Barriers and Drivers Underpinning Newcomers in Agriculture: Evidence from Italian Census Data

Abstract

The present study addresses, for the first time, the difference between older and younger farmers (those aged over and under 40 years) and proposes a methodology to identify factors that affect generational renewal in the Italian agricultural sector in positive and negative ways. The study is carried out using data collected by the General Census of Agriculture of 2020. Firstly, a T-test is used to test the hypothesis of differences between farmers aged under 40 and those over 40. Secondly, linear regression models are constructed to address the factors that affect generational renewal in the Italian agricultural sector. The findings highlight some important initiatives that decision-makers can consider for further action in the Italian agricultural sector at a regional level. Large-scale farming is very likely to attract newcomers to Italian agriculture and has a strong impact on generational turnover. In contrast, sustainable agricultural practices are less attractive, as they require specific responsibilities, knowledge, and technical and organisational solutions that young people may not yet have. Similarly, educational attainment increases the probability that young farmers will move from rural to urban areas. Finally, older farmers, with respect to newcomers, have more capital for innovative investments in the agricultural sector and information technology for business management and have more experience with waste management.

1. Introduction

This paper proposes a methodology to identify drivers and barriers that affect generational renewal in the Italian agricultural sector, starting with the results of the new General Census of Agriculture of 2020 [1]. Its contribution is organised around two main aims. In the first instance, it seeks to highlight the difference between newcomers (under 40 years old) and older farmers that operate in the Italian agricultural sector. The second objective is to understand the determinants that affect, with differing degrees of intensity, the generational turnover in Italian agriculture at the regional level.

Two major features characterise the recent structure of the agricultural sector in Italy and in other European countries: the dramatic ageing of farmers and the large number of small farms [2]. Indeed, the analysis of statistical data shows that Italy has a rural-population-ageing problem, as almost 91% of Italian farms are run by farmers aged over 40 [1]. Another problem is that older farmers are less likely to adopt new technology or to change the organisation and business management of their farms [3]. For this reason, newcomers can make a relevant contribution to promoting rural prosperity, strengthening rural value chains, and investing in rural viability and vitality—three of the ten policy orientations articulated in the Cork 2.0 Declaration “A better life in rural areas” [4].

The European Union (EU), therefore, is committing specific aid to encourage the renewal of generations in rural Europe through direct Pillar II payments for the establishment of young farmers under the age of 40. The Young Farmers Programme is an EU-wide policy measure aimed at addressing the shortage of young farmers and ensuring the long-term sustainability of the primary sector [5]. Furthermore, some authors [6] have pointed out that the shortage of newcomers (farmers aged under 40) is mostly prevalent in countries with small-scale holdings, such as Portugal, Italy, and Greece.

The analysis of the results from the VII General Census of Agriculture of 2020 can help to highlight the recent developments in the structure of Italian farms and can serve as input for future policy discussions. Indeed, according to the results of this census, agriculture has a more modern management model than in the past. There are fewer farms, but they are larger. There is also less land ownership, but land has a more multifunctional role. There are persistent gaps compared to other economic sectors: delays in digitisation, inadequate vocational training for managers, and strong regional and territorial discrepancies. Moreover, there is a slight growth in the number of young businesses, which, as is well known, are characterised by greater competitiveness, productivity, a propensity for innovation, and environmental and social sustainability. Within this dynamic context, which is full of contrasts, the role of the “young” farm manager has not taken off: compared to 2010, in 2020, the percentage of farms with a young farm manager fell from 11.5% to 9.3%. This trend is consistent with what is happening in other sectors of the economy, which also depends on the increased attractiveness of other forms of employment and emigration abroad. However, a more in-depth reading of the census results shows that young managers tend to run specific types of farms, which are still not very numerous. They are also strongly characterised by some identifying factors [1].

The general demographic picture is alarming, and the existence of gaps in the provision of infrastructure and services does not encourage growth in the number of young farmers. More reassuring are the data on farms run by the young; the numbers have grown since 2017, albeit only slightly. This is in contrast to the rest of the economy and the progressive reduction in the number of farms overall. In addition, young farmers in charge of enterprises have a higher degree of competitiveness and productivity and a propensity for innovation and market orientation.

The Registers of Companies from 2017 to today highlight how new farms have been set up every day at the hands of young people (up to 35 years of age), while five have closed their doors. This means that in the balance between registrations and cessations, there is a surplus of more than 6000 holdings, on average, during this five-year period [7]. As a result of these dynamics, the number of farms run by the new generations was 56,172 by the end of 2021, showing a growth of 0.4% per year in the last 5 years. In the same period, the total number of farms decreased at a rate of 0.7% per year and that of “youth” businesses of the entire economy by as much as 2.4%, corresponding to the disappearance of over 70,000 businesses in the period observed.

The involvement of young people in the agricultural sector is an objective of the European Common Agricultural Policy (CAP) and a challenge for Italy [8]. However, the novel literature regarding the Young Farmers Scheme has been oriented towards the success factors of the scheme. A greater presence of young people is necessary in order to accelerate and concretise the renewal that the agricultural sector needs to increase farmers’ incomes [9], be more competitive [10], address environmental challenges, and ensure a contribution to climate change adaptation and mitigation [11]. Young people are also at the forefront of the multifunctional agriculture model, which is changing the perception of the Italian primary sector, often with important repercussions for the environment and the community, such as renewable energy production or social agriculture [11]. Young farmers, from simple producers of food, become creators of services and generators of value for the rural territory through successful examples, such as agritourism, the processing and direct sales of products, educational farms, and agriasili. The incidence of young people in farms with related activities has risen to 19% [1].

