4.1. Descriptive Statistics
This sub-section reports the general descriptive results for female and male farmers, while the next sub-session reports the PCA and MRM results. Both the share of female- and male-operated farms from 2005 to 2016 fell in the EU (
Table 1).
Farms operated by women declined by more than 50% in Slovakia (−63.31%), Czechia (−54.74%), and Denmark (−53.69%), while farms run by men declined in Bulgaria (−65.42%) and Slovakia (−62.37%).
Only three countries recorded positive trends for female-run farms compared with those run by men. The former increased by 13% in Sweden, 10% in Luxembourg, and 4% in Spain, while the latter decreased by 21%, 24%, and 17%, respectively, in the same countries. Only Ireland recorded an increase of more than 4% in male-run farms. Overall, from 2005 to 2016, farms operating within the territory of the EU decreased by almost 29%.
Male-run farms (−31%) rather than female-run farms (−23%) contributed more to this decline. It is possible to hypothesize that in the period of 2005–2016 farms run by women showed greater resilience compared with those run by men.
As can be seen in
Figure 1, the EU countries with a higher incidence of female-run farms (>30%) on the total number of farms are in order of importance: Latvia, Lithuania, Romania, Estonia, Austria, Italy, and Portugal. Conversely, those with a low incidence (<10%) are the Netherlands, Malta, Denmark, and Germany. This can be attributed to the fact that in the first block of countries, especially in Latvia and Lithuania, subsistence agriculture prevails, and farms encounter the marginalisation of productive resources and human capital; while in the second block, agriculture with market-oriented farms prevails, has higher incomes, and is led mainly by men. Furthermore, these trends are related not only to cultural factors but also, and above all, to the type of agricultural activity present in the various geographical areas. So much so that the states with the lowest quotas of female management (<10%) are all characterised by a strong zootechnical specialisation. This has traditionally been the prerogative of the male gender, since the establishment of the first complex and technologically advanced communities (“intensive agriculture passes into the hands of men and breeding is always a male task” (
Arioti 1983)).
The results obtained with the application of descriptive statistical analysis (DSA) highlight that the GEI, on average, is about 65%. The average scores of the GEI range from 51.2% in Greece to 83.6% in Sweden (
Eige 2019). Regarding the educational attainment in the analysis, only those with tertiary education were considered. This indicator highlights that, on average, the percentage of women with this level of education is greater (44% vs. 25%). However, the labour participation rate is higher for males than for females (54% vs. 46%), especially in the service sector that is more important for women than for men in terms of employment (83.7% vs. 60.5%). There is always a significant gap between the level of male and female employment, and for women the service sector is relatively more important than the industrial sector in terms of employment opportunities. The EU member states present, on the basis of the value of coefficient of variation (CV), a homogenous situation regarding the labour participation rate for both sexes. For women, the situation in EU countries ranges from a minimum of 40.56% in Malta to a maximum of 50.53% in Lithuania. Working women, however, contribute more to housework than men (1.32% vs. 0.67%). However, the high value of the CV shows that this contribution varies substantially across EU countries. In addition, on average, the unemployment rate is higher for women than for men (89% vs. 82%). Data show that working women were generally more often salaried than their male counterparts (6.7% vs. 6.3%), but important differences, as shown by the values of the CV, exist in the number of female salaried workers from country to country with respect to the number of men. These differences reflect the distribution of jobs between different sectors of the economy, since women tend to be concentrated in the tertiary sector. On the other hand, the life expectancy at birth, on average, is higher for women than for men (83 vs. 77 years). Finally, on average, just over one quarter (26%) of the EU population lived in a rural area in 2016, with some large dissimilarities between the 27 EU countries. The percentage varies from a minimum of 2% in Belgium to a maximum of 46% in Slovakia (
Table 2).
On average, in the 27 countries of the EU, the agricultural area accounts for almost 43% of the total area. The country with the lowest incidence is Sweden (7%), followed by Finland (7.5%), and Cyprus (14%), while the country with the highest incidence is the United Kingdom (72%), followed by Denmark (66%), and Ireland (66%). On the other hand, in relation to arable area, the value of the CV (51%) highlights a greater heterogeneity between EU countries with respect to the incidence of the agricultural area on the total (39%). On average, within the EU, 25% of the area is arable, and the situation varies from a minimum (6.3%) in Sweden to a maximum (60%) in Denmark. However, in relation to the cultivation systems, the greatest dissimilarity (with a CV value equal to 135%) among the 27 EU countries is found in the incidence of the area invested in permanent crops (vineyards, olive groves and orchards) that, on average, represents almost 3% of the total area of the EU-27, with peaks of 0.01% for the Nordic countries (Finland, Ireland, and Sweden) and with peaks ranging from 9.8% to 8% in Mediterranean countries (Spain, Portugal, Greece, and Italy). On the one hand, the Nordic states (Finland and Sweden but also Slovenia) have a greater incidence of forestry area on the total area. On the other hand, Denmark and Eastern European countries (Hungary, Poland, Romania, and Lithuania) have a greater incidence when compared with the area devoted to cereals. The forestry area, that, on average, represents 34% of the surface of the EU-27, shows less heterogeneity between countries than the average of 12% invested in cereals. Finally, the added value of the agricultural, forestry, and fish sectors accounts for an average of almost 2% of the GDP. The minimum value (0.23%) is recorded for Luxembourg while the maximum (4.3%) is for Romania. The value of the CV attests that there is heterogeneity between the 27 EU countries, and in particular between the poorest countries of the EU (Romania, Greece, Latvia, Hungary, and Bulgaria) and the richest nations (Luxembourg, Belgium, the United Kingdom, and Germany). The agricultural sector, therefore, continues to represent the most important economic sector for low-income countries (
Table 3).
