Revisiting the diversification and insurance relationship: Differences between on- and off-farm strategies

Abstract

Crop insurance is an important instrument for farmers to cope with climate risks. Yet, also, diversification plays a crucial role. The use of crop insurance and diversification is interrelated. For example, an increasing support and uptake of insurance solutions may discourage farmers to take diversification measures on the farm, given that they now have an insurance dealing with risks. This may have unintended consequences, such as biodiversity decline.

We examine the relationships between seven different kinds of income diversification and the uptake of crop insurance. Theoretically, both substitutive and complementary relationships are possible. The reason is that, depending on their preferences and constraints, farmers choose income bundles that optimally balance profits and risks. Here we provide the first systematic empirical examination of this issue. Our analysis is based on our own survey data from 1176 Swiss fruit growers. We consider on- and off-farm diversification strategies, namely inter-varietal diversity, agro-tourism, processing and direct marketing of products, creation of financial reserves for bad times, forestry work, off-farm investment, off-farm income and their association with insurance uptake. In line with our theoretical reasoning, we do indeed find both substitutive and complementary relationships. In general, on-farm diversification is associated with positive insurance demand, whereas off-farm diversification has a negative association.

Full text

Journal Pre-proofs
Revisiting the diversification and insurance relationship: Differences between
on- and off-farm strategies
Ladina Knapp, David Wuepper, Tobias Dalhaus, Robert Finger
PII: S2212-0963(21)00044-9
DOI: https://doi.org/10.1016/j.crm.2021.100315
Reference: CRM 100315
To appear in: Climate Risk Management
Received Date: 23 November 2020
Revised Date: 24 April 2021
Accepted Date: 24 April 2021
Please cite this article as: L. Knapp, D. Wuepper, T. Dalhaus, R. Finger, Revisiting the diversification and
insurance relationship: Differences between on- and off-farm strategies, Climate Risk Management (2021), doi:
https://doi.org/10.1016/j.crm.2021.100315
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2021 The Author(s). Published by Elsevier B.V.

Revisiting the diversification and insurance relationship: Differences between on- and off-
farm strategies
Ladina Knapp1, David Wuepper1, Tobias Dalhaus1,2, Robert Finger 1
1Agricultural Economics and Policy, ETH Zürich, Switzerland
2Wageningen University and Research, Business Economics Group, the Netherlands
Corresponding author: Ladina Knapp
Email: ladina.knapp@graduateinstitute.ch
Revisiting the diversification and insurance relationship: Differences between on- and off-
farm strategies
Abstract
Crop insurance is an important instrument for farmers to cope with climate risks. Yet, also,
diversification plays a crucial role. The use of crop insurance and diversification is interrelated. For
example, an increasing support and uptake of insurance solutions may discourage farmers to take
diversification measures on the farm, given that they now have an insurance dealing with risks. This
may have unintended consequences, such as biodiversity decline.
We examine the relationships between seven different kinds of income diversification and the uptake of
crop insurance. Theoretically, both substitutive and complementary relationships are possible. The
reason is that, depending on their preferences and constraints, farmers choose income bundles that
optimally balance profits and risks. Here we provide the first systematic empirical examination of this
issue. Our analysis is based on our own survey data from 1176 Swiss fruit growers. We consider on-
and off-farm diversification strategies, namely inter-varietal diversity, agro-tourism, processing and
direct marketing of products, creation of financial reserves for bad times, forestry work, off-farm
investment, off-farm income and their association with insurance uptake. In line with our theoretical

reasoning, we do indeed find both substitutive and complementary relationships. In general, on-farm
diversification is associated with positive insurance demand, whereas off-farm diversification has a
negative association.
Keywords: Insurance use; on-farm and off-farm diversification; specialized crops; risk
management strategies; Switzerland

1 Introduction
A salient characteristic of agriculture is the intrinsic risk involved. For example, the exposure
to climatic risks and climatic extreme events can create losses in yield quantity (Lesk et al.,
2016) and quality (Dalhaus et al., 2020) and thus increase the volatility of incomes. Climate
change increases the relevance of these risks (e.g. Lunt et al., 2016, Mase et al., 2017, Trnka et
al., 2017, Webber et al., 2018). Two common policy responses are subsidized insurance (e.g.
in the United States) (Coble and Barnett 2012) and support for farm diversification (e.g. in
Europe) (Meraner et al., 2015). Farmers have a number of risk management strategies at hand,
including weather risk management strategies, that can improve the resilience of farms (Vroege
and Finger, 2021). These risk management strategies can include crop insurance and
diversification decisions, which reduce weather risks.
Hail insurance is the dominant insurance type in Swiss and other central European crop
production, therefore understanding its interrelation with other risk management strategies is
crucial to understand how farmers make climate risk management decisions and subsequently
to develop targeted policies. For example, the uptake of an insurance may discourage farmers
to take diversification measures on the farm, given that they now have an insurance dealing
with risks. If a farmer has an insurance, her/his responses to moral hazard may possibly include
a switch to more risky crops (Yu and Sumner 2018) and inputs (Roll 2019) and adjustments in
capital structure and investment (de Mey et al., 2016). This effect of crop insurance on the level
of diversification is relevant for a wide body of stakeholders. For example, it may affect what
and how farmers produce, which has consequences for food supply as well as environmental
outcomes such as pesticide use and biodiversity (e.g. Müller et al., 2017, Möhring et al., 2020).
The current findings on the interrelationship of crop insurance and diversification are
ambiguous.
In this paper, we revisit this relationship by reviewing risk management instruments, including
on-farm and off-farm diversification activities and their association with crop (hail) insurance
uptake. We argue that the relationship between crop insurance and diversification might vary
depending on the type of diversification activity in question. Crop insurance and diversification
might be negatively or positively correlated, depending on the type of diversification we are
dealing with. Based on data from Swiss fruit growers, we examine whether seven different
diversification activities and crop insurance uptake are negatively or positively correlated at the
general farm management level. Crop insurances are market-based risk management
instruments that provide payouts in case of a given loss event (e.g. due to hail). Diversification

is an on-farm risk management instrument. It involves on-farm diversification such as crop
species diversity and off-farm measures such as the re-allocation of the farm’s production
resources to non-farming activities. In this paper, we consider a wide set of diversification
measures, namely: i) inter-varietal diversity, ii) agro-tourism, iii) processing and direct
marketing of the produce on the farm, iv) creation of financial reserves for bad times, v) forestry
work, vi) off-farm investment and vii) share of off-farm income. These diversifications
activities are contributing to climate risk management. More specifically, the inter-varietal
diversity is a strategy used by farmers to reduce possible impacts of extreme weather. In
addition, with off-farm diversification, such as forestry or off-farm work, growers reduce the
exposure of their overall income and liquidity to climate risks.
The Swiss case is particularly interesting because, while in many countries, agricultural
insurances are subsidized (Moschini and Hennessy 2001, Bardají et al., 2016), Switzerland has
(currently) no subsidies on crop insurance premiums allowing for an unbiased analysis of crop
insurance use (Finger and Lehmann 2012).
Former contributions on crop insurance – diversification interrelationships include Mishra and
Goodwin (2003), Chang and Mishra (2012) and Velandia et al. (2009), who find that in the
United States, income diversification and crop insurance are negatively correlated. Similar
patterns are found by El Benni et al. (2012) and Wuepper et al. (2018) in Switzerland and
Ethiopia, respectively. This negative correlation was not only found for income diversification
but also for on–farm diversification, such as the number of crops which are cultivated (Blank
and McDonald 1996, Smith and Goodwin 1996, Di Falco and Chavas 2009, O’Donoghue et al.,
2009, Enjolras and Sentis 2011, Di Falco et al., 2014). However, other studies find opposite
results. For example, studies undertaken by Mishra et al. (2004), Enjolras et al. (2012), Lefebvre
et al. (2014) and Goodwin (1993) regard diversification and crop insurance as positively
associated. In summary, there have been ambiguous findings on the relationship between
diversification and crop insurance uptake.
One main gap in current literature is the lack of a systematic analysis of different kinds of on-
and off-farm diversification measures and crop insurance. Today’s literature on crop insurance
uptake does not distinguish between different types of diversification activities. Studies differ
in how they measure and define diversification. Indeed, generally speaking, only one type of
diversification activity is considered, whereas farmers have a large portfolio of diversification
measures to choose from especially in high value fruit production. For instance, current

literature neglects diversification involving different varieties of fruits which is an important
risk management strategy adopted by fruit growers. Inter-varietal diversity is especially relevant
in fruit production, where mixing early maturing and late maturing varieties is a commonly
used climate risk management practice (Haigh et al., 2015). This varietal diversity also
generates several ecosystem services (e.g. with respect to pest and disease pressure and
pollination (e.g. Hajjar et al., 2008). Therefore, this diversity and its interaction with crop
insurance uptake is of interest for policy-makers. Moreover, insights into the association of crop
insurance uptake and on-farm and off-farm diversification strategies provide additional
information on possible options for tailoring policies and crop insurance offers.
To fill this gap, we investigate how the relationships between a diversification and crop
insurance use differs among different diversification strategies. For this, we collected survey
data from 1176 farmers in Switzerland on a large portfolio of risk management strategies that
have to the best of our knowledge not been considered so far jointly. Our dataset includes
detailed information on seven different types of diversification activities undertaken by farmers,
including on-farm diversification strategies such as the use of different fruit varieties but also
off-farm diversification strategies not related to the production on the farm, such as forestry
activities or off-farm labor. We use regression analysis to estimate the relationship between on-
farm and off-farm diversification activities and crop insurance uptake. To avoid biased
inference we use the approaches by Imbens (2003) and Oster (2019) to avoid omitted variable
bias.
Our main finding is that the relationship between diversification and crop insurance is
heterogeneous across different kinds of diversification. On-farm diversification measures tend
to be positively correlated to crop insurance uptake, whereas off-farm diversification tends to
be negatively correlated. This finding needs to be considered in agricultural policy making,
when incentivizing either insurance or diversification to avoid undesired side effects.
2 Theoretical Background
Farmers’ risk management strategies aim to reduce agricultural risk. They identify the type of
risk, potential risk-reducing management strategies and the welfare effects of risk (Chavas and
Shi 2015). We assume that risk averse farmers maximize their utility with respect to risk
management strategies, including: i) on-farm diversification strategies, i.e. inter-varietal
diversification, processing/direct marketing, agro-tourism and creation of financial reserves, ii)

off-farm diversification activities, i.e. off-farm investments, forestry work, share of off-farm
income and iii) crop insurance uptake.
Crop insurance and diversification are often expected to reduce the anticipated household
profits. A higher level of risk management activities often implies extra costs, for example,
direct monetary expenditure (e.g. for an insurance premium) or opportunity costs (e.g. forgone
efficiency gains from specialization) (e.g. McNamara and Weiss 2005). There might be notable
exceptions in the presence of complementarity benefits, which provide incentives to diversify
(Chavas and Di Falco 2012). In our case study, for instance, by producing a diversity of fruits
and varieties thereof, producers may be able to supply markets continuously and exploit full
market potential. Thus, on-farm strategies, off-farm strategies and insurance uptake may or may
not decrease expected farm profits.
Climate risk management strategies such as diversification of varieties, on-farm strategies, off-
farm strategies and insurance uptake are expected to have risk reducing effects. Assuming that,
on average, farmers are risk averse (see e.g. Iyer et al., 2020) farmers are better off with higher
levels of expected profit but worse off in higher levels of risk exposure. Farmers face trade-offs
as risk management strategies are expected to reduce the variance of their profits but also induce
additional costs that might erode their profits.
The optimal allocation of insurance and diversification activities depends on their effects on
expected profits and on the variance of profit, farmers’ risk preferences, and effects on risk
exposure (Barrett et al., 2001). Moreover, this optimization problem is subject to several
constraints at the household level, i.e. time constraints (for on- and off-farm work), budget
constraints (restricting possible consumption, expenditure and investment) and technology
constraints (e.g. limiting the use of specific production technologies) (e.g. Hennessy and
Rehman 2008, Fernandez-Cornejo et al., 2005). In addition, choices taken by farmers on risk
management strategies are also expected to depend on farm and household characteristics
(McNamara and Weiss 2005, Hansson et al., 2013, Weltin et al., 2017), and more specifically
also on farmers’ risk aversion (Meraner and Finger 2019). Moreover, policies such as decoupled
payments may induce farmers to take riskier production decisions on the farm (as highlighted
by Hennessy 1998), since farm income risk exposure is reduced. Decisions on on-farm, off-
farm diversification and insurance uptake are all interrelated and connected with the farmer’s
utility and additional variables such as farm and farmer characteristics.

