Description

Trex pioneered the process for making composite decking material out of waste wood fibers (from sources like furniture manufacturing plants) and reclaimed polyethylene (primarily from recycled plastic grocery bags). I believe that Trex is an attractive investment opportunity for the following reasons:

1. Small business, with a
2. Unique product, and a
3. Large market potential.
4. Good management, with a
5. Very large equity interest.
6. Outstanding historical financial track record.
7. A business that I can understand.
8. Reasonable valuation.

The valuation has become reasonable as a result of a recent stock price decline, which resulted from the pre-announcement of poor 2001 Q2 results on June 18. At $17 per share, the equity has a market capitalization of approximately $240 million, equal to 13 x 2000 net income of $19 million. Including debt of approximately $80 million, the total valuation of the business is approximately $320 million, equal to 8 x 2000 EBITDA of $39 million.

Below are excerpts from my notes from a meeting with Bob Matheny, President, and Tony Cavanna, CFO. And at the end of these notes are excerpts from my notes from subsequent phone conversations.

1. Unique product

This is my perception, but it is difficult for me to evaluate. However, there are a number of factors that lead me to believe that Trex is unique:

 Reports from lumber yards
 Product award
 National Evaluation Service code listing
 Market share
 Brand name
 I personally think it’s attractive material for a deck

Reports from lumber yards (also referred to herein as dealers)

I have spoken with six lumber yards that carry Trex, and they all said that Trex is by far the best composite decking material on the market, and superior to most wood for decking because it doesn’t require sealing or painting, it doesn’t splinter, and it lasts forever. I selected these lumber yards from Trex’s Web site, the list included what I believe are some of the best lumber yards in Jackson Hole, Idaho Falls and Salt Lake City. They all said that they sell a lot of Trex, and it is a very popular product.

Product award

Trex won Home Magazine’s 2001 American Building Products Award in the building materials category.

National Evaluation Service (NES) code listing

This is the only composite decking material that has received the National Evaluation Service building code listing, because of its excellent physical and mechanical properties. This makes it easy for an engineer or architect to design Trex into a project, and it facilitates the acquisition of building permits. It cost Trex $1.5 million in third party testing fees, and took over 2 ½ years, to achieve code listing in 1995. An NES inspector makes a surprise visit to Trex’s manufacturing facilities once a month. Matheny said they are surprised that no competitors have received this listing, and suggested this might indicate quality problems with competing composites. He said Advanced Environmental Recycling Technologies has been trying to achieve code listing for its product for over a year. Given that it apparently takes some time to achieve listing status, we may see competing products attain this status over the next few years.

Market share

Matheny said they believe Trex has approximately two-thirds of the composite decking market, and composites now account for at least 75%, and perhaps as much as 90%, of the non-wood market (which includes 100% plastic products). This large market share is probably attributable to a number of factors, such as first to market, and heavy advertising and promotion spending to create a brand. However, I would suspect that product characteristics and quality are also important factors in its success. 100% plastic products are rapidly disappearing from the market because of the way they expand and contract in extreme temperatures. One dealer in Idaho told me about a 100% plastic deck that collapsed in the winter because it contracted off of its support structures.

Competition

Trex’s 1999 10-K lists five competitors: Advanced Environmental Recycling Technologies, Crane Plastics, Eaglebrook Plastics, Royal Crown Limited and U.S. Plastic Lumber. In addition, there are two new competitors: CertainTeed and Lousiana-Pacific. Of all of these competitors, the only one that management is concerned about is Louisiana-Pacific. The others have inferior products and small market shares (Advanced Environmental is the largest with sales of $20 million). Trex’s IPO, and the resulting disclosure of its financial performance, is what prompted L-P to enter this market. L-P is already a major supplier of southern yellow pine for wood decks.

In January 2000, L-P announced that it would construct two wood fiber composite decking facilities, which would come on-line in the 3rd quarter of 2000. It also said that it expects to construct more plants over the longer term. Its press release said this move will “leverage strengths of our existing sales force strong industry relationships and distribution channels to quickly become an important industry player.” It expects the composite decking market, which has doubled each year since 1995, to grow to $350 million by 2005, and $600 million by 2010. In May 2000, L-P agreed to acquire the composite decking assets of Hoff Companies, Inc. in Idaho. The press release said this “gives us an established, immediate presence in West Coast markets.” The assets included a 22,000 square foot manufacturing facility that was built in Meridian, ID in late 1998, and some kind of 70,000 square foot facility on the same site that they said allowed for expansion. The press release went on to say that this was an “excellent fit with L-P’s 50,000 square foot greenfield facility that was under construction in Selma, Alabama. Both (facilities) will produce using Strandex proprietary technology.” Matheny said that Mobile/Trex considered licensing Strandex technology in 1993, and concluded it was inferior to Trex. It was similar to plastic extrusion in that it needed very clean raw materials.

Trex believes that its proprietary manufacturing process, which is patented (one of two patents, and the only one that is important), gives it a significantly lower cost of goods sold than that of its competitors, because it allows it to use less clean polyethylene (its most expensive raw material, accounting for almost 50% of COGS), and polyethylene with more variation. In addition, they have increased their rate of production 3-4 times since they started in 1992, which has reduced costs. My visit to the factory convinced me that, in addition to the patented process, there is also a lot of proprietary “know-how” in the production process that would be difficult for a competitor to replicate. (I have highlighted this because I believe this is probably Trex’s most significant competitive advantage.)

I am concerned about L-P’s entry into the market. I am concerned that L-P may price its product very aggressively to break into the market. I am concerned that, if L-P has a good product, Trex will become more like a commodity, and have to lower its prices. No one can predict the impact. In response to my question as to whether L-P was aggressively trying to break into the market, Matheny said that L-P had offered to buy all of the Trex stock of one of their distributors, and replace it with Strandex product on six month consignment. Their distributor declined the offer. Matheny said they were not aware of any other, similar incidences, but I think this indicates that L-P is going to be aggressive.

CertainTeed, the US building products subsidiary of France’s Saint-Gobain Corporation, the world’s largest building products company with almost $30 billion in revenues, is potentially a significant competitor. They have good technology, and a composite product that uses vinyl scrap from CertainTeed’s production of vinyl siding. Cavanna said they believe that CertainTeed’s limited supply of vinyl scrap may limit its composite decking capacity. He also said Trex hasn’t seen them in the market yet, and they appear to be 2-3 years behind where they said they would be at this point in time.

Only time will tell the impact of L-P’s entry into the market. I view this competitive threat as the biggest risk to a long-term investment in Trex. Near term, Trex could suffer operationally, and its stock could decline further, if the US experiences a recession, and demand for decking declines. However, this should be a cyclical decline, as opposed to a long-term secular decline. On the positive side, I think the market may be large enough for two strong competitors. And, in a way, L-P’s entry helps validate the relatively new concept of composite decking. Also, L-P’s product may not have the same qualities and characteristics as Trex, and may not be well accepted. If Matheny and Cavanna are correct about Trex’s production process having substantial cost advantages (unfortunately, this is difficult for me to verify), then L-P may not be able to undercut Trex pricing. Matheny went so far as to say that “economics don’t allow L-P to buy into the market, unless they want to a lose a lot of money.” (Hope he’s right.) Also, L-P is a behemoth ($3 billion in sales), and perhaps not the best environment to nurture a start-up.

Brand name

Matheny and Cavanna believe that creating a brand name has been a very important factor in their success. This is the work of their fourth partner, Andrew Ferrari, a marketing person, who built the Hefty bag brand for Mobile. Ferrari convinced them at the beginning that they should build a brand. Their goal is to create a building materials brand like Corian, which is used for countertops and has been a very successful product. To that end, they have invested $15 million in advertising (90% directed at consumers, through 18 consumer magazines and cable TV) and promotion over the past three years, and expect to spend $35 million more over the next three years. They want to be the Kleenex of composite decking. They want potential customers to walk into lumber yards and ask for Trex when they want composite decking material. My perception is that this is happening. However, $35 million is a lot of money, and I have no idea whether it will generate incremental sales and margins that produce appropriate returns on the investment.

2. Large Market Potential

Matheny said that factory sales of decking material were approximately $2.1 billion in 1999. (This excludes the posts, beams and columns used for a deck’s substructure. Trex does not have the tensile strength of wood, and therefore, is not suitable for these applications.) The market for decking material has been growing at a rate of approximately 8% per annum, which is roughly double the growth rate of the home improvement market (a $125 billion market). Non-wood decking has been growing at 4-6 x the overall decking market’s growth rate. Trex’s sales of $111 million in 2000 would represent approximately 5% of the decking materials market. Other composite decking and 100% plastic products account for 2-3% of the market. The $2.1 billion is primarily in “wood dollars”. If all decking were Trex, which, at factory prices, is approximately 150% the cost of wood, the decking market would be approximately $2.8-3.0 billion. By volume, Trex currently has 3-4% of the decking materials market.

I asked what percent of the wood market was realistically a candidate for substitution. Bob said they originally thought that somewhere between 35% and 70% of the wood market was a realistic target. However, they now believe their target market is the entire 80% of the wood market that is comprised of treated pine and fir. The remaining 20% of the wood market is high-end cedar and redwood, where a homeowner is not likely to substitute Trex. I would suspect that the very cost conscience low end of the market is probably also not a good candidate for substitution.

At retail, Trex costs about twice as much as the treated yellow pine that is used for most decks. However, according to Cavanna, the wood used for deck surfaces represents only about 20% of the total cost of the deck. Therefore, Trex adds only about 20% to the total cost of a deck. They claim Trex is less costly than wood over the entire life of a deck, because it does not need to be maintained. (Estimates of cost differences are difficult, because there are so many variables: the type and grade of wood, the climate, how often the homeowner chooses to sand and reseal the deck, etc. However, Cavanna said that the total cost of the average Trex deck is about $15-16 per square foot (this includes the post, beams and columns used in the deck’s substructure), and the total cost of the average wood deck is approximately $12-13 per square foot. He said it costs approximately $1-2 per square foot to sand, stain and seal a wood deck. This is typically done every 1-2 years, which means that after 2-4 years the total amount spent on the wood deck would be approximately equal the cost of the Trex deck.) According to one lumber yard that I spoke with, even if a wood deck were not maintained (unlikely), Trex would be cheaper, because the wood deck would have to be replaced in 8-10 years. (Trex has been in existence for only about eight years, but in Trex’s R&D lab, they have simulated 35 years of wear, and the product has not changed.) Cavanna said that the life of wood decks is decreasing, because of a decline in the quality of timber. Old growth wood, which has tight growth rings and is resistant to splitting and cracking, is no longer used for most decks. Young trees, which are about all that is available at an affordable price, have wide growth rings, and are fine for pulp, but make poor building material. The price of wood is expected to increase over time. However, prices are cyclical and have declined by 20-30% in 2000.

According to Matheny, the installed base of decks in the US is approximately 30 million. Approximately 3 million decks are built every year, 85% of which are either new decks added to existing homes or replacements of existing decks. Decks on new homes account for approximately 15% of the total decks that are built each year.

3. Good management

It is difficult to judge the quality and capability of managers in one meeting. I met two of the four top managers - probably the two most important (based on their slightly higher salaries and titles - all four started with the same number of shares). I liked both Matheny and Cavanna very much. They seem like nice people, intelligent, hard working and straightforward. A friend of Justin Adams, Jay Wetzel, who is a judge in Winchester, VA, knows Roger Wittenberg well, and thinks highly of him. Winchester is a small town, and Jay said that all four managers have excellent reputations. I first learned about Trex in an April 2000 Barron’s article, about a former corporate recruiter turned fund manager, who specialized in picking small cap stocks with strong managers. His favorite management team was that of Trex. However, I have no way to determine whether this fund manager is really a good judge of operating managers. The one piece of tangible evidence of good management is Trex’s outstanding financial performance. These guys were probably lucky to be in the right place at the right time, but I suspect they created some of that luck. It is worth noting that all four managers are between 50 and 60 years old: Matheny (54), Cavanna (60), Ferrari (53) and Wittenberg (51).

4. Managers have a very large equity interest

The four senior managers currently own 58-59% of the equity.

They each invested $500,000 at the time of the buyout in 1996 for an aggregate equity interest of 75%. Just prior to the IPO, and based on the IPO valuation and Connecticut General’s unrealized return on its investment, the managers’ equity interest was increased from 75% to 90%, and Connecticut General’s equity interest was decreased from 25% to 10%. Also, prior to the IPO, Trex, which was an LLC, distributed all of its capital to its investors. This resulted in each manager receiving a return of his original $500,000 investment, plus an additional $1.7 million after taxes. After the IPO, each manager owned 2.138 million shares, equal to approximately 15% of the equity, for a total management equity interest of approximately 60%.

The managers have sold shares twice since the IPO. In the first sale, the amount sold by each manager ranged from 60,000 shares to 140,000 shares, so their equity interests are no longer equal. In the second sale, the amount sold by each manager ranged from 10,000 to 12,000 shares. (I don’t know how much the managers received in these sales, but assuming a price of $35 per share, the average amount realized after taxes by each manager would have been approximately $3 million.) Matheny said that all four plan to sell more shares over the next 4-5 years. He said that he could not speak for the others, but he expected to retain a permanent stake of somewhere between 8% and 12%. Cavanna said he would probably do the same. They are currently exploring the new SEC insider sale program, where a certain amount of shares are placed in a trust, and sold automatically at certain dates in the future.

One thing that I don’t have to worry about with Trex is these guys giving away huge amounts of employee stock options. From my perspective, the economic value of stock option awards at most high tech companies rules them out as investment opportunities. After factoring in this non-cash compensation, these businesses usually have no profits.

5. Outstanding historical financial track record

Sales

Sales have increased at a compound annual growth rate of 47% over the past four years. Volume increases accounted for most of the growth, with average annual price increases of approximately 5%.

While full year 2000 sales are expected to increase by 49% from $74 million to $111 million, growth has slowed sharply in the second half. Sales increased 26% in Q3, and are expected to be relatively flat in Q4. Cavanna said this is because expanded production capacity finally allowed them to take their wholesale distributors and retail dealers off of allocation in Q3, and wholesalers and dealers who had been stockpiling inventory began “destocking”. They don’t have a good idea of how this “destocking” will effect them. Q4 is normally their slowest quarter, and they are predicting it will be flat with 1999 Q4. They also think this “destocking” may impact 2001 Q1, and are predicting 20% sales growth for this quarter. For all of 2001, they are forecasting sales growth of 25%. I wouldn’t be surprised if this slowdown is also at least partially attributable to the same factors that are adversely affecting Home Depot and Loew’s: a slowing economy, and a decline in home sales and home remodeling. Matheny said they believe that decking material is less cyclical that other building materials categories, but I don’t think there is a lot of empirical evidence to support this. He said that the decking market continued to grow steadily in 1994, 1995 and 1996, years in which housing starts were down. (I’ll take his word on this. I’m too lazy to try to research the facts.) He also said that the typical deck is added five years after a house is built, which, if true, would be positive, given the large stock of housing put in place over the past five years. The bottom line is no one knows what shape this economic downturn might take (if indeed we have one), and how this business will be affected.

Margins

Gross margins were 51%, 51% and 54%, in 1997, 1998 and 1999, respectively. The gross margin in 2000 is expected to decline to approximately 50%, because of increased production at the Fernley, Nevada facility, which has higher cost of goods sold than the Winchester facility. In an effort to ramp up production quickly this year, they produced some poor quality product at Fernley, which, unfortunately they shipped to wholesalers. They ended up replacing all of it. However, Fernley’s cost of goods sold is also higher because it has higher raw material costs, because it has to ship its raw materials from greater distances. For example, most of its waste wood comes from Los Angeles, and it costs $0.05 per pound, compared to Winchester’s cost of $0.015 per pound. The Fernley plant was started to accommodate West Coast customers, who were frustrated that their rail shipments of Trex were taking 6-8 weeks to reach them. Raw materials account for approximately 55% of COGS, manufacturing overhead approximately 35%, and direct labor 15%. Waste wood and polyethylene are used in equal proportions (50/50) in the manufacturing process. However, polyethylene is about 10 x more expensive at $0.14 per pound, compared to waste wood at $0.015 per pound. The prices of these waste materials are not affected to any material degree by wood prices or oil prices. The price of waste wood is primarily a function of the cost of shipment, and waste polyethylene prices are primarily a function of supply and demand, and shipping costs.

EBITDA margins averaged 33% from 1997 to 2000E, with a range from 31% in 1997 to 34% in 1999 and 2000E. Matheny said they expect they can maintain EBITDA margins in the range of 30-36% over the next few years. EBITA margins averaged 27% from 1997 to 2000E, with a range from 25% in 1997 to 29% in 1999. My model assumes future EBITDA margins of 31% and EBITA margins of 25%.

Capital requirements

Working capital requirements have been close to zero. They may increase a bit in the future to, say, 5% of sales, but to simplify modeling I assume that working capital will continue to be zero. Fixed capital requirements, however, have been extremely high (97%) because of construction of the Nevada facility and the addition of production lines in both Nevada and Winchester. This ratio will continue to be high for 2-3 more years, because of expansion at Winchester, the acquisition of a third plant site and construction of a new facility, and the addition of more production lines in all three locations. As a result, it is very difficult for me to estimate a long-term, equilibrium fixed capital ratio. The model assumes a fixed capital ratio of 60%, and shows sensitivities in increments of 20%. A 20% increase in capital requirements reduces the theoretical value of the stock by $2-3 per share, depending upon the other assumptions. For example, under my “base case” set of assumptions, increasing the fixed capital ratio from 60% to 80% would reduce the theoretical value of the stock by $2, from $26 to $24.

6. A business that I can understand

One of the things I like about Trex is that it is a simple business, and I think I understand it. I know what a deck is. I understand why people want them. I think the demand for decks may be cyclical, and I am concerned about the entry of L-P into the market, but decks are not going to go away. And if Trex is right, and I believe they are, the substitution of composite decking material for wood is going to continue to drive their growth at a high rate for many years.

Conclusion

As with any investment opportunity, Trex has some risks, or negative factors. These include:

 A potentially strong new competitor in L-P
 The possibility of composite decking becoming more of a commodity
 The possibility of a cyclical downturn
 The managers reducing their equity interests

However, after reviewing the factors that initially attracted me to Trex, I believe they are all probably true or valid. And I believe that Trex is currently priced to provide an attractive expected return, using a reasonable set of assumptions for its future performance. The stock could obviously decline further, but I believe it offers an attractive investment at its current level.

Catalyst

Near term outlook favorable

Cavanna said that the near term outlook is favorable. As indicated above, sales to distributors in their top ten markets that account for 71% of their total sales are down 9-10%. However, the sales of these distributors are up 12-13%. He said their 2,900 dealers are hard to survey, but they have called about 100 of them, and Trex sales at every one of them is up from 25% to 150% over last year. June sales are expected to be double April’s sales, and approximately equal to April and May combined.

Cavanna said Trex is a much stronger company today than it was a year ago, for the following reasons:

1. they have 2,900 dealers, compared to 2,100-2,200 at this time last year;
2. they have 87 distributors, compared to 62-63 at this time last year;
3. they have 21 sales reps, compared to 13 at this time last year;
4. plant productivity has increased by 11%; and
5. management infrastructure has been put in place to manage a bigger business.