Within this context, to the best of my knowledge, this is the first study that provides statistical evidence of the differences between new entrants and older farmers using the data of the latest General Census of Agriculture. It also explores the determinants that affect, with differing degrees of intensity, the generational turnover in Italian agriculture at the regional level. The rest of the paper is organised as follows: Section 2 offers a framework of the literature that analyses the difference between new entrants and older farmers and on the determinants that affect the generational turnover in the Italian agricultural sector. In Section 3, the variables used, the main descriptive statistics, and the construction of the linear regression models are presented. In Section 4, the results are summarised. In Section 5, a discussion of the results and policy implications are reported. Finally, Section 6 presents some conclusions.

2. The Theoretical Framework

Nowadays, there is much interest in understanding the generational turnover in agriculture, a sector that in many countries is suffering from a fragmented and old structure and consequently cannot be competitive in the European single market. Some authors [12] have used the substitution index (SI) to measure the turnover rate in agriculture. This indicator, given by the ratio between the number of younger and slightly older holders, makes it possible, albeit indirectly, to predict what will happen in the future. The lower this ratio is, the more the degree of substitution between new and old generations is compromised. Moreover, a rapidly growing body of evidence suggests that small-scale farming is much less likely to attract young people than large-scale farming and nonagricultural sectors [13,14,15]. To date, studies of generational renewal [16] have only considered farms in general (i.e., the range of small to large-scale farms), which minimises the differences in the development conditions for small-scale farms.

On the other hand, other authors [17] have underlined that any discussion of generational renewal in agriculture must start from an analysis of the factors that can influence it. In particular, the main factors include income, the role of rural and local development, the attractiveness and prestige of the profession, the quality of work and life, schooling, and the formation of human capital. The entry of the young into farming may be as much a matter of “push” as “pull” factors [18]. These factors consist of economic, social, demographic, institutional, and territorial determinants of young people’s choices, considered at the macro and micro levels.

At the European level, other researchers have identified a number of scale-related issues, which shape the likelihood of generational renewal in agriculture. These include farm size, profitability, total farm assets, farm type, farm location, diversification strategy, and market integration [19,20,21,22]. Moreover, farm size, as highlighted by Lavison (2013) [23], represents one of the important determinants of technology adoption. Indeed, many other studies have reported a positive relationship between farm size and the adoption of agricultural technology [24,25,26,27]. The low supply of land for sale or lease, as well as competition from other farmers, investors, and residential users, pushes up land prices [28,29].

The preferences and skills of farmers and farmers’ families are also important for generational renewal. Furthermore, new entrants to farming tend to be more highly educated than existing farmers [30], but there is some evidence that younger farmers are people with nonagricultural backgrounds [31] and have less access to extension services in both Europe [32] and Italy. The education level of a farmer increases their ability to obtain processes and use information relevant to the adoption of new technology [23,26,33]. For instance, the adoption of organic fertilisers is positively influenced by the level of education [34].

Moreover, several authors have long discussed farm population ageing [35,36,37,38]. Other authors have pointed out that the average age of farmworkers is increasing faster than in other professions [39]. This is probably because the phenomenon of farmer ageing takes on a structural character, also underlining the difficulties for older farmers leaving the sector and those faced by the youngest who are entering [40]. One of the negative factors that has been taken into consideration is the small size of the farms. Secondly, the creation of new farms (and the continuation of existing small farms) is at odds with the many entrance barriers [41] that increasingly characterise the farming sector. Thirdly, this phenomenon (especially in the view of Italian agriculture generally) represents a remarkable contrast with the yearly outmigration of some 25,000 young graduates from Italy [42] to places that promise better job prospects. Fourthly, empirical data show that the new generation of farmers (especially but not only those with nonagricultural backgrounds) is highly educated [41], which apparently makes the choice of farming even more enigmatic: they should know better. As young farmers have limited availability of capital and limited access to formal credit [43], they must rely more on their own labour, skills, and knowledge and, possibly, on support provided by their family and social networks.

Young farmers have to be able (more than previous generations) to “stand on their own”. New innovative elements are needed that, according to Rullani [44], “respond in a proactive way to the different situations of unsustainability that the planet’s accelerated modernisation is driving … and (which) give voice and space for action to new forms of distributed intelligence that are emerging in knowledge society and putting in motion relevant phenomena of learning and social experimentation (social networks, communities, networks and professional groups)”. These are, once again in Rullani’s words, “the reasons that make innovation a must we cannot do without”. These innovative elements necessarily challenge outdated conceptions and undermine antiquated forms of organisation. It might be hypothesised that young farmers are better equipped to engage actively in such a transformation than others.

3. Dataset and Methods

The study was carried out using data collected by the General Census of Agriculture of 2020. This dataset includes a considerable amount of information about all farms that are divided into four main groups. The information includes the physical size of farms, the education level of farmers and their experience in the agricultural sector, the adoption of sustainability practices, the introduction of innovation, and information regarding farmers’ organisation and operations (

Table 1. Blocks of variables used in the linear regression models.

