**4.1 Interest in farm development**

The following analysis is carried out in a linear regression model of will to invest in Norwegian farming. An additive index was built on the three areas of investments shown above, investing in farm buildings, farm machinery and/or increasing farm land. This variable is dependent variable in the model. Table 4 shows that the majority of Norwegian farmers do not plan to invest in their farm in near future.


Table 4. Will to invest. Percent

A combination of farm and farmer characteristics and variables measuring optimism but also potential family succession is included in the forthcoming model.

The size variable has been transformed from an ordinal level variable to interval level variable using real average size instead of scores from 1 to 6.

Farm production was given by farmers as main production. This excludes the possibility of distinguishing farmers with mixed production from mono production. It is does however give a good indication of potential differences between major production groups in Norway – if they exist for the questions analysed. In 2010 the largest group of producers were animal husbandry (39 percent). 29 percent were involved with dairy, 20 percent with grain production. 5 percent were involved with horticulture and 2 percent had forestry as their main production. The final 4 percent had other productions. An analysis of means showed that dairy producers were most willing to invest in their farm. The production variable is recoded into a dummy-set variable for the regression analysis, and dairy represent the control group in the analysis.

Two different measures of income were tested in the regression model. First an ordinal level variable of income was recoded into real average of the income groups. The second model used a recoded version of amount of income from farming into the groups none income, little income, medium income, majority income and all income from farming. In the analysis little income is used as control variable. This value or income group showed in bivariate

Figure 7 shows that will to invest most possibly also were affected by the increased optimism after the global food and price fluctuations in 2007 and 2008. Will to invest in buildings and machinery (equipment/technology) is still higher in 2010 than 2002. Fewer

A farm cannot be maintained without any investments (Almås, 1984). History has also shown that structural change in agriculture is based on a model where fewer farms means that remaining farms need to increase in size and production to uphold domestic

Continued supply of Norwegian food depends on those farmers that will develop their farm. The following analysis aims to reveal in which groups or on wich types of farms

The following analysis is carried out in a linear regression model of will to invest in Norwegian farming. An additive index was built on the three areas of investments shown above, investing in farm buildings, farm machinery and/or increasing farm land. This variable is dependent variable in the model. Table 4 shows that the majority of Norwegian

> Invest in one area

Frequencies 56.5 21.4 15.5 6.6

A combination of farm and farmer characteristics and variables measuring optimism but

The size variable has been transformed from an ordinal level variable to interval level

Farm production was given by farmers as main production. This excludes the possibility of distinguishing farmers with mixed production from mono production. It is does however give a good indication of potential differences between major production groups in Norway – if they exist for the questions analysed. In 2010 the largest group of producers were animal husbandry (39 percent). 29 percent were involved with dairy, 20 percent with grain production. 5 percent were involved with horticulture and 2 percent had forestry as their main production. The final 4 percent had other productions. An analysis of means showed that dairy producers were most willing to invest in their farm. The production variable is recoded into a dummy-set variable for the regression analysis, and dairy represent the

Two different measures of income were tested in the regression model. First an ordinal level variable of income was recoded into real average of the income groups. The second model used a recoded version of amount of income from farming into the groups none income, little income, medium income, majority income and all income from farming. In the analysis little income is used as control variable. This value or income group showed in bivariate

Invest in two areas

Invest in three areas

continuation of family farming and Norwegian food production will take place.

consider increasing the size of productive land.

**4.1 Interest in farm development** 

Table 4. Will to invest. Percent

control group in the analysis.

production when number of mouths to feed is stable or increasing.

farmers do not plan to invest in their farm in near future.

No plans to invest

also potential family succession is included in the forthcoming model.

variable using real average size instead of scores from 1 to 6.

analysis the lowest average score on will to invest. The difference between the two models is commented below.

Optimism related to whether farmers believe future farm income will be improved is included in the model due to its potential effect on will to invest. The variable is coded from 1 (optimistic) 0 and -1 (pessimistic).

Another variable possibly influencing on the will to invest is the prospect of a future family successor. In the model this variable is coded into a dummy variable were value 1 indicates a family successor and 0 is no family successor or farmer does not know. 63 percent believe a family member will succeed the farm. In bivariate analysis those who have successors are significantly more interested in investing in their farm than those who have no successor.

Farmer characteristics are included in the model. Farmer's gender is coded as 1 man and 0 woman. There is no significant difference between men and women in bivariate analysis of this question.

Age is a linear. The variable is found to behave linear in the analysis.

Finally education is included in the analysis. Like several of the variables above, an ordinal level variable was recoded into real average years of education in all school categories. The variable varies from 9 years to 20 years of eduction.


The results of the regression models are shown in table 5 and 6 below.

Constant: Dairy

Table 5. Linear regression model. Dependent variable: Will to invest in farm I.

Exploring the Sociology of Agriculture:

group having no income from farming at all.

