Predictors of Land Privatization: Cross‐Cultural Tests of Defendability and Resource Stress Theory

Abstract

We have known for some time that complex societies are more likely to have land tenure systems based on private rights and less likely to have communal ownership. Less understood is why; more specifically, what are the mechanisms to explain why complex societies have more private property? What are the adaptive advantages of one system rather than the other? Conceptualizing and coding land tenure systems as a “bundle of rights,” this worldwide cross-cultural study suggests that Acheson’s (2015) economic defendability theory in conjunction with some environmental stressors, such as drought, may help us understand cross-cultural variation in land tenure systems. Our results have evolutionary implications. They suggest that if property rights were claimed, communal property systems would have been the default system for any society having substantial degrees of hunting, gathering, or herdable animals. Agriculture by itself is not a strong predictor of private land rights, although irrigation agriculture is.

Full text

[Accepted for publication in the American Anthropologist—forthcoming]
Predictors of land privatization: Cross-cultural tests of defendability and resource stress
theory
Carol R. Ember, Teferi Abate Adem, Tahlisa Brougham, and Emily Pitek
Human Relations Area Files at Yale University
Acknowledgments
This material is based upon work supported by the National Science Foundation under Grant
Number (#SMA-1416651). We thank Joseph P. Feldman for his coding of land tenure systems,
Rachele Pierro for translation assistance, Abbe McCarter for editorial assistance, and Ian
Skoggard for advice. We also want to thank the three anonymous reviewers for the time and
effort they put into their reviews and for their constructive criticisms.
Corresponding Author: Carol R. Ember
Carol.ember@yale.edu
1

Abstract
We have known for some time that complex societies are more likely to have land tenure systems
based on private rights and less likely to have communal ownership. Less understood is why;
more specifically, what are the mechanisms to explain why complex societies have more private
property? What are the adaptive advantages of one system rather than the other?
Conceptualizing and coding land tenure systems as a “bundle of rights,” this worldwide cross-
cultural study suggests that Acheson’s (2015) economic defendability theory in conjunction with
some environmental stressors, such as drought, may help us understand cross-cultural variation
in land tenure systems. Our results have evolutionary implications. They suggest that if property
rights were claimed, communal property systems would have been the default system for any
society having substantial degrees of hunting, gathering, or herdable animals. Agriculture by
itself is not a strong predictor of private land rights, although irrigation agriculture is.
………….
For more than two millennia, beginning from Greek and Roman times, scholars have been
intensely interested in conditions that led to the origin of different property regimes (Rudmin
1988, 1992; Engels and Hunt 1884). As summarized by Rudmin (1988), much of the debate
concerned whether landed property should be held privately by household heads and entitled
heirs or communally by larger groupings. Although Aristotle compiled data on selected societies,
it was not until the early 20th century that the first systematic worldwide cross-cultural
comparisons on land tenure systems were conducted (Hobhouse, Wheeler, and Ginsberg 1915,
Sumner and Keller 1927 and Simmons 1937). After a considerable hiatus in cross-cultural work
on property, these studies were followed by Swanson (1960), Pryor (1977, 2003, 2005) and
Rudmin (1992a, 1992b, 1988, 1995, 1996). Generally speaking, these studies found associations
between some aspect of social complexity and private property systems. Although the present
study continues this cross-cultural tradition, it differs from some of the earlier work in the
following ways: 1) it explicitly measures land tenure excluding other types of property (e.g.,
livestock, objects) or consideration of inheritance; 2) it builds on a more nuanced view of land
tenure systems as “bundles of rights,” each of which can vary on a continuum; 3) it considers
multiple predictors together; and 4) it takes into account environmental and other stressors along
with other social and economic conditions. We have taken inspiration from Acheson’s (2015)
pioneering comparison of selected societies in which he argued that the concept of economic
defendability was critical for understanding privatization. Our goal is to improve on his work by
identifying some objective and clearly measurable indicators of defendability and to use a
standard sample of societies to increase the chances that results would be generalizable.
All societies are unique and also comparable if we focus on dimensions of variation (Ember and
Ember 2009). If one focuses on differences, then land tenure systems—broadly defined as the
terms in which rights in land, including internal water bodies (rivers, lakes) and other natural
resources that provide the basis for human livelihoods, are accessed, altered, controlled and
transferred—can be thought of as quite complex and variable (Bruce and Migot-Adholia 1994).
Alternatively, if one looks more broadly, then one can ask why land tenure systems should be
communal or private. But we can choose mid-way ground, as we do here. More specifically, we
create more detailed measures to better capture existing variation. Following Bruce (1986),
“bundles of rights” in land include the rights to use land, to alter land, to exclude others from the
land, and to transfer land. This understanding allows us to rate each type of right separately along
2

the private to communal dimension. Furthermore, if each individual right can be characterized as
more robustly private or less robustly private, we can also create a summary “robustness” score
that characterizes the degree to which the bundle of rights together tends towards being more
communal or private.
Understanding variation in land tenure systems is not just an academic exercise. Policymakers in
many developing countries have long advocated for privatizing communal land tenure systems.
The presumption was that Western-style private ownership is necessary for economic growth and
modernization (Feder and Noronha 1987; Berry 2002). But this push is usually done without
understanding the role that particular land tenure systems play in people’s ability to successfully
deal with vagaries in their environments (Ribot and Peluso 2003). For example, research in
Africa and elsewhere suggests that communal land holding systems provide needed flexibility
for adjusting to demographic and resource changes (e.g., Lentz 2013; Berry 1993; Peters 1994).
And reduction of mobility brought about by central governments has degraded the resiliency of
savanna ecologies (Scoones, and Kerven 1993; McCabe 1987) as well as the capacity of local
communities to regulate resource use (Little 2003). Given that climate change-induced hazards
are expected to increase in frequency and severity (Team, Pachauri, and Meyer 2014), pushing
populations to privatize may be especially problematic.
While we recognize that there may be a large number of factors affecting property rights, we
focus here on two main parameters--the degree to which land is defendable and possible impacts
of resource stressors and environmental uncertainty. Our study employs both hypothesis-tests as
well as post-hoc exploration of possible predictors. This mixed strategy results from the fact that
our initial research design, which focused on the possible effects of natural hazards on land
tenure systems (as part of a larger grant project on how natural hazards may transform culture),
was not supported in expected ways. Our more exploratory analyses are more consistent with
“economic defendability” theory (Dyson‐Hudson and Smith 1978; Cashdan et al. 1983; Chabot-
Hanowell & Smith 2013; Acheson 2015). We find that natural hazards and other forms of
resource stress may play some role in shaping property systems, but not to the degree and
manner we initially expected.
Theoretical Expectations
Hardin’s (1968) “The Tragedy of the Commons” espouses the view that communal systems are
generally inefficient and maladaptive. However, subsequent research by scholars has found that
Hardin wrongly assumed communal systems are “open” for everyone to access and use. His
critics convincingly showed that “the tragedy of the commons” does not hold when communally
sanctioned regulations govern use and access (Bromley et al. 1992; Ostrom et al. 2002). If this is
so, then communal systems may be adaptive in certain circumstances; private systems more
adaptive in others. So, the question is, under what circumstances are communal systems more
adaptive? As we discuss below, some of the theoretically posited conditions are: difficulty of
defendability and drought.
Building on prior work by Cashdan et al. (1983), Dyson-Hudson and Smith (1978, 23), and
Chabot-Hanowell & Smith 2013, 1 Acheson (2015) theorizes that economic defendability in the
face of scarcity or competition is critical to understanding why communal property regimes
change to private regimes. He goes on to state: “When the costs of defending boundaries are
3

