Analysis of Wheat Yield and Spread in Italy: 1929 Catasto Agrario Data, Guide, Progetti e Ricerche di Storia Economica

Scarica Analysis of Wheat Yield and Spread in Italy: 1929 Catasto Agrario Data e più Guide, Progetti e Ricerche in PDF di Storia Economica solo su Docsity! Wheat production in fascist period. A comparison between high farming, latifundium and sharecropping using the Catasto Agrario of 1929 Giacomo Zanibelli (SSML Unicollege – Departament of Economics, AgrHistoryLab, corresponding author) Vito Ricci (Università di Bari Aldo Moro, AgrHistoryLab) Keywords: Agriculture, Fascism, Battaglia del Grano Abstract In the mid-1920s, the Kingdom of Italy was a strong importer of wheat; In order to reverse this deficit in the trade balance, the fascist regime decided to implement an agrarian policy aimed at achieving, within a few years, full self-sufficiency. This policy is remembered as the "battle for the wheat", which began in 1925 and was mainly directed by Arrigo Serpieri. The measures were aimed to increase the yields per hectare of wheat. The factors that had a positive impact were the progressive spread of mechanization in the countryside, the use of chemical fertilizers, the widening of land credit and the use of selected seed. The present work will analyze the effects of the fascist politics in the first period using the data coming from the Catasto Agrario of 1929; the second experience of this kind after that of 1910. In this case, three case studies (Lombardy, Tuscany and Puglia) will be considered corresponding to three geographical areas of the country (Northern, Central and Southern) and three different forms of land management (high farming, sharecropping and latifundium). After an initial descriptive analysis with provincial data, the focus will be shifted on the yields per hectare by comparing the values of 1929 with the average values of 1923-28 using the data at the municipal level. It will try to understand if there is a link between the yield of wheat and its spread on the territory in the three different forms of conduction and, with different models of linear regression, it will go to identify which explanatory variables had a greater influence on crop productivity levels in different provinces. 1. Introduction The literature has clearly defined how advanced agriculture is the prodrome for the start of an industrial take-off1. If for the northern-European states the start of a new economic cycle took place following the Industrial Revolution for the Italian case, recognized as a late comer, it was necessary to wait for the conclusion of the unitary process. On how and when industrialization began, the literature offers different methodologies of approach and finish points. After the spread of the quantitative methodology, the cliometry, the results of Fenoaltea appear to be the most satisfactory2. If the economist anticipated the Italian take-off at the end of the 1880s of the 19th century, however, it seems indispensable the thesis that sees in the period of the "belle epoque" the true and substantial growth of Italian agriculture; at least if it wants to compare it with the European standards3. As it has been widely demonstrated, the contingencies of the Great War led to slowing this process of expansion. Therefore, it seems more correct to talk about a slowing down than a real shutdown for the agrarian class managed to maintain a certain efficiency, which allowed it to start again decisively at the end of the Conflict, as recent territorial studies have shown. The years following the end of the war were particularly heated and pervaded by strong social tensions and led the ruling class to reflect on how action could be taken to find a viable solution in relation to the demands of the agrarian population. The failure in the lands’ re-distribution had fostered the increase of disagreements that during the conflict had been calmed by the contingencies of the war economy. At the end of the "Biennio Rosso" and with the rise to power of fascism, it was inevitable not to focus on the world of agriculture. In this regard, also through the use of propaganda, a series of policies were initiated that should ruralize Italy4. Mussolini was flanked by highly qualified technicians; and probably fascism benefited from the most dynamic and proactive minds in all those areas related to technical progress; this can also be found in the agricultural sector5. Regardless of the results, it is irrefutable that the fascist agrarian policies were aimed at promoting the development of highly specialized agriculture. In this perspective, there is the so-called "Battaglia del Grano" whose purpose was the increase of the yields per hectare of cereal through the use of new techniques and the use of methodologies proper to applied sciences. In this regard the goal of this paper is to analyze the effects of the wheat battle in three sample regions on a municipal scale: Lombardy, Puglia and Tuscany, as representing the different geographical areas of the peninsula and, at the same time, also of the three main forms of management characterizing the different "Agricultural Italie": high farming, latifundium