Timberland investments in the United States: A review and prospects

Forest Policy and Economics 109 (2019) 101998

Contents lists available at ScienceDirect

Forest Policy and Economics journal homepage: www.elsevier.com/locate/forpol

Review Article

Timberland investments in the United States: A review and prospects

T

Bin Mei Harley Langdale Jr. Center for Forest Business, Warnell School of Forestry and Natural Resources, University of Georgia, Athens, GA 30602, United States of America

A R T I C LE I N FO

A B S T R A C T

Keywords: Alternative investment Asset pricing Forestry Portfolio optimization Real estate

In this study, 68 peer-reviewed journal articles in timberland investments in the United States published after 1980 are reviewed. Prior to the synthesis, the history of modern timberland investments, investment vehicles and return indices are summarized. Then, the literature is categorized into four groups, i.e., role of timberland in a portfolio, pricing of timberland assets, public timber real estate investment trusts and other relevant topics, and discussed respectively. The analysis suggests that timberland is a risk diversifier in a portfolio whether standard deviation or value-at-risk is used as the risk metric, that classic asset pricing models have difficulty in pricing private-equity timberland assets resulting in significant alpha, that timber REITs have some risk-reduction ability but no excess returns, and that bioenergy market and contractual rights and obligations on the properties may affect cash rewards from timberland investments. At the end, some concluding remarks and potential future research issues are presented.

1. Introduction The United States is rich in forest resources, with one third of the county's land area, or 751 million acres, being forestlands.1 Among them, 514 million acres with an estimated market value of $460 billion are considered to be commercial timberlands, which are mainly used to produce timber (Newell and Eves, 2009; Smith et al., 2009). Of all timberlands, about 360 million acres are privately held (Mendell, 2015). The US is also one of the most productive countries in forest resources in the planet, providing 17%–28% of global wood products over 1998–2013 (Prestemon et al., 2015). The US South, in particular, holds about 41% of the US timberland and is the world's largest wood producing region (Brandeis and Hodges, 2015). Timber production of the 13 southern states accounts for around 60% of US total or 10% of the world's total (Brandeis and Hodges, 2015; Prestemon et al., 2015), and the total contribution of the forest sector to the South's gross regional product remains at about 2% in recent years (Brandeis and Hodges, 2015). About 90% of southern timberland is privately owned and pine plantations cover 33% of the total 205 million areas of timberland in the region (Hartsell and Conner, 2013). Regarding the output, pine (softwood) products represent more than 60% of the total harvest (Wear et al., 2007). In short, the forest sector plays a significant role in the US rural economy. In the early 20th century, most industrial timberlands in the US were owned by forest products companies. By vertical integration,

forest products companies were able to insure internal timber supply, alleviating their dependence on the open markets for raw materials. However, in the past few decades, forest products companies have been divesting their timberlands and outsourcing the business of growing and harvesting timber.2 Major reasons for this change are: 1) shift in production from diversification to specialization; 2) reduced insurance value of timberland due to expending and reliable timber supplies; 3) double taxation on forest products companies structured as “C-Corps”; 4) significant undervaluation of timberland under the US generally accepted accounting principles, where book values reflect historical acquisition costs rather than fair market values; and 5) rising demand from timberland investment management organizations (TIMOs) and real estate investment trusts (REITs) (Healey et al., 2005; Mei et al., 2009; Wear et al., 2007). Moreover, recent consolidation in the forest products industry has substantially increased acquirers' financial leverage. To reduce debt, forest products companies liquidated their timberland assets. Hence, the “invisible hand” has separated timberland management business into its own market. Research shows that timberland sales generated negative abnormal returns on both forest products companies as sellers and RETIs as buyers, whereas conversions to REITs had positive effects (Mendell et al., 2008; Piao et al., 2017; Sun et al., 2013). In most cases, timberland sales increased asset return volatility in the short run (Sun et al., 2013). To retain some control on timber supply after timberland divesture, however, many forest products manufacturers negotiate

E-mail address: [email protected]. 1 acre ≈ 0.4047 ha. 2 For details of those timberland transactions, refer to Hood et al. (2015). 1

https://doi.org/10.1016/j.forpol.2019.101998 Received 2 January 2019; Received in revised form 28 June 2019; Accepted 10 August 2019 1389-9341/ Published by Elsevier B.V.

Forest Policy and Economics 109 (2019) 101998

B. Mei

timberland assets. For instance, International Paper established its timberland subsidiary as a master limited partnerships (MLP), i.e., The IP Timberlands, Georgia-Pacific created letter stocks for its timberland assets,5 i.e., The Timber Co., and Plum Creek converted itself from a MLP to a REIT to retain institutional investors of The Timber Co. after it was sold to Plum Creek. In addition, securitizations of timberland assets lowered the entry barrier and enhanced liquidity by requiring only a modest allocation of capital. In 2015, Plum Creek partnered with some institutional investors and created a new investment vehicle, a longterm joint venture. The deal is known as “Twin Creeks”, where Plum Creek manages day-to-day timberland operations and Silver Creek Capital Management serves as fiduciary. The initial equity of $560 million consisted of 25% of Plum Creek's timberland contribution and 75% of investor's cash (to buy Plum Creek's timberlands). Twin Creeks started with 260,000 acres of southern timberland valued at $2150 per acre and was expected to grow to $1 billion in value through acquisitions. Although restructuring activities like MLPs and letter stocks were once favored by the financial markets, these kinds of timberland investment entities proved to be short-lived (Zinkhan et al., 1992). The fate of joint ventures remains to be seen. In contrast, TIMOs and REITs have gradually gained popularity in timberland investments, especially in large-scale transactions. TIMOs do not own timberlands but manage them on behalf of investors. In particular, a separate account is used for one major investor in a single portfolio, whereas a comingled fund allows multiple investors to participate in a relatively large, diversified portfolio of timberlands. Compared to public-equity timberland investment, private placement has higher requirements of minimum capital commitments, but less pressure from dividend payout. Yet securitization introduces systematic risk into timberland assets. Today, there are more than 30 TIMOs, four public timber REITs, and two non-REIT public timber companies (Table 1). Major TIMOs included Hancock Timber Resource Group, The Forestland Group, Campbell Global, Resource Management Service, Forest Investment Associates, BTG Pactual, Molpus Woodlands Group, and Timberland Investment Resources. The four public timber REITs are Weyerhaeuser, Rayonier, PotlatchDeltic, and CatchMark. The two non-REIT public timber companies are Pope Resources and Acadian Timber. Together, TIMOs and public timber companies accounted for about 50% of the investable timberland universe in the US, or about 12% of private US timberlands (Hood et al., 2015; Mendell, 2015). In recent years, the timberland market consolidated considerably. Hancock Timber Resource Group and Molpus Woodlands Group bought Forest Capital Partners in 2012, BTG Pactual acquired RMK Timberland in 2013, Wells Timberland REIT went public as CatchMark in 2014, Weyerhaeuser acquired Plum Creek in 2016, and Potlatch acquired Deltic Timber in 2018.6 Looking forward, consolidation waves in the timber industry will continue. Mergers and acquisitions are expected, joint ventures between public timber REITs and institutional investors may increase, and initial public offerings of TIMOs may be possible. To provide return information of public- and private-equity timberland investments, a number of indices exist. For private-equity timberland investment, the NCREIF Timberland Index (NTI) is widely acknowledged as the primary source for return information. For publicequity timberland investment, the return of a portfolio of public timber REITs or companies (PUBLIC) is often used. There are some fundamental differences in how these return indices are constructed and reported. First, public timber REITs use a considerable amount of debt. Because financial leverage changes the risk-return profile of equity, the

