Tarragon Corporation is a homebuilder/commercial builder/apartment owner with operations primarily in New Jersey, Connecticut, and Florida.  At its recent price of $11.50, TARR’s stock trades at a 41% discount to our estimated net asset value (NAV) even after marking down all of their exposed properties for the recent drop in the Florida condo market.  (Additionally, we note that TARR trades at just 5.1 times the company’s guidance for 2006 earnings of $2.25 per share.  Given timing issues with regard to earnings, however, we believe that the underlying asset value is a more important metric).  In our view, negative perceptions of homebuilders generally, and the Florida market specifically, as well as confusing GAAP accounting rules are improperly impacting TARR’s current valuation. 

 

In getting to our NAV figure we spent extensive time assessing the company’s financials, engaged in detailed discussions with management, and tested assumptions with local market real estate experts.  While our analysis is no doubt imperfect, we believe it is certainly in the ballpark and the valuation gap provides us with significant margin for error.  More importantly, the majority of the gap between current book value and NAV will be realized via a spinoff early next year.  Furthermore, we believe that the company’s CEO, and his 43% ownership, are clear positives for the investment. In addition to generally being helpful and shareholder friendly, the CEO consistently buys shares personally as well as with the company’s active buyback program.  In addition, management historically has generated strong returns for investors by making smart investment decisions.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

As indicated in the chart, Tarragon has two business segments.  The investment division is their legacy business, which owns and manages residential and commercial properties primarily in Connecticut, Texas, Florida, and other parts of the Southeast.  The other business segment is homebuilding.  This division specializes in urban redevelopment projects in New Jersey, Connecticut, and Florida.  The homebuilding segment also contains numerous Florida condo conversions (and a few condo conversions outside of Florida as well).

 

In 2005 the company decided to separate its investment division from its homebuilding segment in order to realize the true value of the investment division and provide capital to help fuel the rapidly growing homebuilding division.  Consequently, the company divided all investment division rental properties into three different buckets: core, non-core, and hold.

 

Core/Ansonia – These residential properties are located in areas that can be efficiently managed by Tarragon’s in-house management company.  The properties are held in an entity called “Ansonia,” a JV that is 89% owned by Tarragon.

 

Non-core – These are commercial and residential properties that are located in areas that make it difficult for Tarragon’s management company to manage them efficiently, and that the company intends to divest.

 

Hold – These are higher-end residential properties that are suitable for condo conversion.  The properties have been transferred to the homebuilding division, and therefore, we will value them along with the other properties in that division.  It is important to note that these were generally built or acquired a number of years ago and therefore have a low book value basis. 

 

 

The core properties are a collection of 25 residential apartments in Connecticut, Texas, and the Southeast.  These properties are currently held in the Ansonia JV.  Since these properties were built or acquired by TARR many years ago, TARR’s current depreciated GAAP basis in the properties is extremely low.  In fact, as we’ll discuss, the GAAP basis is well below the value that lenders were willing to give TARR on these properties.  The difference between this GAAP basis and the true value is a large part of why TARR’s book value is so understated; we think it is a difference of $8.14 per share.

 

Management has discussed realizing the true Ansonia value via a tax-free spin-off of the properties and the property management company in early 2007.  They believe that the property management company could become very profitable if it added outside properties.  More importantly, a spin-off would allow TARR to avoid paying taxes as it realizes the large book value increase.  We’ll now go through the math.

 

As mentioned above, TARR has more non-recourse (to TARR) debt than it has GAAP asset value in Ansonia.  Simply realizing that difference is a $135mm value.  In addition, according to management TARR has another $37mm of deferred tax liabilities that will go with Ansonia.  In other words, even if the value of the Ansonia spin-off is negligible, there is a positive impact on TARR’s book value because it can eliminate $172mm more of liabilities than assets, or $5.93 per share.

