Tuesday, January 20, 2009

Valuing Timberland III

This is the third 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. Today’s 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.

First, let’s look at timber volumes. What does the cruise say? What does the inventory say? What is the difference? How was the acreage calculated? Are volume estimates tied to a GIS? What are the sources of errors? Can volumes be reasonably audited? If the volume estimate isn’t “right”, the value certainly won’t be!

What is a “cruise”? A survey of forestland to locate timber and estimate its quantity by species, products, size, quality, or other characteristics; the estimate in such a survey. Several different sampling techniques can be used in a cruise. (source: Forestry Terms for Mississippi Landowners, Mississippi State University Extension).

What is an “inventory”? See Cruise (same source). H’mmm. I don’t want to aggravate my Mississippi State colleagues or to argue semantics but there is a huge difference between a cruise and an inventory or, more appropriately, a timber cruise and a forest inventory. To keep it simple, the cruise (as correctly defined above) collects data to estimate the timber volume, quantity, etc. that is there right now. The forest inventory collects much more information which can be used to project volumes and harvests (and other stuff too) into the future. A cruise, as defined above, wouldn’t collect any information on an eight year old loblolly pine plantation. There is little or no volume (perhaps a few trees that meet the limits of merchantability) but there certainly is value. A forest inventory would want to know the species, age, trees/acre, site index and acreage at a minimum. That is the information that is necessary to drive the growth and yield models necessary to estimate future wood flows from the tract. The wood flows become the basis for the cash flows in the appraisal. Perhaps this is just semantics but it is very important to understand the difference and to know what you are looking at.

So the “acquisition cruise” necessary to purchase land is more appropriately referred to as a forest inventory than a cruise. A “timber cruise” is more likely used to determine the immediate value of the timber for a timber sale. Data to drive growth models or cash flow models is not necessary. This distinction is more than academic. The inventory data can be “grown” to provide volume information for appraisals, stand level info for thinning or harvest, projected harvest and regeneration sites for planning site preparation and planting, all at multiple forward points in time. A typical cruise is not designed to do this. Merchantability specs are frequently different between cruises and inventories resulting in differences in volumes reported. Merchantability specs for inventories must be consistent from tract to tract, year to year, cruiser to cruiser and they must be consistent with what the growth models require. Merchantability specs for a timber sale cruise will normally be modified to conform to current demand, market pricing and perhaps even the logger that will be doing the harvesting.

The lesson here is that the user of cruise or inventory information had better understand what he or she is looking at. The cruise/inventory designer needs a clear understanding of the objective of the cruise. Otherwise, expect disappointment.

The sampling intensity of most inventories will not provide adequate data for timber sales. Inventories must collect some information on every stand or stratum (a collection of stands with similar characteristics). The objective of an inventory or cruise is to gather the information that is necessary at an acceptable level of accuracy at the least possible cost. Contrasting the two extremes, consider the case of a 100,000 acre land acquisition compared to a 50 acre mature hardwood timber sale. The objective of the land acquisition inventory is to be accurate enough to generate a reasonably accurate cash flow analysis for the perceived life of the investment (and hence, its present value). Another objective may be to assist in management if the tract is purchased. The resulting inventory would be one with a low sampling intensity but one collecting much information at each sample plot. The inventory is “not accurate” at the stand level. Cost is a serious consideration. If the tract is not purchased, the entire cost of the inventory is wasted. The 50 acre timber sale, on the other hand, has a much higher sampling intensity (perhaps a 100% tally), does not require collecting growth modeling information and requires careful inspection of the quality of each tree relative to current market conditions. The cruise cost/acre is much higher but the payback is immediate and ensured.

Statistics 101: Reiterating an earlier point when addressing timber volume reports; you had better understand what you are looking at. Most well done cruise reports will provide some detail with respect to the statistical accuracy provided by the data. You will see something like “2,216 MBF ± 10% of hardwood sawtimber”. This means, that from a statistical standpoint, the actual volume will be within 10% of the volume estimate 95% of the time, or maybe 90% of the time, or maybe 66% of the time. Maybe you better ask which probability level applies! You will probably also see a volume by species report and perhaps that will also provide statistical error reporting too. If it doesn’t, you should understand that because the sampling intensity for any given species is well below that of all of the species combined, the accuracy of the estimate drops pretty dramatically in a mixed species stand or forest. Now, digging into the cruise report a little deeper you will see that volumes may be reported by species and diameter class. How accurate do you think that will be? If the statistics are reported, that is great. If not, beware the value that you place in the numbers. This information can be useful but you have to assign it the appropriate credibility.