wood supply agreements with landowners. A number of studies have examined the economic values of these contacts using the option pricing framework (e.g., Mei and Clutter, 2013; Petrasek and Perez-Garcia, 2010; Shaffer, 1984; Yin and Izlar, 2001). The rights and obligations of each party, and the length of a contract vary case by case. However, if properly designed, both parties of a contract can benefit, leading to a win-win situation. From an investor's perspective, wood supply agreements can alter cash flows from timberland investments and thus affect the returns. Forest products companies being the primary sellers, institutional investors have been unprecedentedly active buyers of timberlands. According to a recent survey, institutional commitments to the US timberland assets have increased tenfold since the early 1990s, and the growth in such investments remains strong (Hood et al., 2015). One major driver in timberland investments dates back to the implementation of Employee Retirement Income Security Act in 1974, which required pension fund managers to diversify their investments to minimize the risk of large losses. As such, a wide range of alternative assets including timberland have been sought after (e.g., Binkley et al., 1996; Conroy and Miles, 1989). Meanwhile, some forest products companies chose to focus on timberland management and elected to be REITs for favorable tax treatment.3 As a result of the interaction between sellers and buyers, more than 40 million acres of timberland changed hands in the United States in the past two decades (Hood et al., 2015). Since the first foray into timberland, institutional investments have developed rapidly. At the end of the 1980s, timberland assets managed by TIMOs totaled merely $1 billion; in 2015, this number exceeded $40 billion (Hood et al., 2015; Zinkhan, 2008). In addition, public timber firms controlled another $35 billion of timberlands. Timberland being a unique asset class with biological growth being the dominant return driver (Mei et al., 2013),4 many studies have examined its financial characteristics from various angles. This article reviews the literature on timberland investments in the United States with an emphasis on journal papers published after 1980, corresponding to the time period with a substantial timberland ownership change. Literature search focuses on peer-reviewed articles published in English and is based on mainstream databases (e.g., AgEcon Search, EconLit, Google Scholar and Web of Science) with key words being forest (forestland) investment, timber (timberland) assets or economics, and timber (timberland) investment. That resulted in 178 catches, of which I reviewed 68 articles related to timberland investments in the United States, excluding non-industrial forest investments. The goal of this review is to summarize past research that has been done in timberland investments in the United States and offer some prospects for future endeavor in this field. While completeness is aimed at, some studies might have been overlooked and the review should not be considered exclusive or related to the significance of contribution by any means. Prior to categorizing the literature into subgroups, I discuss timberland investment vehicles next.

2. Timberland investment vehicles and return indices Prior to the 1980s, timberlands were mostly invested by farmers and large forest products companies, who were perceived not actively involved in timber management (Zinkhan and Cubbage, 2003). Since the early 1980s when institutional investors initially expressed their interests in timberlands, forest products companies have responded by restructuring their business segments and securitizing the illiquid 3 Being a REIT, it is required to maintain a dividend payout ratio of at least 90%, which essentially eliminates income tax at the corporate level. 4 The three return drivers of timberland investment are timber price change, land value change, and biological growth. Biological growth refers to both physical increase in total biomass and transition into higher value product classes (e.g., from pulpwood to sawtimber).

5 Letter stocks are not registered with the security and exchange commissions (SEC), and hence cannot be sold publicly in the marketplace. Letter stocks are sold directly by the issuer to the investor. 6 Plum Creek was a public timber REIT that controlled 6.6 million acres of timberlands, Deltic Timber was a C-Corp that controlled 0.54 million acres of timberland, prior to the merger.

2

Forest Policy and Economics 109 (2019) 101998

B. Mei

Table 1 Private- and public-equity timberland market. TIMOs (private equity)

Mil. acres

Public timber firms (public equity)

Mil. acres

Hancock Timber Resource Group The Forestland Group Campbell Global Resource Management Service Forest Investment Associates BTG Pactual Molpus Woodlands Group Timberland Investment Resources

3.9 3.4 2.7 2.6 2.4 1.7 1.6 0.8

Weyerhaeuser (REIT) Rayonier (REIT) PotlachDeltic (REIT) Acadian Timber CatchMark (REIT) Pope Resources

13.1 2.7 2.1 1.1⁎ 0.4 0.2

Note: Data are compiled from company websites or 10-K filings. ⁎ Including lands in Canada.