 

Of course, the lenders to Ansonia aren’t crazy, and as Ansonia is most likely worth a good deal more than the debt.  Ansonia currently has $438mm of notes payable, all of which are non-recourse to Tarragon.  In late 2005, Tarragon refinanced 23 properties in Ansonia with General Electric Capital Corporation (GECC).  At that time GECC appraised the 23 properties at $506mm using cap rates of 6.25-7%.  Tarragon management estimates the remaining two properties had an approximate worth of $4mm.  Management believes the value of Ansonia is now worth $550mm due to a 7.5% increase in NOI and a market shift that has reduced cap rates since the appraisal.  Due to the high leverage on these properties and the uncertainty of the valuations we are going to use the company’s original estimate of $510mm to value Ansonia.  Therefore, Tarragon’s 89% equity ownership in Ansonia is valued at $64mm, or $2.21 per share.  This is a gain on top of the $5.93 per share value that the company receives from spinning off Ansonia.  Combined the Ansonia adjustment is $236mm or $8.14 and has by far the biggest impact on Tarragon’s NAV.

Tarragon’s non core properties consist of fifteen residential apartments, three office buildings, five shopping centers, and one piece of land.  These properties have a gross book value of $165mm and a depreciated book value of $123mm.  For this analysis we assume the properties are worth their gross book value, which is below management’s estimate of $179mm.  When you consider the strength of the real estate market over the past several years, and TARR holding some of these properties for over a decade, we consider this to be reasonable.  That implies an after tax profit of $26mm or $0.91 per share.  Additional support of that value comes from twelve of these properties being carried with a negative book value (similar to Ansonia).  Simply giving these properties away would earn $0.50 of the $0.91 per share we estimate.

 

Tarragon has also built two new apartment buildings that have substantial unrealized profit.  These properties are being carried at $47mm and have an estimated value of $62mm according to management.  If we take a 5% discount to management’s estimated value we get an after tax profit of $7mm or $0.26 per share.  Combined with the $0.91 above, this results in a tax affected mark-to-market of $34mm, or $1.16 per share.

 

 

Tarragon’s homebuilding division started in 1998 with its first conversion of a rental property to a condominium.  In 2000, the company began its first ground up construction project, the Las Olas River House, which is the tallest building in Ft. Lauderdale, FL.  In 2001 the company began acquiring land in Hoboken, NJ for its Upper Grand development.  Tarragon’s decision to enter the homebuilding market could not have been much better timed as the company benefited greatly from the housing boom that would soon follow.  The company generated substantial returns for shareholders between 1998 and 2005, but the downturn in late 2005 and thus far in 2006 has reduced the profitability of several projects and has left the company with several underwater projects.  In the sections that follow we have broken down the company’s projects into different categories to better display their assets and the future profitability of those assets.  Ultimately our analysis concludes that the company’s homebuilding division should be marked-to-market at $0.20 per share above its current carrying value.  There are several mark ups and mark downs to get there with the big ups coming from older land in NJ as well as other older properties, and the downs coming from recent Florida purchases.

 

The recent downturn in the housing market has left TARR with six rental properties that it planned on converting to condos in markets where there is no longer sufficient condo demand.  TARR has therefore decided to stick with the rental model and deal with the fact that, as rentals, TARR paid too much for the properties. 

 

While this is clearly a negative situation, and even mentioning the words “Florida condo” seems to make some people apoplectic, it is important to keep the size of these losses in perspective.  The company paid approximately $353mm for six properties that it initially planned on selling out for approximately $500mm after incurring conversion expenses.  According to management the company has capitalized approximately $25mm of costs for interest payments, professional fees, and other miscellaneous expenses, which brings the approximate carrying value of these properties to $378mm.

