Results and Discussion: Drivers of agricultural change in each scenario 1. The metabolism of a Minimum Reproduction Unit (MRU)

In document Agroecological Landscape Modelling as a Deliberative Tool, Learning from Social Metabolism Assessment of Historical Transitions to Industrial Agriculture for Future Sustainable Food System (Page 129-140)


5. Results and Discussion: Drivers of agricultural change in each scenario 1. The metabolism of a Minimum Reproduction Unit (MRU)

The first scenario minimizes land requirements of the farming population (Eq. 1) to assess the existing capacity for eco-functional intensification, by identifying the minimum land cost to maintain the three funds. Our representative DU of five people (the average family type in Sentmenat c.1860) would comprise a girl between 0-5 years old, a boy from 5-10 years old, a woman and a man between 18 and 60 years old, and a man older than 60. The results are shown in Figure 6.429. In order to meet their food, fuel and income requirements the surface needed would have been 3.77 has (case 1). Note that, in this case, the family would have had to buy meat and fertilizers from the outside, as there would not have been domestic animals nor nutrients replenishment into the soil.

If we were to include two sheep, a pig and some chickens and rabbits estimated as average for a family farm, we would have had to add 1.81 has more (case 2). The high proportion that in this case would correspond to the rotation with dryland olive groves is consistent with the need to produce fodder legumes (e.g. lupines) to feed livestock (Roca, 2007). Feeding a mule, or providing a share of this feed for the 0.25 mule required for ploughing the minimum land, would have entailed 0.63 additional ha (case 3). If we sum these three surfaces (i.e., if each one would have been devoted to fulfil one goal), our standard family would have required a total amount of 6.21 has of farmland to guarantee the reproduction of both the DU and livestock funds (case 4).

29 As you can see in this figure, we present the three funds as separate goals. While in the case of DU and animals (split into draft power and other domestic animals) we can analyse them separately, fertilization has intrinsically not a cost on surface; you can calculate the increase of the total land cost due to fertilization only with respect to other requirements.

Chapter 6. Beyond Chayanov, SFRA c.1860

However, by taking advantage of the existing multi-purpose options and land-use synergies among funds, together with the possibility of using crop by-products for animal feed, the minimum area actually required for such a family farm would have been reduced from 6.21 to 4.04 has (case 5). This reduction would have been attained thanks to the multipurpose nature of land and livestock uses. For example, farmers used forest for pasture, and for firewood and timber provision, setting complementarities. Feeding livestock in ways that did not compete with human food also played a role. While in the first run of the model animals would be fed with crop main products, such as corn or vegetables, in the second run animal intake would rely on crop by-products such as straw, stubble, husk and pomaces, as well as natural pastures. Other synergies between different compartments of the agroecosystem would appear when soil fertility maintenance was taken into account. As a result, land-use synergies and livestock multi-functionality would allow reducing 54% of the total area required to reproduce the farm unit.

At first glance the maintenance of nutrient cycles in cultivated soils does not change significantly the surface needed: it increases only by 7%, from 4.04 to 4.31 has (case 6). However, this increase of 0.27 has is also associated with a change in the use of pastures in favour to forest as a nutrients’ supplier, as well as a restructuration of land-uses according to soil quality. Thus, LACAS over a purely subsistence condition (case 1) was only 0.54 has, 14% above basic level, while the incorporation of livestock increased farmland surface 8% over farmers’ needs. Here the generation of internal loops through biomass reuses becomes relevant. As the amount of available manure was not enough, farmers had to rely on burying biomass. This was a much more labour-intensive task and represented 37% of the total nitrogen replenishment. This is the reason why the upkeep of soil fertility produces a decrease in the overall energy efficiency (FEROI, the amount of energy product per total input), which drops from 1.18 to 0.43. Likewise, labour energy efficiency (FEROL) decreases from 59.0 to 33.8 MJ produced per MJ of labour performed.

These results reveal that more intensive land-uses required a greater flow of biomass reuses (BR), which in turn led to a decrease in energy efficiency. So, land-use intensity also determines what the energy returns may be. This means that taking as a single criterion the final

Figure 6.4. Land-use distribution for MRU scenario by goals, for a family of five people in Sentmenat c.1860.

Source: Our own.

Chapter 6. Beyond Chayanov, SFRA c.1860

energy efficiency (FEROI), regardless of the reproduction of funds and the LACAS involved, would incur in what Georgescu-Roegen (1971) criticized as an ‘energetic dogma’. Hence, land-use intensification has to be taken into account when assessing the energy efficiency of agricultural systems.

We have also found that the required surface for MRU (case 6 in Figure 6.4) was 4.31 has. Here more than a third was forest, although dryland herbaceous crops also predominated together with olive trees associated to herbaceous rotations (26 and 22% respectively). On the contrary, vineyards only covered 5% (compared to 24% in case 5). Even in this MRU case 5% of production would have gone through market as surplus. From these, a third would have been animal feeding (that would have covered less than 9% of the animal feed requirement). This allows us to draw two conclusions: the adequacy of the prevailing historical rotations fitted the multiple needs of the agroecosystem; and that livestock density proposed in the model was fairly well adjusted to natural endowment—as the average livestock density left few biomass surplus in the optimal conditions.

To find the local possibility frontier to intensify population density, we should weight case 6 for each DU of the municipality (regarding the number of family members). The average surface required was 0.86 ha/person. This value matches a rough estimation made only in monetary terms by Garrabou and Tello (2008)—according to it, 4-5 ha of wheat were necessary to maintain an average family of 4.5 people in the same period and study area. Land distributions of less than 0.86 ha/person would not allow internal reproduction of peasant farms, and families would be compelled to perform other tasks off farm to get money to pay the goods acquired outside to fulfil self-reproduction. Indeed, they might also transfer part of this pressure towards other funds, thus hindering the condition of sustainable reproduction.

Hence, an equitable land distribution would allow 3,228 people to live off the 2,781 has of agrarian land in Sentmenat. In 1860 the population registered was 1,686 people. This potential population increase of 91% would have meant moving from 60 up to 116 inhab./km2, surpassing the threshold of dense-medium population density (7) towards the highest dense rank (8) defined by Boserup (1981). Boserup associated this highest dense rank (8) to the change from short fallow cropping to annual cropping. However, the agricultural system and practices observed corresponds more to her dense population category rather than a medium-dense one (7). Although there were still fallowing practices carried out in Catalonia at that time, it had virtually disappeared in the Vallès County (Garrabou and Planas, 1998). Therefore, we observe a gap between the population density threshold put forward by Boserup and the one registered in Sentmenat in 1860. In other words, there was more intensive land-use than would have been necessary with an equitable distribution of farming resources—a relevant result suggesting that probably social inequality was a major driving force towards land-use intensification, beyond simply a population density pressure (Nell, 1992).

The greatest implication that this strategy would entail in terms of land-use changes at municipal scale would be a shift from vineyards to herbaceous crops of 334 ha. It could seem difficult to reach that in a sustainable manner. However, at least in terms of slope there were around 700 ha of vineyard in slopes lower than 20%, where it would be possible to cultivate cereal crops. Moreover, a study on land suitability of the study area estimated that in 1860 around 60%

of the land in the four municipalities of the Vallès was either moderately suitable or very suitable for wheat cultivation (Rodriguez Valle, 2003). Therefore, we deem that a total surface of around 30% of herbaceous crops would have been quite viable and feasible. In terms of other uses, this would have also required some changes from vineyards to both olive groves and forestland, which does not seem to have presented more difficulties.

Chapter 6. Beyond Chayanov, SFRA c.1860

5.2 An extensification scenario with the Peasant Reproduction Unit (PRU)

MRU defined the possible threshold to minimize land-use requirements and host the maximum population. As Chayanov argued, a peasant economy might respond to different criteria other than labour allocation efficiency, crop diversification and risk minimization (Chayanov, [1925]1986). Peasants might also have been interested in minimizing labour, provided that their land endowment would allow this. So we define objective function 2 (Eq. 2), in which we are minimizing farm labour in the Peasant Reproduction Unit (PRU) scenario.

A reproducible land surface to minimize the labour required in a peasant community equally-endowed of land would be of 5.83 has for a 5-people family type in Sentmenat c.1860.

This is 29% more than for a MRU strategy aimed at accommodating as much population as possible to the available land (see Figure 6.5). As expected, more extensive uses increased over 52% the surface under pasture and forest. Vineyards increased to 14%. The latter makes their role as cash-crop apparent, in contrast with MRU case, where olive groves played a major role. Cereal rotation associated to olive trees yielded higher gross revenues than vineyards, but required 87%

more labour per hectare. Therefore, the aim to minimize labour costs, and get a higher net income, favoured the growth of vines instead of olive trees.

In order to reach the reproduction goals of a peasant family with a lesser labour requirement, a RPU strategy required increasing by 35% the land surface at its disposal.

Conversely, energy labour productivity (FEROL) would have increased 51%, from 33.8 to 51.0 MJ per MJ of labour. This becomes a good trade-off example between labour intensity and land cost (Sahlins, 1971). In the RPU case gains in labour productivity would have been higher than land costs. Therefore, the RPU strategy would have meant an improvement in the population wellbeing with only a moderate increase of land available for each peasant unit.

The comparison between MRU and RPU counterfactuals sets the difference among a hypothetical land distribution (e.g. a land reform) aimed at accommodating as much population as possible on the land, or increasing farmers’ wellbeing. Indeed, the labour-saving RPU case might have allowed a population density up to 86 inhab./km2—still a higher value than the recorded data in 1860. This suggests, once again, that the actual population density was conditioned by the prevailing inequality that allowed a small group of wealthy landowners to accumulate a large share of the better lands, and to manage their estates in an extensive, poly-cultural manner. They rented in tenancy the poorer sloping lands to many smallholders who had no other choice than to farm intensively with vineyards these small plots, so as to supplement the family income by selling their surplus labour to larger landowners (Badia-Miró and Tello, 2014;

Marco et al., forthcoming). Hence, vineyard intensification was not mainly driven by population density. Despite limitations of the counterfactual analysis—as we cannot assume that inequality was responsible for all the divergence between the modelled scenario and the actual situation—, we deem that social conditions to land entitlements played a relevant role. We acknowledge, however, that this should be addressed in further research that would compare the individual farm functioning of different social groups.

5.3. The Maximum Sustainable Specialization (MSS), deepening vine-growing strategy

Using the third scenario, we can go deeper in the analysis of the role of vine-growing specialization in our case study. Land distribution did not respond to the basic needs of peasants’

self-sufficiency. The region was undergoing a process of wine specialization linked to international Atlantic markets (for wine exports) and to inner Iberian Peninsula (for wheat imports). As seen in the previous sections, population dynamics seem were not the main driver behind this agricultural intensification. It seems that the land entitlements that prevailed in the transition from feudalism to agrarian capitalism in Catalonia, and the ensuing social polarization

Chapter 6. Beyond Chayanov, SFRA c.1860

between wealthy landowners and small vine-growing tenants and labourers, played a major institutional role (Garrabou, Tello and Cussó, 2008, 2010; Congost, 2015). The Catalan emphyteutic contract of rabassa morta set a long-lasting lease on the vineyards the tenant planted, by paying to the landowners a third or more of the vintage. It allowed them to plough brushwood and pastureland, and to make a profit from growing grapes without having to carry the burden of planting and hiring labour to manage the vines (Colomé and Valls, 1994). Through these rabassa morta contracts many landless families managed to make a living and stay in the village, providing landowners a local job offer with low wages when they needed to hire day labourers.

As a result, and despite its land-use prevalence (Figure 6.2), vineyard specialization remained a partial option, combined with the poly-culture of the larger estates, as well as with the pluriactivity of smallholder peasant units. In spite of a high market integration level—voluntary for landowners and forced for smallholder tenants and labourers—, all these farms took decisions considering several reproductive dimensions (Chayanov, [1925]1986).

It is interesting to figure out how farmers could have developed further this process of winemaking specialization, under an equitable land distribution and ensuring the agroecosystem’s sustainability. The aim is to identify through this counterfactual analysis whether the actual strategies followed by those socially-polarized farm units created conditions for social and/or ecological instability. Our scenario of Maximum Sustainable Specialization (MSS) is the result of implementing the third objective function (Eq.3) and defining, with the actual population density given, what the highest share of vineyard would have been while ensuring the reproduction of funds, and identifying the factors limiting a further expansion of vines.

Figure 6.5 shows the results, according to different additional assumptions. The standard farm size for five active people would have been 8.25 ha/DU. If we consider the need to satisfy the restrictions of the three funds (MSS1), the highest share of land devoted to vines would have been 26%, far from the 42% in c.1860. Analysing the flow profile, we identify that the limiting factor was labour availability in the key months for vine management such as October (harvest) and April (leaf removal). If we then consider the feasibility of hiring external labour in the peak seasons, as was the case in Catalonia at that time (Garrabou et al., 2015), the vineyard threshold would go up to 58% of cropland (MSS2).