(I) Size by type of farms (ha/n. of farms)
S1	Size of farms
S2	Size of privately owned farms
S3	Size of rented farms
S4	Size of farms for free use
(II) Education and experience in the agricultural sector (% of the total)
E1	No school
E2	Elementary
E3	Middle
E4	Agricultural diploma (2–3 years)
E5	Nonagricultural diploma (2–3 years)
E6	High school diploma in agriculture
E7	High school diploma not in agriculture
E8	Agricultural degree
E9	No agricultural degree
E10	Training courses
E11	Experience in agriculture < 3 years
E12	Experience in agriculture from 3 to 10 years
E13	Experience in agriculture > 10 years
(III) Organic (% of the total)
O1	Livestock farms with organic farms
O2	Farms with organic crops
(IV) Innovation and computerisation (% of the total)
IC1	Farms with at least one innovative investment in the three-year period 2018–2020
IC2	Waste management
IC3	Mechanisation
IC4	Organisation and business management
IC5	Sales and marketing of products
IC6	Connected activities
IC7	Computerised agricultural farms
IC8	Accounting
IC9	Crop management
IC10	Farm management
IC11	Management of connected activities

Source: Owen elaboration on VII General Census of Agriculture in 2020.

and Figure 1). These variables, as suggested by other authors [45], are important for increasing the attractiveness of the agricultural sector. Therefore, it is very useful to explore the effects of the variable that is of interest to this study, i.e., the age of entrepreneurs, as it allows control of a high number of other relevant variables.

The data analysis was organised around three research questions (RQs):

RQ_1: Are there differences between young farmers and older farmers?

RQ₂: Which factors determine the generational turnover in the Italian agriculture sector?

RQ₃: How can young farmers respond in a proactive way to the different situations of unsustainability?

In order to answer the first RQ, the T-test dependent samples given by the statistical programme STATA version 16 were used. The dependent T-test is used to test the null hypothesis that there are no differences between the means of the two related groups. If a significant result is obtained, the null hypothesis that there are no significant differences between the means can be rejected, and the alternative hypothesis that there are statistically significant differences between the means can be accepted. This can be expressed as follows:

HO: μ1 = μ2

HA: μ1 ≠ μ2

With the first step, a T-test was used to test the hypothesis about differences between farmers up to the age of 40 and the groups of farmers over the age of 40.

In the second step, the construction of linear regression models led to an assessment of the relationship between the substation index (SI) variable and each of the predictors, adjusting for the remaining predictors in order to respond to RQ₂ and RQ₃.

4. Results

In Italy, based on the latest Structural Census of Agriculture of 2020, there are 1,130,558 businesses. Most farms are located in Apulia, Sicily, Calabria, and Veneto. On the other hand, the least number of farms are in Valle d’Aosta, Liguria, and in the Autonomous Province of Trento.

In terms of farm size, the largest farms are in Sardinia (25 hectares), in Valle d’Aosta (24), and in Lombardy (20.8). On the other hand, the smallest farms are in Campania (6.2 ha), Calabria (5.4 ha), and Liguria (3.3 ha). The average size of farms in Italy is 10.7 ha of land (Year 2020). See Figure 2.

Table 2. Main descriptive statistics on the variables used in the first model and results of T-test.

Size (y = Young; o = Old)
Variable	Mean	Min.	Max.	Std. Err.	Std. Dev.	CV	Diff = Mean Size Young − Old
sy1	19.272	5.070	44.099	2.114	9.686	0.503	diff = mean (sy1) − mean (so1)
so1	11.096	3.070	22.093	1.223	5.604	0.505	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	8.176			2.442	5.410	0.662	6.9248	1.000	0.000	0.000
sy2	9.293	2.322	21.261	0.947	4.339	0.467	diff = mean (sy2) − mean (so2)
so2	6.437	1.846	13.071	0.639	2.928	0.455	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	2.856			0.443	2.028	0.710	6.4529	1.000	0.000	0.000
sy3	20.445	4.336	50.942	2.398	10.991	0.538	diff = mean (sy3) − mean (so3)
so3	15.606	4.108	32.344	1.618	7.416	0.475	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	4.839			1.054	4.830	0.998	4.5911	0.9999	0.0002	0.0001
sy4	9.145	2.071	25.222	1.127	5.163	0.565	diff = mean (sy4) − mean (so4)
so4	5.813	1.532	15.847	0.732	3.355	0.577	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	3.332			0.577	2.646	0.794	5.7711	1.000	0.000	0.000

Source: Owen elaboration on VII General Census of Agriculture in 2020.

,

Table 3. Main descriptive statistics on the variables used in the second model and results of T-test.

Organic (y = Young; o = Old)
Variable	Mean	Min.	Max.	Std. Err.	Std. Dev.	CV	Diff = Mean Young – Mean Old
oy1	23.002	7.407	35.065	1.343	6.155	0.268	diff = mean (oy1) − mean (oo1)
oo1	76.998	64.935	92.593	1.343	6.155	0.080	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−53.995			2.686	12.310	−0.228	−20.0997	0.000	0.000	1.000
oy2	20.341	14.754	32.880	1.021	4.678	0.230	diff = mean (oy2) − mean (oo2)
oo2	79.659	64.935	92.593	1.021	4.678	0.059	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−59.318			2.041	9.355	−0.158	−29.0568	0.5342	0.9317	0.4658

Source: Owen elaboration on VII General Census of Agriculture in 2020.

,

Table 4. Main descriptive statistics on the variables used in the third model and results of T-test.

