Description

The past few months have seen several optimism-heavy, analysis-light opinion pieces on the value of The St. Joe Company (JOE). Given this recent editorial flow, we felt compelled to outline a case that rests on more traditional valuation techniques. (David Einhorn of Greenlight has made the case eloquently in the past, but it is time for a refresher).

 

JOE owns roughly 638,000 acres of land, most of it in northwest Florida. For about the last decade, JOE was an integrated provider of real estate products and services. The company handled everything, including land sales, development, homebuilding, resort/community management, etc. In October 2007, after curtailing stock buybacks and scrapping the dividend, management announced a dramatic restructuring of the business: going forward, JOE would focus exclusively on selling raw and/or entitled land and would consider minor development work on a case by case basis – essentially a wholesale business model shift. Implied by this transformation is a reduction of capital investment and costs as JOE effectively exits the homebuilding and development businesses, but vastly increased dependence on homebuilders and developers, to say nothing of a reduction in realized prices for (and likely substantial deferral of) future property sales. As a land holder with the express goal of liquidating its holdings, JOE is a depleting asset. Given this inherently finite business model, which presents no problematic terminal value calculation for us to ponder, JOE is a natural candidate for DCF valuation.  

 

A fluffy piece in the April 26^th issue of Barron’s highlighted the essence of the bull thesis on JOE, which to paraphrase is: “how can you go wrong buying Florida real estate at ~$5,700 per acre?” (EV/acreage after the Q1-08 sales).  For some reason, JOE seems to be about the only company that many investors are willing to value on a gross, undiscounted basis, despite the fact that the time value of money is a tremendous driver of equity value, particularly given that most experts describe the development of JOE’s properties as a 30+ year project. Before getting into the specifics of valuing the various real estate assets in question, one overarching point must be made: gross enterprise value per acre does not represent the actual cash value per acre that will accrue to the company, let alone to the shareholders. Just like every other company, even those owning land bought very cheaply years ago, JOE pays operating expenses and taxes and invests in capital expenditures to prepare properties for sale and for miscellaneous development projects.   

 

So the first question to answer is: what’s the real implied price per acre you’re paying at today’s share price after incorporating the various cash costs associated with running the business?  To get a rough sense, we build a very simplified DCF model with a few basic assumptions:

 

(1)   Timing of land sales: all of the land sells within 20 years, implying a rate of 5% per year or ~32,000 acres/year (this is faster than the company would estimate).

 

(2)   Land value appreciation: straight-line price appreciation of 3% per year.

 

(3)   Expenses: overall cash expenses of 20% of sales (this is much lower than historical levels, where gross margins for home-site sales have been around 50%). This includes direct land-related costs and corporate expenses.

 

(4)   Taxes: we use a 25% rate, which is what the company estimates. Management uses various deferral strategies, most frequently installment notes, to get to this low level, and we will ignore future tax liabilities for purposes of this analysis.

 

(5)   Discount rate: we use 10%, which is probably quite generous considering the company is fully equity-financed.

 

(6)   Capital expenditures: for purposes of this analysis, we’re ignoring them, just like we’re ignoring depreciation.  Obviously these items exist, but we’re trying to run a very stripped down DCF here – so essentially net income equals free cash flow.

 

We use these assumptions to back-solve into the starting sales price per acre, which at the current share price is about $18,400. This is more than triple the price that the bulls suggest investors are paying per acre of JOE land at the current share price. We concede that $18,400/acre for Florida land, at least on the surface, still sounds somewhat reasonably priced, which brings us to our next major point.