Family Farmers in Norway – Future or Past Food Producers? 299

substantially. The reason for showing two separate models is the measure of income that is carried out differently in table 6. Here share of income from farming is recoded into a dummy set variable where little income from farming is the control variable in the equation. In bivariate analysis this group was found to be substantially less interested in investing in the farm than the other income groups. This is still valid when controlled for the other variables in the model. Will to invest depends on farm income for the farming household and increase with increase dependence on this income. A deviation from this pattern is the

A combination of the two income variables in the same model does not add new knowledge to the analysis of will to invest in Norwegian farms. There is a strong positive correlation between farm income and share of income from farming. This indicate that relying on a *substantial* amount of off-farm income (and off-farm work) decrease the opportunity to increase farm income and with that further interest in investing in the farm. There seems to be a moment of critical change when off-farm income to the household exceeds 75 percent. Further adaptation in direction to increased dependence on off-farm income in Norwegian

Norwegian agriculture has faced major structural changes in the statistical history since 1969. Close to 70 percent of the farm units have closed down. Remaining farms are getting bigger, on either bought, but most often rented neighbouring farm land. There has been an increase in big farms (relatively in a Norwegian context), and a decrease in small farms, but

Norwegian farms are operated by mostly male heads that on average are getting older. Farmers are gradually losing their farming identity and more and more farmers find their occupational identity in off-farm work. Still, farming in Norway is based on family involvement and wife/husband/partner participates in farm work on most farms. A small majority of farmers expect family succession to take place in the future. The Norwegian

In the first decade of the twenty-first century Norwegian farmers have experienced increased revenue from agricultural production. The subjective experience of the situation is fewer farmers reporting on negative economic development throughout the decade. This could also reflect that many farmers in the red left the statistics when closing down the farm

The income pattern in Norwegian farming households also shows a critical pattern of offfarm income dominating the economic situation on many farms. 38 percent of Norwegian farmers collect a majority (more than 50 percent) of their household income from farming. This pattern is even enhanced by the finding that one out of two farmers report that farm income constitute less than 25 percent of their household income. This is a critical negative

Future agriculture in Norway is depending on farmers' interest in developing and investing in farming. The willingness to invest has increased slightly in 2002, but there is still a

minority of farmers that plan to invest in their farms in the near future.

farming might be an unsustainable development of future Norwegian agriculture.

**5. Summarising trends in Norwegian family farming** 

the middle size segment is still the dominating farm group.

agricultural system is still based on family farming system.

development in Norwegian agriculture.

production.

The regression model shows several interesting correlations. First of all will to invest in farming increase with increasing size of agricultural productive land. Larger farms are more willing to invest than smaller farms. This is not connected to any particular production; rather it is valid across large size holdings in all production groups. Willingness to invest in farming is further related to income and prospects of the future income situation in agriculture. Willingness to invest is higher in groups that have high income from farming in real value and increase with optimistic views on the economic development of farm income. When it comes to characteristics of the farmers themselves the model do not reveal significant differences between men and women nor of educational level. Age is negatively correlated with will to invest. Young farmers are more willing to invest and this desire decline with increasing age. This might indicate that investments takes place in the beginning of a farming career. On the other hand, knowledge or prospects of a family successor also influence heavily on will to invest in the farm. Bivariate correlation analysis do show that there is a positive correlation between age and knowledge of a family successor, but it is not particularly strong. This means that will to invest due to successors does not necessarily take place in the final stage of one's own farming career. Table 6 shows how the model changes with a different measure of income.


Constant: Dairy and Little income from farming.

Table 6. Linear regression model. Dependent variable: Will to invest in farm II.

The regression model in table 6 show very similar results to the model shown in table 5 above. No variables have strengthened or weakened their position in the model

The regression model shows several interesting correlations. First of all will to invest in farming increase with increasing size of agricultural productive land. Larger farms are more willing to invest than smaller farms. This is not connected to any particular production; rather it is valid across large size holdings in all production groups. Willingness to invest in farming is further related to income and prospects of the future income situation in agriculture. Willingness to invest is higher in groups that have high income from farming in real value and increase with optimistic views on the economic development of farm income. When it comes to characteristics of the farmers themselves the model do not reveal significant differences between men and women nor of educational level. Age is negatively correlated with will to invest. Young farmers are more willing to invest and this desire decline with increasing age. This might indicate that investments takes place in the beginning of a farming career. On the other hand, knowledge or prospects of a family successor also influence heavily on will to invest in the farm. Bivariate correlation analysis do show that there is a positive correlation between age and knowledge of a family successor, but it is not particularly strong. This means that will to invest due to successors does not necessarily take place in the final stage of one's own farming career. Table 6 shows

Constant 1.463 8.392 .000 Area in use .001 7.396 .000 Husbandry -.084 -1.397 .163 Grain -.098 -1.362 .174 Horticulture -.012 -.114 .909 Forestry -.118 -.801 .423 Other -.054 -.406 .685 No income .158 1.907 .057 Medium income .157 2.454 .014 Majority income .186 2.788 .005 All income .211 1.956 .051 Economic optimism .373 11.999 .000 Men .029 .453 .651 Age -.024 -11.174 .000 Education .014 1.870 .062 Family successor .194 4.205 .000

Table 6. Linear regression model. Dependent variable: Will to invest in farm II.

The regression model in table 6 show very similar results to the model shown in table 5 above. No variables have strengthened or weakened their position in the model

B t Sig.

how the model changes with a different measure of income.

Constant: Dairy and Little income from farming.

Model Sumamry: R Square .301 Sig. .000

substantially. The reason for showing two separate models is the measure of income that is carried out differently in table 6. Here share of income from farming is recoded into a dummy set variable where little income from farming is the control variable in the equation. In bivariate analysis this group was found to be substantially less interested in investing in the farm than the other income groups. This is still valid when controlled for the other variables in the model. Will to invest depends on farm income for the farming household and increase with increase dependence on this income. A deviation from this pattern is the group having no income from farming at all.

A combination of the two income variables in the same model does not add new knowledge to the analysis of will to invest in Norwegian farms. There is a strong positive correlation between farm income and share of income from farming. This indicate that relying on a *substantial* amount of off-farm income (and off-farm work) decrease the opportunity to increase farm income and with that further interest in investing in the farm. There seems to be a moment of critical change when off-farm income to the household exceeds 75 percent. Further adaptation in direction to increased dependence on off-farm income in Norwegian farming might be an unsustainable development of future Norwegian agriculture.