Scholtens and Spierdijk (2010) found that the efficiency improvement from adding private-equity timberland assets reduced significantly after the appraisal smoothing bias was corrected. Matthies et al. (2015) examined how price dynamics and the structures of economic returns for different payments for ecosystem services schemes affect diversification benefits. Specifically, they studied optimal allocations of all stand units and forest management regimes at the landscape level both including and excluding financial assets in the portfolio, and demonstrated that payments for ecosystem services contracts could provide diversification benefits to investors. Martinez-Oviedo and Medda (2017) found that allocations in timbreland and farmland in an oil-based sovereign wealth fund portfolio had a positive effect on the performance when replacing equity investments, enhanced risk-return profile of the portfolio, and provided hedging aginst oil risk. With respect to the M-V framework, variance or standard deviation of asset returns is used to measure the portfolio risk under a multivariate normal distributional assumption. However, returns of financial assets such as stocks and bonds are often not normally distributed (Sheikh and Qiao, 2009), causing variance or standard deviation invalid as a risk measure. For instance, negatively skewed return distribution implies that that the probability of negative returns is higher than positive ones, and fat-tailed return distribution implies that both extreme negative and positive returns occur more likely than those in a normal distribution. Moreover, investors are more concerned about potential losses from extreme events such as financial crises. Therefore, more attention should be paid to the downside risk in practice. Timberland returns also depart from normality to some degree (Mei, 2015a; Petrasek et al., 2011). Hence, alternative methods are needed to derive the efficient frontier. Caulfield and Meldahl (1994) used the downside risk semivariance to construct timberland portfolios and identified more efficient frontiers than the M-V approach. Hildebrandt and Knoke (2011) pointed out that downside risk models provided an important implication for risk management in forestry. Because value at risk (VaR) is easy to calculate and interpret, it has become a popular downside risk metric. Although VaR gives the maximum loss that will not be exceeded with a given probability over a period of time, it cannot measure risk that exceeds VaR. Thus, conditional VaR (CVaR), which is capable of measuring the loss greater than VaR by incorporating the left-tail distribution, is used instead (Hull, 2012). The mean-conditional value at risk (M-CVaR) optimization method minimizes CVaR for a given target return without assuming a multivariate normal distribution. It has been used to construct hedge fund portfolios with minimum tail risks (Giamouridis and Vrontos, 2007) and to demonstrate the extent to which the M-V approach underestimates the tail risks (Agarwal and Naik, 2004). Washburn et al. (2012) first introduced the M-CVaR approach to assess how timberland allocations performed in a hypothetical portfolio from the risk perspective. Wan et al. (2015) further formulated both static and dynamic optimizations to assess whether asset allocation is consistent. Risk decomposition was introduced to understand how portfolio risk was contributed by individual assets. First, the efficient frontier was

PUBLIC should not be compared with the NTI directly as the NTI is an unlevered index (NCREIF, 2018). Second, liquidity of the public- and private-equity timberland is quite different. Shares of stocks of public timber REITs can be bought and sold easily without any appreciable impact on the price, while private timber funds require a minimum commitment and spend more time executing transactions. To put it another way, the PUBLIC is a constant-liquidity index but the NTI is a variable-liquidity index. Third, public timber firms often have some non-timber business segments such as wood products manufacturing that provide substantial operating income, whereas TIMOs focus more on timberland investments. Thus, the two indices essentially reflect returns on two different business models rather than on the underlying timberland assets. Lastly, the PUBLIC is based on real transactions on security exchanges, while the NTI is based largely on periodic appraisals. Realizing these differences and to make return indices more comparable, Mei (2016) constructed a transaction-based timberland index using NCREIF property-level transaction data, and Mei (2015b) constructed a pure-play timberland index based on securitized timberland data. Mei (2017) further compared different timberland index construction methods and results, concluding that the NTI had higher mean and lower volatility compared with the transaction-based timberland index, that separate accounts outperformed comingled funds in the private timberland market, that the pure-play timberland index exhibited higher return and lower risk than the corresponding portfolio of public timber companies, and that abnormal performance of timberland assets became less significant after controlling for the appraisal smoothing or by using real transaction data. Based on these research findings, TIMO and REIT assets are differentiated in this review whenever possible as they present many different financial characteristics.

3. Role of timberland in a portfolio A number of studies have employed the modern portfolio theory to evaluate timberland's diversification potential, with Mills Jr. and Hoover (1982) and DeForest et al. (1991) being the pioneers. Using the mean-variance (M-V) optimization approach, Caulfield (1998a), Caulfield (1998b), and Zinkhan et al. (1992) demonstrated that adding timberland assets to a portfolio could improve investment performance. Thomson (1997) employed the multiperiod portfolio optimization approach and showed that timber had been valuable either as a single asset or as an addition to a financial portoflio. Newell and Eves (2009) analyzed timberland assets in real estate portoflios and concluded that timberland was a strongly performing asset over 1987–2007. Waggle and Johnson (2009) compared timberland, farmland, and commercial real estate in a mixed asset portfolio with stocks and bonds, and found that timberland entered all portfolios, whereas farmland only entered low risk portfolios. Overall, the appealing nature of timberland in a mixed asset portfolio can be owed to its low return correlations with other asset classes (Lutz, 2004, 2018). 3