Our goal is to quantify the sign and magnitude, of the relationship between crop insurance and
different diversification measures. Based on the framework presented above, we hypothesize
that on-farm and off-farm diversification activities are unlikely to have the same relationship
with insurance uptake. Crop insurance to cope with climatic risks will probably change the use
of other risk management strategies such as on-farm and off-farm diversification activities
(Wright and Hewitt 1994). There is a range of literature which suggests that insured farmers
are less likely to adopt on-farm diversification, because diversification and insurance are
substitutes (O’Donoghue et al., 2009). Insured farmers may increase specialization if it reduces
the need for self-insurance (O’Donoghue et al., 2009). Thus, the uptake of insurance may reduce
optimal diversification on and off the farm. However, a risk averse farmer using on-farm
diversification strategies might also be insured in order to reduce agricultural production risk
(Mishra et al., 2004). Since farmers choose income source bundles that optimally balance profit
and risk considerations according to their preferences and constraints, insurance and
diversification activities can have a positive (i.e. complementary) and negative (i.e. substitutive)
relationship (Richards 2000, Enjolras et al., 2012).
3 Data
Data collection
This paper is based on an online survey undertaken with 1176 Swiss fruit growers (Knapp et
al., 2018). Swiss fruit growers are often characterized by small farm sizes and usually family
farms. In our data collected, we have information on a large sample of small producers of
special crops, i.e. grapes, plums, berries and cherries.
The data has been pooled as there is little variability in insurance uptake and diversification
activities over the three years 2016, 2017, 2018 (see Appendix Table A2 and A3). Specifically,
if a farmer is present more than once in the dataset, we use the survey records filed for the first
year of his/her participation.
Farmers are asked whether they use crop insurance. In Swiss fruit production, this insurance
usually comprises coverage against hail damage as well as coverage against other elementary
risks such as storms, landslides and floods (see hagel.ch for details)
1
..For the whole country,
this results in a total sum of insured value of about 2 billion CHF with a yearly premium of
1
We also asked more details about different types of insurance, e.g. on insurance of frosts risks. But in general,
only few producers deviated from stand contracts. Thus, we here focus on one binary adoption decision: crop
insurances yes or no.

around 50 million CHF
2
. Farmers in Switzerland can choose to take insurance for a single crop
or the entire farm. Crop insurances are prolonged for the next year if the farmer does not cancel
it by the end of September. A list of the crops that are chosen by farmers should be sent to the
insurance company before the end of April (see www.hagel.ch for more information). The
Swiss case is particularly interesting for three reasons: a) Switzerland is considered to be one
of Europe’s hail hotspot (Schemm et al., 2016), b) Switzerland has a high share of special crops
that are vulnerable to hail
3
, c) whereas in many countries, agricultural insurances are subsidized
(Moschini and Hennessy 2001), Switzerland has currently no subsidies on insurance premiums
allowing for an unbiased analysis (Finger and Lehmann 2012).
The dataset includes an overview of fruit varieties (36 berry varieties, 20 cherry varieties, 13
plum varieties and 30 grape varieties) and the surface area for each variety. We create a Shannon
Diversity Index
4
to represent the evenness and diversity in the type of fruit varieties used per
fruit grower (see Smale 2005). The Shannon Diversity Index is maximal if the same acreage of
all species considered are grown on a farm. Thus, the maximum value differs across crops
ranging from 3 (cherries) to 2.25 (plums). Information was collected on further different types
of on-farm/off-farm management strategies (1/0) adopted by the fruit growers. These included
commonly adopted strategies, such as processing and direct marketing activities, agro-tourism,
creation of financial reserves, off-farm investments, forestry work, and share of off-farm
income
5
. These diversification activities are relevant to the Swiss case study and were identified
based on expert interviews.
As control variables we use farmers’ risk preferences, based on a self-assessment of risk as
proposed by Dohmen et al., (2011), farm size (100 square meters), focus of the farm, land
tenure, production system, the travel distance in seconds to the next urban center
6
, farm-specific
2
Numbers are taken from this source https://www.gabot.de/ansicht/schweizer-hagel-anspruchsvolles-
schadensjahr-402989.html.
3
The total acreage under fruits in Switzerland is circa 6258 ha. Of this area under fruits, cherries are grown on 9%,
plums on 4% and table grapes on 0.29%. Grapes cultivated for wine production cover 14712 ha (for more
information s ee: Swiss Federal Office for Agriculture (FOAG) Obst‐ und Tafeltraubenanlagen der Schweiz 2019,
Bern (2019). The production of fruit and grapes for wine production represented 4% and 7% of the total agricultural
production value in 2018 (see Knapp et al. 2019).
4
To calculate the Shannon Diversity Index, we took the area of parcel i planted with variety j divided by the total
area of the parcel i multiplied by the log area share (αji). This was performed for every variety the fruit grower has
and the sum of these values results in the Shannon Index. See formula, Shannon Diversity Index :

  

5
If the percentage of earnings originating from farm is ≤50% then 0 if percentage of earnings originating from
farm >50%,then 1). We performed the regression with the categorical variables: 0-25%, 26-50%,51-75%, 76-
100%. Results are available in the Appendix in Table A7.
6
Urban center is defined here as centers having more than 10'000 inhabitants. The data is taken from google maps.

annual average temperature and annual rainfall derived from high-resolution gridded maps for
the period 1961-2016 (Frei and Schär 1998, Frei 2014). Unfortunately, due to the structure of
our data, we are unable to estimate short term impacts
7
of within season weather shocks on
climate risk management. As the main risk covered in the crop insurance is hail, we also account
for both spatial and temporal heterogeneity of hail risk exposure. To this end, we control for
hail events in recent years and also include an indicator for long-term hail risk exposure of each
municipality (Finger and Lehmann, 2012). Table A4 of the Appendix provides a detailed
description of the variables.
Summary statistics
Table 1 provides an overview of some important variables. Table 2 provides summary
statistics. In our sample of fruit growers, 37% (436 fruit growers) have an insurance. This
reflects that the share of adopters is lower than for the general sample of farms in Switzerland
(e.g. Finger and Lehmann (2012), with reported uptake rates of about 60% of farms in the
bookkeeping dataset for the year 2009). As shown in Table 2, the mean for the diversification
activities varies between insured and uninsured farmers, especially for the Shannon Diversity
Index, processing and direct marketing, forestry work and the share of off-farm income. We
undertook the Mann Whitney test to compare the means of insured as well as the chi square test
for the binary variables.
7
In case a dataset with more temporal heterogeneity in the risk management decision, i.e. hail insurance and
diversification, becomes available, future research could further assess the impact of short term weather
fluctuations, i.e. weather shocks, on risk management decisions. Unfortunately, due to the structure of our data
we are unable to quantify this impact here.

Table 1. Overview of variables
Variables
Data collected
Shannon Diversity Index
Shannon Diversity Index for varieties represents the evenness and diversity of
the varieties the fruit grower has. We use the surface area for each variety and
calculate a Shannon Diversity Index for each fruit grower (0= low diversity).
Processing and direct marketing of
products
Processing and direct marketing are defined as processing the fruits or selling
them directly to the customers.
(1=yes,0=no):
Agro-tourism
Other activities on the farm such as agro-tourism.
(1=yes,0=no)
Creation of financial reserves
Creation of financial reserves is defined as saving money for bad periods.
(1=yes,0=no)
Forestry work
Forestry work is defined as undertaking forestry work
(1=yes,0=no)
Off-farm investment
Off-farm investment is defined as investing in other companies or real estate.
(1=yes,0=no)
Share of off-farm income
Share of off-farm income is 1 if less than 50% of earnings originate from farm
and 0 if more than 50% originates from farm.
Risk preferences
Mean of fruit growers risk aversion over four domains: agriculture, production,
market and price, external financing.
Likert scale : 10= not willing to take a risk
0 = very willing to take a risk.
Organic farming
Production system is organic or not (1=yes,0=no)
Focus specialized
Focus specialized category indicates the farm type, i.e. whether a fruit is source
of main income (i.e. berries, cherries, plums or grapes).
(1=yes,0=no):
Farm size
Total farmland in 100 sq meters
Land tenure
Percentage of land leased by the fruit grower
(0 if ≤ 50%, 1 if >50%)
Mean precipitation
Mean temperature
Mean precipitation is the average annual rainfall over the years 1961-2016 for
the municipality (Frei and Schär 1998, Frei 2014).
Mean temperature is the average annual temperature over the years 1961-2016
for the municipality (Frei and Schär 1998, Frei 2014).
Hail last year
Hail years
Hail last years is the number of hail events in the year prior to data collection for
each municipality. (Finger and Lehmann, 2012).
Hail years is the number of years with hail events for each municipality in the
period 1961–2004. (Finger and Lehmann, 2012).
Distance
Distance measured in seconds from the location of the fruit grower to the next
town /urban center (Data retrieved from Google maps)
Year
Year the fruit grower filled out the survey
Fruit
Type of survey the fruit grower responded to (berries, cherries, plums or grapes)
Cantons
Cantons where the fruit growers are located.
Note: T he table above provides an overview of variables considered in our regression analysis and for each it includes a definition and the
name of the variable. All variables originate from the surveys undertaken with farmers except for the variables where references are provided.

Table 2. Summary statistics
Overall N=1176
Insured
N=436
Uninsured
N=728
All
N=1176
Diversification activities
Mean
Mean
Mean
Shannon Diversity Index (0=low diversity,
3=high diversity)
1.05*** (sd=0.7)
0.88 (sd=0.8)
0.94 (sd=0.77)
Processing, direct marketing (1=yes/0=no)
0.51***
(0.001)
0.41
0.44
Agro-tourism (1=yes/0=no)
0.14
(0.667)
0.13
0.14
Creation of financial reserves (1=yes/0=no)
0.26
(0.662)
0.27
0.27
Forestry work (1=yes/0=no)
0.016**
(0.03)
0.04
0.03
Off-farm investment (1=yes/0=no)
0.094
(0.495)
0.082
0.086
Share off-farm income
(earnings originating from
non-farm income ≥ 50% =1 if <50%=0)
0.24 ***
(0.000)
0.42
0.35
Control variables
Risk preferences (Numeric, 0=Risk loving,
10=Risk averse)
5.95 (sd=2.14)
5.95(sd=2.29)
5.96 (sd=2.25)
Organic production(1=yes/0=no)
0.06***
(0.008)
0.11
0.094
Focus specialized (1=yes/0=no)
0.49**
(0.108)
0.54
0.52
Land tenure (% of land leased) (≥50%=1
<50%=0)
0.28
(0.351)
0.31
0.3
Farm size (100 square meters)
1689***(sd=2897)
1007.41 (sd=202
3)
1256.36(sd=2400)
Mean precipitation (numeric)
1124.58***(sd=202)
1207.2 (sd=329)
1176.91(sd=291)
Mean temperature (numeric)
9.11*** (sd=1.5)
8.89 (sd=1.7)
8.98 (sd=1.64)
Hail years (numeric)
26.28***(sd=9)
23.5 (sd=10.6)
24.6 (sd=10)
Hail last year (numeric)
1.5**(sd=1.4)
1.82 (sd=1.7)
1.7 (sd=1.7)
Distance (numeric, time in seconds)
918.99***(sd=770)
985.82 (sd=682)
961.63 (sd=714)
Note: *,**,*** denote significance at the 10%,5%,1% level of the Mann Whitney Test highlighting whether there is a significant difference
between the two means insured and non-insured. In brackets are the p values of the Chi square test (undertaken only for the binary variables)
comparing the insured to non insured mean values.
4 Methodology
Our interest focuses on the relationship between insurance uptake and inter-varietal diversity,
processing and directing marketing of produce on the farm, agro-tourism, creation of financial
reserves, forestry work, off-farm investments and the share of off farm income. We regress
these diversification activities on insurance uptake, accounting for several control variables.
Specifications are estimated for every individual diversification activity. In addition, we
estimate a specification including all seven diversification activities to test whether the initial
coefficient changes.

