Description

Introduction

Are you interested in buying a small company with almost no analyst coverage, no quarterly conference calls, and about to benefit from both secular and cyclical trends? And to top it all off, we believe the company will realize the benefit of a significant one-time earnings event (that few people know about) during the fourth quarter, The event will become common knowledge when the company reports February 18^th.

We hope you are intrigued. So without further ado, we introduce to you Stewart Information Services (STC). I had hoped to post this earlier and unfortunately the stock is running hard ahead of the quarter, but there is a compelling long term story here as well.

First, title insurance is not an area people are commonly familiar with. So let's start off with a little industry tutorial:

Industry Dynamics

Title insurance protects lenders and buyers against defects in real estate titles. For example, an unpaid electrician may have filed a lien against a house. Or an unknown heir may show up years later and try to claim his inheritance. Or maybe someone can commit ID fraud and sell a property that doesn't belong to them. The situations are varied, but if there are any title issues, the title insurer has to pay. Given that a house is often the most expensive item most families ever buy, it's worthwhile to pay a small sum to protect against risks that might cost hundreds of thousands of dollars.

However, in practice, claims are very rare. Over a long cycle, claims run about 7% of total premiums. For a typical property and casualty insurer (example: auto insurance), claims can run well over 70% of premiums and is by far the largest part of the cost structure. How can this be? Is the title insurance company just ripping people off and making boatloads of money?

Well, the answer is no. Most P&C insurers protect against future events: an auto insurer has no idea if you will crash your car next year. All GEICO can do is to take some basic information from you, pop it into a computer spreadsheet, and out comes an insurance policy. This is a very simple and cheap procedure.

In contrast, title insurers mostly protect against past events that its customers may not be aware of: did the county levy a tax lien against the house? Did the owner fail to pay the carpenter? Is the property mired in a divorce lawsuit? It takes a lot of work to check into and resolve the looming issues. The goal is to catch any existing problems and contain them before the real estate transaction takes place and degenerates into a messy lawsuit. So, instead of filling out a 5-minute form to generate a title policy, an underwriter has to access court houses, tax records, land records, and so on. Sometimes, a lawyer is needed. This is very costly--in fact, it's the biggest part of a title insurer's cost structure.

Over a long cycle, the title industry is profitable, but not as profitable as its claim numbers suggest. Industry underwriting margins fluctuate between 5% and 10% over the long run, which is not far from a normal P&C insurer.

That said, profits have been very weak since 2007 as revenues are tied to both transaction volumes and house prices. Even worse, insurance companies historically received sizable revenues from commercial real estate transactions (20% to 25% of revenues in good years), which tend to be more profitable than residential policies. The commercial market is currently comatose. As a result of these factors, industry revenues fell almost 50% from 2006 to 2009.

Title insurance companies tend to have highly fixed cost structures, especially in the short run. Consequently, operating profits fell even faster revenues.. They employ legions of title examiners, lawyers, people to maintain in-house databases (title plants), closing agents, and so on. Many of these employees are highly trained and specialized, and companies are loath to cut them in a downturn because it's hard to rehire and retrain them when the market improves. Therefore, the industry was relatively slow to cut costs, hurting operating margins.

Unfortunately, this isn't the end of the bad news. As the downturn worsened, another problem reared its ugly head. National title insurers (which are large and publicly traded), usually conduct about half of their business through independent agents. These agents source clients, do the underwriting work (checking court records, etc), then submit policies to the title insurer. During the boom, business was so plentiful that underwriting standards loosened substantially, causing the big title insurers to significantly under-estimate their ultimate losses. Essentially, agents and underwriters were sloppy in their work and failed to uncover many potential issues that are now cropping up. Title insurers, in response, have been furiously strengthening their reserves to cover these losses.

Beyond the underwriting issues, the independent agents have been engaging in fraud, through a process known as defalcation. Due to their  role in the real estate closing process, independent agents have access to escrow accounts. When times got bad, some agents dipped into this pool (illegally) to fund their own operations--such as to meet payroll. When the tide went out, all sorts of criminal activity was exposed. This behavior wasn't unique to the recent boom. It happens every cycle and (fraud) can add 1 percentage point or more to an insurer's long run loss experience.

So to sum up, the industry had experienced free-falling revenues, negative operating leverage, and increased fraud. Operating margins are just stabilizing at a fraction of their previous levels, and the real estate market remains sick. So, why are we playing in this sandbox?

A Transformational Deal

One big reason for our optimism is a seminal event that took place in late 2008. Facing bankruptcy, the third largest title insurer LandAmerica, agreed to be bought by the second largest title insurer Fidelity National (FNF), for the princely sum of $238 million. To put this number in context: before the bubble burst, LandAmerica had a book value of about $1.4 billion, took in $4 billion of revenues, and generated $170 million in net income per year. How quickly things change!

The LandAmerica acquisition was vital because it created the largest title insurance company in the country, with a pro-forma market share of 46% (though we expect it to fall a bit). The industry is now dominated by just two big companies, Fidelity and First American (FAF), who have a combined market share in the mid-70%'s. They are followed by Stewart Information Services (STC) and Old Republic (ORI), with market shares in the low teens and high single digits, respectively (though that's changing, as we will explain later). A host of tiny regional companies fight over the remaining scraps.

LandAmerica was not a wonderfully run company, but it wasn't broken either. There's a good chance that it would still be around had it not made one crucial mistake. As a part of its services, LandAmerica handled escrow funds in commercial real estate (1031 exchange) transactions. Foolishly, LandAmerica decided to invest its clients' cash in auction-rate securities, right before that market froze over. Of course, when clients called to redeem the money, LandAmerica was stuck. This severe liquidity crunch, combined with poor industry trends, did LandAmerica in.

Bad news for LandAmerica was great news for Fidelity. Post acquisition, FNF ruthlessly cut costs--it laid off people, closed redundant offices, shuttered the corporate headquarters, merged the back office, right sized the agent network, and much more. Fidelity fired around 40% of LandAmerica's workforce and brought the company back to profitability. In fact, due to these cost cuts, Fidelity Title is currently earning 8%-9% operating margins, an impressive feat.

We think the LandAmerica deal will bring strategic benefits to the entire industry. First, having two 600-pound gorillas atop the title insurance food chain augurs very well for pricing in the long run. Less price competition generally equals higher margins. Furthermore, industry consolidation also increases the bargaining power title insurers have over independent agents. Currently, depending on size and locality, independent agents can have significant clout as they choose who gets their business, which is frequently the company that gives them the highest premium split. Industry wide, this split is about 80% (though individual agent rates vary widely). With fewer and more powerful underwriters, I wouldn't be surprised if this premium split erodes slowly over time. Although these trends will be slow to manifest, a few percentage points of extra operating margin makes a big difference to profitability. However, in our valuation, we are not baking in anything explicitly from these trends. So this is just icing on the cake.

Stewart Company Profile

Stewart is the third largest title insurer and traditionally controlled around 12% of the market. Unfortunately, as a rather insular, family-controlled business, it's also one of the less well run title insurers. Over a long cycle, its operating margins are closer to 5%. In contrast, Fidelity's long run operating margins are 10% or higher, and First American is somewhere in between. Our conversations with industry participants also indicate that Stewart has poorer risk controls than the big boys. However, just because Stewart is a mediocre company doesn't mean STC is a bad investment.

Specifically, STC looks very, very cheap. Stewart is trading for $12 today. In comparison, it has a book value of over $24 per share; and EPS averaged around $4.50 between 2001 and 2006, though of course that was during a very good real estate cycle. The balance sheet isn't bad either--gross debt is $82 million, and is essentially offset by liquid assets. But, as a poor operator in a horrible industry downturn, Stewart fared materially worse than its bigger rivals. In 2008, the company lost a mind-boggling $242 million, or $13.37 per share, without goodwill or other impairments! Needless to say, the stock price got crushed and has fallen over 80% from its all time highs of over $50 reached in 2006.

That said, if the real estate cycle is somewhere near its bottom, we think Stewart will survive. Its underwriting subsidiaries are well capitalized and the balance sheet is in good shape. The company has cut costs drastically during the last few years, though probably not as fast as it should have. Today, at a greatly reduced revenue level, it's basically about break-even, excluding retroactive reserve strengthening charges.

Let's leave these risks aside for a moment and think intuitively about Stewart's sustainable earnings power. Here, there are several tail winds. First is the benefit to the entire industry created by the recent union between Fidelity and LandAmerica. The second is that Stewart is gaining market share from Fidelity. Through the third quarter of 2009, its market share expanded from about 12% to around 14.5%. For a small company like Stewart, a modest gain in market share can translate into a material revenue boost. Although we won't know fourth quarter market shares until industry data is published in a few months, I think these gains are sustainable. Fidelity is losing share for three reasons:

• It's cutting its agent network after acquiring LandAmerica
• Big clients (like banks) don't like to concentrate their business in a single underwriter. Given Fidelity's larger presence, we expect customers will take some of their business and "spread the wealth around".
• Some clients are spooked by LandAmerica's old financial troubles and by Fidelity's relatively worse credit ratings.

We think all three reasons are legitimate and enduring. We expect smaller title insurers, including Stewart and Old Republic, to benefit disproportionately from this trend.

Anyway, Stewart has about $1.5 billion in run-rate revenues in 2009. This is down dramatically from the peak of $2.5 billion reached in 2006. In two or three years, as the real estate market thaws, we think revenues can recover to $2 billion. During the "bubble years", operating margins ranged from the low 6%'s to nearly 9% (which is again much worse than peers). We think a more realistic margin is 5%, which is what the company earned in the 1990's. Still, after putting together these assumptions and doing some arithmetic, I think Stewart can earn between $2.50 and $3.00 per fully diluted share in a "normal" economy. While this is not the $4.50 that it earned during the bubble, these earnings can easily justify a stock price of $25 to $30.

The "Secret Event"

As it turns out, the $242 million Stewart lost in 2008 has a silver lining. Because of these losses, the company recorded a huge tax benefit of over $80 million. However, as the company was not expected to make money anytime soon, it had to record an equally large valuation allowance. Thus, this asset is essentially invisible on the balance sheet, and in normal circumstances, it might take many years for Stewart to recoup cash from the government as it slowly returns to profitability.

However, things took a surprising turn on November 6^th, 2009. On that day, President Obama signed into law the Worker, Homeownership, and Business Assistance Act of 2009 (HR 3548). Among this bill's many provisions is one that allows all firms to extend their "net operating loss carryback" from two years to five years. This means companies like Stewart can get immediate cash refunds for their current operating losses instead of getting dribs of money over many years. Since Stewart paid a lot of taxes during the boom years, it's also entitled to a lot of refunds.

Unfortunately, Stewart won't receive over $80 million, at least not right away. However, from my conversations with the company (they won't explicitly disclose the number), I deduce they will get around $45 million. This might not seem like much, but Stewart's market cap is less than $200 million, and the refund will be about $2 per share. In the event that 2010 industry conditions are worse than I forecasted, this cash will also be a valuable cushion.

I believe the Street is not unaware of this event, but it's almost certainly under-aware. For example, FBR Capital initiated coverage of Stewart in December, after the law was passed. Its report did not mention this issue at all. The company is very lightly covered and I am not aware of any sell side report that mentions this tax benefit.

Near Term Outlook and Concerns

Our biggest concern is obviously the housing market, and 2010 could turn out to be a pretty lean year. 2009, though very poor, was helped by two mighty forces. One, super low interest rates generated a mini-refinancing boom, which is going to go away this year with higher interest rates. Two, the homebuyer's tax credit likely drove a lot of people to market, pulling demand forward.

Of course, if the economy does well, home prices and purchases should pick up somewhat. And industry pricing is likely going up this year as well. However, I don't think these positive forces will counteract the negatives. We should be prepared for lackluster revenues and profitability in the short term.

We also have some concerns about Stewart in particular. First, it is an inferior operator and seems to have taken on more risk than peers during the boom. As a family controlled business, its management team is also well insulated (the company is essentially run by a pair of first cousins, descendants of the founder). This is important because if 2010 turns out to be a pretty lean year for the industry, more skeletons can come tumbling out of its closet, and new skeletons may be created as independent agents may engage in more fraudulent activity.

How the company would perform in 2010 is also an important issue. Stewart clearly made a very big mistake in not cutting costs quickly enough as the cycle turned in 2007. The company tells me they will cut hard and fast if conditions worsen. We can only hope that it will do so, but there is really no way to be sure.

Furthermore, Stewart was behind the ball in strengthening its loss reserves. While Fidelity and First America already finished their reserve strengthening measures, Stewart is still hard at work. I currently expect 2010 earnings to be quite weak, impacted by further charges. Should these problems become even worse, our thesis would take a serious hit.

Lastly, the tax refund aside, we expect the balance sheet to get worse before it gets better. Many of the reserve charges Stewart took in 2009 were not in cash. However, these "paper" charges will turn into cash losses as claims pile up. We expect $30 to $50 million to drain out of the company in the coming year.

Miscellaneous Notes

STC recently issued convertible debt with a strike of $12.88. Therefore, EPS figures are on a fully diluted basis of about 23 million shares. Current shares are about 18 million.

On STC's balance sheet you will see a large line item called "Investments-Pledged". This is offset by a liability called "Line of Credit" in the exact same amount. The "Investments" are a large amount of Auction Rate Securities that almost blew the company up a few years ago (similar to what happened at LFG). But in STC's case, their bank essentially took the ARS's back by issuing them a line of credit. This arrangement has zero financing cost to STC, and no credit risk. The line of credit is only recourse to the Investments. I believe the bank made this arrangement to avoid taking ARSs onto its own balance sheet. This arrangement should be unwound in mid 2010.

When we say the underwriting subsidiaries are well capitalized, we are relying a lot of the credit rating agencies. Here are some highlights from the October 12 2009 report by Fitch:

                        2008 Risk Adjusted Capital Ratio

FNF                             73

FAF                             112

STC                             143

ORI                             224

Fitch's IFS rating = BBB+; FAF = A-; FNF = BBB-; ORI = A+

Here is a key metric the rating agencies look at: premiums to statutory surplus. Higher is worse. STC's ratios will get better because the recent convertible issuance plus the tax benefit will increase its statutory surplus by about $100 million.

Surplus Leverage: Premiums/Surplus

STC                                                  3.6

FNF     Chicago Title                            5.5

          Fidelity national Title                 5.0

          Commonwealth Land Title         3.5

          Lawyers Title                           6.6

ORI                                                 3.8

FAF                                                  3.0

FNF's ratios are bad because the LFG acquisition really stretched its balance sheet.

Outlook for STC Quarter

Stewart is releasing results before the market opens on Thursday, so I thought a preview would be helpful.

First, the good news: we have already seen quarterly reports from Fidelity and from Old Republic (ORI), two of Stewart's biggest competitors. We know that the fourth quarter was a decent one for the title industry, as both firms reported good revenue numbers. Old Republic, however, notched up some solid market share gains (I would ballpark it at 3%), allowing its title-related revenues to rise an astronomical 70% over last year. Old Republic is the smallest of the "top four" insurers, so small market share gains translate into a big revenue boost.

I suspect Stewart also had some success at gaining market share, though probably not as much as Old Republic. That said, the favorable industry conditions combined with market share gains will probably allow Stewart to post revenues solidly above Street expectations. Due to the short term operating leverage in the industry, this will be good for earnings as well.

Though this is good, the real test will be the earnings number. Unfortunately, our outlook here is murkier.

The biggest wild card is loss reserves and reserve strengthening. As we said before, Stewart was behind the curve in reserve strengthening (which shows up as a charge on the income statement). While its rivals have largely returned to a "normalized" loss reserve charge of around 7% of revenues, Stewart still had above-average charges as of Q3 2009.

I suspect that this quarter will be more of the same. We already know that the GAAP EPS numbers will be very good, thanks to the tax refund. The big temptation for Stewart management will be to use this tax benefit to create a "cookie jar" of losses and smooth out earnings. In fact, Fidelity did a very similar thing two quarters ago. Furthermore, Stewart has always been a poor operator, so its losses probably would have been comparatively worse regardless.

It's impossible to tell exactly how big the loss reserves will be from quarter to quarter. This depends on market conditions, but also greatly on management discretion. Over the short run, the number can be very volatile.

If losses are abnormally large, it is very hard to predict how the market will interpret the results. People will see the tax benefit as a one-time event, but are much more likely to extrapolate the losses into the future. Therefore, when looking at "clean" EPS figures, after stripping out the tax benefit, STC may not look very good. Lastly, since Stewart doesn't do conference calls, it will be much harder for management to put their own spin on things.

So, I think the quarter could turn out to be a battle between excellent revenues and GAAP earnings and bad loss reserves. My feeling is that the stock will react positively, because the tax benefit is so substantial, but anything can happen.

Long Term risks:

Real estate downturn much longer than expected

underwriting quality much worse than expected

Catalyst

Big Tax refund