The third set of variables describes the characteristics of the agricultural sector in the 27 EU countries. The data reported in
Table 4 show that, on average, more than three quarters of farms are run by men (78%). These farms are generally larger than those run by women as they incorporate, on average, more than 86% of the agricultural area used for production purposes (UAA). Agriculture tends to be much more important for men than for women in terms of the percentage of employment (5.79% vs. 3.39%). This importance, as shown by the high values of the CV, varies substantially across countries, especially for women. About 80% of farms with livestock are run by male farm holders, while only 21% are run by female farm holders. The differences are greater with regard to the units of live livestock (87% for male vs. 16% for female). Differences between men and women in the average area of farmland owned are also reflected in the output per holding, where women farmers also fare much less well than their male counterparts (0.15% vs. 84%). Finally, there is an evident gender gap in the farms whose household consumption exceeds 50% of final production (40% vs. 19%) (
Table 4).
4.2. PCA and MRMs Results
PCA and MRM were applied in order to bring out the differences between male- and female-operated farms in the 27 EU countries, and in order to respond to the RQ
3 (
To what extent are the characteristics of farms run by women affected by different socio-economic and labour market determinants?). The analysis of the main PCs highlighted the differences in the variables of the gender gap in the 27 countries of the EU, and led to the identification, based on the Kaiser criterion theory, of three main components for women and four components for men. Overall, the three components accounted for more than 70% of the total variability for women and the four components accounted for 77% of that for men. There was a low information loss of 30% and 23%, respectively (
Table 5).
Table 6 contains the loadings of every variable of the retained components (score coefficients). To interpret the meaning of every factor, the variables that had the greatest loadings on one factor were analysed in terms of their similarity regarding the measured construct. Following this, it was possible to label the PCs according to their relevant meaning. Significant loadings on a PC are defined as those with a loading greater than 0.30 in absolute value. The higher the loading of a variable, the more contribution is reflected by that variable within a particular PC. The PCA suggested three components with positive signs for the situation concerning women and four for that of men, which meant that the three and the four latent dimensions in the component space accounted for 73% and 77% of the variance, respectively. The first component identified (the presence of educated women in industrial and market-oriented agriculture vs. the presence of rural males in industrial and market-oriented agriculture) included three items concerning the situation for women (SL.IND.EMPL.FE.ZS, F-EA.TE and NV.AGR.TOTL.ZS) and three for the situation concerning men (SL.IND.EMPL.MA.ZS, SP.RUR.TOTL.ZS and NV.AGR.TOTL.ZS). The second component emphasised the presence of women with wages or salaries in the agricultural sector vs. the presence of educated men in permanent cropland, and included one item (SL.EMP.WORK.FE.ZS) for the former and four items for the latter (AG.LND.CROP.ZS, SL.UEM.TOTL.MA.ZS, M-EA.TE and MLPR). The third component, the female use of agricultural land vs. the male use of agricultural and arable land under cereal production, included one item for the situation concerning women and related to the share of agricultural land on total land (AG.LND.AGRI.ZS) and three items concerning the situation for men, relating to the share of agricultural land on the total land in the country (AG.LND.AGRI.ZS), the share of arable land on total land (AG.LND.ARBL.ZS), and the share of land under cereal production on total land (AG.LND.CREL.ZS). The fourth component, the presence of male unemployment in the market-oriented agriculture sector, included two items: SL.UEM.TOTL.MA.ZS and NV.AGR.TOTL.ZS. However, in the countries where more women are employed in the industrial sector, the percentage of value added derived from the agricultural sector is higher. This suggests, in line with the
European Parliament (
2019), that well-educated women are deciding to move to the countryside to carry out their professional activities.
In
Table 7, the new latent factors and their denomination are reported.
As seen in
Table 8 and
Table 9, PC
1 is the most important independent variable that characterises female and male roles in the agricultural sector in the best way. In contrast, the
t-Test was used to assess the significance of the regression coefficients and showed that PC
1 is statistically significant for almost all the dependent variables relating to the characteristics of the agricultural sector (
p < 0.05). According to the t-statistic values, the item that most affects the role of women and men in the agricultural sector is their presence in terms of employment in this sector. The principal difference between women and men is that for the former the most important percentage is that relating to farms owned, while for the latter, it is that relating to the management of farms with livestock. Indeed, for women, all variables that describe the characteristics of the agricultural sector in the 27 EU countries depend positively on the Presence of educated females in industrial and market-oriented agriculture (
Table 8). In contrast, regarding the situation for men, only the variables SL.AGR.EMPL.MA.ZS and M-FWHC > 50% depend positively on the PC
1 (presence of rural male in the industrial and in the market-oriented agriculture). Regarding the second PC, for woman, the relationship between the presence of women with wages and salaries in the agricultural sector and the dependent variables is significant to a small degree and only in relation to F-FWLL. The independent variable SL.AGR.EMPL.MA.ZS is, however, statistically significant for the presence of educated men in the permanent cropland. The female use of agricultural land (PC
3) is influenced to a small degree only by the share of women employed in the agricultural sector, while almost all the dependent variables (except SL.AGR.EMPL.MA.ZS and M-FWHC > 50%) are positively correlated with the male use of agricultural and arable land under cereal production, especially the M-FWLL. Finally, concerning the situation for men, according to t-statistic values, the presence of male employment in the market-oriented agricultural sector is positively affected by only two dependent variables: SL.AGR.EMPL.MA.ZS and M-SO (
Table 9).