We estimate specifications without control variables to test the sensitivity of the estimates when
control variables are included. The Variance Inflation Factor is calculated to test for
multicollinearity between variables. In addition, we calculate correlation coefficients between
diversification activities and insurance uptake to check if there are high correlations between
these activities. Furthermore, diversification activities were summarized in one indicator to see
if the resulting interaction effects are relevant.
A probit model was chosen for our main analysis.
8
, We estimate the variations of the following
equation:
      
      
         (1)
 indicates whether or not the farmer has crop insurance, is the inter-
varietal diversity (i.e. Shannon Index),   is a dummy variable for
processing and direct marketing,  is a dummy variable indicating whether the
farmer offers agro-tourism on the farm,    indicates the creation of
financial reserves, is forestry work,    is off-farm investment not
related to agriculture,    is the share of off-farm income,  is a vector of
control variables including risk preferences, organic farming, specialization of farm, farm size,
land tenure, weather variables (precipitation, temperature, hail events), distance to the next
center,  are production fixed effects (berries, cherries, plums and grapes),  are period fixed
effects (2017 and 2018) and  is the residual term. Moreover,  is a constant,  ,  , 
   identify the relationship between insurance uptake and diversification activities,
respectively, and  controls for the influence of control variables. The standard errors are
clustered at the cantonal level, reflecting possible structural differences in extension and
organization of producers.. Clustering at the cantonal level, which is a sub-national geographic
boundary level in Switzerland, allows for correlation between observations within one canton,
thus also spatial autocorrelation. Additionally, as we pool observations from different years,
correlation of observations within one canton over the years is also allowed. Clustering the error
terms, i.e. also allowing for spatial autocorrelation, affects the standard errors of our regression
estimates and therefore the significance of our results. In addition, given our relatively small
8
Our results do not change when using logistic or OLS regression.

number of clusters, i.e. there are only 26 cantons in Switzerland, we use the wild bootstrap to
compute standard errors (Cameron et al., 2008).
The Oster bounds (Oster 2019) approach and the Generalized Sensitivity Analysis of Imbens
(2003) serve as robustness tests. The Oster bounds analysis (Oster 2019) indicates whether
coefficients for significant variables are robust against omitted variable bias. The Generalized
Sensitivity Analysis of Imbens (2003) quantifies the extent of the bias which would result from
the absence of each of the control variables included in our analysis.
5. Results
Table 3 summarizes our main results. Columns 1-7 present the results of the restricted
specifications for every diversification measure independently. Column 8 shows the full
specification with all diversification measures included.
We find that there is a pattern:
i) On-farm diversification activities are positively associated with insurance uptake,
except for the creation of financial reserves
ii) Off-farm diversification activities are negatively associated with insurance uptake
iii) Farmers with higher financial reserves are less likely to buy an insurance.
As can be seen in column (1) and column (8), the Shannon Index is positively and significantly
associated with insurance uptake. A one-unit increase in the Shannon Index (increased inter-
varietal diversity), ceteris paribus, is associated to a 10 percentage points increase in the
probability that a farmer has an insurance. Column (2) and column (8) indicate that processing
and direct marketing is positively and significantly associated with insurance uptake. There is
a 7 percentage point (column (8)) higher likelihood that farmers who process and market their
products directly will have insurance. This highlights that the Shannon Index and the processing
and direct marketing of farm produce are complements to insurance uptake. As shown in
column (3) and column (8), the coefficient for agro-tourism is not statistically significant.
Creation of financial reserves is not statistically significant in the restricted specification
(column 4), but it gains significance in the full specification in column (8). Farmers who create
financial reserves are 8 percentage points (column 8) less likely to have insurance.

Table 3. Marginal correlation results insurance probit: Specification 1-8
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Diversification activities
Shannon
Index
0.13***
(95% CI:
0.06, 0.2)
0.10**
(95% CI :
0.02,0.18)
Processing
and direct
marketing
0.11***
(95% CI:
0.05,0.16)
0.07***
(95% CI :
0.02,0.12)
Agro-
tourism
0.05
(95% CI:
-
0.04,0.14)
-0.01
(95% CI :
-0.08,0.06)
Creation of
financial
reserves
-0.05
(95% CI :
-
0.15,0.04)
-0.08**
(95% CI :
-0.16,-
0.00)
Forestry
work
-0.23***
(95% CI:
-0.29,-
0.15)
-0.23***
(95% CI :
-0.33,-
0.12)
Off-farm
investment
0.02
(95% CI:
-0.13,-
0.17)
-0.02
(95% CI :
-0.14,0.11)
Share of off-
farm income
-0.17***
(95% CI:
-0.25,-
0.08)
-0.13***
(95% CI :
-0.19,-
0.07)
Control
variables
Risk preferences, organic production systems, focus specialized9, land tenure, farm size, mean
precipitation, mean precipitation squared, mean temperature, mean temperature squared, hail years, hail
last year, distance
Mcfadden R2
0.128
0.117
0.109
0.110
0.108
0.123
0.123
0.150
Mcfadden
Adjusted R2
0.096
0.085
0.077
0.078
0.076
0.086
0.088
0.103
Number
observations
881
881
881
881
881
787
787
787
AIC*n
1080.304
1093.251
1102.539
1101.619
1103.381
1103.381
974.655
948.581
BIC*n
1171.144
1184.091
1193.379
1192.459
1194.221
1194.221
1063.351
1041.946
Deviance
1042.304
1055.251
1064.539
1063.619
1065.381
1065.381
936.655
908.581
Note: Estimation method with a probit regression with the dependent variable insurance use (1/0). Coefficients are marginal correlations.
Confidence intervals (95%) are in parenthesis. *,** and *** denote significance at the 10%, 5% and 1% level. Standard errors are clustered at
cantonal level. On-farm diversification activities (except for the Shannon Index) and off-farm diversification activities are dummy variables.
Column (5) illustrates the restricted specification for forestry work. The coefficient for forestry
work remains the same in the full specification (8). Farmers who undertake forestry work are
23 percentage points less likely to have an insurance. As shown in column (6), off-farm
investment is not statistically significant in the restricted specification and this also applies to
the full specification (8) where off-farm investment is likewise not statistically significant.
Finally, as shown in column (7), the share of off-farm income is negatively associated with
insurance uptake and is also statistically significant and negatively associated with insurance
9
According to a reviewer’s comment, the “focus specialized” variable might itself be an indication for
diversification. We therefore point the interested reader to the appendix, where we the results of our full model
and multiple sub-variants thereof can be found.

uptake in column (8) of the full model. Farmers who have a higher share of off-farm income
are 13 percentage points less likely to have an insurance.
As can be seen in Table 3, control variables are included in all presented models (see Appendix
Table A6 for marginal effects of control variables). When performing the full specification
without control variables, the coefficients were similar in sign, magnitude and significance (see
Table A6). In another sensitivity tests, we ran OLS, logit regressions and OLS regression with
wild boot cluster (see Appendix Table A8-A10, Table A13)
10
and found no differences in
results. In addition, Table 3 shows that the coefficients which are statistically significant in the
restricted specification remain statistically significant in the full specification, which
demonstrates the robustness across various specifications of the results presented here.
As a robustness check, we use a cluster analysis to find potential interrelations between the
different diversification strategies. The results are to be found in the Appendix (see Table A12)
and indicate that certain diversification measures tend to be adopted in bundles more often than
others, namely i) processing and direct marketing and agro-tourism, ii) creation of financial
reserves and off-farm investment, iii) work off-farm and agro-tourism. Since these results
indicate slight interactions between the activities themselves, we use the Variance Inflation
Factor to show that our main set of results, i.e. the regression analysis is not affected by
multicollinearity. The Variance Inflation Factor is below 2 and the correlations between
diversification activities are ≤ 0.3 (see Appendix Table A10-A11, Fig.A1). Therefore, we
conclude that our results on the relation between the different diversification strategies and the
hail insurance uptake are unaffected by potential correlations between the diversification
strategies itself.
The results of an Oster bounds analysis (Oster 2019) indicate that all significant coefficients
are quite robust and selection on unobservables would need to be at least twice as important as
selection on observables to meaningfully change our results (see Appendix, Table A14).
The Generalized Sensitivity Analysis (see Appendix, Figure A2) based on Imbens (2003) and
Harada (2013) quantifies the extent of the bias that would result from the absence of each
included covariate, thus allowing the researcher to consider the plausibility of the existence of
10
Results of correlation show low collinearity between the on-farm/off-farm management strategies variables. We
assess multicollinearity with the general variance of inflation factor for categorical variables and the variance
inflation factor for the continuous variables chosen for our specifications (see Appendix, Table A9-A10).

actually omitted confounding variables with the same effect. The plotted contour shows at
which point such confounding variables would have the potential to meaningfully change our
results. In our study, selection is not indicated to be a first order issue (see Appendix, Figure
A2).
6. Discussion
Our analysis sheds new light on farmers’ management of climate risks. More specifically, we
revisit the relationship between crop insurance use and diversification, on and off the farm. The
here considered diversification strategies can all be used to reduce weather risks on the farm.
For instance, the inter-varietal diversity is a way for the farmers to spread the growth period of
different fruit varieties over the year. Other diversification strategies which we include here are
off-farm strategies, such as forestry work, reducing the exposure of the general income
(including all off- and on-farm activities) to weather risks.
Our results highlight the value of breaking down diversification activities instead of bundling
them into one single index, as there is a tendency towards a positive association between on-
farm diversification activities and insurance uptake while the association between off-farm
activities and insurance uptake tends to be negative. Our results reveal one exception to this
pattern, namely the creation of financial reserves. In fact, saving money for bad times follows
the off-farm diversification pattern and is negatively correlated with insurance uptake. Note that
saving money for bad times is not a clear-cut on- or off-farm diversification activity, as the
money saved can be used for purposes other than agricultural activities. Money itself is directly
accessible to the farmer, whereas this is not the case with processing and direct marketing. In
addition, resources saved could technically also be shifted out of agriculture. This is further
suggested by the results of the cluster analysis performed, which identifies that off-farm
investment and creation of financial reserves are part of the same cluster. We presume that
farmers who build up financial reserves are saving money for off-farm investments.
Farmers choose diversification strategies from a range of portfolios. There is clear evidence
that farmers may still be insured even if they pursue diversification activities on the farm. Our
results highlight that policies encouraging insurance uptake will not necessarily have a negative
impact on diversification on the farm. Diversification itself is an important strategy for farms,
especially in Europe as it not only enhances the farms’ social cohesion, but is also strongly
linked to their multifunctionality (Van der Ploeg and Roep 2003, Boncinelli et al., 2017, Weltin

