Seasonal variation of price and milk production in Mato Grosso

67 Scientific Electronic Archives Issue ID: Sci. Elec. Arch. Vol. 13 (7) July 2020 DOI: http://dx.doi.org/10.36560/13720201124 Article link http://sea.ufr.edu.br/index.php?journal=SEA&page=article&o p=view&path%5B%5D=1124&path%5B%5D=pdf Included in DOAJ, AGRIS, Latindex, Journal TOCs, CORE, Discoursio Open Science, Science Gate, GFAR, CIARDRING, Academic Journals Database and NTHRYS Technologies, Portal de Periódicos CAPES, CrossRef


Introduction
According to the Food and Agriculture Organization of the United Nations | FAO (2016), approximately 150 million farms worldwide are inserted in the dairy chain. In most developing countries, the same as in Brazil, small dairy farmers have a close relation to the regional industries, and their income from livestock contributes significantly to the subsistence of these families. In contrast, the most significant volumes still being provided by another tiny fraction of Brazilian properties (Marcondes et al., 2017), mainly due to technology application and high herd genetic profile.
Although the milk production per cow has increased in Brazil during the past two decades, the national productivity on average still low (1,304 liters/cow/year) comparing to the largest milk producers countries. For instance, the United States stands out with 10,327 liters, New Zealand with 4,317, and European Union countries with 6,434 per cow milked (USDA, 2016). Regarding low production in some extracts of dairy farms, another significant characteristic of the Brazilian dairy chain involves the fluctuation of raw milk stock, with falls of around 40% depending on the season (Gomes, 2011). The oscillation due to the seasonality compromises the processing capacity of industries all over Brazil (Junqueira et al., 2008;Marin et al., 2011), directly affecting various sectors of the national dairy chain. Issues related to facilities, equipment, and laborers' idleness frequently are reflected in the supply of dairy products for the domestic market and the medium-and long-term strategic planning of these companies.
Moreover, regarding dairy farmers, seasonality usually reflects the reduction in profits due to the drop in the volume milked throughout the off-season (Novo et al. 2013). At the same time, there is an increase in production costs, whether due to the need to provide cattle with roughage supplementation or the greater use of grains to feed them (Junqueira et al., 2008).
Unfortunately, livestock productivity and seasonality are often closely related to the low quality of the raw milk in Mato Grosso. That could be a crucial issue to the non-development of a competitive regional agro-industrial complex, given that of the biggest industries established in Brazil, none of them has activity in the state (Milkpoint, 2017).
This study aimed to characterize the monthly raw milk yield in Brazil, in Mato Grosso, and seven other Brazilian states from 1997 up to 2018. Moreover, the study aimed to determine seasonal variations of the price paid by dairy industries to dairy farmers from 2011 up to 2018.