and sharecropping6. Through a predominantly quantitative approach, but in harmony with the traditional historical methodology, it will descend into the territorial microhistory to acquire additional elements in order to deepen the national dimension. The contribution presents the following structure: a continuation of the contextualization of the historical period, a reconstruction and presentation of the used dataset, an analysis on the evolution of yields and finally some models will be presented to offer an economic and historical interpretation of the case study. If, as we introduced, on the one hand, fascism really bet on the wheat battle, on the other, the historians always looked at the battle a traumatic phenomenon for the Italian agrarian economy7, in particular by focusing on the slowdown in investment, economics and wood crops that made Italy a significant competitor in the international market for quality products. A process that had seen its take-off in the Giolitti’s decade. Between 1891 and 1911 there had been an increase in agricultural production, in particular for some specialized products such as citrus fruits (116.75% lemons; 76.85% oranges). These must be added to the increase in potatoes (131.46%) and rice (54.86%); products such as wheat (31.08%) and corn (36.68%) and wine (61.34%) also showed a markedly positive increase, excluding oil that, instead, has a decrease (-11.61%). It treated about, then, a very prosperous scenario, also due to improvements, which are essential for shifting from an extensive to an intensive agriculture8 system. This was also possible in land settling operations that led to an increase in arable land. If, as it has been anticipated, the Great War slowed down agriculture by favoring traditional crops such as wheat, it must also say that the post-war situation was not flourishing for any of the countries involved, particularly Germany, whose economy came out destroyed by conflict. Notwithstanding its official victory in the war and the institutions’ interventions in the attempt to contain the inflation always in rise since the beginning of the conflict, Italy did not present an economic situation of great prosperity. In addition, the instability of the "Biennio Rosso" years prevented de facto the propulsive climate of the propulsive period of the years before the First World War from reoccurring. Once fascism came to power, it had to decide which way to go to try to revive agriculture, and the technocrats chose what seemed to be most viable one at that time, even on the basis of the echo these initiatives would have had all over the peninsula. Hence the idea of ruralizing Italy, but how could this process have taken place in a time frame useful to the propaganda machine? Agriculture in relation to other production sectors has a significant variable within the production function, the natural cycle; if, on the one side the natural cycle can be beneficial for the cultivations, on the other side, it often lengthens the times of an agricultural specialization process. It was therefore decided to invest in a crop that, thanks to the use of chemical fertilizers and scientific experiments, could soon bring the desired results to show that Italian agriculture had taken a significant step forward. Reducing the imports of wheat and increasing its production was also part of that classic “ruralismo” from the but above all with an improvement in yields per hectare. Factors that had a positive impact, as often reported in the provincial volumes of the Catasto Agrario, were the progressive spread of mechanization in the countryside, the increase in the use of chemical fertilizers, the enlargement of land credit and selected seeds14. If we look at the productivity trend per hectare of wheat (Fig. 1) in the decade 1921-1931 it immediately notices a growing trend going from the initial 10.5 to the end 13. The agrarian policy of the fascist regime, in the short term, seemed to be successful, but it resulted in a significant reduction in the yields of other crops such as olive and vine. Tab. 2 shows data of a linear regression in wheat production and imports in Italy on the WheatBatt dummy variable is 1 in the years following the "Battaglia del Grano" and 0 in previous years. As it can see in both cases, the model is significant, especially in terms of production, highlighting the positive effects of the fascist agricultural policy in the short term. Figure 2 shows the graph of the yield per hectare of wheat in some European countries between 1889 and 193915; it immediately emerges that the yield in Italy was significantly lower than in other more developed countries such as England, France and Germany, but surely higher than a country almost similar in agriculture and development process types, such as Spain. This confirms to us how it seems more significant to compare the Italian case with those realities having strong similarities with those of the Peninsula in respect to advanced agriculture types, such as those in Northern Europe. Figure 1. Trend of wheat yield per hectar in Italy from 1921 to 1931 Source: our elaborations on data of ISTAT, http://seriestoriche.istat.it/ Figure 2. Wheat yield per hectare in some European Countries 1889-1939 Source: Our elaboration on Pinilla (2004). Table 2. Wheat production and import in Italy: before and after the “Battaglia del Grano” Production Import (Intercept) 46.735*** 20.123*** (2.241) (1.848) Dummy WheatBatt 20.958*** -7.004** (3.068) (2.531) R2 0,62 0,21 R2 adjusted 0,61 0,18 F statistic 46,45*** 7,659** N. obs 30 30 Source: our elaboration on dati of ISTAT, Catasto agrario 1929. Notes. Signif. codes: 0 ‘***’ 0,001 ‘**’ 0,01 ‘*’ 0,05 ‘.’ 