relatively low and the value of resources in an area is high, people have a strong incentive to
generate a private property regime. When the costs of exclusion are high relative to the value of
resources, the probability of a common property regime being developed or maintained is high
(Acheson 2015, 29).” Looking at both sea and land property systems in a purposive sample of
societies that had communal property systems at least at some point in their histories, Acheson
claims support for economic defendability theory. While we acknowledge Acheson’s pioneering
work, we believe that it is important to: 1) systematically test economic defendability theory
using a non-purposive sample of societies; and 2) employ objective measures of defendability.
In the absence of strong state enforced property laws or other cultural mechanisms, economic
defendability theory suggests that the incentive to establish exclusive private rights over a given
tract of land stems from the economic value of the resources one seeks to reap from that land
(Acheson 2015). If a social group, whether constituted as a territorial band, village community,
lineage corporation or other kinship-based group, is moving from place to place throughout the
year, presumably to obtain an adequate food supply, one can surmise that the economically
valuable resources people seek in the land are better defended collectively. Therefore, we expect
that frequent mobility and absence of year-round permanent settlement will favor communal
property because the cost of defending large territories by individual households is prohibitive.
In addition, a collective group could serve as a relatively large armed force against external
intervention, while also providing enforceable cultural rules for dividing and transferring land
rights within the group.
We adopt a broad evolutionary view and assume that while not all individuals and/or social
groups make decisions and adopt behaviors that are adaptive, in the long-run behaviors
(including customary behaviors), that are adaptive for particular social and physical
environments will generally become more prevalent in those environments. Whether we use
principles from individual or group selection2, individuals and/or social groups practicing
maladaptive behaviors should have lower reproductive success. If individuals or social groups
practicing more adaptive behaviors reproduce more successfully and are imitated by others,
behaviors can become typical or customary for a larger cultural group. Our concern here is not
with the specific evolutionary mechanisms that might favor communal or private land tenure
systems, but rather with what varying conditions might favor one more than the other.
We have proceeded in a stepwise fashion, first asking what conditions may favor mobility or the
lack thereof. Then we ask how much mobility and/or predictors of mobility predict property;
then we add some environmental variables.
As for what may favor mobility, it is widely agreed that hunter-gatherers and pastoralists tend to
employ mobility strategies to access dispersed resources, so these subsistence strategies are an
obvious choice to consider as predictors of communal land tenure. More specifically, we suggest
that it is high dependence on hunting and/or high dependence on gathering (excluding fishing)
that favors mobility; pastoralists, particularly those in arid and semi-arid ecologies dependent on
grazing and browsing animals need frequent, and sometimes rapid mobility to obtain adequate
pasture and water for their herds (e.g., McCabe 1990) in addition to extracting resources through
other subsistence strategies (Salzman 1971). In contrast, we believe that irrigation should be
particularly predictive of permanent year-round settlements because irrigation channels and
irrigation systems involve considerable labor input and have to be maintained (Geertz 1980).
4

Communal land tenure is believed to provide needed flexibility in drought-prone environments
(e.g., Berry 1993; Peters 1994). With communal rights, migrants who temporarily lose their land
rights could usually claim them later during recovery. In pastoral societies, mobility in large
tribal territories provides a way of moving to other pastures when drought strikes (McPeak,
Little, and Doss 2012; Fratkin and Roth 2006). Nugent and Sanchez (1999) found empirical
support for the relationship between drought and communal ownership in a regional cross-
cultural comparison of 41 Sudanese societies. Land tended to be held communally in drought-
prone societies and conversely, societies in rainfall sufficient areas developed strongly private
land ownership. But, hazards that destroy food supplies are broader than droughts and may
include flooding, storms, and pest invasions. Generalizing from these findings on drought, we
expected that a society’s culturally preferred land tenure choice, along the private vs. communal
continuum, would be shaped to some extent by its cumulative experiences of responding, and
adapting, to past food-destroying natural hazards—more specifically, in risky ecologies with
frequent hazards, we expect more flexible and broadly inclusive communal land tenure.3
Based on the literature discussed above, we address the following main research questions: 1)
Does economic defendability theory help explain private versus communal property systems in
land or robustness of individual land rights? 2) Does resource stress and environmental stress
increase the likelihood of communal property?
Methods
This section outlines our sample, measures of our variables (those we newly coded and those
previously coded by others) , and analysis strategies; details are provided in the Supplement (S).
Sample and General Coding Procedures
As mentioned earlier, this research was part of larger project exploring possible effects of natural
hazards and other resource stressors on culture, including land tenure systems. Land tenure
variables were newly coded for this project; we also made use of new climatological measures of
drought compiled by climatological colleagues (see S1 and S.5.3 for details about the new
measures). All of the other variables were based on previously coded information. Our starting
sample was the Standard Cross-Cultural Sample (SCCS) of 186 largely nonindustrial societies;
the sample was constructed to maximize historical independence of cases (Murdock and White
1969). Because frequency of natural hazards was an essential part of our initial research design,
we only coded land tenure variables for the SCCS societies that had been reliably coded on
natural hazards; this resulted in 89 societies coded for land tenure where the sample societies are
briefly described with the SCCS identifier, the SCCS name, the ethnographic present, and the
place focus (see S.9). But, where data on natural hazards are not needed, we have used
additional codes of property systems made by Pryor (2005, 2011) and the CONAN project (Lang
1995) for the SCCS sample. This gives us a way to expand the sample size to 133 societies for
some analyses.
While coding our land tenure variables, we adhered to the same time and place foci previously
used by Ember and Ember (1992a, b) to code natural hazards. Wherever possible we used
5

eHRAF World Cultures (http://ehrafworldcultures.yale.edu) or the HRAF Collection of
Ethnography (in paper) for our source material and Ember’s (2007) concordance that specifies
document matches between the SCCS time and place focus
(https://hraf.yale.edu/resources/reference/sccs-cases-in-ehraf/ provides updates for document
matches.)
Dependent Variables—Land Tenure
In this paper we concentrate on two measures of land tenure constructed from our own newly
coded data (for details see S.1). The first contrasts whether the typical land-holding unit that
controls customary legal claims over land or territory is primarily a family household unit (which
we consider private ownership) or is a communal unit (larger than an extended family such as a
corporate descent group, another type of kin group, or a residence-based group). The second
sums the degree of private versus communal control across three different land tenure rights—
rights to alter (L5), rights to exclusive use (L6), and rights to transfer (L7). The summary score
is labelled “robustness of individual land rights.” To increase the sample size for some of our
analyses we combined our coded variables with previous property codes by Pryor (2005, 2011)
and the CONAN (Lang 1995) project (see S.2-4).
Independent Variables
Mobility and Subsistence Strategies
The information on mobility comes from Murdock and Provost (1973) and contrasts societies
with year-round permanent settlements versus nonpermanent settlements (see S.5.1) To measure
the degree of dependence on hunting and gathering (fishing excluded) we used data from the
Ethnographic Atlas (originally column 7; see Ethnology 1962ff) that estimated the degree of
dependence on each of five types of subsistence--gathering, hunting, fishing, animal husbandry,
and agriculture. The scores for all five types sum to 10, so a summary score of 5-10 for gathering
plus hunting represents about 50-100% dependence on hunting and gathering. Note that we only
used data from the Atlas that matched the time and place focus given in the SCCS. For a
dependence on herding animals, as opposed to other kinds of animal husbandry, we used a
combination of degree of dependence on animal husbandry (5-10) in conjunction with the Atlas
variable on the main type of domesticated animal. More specifically, we considered high
herding societies to have scores of 5-10 on dependence on husbandry (omitting societies
depending heavily on pigs or small animals—column 39 of the Atlas) and contrasted them with
societies with a lower dependence on animal husbandry and/or a high dependence on pigs or
small animals.4 We also used the Atlas variable “type and intensity of agriculture” (column 28)
to contrast intensive irrigation agriculture societies with those that had other types of agriculture
or little or no agriculture (see S.5.1 for details).
Resource Stress Based on Ethnographer Reports
We employ four measures of resource stress based on ethnographic reports: 1) the threat or
actuality of natural hazards that seriously disrupts food supply; 2) the threat or actuality of
famines; 3) a combination of the first two measures (hazards and famine), and 4) a factor score
6