Methods
Seasonality is a feature of a time series in which the data experiences regular and predictable changes that recur every calendar year.. Factors related to the time that influences the series behavior can be identified and segregated, and a model might contain four patterns: tendential, cyclical, seasonal, or irregular/random (Mendenhall, 1993).
To determine the state and national seasonal variation pattern, we used monthly production and milk´ price series to evaluate the Brazilian dairy market. The same analyzes were done in the states of Mato Grosso (MT), Rio Grande do Sul (RS), Santa Catarina (SC), Paraná (PR), São Paulo (SP), Minas Gerais (MG), Goiás (GO), and Bahia (BA). We gathered the average milk price from historical data of the Center for Advanced Studies in Applied Economics (CEPEA). To determine de price variation in Mato Grosso, the authors accessed information from the IMEA dataset. We estimated the price of chilled raw milk in Reais per liter for each state. Gross prices include transportation cost and taxes (paid by dairy companies/cooperatives), and net prices represent the values received by farmers.
Our final dataset included amounts paid/received for raw milk for every month. The average price for each state was defined based on the participation of weighted milk production in the mesoregions, as pointed out by the most recent edition of the quarterly milk survey from the Brazilian Institute of Geography and Statistics (IBGE, 2018). The national average price was composed by the weighting of average in each state included in the study.Moreover, the weighting of production was calculated according to the last ten years of data published in the quarterly milk survey (IBGE, 2018), based on the average share of formal milk production in each state in the total sampled per month. Thus, we have a weighting panel that takes into account the seasonal production pattern in each state.
To analyze the price seasonality in Mato Grosso, we used the weighted average monthly price series paid to dairy farmers. Complementary, we included in the analysis of other studies and systematical analysis of the milk production chain by Instituto Matogrossense de Economia Agropecuária (IMEA, 2019). The collection of prices for the database follows a method similar to that practiced by CEPEA. The price in Mato Grosso was based on the farmers´ income (without tax rebates). The value of the liter of milk was established as that paid to the farmers by the regional dairy industries (R$/liter).
We gathered data on the second half of each month after the financial statement of cooperatives, considering milk price established to farmers in that month. The method excludes outliers and calculates the weighted price for the participation of each region of the state concerning the proportion of uptake of dairy products. The division of the areas took into account the limits of the municipalities, the formation of the economic centers, geographic coordinates, and the farm localization, according to the Region Map (IMEA, 2008).
To reach the seasonality production index, we set a monthly time series concerning milk yield (liters) from each state, based on data from IBGE according to Quarterly Milk Survey (PTL), from 1997 up to 2018. This period considers the monthly series within PTL, with the most extended available data period.
To eliminate the inflation effects the dataset was adjusted by the General Index Price -Internal Availability (IGP-DI), considering December 2018 as the base month/year for deflation. Then, the Centralized Moving Geometric Average as described by Hoffmann (2002), according to equation (1): Where: Gt = Geometric mean/month t; Pt = Price or production/month t; t = monthly mean. From that average, the seasonal index (It) (Equation 2) was estimated. The seasonal index was calculated, dividing the values of the price series (Pt) by their respective moving geometric mean (Gt), then multiplied by 100.
Subsequently, the seasonal index for each month (Ԑ) was taken (Equation 3). This index indicates the pattern of seasonal price variation (Hoffmann, 2002).

Where equation (4):
