The purpose of this section is to describe the results of the Spearman correlation analysis, first at the aggregate level (EU-28) and then at Member States level. Table 2 shows Spearman coefficients that examine the relationship between the distribution of allocation parameters (and therefore of decoupled direct payments) and targeting variables, such as farm income level, and expenditures for fertilizers and pesticides at the EU-28 level.
Above all, coefficients at the EU-28 are largely significant for all the combinations under analysis. First, we look at the UAA parameter, which represents the status quo (that is, the current allocation criteria for decoupled direct payments1). What emerges is a positive association with farm income (ρUAA−FI = 0.329) as well as a strong and positive correlation with the use of fertilizers and pesticides (ρUAA−EXP_FP = 0.592). This result reveals a strong relationship between farm size and the use of chemical input with negative impacts on the environment.
When VA is used as allocation criteria for direct payment in a flat rate scenario, there is a positive and strong correlation with farm income level (ρVA−FI = 0.821). Likewise, when productivity of work and land are used for allocating payments (ρVA/AWU−FI = 0.808 and ρVA/UAA−FI = 0.578), respectively. At the same time, VA is also strongly and positively correlated with the level of expenditure for fertilizers and pesticides (ρVA−EXP_FP = 0.532). The correlation is still positive but lower when labour productivity is used (ρVA/AWU−EXP_FP = 0.406), and close to zero when land productivity is adopted (ρVA/AWU−EXP_FP = 0.024).
Using AWU as allocation criterion we observe a positive correlation with the expenditure for chemical input (ρAWU−EXP_FP = 0.400) and farm income level (ρAWU−FI = 0.355). However, when labour is linked to land (AWU/UAA) coefficients reveal different findings. What emerge are negative correlations both with chemical input expenditure (ρAWU/UAA_EXP_FP=-0.423) and with farm income, even if lower (ρAWU/UAA_FI=-0.174).
Table 2 - Spearman Correlation Coefficients (ρ) between allocation criteria and targeting parameters - EU 28.
[please insert here]
When looking at the parameters for the allocation of direct payments at MS level, despite country specific characteristics which heterogeneously affect Spearman rank coefficients (see appendix - Table A3), it turns out that there are some relevant similarities across countries. Measuring direct payments performance based on the value and the sign of Spearman correlation coefficients at MS level, Table A4 in the appendix ranks the top and the worst five European countries for each allocation criteria under investigation. For space reason and by means of a map, Fig. 2 directly shows, for each policy target, the allocation criteria able to ensure a best/worst distribution of the targeting parameters (farm income and expenses for chemical inputs) in each MS.
Starting from the consumption of chemical inputs, as shown in picture A, it clearly emerges that the parameter with best performances for all MSs is land associated with labour (AWU/UAA), which always shows the lowest Spearman coefficient. Likewise, regarding the most equitable distribution of direct payments (picture B), the parameter that represents the best solution is AWU/UAA for all MSs except France and the Netherlands, where the best parameter is represented by land (UAA). On the other hand, allocating criteria with the worst performance show a higher level of variability across MSs. In details, the Spearman coefficient for aggregate expenditure on fertilisers and pesticides (picture C) exhibits the highest positive value in 20 out of 28 MSs when land (the status quo) is used as allocation criterion. The value added (VA) parameter reveals the lowest performance across 7 MSs; lastly, only in Cyprus is labour (AWU) which perform the worst. As far as farm income support is concerned, the parameter that exhibits the worst performance in most of MSs (18 out of 28) is VA (picture D). It is followed by other criteria including value added, such as VA/AWU (in 9 countries), and VA/UAA (in Sweden only).
Figure 2- Best and worst allocation criteria for chemicals expenditure and for farm income at MS level.
[please insert here]
What emerges from the analyses we carried out at EU-28 level is synthetically represented in Fig. 3, fostering the interpretation of the results based on the conceptual framework we proposed as well as allowing to answer the research questions we elaborated.
This figure shows a Cartesian graph where, for each of the six allocation criterion of direct payments under investigation, it is possible to identify a unique point in the graph which is obtained as a combination of two Cartesian coordinates (respectively, the correlation coefficients with the level of expenditure for chemical input on the y-axis, and with farm income on the x-axis). Each of the six points can be located in one of quadrants which we already identified and labelled in the conceptual framework based on the different characteristics of the recipients of direct payments therein, both in terms of environmental sustainability and farm income level. The position of each point in one of the four quadrants is particularly relevant since it reveals whether a given distribution of direct payments is more or less coherent and effective, given the CAP goals.
Figure 3 - Correlations with the two targeting parameters (FI and EXP_FP) for different allocation criteria (UAA, WU, VA, VA/WU, AWU/UAA, VA/UAA) - EU 28.
[please insert here]
With regard to the first research question, we observe that when land (UAA) is adopted as allocation criteria for direct payments, correlations with both level of farm income and consumption of chemical inputs is positive and quite relevant. As a consequence, the corresponding point is located in the fourth quadrant (“d”) of the Cartesian graph entailing a non-optimal allocation of public aids in terms of coherence and effectiveness, since it mainly favours farms with high level of income and high expenditure for fertilisers and pesticides. Consequences are twofold which is worth discussing. First, what emerges is that the current allocation criteria for direct payments is not neutral regarding both the level of farm income and the use of chemical input in agriculture, confirming previous evidence in single Member States (Ciliberti et al., 2022a). Going beyond, correlation analysis highlights that UAA (i.e. the status quo) negatively impact on policy targeting and effectiveness because it generates an allocation of the aid which is significantly misaligned with the main CAP targets (provision of public goods and ensuring a fair income and standard of living to farmers). This evidence is in line with previous literature in this field, confirming that the current distribution criterion tends to amplify inequalities benefiting larger farms and landowners (Baldoni and Ciaian, 2021; Baldoni et al., 2021; Bateman and Balmford, 2018; Ciaian et al., 2021; Ciliberti and Frascarelli, 2018; Ciliberti et al., 2022b; Guastella et al., 2021; European Court of Auditors, 2018; Varacca et al., 2022; Valenti et al., 2020) rather than smaller or medium sized family farms. In this regard, Grochowska et al. (2021) analysed the fluctuations in income disparities among Polish farmers, highlighting that unless the CAP stops linking payments with agricultural land, measures like capping and degressivity had no effect on changing the situation. On the other hand, regarding the achievement of environmental objectives, an analysis by the European Court of Auditors (2023) on the effectiveness of greening and cross-compliance for sustainable soil management showed that cross-compliance represents a potentially useful tool for soil health, even if it has so far brought modest improvements due to the minimalist choices of Member States.
As far as the second research question is considered empirical evidence provide interesting and diversified insights. Looking at Fig. 3, it is straightforward that the majority of the alternative criteria for the allocation of direct payments do not perform well. In detail, the Cartesian graph displays that three out of five alternative allocation parameters under investigation (i.e. VA, VA/AWU and AWU) are located in the quadrant “d”, revealing medium-high and positive correlations with the use of chemical input and level of farm income. While AWU slightly improve the environmental performance vis à vis UAA, VA and VA/AWU achieve worst results. However, all these alternative criteria would generate an incoherent and ineffective distribution of direct payments towards less sustainable recipients with high income as well to the detriment of both policy targeting and effectiveness of CAP direct payments.
Land productivity (VA/UAA) determines a peculiar distribution of the public support, which approximates the so-called “zero” scenario. Indeed, regardless of the (positive and medium) correlation with farm income, this criterion leads to a neutral (and therefore random) distribution of public aids with regard to the expenditure for fertilisers and pesticides at farm level. As a result, compared to the status quo, such an alternative criterion would improve the environmental impact of the distribution of the aids but at the cost of worsening the inequality of its distribution towards less disadvantaged farms.
Last but not least, the amount of work for unit of utilised agriculture land (AWU/UAA) represents the only allocation parameter under investigation which is able to generate a coherent allocation of direct payments (that is, a point located in the quadrant “b). Such a criterion therefore allows to distribute public aids across the neediest farms using less chemical inputs and therefore farming in a more sustainable way. As a result, this is the only alternative scenario that, in addition to improve the status quo, would also positively impact both policy coherence and effectiveness, because it is completely aligned with both main CAP targets (provision of public goods and ensuring a fair income support to farmers). These findings are in line with intuitions from previous works in this field (Saman, 2021; Severini and Tantari, 2015) which, without going into detailed analysis, at least recognised the need to stop linking direct payment with land and trying to consider other allocation criteria in order to improve the coherence of the policy tools and the CAP effectiveness by means of a more targeted design.