on famine variables that we created from three Dirks’ (1993) variables (see S.5.2). The codes for
the first two came from Ember and Ember (1992a,b), the third combined measure was
constructed from 1 and 2 (see details in Table S3 in S.5.2).
Drought
A team of climate researchers working with us has developed a series of drought measures based
on monthly gridded climatic data from the Climatic Research Unit for drought events from 1901
to 2009 (Felzer, Ember, Cheng, & Jiang, under review). Using slightly different methods, two
measures were based on precipitation minus evaporation, a third was the Palmer Drought
Severity Index (PDSI). After standardizing for biome, each measure estimates whether or not a
drought occurred in a particular year and a proportion score was created for how many years
droughts were estimated to occur (divided by the number of years) for each of the 98 SCCS
cases in our beginning sample. To have a more conservative measure, we estimated drought for a
particular year only when all three drought measures concur (Combined Drought Measure). In
addition, we also employed a measure of the inter-annual coefficient of variation that indicates
the degree to which rainfall varies from year to year (see S.5.3). While we would have ideally
preferred to use weather data for the same time period as the hazard and famine measures, we
would have had to drop too many cases. We note however that many of these indices did not
vary substantially over different time frames.
Modernization Variables
Because state governments commonly try to settle people down and impose Western land tenure
systems including land title registration (e.g., Scott 1998; Berry 1993), we would ideally want to
know how much modernization affected land rights in focal communities. Lacking a direct
measure, we used a few “modernization” codes developed by Divale and Seda (2000; retrieved
from Divale 2004) in our multivariate analyses. Although these variables were not about property
rights per se, they give us a general idea of degree of change in a few domains of life. These
included agricultural improvements (Agriculture 3.1-v1812), changes or implementation of a
foreign judicial system (Government, Political, and Legal System 5.2-v1819), and large scale
projects introduced by outside agencies (Government, Political and Legal System 5.4-v1821).
Analyses
For all bivariate relationships we first assessed whether the relationships looked linear using box
plots or scatterplots. For those linear relationships we have compared two types of correlation
coefficients—gamma and rho. This comparison gives us a rough idea of whether the relationship
is asymmetric or symmetric. The gamma coefficient goes to +/-1.00 with one empty cell or
quadrant (Weisberg, Krosnick, and Bowen 1996) and rho goes to +/-1.00 with two empty cells or
quadrants. If the gamma is close to one and the rho is much lower, the pattern is asymmetric and
is consistent with a multicausal model where a predictor is either necessary or sufficient, but not
necessary and sufficient. If the gamma and rho are roughly similar, symmetry is suggested and
this pattern would be consistent with single-causal model (the predictor is necessary and
sufficient). To evaluate the possible independent effects of a number of independent variables we
use logistic regression when the dependent variable is dichotomized as communal or private.
When the dependent variable has a large number of ordinal scale score positions we perform two
types of analyses: linear multiple regression (see Labovitz 1970 for justification) and ordinal
7

logistic multiple regression. Because the modernization variables were not available for that
many societies (and reduced our sample size considerably), we introduced the variables only in a
next step in the regression models to assess whether the modernization variable added
significantly to the model. None of the modernization results were significant (not shown).
Reproducibility Statement
All data are available in the Data Repository section of hrafarc.org [Will be made available upon
publication.] Both R and SPSS were used for our analyses. Figure 1 was created with the R
package ggplot2 (Wickham 2016). Linear model summaries and standardized betas were
extracted using lm.beta (Behrendt 2014). The package car (Fox and Weisberg 2019) was used to
determine Variance Inflation Factors (VIF). Gammas were calculated with MESS (Ekstrøm
2019), and Pearson and Spearman correlations were calculated using DescTools (Signorell et. al.
2019). To obtain information on AIC values, AICcmodavg (Mazerolle 2019) was used. Ordinal
logistic regressions were created using MASS (Venables and Ripley 2002). SPSS 26 was used
for the binary logistic regressions and the Kruskal-Wallis tests.
Results
Defendability and Communal vs. Private Property
Our first step was to see if there is support for the idea that mobility would make it hard for an
individual or household to defend land and therefore lessen the likelihood of private ownership.
Indeed, mobility is a strong asymmetric bivariate predictor of communal (versus private)
property when we contrast some mobility versus year-round permanent settlements (the
respective gamma and rho are: -.745, -.370, p < .001, N=88). Only 4 societies with non-
permanent settlements have private property, whereas all the other combinations have 24-32
societies, consistent with the idea that non-permanence is a sufficient predictor of communal
property. An even higher gamma (-.861, p < .001, N=133; the rho is -.457) is found when we
combine our coded data with data from other samples (see S.2 to S.4 for how samples were
combined). Supporting the connection between mobility and more territory is the high degree of
asymmetric relationship between our measure of territorial size (see S.7) and mobility (gamma=
-.742, rho= -.574, p < .001, N =88). Virtually all the societies with small territories consisting of
a homestead with surrounding fields, pasture and woodland are privately owned.
Our second step is to ask what predicts mobility. Earlier we suggested specific subsistence
practices that would predict high mobility: high hunting and gathering (excluding fishing) and
high dependence on herding animals. The asymmetry is particularly pronounced when we relate
high hunting and gathering dependence of roughly 50% or more and correlate it with non-
permanent versus year-round permanent settlements (gamma = -.947, rho = .507, p< .001, n = 88
in our sample; gamma = -.974, rho = -.580, p < .001, N =133 in the combined sample). Virtually
all the societies with high dependence on hunting and gathering (15 of 16 in our sample) lack
permanent year-round settlements (that is, are nomadic, seminomadic, or semisedentary). Similar
to hunting and gathering, high dependence on herding animals is a strong asymmetric predictor
of some mobility versus year-round permanent settlements (gamma = -.889, rho = -.358, p < .01,
N=88 in our sample; gamma = -.859, rho = -.265, p < .01, N = 133 in the combined sample). Of
the 10 high herding societies in our sample, only one lacks permanent year-round settlements. 5
We also expected that irrigation would be a negative predictor of communal property. Indeed, the
presence of irrigation is an asymmetric negative predictor of mobility, albeit not as strong as
either hunting and gathering or high dependence on herding animals (gamma = .660, p < .01; rho
8

= .268, p < .05, n = 88 in our sample; gamma = .720, rho = .295, p < .001, N = 133 in the
combined sample). In our sample, 16 of 19 societies with irrigation have permanent year-round
settlements. We analyzed our results using binary logistic regression for the dichotomized land
tenure variable (private versus communal). For the “robustness of individual land rights” scores
we used multiple regression and ordinal regression. We primarily discuss the results regarding
the “robustness” score here and show the binary logistic regression results for the private versus
communal score in S.6. Since there are no cases where high dependence on hunting and
gathering and high dependence on herding animals both occur, we created a combined variable
where one of the conditions is present versus none of the two conditions are present to include in
the tested model along with irrigation agriculture.
Defendability and Robustness of Individual Land Rights
Our “robustness of individual land rights” score not only gives us a more nuanced way of
assessing the degree to which rights are private or communal, but it also provides us with
analytical advantages in that the dependent variable has more scale points. We also use data
from (Pryor 2005, 2011) because he also measured robustness of individual land rights (see S.4
for how we combined his codes from different types of economies). Finally, to get a larger
sample we transformed our scale to merge with Pryor’s (see S.2 and S.4). This combined result
(column 3 of Table 1) must be viewed more cautiously for the multiple regression because the
dependent variable for the combined scale has only 4 scale positions. But this combined result
(column 3) meets the requirements for ordinal regression. All the models in Table 1 (cols. 1-3)
show significant effects of mobility—year-long permanent settlements predict more robust
individual land rights. Moreover, high dependence on herding animals or hunting/gathering also
significantly predicts less private land rights. Irrigation is less consistent and only significantly
predicts more private rights in our sample (col. 1) and in the combined sample using ordinal
regression (col. 3). Note that when the dependent variable is robustness of individual land rights,
in contrast to private vs. communal, both mobility and high herding or high hunting and
gathering both independently predict more communal rights.6 Similar factors predict larger
territories (see S.7).
Table 1. Multiple regression models predicting “robustness of individual land rights.” The p-
value in parentheses reflects the p-value of the estimates’ log-likelihood in an ordinal regression
(see Figure S1 for effect sizes and S.8.2.1 for ordinal regression details).
(1) (2) (3)
irrigation .192*(**) .049 (n.s) .123+a (*)
9