Educational Level and Experience in the Agricultural Sector
Variable	Mean	Min.	Max.	Std. Err.	Std. Dev.	CV	Diff. Mean Young – Mean Old
ey1	1.883	0.000	10.088	0.488	2.236	1.187	diff = mean (ey1) − mean (eo1)
eo1	98.117	89.912	100.000	0.488	2.236	0.023	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−96.234			0.976	4.472	−0.046	−98.622	0.000	0.000	1.000
ey2	6.928	2.284	28.571	1.266	5.804	0.838	diff = mean (ey2) − mean (eo2)
eo2	1717.940	489.277	5600.000	250.592	1148.354	0.668	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−1711.012			249.622	1143.913	−0.669	−6.8544	0.000	0.000	1.000
ey3	208.857	36.612	1514.286	69.555	318.742	1.526	diff = mean (ey3) − mean (eo3)
eo3	3068.777	687.373	13100.000	601.309	2755.544	0.898	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−2859.920			536.161	2456.997	−0.859	−5.3341	0.000	0.000	1.000
ey4	59.666	4.348	296.053	20.157	92.371	1.548	diff = mean (ey4) − mean (eo4)
eo4	290.060	26.624	1476.754	83.824	384.130	1.324	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−230.394			64.380	295.026	−1.281	−3.5787	0.0009	0.0019	0.99910
ey5	95.194	7.716	485.714	30.181	138.307	1.453	diff = mean (ey5) − mean (eo5)
eo5	565.527	54.445	2400.000	145.240	665.575	1.177	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−470.333			116.234	532.652	−1.132	−4.0464	0.0003	0.0006	0.99970
ey6	200.711	18.356	1371.429	66.152	303.147	1.510	diff = mean (ey6) − mean (eo6)
eo6	423.285	66.954	1428.571	81.891	375.271	0.887	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−222.575			39.427	180.675	−0.812	−5.6453	0.000	0.000	1.000
ey7	318.239	80.802	1385.714	64.219	294.289	0.925	diff = mean (ey7) − mean (eo7)
eo7	1415.011	321.701	4642.857	223.818	1025.662	0.725	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−1096.772			162.567	744.977	−0.679	−6.7466	0.000	0.000	1.000
ey8	51.310	5.943	214.286	11.409	52.281	1.019	diff = mean (ey8) − mean (eo8)
eo8	113.571	18.428	314.286	19.509	89.401	0.787	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	62.261			10.270	47.064	0.756	6.0622	0.000	0.000	1.000
ey9	125.160	30.786	371.429	19.492	89.322	0.714	diff = mean (ey9) − mean (eo9)
eo9	546.680	125.508	1557.143	77.583	355.532	0.650	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−421.520			−59.180	271.199	−0.643	−7.1226	0.000	0.000	1.000
ey10	20.341	14.754	32.880	1.021	4.678	0.230	diff = mean (ey10) − mean (eo10)
eo10	79.659	67.120	85.246	1.021	4.678	0.059	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−59.318			2.041	9.355	−0.158	−29.0568	0.000	0.000	1.000
ey11	183.303	26.264	942.857	45.279	207.492	1.132	diff = mean (ey11) − mean (eo11)
eo11	309.183	59.454	1142.857	53.905	247.024	0.799	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	125.880			15.837	72.573	0.577	0.0000	0.5000	1.0000	0.50000
ey12	531.563	94.752	2814.286	133.505	611.796	1.151	diff = mean (ey12) − mean (eo12)
eo12	1331.308	316.942	4914.286	224.378	1028.228	0.772	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−799.745			100.899	462.375	−0.578	−7.9262	0.000	0.000	1.000
ey13	333.576	64.467	1814.286	86.035	394.263	1.182	diff = mean (ey13) − mean (eo13)
eo13	6483.323	1531.472	23528.570	1104.571	5061.780	0.781	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−6149.747			1021.754	4682.266	−0.761	−6.0188	0.000	0.000	1.000

Source: Owen elaboration on VII General Census of Agriculture in 2020.

and

Table 5. Main descriptive statistics on the variables used in the third model and results of T-test.