You should also understand that there are other sources of error in a cruise in addition to the statistical sampling error. Examples are the use of incorrect volume tables, measurement errors, poor sample design, calculation errors and, yes, a poor cruiser.

The statistical accuracy of large inventories that have been conducted over several different years and have been brought current using growth models is a different story. I’m not going to get into how to do this because I have no earthly idea how to do it! It can be done by a fairly limited group of forest statisticians, mensurationists and biometricians but it is clearly a difficult task. As a rule of thumb, it is safe to say that the accuracy is somewhat less than what would be calculated using the “ungrown” plots and declines as the number of “grown” years increases.

That’s enough said about cruise reports, inventories and the timber volumes that they report. Now let’s talk about volumes that what will be there tomorrow and the fundamentals of creating a future wood flow.

Future harvest volumes and timing must be done on a stand by stand basis utilizing stand volumes that have been “grown” using yield tables and growth and yield models. The starting base is the stand or strata level inventory and each stand or stratum is grown into the future. Some rule or methodology is then used to select the harvest date to determine the harvest volume for each future year. The result is a wood flow model that, when combined with values, becomes a key component of the cash flow model. More detail about this will be provided in the next post.

This is probably the appropriate place to discuss the integration of a GIS (Geographic Information System) with the timber inventory system when creating the information system used for timberland management. GIS based inventory systems are very common among managers of large tracts of timberland today and sellers normally provide output from these systems to potential buyers. Inventory or cruise volumes may say what is out there but the GIS says exactly WHERE it is on the surface of the earth. A quick check of the validity of the data provided by the seller can be made by auditing selected stands to compare inventory acreage, stocking, volumes, forest types, SI, etc. to the audited values. It doesn’t take much of this type of checking to determine how serious the forest managers have been about the quality and timeliness of their inventory efforts and hence the credence that can be placed in the data provided. The GIS data can be “read” into the GIS based harvest scheduling/ cash flow models and will eventually provide the basis for future management (activity schedules, maps and annual cash flow projections). These are the tools that become the foundation for communications and expectations between the new landowner and the timberland managers. The schedules tell the owner what to expect financially and form the basis of the management plan for the manager. A final point about the GIS based inventory system is that it provides the basis for a field audit in conjunction with the traditional financial audit. In the timberland sector, a financial audit without a corresponding field audit is a serious mistake; in fact, it's not really an audit.

Now let’s discuss timber values. They are down! A lot! If you are selling timber today, that is very important but how important is it with respect to timberland valuation? Should current market conditions be used for valuation?

The comments below are by Rick Holley (from Plum Creek 2008 Q3 Earnings Conference call) answering an analyst’s question about why he thinks timberland values are holding up while other asset classes are declining in value. The quote is a little messed up but his point is clear.

“I think largely because the investor in timberlands, and there’s still a fair amount of capital on the sidelines trying to invest in timberlands, is longer term. They’re looking through this cycle. If anything they’re starting to look at this cycle and lower prices as being more upside when you look out a year or two years or whatever these markets improve. I know there was a recent [Reese] article talking about some transactions that may or may not get done. We’re in the marketplace and we talk to a lot of buyers. We know who the sellers are, and we think all those transactions will in fact get done at very, very good prices, so I think the market is holding up and there’s still a lot of capital chasing very little opportunity…”

The key words above are “looking through…” with respect to timberland prices. To me, that implies that they are also “looking through” with respect to the timber values used in the timberland valuation. I think that if stumpage prices were at all-time highs, buyers would “look through” and see lower prices in the future as well. As Andy Malmquist, my good friend, former employee, and current TIMO guy, used to remind me about so many things… “It’s regressing toward the mean”! Andy actually talks like that.

There are at least three factors to consider for determining future timber values to apply to future volumes – current values, real price increases and inflation. We will discuss how to address inflation in the next post so for now, we will confine the discussion to the first two items.

I am defining “current price” as a “look through” price that is on the trend line for the species,local area and region (huge differences between NE and NW, cherry and red maple). In other words, I am looking for the price that has “regressed to the mean”. This data can be obtained from several different pricing services such as Forest2Market, RISI and Timber-Mart South. In addition, most state forestry commissions (South Carolina Forestry Commission pricing report) or forestry university extension services also provide some level of pricing reports. Local consulting foresters usually know what is available and can also provide first hand local knowledge. Pricing from these sources will provide the baseline for projecting the value of future harvests in the cash flow model.