Forest Policy and Economics 109 (2019) 101998

B. Mei

had a significant positive equilibrium relationship with lumber futures and building permits, and a significant negative relationship with capitalization rates. In the short run, unexpected shocks in the three variables caused a permanent change in timberland values. Considering multiple state variables in the economy, Sun and Zhang (2001) applied arbitrage pricing theory (APT) to forest-related assets and illustrated that APT was superior to CAPM in model fitting. Interestingly, they found that stumpage price did not resemble return generation process of timberland. This contradicting result to Liao et al. (2009) comes from different methodologies applied. Sun and Zhang (2001)’s finding is for their particular sample period from the crosssectional analysis, whereas Liao et al. (2009)’s finding is for the longrun equilibrium from time series analysis. Yao et al. (2014) assessed the financial performance of timberland investments in the United States by APT. Results showed that public-equity timberland assets had higher mean excess returns in general. Expected returns of timberland assets were found to be declining over time. Yao and Mei (2015) used intertemporal CAPM to assess the risk-return relationship between forestryrelated assets and innovations in state variables. Market excess returns and innovations in the size (small-minus-big) and value (high-minuslow) factors, interest rate, term spread, default spread and aggregate consumption explained about 80% of the variation in cross-sectional returns of 16 forestry-related assets. Beta loadings on innovations in the value factor, interest rate and term spread induced significant risk premiums, and should be priced to determine the expected returns of forestry-related assets. In general, average excess returns of forestryrelated assets decreased over time. Significant positive excess returns were obtained for private- and public-equity timberland assets over 1988–1999. Insignificant excess returns were obtained for forest products and timber products in the whole sample period of 1988–2011. Some latest studies have borrowed the ideas in information economics into pricing timberland assets. Although there are overabundant sources of information, only a small portion is consumed. Information of interest is usually buried in distracting alternatives. As a result, we live in a world with a wealth of information but a poverty of attention (Greenberger, 1971). In theory, a framework has been proposed in which limited attention can impact asset returns; in practice, however, direct proxies of investor attention are hard to find and a variety of indirect proxies such as extreme returns, trading volume, and news and headlines have been used (Barber and Odean, 2008; Hirshleifer and Teoh, 2003; Peng and Xiong, 2006; Sims, 2003). In the internet era, when there are more and more electronic information sources, human being's capability to acquire information has been improved, especially with the help of internet search engines. With about 66% internet search market share,7 Google provides aggregate information acquisition statistics, which could be the best available attention proxy. By and large, relevant internet search has been found to affect asset pricing and the real economy. This additional information can be factored into our decision-making. Da et al. (2011) put forward a new and direct measure of investor attention by using the Google search frequency, known as the search volume index, and found that variations in that index affected both short- and long-run stock prices. Gao and Mei (2013) used the search volume index of timberland-related terms to test investor attention on timberland asset pricing. Results revealed that investor attention to finished wood products market, climate change, emerging woody bioenergy market, and overall manufacturing industry had significant effects on the abnormal returns of private-equity timberland investment. Similarly, Yao et al. (2016) used the orthogonalized investor sentiment index formed by Baker and Wurgler (2006) to examine the relationship between investor sentiment and returns of private-equity timberland investment. Results showed that current investor sentiment was an important factor that determines the one-quarter future returns

constructed by the M-CVaR optimization approach, which minimizes the downside risk CVaR and accounts for the non-normality of asset returns. Results indicated that the M-CVaR approach led to more efficient frontiers than the M-V approach. Second, both static and dynamic optimizations showed that timberland assets maintained a significant allocation in the mixed asset portfolio. Third, three metrics of the portfolio risk were compared and standard deviation resulted in underestimation compared with VaR and CVaR. Finally, the portfolio risk was decomposed via back testing under four scenarios. Both large-cap stocks and small-cap stocks were risk intensifiers, whereas Treasury bonds and timberland were risk diversifiers. Zhang and Mei (2019) demonstrated that M-CVaR framework could more precisely estimate the downside risk and that farmland replaced timberland as risk increased along the efficient frontier. In summary, M-CVaR framework can better define the efficient frontier than the M-V framework. Regardless of the choice of the risk metric, timberland tends to be a risk diversifier in a mixed asset portfolio. 4. Pricing of timberland assets With the advance in asset pricing theories and models in finance, many studies have attempted to price timberland assets and form expected returns. In particular, the capital asset pricing model (CAPM) has received much attention in the forest finance literature (e.g., Conroy and Miles, 1989; Redmond and Cubbage, 1988; Wagner et al., 1995; Wagner and Rideout, 1991; Washburn and Binkley, 1990; Zinkhan, 1988). However, these early work suffers from a short time series of data or the oversimplification in using timber price to derive timberland investment returns (Washburn and Binkley, 1990). Cascio and Clutter (2008) applied CAPM to synthetic timberland return indices and calculated required returns to be 4.5%–6.3% in the US South. Mei and Clutter (2010) compared private- and public-equity timberland investment using the NTI and PUBLIC. Results from the CAPM and Fama-French Three-Factor model revealed that private-equity timberland exhibited a low systematic risk and a positive abnormal return, whereas public-equity timberland fared similarly as the market. Results from the nonparametric stochastic discount factor approach revealed that both private- and public-equity timberlands had high excess returns. One common issue of the application of CAPM as well as FamaFrench Three-Factor model to the NTI is the low R-squared. That is, a large portion of variation in private-equity timberland returns remains unexplained. Hence, alternative multi-factor, multi-period asset pricing models are needed. Timberland returns have been positively correlated with inflation (Lutz, 2017) and early research shows that forests in the US West and South are effective inflation hedges (Washburn and Binkley, 1993). Expressing all returns in real terms, the real CAPM has an inflation factor in addition to the market factor. Actual inflation can be further divided into expected and unexpected inflations, and the estimated coefficients can be tested for an asset's ability to hedge against inflations (Zhang et al., 2011). Wan et al. (2013) found that private- but not public-equity timberland was able to hedge against expected and unexpected inflations. Hedging effectiveness depended on the states of the economy. Private-equity timberland assets were effective in hedging inflation during booms and less effective during recessions. Investment horizon also played a significant role in inflation hedging. The longer the timberland investment horizon, the stronger and more consistent the hedging ability held. Liao et al. (2009) and Clements et al. (2011) demonstrated the impact of cointegration on expected private-equity timberland returns and portfolio optimization. Liao et al. (2009) found that timberland return was largely driven by stumpage price in the long-run equilibrium. Clements et al. (2011) examined the impacts of lumber futures, capitalization rates (risk premium) and anticipated construction on timberland value. They found that, in the long run, timberland values

7

4

Average from 2005 to 2011 according to comScore.com.

Forest Policy and Economics 109 (2019) 101998

B. Mei

REITs, timber REITs have some risk-reduction ability but no excess returns. Lumber futures price tends to be a lead variable of timber REIT returns.

of timberland and the predicting power persisted over the next 1–5 years. Both the short- and long-term analyses showed negative coefficients on investor sentiment, indicating that current increase in investor sentiment drove prices up and lowered future returns. In addition, significantly different variances and insignificantly different means of timberland investment returns were obtained between lowand high-sentiment periods, which further confirmed the ability of earning long-term stable returns by private-equity timberland investment. In summary, most classic asset pricing models in finance can price public timber REITs well and suggest that public timber REITs behave like large-cap stocks. Nevertheless, the same models have some difficulties in pricing private-equity timberland assets and suggest significant abnormal returns. The difference diminishes when unsmoothed NTI or transaction-based timberland index is used. Overall, multifactor, multi-period asset pricing models outperform CAPM in pricing timberland assets. State variables like the value factor, interest rate, and term spread help form expectations of timberland investment returns. In addition, indices that measure investor attention or sentiment have marginal prediction power after controlling for traditional asset pricing factors. Yet the persistency of these indices in making forecast remains to be seen in the future. The dynamics of timberland investment returns could be related to recent consolidation of the timber industry and improved efficiency of timberland market as more capital flows into this market.