0,1 ‘ ’ 1. The average production of wheat per hectare is reported as growing in all the three regions: 32,4% in Puglia, where yields continue to remain low compared to intensive agriculture such as Lombardy (24,6% in Lombardy and only 6,6% in Tuscany); this is because, as it was well emphasized by Galassi, the sharecropping already had a form of efficiency within it, and it seemed difficult to increase the yields for the structure of the system itself16. It is during this period that it is interesting to study the Tuscan agricultural system, in order to see the various stages of the introduction of mechanization in the territory; in two provinces only, there is a decline in yields: Sondrio (-2%), characterized by a predominantly mountainous territory, and Arezzo (-1,5%). In Puglia the largest increases occur in the provinces of Salento. 4. The yield per hectare: an analysis The present paragraph is dedicated to a specific topic in the field regarding the wheat yields per hectare. The reasons for such a specific choice are to be found in the importance of the cultivation under an economic perspective. First, it needs to remind that the fascist regime, through the so-called “Battaglia del Grano” implemented an agrarian policy aimed at achieving cereal self-sufficiency in our country17; in addition, wheat is the only crop, among those examined here, to be present to a significant extent in all the three regions, also allowing to assess the effects of the three different forms of agricultural management. The localization quotients18 were calculated by comparing some size (number of farms per municipality, average agrarian area per farm, percentage incidences of crops and yields per hectare) at a municipal level (i) with the corresponding at the level regional (j): These indicators allow to measure and identify those municipalities which, within a region, present a particular concentration of the phenomenon and, therefore, are far from the average regional profile; they can also be interpreted as a measure of specialization. In the event that Qij>1 the municipality i.esmo of the j.th region presents a specialization with respect to the overall regional context for the variable examined. Moreover, as they are pure numbers and the effect of the different average value at a regional level is eliminated, I allow for a better comparison. From the cartograms, (Fig. 3) it emerges that the provinces specialized in the cultivation of wheat are Foggia, the Padane provinces of Lombardy: Cremona, Mantua, Milan and Pavia, which are all flat areas and historically characterized by wheat production; In Tuscany, provinces specializing in the cultivation in question do not seem to emerge. 𝑄𝑖 𝑗 = 𝐴𝑖 𝐽 𝐵𝑖 𝐴𝑗 𝐵 Source: our elaboration on data of ISTAT, Catasto agrario 1929 On the axis of the x it is reported the yield of the initial period (1923-28), while on the orderly axis that of the final period (1929); In blue are indicated the municipalities of Puglia, in green the Tuscan ones and in red the Lombardy ones. There are also two lines: the dotted one is the bisector, the points that are above this indicate that an improvement in yield for the municipality, while the continuous one is the regression line expressing an indication of the average growth between the two periods with an angular coefficient of 1,189546, from which an average growth of 18,9% can be retrieved; the points above the regression are those pertaining to those municipalities for which the yield has increased more than the average of the data of the three regions. The increase in municipal yields is not uniform in the three regions: in Puglia in 98,8% of municipalities there is an improvement in productivity per hectare, whilst in Lombardy this percentage is lower and equal to 85,7% and in Tuscany to 79,7%. The Puglia case is symptomatic, for it shows a good response to rural policies, and the same reasoning can be extended to Lombardy. Although the Tuscan case has its own peculiarities, already stated above, it is noted that the percentage change is significant and this allows us to say, that, even if the sharecropping soured by nature an efficiency within it, there was a rural sector's effort to enhance cereal farming by showing a similarity to the national trend. However, if it examines how many municipalities have grown more than the average, it is that they are 91,7% for Puglia, 46,5% for Lombardy and only 18,1% for Tuscany. These percentages show that the increase in yields per hectare of wheat was particularly strong for Puglia, a region presenting the lowest average yield in the period 1923-28 and, therefore, with better room for improvement. Another