Mobility (Non-Permanent=0
vs. Permanent=1)
.367**(**) .396***(***) .303**(**)
High Hunting/Gathering =>5
or High Anim.Herding=>5
(Present=1; Absent=0)
-.281*(*) -.337**(**) -.326**(**)
R2.45*** .483*** .397***
adj R2.430*** .467*** .382***
N 88 98 123
+p<.10, *p<.05, **p<.01, p<.001; all two-tailed unless marked with a superscript “a” indicating
one-tailed
(1) Dependent variable from our sample (Com567)
(2) Dependent variable constructed from Pryor’s robustness score (see S.2)
(3) Dependent variable is a combined measure of our cases and Pryor’s cases (see S.4)
Hazards, Other Resource Stressors and Drought
Contrary to our initial expectations (not shown), we found no significant linear relationships
between natural hazards and property (private or communal) or robustness of individual rights.
However, exploratory evidence suggests a curvilinear relationship between hazards/resource
stress and robustness of individual land rights (shown in Figure 1). In this analysis we use a
combined variable of hazards and famine (S.5.2) that presumably reflects the seriousness of
hazards. While this relationship is suggestive of a curvilinear relationship with robustness of
individual rights (Com567), it is not significant by a Kruskal-Wallis comparison of groups
(p=.193) or by a quadratic regression model (p= .12). However, using Ember and Ember’s
(1992b) measure of famine (dichotomized as none vs. threat of or an actual famine), we find a
somewhat asymmetric relationship between higher threat of famines and more communal (vs.
private) ownership (gamma = .475, p < .05, one tail, N= 61; rho = .210, p < .10, one tail, N=61),
providing partial evidence of a relationship between famine and more communal ownership as
predicted. However, when we look at a factor score based on three of Dirks’ (1993, 2004) famine
measures we do not find a significant relationship with robustness of individual land rights. We
know that mobility is a very strong predictor of communal ownership, so mobility could be
obscuring a relationship. Indeed, if we just look at societies with year-round permanent
settlements, we find that the relationship is now significantly linear in the predicted negative
direction (gamma=-.24, p < .05, one-tailed, n=29; rho = -.344, N=29, p < .05, one-tailed). But
interestingly, the result for non-permanent settlement societies is in the non-predicted positive
direction (gamma=.343, p <.10, two-tails; rho = .401, N=20, p <.10, two tails), albeit only
marginally significant. Figure 2 shows the lines of best fit differentiated by the green circles
representing non-mobile settlement societies and the red circles representing mobile settlement
10

societies. We note here that “mobile” societies vary considerably on mobility from nomadic to
sedentary with impermanent settlements. Five of the 6 societies in the lower left corner are fully
nomadic or semi-nomadic and moreover have either high herding or high hunting and gathering.
Thus, their more communal rights might be better explained by mobility and subsistence, rather
than famine. In sum, the picture regarding famine is more complex than we imagined.
Figure 1: Violin and box plot of the relationship between robustness of individual land rights
(Com567) and hazards modified by famine. Horizontal bar within the box plot represents the
median, with the lower and upper hinges corresponding to the first and third quartiles (25th and
75th percentiles, respectively). Upper whisker extends from hinge to largest value (within 1.5 *
IQR from hinge); lower whisker extends from hinge to smallest value (within 1.5 * IQR of
hinge). IQR= interquartile range, or distance from first and third quartiles. Data outside the
whiskers are "outlying" points and plotted individually.
11

Figure 2: Scatterplot of relationship between robustness of individual land rights and a famine
factor score with linear relationships plotted by variation in mobility (green = permanent
settlements; red = non-permanent settlements)
Drought is more clearly related to communal land tenure systems. Table 2 shows the correlations
between the measure of robustness of individual land rights with measures of yearly rainfall
variation (inter-annual coefficient of variation) or drought (the last four rows) in our sample.
Four of the 5 measures are significant or marginally significant and the combined measure is the
strongest of the five (r= -.272) indicating that the presence of drought may work against
privatization. We use the combined drought measure in subsequent multiple and ordinal
regression analyses.
Table 2. Relationships between measures of drought and unpredictability of rain and robustness
of individual land rights (Pearson Rs)
interannual coefficient of
variation in precipitation
-.166 +a(N=88)
P-E_th -.262* (N=88)
P-E_pm -.200*a (N=88)
PDSI -.047 (N=86)
Combined Drought Measure -.272*(N=86)
+p<.10, *p<.05, **p<.01, p<.001; all two-tailed unless marked with a superscript “a” indicating
one-tailed.
12

Does drought add predictability to the regression models shown above in Table 1? As mentioned
above, we only have drought measures for our sample. In Table 3 we show the standardized beta
coefficients for the multiple regressions along with the p-values for the regression; the ordinal
regression p-values are shown in parentheses and the full ordinal regression results are shown in
S.8.2. Looking first at column 1 of Table 3, we see that neither in the multiple regression result
nor in the ordinal regression result does drought add significantly to the model on its own, so we
explored whether the presence of irrigation with permanent water sources (rivers or wells) may
detract from the predictive effect of weather-caused drought. After all, with a viable water source
the effect of drought on crops and herds should be mitigated. If we remove societies with
irrigation and reanalyze the effect of drought (see Table 3, column 2), the result for drought in
the multiple regression improves somewhat and the ordinal regression result for drought is now
significant. Because the result with just irrigation societies is not significant and is in the
opposite direction (not shown), it appears that there is an interaction effect depending on the
status of irrigation. Accordingly, after standardizing both drought and irrigation, we created an
interaction term of irrigation by drought. The model including the interaction term is shown in
column 3 of Table 3. Looking at the ordinal regression results first (p-values in parentheses), the
drought by irrigation interaction term is significant as is the combined drought measure (one-
tailed in the direction predicted). The multiple regression VIFs for the subsistence term and
mobility are 1.878 and 1.926 suggestive of some multicollinearity so in columns 4 and 5 we
remove either the subsistence variable or mobility. However, the results do not change
substantially. The fact that the interaction term is significant or marginally significant is
consistent with the idea that drought, in the absence of irrigation, may also be an important,
albeit modest, predictor of more communal rights. The R-squared results (and the adjusted R-
squared results) tell us that all the models are predicting a substantial amount of variance in the
robustness of individual land rights, predicting at least 40% of the variation in the dependent
variable. The AIC values, change in AIC and the AIC weights suggest that Model 3, which
contains all the variables, is the best model.
Table 3. Multiple regression models predicting “robustness of individual land rights” adding
drought to the models in Table 1.The p-value in parentheses reflects the p-value of the estimates’
log-likelihood in an ordinal regression (for details see S.8.2.1).
(1) (2)a(3) (4) (5)
irrigation .19*(**) NA .21*(**) .22*(**) .26**(**)
Mobility (Non-
Permanent=0 vs.
Permanent=1)
.36**(**) .37**(**) .33**(**) .50***(***)
13