Since D j * is the geometric mean of Dj values for the D ij month, we had the following equation (Equation 5): The product of the 12 estimates of Ԑ_j should be equal to 1. So, if the product of Dj * is different from 1, each of them is divided by correction (C) (Equation 6) (Hoffmann, 2002):

)
The descriptive analysis of means and the standard deviation was performed by the CORR procedure of the statistical package SAS ® online.

Results and discussion
Milk yield had an absolute growth within 1997 -2017 in Mato Grosso, presenting an average annual rate of 2.2%. However, in the last decade, there was a reduction of 1.2% in production, with a sharp drop from 2012. Figure 1 shows the evolution of production in recent years, pointing toward a switch in the direction from 2012, with a more accentuated tendency of a drop from 2015. The rate of geometric growth from 1997 to 2007 was 10.3%. However, from 2007 to 2018, this rate decreased to 2.4%, indicating that the growth of milk production was quite lower than in the last decade. The amplitude of the seasonal index observed in our study from 1997 to 2018 was 43.3% for Mato Grosso, while in Brazil, in the same period, this amplitude was 22.9%. The seasonality of Mato Grosso was superior if compared to the other seven states, according to PPM (Table 1).  According to Oviedo-Pastrana et al. (2016), São Paulo and Paraná got seasonality below than the national average. Paraná holds three cities with the highest productive density in the country: Castro, Toledo, and Marechal Candido Rondon. In this case, less seasonality is linked to higher animal productivity. Castro has an average per animal of approximately 7,000 liters per cow per year. In São Paulo, low seasonality was associated with the adoption of intensive production systems, where some of the biggest industrial-scale farms are located (MILKPOINT, 2015).
As seen in Figure 2, the seasonality in Mato Grosso gradually increases in a specific period, with lower milked volumes between the last fortnight of August and early September. It can be concluded that with a relatively short period of feed supplementation, it would be possible to minimize the seasonality of production in Mato Grosso. Although the recommendation seems reasonable, it is hard to find the cause for dairy farmers not adopt that strategy in a state heading crop production in Brazil. Moreover, several authors highlighted the viability of livestock supplementation (Shalloo et al., 2014;Pancoti, 2009;Costa et al., 2011); maybe it is undoubtedly a cultural issue or even a lack of a feasible strategy to transfer technologies. Also, observing the upward curve between September and October, it is possible to check that production already exceeds the cutpoint of the index (100). It can be seen as a gradual droop and then a rapid reestablishment in production levels after an increase in forage offer on pastures.
Pearson's correlation (r 2 ) from 2000 to 2018 (Figure 3), indicates that there is a high correlation between monthly rainfall and milk production (0.8284, p<0.01%). Our results reinforce the hypothesis that the primary vector of the seasonality of raw milk in Mato Grosso is closely related to annual climatic variations and the level of technology adopted (Junqueira et al., 2008). Mato Grosso has a well-defined rainfall regime, with two distinct seasons: a long period of drought, which historically varies from April to the beginning of October; and a long period of rain, between October and March. The climate was classified according to the criteria of Koppen, as Ammonsoon climate, which alternates between a rainy and a dry season (Alvares et al. 2013). During the rainy season, pastures are naturally more abundant, leading to higher volumes of production. On the other hand, during the dry season, there is an acute reduction in milked volumes due to the low availability of forage in pastures (de Castro et al. 2016).
Variations in the rainfall regime are an inherent feature of the Mato Grosso as well as in the entire Brazilian Midwest region. Thus, be prepared to adopt some kind of supplementation strategy for livestock and face clime challenges should not be a big issue for most dairy farmers, but an ordinary necessity for all of them.

