focused business leader.

My only concern about Willis’ management is a meaningful one; Joe Plumeri is selling a large percent of his shares. On May 4, 2005, he initiated a 10b5-1 trading plan to sell 2.4 million shares. While he will be left with approximately 2.4 million shares even after this trading plan is fulfilled, it certainly is concerning. Plumeri pumps the stock in meetings, encourages employees to buy, and then sell his own shares? Plumeri also indicated in the past that he would never sell shares while he was CEO. Willis hasn’t provided significant disclosure into his rationale; in the 8-K, Willis indicated that the Board of Directors is encouraging senior officials to periodically sell shares. This isn’t entirely unreasonable but doesn’t fully satisfy me. It is possible that, with their $300 million share repurchase underway (and more potentially coming with the recent issuance of debt), the Board encouraged Joe to execute the plan beforehand to not be caught by 10b5-1 / 10b-18 restrictions and the optics of boosting the stock through a buyback only for management to then sell shares.

Ultimately, what provides comfort is the fact that Plumeri will hold approximately $80 million of stock even after the plan is executed. That should function as sufficient incentive. Regarding the issue of selling when the stock is cheap, on a more general philosophical level, I’ll add that if I were in his position, I would probably trim the position as well. Joe received the shares to turn the business around, and he has accomplished that feat remarkably well. It only makes sense at age 61 to have some diversification.


CONCLUSION

Willis is a strong company in an industry undergoing significant change. Even if there is virtually no growth over the next couple years and the company isn’t able to regain profitability through MSAs, a mid-7% free cash flow yield isn’t too shabby. In the very plausible/realistic scenario in which the company achieves single digits revenue growth, Willis’ valuation is extremely compelling and warrants a stock price at least in the mid-$40s.

Catalyst

-Ability to gain incremental revenue through MSAs
-Steal marketshare / quality producers with Marsh in flux
-Industrywide issues settle down