aspect of yields that has been addressed is the verification of, whether or not, there is a link between the percentage incidence of the area destined for wheat and the yield of the same crop per hectare, i.e. whether or not the largest or least spread over the territory of the crop affects yield levels. Figure 5 shows the scatterplots of the regions under examination. Figure 5. Wheat yield per hectare vs wheat percentage agrarian surface in Puglia, Lombardia and Toscana Source: our elaborations on data of ISTAT, Catasto agrario 1929 In Tab. 3, it is reported the regression data between the two variables. It can immediately observe that there is a positive relationship between the spread of culture and productivity per hectare in all the regions examined, albeit with some differences. Before proceeding with the analysis of the models, it is important to remember, that the incentive to grow yields could also be attributable to the prizes of the various competitions on productivity19. In particular, by examining the parameters of the regression lines, a different value is found in the intercept reflecting the different value of yield per hectare in the regions; The highest regression coefficient is recorded for Lombardy, where an increase of 1% in the cultivated area averages (0,425) tons in yield, while Puglia (0,163) and Tuscany (0,110). If, on the other hand, it intends to measure the intensity of the bond through the R2 determination coefficient, the highest value is for Puglia, with 0,392, slightly lower than that of Lombardy (0,339), while that of Tuscany is almost completely irrelevant. It can admit, therefore, that there is a direct link between the two variables, and it has a level of intensity of a certain importance, although below the 0.5 threshold, for regions characterized by extensive and intensive agriculture, while the bond is practically negligible in the case of the Tuscan sharecropping. This is because Tuscany had a defined and consolidated territorial connotation over the time. The regression also confirms that productivity was still linked to typical characteristics that prevented a momentum comparable to that of other systems; to this must be added the peculiarity of the work factor. It is therefore clear that in this first phase of the verification of the effects of ruralism, the sharecropping system, notwithstanding the improvements seen above, was already efficient, and such a model of conduct could hardly lead to the hypothesis of a yields’ exponential growth. Table 3. Estimates of parameters of linear regression wheat yield per hectare vs wheat percentage agrarian surface in 1929 per region Predictors Lombardia Puglia Toscana (Intercept) 14,604*** 9,351*** 9,951*** (0,3166) (0,250) (0,290) % wheat surface 0,425*** 0,163*** 0,11*** (0,0162) (0,013) (0,015) Observations 1351 241 160 R2 0,339 0,392 0,166 R2 adjusted 0,338 0,389 0,162 Source: our elaborations on data of ISTAT, Catasto agrario 1929; Note. Signif. codes: 0 ‘***’ 0,001 ‘**’ 0,01 ‘*’ 0,05 ‘.’ 0,1 ‘ ’ 1 The Linear regression analysis was also used with the provincial data, using the 1929 wheat yield per hectare as a response variable together with some explanatory variables that can affect yield levels at a provincial level. It has already been seen in a previous paragraph how the determinants that positively influenced, as often reported in the provincial volumes of the Catasto Agrario, were the progressive spread of mechanization in the countryside, the greater use of fertilizers enlargement of land credit, the use of selected seeds. Among these variables, the only one for which data are available for all the provinces is the amount of chemical fertilizers used; A number of inputs have been taken into account: agricultural workers and livestock, the latter, at the time, was also an important workforce on many farms; only at an example level, it is found that in 1929 in the province of Foggia there were 5,482 agricultural vehicles, instead in the Padane provinces of Cremona and Pavia this figure was around 18,000, in the Tuscan provinces (Siena, Pisa, Grosseto) was slightly lower and 10- 13,000 units. The explanatory variables have been properly normalized to make them homogeneous: the use of chemical fertilizers has been compared to the agrarian surface, thus having the figure per hectare (Fertilizer); the number of agricultural workers has been compared to the number of companies, thus having the average number of workers per company (Workers); the population with head of the agricultural family was compared to the present population (AgrPop), in order to obtain an index measuring the degree of dependence of the provincial economy from agriculture; the number of cattle (cattle and equines) was divided by the number of farms, obtaining the average number of cattle per farm (Livestock). Tab 4 shows the average values of the variables used in linear regression for the three regions. The main differences, in addition to the wheat yield, are found in the average amount of chemical fertilizer per hectare used, and the number of livestock per farm; in high farming in Lombardy the amount of chemical fertilizers used is