Innovation and Computerisation
Variable	Mean	Min.	Max.	Std. Err.	Std. Dev.	CV	Diff. Mean Young − Mean Old
icy1	21.741	15.768	29.637	0.824	3.776	0.174	diff = mean (icy1) − mean (ico1)
ico1	111.322	85.824	146.177	3.517	16.116	0.145	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−89.581			3.684	16.881	−0.188	−24.318	0.000	0.000	1.000
icy2	0.404	0.180	0.973	0.042	0.191	0.473	diff = mean (icy2) − mean (ico2)
ico2	80.377	58.341	112.352	3.059	14.018	0.174	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−79.973			3.063	14.037	−0.176	−26.108	0.000	0.000	1.000
icy3	12.744	7.984	18.750	0.542	2.483	0.195	diff = mean (icy3) − mean (ico3)
ico3	30.196	10.685	38.318	1.382	6.334	0.210	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−17.453			1.705	7.811	−0.448	−10.239	0.000	0.000	1.000
icy4	2.296	1.008	3.710	0.158	0.726	0.316	diff = mean (icy4) − mean (ico4)
ico4	18.155	6.532	46.334	2.126	9.743	0.537	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−15.858			2.134	9.778	−0.617	−7.4319	0.000	0.000	1.000
icy5	1.900	0.629	3.933	0.157	0.721	0.380	diff = mean (icy5) − mean (ico5)
ico5	17.918	9.588	40.741	1.506	6.901	0.385	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−16.018			1.496	6.855	−0.428	−12.897	0.000	0.000	1.000
icy6	1.691	0.952	2.883	0.119	0.544	0.321	diff = mean (icy6) − mean (ico6)
ico6	5.983	3.099	10.593	0.372	1.703	0.285	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−4.291			0.271	1.244	−0.290	−15.813	0.000	0.000	1.000
icy7	29.811	20.608	47.246	1.518	6.956	0.233	diff = mean (icy7) − mean (ico7)
ico7	111.322	85.824	146.177	3.517	16.116	0.145	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−81.511			3.677	16.852	−0.207	−22.165	0.000	0.000	1.000
icy8	22.039	15.370	34.274	1.054	4.831	0.219	diff = mean (icy8) − mean (ico8)
ico8	80.377	58.341	112.352	3.059	14.018	0.174	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−58.338			3.001	13.751	−0.236	−19.442	0.000	0.000	1.000
icy9	9.123	5.198	17.153	0.589	2.700	0.296	diff = mean (icy9) − mean (ico9)
ico9	30.196	10.685	38.318	1.382	6.334	0.210	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−21.073			1.269	5.815	−0.276	−16.607	0.000	0.000	1.000
icy10	6.439	1.827	18.548	0.936	4.291	0.666	diff = mean (icy10) − mean (ico10)
ico10	18.155	6.532	46.334	2.126	9.743	0.537	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−11.715			1.515	6.943	−0.593	−7.7324	0.000	0.000	1.000
icy11	4.610	2.613	8.035	0.309	1.415	0.307	diff = mean (icy11) − mean (ico11)
ico11	17.918	9.588	40.741	1.506	6.901	0.385	t	Pr (T < t)	Pr (ITI > t)	Pr (T > t)
Difference	−13.308			1.276	5.848	−0.439	−10.428	0.000	0.000	1.000

Source: Owen elaboration on VII General Census of Agriculture in 2020.

show the main descriptive statistics and the T-test analysis relating to the indicated variables and to the differences between the two groups (young vs. old) in the size of farms, organic farming, educational level and experience in the agricultural sector, and innovation and computerisation. The differences between the average values are mainly due to asymmetry in the farm distribution in Italian regions. The coefficient of variation (CV) shows the subtle differences between the Italian regions. The two dimensions of organic (O) and size (S) have the lowest relative variability in the set of variables considered for the analysis (Table 2 and Table 3). In contrast, the highest values are recorded for educational level and experience in the agricultural sector (E) and innovation and computerisation (IC) of the farmers’ activities (Table 4 and Table 5).

To begin with, the first block of variables, the size of farms owned by older farmers (so2), has the lowest variability: the highest values are recorded in Sardinia, Emilia Romagna, and Lombardy, while the lowest values are registered in the Autonomous Province of Trento, in Liguria, and in Campania. The results of the T-test suggest that the null hypothesis that there is no significant difference between the means can be rejected, and the alternative hypothesis that there are statistically significant differences between the sizes of farms run by young people compared to those run by older farmers can be accepted. As expected, younger farmers make use of a higher amount of rented land, which could be interpreted as a signal of a greater effort employed in the activity of the farmer. Furthermore, large-scale farming is more likely to attract young people [14,15].

For the two variables related to sustainable practices (livestock farms with organic farming methods and farms with organic crops), the CV highlights that there are differences between the Italian regions. The highest values are recorded for farms run by older farmers in Valle D’Aosta and in Lombardy, while the lowest values are registered in Molise and Sardinia. Molise and Campania, however, registered the highest values for farms run by younger farmers. In contrast, the T-test results suggest that the null hypothesis that there is no significant difference between the means can be rejected, while the alternative hypothesis that there are statistically significant differences between the younger and the older farmers can be accepted. On average, older farmers are more likely to invest in organic farming when compared to younger farmers. This may be because these practices require responsibility, knowledge, and specific technical and organisational solutions.

The CV shows the differences between the Italian regions, especially in relation to the higher education level of younger farmers (agricultural diploma (2–3 years), middle, and high school diploma in agriculture). The T-test results suggest that the null hypothesis that there is no significant difference between the means can be rejected, whereas the alternative hypothesis that there are statistically significant differences between younger and older farmers can be accepted. As expected, younger farmers, on average, have a lower level of education. It is likely that for farmers that are more educated, the cost of leaving farming is higher than for their lesser-educated counterparts. As the elasticity of education in comparison to salary is much higher in urban areas than in rural areas, education can increase the probability of farmers migrating [46]. Moreover, it is true that the well-educated young are likely to move away, but the decision to study could also be related to a choice made because of the low revenue of the farm or because they do not consider farming as their future career.

Finally, regarding the set of variables that describes the adoption of innovation and the use of information technology in the agricultural sector, as the values of the CV suggest, there is quite a variability at the regional level. On the one hand, the higher values are in relation to farm management, organisation, business management, and waste management, especially for farms run by older farmers. For farm and organisation business management, the highest values are recorded in Lombardy, Valle d’Aosta, and Sardinia. While the lowest values are registered in Liguria, in Apulia, and in the Autonomous Province of Trento. For waste management, the highest values are recorded for the Autonomous Province of Trento and for Lombardy, while the lowest values are registered in Molise and in Basilicata. On the other hand, the lowest variability is in correspondence with farms with at least one innovative investment in the three-year period of 2018–2020. The highest values are recorded in Lombardy and in Tuscany. In contrast, the lowest values are registered in Abruzzo and in Calabria. The T-test results suggest that the null hypothesis that there is no significant difference between the means can be rejected, and the alternative hypothesis that there are statistically significant differences between younger and older farmers can be accepted. As expected, on average, older farmers introduced more innovative investments in the three-year period of 2018–2020. They also introduced more information technology and have more experience with waste management.

The above results corroborate the different position of young farmers in the Italian regions. If, on the one hand, the substitution index (SI) is higher for Valle d’Aosta, Trentino-Alto Adige and Sardinia, on the other hand, the generational turnover is more compromised for Abruzzo and Apulia (Figure 3).

As is well known, the agricultural entrepreneurship of young farmers is essential to the sustained, inclusive, and sustainable economic growth of developing agrarian economies. Moreover, it offers great potential for providing full and productive employment and decent work. Nowadays, young farmers must be the main actors in the conservation, management, and protection of the environment.

From

Table 6. Parametric estimation results of the linear regression models.

Size
SI	Estimate	Std. Err.	t Value	Pr > (ItI)	Number of obs.	=	21
(Intercep)	0.114	0.016	7.030	0.000	F(4,16)	=	3.580
S1	0.012	0.004	3.000	0.009	Prob > F	=	0.029
S2	−0.014	0.005	−2.710	0.016	R-squared	=	0.472
S3	−0.005	0.003	−1.900	0.075	Adj. R-squared	=	0.340
S4	0.005	0.003	1.480	0.158	Root MSE	=	0.027
Educational level and Experience in the Agricultural Sector
SI	Estimate	Std. Err.	T Value	Pr > (ItI)	Number of obs.	=	21
(Intercep)	0.139	0.015	9.200	0.000	F(13,7)	=	2.130
E1	0.018	0.016	1.160	0.284	Prob > F	=	0.160
E2	0.055	0.109	0.500	0.631	R-squared	=	0.7982
E3	0.211	0.178	1.180	0.275	Adj. R-squared	=	0.4235
E4	0.002	0.012	0.200	0.850	Root MSE	=	0.02532
E5	0.033	0.028	1.150	0.287
E6	0.026	0.031	0.840	0.428
E7	0.012	0.091	0.130	0.902
E8	−0.017	0.014	−1.230	0.257
E9	0.069	0.042	1.660	0.141
E10	0.008	0.006	1.410	0.202
E11	0.000	0.023	0.020	0.983
E12	−0.104	0.110	−0.950	0.374
E13	−0.318	0.367	−0.870	0.415
Organic
SI	Estimate	Std. Err.	T Value	Pr > (ItI)	Number of obs.	=	21
(Intercep)	0.127	0.010	12.270	0.000	F(2,18)	=	3.780
O1	0.003	0.002	1.750	0.097	Prob > F	=	0.043
O2	−0.006	0.002	−2.750	0.013	R-squared	=	0.296
					Adj. R-squared	=	0.217
					Root MSE	=	0.030
Innovation and Computerisation
SI	Estimate	Std. Err.	t Value	Pr > (ItI)	Number of obs.	=	21
(Intercep)	0.130	0.015	8.510	0.000	F(11,9)	=	1.060
IC1	0.019	0.024	0.790	0.450	Prob > F	=	0.476
IC2	0.000	0.006	0.010	0.993	R-squared	=	0.563
IC3	−0.027	0.021	−1.250	0.243	Adj. R-squared	=	0.030
IC4	−0.027	0.033	−0.830	0.430	Root MSE	=	0.033
IC5	0.018	0.019	0.940	0.371
IC6	−0.018	0.032	−0.560	0.586
IC7	0.121	0.108	1.120	0.292
IC8	−0.057	0.081	−0.700	0.499
IC9	−0.043	0.025	−1.690	0.125
IC10	0.005	0.009	0.550	0.594
IC11	0.006	0.022	0.250	0.810

Source: Owen elaboration on VII General Census of Agriculture in 2020.

, it is possible to see if there are differences between the groups of farmers under the age 40 and those over 40 for the four blocks of variables: size (S), education level and experience in the sector €, organic farming (O), and innovation and computerisation (IC). The equations of the linear models used for the estimation of each block of variables, in the explicit form, are as follows:

 

(I) SI = β0 + β1S1 + β2S2 + β3S3 + β4S4

 

(II) SI = β0 + β1E1 + β2E2 + β3E3 + β4E4 + β1E5 + β2E6 + β3E7 + β4E8 + β1E9 + β2E10 + β3E11 + β4E12 + β1E13

 

(III) SI = β0 + β1O1 + β2O2

 

(IV) SI = β0 + β1IC1 + β2IC2 + β3IC3 + β4IC4 + β1IC5 + β2IC6 + β3IC7 + β4IC8 + β1IC9 + β2IC10 + β3IC11

The meaning of the abbreviations of the variables is reported in Table 1.

With the aid of the four linear regression models, it is possible to attempt an estimate of the effects of each determinate (variable) on the SI. The SI is the dependent variable in the regression analysis, while explanatory variables were selected in consideration of their aptitude to explain the causes of differences between groups of farmers under the age of 40 and those over 40.

Starting with the first model, this explains 47% of the observed variability, confirming the regional differences in farm size. Given the low diffusion of available properties and land to lease, this is an almost insurmountable barrier for those who do not have a priori the necessary capital. Indeed, the low propensity of the credit system to finance entrepreneurs on the basis of a productive project rather than on the basis of capital guarantees is an obstacle. Furthermore, both the size of privately owned farms (S2) and the size of rented farms (S3) are significant and have a negative sign. The propensity to enter into the agricultural sector is negatively associated with the size of privately owned and rented farms.