 

Only ~45,000 acres of the company’s land is entitled, or in the entitlement process. The other ~593,000 acres out of JOE’s total ~638,000 acres are un-entitled rural pine forest and/or swamp, most of which is currently used for timberland. JOE plans to sell off this land for either “rural recreation/residential” or for timber value. Such values are far lower than $18,400 per acre, and are significantly lower than even $5,700 per acre. First, a few timber value comparables from transactions completed during 2007 in the southeast: MeadWestvaco sold 323k acres to Wells Timberland REIT for $1,238/acre; Temple Inland sold 1.55mm acres to The Campbell Group for $1,535/acre, and iStar Financial sold 900k acres to Hancock Timber Resources for $1,900/acre. Second, a few comparables for “rural recreation” (the hope that rich outdoorsmen will buy large tracts to hunt quail, deer, and other local wildlife): the most prolific quail country in the US, located in South Texas, goes for about $1,500 to $2,000 an acre. Some might try to point to the quail plantations selling in the “Red Hills” region (between Tallahassee and Thomasville) for $3,000 to $5,000/acre, but none of JOE’s land lies in that desirable hunting region. Rural developers like Southern Pine Plantations are selling land in JOE’s region for under $2,000 per acre. Third, the most telling indicator of land value: the prices at which JOE has actually sold its rural land over the last few years:

 

                        Year                Acreage          Price                Price/Acre

                        Q1-08              57k                  $91.1m            $1,586

2007                106k                $161m             $1,522

                        2006                34k                  $90m               $2,618

                        2005                29k                  $69m               $2,379

                        2004                20k                  $68m               $3,375

                        2003                65k                  $97m               $1,487

 

We would note that the 2004 figure is skewed upward by the sales of two prime pieces of waterfront property (~650 total acres) for $17.4m.  Excluding these unusual pieces of land, the average per acre price in 2004 was ~$2,500.

 

Knowing that rural holdings comprise the vast majority of JOE’s land, we adjust the DCF analysis slightly. We separate the revenue stream from the rural land (593k acres), and make two generous assumptions: first, this land sells for an average of $2,500/acre and second, JOE sells this land at a rate of 20% per year, or ~119,000 acres per year (assuming a faster sales rate of ~176k acres in 2008, given that ~57k acres have already been sold). With these assumptions, the implied price per acre for the remaining ~45k acres of land in varying stages of development is about ~$210,000 per acre, which no longer sounds so cheap. Many analysts allude to JOE’s last “land analysis update” released on May 10, 2006. The following passage is not often mentioned, but it stands out: “JOE’s commercial land sales in Northwest Florida excluding multifamily land, from April 1, 2004 through March 31, 2006 averaged $168,000 per acre. The average price of multifamily land sold during this period was $87,000 per acre.”  Note that on the recent Q1 2008 conference call, JOE CEO Peter Rummell referred to August 2005 as the real estate price peak.

 

Anyone can build the basic DCF analysis described above to see that contrary to the bulls’ argument, you are actually paying quite a rich price for JOE today.  This is admittedly simple math, but it does highlight the fact that JOE’s stock isn’t cheap even using very basic (and generous) assumptions.  A more realistic DCF analysis involves separate valuations for the different categories of developed and entitled acreage.  We have gone through this analysis in substantial detail, separately modeling revenues for each development property, as listed in the front of the 2007 10-K, based on comparable property valuations, input from various industry experts and commentary from the company itself. Key assumptions include per unit (lot) pricing for the residential real estate, per acre pricing for the rural and commercial land, sales schedule (rate of sales), cash expenses, tax rate, capex, and discount rate. We use a 10% discount rate, a 25% tax rate, 3% yearly price increases, and we assume that all properties are sold by 2030. Per management guidance, annual capex never rises above $90mm, and we model corporate expenses well below historical levels. Using conservative selling price assumptions, our analysis implies a share price of around $15.