twice as high as Tuscany and four times that of Puglia, while the number of animals per company is 2,82, compared to 0,99 in Puglia and 2,07 in Tuscany. There are no significant differences, however, with regard to the number of employees per farm, which is around 2,5, and the impact of the agricultural population on the present: for both Puglia and Lombardy it is about 26%, in Tuscany the level is slightly lower (23,4%). Table 6. Parameters estimation of linear regression models SAR with maximum likelihood method A B C D E F G Predictors (Intercept) -0,713 -0,627 -1,733 -1,215 -2,123 -1,370 - 173,5345*** (2,40418) (2,63377) (2,22467) (2,11134) (2,41074) (2,71504) (47,9509) Workers -0,182 -0,710 0,252 -0,857 -0,248 (0,47365) (0,85555) (0,37649) (0,784481) (0,75586) Livestock 1,37549*** 1,5316*** 0,265 0,143 0,040 (0,36784) (0,41245) (0,5408) (0,645308) (0,55264) AgrPop 0,051 0,046 0,113164. 0,050 (0,07264) (0,03414) (0,067869) (0,06819) Fertilizer 3,36529*** 2,96255* 3,60711*** 3,970361** 3,6941** (0,71525) (1,3198) (0,67289) (1,434887) (1,25184) Longitude 1,7121*** (0,4767) Latitude 3,655*** (1,06146) Rho 0,85657*** 0,82815*** 0,85721*** 0,8533*** 0,82837*** 0,77082*** 0,345 (0,09143) (0,10762) (0,08999) (0,092294) (0,10515) (0,13225) (0,32016) Log- likelihood -59,787 -59,565 -57,481 -57,583 -56,855 -56,323 -51,395 AIC 129,570 131,130 124,960 125,170 123,710 126,650 120,790 Nagelkerke pseudo-R- squared 0,722 0,733 0,778 0,776 0,789 0,799 0,869 Observations 23 23 23 23 23 23 23 Signif. codes: 0 ‘***’ 0,001 ‘**’ 0,01 ‘*’ 0,05 ‘.’ 0,1 ‘ ’ 1; Notes. Source: our elaborations on data of ISTAT, Catasto agrario 1929 The figure 6 shows scatterplots between yield per hectare and four explanatory variables used in models; in all cases a positive bond is observed, it has a greater intensity for the variables Fertilized and Livestock. Figure 6. Scatterplot wheat yield in 1929 vs predictors used in linear regressions per province Using the same data already used for regression, with the exception of geographical variables, the DEA analysis (Data Envelopment Analysis) was conducted to estimate the frontier of the production function and calculate the production efficiency of the wheat yields for each province examined23. It was decided to use the DEA to observe the efficiency of the provinces, in order to identify further confirmations of what emerged from this study regarding the development of wheat yields in the three sample regions. In Tab. 7, the efficiency indexes at a provincial level are presented, whilst in Tab. 8 the frequency distribution (absolute and percentage) of these indices is reported, and in Tab. 8 the percentage impact of the provinces with an efficiency index of 1 on the provinces, to obtain an efficiency assessment for the different types of agricultural management examined. The efficiency analysis shows no diverging results in respect to those previously emerged. One striking fact is that the average efficiency is 0,858, a particularly high value showing how the battle for wheat was also aimed at achieving agricultural efficiency. The most efficient region is Lombardy, followed by Puglia and Tuscany, which would be inefficient if placed in a wider context. What it has emerged from the DEA, however, must not divert the attention from the fact that it seems difficult to carry out comparative analyses of such different production realities, but it allows to observe that the agricultural land settling policies in Puglia, in particular in the captaincy area, had brought most of the provinces into a full efficiency condition. It has also verified whether the efficiency affected the change in yields between 1923-28 and 1929. This analysis also confirms that Tuscany is below the growth of other regions. The only exceptions were the provinces of Massa Carrara and Livorno, probably due to a greater small farms presence that favoring an increase in efficiency compared to the other provinces. The DEA analysis allows to confirm what it emerged from the sharecropping literature, i.e. the comparison of the sharecropping system with other realities doesn’t seem to be so coherent, and this would also be confirmed by the Sienese situation, which is in a position of countertrend towards the real conditions of the Sienese agriculture. Table 7. Wheat yield efficiency indicators according the province Province Efficiency BARI 1,000 BRINDISI 1,000 FOGGIA 0,887 LECCE 1,000 TARANTO 0,975 BERGAMO 1,000 BRESCIA 1,000 COMO 1,000 CREMONA 1,000 MANTOVA 0,452 MILANO 1,000 PAVIA 0,754 SONDRIO 1,000 VARESE 1,000 AREZZO 0,614 FIRENZE 0,761 GROSSETO 0,356 LIVORNO 1,000 LUCCA 0,954 MASSA CARRARA 1,000 PISA 0,657 PISTOIA 0,931 SIENA 0,396 Mean Efficiency 0,858 Source: our elaborations on data of ISTAT, Catasto agrario 1929 Table 8 Distribution of wheat yield efficiency indicators among provinces Efficiency range N. % 0,31-0,4 2 8,7 0,41-0,5 1 4,3 0,51-0,6 0 0,0 0,61-0,7 2 8,7 0,71-0,8 2 8,7 0,81-0,9 1 4,3 0,91-0,99 3 13,0 1 12 52,2 Total 23 100,0 Source: our elaborations on data of ISTAT, Catasto agrario 1929 Table 9 Distribution of province with wheat yield efficiency indicator ugual to 1 among regions Regione N. efficiency=1 Total %