et al., 2017). We find a positive association between inter-varietal diversity and insurance
uptake. In current literature, crop diversity or species diversity is associated negatively with
insurance. In their study on Italy, Di Falco et al., (2014) state that diversification of crops (i.e.
monoculture vs. diversified crops) is a negatively correlated with insurance. However, the
context of our case-study must be considered, as we analyze extreme events such as hail.
Enjolras and Sentis (2011) show that while insurance can act as a substitute for less diversified
farms, it can also be seen as a complement when facing major hazards. Single hazard events,
such as hail, constitute high impact but often low probability losses (Bocquého et al., 2014).
Farmers might overestimate this risk and thus take an insurance (Bocquého et al., 2014).
Consequently, although farmers go in for on-farm diversification to manage risk, they still rely
on a crop insurance in this context. In addition, extreme events are predicted to increase due to
climate change. This might also motivate farmers to have insurance in spite of their
diversification activities (Botzen et al., 2010). Finally, it is important to highlight that in
addition to its effects on risk management, varietal diversity may provide other private and
public benefits, such as lower pest pressure and other ecosystem services (Baumgärtner 2007,
Di Falco et al., 2014, Schaub et al., 2020). Thus, risk management strategies may not be the
only motivation for fruit growers in our sample to go in for inter-varietal diversity.
Entrepreneurial literature indicates that processing and direct marketing can be regarded as a
risky activity. Farmers face a number of challenges when establishing a new business model.
They must identify their available resources and products, as well as customer requirements
(Ferguson and Hansson 2015). Thus, given the riskiness of this business, farmers might also
decide to take out an insurance to reduce the risk. From a moral hazard perspective, insured
farmers might decide to go in for more risky diversification activities, such as processing and
direct marketing. On the other hand, diversification is usually perceived as risk decreasing. To
contextualize our results, we suggest that although processing and direct marketing is
considered as a diversification activity, in the case of farmers who focus on fruit production, it
is rather more closely related to the production of fruit on the farm. Assuming that farmers
process fruit, the harvest might not be available for this diversification activity if it is damaged
by hail. Given that this activity relies on the farm’s own produce, it might explain the positive
correlation with insurance uptake. Along these lines, Mishra and Goodwin (2003) suggest that
farmers who earn most of their income from farming activities are more likely to be insurance
users.

There is a close relationship between our research and studies that identify the determinants for
insurance uptake. For example, this literature puts forward that off-farm income diversifies a
farm’s income and thus reduces the probability that a farmer will take an insurance, given that
these farmers have more self-insurance possibilities (Mishra and Goodwin 2006, Velandia et
al., 2009). In the context of on-farm, off-farm diversification activities and insurance uptake, it
is important to note that insurance uptake deals with yield risk, while diversifying per se deals
with income risk, of which yield risk is a component (Blank and McDonald, 1996). Note that
in the case of crop insurance, there exists an alternative hedging instrument for hail risk for
some types of crops, for example hail nets (Rogna et al., 2019) which we did not include in our
study. However, hail risks remain even if using hail nets (e.g. Gandorfer et al., 2015)). In fact,
in Switzerland insurance contracts incentivize the combination of nets and insurance by offering
lower premiums. Thus, these are often used in combination. Off-farm management strategies
reduce the variance of the total farm income as these income sources are independent of weather
and hail (Finger and Lehmann 2012). Our results indicate that off-farm diversification activities
tend to be negatively correlated with insurance uptake. However, it can go both ways, namely
the fruit grower has off-farm activities and therefore he/she decides not to take an insurance, or
it is impossible for the fruit grower to take an insurance due, for instance, budget constraints
and decides to pursue off-farm activities to meet the risk.
7. Conclusion
In this paper, we analyze a rich dataset from 1176 Swiss fruit growers and investigate the
relationship between insurance uptake and various diversification strategies, namely, inter-
varietal diversity, processing and direct marketing of the produce, agro-tourism, creation of
financial reserves, forestry work, off-farm investment and share of off farm income. The hail
insurance but also the diversification activities contribute to manage climate risks of farmers.
Hail insurance is the dominant insurance type in Swiss and other central European crop
production, so that quantifying its interrelation with other risk management strategies is crucial
to understand farmers’ climate risk management decisions, and subsequently to develop
targeted policies. We include a vector of control variables with farm and farmer characteristics,
risk exposure and weather data. Up until now, literature rates insurance as a replacement to
diversification measures and sometimes as a complement to diversification measures. In this
paper, we revisit this relationship by considering different types of management strategies,

namely both on-farm and off-farm diversification activities, adopted by fruit growers in
Switzerland.
Our results show that on-farm activities are positively correlated with insurance uptake and are
thus complementary risk management strategies for farmers, while off-farm activities have a
negative correlation with insurance uptake, and thus represent substitutes in a farmer’s risk
management portfolio.
For policy makers, our findings show that interdependencies between risk management
strategies must be taken into account when providing support for such strategies. If more
successful policies are to be developed, it is essential to understand the association between
different diversification activities and insurance uptake and indeed to recognize that the
associations themselves differ depending on the type of diversification activity. For example,
support for both on-farm diversification and insurance might be efficient because both are
substitutes. Along these lines, this could also mean that insurance does not necessarily reduce
varietal diversity on farms. In turn, this suggests that insurance does not per se hinder optimal
diversification on the farm. Moreover, insurance providers must understand how diversification
activities relate differently to insurance uptake as it may help them design insurance for farmers
who are currently not well served by the products on offer. We suggest that insurance providers
may need to develop tailored options for farmers who rely on off-farm income.
In addition, the types of strategies chosen by farmers will have a certain impact on the structure
of the farming sector in future. When incentivizing insurance or the use of other on-farm and
off-farm diversification strategies, policy makers should allow for the interdependencies, e.g.
pay attention to farmers’ responses to avoid contradictory policy instruments.
Future research should seek to further probe the generality of our results. In particular, the
interplay of crop and inter-variety diversity and the associated ecological benefits, as well as
different crop insurance schemes merit more attention. Moreover, observed diversification
strategies and insurance uptake across different types of farm production must be investigated
to test whether different patterns emerge.

References
Bardají, I., A. Garrido, I. Blanco, A. Felis, J. Sumpsi, T. García-Azcárate, G. Enjolras and F. Capitanio
(2016). "Research for Agri committee-State of play of risk management tools implemented by member
states during the period 2014-2020: national and European frameworks." Agriculture and Rural
Development, European Parliament
Barrett, C. B., T. Reardon and P. Webb (2001). "Nonfarm income diversification and household
livelihood strategies in rural Africa: concepts, dynamics, and policy implications." Food policy 26(4):
315-331.
Baumgärtner, S. (2007). "The insurance value of biodiversity in the provision of ecosystem services."
Natural Resource Modeling 20(1): 87-127.
Blank, S. C., & McDonald, J. (1996). Preferences for crop insurance when farmers are diversified.
Agribusiness: an international journal, 12(6), 583-592.
Bocquého, G., F. Jacquet and A. Reynaud (2014). "Expected utility or prospect theory maximisers?
Assessing farmers' risk behaviour from field-experiment data." European Review of Agricultural
Economics 41(1): 135-172.
Boncinelli, F., F. Bartolini, L. Casini and G. Brunori (2017). "On farm non-agricultural activities:
geographical determinants of diversification and intensification strategy." Letters in Spatial and
Resource Sciences 10(1): 17-29.
Botzen, W., L. M. Bouwer and J. Van den Bergh (2010). "Climate change and hailstorm damage:
Empirical evidence and implications for agriculture and insurance." Resource and Energy Economics
32(3): 341-362.
Cameron, A. C., Gelbach, J. B. and Miller, D. L. (2008). "Bootstrap-based improvements for inference
with clustered errors." The Review of Economics and Statistics 90(3): 414-427.
Chang, H.-H. and A. K. Mishra (2012). "Chemical usage in production agriculture: Do crop insurance
and off-farm work play a part?" Journal of environmental management 105: 76-82.

Chavas, J.-P. and G. Shi (2015). "An economic analysis of risk, management, and agricultural
technology." Journal of agricultural and resource economics: 63-79.
Chavas, J. P. and S. Di Falco (2012). "On the role of risk versus economies of scope in farm
diversification with an application to Ethiopian farms." Journal of agricultural economics 63(1): 25-55.
Coble, K. H. and B. J. Barnett (2012). "Why do we subsidize crop insurance?" American Journal of
Agricultural Economics 95(2): 498-504.
Dalhaus, T., Schlenker, W., Blanke, M. M., Bravin, E., & Finger, R. (2020). „The Effects of Extreme
Weather on Apple Quality.” Scientific reports, 10(1), 1-7.
Di Falco, S., & Chavas, J.-P. (2009). On crop biodiversity, risk exposure, and food security in the
highlands of Ethiopia. American Journal of Agricultural Economics, 91(3), 599-611.
Di Falco, S., F. Adinolfi, M. Bozzola and F. Capitanio (2014). "Crop insurance as a strategy for
adapting to climate change." Journal of Agricultural Economics 65(2): 485-504.
Dohmen, T., A. Falk, D. Huffman, U. Sunde, J. Schupp and G. G. Wagner (2011). "Individual risk
attitudes: Measurement, determinants, and behavioral consequences." Journal of the European
Economic Association 9(3): 522-550.
El Benni, N., R. Finger and S. Mann (2012). "Effects of agricultural policy reforms and farm
characteristics on income risk in Swiss agriculture." Agricultural Finance Review 72(3): 301-324.
Enjolras, G., F. Capitanio and F. Adinolfi (2012). "The demand for crop insurance: Combined
approaches for France and Italy." Agricultural Economics Review 13(389-2016-23488): 5.
Enjolras, G. and P. Sentis (2011). "Crop insurance policies and purchases in France." Agricultural
Economics 42(4): 475-486.
Ferguson, R. and H. Hansson (2015). "Measuring embeddedness and its effect on new venture
creation—A study of farm diversification." Managerial and Decision Economics 36(5): 314-325.

Fernandez-Cornejo, J., C. Hendricks and A. K. Mishra (2005). "Technology adoption and off-farm
household income: the case of herbicide-tolerant soybeans." Journal of Agricultural and Applied
Economics 37(3): 549-563.
Finger, R. and N. Lehmann (2012). "The influence of direct payments on farmers’ hail insurance
decisions." Agricultural economics 43(3): 343-354.
Frei, C. (2014). "Interpolation of temperature in a mountainous region using nonlinear profiles and
non‐Euclidean distances." International Journal of Climatology 34(5): 1585-1605.
Frei, C. and C. Schär (1998). "A precipitation climatology of the Alps from high‐resolution rain‐gauge
observations." International Journal of climatology 18(8): 873-900.
Gandorfer, M., Hartwich, A., & Bitsch, V. (2015). Hail risk management in fruit production: anti-hail
net versus hail insurance in Germany. Paper presented at the XVIII International Symposium on
Horticultural Economics and Management 1132.
Goodwin, B. K. (1993). "An empirical analysis of the demand for multiple peril crop insurance."
American Journal of Agricultural Economics 75(2): 425-434.
Haigh, T., Takle, E., Andresen, J., Widhalm, M., Carlton, J. S., & Angel, J. (2015). “Mapping the
decision points and climate information use of agricultural producers across the US Corn Belt.”
Climate Risk Management, 7, 20-30.
Hajjar, R., D. I. Jarvis and B. Gemmill-Herren (2008). "The utility of crop genetic diversity in
maintaining ecosystem services." Agriculture, Ecosystems & Environment 123(4): 261-270.
Hansson, H., R. Ferguson, C. Olofsson and L. Rantamäki-Lahtinen (2013). "Farmers' motives for
diversifying their farm business–The influence of family." Journal of Rural Studies 32: 240-250.
Harada, M. (2013). Generalized sensitivity analysis and application to quasi-experiments, Working
Paper, New York University.