Moving to the second model, this explains about 80% of the observed variability and demonstrates how individual schooling levels have a statistically significant influence on the generational turnover. On the one hand, the variables “E1, E2, E3” are significant and have a positive sign. These results mean that with this level of schooling (no school, elementary, and middle) the cost of a young person’s move away from agriculture is high because the opportunity to find another occupation is limited. On the other hand, well-educated farmers (E8, E12, and E13) are negatively associated with the generational turnover. It is conceivable that a higher level of education leads to an exit from the labour force of the agricultural sector. This is because a higher level of education makes it possible to access jobs with higher pay outside the agricultural sector, thus there can be a move in a direction contrary to the generational turnover. This effect is worth less when education is in the field of agriculture (Agricultural Technical Institutes, Faculty of Agriculture), but it does not disappear even for graduates in general and graduates in these disciplines. Regarding experience in the agricultural sector and variables E11–E13, while less experience in the agricultural sector (<3 years) is positively correlated to SI, more experience (from 3 to 10 and >10 years) is negatively associated with it. In relation to this, a greater difference is to be assumed in rural regions with relatively older farmers, while in the regions where the number of younger farmers is higher, the distribution of SI appears to be less concentrated, and this occurs essentially because older farmers have more experience in the agricultural sector.

The third model explains a low percentage (about 30%) of the observed variability and indicates how the organic practices in the agricultural sector do not affect the generational turnover in the same sector. This may be due to the fact that these practices require responsibility, knowledge, and specific technical and organisational solutions.

Finally, the last model explains 56% of the observed variability, confirming the strong regional characterisation of farms with innovation and computerisation. A first consideration is that all the selected variables are quite significant. In detail, a t ratio between coefficients and standard errors reports the relative variation of the innovation and computerisation, which is associated with a unitary variation of the explanatory variables (SI). The p value and the relative significance indicate that when estimated coefficients have positive values, a variation in the reference variables will have a consequence in the same direction as the dependent variable. In contrast, the negative values in the coefficients show the opposite behaviour between independent and dependent variables. The relationship between SI and sales and the marketing of products (IC5) is clearly significant from a gestational point of view and indicates a great difference in the Italian regions where there is a sale of local products, which is probably due to the greater weight of local production [47]. On the one hand, with negative relationships, other predictors that strongly affect the dependent variable (SI) are farms with organisation and business management (IC4) and accounting (IC8). On the other hand, with positive values, are farms that have made at least one innovative investment in the three-year period of 2018–2020 (IC1) and in farm management (IC10). Digital infrastructures are present in Italy, especially in urban regions. Information technology is associated with the phenomenon of skill-biased technological change, which represents one of the main drivers of differences. As a result of this phenomenon, farmers who have greater access to advanced technology enjoy better career opportunities and innovation growth than the less skilled farmers, with the risk that the most innovative regions will be affected by a new malaise, one definable as “technological unemployment”.

5. Discussion

The purpose of this research is to identify the positive and negative factors influencing generational renewal in Italian agriculture. The relevance of understanding the determinants is important to give scientific support to the policy-makers who can facilitate the entry of young people into the agricultural sector. In support of this aim, descriptive statistical analysis, Student’s T-test, and linear regression models were applied.

For the analysis, four groups of factors (the size of the farms, the adoption of sustainable practices, the educational level and experience of holders, and the adoption of innovation and information technology) were created. As argued by Simeone and Spigola (2004) [45], these factors are essential to increase the attractiveness of the agricultural sector. Indeed, the results show that the SI appears greater in the Italian regions where the agricultural sector is more attractive (Trentino-Alto Adige and Valle d’Aosta) or represents the most important sector from an employment point of view (Sardinia). In contrast, the same index is more of a compromise in southern Italian regions (Abruzzo and Apulia), where there is more competition in terms of employment between agriculture and other productive sectors, above all tourism.

Furthermore, the results reported in Section 4 show that the ageing of farm holders is a widespread, problematic phenomenon in Italy, with variable intensity across regions. Firstly, the lack of young farmers puts the survival of the sector itself at risk [48]. This is because the main effect of an inadequate rate of generational turnover is that the exit of farmers from the sector due to old age is not balanced by the entry of new farms run by young farmers. One important factor that has a strong positive impact, in accordance with Glauben et al. (2005) [21], is the greater size of farms. Moreover, other authors [14,15] have pointed out how large-scale farming is much more likely to attract young people to the agricultural sector than small-scale farming.

Another braking factor is represented by the fact that in Italy, compared to other European countries, the cost of leasing and buying arable land is high. Indeed, the stagnant land market, due to an Italian agricultural tradition still focused on small peasant properties and a culture linked to the preservation of such properties, has pushed prices up. This is particularly damaging for young people who are unable to inherit land for farming [49]. One action that could break down this barrier could be to encourage free or facilitated access to public land. Some local and regional governments are pushing in this direction: through the establishment of the Bank of the Earth and the introduction of national measures that favour young farmers, they have managed to encourage the sale of land at more affordable prices. However, access to inputs (land, water, labour and quotas, and production rights) is far from easy, especially in terms of land holdings. The amount of land available is decreasing, and the price of access is continually increasing. Indeed, in Italy, available locations to rent are scarce and often prohibitively expensive. In addition, the competition from other sectors, which often generates more income (tourism, industry, and commerce), is strong both for the land and for other production factors.