BR 6,3%
The montly milk price seasonality index in Mato Grosso has a range of 11%, and the variation in Brazil was 14% in the same period. The analysis indicated two critical periods of low (January to March) and high variation (July to September), with peaks of amplitude in August and February. The same kind of pattern was observed in Brazil for minimal prices; however, the high peak occurs between July and August (Table 2). That happened in a series of price seasonality index in the Rio Grande do Sul for the period from 1995 to 2007 (Viana et al. 2010). The differences between the minimum, maximum, and average prices are 23.28%, 27.81%, and 23.29% lower in Mato Grosso, respectively. The states of São Paulo, Minas Gerais, Goiás, and Paraná had the highest prices, and Santa Catarina, Rio Grande do Sul and Mato Grosso, got the lowest (Figure 4).
The price of raw milk seems not to encourage the sustenance of production in Brazil and Mato Grosso as well. It should be considered different variables in regional scenarios, such as low technological level adopted in dairy ranches, the low quality of milk produced, health problems of herds, the low genetic potential of cows, inadequate and insufficient feeding, reduced reproductive rates, lack of planning and management, and logistical issues. These characteristics contribute to rising production costs, and to impose barriers to dairies commercialization.
The lower amplitude of raw milk prices in Mato Grosso can be a sign of less competition among dairy industries. This fact, combined with a 43% seasonality in production, reinforces the hypothesis that there is a sharp reduction in raw milk supply which it is not followed by an increase in price. This aspect of the dairy market has already been pointed out by Sexton & Zhang (2001). As the supply of raw milk is inelastic, there are structural barriers to production growth. In Mato Grosso, these barriers include climatic variables and technological limitations.
In any productive chain, multiple issues are able to influence the formation of prices, such as external factors (e.g., world supply) as well as internal factors (e.g., climatic risks and commercialization capacity). We must also consider difficulties in storage due to the perishable characteristic of dairy products. These factors indicate that milk price is influenced by the amount that each company receives back for the commercialized dairies, reflecting directing on demand for the products in the retail and wholesale market.
Moreover, the continental dimensions of Mato Grosso (more than 90 thousand km²) contribute to logistical barriers due to enormous distances between municipalities, promoting even more low competitiveness among industries also. Although few companies take a large volume of milk to manufacture, the low quality of dairies and seasonal distortions of milk supply in the domestic market discourages most multinational companies from establishing itself in Mato Grosso.
Based on the Imea dataset, the analysis of the seasonality of manufacturing raw milk points out a range of 35% between the peak and the lower levels (35% -70%) for each year, from 2015 to 2018. Approaching this same data bank, according to the mean test, we found a maximum operation level of 70%, a minimum of 40%, a mean and median of 52% and 53%, respectively. This increase in inactivity, especially in July, August, and September, can be one of the factors that limit prices to increase concerning the domestic market, even in need of raw milk (Table 3). *means with same letters are not significative according to Tukey test p<0,01   Silva et al. (2013) highlight the great difficulty that the sector has encountered in the high seasonality of production, a consequence of different technological approaches on farms, and the influence of climatic factors. Some studies show how the industry in developed countries such as New Zealand and the USA have adapted to address the issue of seasonality adjusting their operational process, offering a wider variety of products, or even halting the production at different times during the year (Nicholson & Fiddaman, 2003).
Based on our results, the industrial inactivity in Mato Grosso was higher than the Brazilian average. This characteristic of the inactiveness rate above 50% dramatically reduces the payback capacity, especially in machinery and equipment. That may be a decisive factor in flattening prices during the off-season since higher incomes are not effective in stimulating growth production.
A strategy to stimulate production in the offseason would be the establishment of quotas. Quota milk is considered the volume in liters delivered by farmers during a particular period. In the states of São Paulo, Rio de Janeiro, Espírito Santo, Minas Gerais, Goiás, Mato Grosso, Mato Grosso do Sul and the Federal District, the quota is established from June to September. The primary reason to adopt quota and an extra-quota is to stimulate milk production along the dry season, when production costs are, in general, higher than those in the rainy months (Gomes, 1988). The lack of pricing policies established by industries and cooperatives in Mato Grosso, including payment for quality of raw milk and incentives to reduce seasonality, or even policies that remunerate the volume produced per farm, can be a factor to non-establishment of a reference price.
As presented before, the Brazilian dairy chain is composed of micro and small companies, with a processing capacity lower than 100 thousand liters/day. In general, these establishments have limited technological potential (Lima et al., 2018a). Foward the same tendency, most industries in Mato Grosso are characterized as a medium or small company, processing from 200 to 5,000 liters daily. Even with inactivity issues, the cooperatives are responsible for approximately 30% of the raw milk manufactured, being always a valuable player and determinant to influence the bottom price. In contrast, the market share of private companies had decreased in the last decade (Figure 4). Ferreira et al. (2008) report that, in general, the cooperatives had fewer problems with inoperability due to its members are closely associated with the corporation, which generates loyalty and commitment. Based on that complex scenario, composed by climatic, public policies, and dairy company issues, a key factor for the dairy ranches is technical assistance. Some Brazilian programs such as "Balde Cheio" (Embrapa), "Educampo" and "Nosso Leite" (Sebrae), Senar Technical and Managerial Assistance Program (ATeG) show positive results all over the country (Lopes et al., 2015). However, Durigon (2019) showed in his study that the actual actions in Mato Grosso to improve the dairy chain are uncoordinated and without alignment. That study also pointed out several barriers faced by local dairy farmers, and some of them were highlighted: genetics limitation of herds, technical assistance privation, lack of nutritional strategies, poor milk quality, absence of infrastructure, and finally, easy access to rural credit.
The need for technical support in dairy ranches is reinforced by checking out data from the last national agricultural census (IBGE, 2017). Bult from census data, Table 4 includes the percentages of establishments receiving or not professional assistance in Mato Grosso, in Brazil, and Santa Catarina. According to the numbers, the state of Santa Catarina has the highest proportion of farms receiving technical support (63%). In contrast, only 19% of farmers received some kind of support in Mato Grosso, private or not. There is a lack of adherence to hiring professionals, not only concerning farms in general (81%) but on livestock ranches, reaching 83% of properties.

Conclusions
The seasonality of raw milk production is higher in Mato Grosso than the other principal milk-producing states in Brazil, however, regarding mik price, the seasonal variation is lower. That oscillation in the supply compromises the processing capacity of regional dairy industries, impacting the price paid to the farmers, and reducing dairies production and commercialization, right in the period of highest output. These issues negatively contribute to the promotion of better policy concerning prices paid to dairy farmers in Mato Grosso.