Hennessy, D. A. (1998). "The production effects of agricultural income support policies under
uncertainty." American Journal of Agricultural Economics 80(1): 46-57.
Hennessy, T. C. and T. Rehman (2008). "Assessing the impact of the ‘decoupling’reform of the
common agricultural policy on Irish farmers’ off‐farm labour market participation decisions." Journal
of Agricultural Economics 59(1): 41-56.
Imbens, G. W. (2003). "Sensitivity to exogeneity assumptions in program evaluation." American
Economic Review 93(2): 126-132.
Iyer, P., Bozzola, M., Hirsch, S., Meraner, M., Finger, R. (2020) “Measuring Farmer Risk Preferences
in Europe: A Systematic Review.” Journal of Agricultural Economics. 71(1): 3-26
Knapp, L., E. Bravin and R. Finger. (2018). "DROSOPHRISK: Data from surveys on agricultural risk
management strategies in Switzerland in response to Drosophila suzukii." ETH Research Collection.
Lefebvre, M., D. Nikolov, S. Gomez-y-Paloma and M. Chopeva (2014). "Determinants of insurance
adoption among Bulgarian farmers." Agricultural Finance Review 74(3): 326-347.
Lesk, C., Rowhani, P., & Ramankutty, N. (2016). “Influence of extreme weather disasters on global
crop production.” Nature, 529(7584), 84-87.
Lunt, T., Jones, A. W., Mulhern, W. S., Lezaks, D. P., & Jahn, M. M. (2016). “Vulnerabilities to
agricultural production shocks: An extreme, plausible scenario for assessment of risk for the insurance
sector.” Climate Risk Management, 13, 1-9.
Mase, A. S., Gramig, B. M., & Prokopy, L. S. (2017). “Climate change beliefs, risk perceptions, and
adaptation behavior among Midwestern US crop farmers.” Climate Risk Management, 15, 8-17.
McNamara, K. T. and C. Weiss (2005). "Farm household income and on-and off-farm diversification."
Journal of Agricultural and Applied Economics 37(1): 37-48.

Meraner, M. and R. Finger (2019). "Risk perceptions, preferences and management strategies:
evidence from a case study using German livestock farmers." Journal of Risk Research 22(1): 110-
135.
Meraner, M., W. Heijman, T. Kuhlman and R. Finger (2015). "Determinants of farm diversification in
the Netherlands." Land Use Policy 42: 767-780.
Mishra, A. K., H. S. El-Osta and C. L. Sandretto (2004). "Factors affecting farm enterprise
diversification." Agricultural finance review 64(2): 151-166.
Mishra, A. K. and B. Goodwin (2003). "Adoption of crop versus revenue insurance: a farm-level
analysis." Agricultural Finance Review 63(2): 143-155.
Mishra, A. K. and B. Goodwin (2006). "Revenue insurance purchase decisions of farmers." Applied
Economics 38(2): 149-159.
Möhring, N., Dalhaus, T., Enjolras, G., & Finger, R. (2020).” Crop insurance and pesticide use in
European agriculture.” Agricultural Systems, 184, 102902
Moschini, G. and D. A. Hennessy (2001). "Uncertainty, risk aversion, and risk management for
agricultural producers." Handbook of agricultural economics 1: 87-153.
Müller, B., Johnson, L., & Kreuer, D. (2017). Maladaptive outcomes of climate insurance in
agriculture. Global Environmental Change, 46, 23-33.
O’Donoghue, E. J., M. J. Roberts and N. Key (2009). "Did the federal crop insurance reform act alter
farm enterprise diversification?" Journal of Agricultural Economics 60(1): 80-104.
Oster, E. (2019). "Unobservable selection and coefficient stability: Theory and evidence." Journal of
Business & Economic Statistics 37(2): 187-204.
Richards, T. J. (2000). "A two-stage model of the demand for specialty crop insurance." Journal of
agricultural and resource economics 25(1835-2016-149081): 177-194.

Rogna, M., Schamel, G., & Weissensteiner, A. (2019). Modeling the Switch from Hail Insurance to
Anti-Hail Nets. Available at SSRN 3424071. doi:http://dx.doi.org/10.2139/ssrn.3424071
Roll, K. H. (2019). "Moral hazard: the effect of insurance on risk and efficiency." Agricultural
Economics 50(3): 367-375.
Schaub, S., N. Buchmann, A. Lüscher and R. Finger (2020). "Economic benefits from plant species
diversity in intensively managed grasslands." Ecological Economics 168: 106488.
Schaub, S., Finger, R., Leiber, F., Probst, S., Kreuzer, M., Weigelt, A., Buchmann, N., Scherer-
Lorenzen, M. (2020). “Plant diversity effects on forage quality, yield and revenues of semi-natural
grasslands.” Nature Communications 11 (768)
Schemm, S., L. Nisi, A. Martinov, D. Leuenberger and O. Martius (2016). "On the link between cold
fronts and hail in Switzerland." Atmospheric science letters 17(5): 315-325.
Smale, M. (ed) (2006) Valuing Crop Biodiversity: On-Farm Genetic Resources and Economic
Chance, CABI, Cambridge,MA
Smale, M. (ed) (2006) Valuing Crop Biodiversity: On-Farm Genetic Resources and Economic
Chance, CABI, Cambridge,MA
Smith, V. H., & Goodwin, B. K. (1996). Crop insurance, moral hazard, and agricultural chemical use.
American Journal of Agricultural Economics, 78(2), 428-438. doi:https://doi.org/10.2307/1243714
Swiss Federal Statistical Office 2018. (2018). "Atlas statistique de la Suisse: Exploitations et surface
agricole utile 2018." From
https://www.atlas.bfs.admin.ch/maps/13/fr/14338_5882_5872_4801/22888.html.
Trnka, M., Rötter, R. P., Ruiz-Ramos, M., Kersebaum, K. C., Olesen, J. E., Žalud, Z., & Semenov, M.
A. (2014). « Adverse weather conditions for European wheat production will become more frequent
with climate change.” Nature Climate Change, 4(7), 637.

Van der Ploeg, J. D. and D. Roep (2003). "Multifunctionality and rural development: the actual
situation in Europe." Multifunctional agriculture: a new paradigm for European agriculture and rural
development: 37-53.
Velandia, M., R. M. Rejesus, T. O. Knight and B. J. Sherrick (2009). "Factors affecting farmers'
utilization of agricultural risk management tools: the case of crop insurance, forward contracting, and
spreading sales." Journal of agricultural and applied economics 41(1): 107-123.
Vroege, Willemijn, and Robert Finger (2020). "Insuring Weather Risks in European Agriculture."
Eurochoices 19(2): 54-62
Webber, H., Ewert, F., Olesen, J. E., Müller, C., Fronzek, S., Ruane, A. C., Bourgault, M., Marte, P.,
Ababaei, B., Bindi, M., Ferrise, R., Finger, R., Fodor, N., Gabaldón-Leal, C., Gaiser, T., Jabloun, M.,
Kersebaum, K.C., Lizaso, J.I., Lorite, I.J., Manceau, L., Moriondo, M., Nendel, C., Rodríguez, A.,
Ruiz-Ramos, M., Semenov, M.A., Siebert, S., Stella, T., Stratonovitch, P., Trombi, G., Wallach, D.
(2018). “Diverging importance of drought stress for maize and winter wheat in Europe.” Nature
Communications, 9(1), 4249.
Weltin, M., I. Zasada, C. Franke, A. Piorr, M. Raggi and D. Viaggi (2017). "Analysing behavioural
differences of farm households: An example of income diversification strategies based on European
farm survey data." Land use policy 62: 172-184.
Wright, B. D. and J. A. Hewitt (1994). All-risk crop insurance: lessons from theory and experience.
Economics of agricultural crop insurance: theory and evidence, Springer: 73-112.
Wuepper, D., H. Yesigat Ayenew and J. Sauer (2018). "Social Capital, Income Diversification and
Climate Change Adaptation: Panel Data Evidence from Rural Ethiopia." Journal of Agricultural
Economics 69(2): 458-475.
Yu, J. and D. A. Sumner (2018). "Effects of subsidized crop insurance on crop choices." Agricultural
economics 49(4): 533-545.

9 Appendix
Table A1 provides an overview of fruit growers who participated to our survey per year and
per fruit type. Table A2 and Table A3 show that there is little variation regarding insurance
coverage and diversification measures. Therefore, we chose to use the cross-sectional dataset
and pooled the data.
Table A1. Overview data: Number of fruit growers (per fruit) participating in the survey
Fruits
2016
2017
2018
Total sample
Berries
-
X
N=47
-
47
Cherries
-
X
N=84
X
N=67
151
Plums
X
N=99
X
N=37
X
N=54
190
Grapes
X
N= 370
X
N= 205
X
N=213
788
Total
476
373
334
1176
Note: The table above provides an overview of the number of participants, per survey, per year. The survey for berries was
only undertaken in 2017 and the survey for cherries was only undertaken in 2017 and 2018.
Table A2. Unbalanced panel data: percentage switching with regard to insurance contract
Fruit
2016-2018
Switching
Not switching
Berries
-
-
-
Cherries
37
2 (5%)
35(95%)
Plums
55
9(16%)
46(84%)
Grapes
230
14 (6%)
216 (94%)
Total
322
25† (8%)
297(92%)
Note: The table above provides an overview of the number of fruit growers for whom we have unbalanced panel data between 2016 and 2018.
†Of the 25 switching, 8 fruit growers adopt insurance and 25 exit the insurance scheme.

Table A3. Unbalanced panel data: percentage switching for diversification activities
Total
N=322
Agro-
tourism
Processi
ng
Creation of financial
reserves
Forestry
work
Off-farm
investment
Share of off-farm
income
Berries
-
-
-
-
-
-
Cherries
N=37
5
5
8
1
4
3
Plums
N=55
5
3
4
1
0
6
Grapes
N=230
31
44
47
4
18
37
Total
41
52
59
6
22
46
%
13%
16%
18%
2%
7%
14%
Table A4 provides an overview of variables included in the probit specifications and provides
a detailed description together with the hypothesis related to these variables as found in current
literature.
Table A4. Overview of variables included in probit specifications
Variables
Hypothesis
Literature
Data collected
Shannon
Diversity Index
Crop diversification is a way to buffer
risk. We refer to diversification within a
same crop via different varieties.
-
Bradshaw et
al., (2004)
Hazell (1992)
Fujisawa and
Kobayashi
(2011)
(Di Falco et
al., 2014)
Shannon Diversity Index for varieties
represents the evenness and diversity
of the varieties the fruit grower has.
We use the surface area for each
variety and calculate a Shannon
Diversity Index for each fruit grower
(0= low diversity).
Processing and
direct
marketing of
products
Agro-tourism
Creation of
financial
reserves
Forestry work
Off-farm
investment
Share of off-
farm income
Income diversification can substitute the
use of insurance. However, both
insurance and farm diversification can be
considered for very risk averse farmers.
The relationship doesn’t need to be
negative. A farmer might use a number
of risk management tools rather than just
one.,
+/-
Goodwin
(1993)
Mohammed
and Ortmann
(2005)
Velandia et
al.,(2009)
Lefebvre et
al., (2014)
Processing and direct marketing are
defined as processing the fruits or
selling them directly to the customers.
(1=yes,0=no):
Other activities on the farm such as
agro-tourism.
Creation of financial reserves is
defined as saving money for bad
periods.
Forestry work is defined as
undertaking forestry work.
Off-farm investment is defined as
investing in other companies or real
estate.
(1=yes,0=no)
Farmers having more off-farm income
(i.e. income not linked to the farm
activities) are likely to self-insure
-
Mishra and
Goodwin
(2003)
Share of off-farm income is 1 if less
than 50% of earnings originate from
farm and 0 if more than 50%
originates from farm.
Control variables
Risk
preferences
The more risk averse, the more willing
farmers are to use insurance.
+
Lefebvre et
al. (2014)
Hardaker
(2004)
Mean of fruit growers risk aversion
over four domains: agriculture,
production, market and price, external
financing.