Figure 6.5. Land-use distribution for MRU, RPU and MSS scenarios according to the limiting factors considered, and actual data for Sentmenat c.1860. Source: Our own.

Chapter 6. Beyond Chayanov, SFRA c.1860

The reluctance to depend riskily on the market may have also played a role, as shown by comparing MSS2 with model MSS3. In the MSS3 model, crops would have been aimed at meeting only a basic subsistence fraction of the diet of the local population consisting of vegetables and fruit, as well as firewood for heating their homes. In this case, vineyards could have reached 66%

of the whole municipal area (MSS3), but the village would have had to import 60% of its diet, mainly in the form of wheat and potatoes. In this case, therefore, an important share of reproducibility would be relying on imports, not only of labour or food, but as well regarding nutrient soil cycles. We estimate humanure, under this scenario, would satisfy 19% of nitrogen requirements, thus setting sustainability on external imports of food. The actual figure of wheat consumed covered by imports was 35% in Sentmenat c.1860 (chapter 4), while vineyards covered 42% of the area. The difference can be attributed, once again, to the unequal land distribution that concentrated irrigated facilities and woodland among wealthy landowners, depriving smallholders from the vegetable gardens and firewood self-provision supposed in MSS3 (Garrabou, Tello and Cussó, 2008, 2010).

If we wonder why winegrowing specialization did not reached such a high level in 1860, we immediately see that it would have been necessary to clear more forestland, brushwood and pastureland. About 37% of the total surface c.1860 was forest and pasture, so cultivating these additional 416 hectares would had involved a huge amount of labour, and an important opportunity cost in terms of firewood, timber, wood pasture and nutrients for cropland. Again, we care for the land suitability of this extreme scenario. There were 440 ha of forests and pastures located in slopes lower than 30%. Therefore, regarding this constraint it seem not difficult to have changed them into vineyards, despite the effort this would have supposed in terms of building terraces in the steepest areas (Olarieta et al., 2008). Using again the suitability assessment of for the whole study area (including also the other three municipalities we used in other chapters), only around 21% of the available surface was definitely not suitable for cropping vines. Despite that, 20% of vineyards were actually located within this category, and vine-growers had to deal with all the difficulties that would entail. Thus, we consider that this scenario, although we cannot totally ensure its suitability, might have been reliable under the conditions of land-use intensification we observe for that period—always keeping in mind that this would only had been possible if population would have chosen that goal.

In short, our counterfactual programming models reveal that market vineyard specialization might still have had room for reproducible development, but only under an equitable land distribution, and at the expense of reducing farm self-sufficiency by increasing even further market integration of labour and food.

5.4 Discussion of the scenarios

It has long been criticized that in the study of carrying capacities (Murray, 2009) defining biophysical limits for the anthropogenic use of ecosystems has to consider explicitly which objective is going to be optimized. This holds true for agroecosystems as well. The comparison between our counterfactual SFRA scenarios and the actual ones makes the usefulness of linear optimization models apparent, when they are built from a socio-metabolic accountancy to reveal many of the configurations that agroecosystems and land-uses could have, all sustainable, depending on the main goals farmers adopted.

For example, the upper agroecological limit of 116 inhab./km2 set by our MRU model under an equitable land distribution suggests that the existing cultural and biophysical conditions allowed greater population density. Yet, it is quite likely that social inequality could have prevented this. The actual path was a commercial wine specialization (Figure 6.5) in a socially-polarized rural society ruled by medium and large landowners who offered sharecropping contracts to smallholder winegrowers, forcing them to carry out a highly-intensive vine cultivation. Despite limitations of current model regarding land-use allocations, it seem reliable

Chapter 6. Beyond Chayanov, SFRA c.1860

that this vineyard specialization would still have had a potential for a sustainable expansion, but under an equal social condition. By making these contrasts apparent, our programming models also help us to identify which factors may explain the differences between the counterfactual scenarios and the real data observed, bringing to light the role that institutional settings, social inequalities, market asymmetries and coercive forces might have played (Marco et al, forthcoming).

Following the classification proposed by González de Molina and Toledo (2014), those ruling social agencies were: i) competitive exclusion in access to the land; and ii) parasitic ways

Following the classification proposed by González de Molina and Toledo (2014), those ruling social agencies were: i) competitive exclusion in access to the land; and ii) parasitic ways

In document Agroecological Landscape Modelling as a Deliberative Tool, Learning from Social Metabolism Assessment of Historical Transitions to Industrial Agriculture for Future Sustainable Food System (Page 129-140)