 

According to work done for us by a 3^rd party property broker, five of these six units have an approximate market value as rental units of $214mm.  We chose to value the properties as rentals rather than condo conversions because, in our view, the conversion days are probably over for some time.  The remaining property has received serious interest from buyers at a break-even level.  Due to deal uncertainty we are marking this property at a 5% discount to its current carrying value of $63mm.  Therefore, if Tarragon were to sell all of these properties, they would have a -38% gross margin, which translates into an after tax loss of $65mm or $2.23 per share.  It is important to note that for TARR to come out worse than that would require a drop in the Florida apartment market, not the condo market.

 

 

Tarragon also has three Florida condo properties with an estimated sell out value of $80mm, which are selling very slowly.  These properties are realizing an average gross margin of about 10%.  Because these properties are selling slowly, we assume the prices on these properties will need to be reduced significantly to complete the sell out.  Consequently, we are marking these properties down to a 25% gross margin.  This results in a mark-to-market of -$8mm or -$.26 per share.

 

Tarragon has three Florida properties with an estimated sell out value of $236mm that are converting with very little to no profit.  In normal times, Tarragon’s hurdle rate for condo conversions is 15%, but the company has recently increased the hurdle rate to 20% due to the uncertainty currently in the market.  However, since these properties have already sold a significant number of units, and are still experiencing strong sales, a lot of the risk has been taken out of these projects.  Consequently, we are marking these properties to a 15% gross margin.  This results in a tax affected mark-to-market of -$22mm or -$0.76 per share.

 

 

The seventeen remaining properties have an estimated sell out of $578mm and are all selling well or are located in areas where sales are still relatively strong.  These properties include the hold properties, which have excellent margins because the company was depreciating the assets while the assets were appreciating in value.  The remaining properties have gross margins of approximately 15%.  This gives a total blended margin of 32% for all of these properties.  Because the vast majority of these properties have already sold a significant number of units and sales on the properties currently available for sale remain strong,we are marking these properties to a 17.5% gross margin.  This results in a tax affected mark-to-market of $51mm or $1.75 per share.

 

 

Tarragon has five projects, located in Florida, New York, and Tennessee, with an estimated sell out value of $191mm.  Tarragon will typically build only a few units ahead of orders on these projects, minimizing the chances of having unsold inventory.  Even though this does not prevent significant land write-offs, it does help minimize the losses in a housing market slow down.  The blended gross margin of these properties is 19%.  Tarragon’s hurdle rate for all new development projects is 20%, which we believe is appropriate considering the risks involved in new development projects.  Marking the properties to values that would yield a 20% gross margin results in a tax affected mark-to-market of -$2mm or -$0.06/share.

 

Tarragon has six New Jersey projects, three Florida projects, and one Tennessee project with a combined estimated sell out value of $688mm.  These properties are all urban redevelopment projects, which take three-to-eight years to complete once the company has gained control of the land by entering into an option contract.  As this is a significant piece of the TARR’s story, it is helpful to understand some background about the urban redevelopment business a bit before getting into the numbers.

 

Getting entitlement and preparing the ground for development takes one-to-five years.  The entitlement process is lengthy because it usually involves getting land rezoned and getting approval from local municipalities whose most vocal constituents are typically against any form of new development.  Therefore, it is critical for a developer to be able to show the local officials that it has successfully completed similar deals.  At this point, Tarragon has built a positive reputation such that the company actually has some mayors contact the company about projects (rather than marketing the other way).  While we have not included it in our NAV, we believe that this side of TARR’s business has real franchise value.

 

In addition to the risk on the project itself, the lengthy entitlement and ground preparation phase allows considerable time for the land value to appreciate or depreciate dramatically.  Fortunately for TARR, market values at their giant Hoboken, NJ project have improved considerably, leaving the project with a gross margin of around 40%.  In addition to the general market appreciation, the entitling process itself has increased this land value, which in turn has increased the gross margin.

 

In instances where land value appreciation does not occur, TARR has some protection from land depreciation as their options do not require them to go ahead and purchase the land.  Once TARR does purchase the land, the build and sell out takes between one and two years for single building developments and up to five years for multi-building projects.  Before construction begins Tarragon mitigates some of the risk by locking in all costs other than finished items such as cabinets, counter tops, and appliances.  Therefore, the major risk involved in these projects is a change in market conditions during the sales process.