Variables
Hypothesis
Literature
Data collected
Likert scale : 10= not willing to take a
risk
0 = very willing to take a risk.
Organic
farming
Organic farmers are less risk averse than
conventional farmers and are less likely
to buy an insurance contract
-
Gardebroek
(2006)
Production system is organic or not
(1=yes,0=no)
Focus
specialized
Insurance can be a substitute for less
diversified farms (on-farm activities)
+
Enjolras and
Sentis (2011)
O’Donoghue,
et al., (2009)
Focus specialized category indicates
the farm type, i.e. whether a fruit
is source of main income (i.e.
berries, cherries, plums or grapes).
(1=yes,0=no):
Farm size
Farmers with larger farms are more
likely to have an insurance.
On the other hand, Nance et al. (1993)
suggest that smaller farms are more
likely to take insurance whereas big
farms might be able to diversify more.
+/-
Goodwin
(1993)
Lefebvre et
al. (2014)
Nance et al.,
(1993)
Total farmland in 100 square meters
Land tenure
A higher ratio of owned surface is
associated with a larger capacity to bear
risk and a lower need for insurance use.
+
Sherrick et
al.,(2004)
Velandia et
al. (2009)
de Mey et al.
(2016)
Percentage of land leased by the fruit
grower
(0 if ≤ 50%, 1 if >50%)
Mean
precipitation
Mean
temperature
Weather variables must be considered
Di Falco et al.
(2014)
Mean precipitation is the average
rainfall over the years 1961-2016 for
the municipality. (Frei and Schär
1998, Frei 2014).
Mean temperature is the average
temperature over the years 1961-2016
for the municipality. (Frei and Schär
1998, Frei 2014).
Hail last year
Hail years
For example, growers might also use a
crop insurance years after a major hail
event (Goodwin, 1993). In fact, Rydant
(1979) suggests that uptake of crop
insurance is linked to recent perception
of hail damage. Thus, if there were no
extreme hail events, the adoption of crop
insurance might be lower.
Goodwin
(1993)
Rydant
(1979)
Gardebroek
(2006)
Hail last years is the number of hail
events in the year prior to data
collection for each municipality.
(Finger and Lehmann 2012)
Hail years is the number of years with
hail events for each municipality in the
period 1961–2004.( Finger and
Lehmann 2012)
Distance
Farmers who attend annual meetings
organized by the local crop insurance are
more likely to purchase insurance. In our
case, we do not have access to this
information but it is interesting to see
whether distance to local centers has an
effect since, proximity to a center means
better access to information.
Menapace et
al., (2015)
Distance measured in seconds from
the location of the fruit grower to the
next town /urban center.
Year
Year the fruit grower filled out the
survey
Fruit
Type of survey the fruit grower
responded to (berries, cherries, plums
or grapes)
Cantons
Cantons where the fruit growers are
located.
Note: T he table above provides an overview of variables considered in the probit specifications and for each it includes a hypothesis with
literature references, a definition and the name of the variable.

Table A5 provides an overview of the number of fruit growers insured per fruit type
Table A5. Overview of insurance uptake per fruit grower
Fruit
Insured
Not insured
Berries
20
(43%)
27
(57%)
Cherries
44
(30%)
104
(70%)
Plums
76
(40%)
114
(60%)
Grapes
296
(38%)
483
(62%)
Total
436
(37%)
728
(63%)
Note: T he table above gives the number and percent age of insured growers and non -insured growers per fruit.
Table A6 provides the probit specifications including the coefficients for control variables

Table A6. Probit Specification 1-9
Variable s
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Diversification activities
Shannon Index
0.13***
(0.04)
0.10**
(0.04)
0.09**
(0.04)
Processing and direct
marketing
0.11***
(0.03)
0.07***
(0.03)
0.06
(0.04)
Agro-tourism
0.05
-0.01
-0.02
(0.05)
(0.04)
(0.03)
Creation of financial
reserves
-0.05
(0.05)
-0.08**
(0.04)
-0.08*
(0.05)
Forestry work
-0.23***
-0.23***
-0.22***
(0.05)
(0.06)
(0.06)
Off-farm investment
0.02
(0.08)
-0.02
(0.07)
0.03
(0.06)
Share of off-farm
income
-0.17***
(0.04)
-0.13***
(0.03)
-0.16***
(0.04)
Control variables
Risk preferences
0.04**
(0.02)
0.03*
(0.02)
0.03*
(0.02)
0.03
(0.02)
0.03
(0.02)
0.03*
(0.02)
0.02
(0.02)
0.03**
(0.02)
Organic product ion
-0.14*
(0.07)
-0.15**
(0.07)
-0.15**
(0.07)
-0.15*
(0.07)
-0.15**
(0.07)
-0.14**
(0.07)
-0.14*
(0.08)
-0.14*
(0.08)
Focus specialized
-0.13***
(0.04)
-0.09*
(0.05)
-0.09*
(0.05)
-0.08*
(0.04)
-0.10**
(0.05)
-0.09*
(0.05)
-0.06
(0.05)
-0.11**
(0.04)
Farm size
0.06**
(0.03)
0.07**
(0.03)
0.07**
(0.03)
0.08**
(0.03)
0.07**
(0.03)
0.07**
(0.03)
0.09**
(0.05)
0.10*
(0.05)
Land tenure
0.03
(0.06)
0.02
(0.06)
0.02
(0.06)
0.01
(0.06)
0.02
(0.06)
0.02
(0.06)
0.02
(0.05)
0.01
(0.05)
Hail last year
0.02
0.00
0.00
0.01
0.01
0.01
0.02
0.02
0.01
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
Hail years
0.09**
0.08***
0.08**
0.08**
0.08**
0.08**
0.08**
0.09**
0.09**
(0.03)
(0.03)
(0.04)
(0.03)
(0.03)
(0.03)
(0.04)
(0.03)
(0.03)
Mean precipitation
squared
-1.19***
(0.23)
-1.08***
(0.21)
-1.10***
(0.20)
-1.09***
(0.20)
-1.10***
(0.20)
-1.09***
(0.20)
-1.02***
(0.26)
-1.09***
(0.25)
-1.07***
(0.25)
Mean precipitation
1.04***
(0.25)
0.90***
(0.22)
0.93***
(0.21)
0.92***
(0.21)
0.93***
(0.21)
0.92***
(0.21)
0.87***
(0.27)
0.95***
(0.27)
0.91***
(0.27)
Mean temperature
-0.28***
(0.09)
-0.29***
(0.10)
-0.30***
(0.11)
-0.30***
(0.11)
-0.29***
(0.10)
-0.27***
(0.10)
-0.28***
(0.10)
-0.27***
(0.09)
-0.25***
(0.08)
Mean temperature
squared
0.32***
(0.10)
0.32***
(0.11)
0.34***
(0.12)
0.33***
(0.11)
0.33***
(0.11)
0.33***
(0.12)
0.31***
(0.11)
0.30***
(0.10)
0.28***
(0.09)
Distance
0.06*
0.07**
0.07*
0.06
0.07*
0.06*
0.05
0.05
0.04
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
Year2017
-0.05
-0.05
-0.05
-0.04
-0.04
-0.05
-0.06
-0.05
-0.05
(0.06)
(0.06)
(0.05)
(0.05)
(0.05)
(0.05)
(0.06)
(0.06)
(0.05)
Year2018
-0.12**
-0.12**
-0.12***
-0.12**
-0.12***
-0.12***
-0.12***
-0.09**
-0.09**
(0.05)
(0.05)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
Berries
0.27**
0.12
0.14
0.13
0.14
0.13
0.19*
0.28**
0.21**
(0.12)
(0.10)
(0.09)
(0.10)
(0.09)
(0.09)
(0.10)
(0.11)
(0.10)
Plums
0.11
0.02
0.01
0.02
0.02
0.02
0.02
0.10
0.12*
(0.07)
(0.05)
(0.05)
(0.05)
(0.05)
(0.05)
(0.05)
(0.07)
(0.07)
Grapes
0.36***
0.22***
0.22***
0.22***
0.23***
0.22***
0.27***
0.37***
0.27***
(0.07)
(0.07)
(0.07)
(0.07)
(0.07)
(0.07)
(0.08)
(0.08)
(0.07)
Mcfadden R2
0.128
0.117
0.109
0.110
0.108
0.123
0.123
0.150
0.129
Mcfadden Adjusted R2
0.096
0.085
0.077
0.078
0.076
0.086
0.088
0.103
0.096
Number observat ions
881
881
881
881
881
787
787
787
895
AIC*n
1080.304
1093.251
1102.539
1101.619
1103.381
1103.381
974.655
948.581
1096.196
BIC*n
1171.144
1184.091
1193.379
1192.459
1194.221
1194.221
1063.351
1041.946
1192.133
Deviance
1042.304
1055.251
1064.539
1063.619
1065.381
1065.381
936.655
908.581
1056.196
Note: Estimation method with a probit regression with the dependent variable insurance use (1/0). Coefficients are marginal correlations.
Standard errors are in parenthesis. *,** and *** denote significance at the 10%, 5% and 1% level.
Table A7 Probit results for specifications where the variable Farm earning is kept at a
categorical level
(1)
(2)

Shannon
Index
0.10**
(0.04)
0.09**
(0.04)
Processing
and direct
marketing
0.07***
(0.03)
0.07***
(0.02)
Agro-
tourism
-0.01
(0.04)
-0.02
(0.04)
Creation of
financial
reserves
-0.08**
(0.04)
-0.08**
(0.04)
Forestry
work
-0.23***
(0.06)
-0.25***
(0.05)
Off-farm
Investment
-0.02
(0.07)
-0.02
Off-farm
income
-0.13***
(0.03)
Farm
earning
percentage
0.199**
(0.09)
0.27***
(0.08)
0.18***
(0.04)
Control
variables
Mcfadden
R2
0.150
0.157
Mcfadden
Adjusted R2
0.103
0.107
Number
observations
787
787
AIC*n
948.581
940.267
BIC*n
1041.946
1033.631
Deviance
908.581
900.267
Note: Estimation method with a probit regression with the dependent variable insurance use (1/0). Coefficients are marginal correlations.
Standard errors are in parenthesis. *,** and *** denote significance at the 10%, 5% and 1% level. Farm earning percent represents the
percentage of earnings originating from the farm. The question in the survey was “ which percentage of your earnings are generated by farming?
0-25%,26-50%,51-75%,76-100%
Table A8 Overview of specifications for the OLS regression, including the coefficients for the
control variables. Table A9 provides the logit regressions.
Table A8. OLS results for specifications 1-9