High
Hunting/Gathering
=>5 or High
Anim.Herding=>5
(Present=1;
Absent=0)
-.26*(*) -.32**(**) -.26*(*) -.46***(***)
Combined drought
measure
-.09(+b)-.15+b (*) -.09(*b) -.12+b(*b) -.12+b(*b)
Drought by
Irrigation
.14+(*) .15+(*b) .17+(*)
R2.455*** .474*** .474*** .438*** .416***
Adj R2.428*** .450*** .441*** .411*** .387***
AICc 358.66
(460.11)
c 357.99
(457.90)
361.23
(462.27)
364.60
(463.03)
Δ
AICc .67
(2.21)
c 0
(0)
3.24
(4.37)
6.61
(5.13)
AICc Weight .37
(.22)
c .51
(.66)
.10
(.07)
.02
(.05)
N 86 68 86 86 86
+p<.10, *p<.05, **p<.01, p<.001; all two-tailed unless marked with a superscript “b” indicating
one-tailed
aSocieties with irrigation are removed
c Due to having a different n-value, this model’s AICc values should not be compared to those of
columns 1 and 3-5
To see if modernization may have affected our results, we added each of the modernization
variables (one at a time) to the regression models shown in columns 4 and 5 in Table 3. None of
the modernization variables added significance to the regression equations (not shown).
14

Summary
Using defendability theory, we explored a number of specific conditions that we thought would
particularly favor communal ownership. These factors include: settlement impermanence, high
dependence upon hunting and gathering (not fishing)7 and high dependence on herding animals.
We suggest that all these conditions not only increase the need for large and hard to defend
territories, but they also decrease the value for an individual to exert claim to a particular parcel.
In contrast, we suggested that irrigation would favor private ownership because it makes small
parcels of land more valuable over time. Indeed, our cross-cultural results generally find that all
of these conditions are strong or moderately strong (irrigation) asymmetric predictors and the
distributions suggest they may be sufficient predictors of communal property systems.
Combined regression models using these four predictors and either a dichotomous dependent
variable (communal versus private) or “robustness of individual land rights,” predict
approximately 40% (range from .43 to .46 of the variance depending upon the model), lending
support to defendability theory. Although we initially expected that riskier and more hazard-
prone environments would favor the flexibility of communal ownership linearly, aside from
drought (see Tables 2, 3, S8, S9) other measures (natural hazards, famine) did not consistently
predict communal ownership in the expected way. Exploratory analyses suggest that famine may
have a more complicated relationship to land rights, inasmuch as the relationship is in the
opposite direction when comparing non-mobile societies and mobile societies. Only in non-
mobile societies does more famine predict more communal rights.
Discussion
Before turning to some limitations of our research and some unanswered research questions, we
discuss a couple of societies (Burusho and Havasupai) that score in the middle of the range for
robustness of individual land rights. These cases suggest future avenues to pursue to increase our
understanding of land tenure systems.
The Burusho of northern Pakistan have both private and communal land rights apparently
because of two different subsistence activities. All grazing pastures, typically on higher ground,
were collectively owned by patrilineages, while all agricultural lands in the lower zones were
privately owned by individual households (Sidky 1993: 159). The mixed pattern conforms to
one of the principles we have laid out--territory for herding tends to be communal. The Burusho
lack irrigation, so we have not correctly predicted private ownership of land. However, they do
make other investments to make their land arable such as terraces, retaining walls and conduits
(Sidky 1993: 193). Future research should consider other labor investments that might have the
same effect as irrigation.
The Havasupai of the Grand Canyon region of the U.S. also have a mixed land tenure system (in
pre-reservation times) and, like the Burusho, do not fit one of our predictions. In this case, they
practice irrigation but land rights are not private. But, irrigated plots are not that dependable.
Leslie Spier (1928) reports that land was collectively held by communities and household heads
obtained the right to cultivate particular fields through their community membership. Household
heads also had some private rights, which included the ability to plant fruit trees and exclude
other community members from taking resources without permission. However, a household’s
ability to exercise these rights strongly depended on its ability to use the land actively and
continuously. Plots were largely abandoned due to the vagaries of flooding along what Spier
15

referred to as the banks of the “marauding creek (1928:230).” Periodically, following a big flood,
the stream would change its level or position, often wiping out some fields and leaving others
beyond the reach of irrigation. To recoup, households needed to start new plots on uncleared
lands or make use of previously abandoned patches now accessible for irrigation. This example
helps explain why irrigation may not always predict private ownership. Given moderate hazards
in the form of river flooding, completely private rights would put many households in serious
jeopardy if the land was no longer able to be irrigatable. Private ownership for fruit trees might
be explained by the fact that fruit trees are relatively long-lived and have more permanent value
than land subject to river flooding. We surmise that if it were not for periodic river changes,
irrigation would normally have pushed land rights to be much more private.
In any research, there is a practical need to both limit the theoretical parameters examined and to
tailor measures to fit the data that are available. A worldwide cross-cultural study, such as the
one we have undertaken, has the advantage of generalizability, but refined data on ecological
variables (such as the spatio-temporal variability of resources or drought, such as discussed by
Baker [2003], Cashdan [1983], Dyson-Hudson & Smith [1978], Jaeggi et al. [2016], Wrangham
[1980] ) are usually not available on a wide scale. Similarly, the approach suggested by Adger
(2000), Folke (2006) and Folke et al. (2010) to consider the resilience of human systems in
varying in environmental conditions require richer ecological and socio-economic data. In
addition, as our Havasupai discussion indicates, other forms of intensification besides irrigation,
such as arboriculture, may favor private family ownership, but our sample lacks a sufficient
number of societies with arboriculture and other types of long-term resources or infrastructure
(such as fish weirs) to provide adequate tests. Other permanent structures, such as terracing,
practiced by the Burusho, also require analysis. Terracing on a broad scale probably requires
communal labor, but future research could investigate how terracing predicts property rights.
Some of these questions require smaller-scale comparisons limited to societies where researchers
have collected rich ecological descriptions; others that focus on less common intensification
practices require different sampling strategies to get sufficient numbers of cases.
The Burusho case raises the issue of other intensification practices that might push towards
privatization besides irrigation (and irrigation is not as strong a predictor of private land rights
compared to mobility, high dependence on hunting and gathering, or high dependence on herding
animals, as predictors of communal land tenure). But our data do suggest that a transition to
agriculture alone is not likely a sufficient condition for privatization. First, about 25% of the 186
societies in the SCCS that rely on agriculture for at least half of their subsistence do not have
year-round permanent settlements. Second, many agriculturalists also rely substantially on
herding, hunting and gathering as secondary activities, which we have noted increase the need
for larger territories and presumably some defense of those territories.8
Some of our results regarding food-destroying natural hazards and other resource stressors such
as famine were puzzling. Why should drought in particular appear to predict more communal
ownership, but an overall measure of number of natural hazards does not? Why does famine
appear to predict private rights differently in societies with permanent vs. non-permanent
settlements? And why is there a possible curvilinear relationship (Figure 1) such that moderate
degrees of hazards/famine predict more private rights than low hazards and more serious hazards
and famine?
16