 

All in all, TARR’s gross margin on urban development projects isabout 25%.  Tarragon uses a hurdle rate of 20% on these projects, which we believe is appropriate considering the risks and rewards these type of projects offer (where the entitlements are done and the projects are further along the 20% hurdle is arguably conservative).  Marking these projects to values that will yield a 20% margin results in tax affected mark-to-market of $21mm or $0.71 per share.

 

 

Tarragon has three projects inFlorida and Tennesseewith an expected sell out value of $67mm.  Tarragon has a contract to sell Lincoln Pointe, which represents $59mm of this amount, at what should be a $23mm profit.  The other projects have an expected gross margin in total of $1mm.  This produces a blended gross margin of 36%.  Since Lincoln Pointe is already under contract we are using a 5% required gross margin on this property and the standard 20% hurdle rate on the remaining projects.  This produces a blended required gross margin of 7% and results in a tax affected mark-to-market of $12mm or $0.42 per share.

 

 

In addition to the previously mentioned Hoboken projects, Tarragon has a 65% ownership in other JVs with approved development rights that total 1,400 units.  Because Tarragon locked up these units with option contracts several years ago, management believes they could sell the development rights to these units for $60k per unit above TARR’s cost.  We estimate the gross margin TARR would receive if it sold these development rights would be 60%.  We are using a 15% hurdle rate on these land right sales because the estimates come from management and to account for the volatility of land prices.  This results in a tax affected mark-to-market of $26mm or $.88 per share.

 

 

Tarragon has purchased land in five identifiable locations for a ball park estimated total cost of $75mm.  A significant portion of the land was put under the company’s control several years ago.  Even though the land should have appreciated significantly in the last few years, because of rising construction costs, the cooling of the housing market, and other land purchases that we were unable to identity, we are arbitrarily assuming these land projects are underwater by 15%.  This results in a tax affected mark-to-market of -$7mm or -$0.24 per share.

 

 

Tarragon recently announced that it has made material progress on four projects.  Due to a lack of detailed information we have left these projects out of the above analysis.  However, management has told us that it expects gross margins in the neighborhood of 30% on three of these projects and has yet to finalize the expected gross margin on the remaining project.  If management’s estimates are correct this is a positive for two reasons.  First, it demonstrates TARR’s ability to put on new profitable projects even in a slowing housing market.  Second, because our analysis uses a 20% hurdle rate for new development projects, including the three projects that we have gross margin estimates for would add approximately $.75 per share to our mark-to-market analysis.

 

 

As explained above, our estimation of Tarragon’s total sell out value is approximately $2,960mm.  This is made up of a variety of assets, with a variety of sensitivities to changes in the housing and rental markets across different regions of the country.  In our valuation we tried to view the housing market as it stands today, however, given the current negative perception of home values, it is prudent to run some downside scenarios based on declining home sale prices.

 

In the downside case, we assume that prices in all markets decline by 10% from here (for apartments as well as condos and town homes), this would translate into a $185mm or $6.38 per share of tax affected mark-to-market loss.  Since approximately 50% of Tarragon’s properties are located in Florida, a 20% drop in Florida home and rental prices combined with stable prices in all other areas would lead to the same result.  Consequently, a 12% across the board decline in prices is required for Tarragon’s mark-to-market of assets to be equal to its current stock price of $11.50.  It is important to keep in mind that even though this price decline would bring our mark-to-market in line with the current stock price, it would also leave the company with very profitable projects going forward. 