Variable s
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Shannon Index
0.12***
(0.03)
0.09**
(0.04)
0.08**
(0.03)
Processing and direct
marketing
0.11***
(0.02)
0.07***
(0.02)
0.06
(0.04)
Agro-tourism
0.05
-0.01
-0.03
(0.04)
(0.03)
(0.03)
Creation of financial
reserves
-0.05
(0.02)
-0.08*
(0.04)
-0.07
(0.05)
Forestry work
-0.20***
-0.19***
-0.20***
(0.04)
(0.05)
(0.06)
Off-farm investment
0.03
(0.07)
0.00
(0.06)
0.02
(0.06)
Share of off-farm
income
-0.17***
-0.13***
(0.03)
-0.15***
(0.04)
Control variables
Risk preferences
0.03*
(0.01)
0.02
(0.01)
0.02
(0.01)
0.02
(0.02)
0.02
(0.01)
0.02
(0.01)
0.02
(0.01)
0.02
(0.01)
Organic product ion
-0.12*
(0.06)
-0.14**
(0.07)
-0.14*
(0.07)
-0.14*
(0.07)
-0.13*
(0.07)
-0.13*
(0.07)
-0.12
(0.07)
-0.12*
(0.07)
Focus specialized
-0.13***
(0.04)
-0.09**
(0.04)
-0.09*
(0.04)
-0.09**
(0.04)
-0.10**
(0.04)
-0.09*
(0.04)
-0.08*
(0.04)
-0.11**
(0.04)
Farm size
0.04**
(0.02)
0.05**
(0.02)
0.05**
(0.02)
0.05**
(0.02)
0.05**
(0.02)
0.05**
(0.02)
0.05*
(0.02)
0.05*
(0.03)
Land tenure
0.02
(0.05)
0.01
(0.05)
0.02
(0.05)
0.01
(0.05)
0.02
(0.05)
0.02
(0.05)
0.01
(0.05)
0.01
(0.04)
Hail last year
0.01
0.00
0.00
0.00
0.00
0.00
0.01
0.02
0.01
(0.02)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.02)
Hail years
0.08**
0.08**
0.07**
0.07**
0.07**
0.07**
0.07
0.07**
0.08**
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
Mean precipitation
squared
-0.84***
(0.18)
-0.74***
(0.15)
-0.77***
(0.17)
-0.76***
(0.16)
-0.76***
(0.16)
-0.76***
(0.16)
-0.70***
(0.18)
-0.75***
(0.18)
-0.74***
(0.18)
Mean precipitation
0.72***
(0.20)
0.60***
(0.17)
0.62***
(0.18)
0.61***
(0.18)
0.62***
(0.18)
0.62***
(0.18)
0.59**
(0.21)
0.64***
(0.20)
0.62***
(0.20)
Mean temperature
-0.21**
(0.08)
-0.22**
(0.08)
-0.23**
(0.09)
-0.23**
(0.09)
-0.22**
(0.08)
-0.23**
(0.09)
-0.22**
(0.09)
-0.21**
(0.08)
-0.19**
(0.08)
Mean t emperature
squared
0.25**
(0.09)
0.25**
(0.10)
0.26**
(0.10)
0.26**
(0.10)
0.26***
(0.10)
0.26**
(0.10)
0.24**
(0.10)
0.23**
(0.09)
0.21**
(0.09)
Distance
0.06**
0.06**
0.06**
0.06**
0.06**
0.06
0.05
0.04
0.04*
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.02)
Year2017
-0.04
-0.04
-0.04
-0.03
-0.04
-0.04
-0.05
-0.04
-0.04
(0.05)
(0.05)
(0.05)
(0.05)
(0.05)
(0.05)
(0.05)
(0.03)
(0.04)
Year2018
-0.10**
-0.10**
-0.11**
-0.10**
-0.10**
-0.11**
-0.10
-0.08**
-0.09**
(0.04)
(0.04)
(0.04)
(0.04)
(0.03)
(0.04)
(0.04)
(0.04)
(0.04)
Berries
0.25**
0.11
0.13
0.12
0.12
0.12
0.17*
0.25**
0.18**
(0.11)
(0.08)
(0.09)
(0.09)
(0.09)
(0.09)
(0.09)
(0.1)
(0.08)
Plums
0.10
0.02
0.02
0.02
0.03
0.02
0.02
0.09
0.11*
(0.06)
(0.05)
(0.05)
(0.05)
(0.05)
(0.05)
(0.05)
(0.06)
(0.06)
Grapes
0.36***
0.22***
0.22***
0.22***
0.22***
0.22***
0.27***
0.35***
0.26***
(0.08)
(0.07)
(0.07)
(0.07)
(0.07)
(0.08)
(0.09)
(0.08)
(0.07)
R2
0.152
0.139
0.129
0.130
0.134
0.134
0.146
0.176
0.155
Adjusted R2
0.134
0.122
0.111
0.112
0.116
0.116
0.126
0.150
0.137
Number observations
881
881
881
881
881
881
787
787
895
AIC*n
1144.460
1157.524
1167.822
1167.216
1163.457
1168.617
1033.286
1007.025
1160.166
BIC*n
1235.300
1248.364
1258.662
1258.056
1254.297
1259.458
1121.982
1100.390
1256.103
Deviance
1106.460
119.524
1129.822
1129.216
1125.457
1130.617
995.286
967.025
1120.166
Note: Estimation method with an OLS regression with the dependent variable insurance use (1/0). Standard errors are in parenthesis. *,** and
*** denote significance at the 10%, 5% and 1% level.
Table A9. Logit Specification 1-8
Variable s
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Diversification activities
Shannon
Index
0.13***
(0.04)
0.11**
(0.04)
0.09**
(0.04)

Processing
and direct
marketing
0.11***
(0.03)
0.08***
(0.03)
0.06
(0.04)
Agro-tourism
0.05
-0.01
-0.02
(0.05)
(0.04)
(0.03)
Creation of
financial
reserves
-0.06
(0.05)
-0.09**
(0.04)
-0.08
(0.05)
Forestry work
-0.22***
-0.22***
-0.22***
(0.03)
(0.05)
(0.06)
Off-farm
investment
0.02
(0.08)
-0.01
(0.07)
0.03
(0.06)
Share of off-
farm income
-0.17***
-0.13***
(0.03)
-0.16***
(0.04)
Control variables
Risk
preferences
0.04**
(0.02)
0.03*
(0.02)
0.03*
(0.02)
0.03
(0.02)
0.03*
(0.02)
0.03
(0.02)
0.03
(0.02)
0.04**
(0.02)
Organic
production
-0.14*
(0.07)
-0.15**
(0.07)
-0.15**
(0.07)
-0.15**
(0.07)
-0.15**
(0.07)
-0.15**
(0.07)
-0.14*
(0.08)
-0.14*
(0.08)
Focus
specialized
-0.13***
(0.04)
-0.09*
(0.05)
-0.08*
(0.05)
-0.08*
(0.05)
-0.09*
(0.05)
-0.08*
(0.05)
-0.06
(0.04)
-0.11**
(0.04)
Farm size
0.07
(0.04)
0.09*
(0.05)
0.09*
(0.05)
0.09*
(0.05)
0.09*
(0.05)
0.09*
(0.05)
0.11**
(0.05)
0.11*
(0.06)
Land tenure
0.03
(0.06)
0.02
(0.06)
0.02
(0.06)
0.01
(0.06)
0.02
(0.06)
0.02
(0.06)
0.02
(0.05)
0.01
(0.05)
Hail last year
0.02
0.01
0.01
0.01
0.01
0.01
0.02
0.02
0.00
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
Hail years
0.09**
0.08**
0.08**
0.08**
0.08**
0.08**
0.08**
0.09**
0.09**
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.04)
(0.04)
(0.03)
Mean
precipitation
squared
-1.25***
(0.24)
-1.16***
(0.23)
-1.20***
(0.22)
-1.19***
(0.22)
-1.20***
(0.22)
-1.18***
(0.22)
-1.07***
(0.27)
-1.14***
(0.25)
-1.10***
(0.25)
Mean
precipitation
1.09***
(0.26)
0.97***
(0.23)
1.01***
(0.22)
1.00**
(0.22)
1.01***
(0.22)
1.00***
(0.22)
0.90***
(0.28)
0.98***
(0.26)
0.93***
(0.26)
Mean
temperature
-0.30***
(0.10)
-0.31***
(0.11)
-0.32***
(0.11)
-0.32***
(0.11)
-0.32***
(0.11)
-0.32***
(0.11)
-0.30***
(0.11)
-0.29***
(0.10)
-0.26***
(0.09)
Mean
temperature
squared
0.34***
(0.11)
0.35***
(0.12)
0.37***
(0.12)
0.36***
(0.12)
0.36***
(0.12)
0.36***
(0.12)
0.33***
(0.12)
0.33***
(0.10)
0.29***
(0.10)
Distance
0.06*
0.07**
0.07*
0.07*
0.07*
0.07*
0.05
0.05
0.04
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.03)
Year2017
-0.05
-0.05
-0.05
-0.04
-0.04
-0.04
-0.06
-0.05
-0.05
(0.06)
(0.06)
(0.05)
(0.05)
(0.05)
(0.05)
(0.06)
(0.06)
(0.05)
Year2018
-0.12**
-0.12**
-0.12**
-0.11**
-0.12***
-0.12***
-0.12***
-0.09**
-0.09**
(0.05)
(0.05)
(0.05)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
Berries
0.29**
0.13
0.15
0.13
0.15
0.14
0.20**
0.29**
0.21**
(0.12)
(0.10)
(0.10)
(0.10)
(0.10)
(0.10)
(0.10)
(0.12)
(0.10)
Plums
0.11
0.02
0.02
0.02
0.03
0.02
0.02
0.11
0.12*
(0.07)
(0.05)
(0.05)
(0.05)
(0.06)
(0.06)
(0.06)
(0.08)
(0.07)
Grapes
0.36***
0.23***
0.22***
0.22***
0.23***
0.23***
0.27***
0.37***
0.27***
(0.07)
(0.07)
(0.07)
(0.07)
(0.07)
(0.07)
(0.08)
(0.08)
(0.07)
Mc Fadden’s
R2
0.128
0.118
0.111
0.111
0.130
0.110
0.124
0.150
0.129
Mc Fadden’s
Adjusted R2
0.096
0.086
0.079
0.080
0.083
0.078
0.089
0.104
0.096
Number
observations
881
881
881
881
881
881
787
787
895
AIC*n
1079.434
1091.625
1100.601
1099.547
1095.434
1101.610
973.761
947.706
1096.325
BIC*n
1170.275
1182.465
1191.441
1190.387
1186.274
1192.450
1062.458
1041.070
1192.262
Deviance
1041.434
1053.625
1062.601
1061.547
1057.434
1063.60
935.761
907.706
1056.325
Note: Estimation method with a logit regression with the dependent variable insurance use (1/0). Coefficients are marginal co rrelations.
Standard errors are in parenthesis. *,** and *** denote significance at the 10%, 5% and 1% level.
Table A10 provides an overview of the correlation matrix for on-farm and off-farm
diversification activities and insurance uptake. Table A11 provides results of the variance
inflation factor which tests for multicollinearity among the variables of interest. Table A12

provides the results of a clustering of the diversification activities and shows which
diversification activities are adopted bundled together, or put differently, diversification
activities which are often set going simultaneously by fruit growers in our sample.
Table A.10 Correlation matrix for on-farm and off-farm diversification activities
Variables
Shannon
Index
Agro-
tourism
Creation of
financial
reserves
Off-farm
investment
Processing
and direct
marketing
Forestry
work
Share of
off-farm
income
Insurance
Shannon
Index
1
Agro-tourism
0.09**
1
Creation of
financial
reserves
0.17***
-0.01
1
Off-farm
investment
off-farm
0.04*
0.00
0.12***
1
Processing
and direct
marketing
0.20***
0.18***
-0.04
0.03
1
Forestry
work
-0.01
-0.05*
0.05*
0.00
0.01
1
Share of off-
farm income
-0.33
-0.10
-0.19
-0.04
-0.15
0.00
1
Insurance
0.1***
0.01
-0.01
0.02
0.09***
-0.06**
-0.18***
1
Note: *,** and *** denote significance at the 10%, 5% and 1% level for the correlation coefficients. The correlation coefficients between
different on-farm and off-farm diversification act ivities do not exceed 0.2.
Figure A1. Correlation matrix for on-farm and off-farm diversification activities
Table A11. Variance Inflation Factor

Variables
GVIF
DF
GVIF^(1/(2*Df))
Shannon Index
1.44
1
1.20
Processing and direct marketing
1.20
1
1.09
Agro-tourism
1.12
1
1.06
Creation of financial reserves
1.24
1
1.11
Forestry work
1.06
1
1.03
Off-farm investment
1.11
1
1.05
Share off-farm income
1.65
1
1.28
Risk preferences
1.09
1
1.04
Organic production
1.11
1
1.05
Focus specialized
1.56
1
1.25
Farm size
1.45
1
1.20
Land tenure
1.12
1
1.06
Note: If VIF exceeds 5 for continuous variables, there are multicollinearity issues. If GVIF exceeds 5 for categorical variables there are
multicollinearity issues.
Table A12. Clustering diversification activities and insurance
Variables
v.test
Mean in
category
Overall mean
P-value
Cluster 1, n=360
None of these strategies
20.3
0.5
0.2
0
Cluster 2, n=35
Forestry work
34.27
1
0.03
0
Cluster 3, n= 266
Creation of financial reserves
24.3
0.8
0.3
0
Off-farm investment
14.0
0.3
0.1
0
Cluster 4, n= 348
Processing and direct marketing
20.9
0.9
0.4
0
Agro-tourism
12.2
0.3
0.1
0
Insurance
2.2
0.4
0.4
0
Cluster 5, n= 167
Off-farm work
32.8
1
0.2
0
Agro-tourism
2.5
0.2
0.1
0
Note: The p value corresponds to the testing of the hypothesis “the mean of the category is equal to the overall mean”. Only the variables which
differ significantly from the overall mean (95% confidence level) are shown. The sign of the v.test highlights whether t he mean is lower or
higher than the overall mean.53 See Supporting Information for Table 3 for the complete output table. The n here represents the number of plots,
the overall n being equal t o 1708 plots.