First, with regard to why drought may linearly predict land tenure systems, but an overall
measure of hazards does not, we should consider the possibility that different types of hazards
have different effects. In our data, drought is the most common food-destroying hazards; flood is
the next most frequent. Drought is a relatively slow-onset hazard that often has widespread
effects, including soil degradation over a broad region and lowering of productivity, necessitating
mobility and/or claiming larger and diverse tracts of land. If so, drought may be a factor that
makes land less defendable. Floods are often quick-onset events that primarily affect river basins
and coastal areas; while some crops may be damaged and some regions show excessive soil
erosion, other areas, such as broad swaths of nutrient deposits from river overflow, may see the
land gain fertility. Subsequent research should look at types of hazard in more detail.
Second, with regard to the curvilinear relationships suggested by Figure 1, it is possible that in
risky environments (with crop-destroying pests, seasonal droughts, or occasional famines) there
may be adaptive advantages in allowing households to innovate to try to increase productivity. In
agrarian societies these strategies might include applying manure and other inputs, rotating
crops, fallowing plots, constructing erosion-slowing terraces, windbreaks, etc. Additional
strategies include constructing water reservoirs, digging wells, building irrigation schemes, and
constructing flood and landslide management walls. While some of these solutions (such as
terraces, reservoirs, irrigation) require large-scale cooperation and therefore might favor
communal land ownership, households might nonetheless derive maximum benefits with some
degree of autonomy in undertaking their day-to-day and season-to-season farming activities. But
when the risk is very large and includes severe or frequent famine, leaving households on their
own may be maladaptive, particularly if people have to migrate. With collective ownership
people are more likely to retain their ability to return through kinship or residential connections.
Third, with regard to famine and property rights, we noted in the results section, that the
relationship between famine and robustness of individual rights appears negative as predicted
when settlements are permanent (Figure 2). In societies that are somewhat mobile the result
appears to be the opposite. However, degree of mobility may still be the key and we suggest that
much of the “opposite” relationship can be explained by the fact that the most mobile of these
cases tend to be communal, regardless of their famine vulnerability. Our sample size is not large
enough to look at each type of mobility separately and a firmer understanding requires a larger
sample.
The whole point of establishing land rights is to exclude competing claims, reduce risk of
conflicts over unclear rights, and set conducive conditions for creating assets on it. There would
be no need for defendability without concerns, real or imagined, about unwanted intervention by
competing claimants. Therefore, it is important to ask who would help defend land in a
competitive situation. In non-state societies, the most likely individuals to come to your aid are
either kin or community members, either of which can act as a defensive body. Not all societies
with a unilocal rule of residence have unilineal descent, but warfare is a likely catalyst for
unilineal descent, perhaps because unilineal descent provides non-overlapping and unambiguous
sets of males who can provide a defensive force (Ember, Ember, and Pasternak 1974). In bilocal
or multilocal societies, residential entities become a more likely entity for collective defense
since even if unilineal groups are present, the community will not be comprised of a homogenous
kin group. If kin groups or communities act defensively to protect land resources then such
groups are likely to claim ownership of those resources, at least to some degree. We do know that
17

much of the warfare in preindustrial societies involves victors taking the land of the defeated
(Ember and Ember 1992a) which suggests that some groups tried to defend their claims, but to
test economic defendability further it would be important to assess whether, and to what degree,
land is actually defended and by what social entities.
Conclusion
It has been known for some time from previous cross-cultural research that private property is
more common in societies with higher levels of complexity, but the theoretical explanation for
this general tendency has not been very clear. We have found Acheson’s (2015) synthesized
discussion of economic defendability enormously helpful in our research. As noted at the outset,
our goal was to identify some objective and clearly measurable indicators of defendability using
a representative sample that would increase the chances that the results would be generalizable.
One of the most important factors appears to be the need to move, whether it be frequently every
few days, seasonally across the year, or cyclically after a few years. Mobility appears to work
strongly against private property—virtually all our sample societies lacking year-round
permanent settlement have communal systems. As we have discussed, mobility is likely to make
it very difficult for an individual household to defend all of the land they would need for
subsistence. This does not mean that defense of a large territory is not needed, but it does mean
that in nonstate societies defense would be virtually impossible without institutionalized
mechanisms for collective action by a group such as a corporate kin group, a territorial band or
village community. Mobility, in turn, is very strongly predicted by either a high dependence on
hunting and gathering, or on herding animals or combination of both. The lack of mobility is
strongly predicted by irrigation. All of these conditions are strongly related to societal
complexity and suggest that if property rights were claimed at all, communal ownership was
probably the default land system prior to agriculture. Some form of intensification, such as
irrigation, is probably essential to push a society to privatize land tenure.
Our research suggests that defendability theory goes a long way in helping us understand
property systems of societies in the recent past. But given that so many governments are
pressuring or requiring property to be privatized, it is especially important to undertake
diachronic research to evaluate the effects of privatization on societies that previously had
communal ownership. As we noted earlier, our modernization controls were very indirect
measures of government intervention. Does the switch to private property where conditions
favored communal ownership increase vulnerability? Do such societies adopt informal
cooperative practices to mitigate privatization? As in all research, new questions arise. We hope
our findings will encourage others to pursue some of the questions we have raised here and
inspire new questions and new strategies for testing them.
Notes
This material is based upon work supported by the National Science Foundation under Grant
Number (#SMA-1416651). We thank Joseph P. Feldman for his coding of land tenure systems,
Rachele Pierro for translation assistance, Abbe McCarter for editorial assistance, and Ian
Skoggard for advice. We also want to thank, the three anonymous reviewers for the time and
effort they put into their reviews and for their constructive criticisms.
18

1 See also Baker (2003) for a mathematical model. Much of the scholarship regarding economic
defendability theory comes from behavioral ecology research on nonhuman primates and other
animal species (for reviews see Brown 1964, Cashdan 1983, Dyson-Hudson and Smith 1978,
Jaeggi et al. 2016). Territorial behavior has largely been the focus of this research. Economic
defendability models have also been considered as explanations of territoriality in selected
comparisons of foraging societies and other small-scale societies (Baker 2003; Cashdan 1983).
However, Acheson (2015) noted that territoriality and property are very different concepts.
Territoriality is primarily about defensive behavior or control, whereas property concepts
primarily involve jural rights and the cultural assignment of those rights. Societies with
communal land tenure and private land tenure, the focus of our study, both claim rights to
property; what differentiates these property systems is the nature and size of the group that is
making socially-agreed upon (customary) claims to land. Nonetheless, the behavioral ecology
literature has provided important insights into the impacts of, and interactions with the
environment, particularly risky environments, that may shape social structure.
2There is a large literature on the different evolutionary mechanisms that might account for
human cooperation and collective action (see, for example, Blanton 2016, Boehm 2012, Bowles
and Gintis 2013, Cronk and Leech 2013, and Richerson et al. 2016). Presumably communal
property systems require more cooperation than private property systems.
3 Some forms of resource stressors, such as frequent drought, may also make land harder to
defend and less likely to be worth defending. First, drought may require more mobility when
water or pasture is scarce. Second, frequent drought may degrade land and may make particular
parcels of land less valuable.
4 By these definitions, both the Masai and the Nama are considered “high herders,” but the Masai
depend about 90% on herding and the Nama about 50% (with 30% dependence on hunting and
about 10% dependence on gathering and 10% on hunting). The Kurds are considered “low
herders” and they depend 50% on agriculture and have about 40% dependence on herding.
5We note that the simple dependence on animal husbandry (without accounting for dependence
on pigs) is not a significant predictor of mobility. For example, in our sample (n=88), the gamma
for the dependence on animal husbandry (v206) and non-permanent versus permanent
settlements is only .19 and the rho = .15 (both not significant). If we remove societies that have
animal husbandry mostly dependent on pigs, the correlations become significant (gamma= -.430;
rho=-.327; p<.05; N=51), but they are not nearly as strong as when we dichotomize societies
depending mostly (5-10 or about 50% or more) on herding animals versus those that depend less
on herding animals (0-4) or on pig husbandry.
6 Some scholars have expressed concern about possible non-independence of cases in cross-
cultural comparisons and ask for controls, such as on language family, to see if the main
predictors survive such tests. A control on language is assumed to take out the effect of historical
inertia, a bias that might falsely lead to clustering of variables not functionally related. However,
as Ember et al. (2018:386) have pointed out, controlling on language family can alternatively be
controlling on similarity of ecological environments since people tend to stay within familiar
ecozones. Thus controlling on language might inadvertently get rid of important ecological
19

parameters that are not otherwise being measured. We share these reservations, but we have
controlled on language family for our multiple and ordinal regression analyses by adding dummy
variables for language families containing two or more societies. There are only two instances
out of 29 in the multiple regressions (Tables 1 and 3) where the main effects are reduced in
significance; one case become marginal (previously significant) and one becomes marginal one-
tailed (previously marginal two-tailed)--see S.8.1). Our main predictors remain significant using
ordinal regression models (see S.8.2.2).
7Indeed, dependence upon fishing is not a significant predictor of the dichotomized mobility
score; the gammas and rhos are all less than .09 in our sample and in the combined sample.
8 This is probably why subsistence type and mobility both add predictability.
20

