NONINDUSTRIAL PRIVATE FOREST OWNERS IN MICHIGAN:
EXAMINATION AT A FINER SCALE
By
Lee S. Mueller

A THESIS
Submitted to
Michigan State University
in partial fulfillment for the requirements
for the degree of
MASTER OF SCIENCE
Forestry
2011

ABSTRACT
NONINDUSTRIAL PRIVATE FOREST OWNERS IN MICHIGAN:
EXAMINATION AT A FINER SCALE
By
Lee S. Mueller
Non-Industrial Private Forest owners are a diverse group of Michigan’s landowners. This study
seeks to manage NIPF variation through the creation of groups which characterize landowner behavior
and attitudes. This thesis is divided into two papers each examining a different method of grouping
Michigan’s NIPF landowners. Both papers utilize information from a 2003 survey of 1600 Michigan
NIPF landowners.
The first paper utilizes Michigan’s four regions: the Southern Lower Peninsula (SLP), Northern
Lower Peninsula (NLP), Eastern Upper Peninsula (EUP), and the Western Upper Peninsula (WUP), to
place landowners into sub-categories based on location. Distinct differences are found across regions in
regards to landowner demographics, parcel characteristics, reasons for owning forestland, and
management behavior. These differences suggest the need to take into account the effects of variation
of policy success and informational outreach programs.
The second paper utilizes a market-segmentation approach to place landowners in segments
based on their reasons for owning forestland. Three segments are developed: Game Wardens, Tenants,
and Timber Barons. Each segment differs in relation to demographics, parcel characteristics, reasons
for owning forestland, management activities, and preferences for policies or programs. These results
suggest the ability to target specific landowner segments to achieve policy goals. Further, policymakers and outreach extension specialists may wish to account for landowner differences when seeking
to develop policies or programs with specific objectives or targets.

Copyright By
LEE MUELLER
2011

ACKNOWLEDGEMENTS

This text represents a body of hard work supported by many individuals. First, I would like to
thank my advisor, Dr. Karen Potter-Witter, for her unwavering support and thoughtful guidance.
Additionally, the consistent participation of my committee members, Dr. Andrew Finley and Dr. Pat
Norris helped to retain direction and purpose in this work.
I would specifically like to thank Shivan G.C. for her help in compiling resources and
information. Without her patience and continued efforts, this may never have been completed.
Additionally, Brett Kuipers helped in providing information relating to Michigan wood-energy industry
and mill output levels. I thank my parents, Carol and Ric Mueller for their constant support and for all
things that only family do well. My mom, specifically, spent many hours pouring over grammatical and
punctuation errors. Without her hapless and tireless work, this document would be far more difficult to
read.
My sanity was protected by an assortment of the following individuals, who provided insight,
advice, or conversation aside from the monotony of constant research: Dr. Mike Walters, Dr. Tom
Baribault, John Willis, Bene Bachelot, David Minor, Mike Cook, Julie Kerr, Carol Graysmith and Sue
Plesko. The camaraderie and assistance provided by these friends was nothing short of extraordinary
and, on my part, necessary.
Finally, my part in this research would not have been possible without the grants and financial
support provided in part or in whole by the U.S. Department of Energy, Michigan State University
AgBio Research, and the Michigan Economic Development Corporation. This material is based upon
work supported by the Department of Energy under award number DE‐EE‐0000280.

iv

DISCLAIMER
This report was prepared as an account of work sponsored by an agency of the United States
Government. Neither the United States Government nor any agency thereof, nor any of their
employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for
the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed,
or represents that its use would not infringe privately owned rights. Reference herein to any specific
commercial product, or service by trade name, trademark, manufactured, or otherwise does not
necessarily constitute or imply its endorsement, recommendation, or favoring by the United States
Government or any agency thereof. The views and opinions of authors expressed herein do not
necessarily state or reflect those of the United States Government or any agency thereof.

v

TABLE OF CONTENTS

LIST OF TABLES

VII

LIST OF FIGURES

VIII

INTRODUCTION
REFERENCES

1
8

REGIONAL VARIATION OF MICHIGAN'S NON-INDUSTRIAL PRIVATE FOREST OWNERS
1.1 INTRODUCTION
1.2 STUDY OBJECTIVES
1.3 STUDY METHODS
1.4 RESULTS
1.4.1 Characteristics of Michigan NIPF lands
1.4.2 Regional demographic comparison
1.4.3 Importance of reasons for owning forestland
1.4.4 Performance of management activities
1.5 DISCUSSION AND CONCLUSION
1.5.1 Policy Implications
REFERENCES

11
11
13
13
15
15
16
19
22
22
24
27

A MARKET SEGMENT APPROACH TO IDENTIFYING MICHIGAN NONINDUSTRIAL
PRIVATE FOREST OWNERS
2.1 INTRODUCTION
2.2 STUDY OBJECTIVES
2.3 METHODS
2.4 ANALYSIS
2.4.1 Variable reduction using principal component analysis
2.4.2 Identifying NIPF owner segments
2.4.3
Segment differentiation using discriminant analysis
2.5 NIPF OWNER SEGMENTS
2.5.1 Demographic differences between segments
2.5.3 Reasons for owning forestland
2.5.4 Landowner segment perception of policy tools
2.6 DISCUSSION
REFERENCES

30
30
32
33
36
36
40
41
42
44
46
48
49
53

CONCLUSION
REFERENCES

57
60

vi

LIST OF TABLES

Table 1.1

Survey response rate by region.

14

Table 1.2

Median demographic and parcel characteristics by Michigan region.

17

Table 1.3

Median response of importance of reason for owning forestland by region.

21

Table 1.4

Percentage of respondents performing management activities by region.

22

Table 2.1

Rotated component loadings and Cronbach's alpha for Principal Components.

39

Table 2.2

Summary of discriminant analysis and explanatory power of variables used to create
clusters.

42

Table 2.3

Group means at center for NIPF segments on each discriminant function.

43

Table 2.4

Demographic characteristics of NIPF landowner segments.

44

Table 2.5

Forest parcel size by NIPF owner segment.

45

Table 2.6

Performance of management activities (percent) by NIPF owner segment.

46

Table 2.7

Median importance of reasons for owning forestland by NIPF owner segment.

47

Table 2.8

Importance of incentives and programs to aid landowners in meeting forestland
ownership objectives.

49

vii

LIST OF FIGURES

Figure 1.1

Histogram of parcel size category by region.

viii

19

Introduction
Non-Industrial Private Forest (NIPF) owners comprise nearly two thirds of United
States’ forestland (Butler and Leatherberry, 2004). The NIPF category contains forestlands which are
owned by neither government nor forest product industry (Butler and Leatherberry, 2004). Privately
owned forests contribute to societal benefits, such as recreation, ecosystem services and timber
production (Finley and Kittridge, 2006). In Michigan, 47.3% or 8.8 million acres of the state's 19
million acres of timberland is controlled by NIPF owners (Leatherberry et al., 1998). Michigan's forest
product industry contributes as much as $12 billion in annual revenue and an estimated 10% of the
state's manufacturing jobs (Korpi, 2010). Almost half of the state’s wood fiber originates from NIPF
lands (G.C. and Potter-Witter, in progress). Michigan’s forests and recreational opportunities are a draw
for tourists both within and outside the state and the tourist industry contributes an additional $11.5
billion annually in revenue (Stynes, 2002). Considering the importance of Michigan's forests,
understanding and influencing the behavior and attitudes of NIPF landowners has been a priority for
extension and state forest management organizations.
NIPF owners are facing several challenges to owning forestland. In recent years, pressure has
increased on landowners to sell and split their holdings into smaller parcels for development (Egan and
Luloff, 2000). Smaller parcels complicate forest management activities, and in some cases, may reduce
the financial viability of conducting timber harvests (Cleaves and Bennett, 1995). Further, parcelization
reduces recreation access on private lands (Gobster and Rickenbach, 2004). Additionally, the last
decade has seen an increase in inter-generational land transfer (Butler and Leatherberry, 2004). Many
of the new property owners have a less active interest in their forest property or may not understand the
proper care and maintenance of a healthy forest (Egan and Luloff, 2000). A whole new generation of
landowners may need education, information, or assistance in managing their forest parcels (Butler and
Leatherberry, 2004).
1

The factors affecting private landowner behavior are quite complex (Dennis, 1989). Nonresident ownership, forest parcel size, and non-timber preferences have all been studied in an effort to
predict private landowner behavior (Egan, 1997). However, landowners often have different
motivations for owning forested property (Butler and Leatherberry, 2004). The reasons for owning
forest property have been shown to directly influence a landowners' forest management decisions
(Janota and Broussard, 2008). Different management behavior and attitudes on adjoining parcels can
manufacture a patchwork of forest systems (Best, 2002). This patchwork can advance fragmentation
and may have unexpected consequences on habitat and opportunities for amenities or benefits from
private forestland (Gobster and Rickenbach, 2004).
To ensure the continuity of societal benefits from private forests, policy-makers have examined
programs to influence forest management objectives on NIPF lands (Schaff and Broussard, 2006).
Often these programs use financial incentives to encourage specific landowner behavior (Serbruyns and
Luyssaert, 2006). However, a landowner’s motivations for participating in forest management activities
are generally not financial (Koontz, 2001; Kendra and Hull, 2005). Instead, recreation or privacy often
are primary motivators in forest management decisions (Kluender et al., 1999; Koontz, 2001; Kendra
and Hull, 2005). Rather than encouraging management activities, financial incentives often offset
personal capital with public funding for those already performing the activities in question (Koontz,
2001).
In Michigan, several incentive programs have been developed to reach the private forest owner.
The Commercial Forest Program (CFP) offers tax reductions to those with 40 acres or more of
forestland (MiDNRE, 2009). These landowners must have a stewardship plan and be willing to allow
non-motorized recreation on their land (MiDNRE, 2009). Additionally, the Qualified Forest Program
(QFP) waives certain property taxes for enrolled parcels (MiDNRE, 2009). Enrolled forestland is
required to be 20 acres or greater in size and a stewardship plan must be maintained, but no open2

access recreation is required (MiDNRE, 2009). Withdrawal penalties for both programs require
repayment of ad-valorum taxes dependent on time enrolled (MiDNRE, 2009). Finally, some
opportunities for cost-share are available through the Forest Stewardship Program (FSP). This program
provides cost-share assistance for stewardship plans, up to $200 (Deb Huff, Nov. 5 2010, personal
communication). However, funding for this program is on a yearly basis and may fluctuate
unpredictably (Potter-Witter, 2005).
NIPF owners are not likely to respond to programs with any consistency (Kline et al., 2000).
When they do respond, it often varies across groups of landowners (Potter-Witter, 2005). Oftentimes
programs and policies can influence specific sectors of landowners, based on their ownership
objectives, while ignoring others (Serbruys and Luyssaert, 2006). Additionally, landowners hold
different viewpoints on governmental right or responsibility to influence specific objectives (Finley and
Kittridge, 2006). Further, cost share and other financial incentives to forest management tend not to
increase the number of landowners performing forest management, but rather increase the intensity of
the management performed (Kluender et al., 1999). Educational and informational programs generally
attract landowners who are already interested in performing forest management (Serbruyns and
Luyssaert, 2006). Reaching out to or providing assistance and information to those who are ill informed
or unaware of forestry practices can be far more difficult (Kendra and Hull, 2005). As a result of these
challenges, success and participation in various programs and policies across the United States varies
(Potter-Witter, 2005; Finley and Kittridge, 2006; Kluender et al., 1999; Kline et al., 2000)
In efforts to improve the effectiveness of future policies or programs, several studies have
attempted to identify sub-groups of NIPF owners. In 1998, Karpinnen examined NIPF landowners in
Finland grouping them by variables describing forest values. The study identified four groups of NIPF
owners: multi-objective owners, recreationists, self-employed owners, and investors. Relationships
between groups, values held by owners, silvicultural activities and landowner characteristics were
3

