1
LS1622 - Livestock farmers’ attitudes towards consequential loss insurance
Contents
Acronyms and abbreviations
Scientific Report
1. Background and Terms of Reference
2. Methods
3. Results and Discussion
4. Conclusions
Acronyms and abbreviations
B / behaviourBI / behavioural intent(ion)
CA / calculated attitude
CLI / consequential loss insurance
CSN / calculated subjective norm
Defra / Department for Environment Food and Rural Affairs
FGD / focus group discussion
FMD / foot and mouth disease
IQR / inter-quartile range
OA / outcome attitude
RSN / referent subjective norm
SA / stated attitude
SN / subjective norm
SSN / stated subjective norm
TB / tuberculosis
TORA / Theory of Reasoned Action
TPB / Theory of Planned Behaviour
Note: terms in italics are elements of the TORA model
Scientific Report
1. Background and Terms of Reference
Relatively few farmers in England insure their businesses against consequential loss from outbreaks of notifiable diseases – i.e. losses that cannot be compensated under current statutory arrangements. There are few products in the insurance market offering cover for this kind of loss. Experience from recent events, particularly the 2001 FMD outbreak, shows that farmers who are not directly affected by a disease outbreak and are therefore not eligible for compensation can suffer substantial consequential loss, for example from restrictions on movement and marketing of livestock. Defra is in discussion with the insurance industry on the possibility of increasing the availability of appropriate products. This research addressed the question of how farmers are likely to respond and whether they could be encouraged to buy such products.
The objectives of the research were to:
(1)determine livestock farmers’ attitudes towards the taking out of consequential loss insurance, using a methodology based on the Theory of Reasoned Action (TORA)
(2)identify cognitive barriers and drivers towards insuring against consequential loss
(3)suggest elements of a communication strategy to enable livestock farmers to make informed decisions about buying appropriate insurance products covering consequential loss
(4)provide comparative data from farmers within the arable sector facing the risk of consequential loss from disease outbreaks.
2. Methods
2.1 Conceptual framework
The Theory of Reasoned Action (TORA) is a series of concepts and hypotheses linked together as postulated by social psychologists to understand and predict human behaviour (Fishbein & Manfredo, 1992). TORA is one of the ‘expectancy-value’ genre of human behaviour models and is not very different from the vernacular of the subjective expected utility model in economics (Lynne, 1995). TORA has a long-established tradition of research on the relationship between attitudes and behaviour in social psychology. It was introduced by Fishbein (1967), later formalised by Fishbein & Ajzen (1975) and Ajzen & Fishbein (1980), and it has been used successfully in disciplines concerned with volitional human behaviour. The theory of reasoned action postulates “ … that human beings usually behave in a sensible manner; that is, they take account of available information and implicitly or explicitly consider the implications of their actions ... [further] … a person’s intention to perform (or not perform) a behavior [sic] is the immediate determinant of that action. Barring unforeseen events, people are expected to act in accordance with their intentions” (Ajzen, 1988).
The intent to perform is the immediate antecedent of any behaviour. The stronger a person’s intention, the more (s)he is expected to try and thus the greater the possibility that the behaviour is performed (Ajzen & Madden, 1986). The primary concern is to identify the factors that shape and change the behavioural intent (Fishbein & Manfredo, 1992). A person’s intention to behave in a certain way is based on the ‘attitude’ toward the particular behaviour and the perception of the social pressures on them to behave in that way. Such influences are the ‘subjective norms’. The relative contribution of attitudes and subjective norms varies with the behavioural context and the individual. Attitudes are determined by beliefs about the outcomes of the behaviour and evaluation of these expected outcomes. The subjective norm is dependent on beliefs about how
Figure 1 Ajzen and Fishbein Theory of Reasoned Action (1980)
others feel the individual should behave and the motivation to comply with these ‘others’ or referents (Ajzen & Fishbein, 1980), as summarised in Figure 1.
The strength of the relationships between the variable constructs within TORA is measured by analysing the correlation coefficients. In this study, we have used the non-parametric Spearman rank correlation coefficient (rs) as the measure of the strength of association between attitudes and intention, and subjective norm and intention because the scales used to construct these elements are based on ordinal rather than interval measurements (see section 2.3 below).
The relationship between attitudes, subjective norm and behaviour can be expressed as:
where A is attitude toward the behaviour, bi is a belief about the likelihood of outcome i, ei is the evaluation of outcome i, n is the number of salient beliefs, SN is the subjective norm, sbj is a normative belief (that is the referent group or an individual, j, who think(s) the person should, or not, perform the behaviour), mj is motivation to comply with referent j, B is the behaviour, BI is the behavioural intention and w1 and w2 are empirically determined weights based on the respective correlations between A and BI, and SN and BI.
2.2 Methods for data collection
A review of the scientific literature[1] confirmed that little research had been done in the UK on farmers’ attitudes towards insurance in general, and none towards CLI to manage financial risk from notifiable diseases. Research elsewhere in Europe, where insurance is more common as a risk management measure, suggests that socio-economic variables such as gross farm income and farmer’s education have direct and significant relationship with attitude to risk, whereas solvency had an inverse relationship (Meuwissen et al. 2001).
Three focus group discussions were held, with the objective of identifying salient outcome beliefs and referents relating to the taking out of CLI against notifiable disease outbreaks. Telephone interviews were also conducted with several farmers and other stakeholders. An re-analysis, using NVIVO software, was done of focus group discussions from a previous study (Hovi et al. 2004). From these focus groups, salient outcome beliefs and referents were identified.
The outcome beliefs and referents were incorporated into a structured questionnaire, which was administered to a sample of 1500 livestock farmers. A similar questionnaire was administered to a sample of 500 potato farmers, to provide a comparison with farmers in the arable sector who are also faced with the possibility of consequential loss due to future outbreaks of notifiable disease on their own or neighbouring farms. The questionnaire had seven parts covering respectively: basic details of the farm business and enterprises; experience of loss due to notifiable diseases; perceptions of risk and risk management; behaviour and intentions towards insurance and CLI; outcome beliefs related to CLI; subjective norms and sources of advice and information; socio-economic information. Data from the first and last sections was used to identify sub-sets of respondents for comparative analysis.
The samples were selected by simple random sampling by Defra’s Food Chain Analysis 4 Division from the June Census database, to specifications provided by the research team: 500 mainly dairy holdings, 500 beef/sheep holdings, and 500 pig holdings, and 500 arable holdings with more than 8ha. of potato. Defra provided names and addresses to a contractor who mailed the questionnaires, with a covering letter and reply-paid envelope. Under the arrangements for release of the sample, it was not possible to send a reminder inviting non-respondents to complete the questionnaire.
The response rates for livestock and potato surveys were 8.5% and 12.6% respectively. Of the 127 livestock questionnaires received, only 106 were used in the analysis: the remaining 21 were dropped because of incompleteness of responses. Similarly, of the 63 potato questionnaires received, 53 were used in the analysis. Among the livestock farmers, a mix of enterprises was quite common (dairy, beef and sheep). For making comparison between livestock enterprises they have been categorised into dairy (n=43), beef/sheep (n=33) and pigs (n=28), with two respondents not classified as they failed to give sufficient details about their farm animals. The data were analysed using SPSS.
Numbers of livestock per respondent were compared with figures from the June Census (Table 1).
Table 1 Comparison of average livestock numbers between sample and population
Types of animals / Mean from Census 2003 (no of animals/holding) / Mean from the study (no of animals/respondent) / Median value (no of animals/respondent)Dairy / 90 / 148 / 120
Beef / 103 / 100 / 70
Sheep / 343 / 235 / 158
Pig / 514 / 839 / 230
Source: Census data from Defra website (accessed 30/3/2005)
Exact comparison of data from the sample and the census is not possible because the census data were collected and presented by holdings whereas sample data were collected on farm business basis. Nevertheless, since a farm business can comprise several holdings we can conclude that averages for the sample should be equal to or greater than those from the census data. Mean as well as median values of animal numbers have been presented in Table … Because of the presence of a few large farm businesses, median values rather than means would give more realistic picture of the sample parameters. Both the mean and median number of dairy cows for the sample is larger than the national average for England. In case of beef herds the sample mean reflects the population mean but the median figure is smaller than the population mean. For sheep the mean herd size for the sample is lower than the census figure. For pig farms the mean number of pigs in the sample is much larger than the census figure but the median it is less than half the national mean, indicating that our sample has a few very large producers. Similarly for potato farms the average area under potato for study respondents was 70 ha., which was more than four times the national average of 16 ha. The median value for the sample was 39 ha. which is still two and a half times the national mean. Except for beef/sheep category the sample averages work out to be higher than the census figures. This reflects our deliberate choice of sample in which ‘hobby farmers’ have been excluded and suggests we have captured the views of farmers who represent the bulk of commercial production in these enterprises in England.
2.3 Measuring the main TORA variables
Current insurance behaviour was assessed through five questions on insurance practice, identified through focus group discussion (FGD) and telephone interviews. Each practice was allotted a score of (+1) if undertaken and (-1) if not. The sum of the five response scores provided a behavioural index (range of -5 to +5).
The respondent’s strength of intention (I) to take out CLI, if made available next year, in order to manage financial risk resulting from an outbreak of notifiable disease was measured on a five point bi-polar scale, from no intention (-2) to very strong intention (+2). Two measures of intentions were used, relating to notifiable diseases occurring on-farm and notifiable diseases occurring on other farms in the area.
Two measures of attitude held toward the CLI were taken, the stated or emotive attitude (SA) and the Calculated or reasoned attitude (CA). The SA is a measure of how good or bad the respondent feels it would be to take out insurance to cover against losses caused by notifiable disease, if available, in the next year, on a five point bi-polar scale.
The calculated attitude (CA) is arrived at by taking the sum of the individual outcome attitudes (OA). OAs are calculated by taking the product of the specific outcome belief (bi) and the corresponding attributed value (ei ) (i.e. CA = Σbi*ei ). In this case 19 separate OAs are considered, based on outcome beliefs regarding insuring against consequential losses that were drawn from recurring opinions identified during the preliminary focus groups. Each OA has a possible mean score of -4 to +4. Therefore the possible range of the CA score in this case is -76 to +76. A Cronbach’s Alpha coefficient of 0.83 indicates that the 19 item CA scale is reliable.
Similarly, two measures of the Subjective Norm (SN) were taken, the Stated Subjective Norm (SSN) and the Calculate Subjective Norm (CSN). For the SSN, respondents were asked how supportive would ‘other farmers you respect’ be of their taking out insurance cover, if available, against ‘notifiable disease’ for your farm in the next year, with responses given on a five point scale from -2 (very opposed) to +2 (very supportive).
The CSN is based on respondents’ assessment of the views of 13 specified referents identified in focus groups and the extent to which they feel motivated to behave in ways that each referent would approve. The product of these two – subjective belief (sb) and motivation to comply (m) – gives a referent subjective norm (RSN). The CSN is the sum of these individual RSNs.
3. Results and Discussion
3.1 Livestock farmers
3.1.1 Farm businesses and losses due to notifiable diseases
Table 2 presents the mean, median and inter quartile range (IQR) for area farmed by type of farm business. The average livestock farm size was 108 ha. with dairy farmers recording the largest mean farm size of 127 ha. and beef/sheep farms the smallest (73 ha.). Arable farms were almost five times the size of livestock holdings. Potato was cultivated on about 70 ha. on average and farmers planted about three different varieties for a range of purposes – chips, crisps, ware and seed potato. Almost 83% of potato seed is purchased every year by the farmers. About 59% of the livestock businesses are solely owner operated, 24% both owner and tenant and 17% tenant operated. In case of arable farming, 30% farms are owner operated, the majority (51%) with owner and tenant, and remaining 19% as tenants.