 

 

Sell Out Value	Price Decline	Gross Margin Lost	Tax Rate	Value Lost	Value Lost / Share	Price Decline to Bring Asset to Current Stock Price
2,960	10%	296	37.5%	185	6.38	12%

 

Even though we believe it is more likely that the housing market will continue to weaken rather than improve, Tarragon will benefit from not holding several years of land on its books like many other homebuilders.  Instead, Tarragon holds approximately 1.5 years of properties, and therefore, even if prices slowly decline over the next few years, the vast majority of Tarragon’s properties, excluding its Hoboken developments and a few of the recently announced mixed use developments, will already be sold.  These projects will then be replaced with projects that have additional built in margin since they will be underwritten using current prices. Tarragon’s duration of inventory will begin increasing as the company continues to move away from shorter condo conversion to more urban redevelopment projects, but in the current environment we view TARR’s shorter duration as a positive.

 

The CEO of Tarragon, Bill Friedman, owns 43% of the company (substantially more if you consider all of his friends and family) according to SEC filings. Despite Bill’s hefty stake, management compensation has been very reasonable considering the size and success of the company.  In addition, Bill actually has provided the company a personal line of credit, on which he doesn’t charge a credit line fee, so that the company will not miss out on any potential opportunities.

 

Together with Bill, management is made up of senior people who are creative in developing projects and have the connections to push them through.  Furthermore, it is our belief that Tarragon uses strict underwriting standards when approving projects.  To that point, we have confirmed with third parties that the condominium conversions that are currently underwater were written with very healthy margins based on then prevailing market conditions.  While we wish Tarragon had been more careful about committing to projects in a housing market that looked to be on the brink of slowdown, when you consider how much money they made doing condo conversions during prior years it is hard to completely blame them for taking these losses.  Also, as we noted, management recently increased its hurdle rate to 20% on condo conversions (and are of course now baking in current housing market conditions), demonstrating to us that they have prudently become more conservative of late.

While we do think management operates with integrity there have been two legal cases recently that are quickly worth mentioning.  The first case involved accusations that Tarragon Management, Inc. disturbed asbestos-containing materials during the renovations at Pine Crest Village at Victoria Park in 2003.  This case was settled for $1mm and the company spent an additional $1.6mm on legal fees, professional fees, and remediation.  The second case involves buyers of Las Olas River House suing Tarragon because the square footage posted on advertisements was of the total square footage of the apartments instead of useable square footage.  The difference in square footages ranges from 5.5% to 18.7% depending on the unit, which is an unusually large difference that is caused by the specific design of the building.  The useable square footage was presented correctly in the condo documents.  Also, when the company realized that this large difference existed between the two square footage measurements, it offered all buyers a chance to get out of their contracts.  Management and a very knowledgeable and unbiased local lawyer that we talked to believe this case is without merit.  Although, the lawyer did think this case could have some settlement value due to the bad publicity.  If Tarragon were to lose this case and be obligated to pay back all current owners for the difference between the square footages it could amount to around $.50 per share.  However, we view this as a very unlikely scenario.  In addition, both of these cases are from several years ago and it seems clear that Tarragon has tightened up its governance controls during the last few years as it has expanded its homebuilding division.

The company is currently generating a lot of free cash flow and reducing debt as it sells off properties in both divisions.  They have been using a significant amount of this cash flow to purchase stock in the last year and have accelerated their repurchases substantially since the stock price began to fall sharply in 2Q06.  Bill Freidman commented on last quarter’s conference call that it is hard to justify any new project when the stock is trading at the current price.  He also consistently buys shares personally, again displaying his belief in the company’s value.

 

 

It is not clear to us the exact value of Tarragon’s franchise or for that matter the franchise of any homebuilder.  Considering, however, that we believe we are purchasing Tarragon at around a 41% discount to its NAV, we aren’t staying up at night worrying about the value-add of homebuilders.  The sheer magnitude of the gap also provides enough cushion to weather a lot of little nicks.  Furthermore, with TARR’s management buying stock both personally and for the company we take additional comfort that the value is there and they are doing the right things for investors.  Barring a further meltdown in the residential real estate market from the current slow levels, we expect TARR to provide a substantial return from current prices.