examined for significant differences. Karpinnen describes variance between groups in harvesting
timber as well as the relative importance of non-timber forest management objectives. He concludes a
landowner's association with one of the identified groups can be useful in determining landowner
behavior or tailoring extension and outreach materials to meet their needs.
Potter-Witter examined four groups of NIPF owners in Michigan (2005). Landowners were
surveyed based on their participation in FSP, CFP, Michigan Forest Association (MFA) and the TwoHearted River Watershed (TRW). She found that landowner demographics and parcel characteristics
differ significantly by group. Participants in MFA and CFP were significantly more likely to perform
management activities than landowners enrolled in FSP and TRW. Potter-Witter notes that program
managers should be cautious in expecting similar results across different programs, as different types of
programs are likely to attract landowners interested in different facets of owning forest property.
Investigating concerns over the effects of parcelization and land-transfers, Kendra and Hull
(2005) examined the attitudes of new NIPF owners in Virginia. The study used survey data to separate
landowners into groups. Cluster analysis produced six segments: investors, professionals,
preservationists, young families, forest planners and farmers. No identified segments indicated profit or
financial reasons as being important for owning forest property; instead lifestyle concerns
predominated. Each segment exhibited different concerns and beliefs towards their forestland. The
results indicate the challenges to working with new landowners, especially in regards to traditional
approaches of influencing landowners utilizing financial incentives. Kendra and Hull indicate the need
for further research to group landowners to provide more useful and defined categories for policy and
decision makers.
Massachusett's NIPF owners have been examined to determine landowner segments and
differences in participation in the state's Chapter 61 forest property tax program (Finley and Kittridge,
2006). The study highlights specific barriers to Chapter 61 enrollment based on three different
4

ownership groups. The groups identified by Finley and Kittridge (2006) are Jane Doe, Thoreau and
Muir. Where Thoreau owners may have primary objectives outside timber management, the two goals
are not necessarily mutually exclusive. However, Muir owners wish to allow nature to take its course
and programs requiring active management are likely to result in failure. The authors believe the
segmentation approach offers explanation on otherwise confusing patterns in program participation.
They also note the distinct differences in regard to management attitudes, objectives and participation
in programs and the necessity to consider this information when crafting and evaluating policies.
Serbruyns and Luyssaert (2006) examined private forest owner’s feelings and likelihood of
changing behavior under different policy alternatives. The authors identified four different groups on
which their analysis focused: profit-seekers, satisfied recreational, dissatisfied recreational and
materialistic. The study examined the differences between policy tools such as “sticks” (regulation),
“carrots” (incentives) and “sermons” (information, education or outreach). Different preferences for
policy tools were identified by landowner segments. Further, Serbruyns and Luyssaert determined
landowners were unlikely to change behavior, but would support programs or policies which reflected
current management attitudes and activities. NIPF owners have participated and supported programs
which aligned with their objectives and avoided those which are divergent. Thus, it may not be
program characteristics which affect group participation, but a landowner's objectives for their forest
property.
Finley and Kittridge (2006) highlighted the failure of programs in Massachusetts to reach a
large proportion of forest-owners. Much of this failure could be attributed to not taking into account
NIPF segment differences. Likewise, differences in participants across programs have been noted
(Potter-Witter, 2005). Further, the participation of landowners in government offered programs was
affected by a landowner’s beliefs and attitudes (Serbruyns and Luyssaert, 2006). Landowner groups
have a different set of beliefs and objectives regarding their forested parcel (Butler et al., 2007).
5

Finally, programs and information are products which are attempting to reach a customer, the
landowner. Therefore, understanding the landowner is essential to program and extension success.
This study seeks to examine NIPFs at a finer scale in hope of refining landowner groups for
future analysis, outreach development and program creation. The study is organized as a series of two
papers, each examining NIPF owners at a different scale. Both papers utilize information obtained from
a 2003 survey of Michigan NIPF owners. The survey instrument was designed to determine parcel
characteristics, demographics, reasons for owning forest property, past management activities,
preferences for policies and programs and the past utilization of extension and outreach materials.
The first paper breaks NIPF owners into categories based on location. The paper identifies
regional differences in landowner demographics, parcel characteristics, reasons for owning forestland
and management activities. Regional differences may be the result of many different factors and
landowner characteristics which are not determinable by this study. However, the presence of these
differences indicates the potential of policies to affect regions differently.
The second paper uses market segmentation to identify different sectors of NIPF owners in
Michigan. The segments are developed based on survey respondent rankings of 11 reasons for owning
forestland. The result is three distinct segments of Michigan NIPF owners based on their reasons for
owning forest parcels. Differences between these segments in regards to management activities,
demographics, parcel characteristics and perceptions of potential programs are discussed to aid policymakers in future development.
The information contained within this examination of NIPF owners will be useful to extension
and outreach managers, policy-makers and researchers. Extension specialists can utilize this
information to create outreach materials and information sources which are targeted to specific
landowner groups. Policy-makers will want to examine the issues associated with landowner segments
or regional differences where they seek even distribution of incentives or programs. Researchers will
6

continue to examine NIPF landowners and their behavior, especially in regards to timber production or
providing of ecological services. The information contained within this study may help future studies to
dissect landowners into more readily manageable categories. Regardless of how this information is
utilized, it describes methods in which the issue of heterogeneity associated with the NIPF category
may be managed.

7

REFERENCES

8

Best, C. 2002. America’s private forests: challenges for conservation. Journal of Forestry: Vol. 100, No.
3, pp. 14-17.
Butler, B.J., and E.C. Leatherberry. 2004. America’s family forest owners. Journal of Forestry: Vol.
102, No. 7, pp. 4-14.
Butler, B.J., M. Tyrell, G. Feinberg, S. VanManen, L. Wiseman, and S. Wallinger. 2007. Understanding
and reaching family forest owners: Lessons from social marketing research. Journal of Forestry:
105 (7). pp. 348-357.
Cleaves, D.A. and M. Bennett. 1995. Timber harvesting by nonindustrial private forest landowners in
western Oregon. Western Journal of Applied Forestry 10(2). pp. 66-71.
Dennis, D.F. 1989. An economic analysis of harvest behavior: integrating forest and ownership
characteristics. Forest Science: 35(4), pp. 1088-1104.
Egan, A.F. 1997. From timber to forests and people: A view of nonindustrial private forest research.
Northern Journal of Applied Forestry: 14(4), pp. 189-193.
Egan, A.F. and A.E. Luloff. 2000. The exurbanization of America’s forests: research in rural social
science. Journal of Forestry: 98(3), pp. 26-30.
Finley, A.O. and D.B. Kittridge. 2006. Thoreau, Muir, and Jane Doe: Different types of private forest
owners need different kinds of forest management. Northern Journal of Applied Forestry: 22
(1). pp. 27-34.
G.C., S. and K. Potter-Witter. In progress. Michigan’s forest products industry: are they ready for
promoting bioenergy?
Gobster, P.H. and M.G. Rickenbach. 2004. Private forestland parcelization and development in
Wisconsin’s northwoods: perceptions of resource-oriented stakeholders. Landscape and Urban
Planning: 69, pp. 165-182.
Janota, J.J., and S.R. Broussard. 2008. Examining private forest policy preferences. Forest Policy and
Economics: 10(3), pp. 89-97.
Kendra, A and R.B. Hull. 2005. Motivations and behaviors of new forest owners in Virginia. Forest
Science: 51(2), pp. 142.
Kline, D., J. Alig, and L. Johnson. 2000. Fostering the production of nontimber services among forest
owners with heterogeneous objectives. Forest Science: 46 (2). pp. 302-311.
Kluender, R.A., T.L. Walkingstick and J.C. Pickett. 1999. The use of forestry incentives by
nonindustrial forest landowner groups: is it time for a reassessment of where we spend our tax
dollars? Natural Resources Journal: 39 (4), pp. 799-818.
Koontz, T.M. 2001. Money talks-but to whom? Financial verses nonmonetary motivations in land use
decisions. Society and Natural Resources: 14 (1), pp. 51-66.
9

Korpi, K. W. 2010. Trends in Michigan's forest products industry 2000-2010. Michigan Forest Product
Council. Available at: http://michiganforest.com/documents/treds2010.pdf
Leatherberry, E.C., N. P. Kingsley and T.W. Birch. 1998. Private timberland owners of Michigan, 1994.
U.S. Forest Service Resource Bulletin NC-191.
Michigan Department of Natural Resources and Environment. 2009. Federal assistance programs for
non-industrial private landowners. Available at:
hhttp://www.michigan.gov/documents/dnr/FederalAssistancePrograms-NIPF2009_275100_7.pdf
Potter-Witter, K. 2005. A cross-sectional analysis of Michigan nonindustrial private forest owners.
Northern Journal of Applied Forestry: 22 (2). pp. 132-138.
Schaaf, A.K. and S.R. Broussard. 2006. Private forest policy tools: A national survey exploring the
American public's perceptions and support. Forest Policy and Economics: 9, pp 316-334.
Serbruyns, I. and S. Luyssaert. 2006. Acceptance of sticks, carrots and sermons as policy instruments
for directing private forest management. Forest Policy and Economics: 9(3), pp. 285-296
Stynes, D.J. 2002. Michigan statewide tourism and spending economic impact estimates 1998-2000.
Working draft, January 2002. Available at:
https://www.msu.edu/course/prr/840/econimpact/MIindex.htm

10

Paper 1.
Regional Variation of Michigan's Non-Industrial Private Forest Owners

1.1 Introduction
The state of Michigan has begun pursuing the development of alternative energy programs,
specifically in wood-based bio-energy (Michigan Department of Labor and Economic Growth, 2006).
However, wood production from both state and federal lands is currently utilized in Michigan's existing
wood products industry (Mueller et al., 2010). Subject to public review and comment, harvest levels on
public lands are planned for five to ten year intervals. Consequently, timber output on public land is not
expected to change. Private lands, therefore, are receiving renewed attention with respect to their
willingness to provide wood fiber for bio-energy uses.
Michigan contains approximately 19.2 million acres of timberland, or forest land “which is
producing or capable of producing more than 20 ft3 per acre per year of industrial wood” (FIA, 2008).
Leatherberry et al. (1998) estimated that there were 384,700 Non-Industrial Private Forest (NIPF)
owners holding an estimated 8.8 million acres, or 47.3% of Michigan timberland. While such a large
percentage of Michigan's timberland is owned by private landowners, the NIPF subcategory has been
used to categorize a diverse array of forest parcels (Argow, 1996). Therefore, there is large variation
among the population of NIPF landowners with respect to their demographic characteristics, forestland
attributes and management behavior.
Certain characteristics have been used to predict forest management behavior. Straka et al
(1984) described relationships between parcel size and forest management. Specifically, the size of a
forest holding was positively related to management intensity. A study by Cleaves and Bennett (1995)
in Oregon corroborated these results. Other factors such as length of ownership, presence of existing
structures and absenteeism have also been found to be correlated with forest management decisions
11