Table 2: Area farmed by type of farm business
Farm type / Valid responses (n) / Area (ha.)Mean / Median / IQR
Dairy / 41 / 127 / 98 / (54 to 142)
Beef and sheep / 33 / 73 / 46 / (30 to 100)
Pigs / 26 / 123 / 40 / (11 to 109)
Potato / 52 / 529 / 266 / (180 to 585)
The distribution of mean number of animals across farm type is shown in Table 3. Dairy farmers had an average of 148 milking cows with 92 heifers. More than 50% of the dairy farmers also maintained beef herd of on average 100 cows and two breeding bulls. Beef/sheep farmers had an average beef herd of 100 animals combined with 235 ewes that produced an average of 262 finished lambs for sale. Five farmers from this category also engaged in pig production producing on average 150 finished pigs for the market. Pig farms were producing a mean of 5509 finished pigs per annum from an average unit size of 839 sows.
Table 3: Mean number of animals by farm type
Types of animals / Farm type (number of animals)Dairy (no.) / Beef/sheep (no.) / Pigs (no.)
Dairy / 148 / - / -
Dairy followers / 92 / - / -
Beef cattle / 100 / 100 / 115
Breeding bulls / 2 / 2 / 1
Breeding ewes / - / 235 / -
Rams / - / 8 / -
Finished lambs / - / 262 / -
Sows / - / 20 / 839
Finishing pigs / - / 150 / 5509
Boars / - / 2 / 29
Table 4: Comparative annual income across farm type and percentage of farm income spent on insurance
Farm type / Annual income from sale of agricultural produce / Percentage spent on insuranceMean / Median / IQR
Dairy / 249,750 / 182,500 / (117,500 to 262,500) / 2.7
Beef/sheep / 47,529 / 40,000 / (8,500 to 64,500) / 6.9
Pigs / 1,144,586 / 350,000 / (140,000 to 1,200,000) / 1.5
Potato / 651,431 / 350,000 / (200,000 to 637,500) / 2.7
Comparison of annual farm income across farm type reveals that beef/sheep farmers have the least annual income and pig farms have the highest income in livestock sector (Table 4). The median gives a more realistic picture of the group’s farm income because of extreme cases in each category. Potato farmers’ median annual income is comparable to that of pig farms. For beef/sheep farms, only one third reported that the enterprise was their principal source of income whereas this figure was over 70% for dairy and pig farms. This is one of the reasons why beef/sheep farmers registered considerably lower annual incomes than other groups. Analysis of the expenditure on farm insurance against total farm income indicates that farms on average spend less than 3% on insurance, with the exception of beef/sheep category.
Table 5 Most important losses reported by farmers due to livestock notifiable diseases on-farm
Types of losses business suffer / Number of respondents / PercentageLoss of stock / 63 / 66
Loss of cash flow / 51 / 53
Loss of income / 28 / 29
Loss of nucleus breeding stock / 11 / 12
Total valid respondent / 96 / 100
Respondents were asked in open-ended questions to list and rank the most important losses their business might suffer under two different circumstances: (i) occurrence of notifiable disease on-farm, and (ii) occurrence of notifiable disease on other farms in the area. The results are presented in Table 5 and Table 6. Farmers reported the loss of stock – the source of their livelihood – to be the most important loss their business would suffer due to presence of notifiable disease on-farm. Other important losses farmers mentioned include: loss of cash flow and loss of income in longer term. Pedigree breeders mentioned the loss of nucleus breeding stock, which are difficult to restock. Some farmers have also mentioned the loss of sanity and increased stress, loss of workforce, credibility at stake amongst the peers and creditors etc.
Table 6 Most important losses reported by farmers due to livestock notifiable diseases in the area
Types of losses business suffer / Number of respondents / PercentageIncreased cost and credit / 42 / 47
Loss of market access / 35 / 39
Movement restrictions / 31 / 35
Loss of income / 29 / 33
Loss of stock sales / 26 / 29
Loss of market value / 22 / 25
Total valid respondent / 89 / 100
Farmers reported increased cost resulting from increased feed cost and inability to pay back the creditors to be the major loss the business might suffer as a result of notifiable disease on other farms in the area. Loss of market access to their produce due to general movement restrictions imposed in the area leading to loss of market value (over weight and size, over fattening) of their stock and consequent loss of income are also seen as important consequences.