References
Acheson, James M. 2015. "Private Land and Common Oceans Analysis of the Development of
Property Regimes." Current Anthropology 56 (1):28-55.
Adger, W. Neil. 2000. Social and Ecological Resilience: Are They Related? Progress in Human
Geography 24(3): 347-364.
Baker, Matthew J. 2003. An Equilibrium Conflict Model of Land Tenure in Hunter-Gatherer
Societies. Journal of Political Economy 111(1): 124-173.
Berry, Sara. 1993. No Condition is Permanent: The Social Dynamics of Agrarian Change in
Sub-Saharan Africa. University of Wisconsin Press.
Berry, Sara. 2002. "Debating the Land Question in Africa." Comparative Studies in Society and
History 44 (4):638-668.
Blanton, Richard E. 2016. How Humans Cooperate: Confronting the Challenges of Collective
Action. University Press of Colorado.
Boehm, Christopher. 2012. Moral origins: The Evolution of Virtue, Altruism, and Shame. Soft
Skull Press.
Bowles, Samuel, and Herbert Gintis. 2013. A Cooperative Species: Human Reciprocity and its
Evolution. Princeton University Press.
Bromley, Daniel W., David Feeny, Margaret A. Mckean, Pauline Peters, Jere L. Gilles, Ronald J.
Oakerson, C. Ford Runge, and James T. Thomson. 1992. Making the Commons Work: Theory,
Practice and Policy: ICS Press.
Bruce, John W. 1986. "Land Tenure Issues in Project Design and Strategies for Agricultural
Development in sub-Saharan Africa." Land Tenure Center, University of Wisconsin-Madison.
Bruce, John W, and S. E. Migot-Adholia. 1994. Searching for Land Tenure Security in Africa:
Dubuque, Iowa: Kendall/Hunt Publishing Company.
Cashdan, Elizabeth, Alan Barnard, M.C. Bicchieri, Charles A Bishop, Valda Blundell, Jeffrey
Ehrenreich, Mathias Guenther, Annette Hamilton, Henry C. Harpending, and Nancy Howell.
1983. "Territoriality among Human Foragers: Ecological Models and an Application to Four
Bushman Groups [and Comments and Reply]." Current Anthropology 24 (1):47-66.
Behrendt, S. 2014. “lm. beta: Add Standardized Regression Coefficients to lm-Objects.” R
package version 1.5-1.
Chabot‐Hanowell, Benjamin, and Eric Alden Smith. 2012. “5 Territorial and Nonterritorial
Routes to Power: Reconciling Evolutionary Ecological, Social Agency, and Historicist
Approaches.” Archeological papers of the American Anthropological Association 22(1): 72-86.
21

Cronk, Lee, and Beth L. Leech. 2013. Meeting at Grand Central: Understanding the Social and
Evolutionary Roots of Cooperation. Princeton University Press.
Dirks, Robert. 1993. "Starvation and famine: cross-cultural codes and some hypothesis tests."
Cross-Cultural Research 27 (1-2):28-69.
Dirks, Robert. 2004. "Hunger and famine." In Research Frontiers in anthropology. New
Directions in Anthropology, edited by Carol R. Ember and Melvin Ember. Upper Saddle River,
NJ: Prentice-Hall.
Divale, William T. 2004. "Codebook of Variables for the Standard Cross-Cultural Sample."
World Cultures (May).
Divale, William T., and Albert Seda. 2000. "Cross-Cultural Codes of Modernization." World
Cultures 11 (2):152-170.
Dyson‐Hudson, Rada, and Eric Alden Smith. 1978. "Human territoriality: an ecological
reassessment." American Anthropologist 80 (1):21-41.
Ekstrøm, C.T. 2019. “MESS: Miscellaneous Esoteric Statistical Scripts. R package version
0.5.6”. https://github.com/ekstroem/MESS
Ember, Carol R. 2007. “Using the HRAF Collection of Ethnography in Conjunction with the
Standard Cross-Cultural Sample and the Ethnographic Atlas.” Cross-Cultural Research 41 (4):
396-427.
Ember, Carol R., and Melvin Ember. 1992a. “Resource Unpredictability, Mistrust, and War: A
Cross-Cultural Study.” Journal of Conflict Resolution 36: 242-262.
Ember, Carol R., and Melvin Ember. 1992b. Warfare, Aggression, and Resource Problems:
Cross-Cultural Codes. Cross-Cultural Research 26:169-226.
Ember, Carol R., and Melvin Ember. 2009. Cross-Cultural Research Methods. 2nd ed. AltaMira
Press.
Ember, Carol R, Melvin Ember, and Burton Pasternak. 1974. “On the Development of Unilineal
Descent.” Journal of Anthropological Research 30 (2):69-94.
Ember, Carol R., Ian Skoggard, Erik J. Ringen, and Megan Farrer. 2018. “Our Better Nature:
Does Resource Stress Predict Beyond-Household Sharing?” Evolution and Human Behavior 39
(4), 380-391. https://doi.org/10.1016/j.evolhumbehav.2018.03.001
Engels, Friedrich, and Tristram Hunt. 1884. The Origin of the Family, Private Property and the
State: Penguin UK. Reprint, 2010.
Ethnology, Editors of. 1962ff. "Ethnographic Atlas." Ethnology.
Feder, Gershon, and Raymond Noronha. 1987. "Land Rights Systems and Agricultural
Development in Sub-Saharan Africa." The World Bank Research Observer 2 (2):143-169.
22

Felzer, B.S., C. R. Ember, R. Cheng, and M. Jiang. n.d. “The Relationships of Extreme
Precipitation and Temperature Events with Ethnographic Reports of Droughts and Floods in
Nonindustrial Societies” Weather, Climate, and Society (under review).
Folke, Carl. 2006. “Resilience: The Emergence of a Perspective for Social–Ecological Systems
Analyses.” Global Environmental Change 16(3): 253-267.
Folke, Carl, Carpenter, S., Walker, B., Scheffer, M., Chapin, T. and Rockström, J., 2010.
“Resilience Thinking: Integrating Resilience, Adaptability and Transformability.” Ecology and
Society” 15(4) http://www.ecologyandsociety.org/vol15/iss4/art20/
Fox, J. and Sanford Weisberg. 2019. An {R} Companion to Applied Regression, Third Edition.
Sage. URL: https://socialsciences.mcmaster.ca/jfox/Books/Companion/
Fratkin, Elliot, and Eric Abella Roth. 2006. As Pastoralists Settle: Social, Health, and Economic
Consequences of the Pastoral Sedentarization in Marsabit District, Kenya. Vol. 1: Springer
Science & Business Media.
Geertz, Clifford. 1980. "Organization of the Balinese Subak." In Irrigation and Agricultural
Development in Asia: Perspectives from the Social Sciences, edited by E. Waltern Coward, 70-
90. Cornell University Press..
Hardin, Garrett. 1968. "The Tragedy of the Commons." Science 162 (3859):1243-1248.
Hobhouse, L. T., Gerald Clair William Camden Wheeler, and Morris Ginsberg. 1915. The
Material Culture and Social Institutions of the Simpler Peoples; An Essay in Correlation,
London School of Economics and Political Science (University of London) Series of Studies in
Economics and Political Science No 3 of the Monographs on Sociology. Chapman & Hall, ltd.
Jaeggi, Adrian V., Klaree J. Boose, Frances J. White, and Michael Gurven. 2016. “Obstacles and
Catalysts of Cooperation in Humans, Bonobos, and Chimpanzees: Behavioural Reaction Norms
Can Help Explain Variation in Sex Roles, Inequality, War Peace.” Behaviour 153(9-11): 1015-
1051.
Kathryn R. Kirby, Russell D. Gray, Simon J. Greenhill, Fiona M. Jordan, Stephanie Gomes-Ng,
Hans-Jörg Bibiko, Damián E. Blasi, Carlos A. Botero, Claire Bowern, Carol R. Ember, Dan
Leehr, Bobbi S. Low, Joe McCarter, William Divale, and Michael C. Gavin. 2016. “D-PLACE: A
Global Database of Cultural, Linguistic and Environmental Diversity.” PLoS ONE, 11(7):
e0158391. doi:10.1371/journal.pone.0158391.
Labovitz, Sanford. 1970. "The Assignment of Numbers to Rank Order Categories." American
Sociological Review 35: 515-524.
Lang, Hartmut. 1995. "CONAN--An Electronic Code-Text Data-Base for Cross-Cultural
Studies." World Cultures 9 (2): 1-38.
Lentz, Carola. 2013. Land, Mobility, and Belonging in West Africa: Natives and Strangers:
Indiana University Press.
23

Little, Peter D. 2003. "Rethinking Interdisciplinary Paradigms and the Political Ecology of
Pastoralism in East Africa." In African Savannas: Global Narratives and Local Knowledge of
Environmental Changes, editors T. Bassett and D. Crummey, Oxford: James Currey.
Mazerolle, Marc J. .2019. “AICcmodavg: Model selection and multimodel inference based on
(Q)AIC(c).” R package version 2.2-2. https://cran.r-project.org/package=AICcmodavg.
McCabe, J Terrence. 1987. "Drought and Recovery: Livestock Dynamics among the Ngisonyoka
Turkana of Kenya." Human Ecology 15 (4): 371-389.
McCabe, J Terrence. 1990. "Turkana Pastoralism: A Case against the Tragedy of the Commons."
Human Ecology 18 (1): 81-103.
McPeak, John G., Peter D. Little, and Cheryl R. Doss. 2012. “Risk and Change in an African
Rural Economy. London: Routledge.
Murdock, George P., and Douglas R. White. 1969. "Standard Cross-Cultural Sample."
Ethnology 8: 329-369.
Murdock, George Peter, and Caterina Provost. 1973. "Measurement of Cultural Complexity."
Ethnology 12: 379-392.
Nugent, Jeffrey B, and Nicholas Sanchez. 1999. "The Local Variability of Rainfall and Tribal
Institutions: The Case of Sudan." Journal of Economic Behavior & Organization 39 (3): 263-
291.
Ostrom, Elinor, Thomas Dietz, Nives Dolšak, Paul C. Stern, Susan Stonich, and Elke U. Weber.
2002. The Drama of the Commons: National Academy Press.
Peters, Pauline E. 1994. Dividing the Commons: Politics, Policy, and Culture in Botswana.
Charlottesville, University Press of Virginia.
Pryor, Frederic L. 1977. The Origins of the Economy: A Comparative Study of Distribution in
Primitive and Peasant Economies: Academic Press.
Pryor, Frederic L. 2003. "Economic Systems of Foragers." Cross-Cultural Research 37 (4): 393-
426.
Pryor, Frederic L. 2005. Economic Systems of Foraging, Agricultural, and Industrial Societies:
Cambridge University Press.
Pryor, Frederic L. 2011. "Appendices to Economic Systems of Foraging, Agricultural, and
Industrial Societies." http://www.swarthmore.edu/SocSci/Economics/fpryor1/
Ribot, Jesse C, and Nancy Lee Peluso. 2003. "A Theory of Access." Rural Sociology 68 (2):
153-181.
24

Richerson, Peter, Ryan Baldini, Adrian V. Bell, Kathryn Demps, Karl Frost, Vicken Hillis, Sarah
Mathew et al. 2016. "Cultural Group Selection Plays an Essential Role in Explaining Human
Cooperation: A Sketch of the Evidence." Behavioral and Brain Sciences 39.
doi:10.1017/S0140525X1400106X, e30
Rudmin, Floyd Webster. 1988. "Dominance, Social Control, Ownership: A History and a Cross-
Cultural Study of Motivations for Private Property." Behavior Science Research 22 (1-4): 130-
160.
Rudmin, Floyd Webster. 1992a. "Cross-Cultural Correlates of the Ownership of Private
Property." Social Science Research 21 (1): 57-83.
Rudmin, Floyd Webster. 1992b. "Cross-Cultural Correlates of the Ownership of Private Property:
A Look from Another Data-Base." Anthropologica 34: 71-88.
Rudmin, Floyd Webster. 1995. "Cross-Cultural Correlates of the Ownership of Private Property:
Two Samples of Murdock's Data." The Journal of Socio-Economics 24 (2): 345-373.
Rudmin, Floyd Webster. 1996. "Cross-Cultural Correlates of the Ownership of Private Property:
Zelman's Gender Data Revisited." Cross-Cultural Research 30 (2):115-153.
Scoones, Ian, and Carol Kerven. 1993. Range Ecology at Disequilibrium: New Models of
Natural Variability and Pastoral Adaptation in African Savannas: Overseas Development
Institute.
Scott, James C. 1998. Seeing Like a State: How Certain Schemes to Improve the Human
Condition Have Failed: Yale University Press.
Sidky, MH. 1993. "Irrigation and the Political Evolution of the High-Mountain Kingdom of
Hunza1." Asian affairs 24 (2): 131-144.
Signorell, Andri et. al. 2019. “DescTools: Tools for Descriptive Statistics. R Package Version
0.99.29”, https://cran.r-project.org/package=DescTools.
Simmons, Leo W. 1937. "Statistical Correlations in the Science of Society." In Studies in the
Science of Society, editor George P. Murdock, 495-517. New Haven: Yale University Press.
Spier, Leslie. 1928. Havasupai Ethnography: American Museum of Natural History New York.
Sumner, William Graham, and Albert Galloway Keller. 1927. The Science of Society. 4 Vols.
New Haven: Yale University Press.
Swanson, Guy E. 1964. The Birth of the Gods. Ann Arbor: University of Michigan Press.
Team, Core Writing, Rajendra K Pachauri, and LA Meyer. 2014. "IPCC, 2014: Climate Change
2014: Synthesis Report. Contribution of Working Groups I." II and III to the Fifth Assessment
Report of the intergovernmental panel on Climate Change. IPCC, Geneva, Switzerland 151.
25

Venables, W.N., & Ripley, B.D. 2002. Modern Applied Statistics with S, Fourth edition. Springer,
New York.
Weisberg, Herbert, Jon A Krosnick, and Bruce D Bowen. 1996. An Introduction to Survey
Research, Polling, and Data Analysis: Sage.
Whiting, John WM, John A Sodergren, and Stephen M Stigler. 1982. "Winter Temperature as a
Constraint to the Migration of Preindustrial Peoples." American Anthropologist 84 (2): 279-298.
Wickham, Hadley. 2016. ggplot2: Elegant Graphics for Data Analysis, 2nd edition. New York:
Springer-Verlag.
Wrangham, Richard W. 1980. An Ecological Model of Female-Bonded Primate
Groups. Behaviour 3-4: 262-300.
26