A real increase in future stumpage prices can and should be incorporated into the model if you think it is justified. U. S. Forest Service reports suggest that sawtimber stumpage prices have increased at a real rate of 2% a year over a very long period of time. An analysis of southern pine stumpage prices by forest economist Jack Lutz reached the following conclusion:

“Our analysis indicates that southern pine sawtimber stumpage prices are mean-reverting, with a 50-year mean of $38.29/ton (based on LDAF data). Those prices have held to that mean through 50 years of timber supply and demand shocks and significant changes in timber harvesting and processing technology. That means we should not expect a significant increase in sawtimber stumpage prices in the near future. This supports the current practice of many timberland investors who are using 0% real appreciation rates in their timberland investment models.”

The future is anybodies guess but the real value increase used should be based upon your view of the future. That’s why there is always a high bidder and a low bidder!

This concludes the discussion on timber volumes and values. The next post in the series will attempt to tie it all together by addressing the methodologies associated with the discounted cash flow models. Comments are welcomed. --Brian

Friday, January 16, 2009

How Much Did Timberland Values Fall in 2008?

My stock portfolio has certainly taken a hit to the downside. The housing market has tanked and the values of many houses and rental units have gone down with it. Stocks of the publicly traded timberland owners have gone down 20 to 60% in the last year or so. The logic is that the decline in housing has caused a drop in lumber price which has created a drop in timber (stumpage) price which has resulted in a corresponding drop in timberland prices. So logically my portfolio has taken a hit on the tree farm side too. But has it?

I do some consulting with market analysts and hedge fund managers interested in valuing timberland and the drivers behind it. The most common question that I have heard recently surrounds the declining value of timberland and just how great that decline is. How can a company like Weyerhaeuser or Plum Creek be properly valued given the logical decline in timberland values? I could give my opinion (and I always do!) but that is not enough.

I have been maintaining a database of the major timberland transactions for over a decade so I wanted to quantify the change for 2008. Below is a chart showing the $/acre sale price trend for the last decade.


















This is pretty interesting. Nationwide, it appears that prices have dropped about 21% following a year where they gained 60%! BUT… it is important to understand the data and what is happening. This database is composed of transactions that total between one and seven million acres in any given year. Also in the database is a “REGIONS” field. Price distortion occurs between years due to the variation in the percentage of sales occurring between regions ($/acre varies significantly between regions). So how does the price per acre change if we just look at a single region? Let’s look at the South.



















Within any particular region, individual transactions will impact any given years weighted average price. In spite of that, the trend is clear and it is difficult to find any argument in the data that supports a decline in timberland pricing – at least in the South. The first reported sale for 2009, Potlatch to RMK, was at $1,745/acre, right in the ballpark.

The Northeast is the only region that did show a drop last year. The acreage sold in the NE was not particularly large (meaning most of the variability was probably due to the variability between tracts) so I would be reluctant to attribute much significance to the decline. There was one significant datapoint though. The Essex Timber to Plum Creek transaction (at $267/acre) pulled the average down. It is important to note that there was a conservation easement on this tract. We should expect to see significantly lower transaction values as more sales occur on tracts with existing conservation easements. The sale of a conservation easement provides revenue in the form of an early payment for HBU land but it can also have a negative impact on future forestry based revenues as well.

So…, what does all of this mean? I can see little or no decline in timberland values. I’ll be curious to see what the NCREIF index on timberland values has to say. The index is based more on appraisals than actual sales but the appraisals SHOULD be based on comparable sales and comparable sales just do not support a decline in appraisal values. If sale prices are not declining, what is wrong with the logic expressed in the first paragraph – declining housing starts means declining timberland values?

Here is my take. First, buyers are “looking through” current timber prices. Sophisticated timberland investors use appraisal techniques that look at cash flow from timber over the planned life of the investment. Those timber values are based on their view of the future, not just today’s market.

Second, more money is chasing fewer acres. A significant amount of timberland is being pulled from the market. Example: almost all of the 161,000 acres of Finch and Pruyn timberland in the Adirondacks will ultimately be withdrawn from production and become a part of the Forest Preserve. An example from the other side of the equation: this week the United Nations announced that its pension fund would diversify its portfolio and seek to acquire timberland. More money chasing fewer acres. Institutions want to diversify their portfolios, particularly by acquiring hard assets.

Third, future wood demand will be impacted by both renewable energy and global warming concerns. We are already seeing pellet mills being built to export pellets to Europe where they are mixed with coal to reduce the amount of carbon tax the utilities must pay. Most forecasters expect to see a similar tax in the U.S. soon.

Fourth, there is evidence that declining interest rates are impacting the discount rates used by institutional timberland purchasers. A declining discount rate drives value up!

These are my thoughts, these are my numbers, and these are my opinions. Comments are welcomed. --Brian