In this context, it is crucial that young farmers are able to overcome difficulties in the purchase of agricultural land due to high costs and the lack of available land to lease [49]. A series of measures has been developed at national and community levels with a view to encouraging new entries into agriculture. However, in Italy, agriculture continues to remain an unattractive sector for young people, especially if they do not come from agricultural families.

As regards the level of education and training in the sector, although the level of training of young farmers under 40 has increased over time throughout Europe, access to knowledge and advice is still considered insufficient. In 2016, 43% of young farm managers at the EU level had experience that went beyond practice compared to an average total of 32%.

Finally, for services, there is the issue of the digital divide between the Italian regions, which is hampering the process of modernisation for agricultural activity. Access to knowledge, skills, and innovation is problematic. For young people, it is complex and expensive to start training processes or participate in specific research programmes. Furthermore, in line with other studies [24,25,26,27], the results of the linear regression models show how the adoption of agricultural technology is positively correlated with farm size. The adoption of these technologies in many Italian regions (especially in Liguria, Calabria, and Campania) is suffering from a fragmented and old structure.

Ensuring the entry of young people into agriculture is of primary importance to encourage the introduction of new production methods and practices, such as agro-ecology, which have less impact on the environment and support the development of a healthy economy in the primary sector. Furthermore, the return of young people into agriculture is important for the sector because young people can be more dynamic and innovative. Therefore, if the human factor is the element that characterises the development potential of agriculture in new scenarios, the need to encourage the establishment of young farmers is one of the prerequisites. Therefore, it is necessary to create the proper conditions to make agriculture “interesting” for young people.

These are the topics of main interest for policy-makers who are paying considerable attention to the promotion of multifunctional agriculture, which is considered the key element of rural development. To promote multifunctional agriculture, the starting point is to create the conditions for a generational turnover. In fact, the abandonment of farms leads to negative externalities for the environment. The first result is a concentration of the production and the degradation of the landscape. In addition, the concentration of production in areas with intensive agriculture has negative effects for the environment related to the use of chemical inputs and the reduction in biodiversity in the abandoned area [11]. A successful policy encouraging generational turnover would favour the entrance of young people to the sector, thus substituting older farmers and reducing the number of farms abandoned and the migration of young adults from the sector. The existence of a generational turnover is relevant for agriculture in Italy, which is one of the European countries with the lowest numbers of young adult farmers. A number of policies aim to encourage the young into the agricultural sector in order to reach the objectives of sustainable agriculture and to maintain employment in the agricultural sector. The latest legislation is based on rules to keep young farmers in the sector in the following way: it encourages the young to take over farms from older farmers or to start new ones. Support has been given only to those aged between 18 and 39 years who are interested in taking over a farm from someone aged over 55 years. Geographically, it is the south and the islands that suffer most from the phenomenon, as a result of a regulatory, economic and social environment that is less and less conducive to generational change.

The obstacles to overcome to encourage generational turnover in the agricultural sectors of the Italian regions remain the same, as highlighted by the report on the financial needs in the agriculture and agri-food sectors in Italy by the European Commission (2020) [50]. In particular, there is a large number of small farms with poor integration into the value chain and a substantial presence of family-run farm prices with little or no formal accounting. For these reasons, some financial instruments are necessary to cover the needs of young farmers and their businesses, such as working capital financing, microcredit, and/or complementarity with grants. At national and community levels, a series of measures has been developed to encourage an increase in the average size of agricultural holdings, linked to the reduction in shares and the fragmentation of agricultural holdings.

6. Conclusions

In this paper, a first attempt was made to interpret the drivers and the barriers that encourage and provide obstacles for new entrants into Italian agriculture, drawing on data from the latest agricultural census. So far, the previous literature has neglected the challenging effects of generation renewal in Italian agriculture, which this paper has attempted to address.

The correlation between the substitution index and each explanatory variable has highlighted the fact that, firstly, a propensity to enter the agricultural sector is negatively associated with the size of privately owned and rented farms.

Secondly, the correlation between the dependent and the independent variables related to education is positive for a low level of schooling (no school, elementary, and middle) and negative for a high level of education. This means that, on the one hand, the cost of a young person’s move away from agriculture is high because there are few opportunities to find another occupation. On the other hand, well-educated farmers leave the agricultural sector because they have more possibilities to work in other productive sectors.

Thirdly, organic practices do not affect the generational turnover in the agricultural sector because they require specific responsibilities, knowledge, and technical and organisational solutions that young people do not yet have.

Fourthly, information technology is associated with the phenomenon of skill-biased technological change, which represents one of the main drivers of differences among Italian regions.

The empirical evidence suggests that the ageing of farm holders is a widespread phenomenon in Italy, with variable intensity across regions. The problem of generational renewal affects small farms in all of the study regions due to their lower potential for the adoption of innovative production and sales technologies.

Regardless of the region, the major challenge for transforming small-scale farms into attractive places to work and live for young people is to provide better access to agricultural land, capital, knowledge, and markets. Consequently, there is an urgent need for agriculture policy-makers and practitioners to re-examine their existing predominant focus on addressing the needs and requirements of younger farming generations and to place a greater or equal emphasis on improving the quality of structures of those most affected by the process, namely, older farmers.

Further studies could provide additional evidence and more variables that affect generational renewal in the Italian agricultural sector by addressing some of the limitations of this study, including the use of a single database and a small number of observations for the analysis.

Another limitation is the lack of a theoretical foundation for identifying a quantitative level at which ageing or an absence of youth becomes a social and economic problem for many Italian regions, especially those located in the south.