Table A13 Overview of specifications with wild bootstrap correction, including the coefficients
for the control variables.
Table A13. Wildbootstrapping results for specifications 1-9
Variable s
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Shannon Index
0.12***
(0.004)
0.09**
(0.02)
0.07*
(0.054)
Processing and
direct marketing
0.10***
(0.004)
0.06***
(0.004)
0.05
(0.148)
Agro-tourism
0.04
-0.01
-0.02
(0.326)
(0.68)
(0.3)
Creation of financial
reserves
-0.04
(0.42)
-0.07
(0.106)
-0.07
(0.196)
Forestry work
-0.19***
-0.19***
-0.19**
(0.002)
(0.002)
(0.018)
Off-farm investment
0.02
(0.698)
0.00
(0.93)
0.02
(0.746)
Share of off-farm
income
-0.17***
(0.002)
-0.13***
(0.002)
-0.14***
(0.002)
Control variables
Risk preferences
0.02**
(0.04)
0.02
(0.128)
0.02
(0.164)
0.02
(0.2)
0.02
(0.194)
0.02
(0.17)
0.01
(0.274)
0.02
(0.1)
Organic production
-0.12
(0.104)
-0.14**
(0.07)
-0.13*
(0.08)
-0.13*
(0.08)
-0.13*
(0.08)
-0.13*
(0.08)
-0.12
(0.134)
-0.12*
(0.124)
Focus specialized
-0.12***
(0.002)
-0.09*
(0.06)
-0.08*
(0.06)
-0.08**
(0.04)
-0.09**
(0.03)
-0.08*
(0.06)
-0.07*
(0.09)
-0.11**
(0.02)
Farm size
0.04*
(0.084)
0.05**
(0.04)
0.05**
(0.04)
0.05*
(0.05)
0.05**
(0.04)
0.05**
(0.04)
0.05
(0.172)
0.04*
(0.2)
Land tenure
0.02
(0.678)
0.01
(0.836)
0.01
(0.72)
0.01
(0.8)
0.01
(0.71)
0.01
(0.72)
0.01
(0.7)
0.005
(0.9)
Hail last year
0.01
0.00
0.00
0.00
0.00
0.00
0.01
0.01
0.00
(0.634)
(0.942)
(0.956)
(0.93)
(0.8)
(0.92)
(0.68)
(0.55)
(0.806)
Hail years
0.07**
0.07**
0.07**
0.07*
0.07**
0.07**
0.06
0.07*
0.07**

(0.058)
(0.04)
(0.06)
(0.06)
(0.05)
(0.05)
(0.11)
(0.06)
(0.05)
Mean precipitation
squared
-0.84***
(0.002)
-0.74***
(0.002)
-0.76***
(0.002)
-0.75***
(0.002)
-0.76***
(0.002)
-0.76***
(0.002)
-0.70***
(0.002)
-0.75***
(0.002)
-0.73***
(0.02)
Mean precipitation
0.72***
(0.00)
0.60***
(0.00)
0.62***
(0.00)
0.61***
(0.00)
0.61***
(0.00)
0.62***
(0.00)
0.58**
(0.01)
0.64***
(0.00)
0.61***
(0.006)
Mean temperature
-0.21***
(0.00)
-0.22**
(0.01)
-0.23**
(0.012)
-0.22**
(0.01)
-0.22**
(0.01)
-0.22**
(0.008)
-0.21*
(0.06)
-0.20**
(0.04)
-0.18*
(0.06)
Mean temperature
squared
0.24**
(0.002)
0.24*
(0.08)
0.26*
(0.06)
0.26**
(0.04)
0.25***
(0.06)
0.25*
(0.05)
0.24
(0.11)
0.23**
(0.09)
0.21
(0.176)
Distance
0.05***
0.06***
0.06***
0.06***
0.06**
0.06
0.04**
0.04*
0.04**
(0.00)
(0.004)
(0.002)
(0.002)
(0.002)
(0.002)
(0.02)
(0.02)
(0.02)
Year2017
-0.04
-0.04
-0.04
-0.03
-0.03
-0.04
-0.05
-0.04
-0.04
(0.412)
(0.414)
(0.384)
(0.48)
(0.42)
(0.39)
(0.34)
(0.04)
(0.354)
Year2018
-0.09**
-0.10**
-0.10*
-0.10**
-0.10**
-0.10**
-0.10
-0.08**
-0.08**
(0.03)
(0.018)
(0.01)
(0.01)
(0.01)
(0.01)
(0.008)
(0.04)
(0.03)
Berries
0.24*
0.11
0.12
0.11
0.12
0.11
0.17*
0.24*
0.17*
(0.07)
(0.172)
(0.14)
(0.18)
(0.13)
(0.164)
(0.08)
(0.05)
(0.06)
Plums
0.10
0.01
0.01
0.02
0.02
0.01
0.02
0.09
0.10
(0.156)
(0.734)
(0.74)
(0.70)
(0.64)
(0.75)
(0.73)
(0.18)
(0.132)
Grapes
0.36***
0.22***
0.21***
0.21***
0.22***
0.21***
0.26**
0.35***
0.26***
(0.0002)
(0.00)
(0.004)
(0.004)
(0.01)
(0.012)
(0.01)
(0.004)
(0.004)
R2
0.152
0.139
0.129
0.130
0.134
0.134
0.146
0.176
0.155
Adjusted R2
0.134
0.122
0.111
0.112
0.116
0.116
0.126
0.150
0.137
Number
observations
881
881
881
881
881
881
787
787
895
Note: Estimation method with a regression with wild bootstrap correction with the dependent variable insurance use (1/0). P values are in
parenthesis. *,** and *** denote significance at the 10%, 5% and 1% level.
Table A14. Oster bounds
Shannon
Index
Processing and
direct marketing
of products
Creation of
financial
reserves
Forestry
work
Share of off-farm
income
Beta
(Delta is set to one)
0.09
0.05
-0.09
-0.18
-0.08
Delta
(Beta is set to 0)
4.73
3.75
5.33
17.13
1.93
Note: R squared max was set to 0.243 for all four scenarios. T his is 1/3 higher than t he actual R squared of OLS 0.1762

Figure A2. Generalized Sensitivity Tests
Note: The figures above represent the sensitivity based on Imbens (2003) and Harada (2013). The x axis represents the partial R square for
assignment. We specified the target size of 1.282 for the t-statistics for all four sensitivity tests. This sets the strength of a variable that would
render the t reatment effect statistically insignificant at the 10% level. The blue circle represents the estimated partial R-squares for every
simulated pseudo unobservables. Figures above show that covariates are plotted below the contour. The contour illustrates the statistical power
that the pseudo variables must have for the treatment effect to become insignificant.

Barrett, C. B., T. Reardon and P. Webb (2001). "Nonfarm income diversification and household
livelihood strategies in rural Africa: concepts, dynamics, and policy implications." Food policy 26(4):
315-331.
References for Appendix
Bradshaw, B., Dolan, H. & Smit, B. 2004. Farm-level adaptation to climatic variability and
change: crop diversification in the Canadian prairies. Climatic Change, 67, 119-141.
De Mey, Y., Wauters, E., Schmid, D., Lips, M., Vancauteren, M. & Van Passel, S. 2016.
Farm household risk balancing: empirical evidence from Switzerland. European
Review of Agricultural Economics, 43, 637-662.
Di Falco, S., Adinolfi, F., Bozzola, M. & Capitanio, F. 2014. Crop insurance as a strategy for
adapting to climate change. Journal of Agricultural Economics, 65, 485-504.
Enjolras, G. & Sentis, P. 2011. Crop insurance policies and purchases in France. Agricultural
Economics, 42, 475-486.
Fujisawa, M. & Kobayashi, K. 2011. Climate change adaptation practices of apple growers in
Nagano, Japan. Mitigation and adaptation strategies for global change, 16, 865-877.
Gardebroek, C. 2006. Comparing risk attitudes of organic and non-organic farmers with a
Bayesian random coefficient model. European Review of Agricultural Economics, 33,
485-510.
Goodwin, B. K. 1993. An Empirical Analysis of the Demand for Multiple Peril Crop
Insurance. American Journal of Agricultural Economics, 75, 425-434.
Hardaker, J. B. 2004. Coping with risk in agriculture, Cabi.
Hazell, P. B. 1992. The appropriate role of agricultural insurance in developing countries.
Journal of International Development, 4, 567-581.
Lefebvre, M., Nikolov, D., Gomez-Y-Paloma, S. & Chopeva, M. 2014. Determinants of
insurance adoption among Bulgarian farmers. Agricultural Finance Review, 74, 326-
347.
Menapace, L., Colson, G. & Raffaelli, R. 2015. A comparison of hypothetical risk attitude
elicitation instruments for explaining farmer crop insurance purchases. European
Review of Agricultural Economics, 43, 113-135.
Mishra, A. K. & Goodwin, B. K. 2003. Adoption of crop versus revenue insurance: a farm-
level analysis. Agricultural Finance Review, 63, 143-155.

Mohammed, M. & Ortmann, G. F. 2005. Factors influencing adoption of livestock insurance
by commercial dairy farmers in three zobatat of Eritrea. Agrekon, 44, 172-186.
Nance, D. R., Smith Jr, C. W. & Smithson, C. W. 1993. On the determinants of corporate
hedging. The journal of Finance, 48, 267-284.
O’donoghue, E. J., Roberts, M. J. & Key, N. 2009. Did the federal crop insurance reform act
alter farm enterprise diversification? Journal of Agricultural Economics, 60, 80-104.
Rydant, A. 1979. Adjustments to natural hazards: Factors affecting the adoption of crop-hail
insurance. The Professional Geographer, 31, 312-320.
Sherrick, B. J., Barry, P. J., Ellinger, P. N. & Schnitkey, G. D. 2004. Factors influencing
farmers' crop insurance decisions. American Journal of Agricultural Economics, 86,
103-114.
Velandia, M., Rejesus, R. M., Knight, T. O. & Sherrick, B. J. 2009. Factors affecting farmers'
utilization of agricultural risk management tools: the case of crop insurance, forward
contracting, and spreading sales. Journal of agricultural and applied economics, 41,
107-123.
Declaration of interests
☒ The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this paper.
☐The authors declare the following financial interests/personal relationships which may be
considered as potential competing interests: