Valuing Timberland IV

This is the forth post of a series on timberland valuation. If you missed the first post, which included an Overview plus a discussion on Disaggregation, you can read “Valuing Timberland I” here. The second post focused primarily on productivity and how that fits into today’s appraisal systems which use discounted cash flow techniques to determine the value of timberland. You can read “Valuing Timberland II” here. The third post focused on estimates of timber volumes and values, how we get them, and how to forecast them for future year’s cash flows. Read “Valuing Timberland III” here. Today, I’ll focus on how all of the previously discussed factors are tied into some form of discounted cash flow (DCF) model.

Foresters and most financial types are well versed in DCF, what it is and why it is so important. Its use can be traced back to Martin Faustmann, a German forester, in an 1849 publication on valuing immature stands. By the 1930’s, the financial community had recognized its importance and began incorporating it into investment analysis.

Here is the definition of DCF according to Investopedia: A valuation method used to estimate the attractiveness of an investment opportunity. Discounted cash flow (DCF) analysis uses future free cash flow projections and discounts them (most often using the weighted average cost of capital) to arrive at a present value, which is used to evaluate the potential for investment. If the value arrived at through DCF analysis is higher than the current cost of the investment, the opportunity may be a good one.

Forget that comment about the “weighted average cost of capital” and we will talk about that in the next post. This post will touch on the following issues.

• Cash flows: Revenue from timber sales, leases, HBU sales, etc.; silvicultural expenses, taxes, management fees and other costs.
• Inflation and taxes
• Spreadsheet models, computer programs, harvest scheduling software
• Location: Important? Can it be quantified? Final sale price: When you sell the land, how much will you get? Existing cons easements
• Discount rate: very, very important and will be discussed in next post

Cash Flows: The costs and revenues, and when they occur, constitute the cash flows used in a DCF analysis. The entire process can be visualized utilizing a timeline such as the one below from a forestry investment example illustrated in the Forest Landowners Guide to the Federal Income Tax, Ag. Handbook No. 718.



The cash flows illustrated in this timeline are shown below.


This is a very simple example but it illustrates the process very well. Note that some of the flows occur in a particular year and some recur every year. Although the costs and revenues occur at different times, they are all discounted back to the present to determine the value today. There is a very thorough explanation of the process accessible by the link above. Note that in this example, the land purchase price is included at the start of the analysis and then the land is sold in the final year. An alternate approach is to create the cash flows in perpetuity (don’t include the land cost) and discount them back to the present. The result is the value of the land BUT only if you hold it forever any your cash flow calculations never change! If you plan to sell off some HBU land in the early years or if you see real increases in land value at the end of the investment, you MUST include land cost. If a conservation easement has been sold, the revenue should show up in the appropriate place AND the final sale price must be reduced by the appropriate amount.

What about inflation and what about taxes? Should they be incorporated into the cash flows? The answer is fairly clear, with respect to taxes, if the owner, such as a pension fund, does not pay taxes! For every investor though, there is a clear answer. Include them IF they are included in the discount rate. Here are a couple of quotes, also from the Forest Landowners Guide to the Federal Income Tax.

“Most forestry costs change at the rate of inflation in the economy; however, stumpage prices may increase (or decrease) at rates exceeding (or less than) inflation when supply/demand relationships change. These differential price trends can cause miscalculations in an investment analysis. Real (exceeding inflation) price appreciation—or price depreciation as the case may be—for some products, such as Southern pine and Douglas-fir sawtimber stumpage, has received much attention. But other product prices, such as those for pine and hardwood pulpwood, and equipment costs, also have been affected. Predicting the future always is uncertain and hazardous, so the best information available for projecting real changes in cash flows should be used.”

And the second quote: “it is imperative that the discount (interest) rate used for the analysis include a similar expectation factor for inflation. In summary, both elements of the analysis—cash flow and discount rate—must be kept in comparable terms (with or without inflation and before or after-tax) for reliable results.”

The take-away here is to include inflation in the cash flow if you think it is appropriate and to include taxes in the cash flow if you think that is appropriate, but you must incorporate those elements in the discount rate if they are included in the cash flow. My solution, when building a model, is always to create the discount rate as a variable. Then model the discount rate so the analysis can instantly be modified by setting the inflation and tax rate variables (either or all) to zero. If you don’t do this, I promise that someone will ask you “What if…”

Let me also add that there are some complexities associated with timber depletion that will create some pitfalls for the model builder not experienced with forestry taxation (depletion and multiple tax rates). The taxation issues are further complicated by the “inflation tax”. Maybe someday I will do a post on that too but not today. For now, let’s look at the types of software available to do DCF analysis of timberland investments.

Spreadsheets, forestry investment software programs or Harvest Scheduling software?

Spreadsheets: This is the “roll your own” option. It has some advantages but some very powerful disadvantages as well. The biggest advantage is that you can do it your way, you already have the software (and you can distribute the model to everyone in your organization) and you understand all of the drivers used. It is a good solution for simple situations like the one above. The disadvantage is that you really have to understand forest management, DCF, forest taxation, depletion accounting, G & Y models, spreadsheet development and probably a few things that I have left out. I have constructed complex models used for evaluating large transactions so it can be done. But the results fall short in many ways.

Forestry Investment Software Applications: These applications, frequently web based, are a good alternative to the use of spreadsheets if the target tract is not to large or complex. There are usually multiple applications in each suite which allow calculations designed to feed the DCF application. Growth and yield models may be built-in or a part of a supporting application. Applications that allow projections of both real and inflated timber prices (and land prices) are available to help feed the DCF analysis (remember to use a discount rate that is consistent with the inflation/real price). They have “help files” that explain the terminology and provide guidance in usage. They are generally easy to use and they are free.

Two examples can be found on web sites at Mississippi State and the Texas Forest Service. See FORest VALuation or FORVAL Online and the Timberland Decision Support System. You can actually use both of these sites to create your cash flows and then enter them into the Texas site.

The MSU site also has G&Y models for loblolly, slash and longleaf as well as some calculators to let you look at real increases in timber values (can also be used for real increases in land values as well). The Texas site also has several supporting “calculators” including The Timberland Management Simulator which integrates a loblolly G&Y model with the financial analysis.

There is other software available that you can find with a Google search. Some of the software is web-based and some requires downloads. I prefer the web-based stuff because it is more likely to work without problems. There are some very good G&Y models that have been “left-behind” because they no longer behave well (or at all) on the current operating systems. Another advantage of the web software, particularly over spreadsheets, is that they have well thought out explanations of the terminology and give you some good advice on how to conduct your analysis.

Harvest Scheduling or Forest Planning Software: What is Harvest Scheduling? First of all, the term is a misnomer. It is a carryover from earlier linear programming models that focused on the best time to harvest stands. Today’s applications, best referred to as Forest Planning applications, are far advance from just projecting harvests. All of the activities from planting through harvest are included while measuring the impact on attributes such as habitat, wood flows and cash flows are reported. The earlier applications have morphed into spatial applications importing GIS data and producing maps that show the “where” of all activities in the plan. The spatial applications also allow the use of spatial constraints, such as “green-up” requirements. The following paragraphs are from NCASI’s HABPLAN, spatial Forest Harvest and Habitat Scheduling software, users manual. I selected them to give you a better understanding of harvest scheduling and its complexities.

************************************************************
“Harvest scheduling entails the application of mathematical programming techniques to determine the allowable cut and/or the cutting budget, for a given area of forest, over multiple rotations or cutting cycles. With sustainability being a buzz word in the forestry industry, a number of harvest scheduling methods have been (and continue to be) developed that help us to manage our forests on a sustainable basis. The basic management unit is the forest stand (or a polygon comprising multiple stands). It is desirable that each management unit be managed in the most environmentally, economically and socially beneficial way. For each management unit, however, there are numerous management regime possibilities. The following are a few variables, which contribute to the wide range of potential management regimes.