(Vokoun et al., 2006). On the other hand, absenteeism and non-hunting recreation were negatively
correlated with a landowner's decision to harvest whereas parcel size, landowner debt, stumpage price
and landowner age had a positive relationships (Conway et al., 2003). Decidedly, the most important
factors for determining a landowners management behavior may be absenteeism and parcel size
(Conway et al, 2003; Vokoun et al, 2006).
Opportunities for cost-sharing, technical assistance and property tax reduction have been
utilized to encourage timber output and support responsible forest management at both state and federal
levels. The effects of these policies have varied by program and group (Potter-Witter, 2005). In
Michigan, landowners may receive tax reductions for enrolling forestland as “Commercial Forests”.
The program requires a tract of at least 40 contiguous acres on which landowners are willing to allow
non-motorized recreation (Natural Resources and Environment Protection Act, 1994). Landowners also
have the opportunity to enroll their forestland as a “Qualified Forest Property”. This program requires
20 contiguous acres and an approved management plan. While no public access is required, there can
be significant costs to landowners who withdraw lands from the program (Qualified Forest Program,
2006). Considering nearly a quarter of Michigan's landowners hold parcels of less than 50 acres, forest
management on a large amount of Michigan's lands may not currently be addressed.
Policy makers are interested in effectively directing conservation and management programs on
private forestlands to meet societal objectives (Serbruyns and Luyssaert, 2006). Yet this requires
information on factors affecting landowner's willingness to consider enrollment in these programs.
Demographic factors alone have been unable to adequately explain forest management behavior on
NIPF lands (Potter-Witter, 2005). Other research has suggested that absentee ownership, forest parcel
size and non-timber use preferences may be important in predicting management activities (Egan,
1997; Cleaves and Bennett 1995). Furthermore, another large contributing factor is a landowner's
reasons for owning forested property (Karpinnen, 1998). Societal, economic and ecological factors
12

vary across Michigan so it is reasonable to expect NIPF characteristics and landowner behavior to
differ as well.

1.2 Study Objectives
The factors affecting private landowner behavior are quite complex (Dennis, 1989). Absentee
ownership, forest parcel size, and non-timber preferences have all been studied in efforts to more
accurately predict NIPF behavior (Egan, 1997). Various degrees of correlation have been found for a
number of characteristics (Peterson and Potter-Witter, 2006; Vokoun et al., 2006; Cleaves and Bennett,
1995; Straka et al., 1984), but many of these have not been examined at regional levels. Many research
questions concerning economics and societal impacts on ecosystems are discussed by region; therefore,
it is important to examine landowner beliefs and decision processes at the same scale.
This study seeks to test the null hypothesis that Michigan regions – the Southern Lower
Peninsula (SLP), Northern Lower Peninsula (NLP), Eastern Upper Peninsula (EUP) and Western
Upper Peninsula (WUP) – do not differ in regards to management behavior, reasons for owning
forestland, landowner demographics or parcel characteristics. The analysis focuses on seven forest
management activities and eleven reasons for owning forested parcels. The examination of regional
differences could aid policy makers in developing programs and incentives which more adequately
target their private forest goals in Michigan.

1.3 Study Methods
In 2003, a survey of Michigan nonindustrial private landowners was conducted. The survey
instrument was based on a review of similar studies in the United States, especially those conducted by
the USDA Forest Service (Birch, 1996; Leatherberry et al., 1998) and an earlier examination of
Michigan landowner assistance programs (Potter-Witter, 2005). The instrument included questions on
13

demographics, parcel characteristics, reasons for owning forested property, assistance programs and
management activities.
The sample consisted of 1,600 randomly selected Michigan landowners stratified by number of
NIPF parcels in each region. Since no database of Michigan forest owners was available, parcels were
randomly selected for inclusion in the survey. First, a legal section of township was randomly selected.
If aerial photography determined at least 50% of the section was forested, it was included in the sample
frame. County and township tax rolls were consulted to obtain mailing addresses for all landowners
within the specified section. Based upon Leatherberry's (1998) estimates of the number of NIPFs in
Michigan and in each region, the target sample size was set at 0.05% of forest parcels and was stratified
such that 204, 235, 706 and 455 owners of WUP, EUP, NLP and SLP properties comprised the mailing
list. Using Dillman's (2009) Total Response Method, follow up postcards were sent to remind those
who had failed to respond. In total, 457 usable surveys were returned for a total response rate of
28.25% (Table 1.1). Adjusted for each region, response rates were 26.1% in the SLP, 30.4% in the
NLP, 28.5% in the EUP and 27.9% in the WUP.
Table 1.1 Survey response rate by region.
Mailed

Returned

Response

Statewide

1600

457

28.25%

SLP

455

119

26.2%

NLP

706

214

30.3%

EUP

235

67

28.5%

WUP

204

57

27.90%

14

To corroborate this survey, it was compared against results obtained by the National Woodland
Owner Survey (2009). Age, income, gender, primary residence location and parcel size were extracted
and compared to the earlier results. For the tested factors, no significant differences were detected.
The data was examined for measures of central tendency, frequency and distribution.
Exploration revealed possible regional variations to examine statistically. Missing or extreme values
identified in exploratory analysis were corrected, where possible, by consulting the original surveys.
Where a response could not be determined with certainty, it was omitted. Extreme values corroborated
by the original survey were retained.
Statistical tests were performed in R (R Project: http://www.r-project.org). Exploratory
analysis revealed age and tenure as normal, continuous variables. For this case, a students t-test was
used to compare regions. The remainder of the variables of interest were found to be non-normal. More
often than not, responses were bi-modal. A Chi-squared test was used for gender, education, income,
employment and management activities, as these are all either binary or categorical. The Wilcoxonranked-sum test was used for variables such as importance of reasons for owning forested property,
distance from nearest forested parcel and forest parcel size categories. While these are not continuous,
they are measured on a scale and therefore such a test is more sensitive to deviations from normality.

1.4 Results
1.4.1 Characteristics of Michigan NIPF lands
The average age for all respondents was 58 and 82.3% were male. Most respondents reported
acquiring “some college” education and nearly all (96.6%) had obtained a high school diploma. The
majority (51.0%) of respondents were employed full time and the remaining 37.1% were retired. The
median income for Michigan forest owners was $40,000-59,999 and 45.5% resided on their forested
parcel.
15

Across all regions, 56.6% Michigan landowners considered “to enjoy beauty or scenery” as a
“very important” reason for owning forestland. Additionally, “for land investment” and “to pass land on
to my children or other heirs” were seen as “somewhat important.” Respondents were neutral on “for
production of firewood or biofuel (energy).” Finally, “for recreation other than hunting or fishing” was
“somewhat important” across all regions. While this information can be useful in determining
characteristics of Michigan landowners, this analysis focuses on differences among Michigan's
geographic regions.

1.4.2 Regional demographic comparison
Landowners in the WUP tended to have a higher formal education than those in the NLP or
EUP (Table 1.2). Educational differences between the EUP and NLP were exhibited in a slightly higher
amount of formal education obtained in the NLP. In regards to employment, more EUP respondents
were employed part time or seasonally compared to those in the SLP who were more often retired or
employed full-time. It is important to note, terms such as “NLP landowners,” “landowners in the WUP”
or “EUP respondents” refer to where the parcel is located, and the landowner in question may not live
in the same region, or even the same state.

16

Table 1.2 Median demographic and parcel characteristics by Michigan region.*
Variable

SLP

NLP

EUP

WUP

Landowner age**

59a

57a

57a

57a

Gender (percent
male)

84.8%a

80.2%a

83.3%a

83.9%a

Education

Some collegeabc

Some collegea

Some collegeb

Bachelors degree or
equivalentc

Income

$40,000-59,999a

$40,000-59,999ab

$40,000-59,999ab

$60,000-74,999b

Employed full timea

Employed full
timeab

Employed full timeb

Employed full timeab

30 yearsa

25 yearsb

28 yearsab

26 yearsab

Forest size category

10-49 acres

10-49 acres

50-99 acres

10-49 acres

Permanent residence
(percentage)

67.9%

35.7%a

44.8%a

40%a

Employment^
Tenure**

15 milesa
82.5 milesbc
25.5 milesab
101.5 milesc
*Like superscripts denote no significant difference between regions. ** Mean was used to report average response. ^ Mode was used to
report average response.
Distance

17

SLP respondents held their forest parcel longer than those in the NLP (Table 1.2). Additionally,
more SLP landowners resided on their forested parcel than did those in other regions. Further, SLP
landowners lived closer to their nearest parcel than those in the NLP or WUP, while WUP respondents
lived furthest. A higher incidence of agricultural land was reported by SLP respondents compared to
those in other regions. Considering that many farms tend to be residences, this may explain the higher
percentage of permanent residences and close proximity to owned forest land found in the SLP.
All regions were significantly different regarding forest parcel size (Table 1.2). The NLP
contained the largest parcels sampled with 1.1% in the 1000-4999 acre category (Figure 1.1). However,
the EUP contained the largest median forest parcels with 32.7% of parcels larger than 100 acres.
Conversely, the SLP contained the most small parcels with 86.9% less than 50 acres in size. In general,
the Upper Peninsula contained more parcels in the 50-99 acre and 100-499 acre categories whereas the
Lower Peninsula was comprised of more parcels in the 10-49 acre category. This difference was
obtained since the parcel size categories are not normally distributed, and in fact are distributed quite
differently from each other. Wilcoxon Ranked Sum takes distribution into account when determining
whether two samples are from the same population.

18

Figure 1.1 Histogram of parcel size category by region.*

Percentage of Responses

70.0%
60.0%

EUP

50.0%

WUP

40.0%

NLP
SLP

30.0%
20.0%
10.0%
0.0%
0-9 ac

10-49 ac

50-99 ac

100-499 ac

500-999 ac

1000-4999
ac

Reported Parcel Size

*For

interpretation of the references to color in this and all other figures, the reader is referred to the
electronic version of this dissertation.

1.4.3 Importance of reasons for owning forestland
SLP respondents ranked “to protect nature and biologic diversity” as a more important reason
for owning forestland than those in the EUP (Table 1.3). In the SLP, 74.0% of respondents considered
“as part of my home, vacation home, farm, or ranch” as a “very important” reason for owning forested
property compared to 56.0% in the WUP and 58.2% in the NLP. Additionally, NLP respondents
indicated “for cultivation/collection of non timber forest products” as more important than SLP
landowners. EUP respondents reported “for production of sawlogs, pulpwood, or other timber
products” as more important than other regions. Where 44.4% of EUP respondents claimed this factor
as important, only 26.9% of those in the WUP, 29.7% in the NLP and 31.2% in the SLP provided
19

similar responses. Further, NLP respondents reported “for hunting or fishing” was a more important
reason for owning forested property than those in the SLP or WUP. Likewise, EUP respondents
considered the same reason as more important than those in the WUP. Finally, all regions felt “for
recreation, other than hunting or fishing” was “somewhat important”.

20

Table 1.3 Median response of importance of reason for owning forestland by region.*
Reason for owning forestland

SLP

NLP

EUP

WUP

To enjoy beauty or scenery

Very important

Very importanta

Very importanta

Very importanta

To protect nature and biologic
diversity

Very importanta

Very importantab

Somewhat
importantb

Very importantab

Somewhat
importanta

Somewhat
importanta

Somewhat
importanta

Somewhat
importanta

As part of my home, vacation
home, farm, or ranch

Very importanta

Very importantb

Very importantab

Very importantb

For privacy

Very importanta

Very importanta

Very importanta

Very importanta

To pass land on to my children or
other heirs

Somewhat
importanta

Somewhat
importanta

Somewhat
importanta

Somewhat
importanta

For cultivation/collection of non
timber forest products

Neutrala

Somewhat
unimportantb

Neutralab

Neutralab

For production of firewood or
biofuel (energy)

Neutrala

Neutrala

Neutrala

Somewhat
Unimportanta

Somewhat
unimportanta

Somewhat
unimportanta

Neutral

Somewhat
unimportanta

Somewhat
importantac

Very importantb

Very importantab

Somewhat
importantc

Somewhat
importanta

Somewhat
importanta

Somewhat
important-

For land investment

For production of sawlogs,
pulpwood or other timber
products
For hunting of fishing

Somewhat
importanta
*Like superscripts denote no significant difference.
For recreation, other than hunting
or fishing

21

1.4.4 Performance of management activities
Finally, the regional differences in management activities were compared. Few significant
differences were found among regions except for those activities generally performed in conjunction
with timber production. A greater number of EUP landowners cited “improved an existing stand for
timber” than those in any other region (Table 4) and was significantly different from SLP landowners.
Additionally, more EUP landowners had “built or improved access road or trail” than those in the NLP.
Moreover, a greater number of EUP and WUP respondents have harvested timber than those in the
SLP. No differences were detected among other management activities.

Table 1.4 Percentage of respondents performing management activities by region. *
Management activity

SLP

NLP

EUP

WUP

Planted trees for timber, wildlife, or
environmental protection

47.8%a

53.6%a

42.5%a

42.9%a

Harvested firewood

58.7%a

53.6%a

51.1%a

48.6%a

Improved an existing stand for timber

37.9%a

48.9%ab

61.5%b

48.6%ab

Fenced livestock or deer out of forested
land

15.1%a

5.1%a

17.9%a

11.1%a

Built or restored a wetland or pond

10.7%a

11.7%a

5.4%a

7.7%a

Built or improved access road or trail

22.2%ab

16.5%a

37.0%b

23.7%ab

37.7%a
*Like superscripts denote no significant difference.

49.6%ab

73.2%b

52.6%b

Harvested timber

1.5 Discussion and Conclusion
This analysis highlights differences in NIPF forest owners across Michigan's regions.
Specifically, the SLP had a higher degree of agriculture as well as a majority of non-absentee
22

landowners, suggesting many of the forest parcels may be influenced by agricultural uses. Additionally,
the higher population density in southern Michigan (Michigan Information Center, 2001) may
influence forest property and landowner characteristics in this part of the state. The lack of participation
in forest management activities is consistent with our findings that SLP landowners find nonconsumptive uses of their forest property as more important factors for ownership.
Northern Michigan is known for its many summer cottage retreats and vacation homes (Stewart
and Stynes, 2006). The amount of absentee landowners and greater distance from nearest forest parcel
corroborate these findings. The ranking of reasons such as “for hunting or fishing” and “for recreation
other than hunting or fishing” further indicate the potential leisure usage of these properties. The
relative proximity to SLP population centers and low population densities make the NLP ideal for
summer retreats.
In contrast, the incidence of management activities and the higher rating of “for the production
of sawlogs, pulpwood, or other forest products” separate the EUP as one of the more unique regions.
The high occurrence of absentee landowners suggested some of the same dynamics as the NLP,
however the EUP tends to be further from metro-Detroit than convenient for weekend cottage use.
Considering 50% of EUP landowners live 25.5 miles or closer to their forest property, it is far more
likely many of these landowners live elsewhere in the region. While interesting, further research is
needed to pinpoint the cause of differences compared to other regions.
Finally, the WUP was very similar to the NLP regarding private forest owners. However, WUP
landowners obtained more formal education than those in the NLP and these landowners considered
hunting and fishing was more important for owning forest property. Other similarities suggested many
of the same factors which drove NLP ownership may have influenced NIPF properties in the WUP.
Specifically, the degree of absentee ownership and distance to forest parcel suggest many WUP
landowners may have traveled from other mid-west population centers in the pursuit of vacation or
23

cottage properties.
The regional differences tended to be centered around specific categories of ownership factors.
For example, where all regions consider bequeathing their land to their heirs and non-hunting
recreation as equally important, regions can be more adequately distinguished based on the importance
of consumptive reasons for owning forest property such as timber production, hunting or fishing, or
biologic protection. These results indicated the possibility of a “spectrum” of forest ownership with
consumptive uses on one end and biologic protection on the other. Considering that preservation and
consumptive uses of forest parcels may be seen by landowners as exclusive or contrary to other
objectives, this can be expected. Future research could create useful indicators or predictor variables for
determining where on the spectrum landowners may fall.
This analysis is limited to reasons for owning forested parcels, basic parcel and landowner
characteristics and past performance of management activities. Regarding NIPF landowners, chief
interest has been in determining their willingness to produce timber (Egan, 1997). While a statement
can be made regarding past management activities, past management decisions do not have a clear
effect on future ones. Interesting future analyses could attempt to determine exogenous factors relating
to a landowners willingness to harvest on a regional scale. Since stumpage price and other harvest
factors vary geographically, such a study could present useful information in further determining, at a
finer scale, a landowners decision process regarding forest management.

1.5.1 Policy Implications
In a state as diverse as Michigan, it is not surprising that factors describing private landowners
and their parcels differ geographically. Where past research and policies have described an entire state,
this study reports unique differences and motivations by region. While the driving forces behind this
variation require further examination, it retains important implications for forest policies and programs.
24

NIPF owner's personal beliefs and objectives have been shown to influence participation in a variety of
forest management programs (Serbruyns and Luyssaert, 2006). Ownership objectives greatly influence
a landowners willingness to participate in government sponsored programs (Janota and Broussard,
2008). Erickson et al. (2002) point that modern forest policies must strive to target wide bases of forest
owners by offering assistance or incentives consistent with a landowner's reasons for owning
forestland. Such reasons are not often financial (Erickson et al., 2002; Beach et al., 2003). Therefore,
where landowners' objectives and forest parcel characteristics differ geographically, we can expect
behaviors, including participation in management and conservation programs, to vary as well.
Considering Michigan's recent investigation of wood-based alternative energy, we can
anticipate increased pressure on private lands to produce wood fiber. Since financial incentives are
ineffective motivations for conducting forest management (Erickson et al., 2002; Beach et al., 2003),
policy makers may wish to examine new methods of encouraging participation in government
programs. Where we have illustrated geographic variation among NIPF lands, past researchers have
examined the relationships between landowners and participation in activities. These factors suggest
Michigan policy-makers must seek to develop policies which take into account regional variation if
they hope to produce more effective participation in forest management incentives and programs.
Since policy objectives are not often regional in nature, these differences could have profound
implications for their effects and success. Where one region may be likely to participate in a proposed
program or incentive, another may fail to actively enroll. Such disconnects in forest management can
only contribute to differences in forest management patterns and practices. Therefore, policy-makers
will wish to carefully consider regional differences when developing policies across large geographic
areas to avoid problems of inequality or potential forest fragmentation.
Knowing differences in regional forest management, policy makers could craft programs
specifically designed to utilize this variation to their advantage. Performance of management activities
25

could be influenced by utilizing landowner characteristics endogenous to specific regions. For example,
where landowners consider “for biologic protection” as an important reason for owning property,
programs could be designed that utilize a landowners interest in bio-diversity or wildlife habitat to
stimulate specific objectives. Programs which are more aligned with a landowner's value of land might
go a long way in increasing participation in these regions.
Regardless of objectives, policy-makers should strive to understand all possible effects of their
programs to avoid unintended consequences. While this study presents one way of looking at NIPF
landowners, some questions remain unanswered. For instance, perhaps there is potential for
participating in various policies or government programs based on reasons for owning forested land.
Further, since financial incentives are not always effective at influencing landowners, what policy
methods may be more attractive to certain landowner groups? Or, based on reasons for owning forest
property, how have landowners viewed management activities which they have performed? Would they
conduct them again? Future research could shed light on these issues and provide further information to
guide policy analysis and creation.
As our world continues to change, we can expect society to place increased demand for multiple
uses on our nation's forestland. Considering the extent of private land control of our forests, it is ever
more important we continue to seek to understand the NIPF owner. While NIPF landowners have
proven to be quite complex, certain factors can help facilitate in the further clarification of their
motivations. Regardless of whether we wish to encourage NIPF landowners to harvest timber, or
provide further recreation services, it may be time to dissect the category and look at this unique sector
of forestland at a finer scale. Perhaps then, we will begin to understand the elusive NIPF owner.

26

REFERENCES

27

Argow, K.A. The potential and diversity of non-industrial private forests. 1996. Journal of Forestry:
94(2), pp. 30-33.
Beach, R. H., S.K. Pattanayak, J. Yang, B.C. Murray and R.C. Abt. 2003. Econometric studies of nonindustrial private forest management a review and synthesis. Forest Policy and Economics: 7, pp.
261-281
Birch, T.W. 1996. Private forest-land owners of the United States 1994. USDA Forest Service Resource
Bulletin: NE-136, 293 p.
Cleaves, D.A. and M. Bennett. 1995. Timber harvesting by nonindustrial private forest landowners in
western Oregon. Western Journal of Applied Forestry 10(2).
Conway, C.M., G.S. Amacher, J. Sullivan and D. Wear. 2003. Decisions nonindustrial forest
landowners make: An empirical examination. Journal of Forest Economics: 9, pp. 181-203.
Dillman D.A., J.D. Smyth and L.M. Christian. 2009. Internet, mail and mixed-mode surveys: the
tailored design method, 3rd edition. John Wiley: Hoboken, NJ 499 pp.
Dennis, D.F. 1989. An economic analysis of harvest behavior: integrating forest and ownership
characteristics. Forest Science: 35(4), pp. 1088-1104.
Egan, A.F. 1997. From timber to forests and people: A view of nonindustrial private forest research.
Northern Journal of Applied Forestry: 14(4), pp. 189-193.
Erickson, D.L., R.L. Ryan, and R. De Young. 2002. Woodlots in the rural landscape: landowner
motivations and management attitudes in Michigan (USA) case study. Landscape and Urban
Planning: 58, pp. 101-112.
Forest Inventory and Analysis. 2008.United States Department of Agriculture, Forest Service.
Janota, J.J., and S.R. Broussard. 2008. Examining private forest policy preferences. Forest Policy and
Economics: 10(3), pp. 89-97.
Karpinnen, H. 2008. Values and objectives of non-industrial private forest owners in Finland. Silva
Fennica: 32(1), pp. 43-59.
Leatherberry, E.C., N.P. Kingsley, and T.W. Birch. 1998. Private timberland owners of Michigan, 1994.
USDA Forest Service Resource Bulletin NC-191.

28

Michigan Department of Labor and Economic Growth. 2006. Clean energy from wood residues in
Michigan. Michigan Biomass Energy Program.
Michigan Department of Natural Resources and Environment. 2009. Federal assistance programs for
non-industrial private landowners. Available at:
hhttp://www.michigan.gov/documents/dnr/FederalAssistancePrograms-NIPF2009_275100_7.pdf
Michigan Information Center. 2001. Land are and population density for Michigan, counties and MCD
April 1, 2000. http://www.michigan.gov/documents/co_mcd_26764_7.PDF. Accessed: Oct 12,
2010.
Mueller, L. S., Shivan G.C., and K. Potter-Witter. 2010 Michigan woody biomass supply snapshot.
Report to the Forestry Biofuels Statewide Collaboration Center. Michigan Economic Development
Corporation.
National Woodland Owner Survey. 2009. United States Department Of Agriculture, Forest Service.
Natural Resources and Environmental Protection Act. 1994. Mich. Stat., 511.
Peterson, G. and K. Potter-Witter. 2006. Stalking the elusive family forest owner in Michigan.
Proceedings of the Society of American Foresters 2006 Convention. Society of American Foresters,
Bethesda, MD.
Potter-Witter, K. 2005. A cross-sectional analysis of Michigan nonindustrial private forest owners.
Northern Journal of Applied Forestry: 22 (2). pp. 132-138.
Qualified Forest Program 2006. Michigan Statue, 378.
Serbruyns, I. And S. Luyssaert. 2006. Acceptance of sticks, carrots and sermons as policy instruments
for directing private forest management. Forest Policy and Economics: 9(3), pp. 285-296
Stewart, S.I. and D.J. Stynes. 2006. Second homes in the upper midwest. In N. McIntyre, D.R.
Williams and K.E. McHugh, Multiple dwelling and tourism: negotiating place, home and identity.
(pp. 180-193). CABI, Oxfordshire, UK.
Straka, T.J., H.W. Wisdom and J.E. Moak. 1984. Size of forest holding and investment behavior of
nonindustrial private owners. Journal of Forestry: August 1984, pp. 495-496.
Vokoun, M., G.S. Amacher and D.N. Wear. 2006. Scale of harvesting by non-industrial private forest
landowners. Journal of Forest Economics: 11(4), pp. 223-244.

29

Paper 2.
A market segment approach to identifying Michigan nonindustrial private forest owners

2.1 Introduction
The state of Michigan contains 19.2 million acres of timberland, or forestland which is “capable
of producing 20 ft3 of wood fiber annually” (FIA, 2008). Michigan's forestland plays an important role
in sustaining the timber industry, outdoor recreation and tourism industries as well as in developing the
state's natural character (Leatherberry et al., 1998). In fact, the Michigan timber industry accounts for
approximately 10% of the state's manufacturing jobs and an estimated $12 billion in annual revenue
(Korpi, 2010). Michigan's tourist industry is estimated to contribute an additional $11.5 billion in
annual revenue (Stynes, 2002) and much of this is centered on the state's bountiful natural resources.
Approximately 47.3%, or 8.8 million acres, of Michigan's timberland is controlled by an estimated
384,800 nonindustrial private forest (NIPF) owners and as much as 49% of Michigan's timber products
originate from private lands (GC and Potter-Witter, in press). Each landowner may have different
objectives and goals for his or her timberland, which can lead to a patchwork of forest management
(Best, 2002). Considering this variation, much interest has been dedicated to understanding and
influencing Michigan's NIPF owners.
Currently, Michigan has several programs designed to target private forest owners. The
Commercial Forest Program (CFP) allows property tax reductions on forested tracts of at least 40 acres
which allow open-access to non-motorized forms of recreation (Natural Resources and Environment
Protection Act, 1994). Only 2.2 million acres and 1,600 private landowners (nonindustrial and
industrial) are currently enrolled in this program (MiDNRE, 2009). Another opportunity for property
tax reduction is available through the “Qualified Forest Program” (QFP). This program does not

30

mandate open-access recreation and requires 20 acres of undeveloped forestland (Potter-Witter, 2008).
The potential recapture tax upon withdrawal of land from the QFP can be up to 14 times the current ad
valorem tax for the property and statewide enrollment is capped at 900,000 acres for the 2010 tax year
(MiDNRE, 2009). For CFP, withdrawal penalties may total a maximum of 15 years of ad valorem tax
(MiDNRE, 1993). While programs differ on some factors including forest parcel size, they both require
commitment to a forest management plan (Potter-Witter, 2005). Finally, Michigan offers federal
assistance through the Forest Stewardship Program (FSP). This program offers cost-share assistance for
the development of a forest management plan (MiDNRE, 2009). However, the FSP does not require
management activities and continuation of funding is year-to-year (Potter-Witter, 2005). Regardless of
the differences between available programs, enrollment remains a small minority of NIPF landowners.
Participation in programs often varies depending on landowner characteristics (Potter-Witter,
2005). The factors that influence a landowner’s performance of management activities can be complex
and difficult to pinpoint (Dennis, 1989). Research has shown that financial incentives often have little
impact on NIPF lands (Beach et al., 2003; Serbruyns and Luyssaert, 2006). A wide variety of attitudes
and beliefs factor into a landowner’s decision to perform forest management (Erickson et al., 2002).
Lifestyle and amenity choice are often of higher consideration than are financial reasons for owning
forested land (Koontz, 2001; Kendra and Hull, 2005). While a landowner may have a variety of reasons
for owning forestland, these do not always exclude wood production (Karpinnen, 1998). Instead,
landowners may be more selective when choosing management activities, including timber production,
which will not affect and can occur along with those objectives he or she deems more important
(Karpinnen, 1998). In other words, a landowner seeks compatibility between management activities
and their primary land ownership objectives.
The primary objectives of a landowner for his or her parcel are paramount in influencing forest
management program participation (Janota and Broussard, 2008). By and large, landowners will adopt
31

programs which fit with the current usage of their land (Serbruyns and Luyssaert, 2005). NIPF owners
are often unwilling to compromise the current objectives for financial incentives or other program
benefits (Kline et al., 2000). If policies seek to achieve equal distribution and maximum effectiveness,
they must strive to target wide bases of NIPF landowners by offering programs consistent with reasons
for owning forest property and management objectives (Erickson et al, 2002). The first step in creating
such policies is to identify target landowner groups for inclusion in incentives, programs, or policies.
Cluster analysis regularly is used in marketing research and analysis to identify and describe
unique market sectors to facilitate effective product marketing (Hair et al., 2006). Considering that
forest management policies are essentially a “product” and policy-makers seek a “customer” to utilize
them, this analysis could be useful in identifying potential NIPF landowner sectors. Once identified,
future policies or outreach programs could be constructed to meet the demands of each unique
landowner sector. Cluster analysis is a relatively new approach to identifying sub-groups of NIPF
landowners (See: Karpinnen, 1998; Finley and Kittridge, 2006) and it has yet to be used in defining
policy targets in Michigan.

2.2 Study Objectives
Considering low participation rates in Michigan's current forest management programs,
identifying key sectors may help in crafting policies with greater reach, accessibility, or influence. This
study attempts to identify landowner segments in Michigan for this purpose. Survey data from 2003
was used to identify clusters (alternatively: segments or sectors) of Michigan NIPF landowners based
on their reasons for owning forested property. Once identified, these clusters were defined and tested
for significant differences in order to provide information for future program construction and
administration. The information generated from this study can then be utilized to understand difference
in participation rates across programs, or create and alter policies which seek to utilize landowner
32

segments to achieve specific objectives.

2.3 Methods
This analysis used data obtained from a 2003 survey of Michigan NIPF landowners. The survey
instrument was designed to obtain information on landowner and parcel characteristics, reason for
owning forest property, management activities and perceptions of management incentives and
extension materials (Peterson and Potter-Witter, 2006; Mueller and Potter-Witter, in progress). The
survey also sought to determine a landowners' motivation for refraining from conducting harvesting
timber. The instrument was mailed to a stratified random sample of 1,600 private forest owners in
Michigan. Since no database of Michigan landowners exist, one had to be constructed. First, legal
sections were randomly selected. Aerial photographs were then consulted to determine if the section
contained at least 50% forest cover. If so, all landowners with more than ten acres of land were added
to the database. Stratification was done by region such that 204, 235, 706, 455 surveys were sent to the
Western Upper Peninsula (WUP), Eastern Upper Peninsula (EUP), Northern Lower Peninsula (NLP)
and Southern Lower Peninsula (SLP) respectively. The construction and distribution of the survey
followed the Total Design Method as outlined in Dillman (2009).
This analysis followed a three phase approach to identify specific NIPF owner segments.
Segments were constructed based on responses to eleven, five-point Likert scale, questions regarding
the importance of reasons for owning forestland. First, principal component analysis (PCA) was
performed to reduce the number of variables where correlation existed. PCA is used to examine the
underlying structure among variables included in an analysis (Hair et al., 2006). It uses the correlation
between variables to make combinations which reduce the number of independent variables in a model
but seeks to retain the variance structure of the dataset (Hair et al., 2006). In other words, the number of
independent variables is reduced based on correlation with each another. This can help to increase the
33

interpretability of a dataset without losing substantial amounts of information. Once principal
components, representing original variables, were created, PCA was used to generate factor scores for
each observation. The scores are a normalized representation of the component based on how a
respondent scored the variables associated with the component in question (Hair et al., 2006). The
number of components needed to adequately explain data is dependent on the complexity of the
dataset. Scree tests, latent root criterion, or prior research knowledge can be utilized to base judgment
on component retention (Hair et al., 2006). A Scree test relies on finding a point where no marginal
increase in the amount of explained variation is found between one set of components and another
(Hair et al., 2006). Latent root criterion seeks to retain the amount of components where the eigenvalue
is greater than one (Hair et al., 2006). An eigenvalue of one indicates the component is explaining as
much variation as an original variable. As the analysis seeks to combine variables into better
representations, the set of component solutions with an eigenvalue greater than one would be retained
(Hair et al., 2006).
Second, cluster analysis (CA) was used to create NIPF owner segments based on the reduced
independent variables from the PCA. CA is a statistical approach often used in market analysis to
identify specific segments (Hair et al., 2006). CA seeks to create “clusters” in which members of one
cluster are more similar to each other than they are to members of other clusters (Hair et al., 2006).
Segment means are used to group responses which are closest to the center of that “cluster” (Hair et al.,
2006). CA analysis attempts to reduce variability of members within a cluster, and increase variability
between clusters (Hair et al., 2006). Researchers using CA are searching for natural structures which
have descriptive significance, rather than creating artificial structures with statistical power (Hair et al.,
2006). To determine the number of clusters, prior information is used or researchers should carefully
consider the implications and meaning of potential solutions (Hair et al., 2006).
Once clusters were identified, multiple discriminant analysis (MDA) was used to determine
34

which factors were most effective in distinguishing each cluster. MDA is a common technique used in
the biological sciences to help explain the differences between two or more groups being sampled
(McGarigal et al., 2000). MDA utilizes methods similar to regression analysis to create functions which
predict an observation's group membership based on utilized measurement scales (Hair et al., 2006).
The number of functions developed is one less than the number of groups defined (McGarigal et al.,
2000). The first function seeks to explain the greatest amount of variance, where the subsequent
functions seek to explain the remainder (McGarigal et al., 2000). As in regression analysis, the function
creates coefficients which measure the effects of each variable within that specific function (Hair et al.,
2006). These functions are essentially linear combinations of the variables which are used to predict
group assignment (McGarigal et al., 2000). The form, is thus:

νj = β0j + β1jx1j + β2jx2j + β3jx3j + … + βijxij

equation 1.1

Where βij represents the coefficient of the ith variable for function j, νj represents the score for function
j, and xij represents the value of that variable for that observation (McGarigal et al., 2000). In this case,
the factor scores from PCA can be used as the variables to determine factor effects on group outcome
(McGarigal et al., 2000). MDA means can be computed which aid in further explaining the average
landowner from that specific segment (Hair et al., 2006). Essentially, this is similar to the “group
means” described in CA relative to the discriminant functions.
After MDA was performed and groups were identified, the groups were compared for
significant differences. In this study, the comparisons centered on parcel and ownership characteristics
as well as on the performance of specific forest management activities. Further, MDA was used to
determine the primary ownership reasons which indicate landowner’s association with a specific NIPF

35

owner segment. The comparison of these factors helped to establish the similarities and differences
among NIPF landowners.
After PCA, CA, and MDA were performed, the data were separated into landowner segments.
Exploratory analysis was performed on the variables of residency, forest parcel size, management
activities, age, education, income and employment to look for measures of central tendency, deviation,
average values and normality. Age exhibited a normal distribution, where other variables such as
management activities, income, education, employment and absenteeism were either binary or
nonparametric. Forest parcel size, due to its non-normal distribution as a larger number of smaller
parcels are owned than are large ones, was broken into forest size categories as used in previous
research (Birch, 1996; Leatherberry et al., 1998). Significant tests were set at ά = 0.05. Wilcoxonranked-sum was used to test differences in parcel size category between groups where chi-squared was
used for all other nonparametric tests. The single normally distributed variable, age, was tested using
the Student's t-test. All statistical methods and analyses were performed in R (R project: http://www.rproject.org)

2.4 Analysis
2.4.1 Variable reduction using principal component analysis
The 2003 survey asked respondents to report their ratings of importance of eleven reasons for
owning forestland. These reasons were “to enjoy beauty or scenery,” “to protect nature and biologic
diversity,”“for land investment,”“as part of my home, vacation home, farm, or ranch,”“for privacy,”“to
pass land on to my children or other heirs,”“for cultivation/collection of non-timber forest
products,”“for production of firewood or biofuel (energy),” “for production of sawlogs, pulpwood, or
other timber products,”“for hunting or fishing” and “for recreation, other than hunting or fishing.”
Respondents were asked to rank the importance of these reasons on a Likert scale of “very important”
36

to “not important.”
Often, some Likert-scale variables can be “linked” with one another. This may be the result of
questions which measure similar variables, such as core reasons for owning forest property. To reduce
the number of variables and simplify the dataset, an exploratory PCA was used to determine the
number of components necessary to retain explanatory variance. A Scree test was unable to determine
the number of necessary components. Using the latent root criterion, which suggests utilizing the
number of components which have an eigenvalue greater than one (Hair et al., 2006), four components
were retained. “To pass on land to my children or other heirs” and “for land investment” did not
adequately load (<0.50) or cross-loaded on several components (>0.30) and were retained as unique
identifiers.
The PCA was performed again, with those variables which inadequately loaded, or cross loaded
removed and examined utilizing a varimax rotation. This method rotates the data axis to achieve a
greater degree of interpretability (McGarigal et al., 2000). The rotation moves the data axes in order to
maximize the similarities within components, and maximize the differences between components. The
axes may not necessarily remain at 90 degrees to one another. This allows researchers to more readily
determine which variables should be included in each component. Important loadings for each
component are represented as numbers closer to one and less important characteristics are represented
as numbers closer to zero. A loading of 0.50 generally indicates significance and loadings of 0.20 to
0.50 may signify cross-loading between several components (Hair et al., 2006).
The rotated factor loadings indicate how each variable is represented by the newly created
component. These components are generally given names based on the original variables they represent
(Hair et al., 2006). The reasons for owning forest property of “as part of my home, vacation home,
farm or ranch” and “for privacy” both loaded on principal component one (PC 1) (Table 1). For this
reason, PC 1 was named “homestead.” PC 2 held responses related to utilitarian reasons for owning
37

forestland and was thus named “consumptive enjoyment.” These variables related to utilitarian uses
were “for cultivation/collection of non-timber forest products,”“for production of firewood or biofuel
(energy)” and “for production of sawlogs, pulpwood or other timber products.”“For hunting or fishing”
and “for recreation other than hunting or fishing” loaded on PC 3 which was named “personal
recreation.” PC 4 provides the information related to the reasons for ownership “to enjoy beauty or
scenery” and “to protect nature and biologic diversity,” therefore it was referred to as the component
explaining “preservation.”

38

Table 2.1 Rotated component loadings and Cronbach's alpha for Principal Components.
Rotated principle component scores
Items

Reasons for owning forestland

Homestead

Consumptive
Enjoyment

Personal
Recreation

Preservation

Cronbach's
Alpha
0.64

1

As part of my home, vacation home,
farm or ranch

0.86

0.05

0.00

0.14

2

For privacy

0.80

0.08

0.16

0.16

3

For cultivation/collection of non-timber
forest products.

0.22

0.75

0.12

0.05

4

For production of firewood or biofuel
(energy)

0.05

0.84

0.17

0.02

5

For production of sawlogs, pulpwood
or other timber products.

-0.09

0.86

0.05

-0.01

6

For hunting or fishing

0.00

0.21

0.84

0.00

7

For recreation other than hunting or
fishing.

0.18

0.08

0.75

0.24

8

To enjoy beauty or scenery

0.36

-0.05

0.09

0.79

9

To protect nature and biologic diversity

0.04

0.09

0.14

0.90

10

To pass land to my children or other
heirs

Retained as unique variable

11

For land investment

Retained as unique variable

39

0.76

0.55

0.72

Finally, Cronbach's alpha was computed for each PC. Cronbach's Alpha is a measure of
reliability which addresses the consistency of a measurement scale (Hair et al., 2006). Based on the
Cronbach's alpha, which suggests measures of 0.70 or greater are reliable (Hair et al., 2006), PC 2 and
PC 4 were reliable estimates of their combination variables. Due to the low alpha of PC 1 caution was
taken when interpreting results associated with this component. Finally, the low alpha associated with
PC 3 indicated the need to develop and refine indicators for testing factors representing recreational
importance of owning forestland. Regardless, this component was retained for further analysis as it
represents an important factor regarding forest ownership that was otherwise unincluded.

2.4.2 Identifying NIPF owner segments
CA seeks to analyze the heterogeneity of subgroups within a larger NIPF population (Hair et
al., 2006). While many studies seek to generalize across NIPF landowners (Egan, 1997; Leatherberry et
al, 1998; Conway et al., 2003) only a few have sought to identify specific segments within the NIPF
group (Karpinnen, 1998; Finley and Kittredge, 2006).
Based on the six factors (four generated from PCA and two unique variables), the k-means
algorithm was used to assign landowners to clusters. In this method, NIPF landowners who exhibited
similar characteristics in their reasons for owning forestland were grouped together to form a segment.
Note that these “segments” were based on reasons for owning forestland and not on other endogenous
or exogenous factors which might influence their characteristics. However, as reasons for owning
forestland often influence a landowner's decision process (Conway et al., 2003), these factors do play
an important explanatory role in the developed segments. To ensure best results, two-member, threemember, four-member, five-member and six-member clusters were analyzed. Following MDA in the
last phase of this analysis, the three-member solution provided results which were consistent with
40

earlier studies and most easily interpreted in regard to significant differences and unique distinguishing
characteristics of the generated segments.

2.4.3 Segment differentiation using discriminant analysis
CA created three segments based on their clustering around six variables (four generated from
PCA and two retained as unique variables) which represented the importance of reasons for owning
forestland. MDA was used to determine the unique profiles of each segment. This is by the
discriminant function which is a linear combination of the descriptive variables which seeks to describe
group membership (McGarigal et al., 2000). The first discriminant function explains the most variance,
while the second function seeks to explain remaining variance between the included variables and
groups. Specifically, MDA was used to examine the impact of each of the six variables on segment
membership.
Both discriminant functions (DF) were significant (p < 0.01) by their approximate F-value. All
included variables were significant in aiding to create the discriminant functions. Associated
standardized canonical loadings and rotated discriminant loadings are contained in table 2.2.

41

Table 2.2 Summary of discriminant analysis and explanatory power of variables used to create
clusters.
Standardized
Coefficients
Discriminant variables

Discriminant

Univariate F

Loadings

DF 1

DF 2

DF 1

DF 2

Ratio

Prob.

PC 1
Homestead

0.149

-0.023

0.071

0.065

176.15

< 0.01

PC 2
Consumptive enjoyment

0.656

-0.728

0.905

-0.083

433.96

< 0.01

PC 3
Personal recreation

0.351

0.550

-0.033

0.438

231.56

< 0.01

PC 4
Preservation

0.375

0.330

0.051

0.249

193.54

< 0.01

0.575

0.623

0.180

0.674

329.40

< 0.01

0.265

-0.226

0.265

0.074

198.20

< 0.01

Item 10
To pass on land to children
or other heirs
Item 11
For investment

2.5 NIPF owner segments
Discriminant function one, which was dominated by consumptive enjoyment, distinguished
among all three segments (Table 2.3). Discriminant function two, which had emphasis placed on
recreational uses for forestland and bequeathing land, distinguished between segment one and the other
two segments. As a result of the importance of personal recreation exhibited in group 1, illustrated by
the group mean and discriminant function, this group was labeled “Game Wardens.” These landowners
represented 29.8% of respondents. As illustrated by the discriminant functions, the Game Wardens
considered consumptive enjoyment of their forestland as less important than did those in segment three.
Due to the relative higher importance of consumptive enjoyment, segment three landowners were
labeled “Timber Barons.” Timber Barons represented 46.4% of respondents. Finally, segment two was

42

distinguished by the lack of relative importance of the factors illustrated in the MDA. From this
standpoint, segment two landowners may feel the same towards the importance for owning forestland
as do the others, with the exception of the strongly discriminating factors (personal recreation and
consumptive enjoyment). These landowners were labeled “Tenants” and comprised 23.8% of
respondents.

Table 2.3 Group means at center for NIPF segments on each discriminant function.
Discriminant Function Means
NIPF segments

n

DF 1

DF 2

Game Wardens
(Segment 1)

104
(29.8%)

-16.53

61.81

Tenants
(Segment 2)

83
(23.8%)

-101.16

-34.96

Timber Barons
(Segment 3)

162
(46.4%)

51.83

-20.07

It is important to note that the MDA analyzed differences between groups. Where Timber Barons may
have been “less interested” in personal recreational uses of their forestland, this does not mean they
were altogether uninterested, only that their ranking was different than was that of Game Wardens.
Likewise, MDA only measures the ability of the measurement scales to distinguish among groups.
Where a specific scale may be unimportant in MDA (ex: homestead or for investment), it could still be
a very important reason for owning forest property. Rather than the said factor distinguishing among
groups, the importance could be held in common among all groups.

43

2.5.1 Demographic differences between segments
At 52.9 years old, Game Wardens were the youngest of the landowner segments (Table 2.4). No
significant differences existed between landowner segments in regard to education, income or
employment. However, Game Wardens were less often permanent residents of their forest property than
were the other segments. Tenants resided on their parcel as permanent residents more often than did
those in other groups.

Table 2.4 Demographic characteristics of NIPF landowner segments.
NIPF segments
Item
Respondent Age
Education
Income
Employment

Game Wardens
(Segment 1)

Tenants
(Segment 2)

Timber Barons
(Segment 3)

52.9

57.2a

55.4a

Some collegea

Some collegea

Some collegea

$60,000-74,999a

$60,000-74,999a

$40,000-59.999a

Employed full timea

Employed full timea

Employed full timea

54.2%

43.8%

Absent Ownership
34.6%
Like superscripts denote no significant difference.

Additionally, Game Wardens tended to hold smaller forest parcels than did Timber Barons,
where Timber Barons held the largest forest parcels (Table 2.5). Game Wardens held no parcels greater
than 500 acres in size, while both of the other segments did so. Considering that recreation values can
be achieved on relatively small parcels (Zhang et al., 2005), there may not be a strong incentive to own
large parcels when recreation is the prime reason for owning forestland. Conversely, utilitarian uses of
forest property often require large parcels (Cleaves and Bennett, 1995). Timber Barons also held
significantly larger parcels than did Tenants. Again, this could be an indication towards the necessity of
Timber Barons to own larger parcels to meet their objectives, than a statement regarding Tenant
behavior.
44

Table 2.5 Forest parcel size by NIPF owner segment.
Forest Parcel Size
0-9
acres

10-49
acres

50-99
acres

100-499
acres

500-999
acres

1000-4999
acres

Game Wardensa
(Segment 1)

25.3%

53.8%

12.1%

8.8%

0.0%

0.0%

Tenantsa
(Segment 2)

28.8%

49.3%

12.3%

6.8%

2.7%

0.0%

19.2%

15.8%

0.0%

1.4%

NIPF segments

Timber Barons
11.6%
52.1%
(Segment 3)
Like superscripts denote no significant difference.

2.5.2 Participation in management activities
All management activities except “fenced livestock or deer out of forested land” were
significantly different by NIPF owner segment. Game Wardens had “planted trees for timber, wildlife,
or environmental protection” more than had Tenants, but less than Timber Barons (Table 2.6).
Conversely, fewer Game Warden respondents that they had harvested firewood than did other
segments. More Timber Barons had harvested timber and conducted a timber stand improvement than
had other landowner segments. However, where more Game Wardens had conducted timber stand
improvements than Tenants, fewer Game Wardens had harvested timber.
The disconnect between managing for timber, as illustrated by performance of timber stand
improvement, versus actual timber production is perplexing. It is possible that the survey question
measuring responses was unclear, or that Game Wardens were considering habitat management or
firewood collection a form of “timber improvement.” Finally, current Game Warden property owners
may have the desire to produce timber, but have not yet conducted a harvest. The survey instrument did
not require information on intentions; as such, we can only speculate on the reasons for this result.
Additionally, Game Wardens had “built or restored a wetland or pond” and “built or improved
access road or trail” more than had Tenants, but less than had Timber Barons. Tenant respondents
45

reported having “harvested firewood” and “harvested timber” more than did Game Wardens. This may
indicate a stronger inclination to “preserve” forestland in the Game Warden segment than that in
Tenants. Where Game Wardens tended to participate in restoration activities more than did Tenants,
Tenant NIPF owners seemed to participate more often in management activities which were centered
on utilization of their forestland.
This result contradicts that of the MDA where Tenants placed less emphasis on consumptive
enjoyment of their forestland than did other segments. However, while Tenants may have considered
“consumptive enjoyment” as less important than did other segments, this does not mean that they did
not partake in active management. Where these activities may not have been important reasons for
owning forestland, at least relative to other segments, “Tenants” could have viewed it as an active part
of owning forestland

Table 2.6 Performance of management activities (percent) by NIPF owner segment.
NIPF segments
Game Wardens
(Segment 1)

Tenants
(Segment 2)

Timber Barons
(Segment 3)

χ2, prob

Planted trees for timber, wildlife, or
environmental protection

65.1%

54.0%

78.8%

10.9, < .01

Harvested firewood

50.7%

55.8%

82.6%

25.1, < .01

Improved an existing stand for
timber

44.7%

23.3%

68.3%

26.8, < .01

Fenced livestock or deer out of
forested land

12.1%

2.5%

16.2%

4.7, 0.09

Built or restored a wetland or pond

20.3%

15.9%

43.7%

14.8, < .01

Built or improved access road or
trail

51.4%

36.7%

66.1%

12.4, < .01

Harvested timber
2.5.3 Reasons for owning forestland

36.8%

40.4%

68.6%

30.4, < .01

Management Activity

46

While MDA illustrates the relative differences of the measures among segments, it does not
report the overall importance of the measures. An examination of the absolute ranking of importance
for reasons for owning forestland by landowner segment provides more insight into the importance of
the measures for each group. The median responses to each of the original 11 measures are reported in
Table 2.7.

Table 2.7 Median importance of reasons for owning forestland by NIPF owner segment.
NIPF owner segments
Game Wardens
(Segment 1)

Tenants
(Segment 2)

Timber Barons
(Segment 3)

Very important

Somewhat
important

Very important

Very importantab

Somewhat
importanta

Very importantb

For cultivation/collection of of non-timber
forest products

Not importanta

Not importanta

Neutral

For production of firewood or biofuel
(energy)

Not importanta

Somewhat
unimportanta

Somewhat
important

For production of sawlogs, pulpwood or
other timber products

Not important

Not Important

Somewhat
important

For hunting or fishing

Very importanta

Neutral

Very importanta

For recreation, other than hunting or
fishing

Very importanta

Neutral

Somewhat
importanta

To enjoy beauty or scenery

Very importanta

Very important

Very importanta

To protect nature and biologic diversity

Very importanta

Somewhat
important

Very importanta

To pass on land to my children or other
heirs

Very importanta

Somewhat
unimportant

Somewhat
importanta

Neutrala

Somewhat
important

Reasons for owning forestland
As part of my home, vacation home, farm
or ranch
For privacy

For land investment

Somewhat
importanta
Like superscripts denote no significant difference.

Game Wardens considered all measures of reasons for owning forested property as important
47

with the exception of those factors within Consumptive Enjoyment. Conversely, only “to protect beauty
or scenery,” “to protect nature and biologic diversity,” “for privacy” and “as part of my home, vacation
home, farm or ranch,” factors which loaded on the homestead and preservation components, were
important to Tenant landowners. The reasons for owning forestland which fell into the consumptive
enjoyment component, “for cultivation/collection of nontimber forest products,” “for production of
firewood or biofuel (energy),” and “for the production of sawlogs, pulpwood or other timber products,”
were more important for Timber Barons than they were for other segments. These results reinforced
that of the MDA. Note, factors which were not important in distinguishing Timber Barons were cited as
being more important than were utilitarian reasons for owning forest property. Again, recall that MDA
only showed which factors were influential in distinguishing between clusters, but made no claims on
the absolute importance of these variables to landowners. As such, table 2.7 shows factors which did
not show strongly in MDA, but were considered “Very Important” to landowners.

2.5.4 Landowner segment perception of policy tools
To determine potential policy effects across landowner groups in Michigan, and to provide
information to aid in future policy creation, differences among landowner desires for incentives or
programs were examined (Table 2.8). Those landowners most actively interested in managing their
forestland (Game Wardens and Timber Barons) liked the idea of “free or low-cost management
assistance or professional advice” and “free or low cost educational materials” more than did the
Tenants who expressed less interest in forest management activities. Financial incentives were more
largely supported by Timber Barons than they were by other landowner groups. Tenants did not
consider financial reasons as important in aiding to meet their management goals with the exception of
“cost-share.” However, other financial incentives measured in the survey were tailored towards timber
management, where Tenants could be more interested in habitat management or improving
48

biodiversity. All landowners were supportive of “property tax reductions for management forest lands,”
however the meaning of “management” likely differed by landowner group.

Table 2.8 Importance of incentives and programs to aid landowners in meeting forestland
ownership objectives.

NIPF Owner Segments
Game Wardens
(Segment 1)

Tenants
(Segment 2)

Timber Barons
(Segment 3)

Free or low-cost management assistance or advice
from a natural resource professional

56.2%

35.2%

55.6%

Free or low-cost educational materials or events on
forest management

54.8%

45.9%

57.7%

Income tax deductions or credits for management
expenses

58.5%ab

30.6%a

63.2%b

Property tax reductions for managed forest lands

66.3%a

61.0%a

72.6%

Low interest loans to help pay for management
expenses until forestry income is achieved

20.1%a

9.8%a

26.6%

Cost-sharing assistance to help pay for forest
management

20.1%a

24.2%a

44.4%

Annual rental payments to provide income while
trees mature

26.9%a

15.8%a

36.1%

Higher payments for my timber
Like superscripts denote no significant difference.

25.3%a

24.5%a

58.9%

Incentives to encourage forest management

2.6 Discussion
The NIPF category of U.S. landowners has always been the target of research and policies since
colonial times (Binkley, 1981). Much of this can be attributed to the difficulty in understanding or
characterizing private landowners due to the diversity contained within the group. This analysis,
however, indicates that it is possible to combine NIPF landowners into segments for which the
variability within the group is less than that between groups. Statistical analysis indicates that some
landowners may have specific traits in common with one another, and thus these land owners can be
49

grouped into identifiable segments. This brings into question the wisdom of utilizing one “catch-all”
group to describe NIPF landowners. Perhaps, instead, policy-makers and researchers should consider
analyzing NIPF ownership at a finer scale in order to effectively understand this ownership group.
A few specific factors seem to be key in distinguishing among private landowners. Identifying
these factors is integral to understanding NIPF ownership in Michigan. Where traits such as homestead
and preservation are held in common across Michigan NIPF owners, consumptive enjoyment,
recreation, and bequeathing land are quite polarizing and aid in distinguishing among particular
ownership groups. While these differences do not illustrate the absolute importance of these factors to
particular landowners, they do illustrate relative differences between each segment; differences which
may be important in understanding landowner behavior. While Timber Barons found consumptive
enjoyment of their forestland to be more important than did other NIPF subgroups, it was still relatively
unimportant compared to other ownership factors. “For privacy”, “as part of my home, vacation home,
farm or ranch” and other characteristics were cited as more important than were the factors which
comprise “consumptive enjoyment.”
Timber Barons reported being the most active in managing their forestland (Table 2.6). This was
consistent with the higher importance they placed on consumptive enjoyment of their forestland. These
landowners may have similar objectives to other segments, but understand the potential of stewardship
to meet their objectives and retain the health of their forest system. Perhaps Timber Barons saw
compatibility between consumptive enjoyment and other reasons for owning forest property where
other segments saw such as divergent. While the absolute reason for such differences is not known, the
characteristics of this group can aid in understanding their motivations and behavior.
Game Wardens also tended to take an active role in managing their land. More of these
landowners had performed all activities, with the exception of “harvested timber” and “harvested
firewood” than had Tenants. The high importance placed on recreational value of their timberland may
50

influence Game Wardens to actively manage their land for such purposes. The enjoyment of wildlife
and nature may have driven these landowners to create nature trails, restore wildlife habitat, or perform
other activities related to forest recreation.
Finally, Tenants actively managed their land less than did other NIPF owner segments.
Conversely, these landowners performed activities that can be related to consumptive enjoyment
(“harvested firewood” and “harvested timber”) more often than did Game Wardens. Perhaps these
activities offer a direct tangible benefit and were thus preferred. However, as a group, these landowners
seemed uninterested in active forest management. It may be that latent ownership factors contained on
components such as Preservation may have been the primary objectives of these landowners. Another
possibility is that the factors most importantly held by these landowners, such as intrinsic values
associated with owning land, were not measured in the survey instrument.
This study focused on the importance of reasons for owning forestland and performance of
management activities as a means to distinguish among NIPF owner segments. While the NIPF results
show distinct groups which are significantly different from one another, some caution is necessary in
interpreting the conclusions. Factors unmeasured in this study may be more important in influencing
and separating NIPF owners. These could include intrinsic value in owning land, holding land for
development or later home development. The ability to distinguish segments based on the traits,
however, determines that there is a significant result which aids in characterizing NIPF landowners.
Therefore, while factors not studied here might help to further categorize or characterize NIPF
landowners, the results reported in this study still report specific and important differences between
segments.
Cluster analysis can be open to interpretation (Hair et al., 2006). There is no absolute
established way to determine how many clusters to utilize in a final solution (Hair et al., 2006).
Several clusters were tested before determining a result which was based on the interpretability of the
51

multiple discriminant analysis. This solution to the cluster problem has been used regularly before
(Finley and Kittridge, 2006). Further, the results based on the importance of consumptive enjoyment,
recreational values and bequeathing land is consistent with other studies (Karpinnen, 1998; Kendra and
Hull, 2005; Finley and Kittridge, 2006). The three-cluster solution, therefore, seems to offer the most
easily interpreted analysis which is also consistent with past examinations.
This and similar analyses open the door for future research into NIPF owner segments. Much
research has been dedicated to attempting to understand NIPF landowner motivations, specifically in
regard to timber harvests (Egan, 1997). However, few studies also seek to determine the informational
and support needs of landowner groups. Such studies could aid extension program managers and policy
makers in reaching out to forest landowners. Likewise, testing different informational sources or
programs across the different segments would help to craft effective policies for targeting specific
groups of landowners. More importantly, these could help determine programs which have greatest
potential for effect across all segments. Either way, the generally dismal performance of policies and
programs aimed at NIPF owners may be improved by further examination of NIPF segments.

52

REFERENCES

53

Beach, R. H., S.K. Pattanayak, J. Yang, B.C. Murray and R.C. Abt. 2003. Econometric studies of nonindustrial private forest management a review and synthesis. Forest Policy and Economics: 7, pp.
261-281
Best, C. 2002. America’s private forests: challenges for conservation. Journal of Forestry: 100 (3), pp.
14-17.
Binkley, C.S. 1981. Timber supply from private nonindustrial forests: a microeconomic analysis of
landowner behavior. Yale University: School of Forestry and Environmental Studies Bulletin
no. 92.
Birch, Thomas W. 1996. Private forest-land owners of the United States, 1994. Resource Bulletin NE134. Radnor, PA: U.S. Department of Agriculture, Forest Service, Northeastern Forest
Experiment Station. 183 p.
Cleaves, D.A. and M. Bennett. 1995. Timber harvesting by nonindustrial private forest landowners in
western Oregon. Western Journal of Applied Forestry 10(2). pp. 66-71.
Conway, C.M., G.S. Amacher, J. Sullivan and D. Wear. 2003. Decisions nonindustrial forest
landowners make: An empirical examination. Journal of Forest Economics: (9), 181-203.
Dennis, D.F. 1989. An economic analysis of harvest behavior: integrating forest and ownership
characteristics. Forest Science: 35(4), pp. 1088-1104.
Dillman D.A., J.D. Smyth and L.M. Christian. 2009. Internet, mail and mixed-mode surveys: the
tailored design method, 3rd edition. John Wiley: Hoboken, NJ 499 pp.
Egan, A.F. 1997. From timber to forests and people: A view of nonindustrial private forest research.
Northern Journal of Applied Forestry 14(4), pp. 189-193.
Erickson, D.L., R.L. Ryan, and R. De Young. 2002. Woodlots in the rural landscape: Landowner
motivations and management attitudes in Michigan (USA) case study. Landscape and Urban
Planning: 58, pp. 101-112.
Finley, A.O. and D.B. Kittridge. 2006. Thoreau, Muir, and Jane Doe: Different types of private forest
owners need different kinds of forest management. Northern Journal of Applied Forestry: 22
(1). pp. 27-34.
Forest Inventory and Analysis. 2008. United States Department of Agriculture, Forest Service
G.C, S. and K. Potter-Witter. In press. Status of Michigan's forest products industry in the changing
market environment. Forest Products Journal.
Hair, J.F., B. Black, B. Babin, R.E. Anderson, and R.L. Tatham. 2006. Multivariate Data Analysis (6th
edition). Pearson Education: Upper Saddle River, New Jersey.
Janota, J.J., and S.R. Broussard. 2008. Examining private forest policy preferences. Forest policy and
economics: 10(3), pp. 89-97.
54

Karpinnen, H. 2008. Values and objectives of non-industrial private forest owners in Finland. Silva
Fennica: 32(1), pp. 43-59.
Kendra, A and R.B. Hull. 2005. Motivations and behaviors of new forest owners in Virginia. Forest
Science: 51(2), pp. 142.
Kline, D., J. Alig, and L. Johnson. 2000. Fostering the production of nontimber services among forest
owners with heterogeneous objectives. Forest Science: 46 (2). pp. 302-311.
Koontz, T.M. 2001. Money talks-but to whom? Financial versus nonmonetary motivations in land use
decisions. Society and Natural Resources: 14 (1), pp. 51-66.
Korpi, K. W. 2010. Trends in Michigan’s forest products industry 2000-2010. michigan forest product
council. Available at: http://michiganforest.com/documents/treds2010.pdf
Leatherberry, E.C., N.P. Kingsley, and T.W. Birch. 1998. Private timberland owners of Michigan, 1994.
USDA Forest Service Resource Bulletin: NC-191. St. Paul, MN.
McGarigal, K., S. Cushman, and S. Stafford. 2000. Multivariate Statistics for Wildlife and Ecology
Research. Springer, 283 pp.
Michigan Department of Natural Resources and Environment. 1993. Michigan’s Commercial Forest
Act. Available at: http://michigansaf.org/ForestInfo/MSUElibrary/CFAact.PDF
Michigan Department of Natural Resources and Environment. 2009. Federal Assistance Programs
Non-Industrial Private Landowners.
Mueller, L.S. and K. Potter-Witter. In progress. Regional variation among non-industrial private forest
owners in Michigan.
Natural Resources and Environmental Protection Act. 1994. Michigan Statute, 511.
Peterson, G. and K. Potter-Witter. 2006. Stalking the elusive family forest owner in Michigan.
Proceedings of the Society of American Foresters National Convention, 2006. Society of
American Foresters, Bethesda, MD.
Potter-Witter, K. 2005. A cross-sectional analysis of Michigan nonindustrial private forest owners.
Northern Journal of Applied Forestry: 22 (2). pp. 132-138.
Potter-Witter, K. 2008. Qualified Forest Property Tax Program. Michigan State University Extension,
Forestry Fact Sheet 32. Available at: http://www.nobleforestry.com/downloads/msu_facts32.pdf

Serbruyns, I. and S. Luyssaert. 2006. Acceptance of sticks, carrots and sermons as policy instruments
for directing private forest management. Forest Policy and Economics: 9(3), pp. 285-296
55

Stynes, D.J. 2002. Michigan statewide tourism spending and economic impact estimates 1998-2000.
Working draft, January 2002. Available at:
https://www.msu.edu/course/prr/840/econimpact/MIindex.htm
Zhang, Y., D. Zhang, and J. Schlhas. 2005. Small-scale non-industrial private forest ownership in the
United States: rationale and implications for forest management. Silva Fennica: 39 (3), pp. 443454.

56

Conclusion
Influencing private lands has been a significant source of interest for researchers and policymakers alike. Policies and programs have tended to center on encouraging the maintenance of an intact
and ecologically healthy forest landscape or on encouraging forest management activities. Many
programs have sought to retain forestland from development for recreation and ecological purposes.
However, past policies and programs have had a variety of effects depending on the characteristics of
the landowner (Potter-Witter, 2005; Serbruyns and Luyssaert, 2006; Tyrell, 2010). This effect has been
seen in Michigan (Potter-Witter, 2005), as well as other locations around the world (Finley and
Kittridge, 2006; Serbruyns and Luyssaert, 2006).
This study highlighted diversity in Michigan private landowners with regard to reasons for
forest ownership, parcel characteristics, management activities and other factors. Recent efforts
indicate that landowners are willing to utilize various programs, policies, or financial incentives to
different degrees based on a myriad of factors including their parcel size and reasons for owning
forestland (Serbruyns and Luyssaert, 2006). Due to differences in landowner characteristics and
behavior, therefore, current policy and outreach efforts may target specific sectors of landowners while
neglecting others.
The first paper examined Michigan's NIPF owners based on the region in which the owner’s
forest parcel was located. SLP respondents placed an emphasis on owning forestland for nature and
biologic protection. Additionally, fewer SLP respondents harvested timber and more resided on their
forested parcel. In contrast, NLP respondents owned their land primarily for recreation, as part of their
home, or for privacy. EUP landowners placed importance on hunting and fishing as reasons for owning
forestland, and many said they were open to timber-harvesting or other management activities. Finally,
WUP landowners placed an emphasis on owning forestland as part of their home or vacation home, for
biological protection and for privacy. For the most part, WUP landowners had conducted timber
57

harvests. These results indicate the large variation between Michigan regions, specifically in regard to
reasons for owning forestland.
The second paper explored specific sectors of NIPF owners based on their reasons for owning
forestland. Game Wardens represented 29.8% of respondents and were more interested in recreational
reasons for owning forestland compared to other groups. Timber Barons comprised 46.4% of Michigan
NIPF owners and were distinguished by the higher importance they placed on consumptive reasons for
owning forest property, such as producing timber or wood for energy. Finally, the last group was 23.8%
of Michigan NIPF owners and was represented by Tenants, who did not exhibit strong distinguishing
factors in regard to reasons for owning forestland. Each of these groups expressed different attitudes
towards various policies and incentives and performed management activities to varying degrees,
further illustrating the differences among Michigan NIPF owners.
As society continues to desire benefits from private forestland, we must continue to understand
the private forest owner in order to meet societal objectives. This study provided information which
should be useful to professionals who seek to influence private landowners. Where even distribution of
a policy or program participation is desired, the information in this study suggests common traits
among all ownership groups can be utilized to influence landowner response. Perhaps incentives or
programs which correspond with ownership reasons such as beauty or privacy (which are held by all
ownership groups) could be utilized to influence the greatest number of landowners. On the other hand,
where responsible timber harvesting is the goal, it may be more prudent to target only those who hold
land for “consumptive enjoyment”. Providing programs or incentives to all landowners might not be
the most effective use of limited public resources.
When we fail to understand those targeted by any policy, we are likely to be disappointed. Only
by seeking to understand the motivations behind NIPF owner behavior can we craft programs which
will pique interest and encourage enrollment or participation in management programs. As we begin
58

the second decade of the 21st century and seek to provide more ecosystem services from our state and
national forestland, from timber to carbon, more and more we will need to examine those private
landowners who control and manage forestland. Only when we seek to help NIPF landowners meet
their objectives, can we aid them in meeting ours.

59

REFERENCES

60

Finley, A.O. and D.B. Kittridge. 2006. Thoreau, Muir, and Jane Doe: Different types of private forest
owners need different kinds of forest management. Northern Journal of Applied Forestry: 22
(1). pp. 27-34.
Kluender, R.A., T.L. Walkingstick and J.C. Pickett. 1999. The use of forestry incentives by
nonindustrial forest landowner groups: is it time for a reassessment of where we spend our tax
dollars? Natural Resources Journal: 39 (4), pp. 799-818.
Potter-Witter, K. 2005. A cross-sectional analysis of Michigan nonindustrial private forest owners.
Northern Journal of Applied Forestry: 22 (2). pp. 132-138.
Sager, E. 2010.. Innovative means for understanding and engaging landowners. Panel presentation at
the Society of American Foresters National Convention, 2010. Albuquerque, NM. Society of
American Foresters, Bethesda MD.
Serbruyns, I. and S. Luyssaert. 2006. Acceptance of sticks, carrots and sermons as policy instruments
for directing private forest management. Forest Policy and Economics: 9(3), pp. 285-296
Tyrell, M. 2010. Innovative means for understanding and engaging landowners. Panel presentation at
the Society of American Foresters National Convention, 2010. Albuquerque, NM. Society of
American Foresters, Bethesda MD. Zhang, D. and W.A. Flick. 2001. Sticks, carrots, and
reforestation investment. Land Economics: 77 pp. 443-456.

61