 

The assumptions that drive our valuation of the company are influenced by conversations we’ve had with industry experts and JOE employees and through site visits in northwest Florida.  A few of the key points include:

 

1.      October 2007 restructuring implies lower realized prices and greater reliance on customers

a.       Lots typically sell for 15-20% of the anticipated home price (waterfront lots are generally higher at about 25% of home price). While expenses are much lower (which is why we implied 80% operating margins in our DCF), revenues will also be much lower.

b.      Before the restructuring, JOE was a vertically integrated real estate provider: land owner, developer, and homebuilder. Now, it is just a land sales operation participating in limited development activities. Thus, JOE’s primary buyers are now homebuilding companies and developers (hard to imagine a weaker or more fickle customer base).

c.       For more on the perils of a strategy dependent on third party builders and developers, read about experience of Calpers with their Newhall Ranch project (article in WSJ, 05/01/08). Worth noting is that Newhall Ranch failed despite an excellent location (about 20 miles from downtown Los Angeles) where one would expect employment opportunities to create demand for residential capacity (unlike northwest Florida, where there is no industry).

 

2.      Recent demand was largely investor-driven, implying massive excess capacity

a.       Most of the buyers over the past few years have been investors. Several of the JOE developments released in the last two years have been selling, but very few people are actually building houses, let alone moving into the neighborhoods.

b.      True buyers (not investors) are reluctant to buy and build in new developments, because it can be years before actual neighborhoods exist, a classic chicken and egg problem. Rivercamps and Windmark Beach, where build schedules have been extended because property owners are balking at beginning (required) construction, are good examples.

c.       There is a glut of “luxury” properties in the area – even the most successful developments have plenty of vacancies, and there are many developments that appear to have been abandoned mid-project. Even if true demand returns, there will be years of excess capacity to soak up. JOE is also not the only game in town – a quick Internet search reveals many other similar projects, many of which are failed or are failing, which suggests prices have room to fall.

d.      On the Q1-08 conference call, CEO Peter Rummell said of Florida resort and residential markets: “Inventories seem to be ranging from anywhere from 12 to 48 months supply.” If northwest Florida is anywhere near the upper end of that range, then JOE will face a very difficult next few years.

 

3.      Circular logic: JOE is attempting to create both housing supply and demand

a.       JOE is attempting to create new commercial zones to attract full-time residents, and many of their new developments are “affordable” housing where this new workforce can live. The obvious question, however, is what will create the explosion of commercial activity JOE seeks?

b.      Apart from real estate/tourism, there is virtually no industry to speak of in the area. The fact remains that most of the commerce in Panama City and Panama City Beach consists of pawn shops and dollar stores (even on the main strip 50 yards from the beach in Panama City Beach). Pier Park Mall, JOE’s attempt at improving the retail environment, still has many vacancies.

 

4.      JOE’s best developments have already been sold

a.       Nearly all of JOE’s remaining development inventory is in Bay and Gulf counties, which are considered to be less desirable than Walton County (the site of JOE success stories such as Watercolor and Watersound)

b.      Most of the remaining residential developments are aimed at a lower price point.

 

5.      Unrealistic value creation is being attributed to the new airport

a.       Many investors seem to be placing their faith in the hypothetical value of a new airport (opening 2010 at earliest), but they seem to forget that Panama City already has a decent-sized airport, only 15 minutes away from the site of the proposed airport. In fact, the current airport is closer to JOE’s properties in Gulf County.

b.      The new airport will have a longer runway, which can accommodate larger aircraft, but it doesn’t sound like the facilities will be much larger than the old airport, as only seven gates are planned.

c.       The existing airport is already served by major carriers, including Delta, Northwest, and USAir. It has direct flights from several of the urban areas that are key markets for JOE, including Atlanta, Charlotte, and Memphis.

d.      Most importantly, the current airport’s website (www.pcairport.com) reveals some rather troubling trends.  Despite Florida’s much hyped growth and JOE’s non-stop promotion of northwest Florida, passenger traffic has been in decline for several years:

 

Period             Passengers     % Change y/y

                                    Jan-08              21k                  -13.4%

2007                338k                -4.8%

                                    2006                355k                -7.3%

                                    2005                383k                -1.0%

                                    2004                387k                +3.8%

                                    2003                372k                +8.5%

 

In fact, 2007 passenger traffic is almost unchanged from 1998 (336k passengers).  We have not studied the airport’s activity levels in extensive detail, so we concede that other factors could be at work, but at the very least these statistics suggest some concerning trends. If there is such demand for air travel to Panama City/northwest Florida, why have passenger numbers been dropping for the last three years, including 2005, the “peak” year of the recent bubble, and stagnant over the last ten?

e.       Bulls like to cite the example of Ft. Myers, which built a larger airport in 1983 and has seen dramatic growth ever since. We doubt this is a great comparison.  Southwest Florida has a more desirable climate and better access to larger population centers.  A recent research piece from Paul Puryear at Raymond James (dated 04/16/08) does a good job explaining why the Panama City area is not likely to repeat what transpired in the Ft. Myers area – and why even if the airport proves to be a boon, the length of time required to absorb all of JOE’s properties is closer to fifty years than to five.

 

6.      Commercial land value data point is misleading

a.       The recent article in Barron’s cites that JOE sold its commercial acreage for an average price of $266,000 per acre last year.

b.      This could be a reasonable estimate for some of the commercial land sales going forward, but JOE has only ~964 acres of entitled commercial land remaining (see 2007 10-K). Using $266k per acre (vs. say, $200k) as the value for JOE’s commercial portfolio barely moves the NPV needle.

c.       Bulls may argue that JOE will entitle more acreage (and they are probably right) but if so, it is a very long way off: as JOE will admit, there is already a multi-year backlog of available for sale real estate in the region. JOE would be lucky to sell its currently entitled commercial portfolio within the next 20 years, let alone entitle additional acreage for commercial development and put it on the market.

 

7.      Rural land is generally low quality

a.       JOE’s management has promoted the idea of elevating their timber land to “higher and better use,” but it must be remembered that this is low quality land.

b.      We won’t hazard a guess at how much of the remaining rural acreage is wetlands, but ecological maps suggest that it is a significant portion. See http://wetlandextension.ifas.ufl.edu/near.htm for county by county detail.

c.       Substantially all of JOE’s remaining land is below the escarpment separating the coastal sandy areas from the uplands with quality topsoil. Because of this, JOE’s land is not ideally suited for hunting purposes (not much game). “Sand pine” forests with relatively low timber value predominate.

d.      We have learned that a single real estate investor, who plans to resell the land, accounted for a large portion of the rural sales in Q1-08. Thus, much of the rural land is not going to end users: it will reappear on the market and compete with JOE’s future sales, effectively depressing prices. 

 

8.      Equity offering eliminates debt fears, but underlines the headwinds facing JOE

a.       Much has been made of the recent equity offering. It is positive that JOE no longer faces the time, capital and liquidity constraints of its former debt load.

b.      However, the deal, priced at $35/share, was highly dilutive to equity holders.

c.       Further, it demonstrates management’s lack of faith in the short- to medium-term environment: it is a tacit admission that the company has no faith in its ability to generate cash flow for many years. This has obvious and negative implications for investors piling in today.

 

9.      If you are looking for cheap Florida real estate, there are better options, like REOs

a.       JOE is not a cheap way of buying land in northwest Florida. If you are looking for a bargain, buy into the REO portfolios that offer foreclosed properties for 50 cents on the dollar.

b.      JOE’s current share price, on the other hand, discounts optimistic real estate prices and an unrealistic development schedule.

 

The risk to our thesis is the opacity/debatability of JOE’s land valuation.  While we have attempted to realistically value the properties at a very granular level, many investors will simply brush away any detailed valuation and succumb to the analytically un-taxing bull thesis.  JOE is also a bit of a cult stock, which can make for dangerous short selling conditions. 

Catalyst

1. Continued deterioration in Florida real estate prices.
2. Financial distress for neighboring properties in northwest Florida.
3. Delays in airport construction.
4. Yet another business model shift from JOE?