• species
• site quality
• age of current stand
• length of rotation
• number of thinnings (& ages at which they occur), and intensity thereof
• regeneration or replanting
• greenup window

The potentially complex procedure of developing and solving a harvest scheduling model can be summarized in the following steps:

• Decide on decision-making variables. In Habplan, where integer programming is used, each decision-making variable represents one whole management unit (forest stand or polygon), i.e. each management unit can only be assigned one management regime. However, in linear programming, it is assumed that each management unit can conceptually split up, and managed under a number of different regimes, thus creating a number of different decision making variables for each management unit.
• Develop the objective function, according to the objectives of the given harvest scheduling problem.
• Incorporate various constraints e.g. land constraints, volume flow constraints, financial constraints and ending inventory constraints.
• Use a mathematical programming technique to solve the problem for the optimal/best solution.
• The solution to such a problem should offer information on which management units (or how much of each management unit) to devote to each of the proposed management regimes.

There is no one computer program in the world that can account for all variables in nature. Therefore, it is important to keep in mind that harvest scheduling is merely man's best effort at simplifying a very complex and dynamic natural phenomenon into a mathematical formula, and does by no means offer the perfect solution in the quest for the optimal management regime. However, it is safe to say that various harvest scheduling methods are capable of providing fairly reliable guidelines, by which land can be managed.”

**********************************************************
The above paragraphs illustrate the complexities and capabilities of harvest scheduling. In addition to Habplan, there are several other applications available. The hands-down favorite of forest planners is Remsoft’s Forest Planning System (Woodstock and Stanley) regarded as the best available by most users (at least managers of 500 million acres on five continents)!

It takes very well trained people to prepare a harvest schedule or plan. From personal experience, I can say that Remsoft also offers excellent training programs and support for its products. Some companies and TIMOs have trained people and do their own plans. Others use forest consultants that have experienced harvest scheduling people on their staff. One company, FORSight Resources, actually specializes in harvest scheduling and has several world-class authorities on the subject.

The take-home for the harvest scheduling discussion is that it is the best DCF methodology for a major acquisition because:

• It seeks optimization of management scenarios.
• The better versions allow spatial constraints and provide spatial output.
• It allows the addressing of very complex issues that other solutions are not capable of doing.
• Optimal solutions form the basis of the future management plan.
• There is some very good software available with tools built that reduce the complexity of the mathematical programming.
• But… it is complex, time consuming, and must be implemented by specialists

Summarizing:

• because of the timeline associated with the cash flows of all timberland investments, DCF is a requirement of all valuations (except the flip!)
• for large timberland transactions, use harvest scheduling software
• for small tracts (up to a couple of thousand acres) and infrequent purchases, use freely available forestry investment software
• for frequent purchases/valuation of smaller transactions, spreadsheets may be the better choice.

This concludes the DCF discussion. The next, and final, post in this series will focus on the choice of the discount rate and how critical that decision is to the outcome of the valuation. Comments welcomed. --Brian

Valuing Timberland II

This is the second post on timberland valuation. If you missed the first post, which included an Overview plus a discussion on Disaggregation, you can read “Valuing Timberland I” here. Today’s post will focus primarily on productivity and how that fits into today’s appraisal systems which use discounted cash flow techniques to determine the value of timberland.

Productivity is key to determining the value of timberland. Some of the productivity is inherent to the land itself and some is a function of the intensity and effectiveness of the silviculture practiced on the tract.

First, let’s look at the inherent productivity of the land. Foresters measure that with a metric referred to as Site Index. Site Index (SI) refers to how tall a given species of tree can grow on that particular site in a specific number of years. Examples might be Loblolly Pine, base age 25 years, SI 60. Or Red Oak, base age 50, SI 60. Or Cottonwood, base age 10, SI 60. All of these are realistic examples and I have worked in stands with these exact site indices. The sites are not similar though. Note that all of the examples are “Site 60” land meaning all of those species will grow to sixty feet in height but it takes the cottonwood only 10 years, the loblolly pine takes 25 years and the oak takes 50. The land supporting these three site indices would also look very different. So SI defines the productivity for a particular species on that site. As a matter of interest, the SI 60 examples given for Loblolly (25 years) and Red Oak (50 years) are reasonable and quite common. The SI 60 (10 years) for cottonwood is actually a very poor site. Cottonwood on a good site can attain 100 or more feet in 10 years! Neither loblolly pine nor red oak could even survive on these sites due to the prolonged flooding during the growing season. The site determines the best species to plant or to favor with natural regeneration.

Let’s look at site productivity in more detail. In the late 60’s and early 70’s I was buying timberland in North Mississippi and West Tennessee. My counterparts (in other regions) and I were tasked with developing a new appraisal system based upon minimizing the cost to the pulp mill rather than the minimizing the bare land value. A key component to valuation is the objective of the owner and I was working for a pulp and paper company that owned land for the purpose of supplying its mills. And remember this, timberland valuation and fair market value are two different things!! That’s why there is always a high bidder! Under this type of a valuation model (delivered cost), it should be very clear that the more productive the land (i.e. the higher the SI), the more wood that would be grown and the lower the delivered cost per ton would be. In addition, there was already a significant acreage of 10 to 15 year old established loblolly pine plantations on private land in the area that had been established by the U.S. Forest Service Yazoo – Little Tallahatchie flood prevention project. Cattle prices were high and the plantations were being cut, cleared and put into pasture. In general, fair market value for timberland in the area was essentially the pulpwood value of these plantations (very low because of their age) plus bare land value.

Our appraisal system team created the details for a model which estimated a future cost per ton delivered to the mill. Obviously, the tracts with the lowest cost per ton were the tracts most desirable to the company. Bare land value was still there and served as a checkpoint against “fair market value” (FMV). The new method was very different from the old in that it represented the beginnings of a cash flow model and timberland productivity became a very important driver. The “Growing Cost” of the model was actually a future value of the cash flows and was very dependent on the productivity of the tract. The impact of the increase in growth alone between SI 60 and SI 70 for loblolly pine is about 30%. Jack it up to SI 80 and the gain is almost 60%! It’s pretty easy to see the significance of productivity gains on valuations dependent on the amount of timber that will be grown in any given amount of time. The “take home” point of the last couple of paragraphs is that the greater the inherent productivity of the land, the more positive the impact will be on an appraisal system employing a discounted cash flow model. And the greater the price that informed buyers will pay for it. We’ll discuss more about the pine plantations mentioned above a bit later. But first we must talk a little more about productivity that we can impact and something called yield tables.

The second factor addressing forest productivity is silviculture. Silviculture can be briefly defined as - the art, science, and practice of controlling the establishment, composition, growth, and quality of forest stands. Silviculture is to the forester what agriculture is to the farmer. It is a broad term encompassing preparation of the site, selection of the species to be planted, selecting the seed with the best genetic capabilities, fertilization, planting, density control, weed control and harvesting. Silvicultural systems employing artificial regeneration, planting or direct seeding, are more in line with agriculture but natural stands employ similar practices to obtain the desired species composition and density control too. The decision to employ a silvicultural system with a focus on natural regeneration versus plantation management has a huge impact on cash flow and profitability (and therefore valuation). Natural forest management regimes have less cash out and less cash in. Which system, natural or plantation, is best depends on the objectives and circumstances of the owner. So…, site quality refers to the natural productivity of the site whereas silvicultural activities are things that we can do to increase the productivity of the forest. At a cost.

To a degree, the current productivity may be what nature provided us (really it is more what past “managers” left us with rather than what nature provided us with). The existing forest type (a classification of forestland based on the species forming a plurality of live tree stocking) may be a 20 year old oak-pine stand or a 10 year-old loblolly pine plantation. The two have very different productivity potentials going forward. There could easily be a five-fold increase in productivity between the two forest types on similar sites. In some cases, the forest manager has the option of changing forest types through the process of harvesting, site preparation and planting. Greater productivity, greater cost.

Inherent in the example above is greater control of stand density or stocking through planting. Control of stand density is critical to maximizing productivity. So what is stand density? I don’t want to get into a forest mensuration short course here so just think of stocking as the combination of the number and size of trees that will provide optimum growth for a particular site. Stocking charts frequently classify stands as “understocked”, “well stocked” or “overstocked”. A “yield table”, mentioned earlier, combines the stocking level with Site Index to forecast what the yield of forest products will be at some future point. This tool is critical to determining the future productivity of the forest and to the timing necessary for the discounted cash flow analysis. For the purpose of explanation and understanding, I have created the super simple yield table below.



A “real” yield table can go on for pages with many different site indices, ages, stocking levels and products but this one will illustrate the points that I am trying to make. Note how the yield (in this case cords/acre at age 25 but could be MBF at age 20) changes with both stocking (trees/acre at Age 1) and site index. Using these variables we can forecast both volumes and cash flow at the time of harvest. And what if we plant genetically improved stock that increases growth that is the equivalent of a 10 point jump in SI? Or phosphorous fertilization on a P deficient soil? Silvicultural activities have a major impact on the yield but “in the old days” we were pretty well restricted to the yield tables and they served us quite well. Today, however, they have generally been replaced with much more sophisticated growth models which are not only capable of predicting growth based on stocking and SI, but on all of the silvicultural practices that impact those metrics.

I feel that I have rambled a little too much in some places and not been clear enough in others so let me summarize the key points.

Forest productivity, both inherent (SI) and developed (silvicultural activities) is probably the most important driver in the cash flow analysis of timberland. Tools are available to measure and project the volumes into the future. Yield tables for volume projection have been replaced by growth models. You can download and learn to use a loblolly growth and yield model from Mississippi State here (CUTOVER LOBLOLLY GYM). It allows you to manipulate the SI and stand density and to see the future output. It also allows you to assign values by product which we will discuss next time.

Timberland valuation and fair market value are two different things! Remember above when I talked about buying 10 to 15 year old established pine plantations? The FMV (based on comparable sales) was about $125/acre in 1970. We began paying $150 - $175/ acre and easily acquired a significant acreage of these plantations. The value that we were assigning was the future value discounted back to the current age as opposed to FMV. Ten years later we were harvesting $1000/acre worth of chip-n-saw from these sites. We recognized that the FMV was well under the actual value of the plantations and were able to capitalize on that. Some of the early TIMOs were able to do the exact same thing with pre-merchantable plantations on a much larger scale. Some of the early TIMOs saw the value in HBU lands and disaggregation and recognized that the sum of the parts was worth more than the whole. These values are now well recognized in a competitive market and the returns, as expected, are dropping. The lesson is that there is a good return to be had if you can recognize a difference between the real value and the FMV.

Location can be very important or fairly minor. In the case of a forest products company valuing timberland, distance to the mill can be critical (as is the case with an institutional investor committed to a fiber supply agreement requiring delivery!). Some states have real estate tax structures which have a significant impact on cash flow evaluations. But frequently, location has minimum bearing on valuations for an investor.

So much for the value of productivity. The next valuation post will focus on estimates of timber volumes and values, how we get them, and how we forecast them for future year’s cash flows. We will also take a quick look at the confidence you should place in them. --Brian

Longleaf Pine Growth Model and Stand Simulator Available

A new longleaf pine growth model and stand simulator is now available from FORSight Resources. The following information is from their news release. For additional details, you can email: info@FORSightResources.com. --Brian
___________________________________________________

FORSight Resources releases FORSim – Longleaf Pine Growth Simulator (LPGS)

North Charleston, S.C., USA – November 16, 2007 – FORSight Resources, a leading provider of decision support services for natural resource management, announced today the release of FORSim – Longleaf Pine Growth Simulator (LPGS). LPGS is a versatile tool that provides biometricians and inventory foresters with the functionality of a longleaf pine growth engine in an easy-to-use, excel-based interface. The growth engine provides for alternative thinning treatments. LPGS provides a means for quickly analyzing and comparing stand-level treatments through graphical and tabular outputs. LPGS also calculates scores for assessing foraging habitat for the endangered red-cockaded woodpecker (RCW), providing foresters and wildlife managers a powerful tool for developing and assessing treatment regimes in support of RCW recovery. Overall, users will find this to be a valuable addition to FORSim product
suite.