6. Other relevant topics Timberland being a long-term asset, its management has various real options embedded. For example, when stumpage price is low, landowners can delay harvesting and simply store values on the stump with a minimum cost (e.g., property tax). Hildebrandt and Knoke (2011) analyzed different decision-making approaches in forest investments and concluded that long-term decisions should consider uncertainty but adequate financial valuation had not been sufficiently developed within forest science. Chaudhari et al. (2016) reviewed the literature on forest investment decisions using the real options analysis published between 1985 and 2014 and found an increasing trend of such publications. A few recent applications are summarized in the next paragraph. Wear et al. (2015) showed that genetically engineered freeze-tolerant Eucalyptus could replace planted pine by as much as 2.8 million acres, with actual adoption depending on uncertain future markets for cellulose, especially for bioenergy feedstock. Mei and Clutter (2015) used a hypothetic fully regulated forest to evaluate timberland investment opportunities under price uncertainty in the US South. They found stumpage prices was close to the temporary suspension threshold, implying reduced harvest activities. Mei et al. (2019) combined financial risk with biophysical risk into entry and exit decision-making of timberland investment via an uneven-aged pine forest in the US South. Results showed that, despite a slight downward drift in price, expected returns for the pine forest fell between entry and exit thresholds, indicating an optimal “hold” strategy. This was explained by an offsetting upward drift in biophysical productivity associated with climate changes across a range of modeled futures. Monte Carlo simulation showed a small positive difference between entry and exit outcomes consistent with observed rates of expansion in timberland investments in the region. Reducing carbon emission being a global effort, many countries are seeking renewable feedstock to replace coal and natural gas for electricity generation. With government subsidies, Europe has consumed around 20 million tons of wood pellets in recent years, about a half of which has been covered by imports (Thran et al., 2017). The United States, in contrast, does not have the same mandates as in Europe but cofiring wood pellets with coal could be a solution. Along this line, Xian et al. (2015) and Mei and Wetzstein (2017) used the real options approach to examine the economic feasibility of low percentage cofiring wood pellets with coal for a power plant under cost uncertainty. Both studies concluded that cofiring was not a viable option without subsidies or carbon taxes. Chudy et al. (2019) showed that current market pricing for forest biomass in the western US generated negative returns from dedicated woody biomass plantations for energy purpose. However, should policies change in the US to trigger the adoption of woody biomass as a renewable feedstock for electricity, expanded demand for biomass energy would increase timber prices and harvests, resulting in new investments in forest stocks (Daigneault et al., 2012). In summary, uncertainty should be incorporated into optimal decisions given the long-term nature of timberland investment and management. Along the supply chain, the emerging woody bioenergy market and existing contractual rights and obligations on the properties may affect realized returns from timberland investments.

5. Public timber REITs Much research has been done on public timber REITs thanks to the abundantly available financial data. Newell and Hsu Wen (2006) categorized timber REITs as specialty REITs and confirmed their risk-reduction and portfolio diversification benefits. Regarding the structural changes in the timber sector, Mendell et al. (2008) and Piao et al. (2017) examined respective short- and long-term market responses to REIT conversion announcements and found significant abnormal returns. Sun (2013b) assessed the joint distribution between daily returns of timber REITs and two market indices and asserted that timber REITs had smaller volatility of tail dependence after the conversion. Sun et al. (2013) found increased volatility coupled with significant abnormal returns related to the conversion. In addition, Sun (2013c) used GARCH and extreme value models to examine price variation and volume dynamics of timber MLPs and REITs and found a positive return-volume relation. Compared to MLPs, REITs exhibited larger total shares and daily turnover rates. Sun (2013a) estimated the 1% level daily VaR of timber REITs to be 13% for the 2008–2009 recession period. La and Mei (2015) investigated the diversification ability of securitized timberland via the cointegration analysis. They found no general trends within timber REITs or between timber REITs and S&P 500. Therefore, there was long-run diversification potential with each timber REIT being a unique candidate. Piao et al. (2016) compared the financial performance of timber REITs, other specialized REITs, and common REITs. They found that timber REITs had larger market capitalizations, smaller unconditional variance, and no excess returns, and that timber REITs were insensitive to recessionary shocks but more vulnerable to idiosyncratic shocks. Lumber being derived from timber, a relationship between lumber prices and timberland prices is expected. Prices of timber REITs are intrinsically determined by the value of timber grown and harvested. Therefore, expected lumber prices should manifest the price discovery of timber REITs. Using vector error correction model and GARCH model, Clements et al. (2017) found a positive long-run equilibrium relationship between lumber futures, capitalization rates, publicly traded real estate equity, and timber REITs. Short-run results generally corroborated those of the long-run. In summary, the REIT structure for timberland management has received a positive response from the market. Among all specialized

7. Concluding remarks Timberland investments in the United States have been unprecedentedly active since the 1980s, when timberland ownership started shifting from industrial firms to institutional investors (via TIMOs) and REITs. Not echoing this trend, top 100 forest products 5

Forest Policy and Economics 109 (2019) 101998

B. Mei

companies in the world have increased timberland holdings between 2007 and 2012 in order to reduce risks associated with the flow of raw materials (Korhonen et al., 2016). An example is IKEA's recent acquisition of US timberlands as part of a broad strategy to invest in the sustainable production of resources that IKEA consumes. The increased vertical integration in the global forest products industry presents both opportunities and challenges for timberland investors in the United States. Forest products companies usually have different valuation models than TIMOs or REITs in that the inclusion of the insurance value of internal wood supply often adds a premium to timberland's intrinsic value. Therefore, this provides existing landowners an incentive to exit and secure a portion of the premium. For new investors, nonetheless, this trend of vertical integration implies a higher demand of timberland and eventually a higher acquisition cost, which hurts expected future returns. In addition to vertical integration, there has been a trend toward timberland investments for climate change mitigation purposes (e.g., carbon offsets) at the global level. In theory, this has similar implications to timberland investors as those from vertical integration; in practice, however, the impact differs substantially by region. In the United States, attitudes toward climate change mitigation are significantly influenced by the emergence of climate change legislation at the federal level (Thompson and Hansen, 2012). In other words, most organizations are taking a passive approach toward climate change mitigation projects in the United States and the development of this market will be an iterative process that requires global collaborations given the integrated nature of climate change mitigation (Thompson and Hansen, 2012). For timber production itself, returns and costs for major timber plantation species as well as other institutional, forestry, and policy factors that affect investments for a set of countries are compared and reported in Cubbage et al. (2010, 2014). Although timberland investments in the United States have attracted many research interests in recent years, many facets of timberland, a special type of real estate, have not been fully explored yet. For instance, public REITs may help price formation of TIMOs and information transition dynamics between those two markets represents a testable hypothesis. Another example is the linkage between lumber prices and TIMO returns. In the US South, lumber prices have been high in the past few years with the recovery of the housing market, whereas sawtimber prices have been stagnating due to the supply overhang. The implied higher profit margin of wood products manufacturing has triggered many greenfield as well as brownfield investments in this region. However, booming manufacturing has not led to higher returns of TIMOs. The exact causes need to be examined in future research but I suspect that they might be related to wood supply agreements attached to TIMO properties that are in favor of wood users and potential oligopsony in the wood products manufacturing sector (e.g., Mei and Sun, 2008; Silva et al., 2019). Finally, with the rapid progress in real estate finance in the academia and timberland investments in practice, I envision a prolific period of research in timberland economics in the future. In addition to traditional forest economics journals, I expect more agricultural economics or real estate economics journals to be the outlets, especially for studies that compare timberland with other types of real estate.

Barber, B.M., Odean, T., 2008. All that glitters: the effect of attention and news on the buying behavior of individual and institutional investors. Rev. Financ. Stud. 21, 785–818. Binkley, C.S., Raper, C.F., Washburn, C.L., 1996. Institutional ownership of US timberland: history, rationale, and implications for forest management. J. For. 94, 21–28. Brandeis, C., Hodges, D.G., 2015. Forest sector and primary forest products industry contributions to the economies of the southern states: 2011 update. J. For. 113, 205–209. Cascio, A.J., Clutter, M.L., 2008. Risk and required return assessments of equity timberland investments in the United States. For. Prod. J. 58, 61–70. Caulfield, J.P., 1998a. A fund-based timberland investment performance measure and implications for asset allocation. South. J. Appl. For. 22, 143–147. Caulfield, J.P., 1998b. Timberland in institutional portfolios and the question of persistence. For. Prod. J. 48, 23–28. Caulfield, P.J., Meldahl, R., 1994. Timberland Investments Return Volatility and Implications for Portfolio Construction. Proceeding of 24th Annual Southern Forest Economics Workshop, New Orleans, LA. pp. 144–162. Chaudhari, U.K., Kane, M.B., Wetzstein, M.E., 2016. The key literature of, and trends in, forestry investment decisions using real options analysis. Int. For. Rev. 18, 146–179. Chudy, R.P., Busby, G.M., Binkley, C.S., Stanton, B.J., 2019. The economics of dedicated hybrid poplar biomass plantations in the western U.S. Biomass Bioenergy 124, 114–124. Clements, S., Ziobrowski, A.J., Holder, M., 2011. Lumber futures and timberland investment. J. Real Estate Res. 33, 49–71. Clements, S., Tidwell, A., Jin, C., 2017. Futures markets and real estate public equity: connectivity of lumber futures and timber REITs. J. For. Econ. 28, 70–79. Conroy, R., Miles, M., 1989. Commercial forestland in the pension portfolio: the biological beta. Financ. Anal. J. 45, 46–54. Cubbage, F., Koesbandana, S., Mac Donagh, P., Rubilar, R., Balmelli, G., Olmos, V.M., De La Torre, R., Murara, M., Hoeflich, V.A., Kotze, H., Gonzalez, R., Carrero, O., Frey, G., Adams, T., Turner, J., Lord, R., Huang, J., MacIntyre, C., McGinley, K., Abt, R., Phillips, R., 2010. Global timber investments, wood costs, regulation, and risk. Biomass Bioenergy 34, 1667–1678. Cubbage, F., Mac Donagh, P., Balmelli, G., Olmos, V.M., Bussoni, A., Rubilar, R., De La Torre, R., Lord, R., Huang, J., Hoeflich, V.A., Murara, M., Kanieski, B., Hall, P., Yao, R., Adams, P., Kotze, H., Monges, E., Perez, C.H., Wikle, J., Abt, R., Gonzalez, R., Carrero, O., 2014. Global timber investments and trends, 2005-2011. N. Z. J. Forest. Sci. 44, 12. Da, Z., Engelberg, J., Gao, P., 2011. In search of attention. J. Financ. 66, 1461–1499. Daigneault, A., Sohngen, B., Sedjo, R., 2012. Economic approach to assess the forest carbon implications of biomass energy. Environ. Sci. Technol. 46, 5664–5671. DeForest, C.E., Cubbage, F.W., Redmond, C., Harris, T.G., 1991. Hedging with trees: timber assets and portfolio performance. For. Prod. J. 41, 23–30. Gao, L., Mei, B., 2013. Investor attention and abnormal performance of timberland investments in the United States. For. Policy Econ. 28, 60–65. Giamouridis, D., Vrontos, I.D., 2007. Hedge fund portfolio construction: a comparison of static and dynamic approaches. J. Bank. Financ. 31, 199–217. Greenberger, M., 1971. Computers, Communication and the Public Interest. The Johns Hopkins Press, Baltimore, MD. Hartsell, A.J., Conner, R.C., 2013. Forest Area and Conditions: A 2010 Update of Chapter 16 of the Southern Forest Resource Assessment, Asheville, NC. pp. 32. Healey, T., Corriero, T., Rosenov, R., 2005. Timber as an institutional investment. J. Altern. Invest. 8, 60–74. Hildebrandt, P., Knoke, T., 2011. Investment decisions under uncertainty—a methodological review on forest science studies. For. Policy Econ. 13, 1–15. Hirshleifer, D., Teoh, S.H., 2003. Limited attention, information disclosure, and financial reporting. J. Account. Econ. 36, 337–386. Hood, H., Harris, T., Siry, J., Baldwin, S., Smith, J., 2015. US Timberland Markets: Transactions, Values & Market Research 2000 to Mid-2015 Timber Mart-South, Athens, GA. Hull, J., 2012. Options, Futures, and Other Derivatives. Eighth edition. Prentice Hall, Boston. Korhonen, J., Zhang, Y., Toppinen, A., 2016. Examining timberland ownership and control strategies in the global forest sector. For. Policy Econ. 70, 39–46. La, L., Mei, B., 2015. Portfolio diversification through timber real estate investment trusts: a cointegration analysis. For. Policy Econ. 50, 269–274. Liao, X., Zhang, Y., Sun, C., 2009. Investments in timberland and softwood timber as parts of portfolio selection in the United States: a cointegration analysis and capital asset pricing model. For. Sci. 55, 471–479. Lutz, J., 2004. The anti-correlation heresy: timberland is negatively correlated with stocks—or is it? Forest Res. Notes 1, 1–4. Lutz, J., 2017. Inflation and timberland returns – update. Forest Res. Notes 14, 1–4. Lutz, J., 2018. Correlation update – the heresy continues. Forest Res. Notes 15, 1–4. Martinez-Oviedo, R., Medda, F., 2017. Assessing the effects of adding timberland and farmland into resource-based Sovereign Wealth Fund portfolios. J. Econ. Bus. 91, 24–40. Matthies, B.D., Kalliokoski, T., Ekholm, T., Hoen, H.F., Valsta, L.T., 2015. Risk, reward, and payments for ecosystem services: a portfolio approach to ecosystem services and forestland investment. Ecosyst. Ser. 16, 1–12. Mei, B., 2015a. Illiquidity and risk of commercial timberland assets in the United States. J. For. Econ. 21, 67–78. Mei, B., 2015b. A pure-play timberland return index based on securitezed timber firms. J. Real Estate Port. Manag. 21, 61–75. Mei, B., 2016. Transaction-based timberland investment returns. Land Econ. 92, 187–201. Mei, B., 2017. Investment returns of US commercial timberland: insights into index construction methods and results. Can. J. For. Res. 47, 226–233.

Acknowledgements United States Department of Agriculture Grant GEOZ-0199-MS funded this research. References Agarwal, V., Naik, N.Y., 2004. Risks and portfolio decisions involving hedge funds. Rev. Financ. Stud. 17, 63–98. Baker, M., Wurgler, J., 2006. Investor sentiment and the cross-section of stock returns. J. Financ. 61, 1645–1680.

6

Forest Policy and Economics 109 (2019) 101998

B. Mei

Sun, C., Zhang, D., 2001. Assessing the financial performance of forestry-related investment vehicles: capital asset pricing model vs arbitrage pricing theory. Am. J. Agric. Econ. 83, 617–628. Sun, C., Rahman, M.M., Munn, I.A., 2013. Adjustment of stock prices and volatility to changes in industrial timberland ownership. For. Policy Econ. 26, 91–101. Thompson, D.W., Hansen, E.N., 2012. Institutional pressures and an evolving forest carbon market. Bus. Strateg. Environ. 21, 351–369. Thomson, T.A., 1997. Long-term portfolio returns from timber and financial assets. J. Real Estate Port. Manag. 3, 57–73. Thran, D., Peetz, D., Schaubach, K., 2017. Global Wood Pellet Industry and Trade Study 2017. IEA Bioenergy. Waggle, D., Johnson, D.T., 2009. An analysis of the impact of timberland, farmland and commercial real estate in the asset allocation decisions of institutional investors. Rev. Financ. Econ. 18, 90–96. Wagner, J.E., Rideout, D.B., 1991. Evaluating forest management investments: the capital asset pricing model and the income growth model. For. Sci. 37, 1591–1604. Wagner, J., Cubbage, F.W., Redmond, C., 1995. Comparing the capital asset pricing model and capital budgeting techniques to analyze timber investments. For. Prod. J. 45, 69–77. Wan, Y., Mei, B., Clutter, M.L., Siry, J.P., 2013. Assessing the inflation hedging ability of timberland assets in the United States. For. Sci. 59, 93–104. Wan, Y., Clutter, M.L., Mei, B., Siry, J.P., 2015. Assessing the role of U.S. timberland assets in a mixed portfolio under the mean-conditional value at risk framework. For. Policy Econ. 50, 118–126. Washburn, C.L., Binkley, C.S., 1990. On the use of period-average stumpage prices to estimate forest asset pricing models. Land Econ. 66, 379–393. Washburn, C.L., Binkley, C.S., 1993. Do forest assets hedge inflation? Land Econ. 69, 215–224. Washburn, C., Keefer, B., Aronow, M.E., Petrasek, S., 2012. Timberland Portfolio Allocation and the Mean-Conditional Value at Risk. Hancock Timberland Investor Q1, Boston, MA, pp. 6. Wear, D.N., Carter, D.R., Prestemon, J., 2007. The U.S. South’s Timber Sector in 2005: A Prospective Analysis of Recent Change, Asheville, NC. pp. 30. Wear, D.N., Dixon, E., Abt, R.C., Singh, N., 2015. Projecting potential adoption of genetically engineered freeze-tolerant Eucalyptus in the United States. For. Sci. 61, 466–480. Xian, H., Colson, G., Mei, B., Wetzstein, M.E., 2015. Co-firing coal with wood pellets for U.S. electricity generation: a real options analysis. Energy Policy 81, 106–116. Yao, W., Mei, B., 2015. Assessing forestry-related assets with the intertemporal capital asset pricing model. For. Policy Econ. 50, 192–199. Yao, W., Mei, B., Clutter, M.L., 2014. Pricing timberland assets in the United States by the arbitrage pricing theory. For. Sci. 60, 943–952. Yao, W., Cheng, B., Mei, B., 2016. Investor sentiment and timberland investment returns. For. Prod. J. 66, 147–154. Yin, R., Izlar, B., 2001. Supply contract and portfolio insurance: applying financial engineering to institutional timberland investment. J. For. 99, 39–44. Zhang, W., Mei, B., 2019. Assessing the risk and return of optimal portfolios of US timberland and farmland. J. Real Estate Port. Manag. 25, 99–113. Zhang, M., Mei, B., Harris, T.G., Siry, J.P., Clutter, M.L., Baldwin, S.S., 2011. Can timber hedge against inflation? An analysis of timber prices in the US south. For. Prod. J. 61, 276–282. Zinkhan, F.C., 1988. Forestry projects, modern portfolio theory, and discount rate selection. South. J. Appl. For. 12, 132–135. Zinkhan, F.C., 2008. History of modern timberland investments. In: International Forest Investment: An Executive Education Conference. Center for Forest Business, Athens, GA, Munich, Germany, pp. 1–19. Zinkhan, F.C., Cubbage, F.W., 2003. Financial analysis of timber investments. In: Sills, E.O., Abt, K.L. (Eds.), Forests in a Market Economy. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 77–95. Zinkhan, F.C., Sizemore, W.R., Mason, G.H., Ebner, T.J., 1992. Timberland Investments: A Portfolio Perspective. Timber Press, Portland, OR.

Mei, B., Clutter, M.L., 2010. Evaluating the financial performance of timberland investments in the United States. For. Sci. 56, 421–428. Mei, B., Clutter, M.L., 2013. Valuing a timber harvest contract as a high-dimensional American call option via least-squares Monte Carlo simulation. Natl. Res. Model. 26, 111–129. Mei, B., Clutter, M.L., 2015. Evaluating timberland investment opportunities in the United States: a real options analysis. For. Sci. 61, 328–335. Mei, B., Sun, C., 2008. Assessing time-varying oligopoly and oligopsony power in the U.S. paper industry. J. Agric. Appl. Econ. 40, 927–939. Mei, B., Wetzstein, M., 2017. Burning wood pellets for US electricity generation? A regime switching analysis. Energy Econ. 65, 434–441. Mei, B., Cascio, A.J., Sun, C., Clutter, M.L., 2009. Mergers and acquisitions in the U.S. forest products industry: motives, financing, and impacts. In: Tobin, J.B., Parker, L.R. (Eds.), Joint Ventures, Mergers and Acquisitions, and Capital Flow. Nova Science Publishers, Inc, Hauppauge, NY, pp. 143–168. Mei, B., Clutter, M.L., Harris, T.G., 2013. Timberland return drivers and timberland returns and risks: a simulation approach. South. J. Appl. For. 37, 18–25. Mei, B., Wear, D.N., Henderson, J.D., 2019. Timberland investment under both financial and biophysical risk. Land Econ. 95, 279–291. Mendell, B., 2015. Splashing in the Bathtub: Putting the Proposed Merger of Plum Creek and Weyerhaeuser into Context. Forisk Consulting. Mendell, C.B., Mishra, N., Sydor, T., 2008. Investor responses to timberlands structured as real estate investment trusts. J. For. 106, 277–280. Mills Jr., W.L., Hoover, W.L., 1982. Investment in forest land: aspects of risk and diversification. Land Econ. 58, 33–51. NCREIF, 2018. National Council of Real Estate Investment Fiduciaries. Newell, G., Eves, C., 2009. The role of U.S. timberland in real estate portfolios. J. Real Estate Port. Manag. 15, 95–106. Newell, G., Hsu Wen, P., 2006. The role of non-traditional real estate sectors in REIT portfolios. J. Real Estate Port. Manag. 12, 155–166. Peng, L., Xiong, W., 2006. Investor attention, overconfidence and category learning. J. Financ. Econ. 80, 563–602. Petrasek, S., Perez-Garcia, J.M., 2010. Valuation of timber harvest contracts as American call options with modified least-squares Monte Carlo algorithm. For. Sci. 56, 494–504. Petrasek, S., Keefer, B., Aronow, M.E., Washburn, C., 2011. Statistical Distributions of Timberland Returns. Hancock Timber Resource Group, Boston, MA, pp. 29. Piao, X., Mei, B., Xue, Y., 2016. Comparing the financial performance of timber REITs and other REITs. For. Policy Econ. 72, 115–121. Piao, X., Mei, B., Zhang, W., 2017. Long-term event study of timber real estate investment trust conversions. For. Policy Econ. 78, 1–9. Prestemon, J.P., Wear, D.N., Foster, M.O., 2015. The Global Position of the U S Forest Products Industry, Asheville, NC. pp. 24. Redmond, C.H., Cubbage, F.W., 1988. Portfolio risk and returns from timber asset investments. Land Econ. 64, 325–337. Scholtens, B., Spierdijk, L., 2010. Does money grow on trees? The diversification properties of U.S. timberland investments. Land Econ. 86, 514–529. Shaffer, R.M., 1984. Valuation of certain long-term timber cutting contracts. For. Sci. 30, 774–787. Sheikh, A.Z., Qiao, H., 2009. Non-normality of market returns: a framework for asset allocation decision making. J. Altern. Invest. 12, 8–35. Silva, B.K., Cubbage, F.W., Gonzalez, R., Abt, R.C., 2019. Assessing market power in the U.S. pulp and paper industry. For. Policy Econ. 102, 138–150. Sims, C.A., 2003. Implications of rational inattention. J. Monet. Econ. 50, 665–690. Smith, W.B., Miles, P.D., Perry, C.H., Pugh, S.A., 2009. Forest Resources of the United States 2007: A Technical Document Supporting the Forest Service 2010 RPA Assessment. US For. Serv, Washington, DC, pp. 349. Sun, C., 2013a. Extreme risk of public timber REITs during the global financial crisis. J. Real Estate Port. Manag. 19, 73–88. Sun, C., 2013b. On the market risk of securitized timberlands. J. For. Econ. 19, 110–127. Sun, C., 2013c. Price variation and volume dynamics of securitized timberlands. For. Policy Econ. 27, 44–53.

7