Economic Perspectives on Analysis of Ensuring Cereal Production and Consumption Security

Cereals are essential for human nutrition. However, the ever-increasing world population makes it difficult to maintain the cereal production and consumption security. It is essential to overcome this situation by increasing cereal yield. An additional issue is the fact that while some countries suffer from hunger, a significant amount of food is discarded in others. This study analyses both production and consumption of cereals with the goal of ensuring food security. On the production side, many developing countries lack production capacity in contrast to developed countries. This is mainly due to a lack of capital, technology and human resources skills. In this study, we first theoretically demonstrated a cereal production gap between developing and developed countries. Second, we performed an empirical analysis to confirm the theoretical demonstration. On the consumption side, we focused on the cause of food loss and waste. We apply economic theories to demonstrate the situation where food loss and waste are occurring in the market. Then we introduced the related data to interpret the current world situation. Finally, we discussed the potential measures to improve cereal production and consumption.


Introduction
Cereals are essential for human nutrition. However, the ever-increasing world population makes it difficult to maintain the security of this food source. It is essential to overcome this situation by increasing cereal yield. An additional issue is the fact that while some countries suffer from hunger, a significant amount of food is discarded in others. This study analyses both production and consumption of cereals with the goal of ensuring food security.
On the production side, many developing countries lack production efficiency mainly due to a lack of capital and human resources skills. First, we theoretically demonstrated a cereal production gap between developing and developed countries. Second, we performed empirical analysis to confirm the theoretical demonstration. We also clarified how the production and consumption status of three typical cereals (wheat, rice and maize-or corn in United States English) are related to the regional food cultures.
On the consumption side, we investigated the cause of food loss and waste. We applied economic theories to demonstrate the market situation where food loss and waste are occurring. We then introduced the data relating to food loss and waste in G20 and the least developed countries in the world to interpret the current world situation. In addition, we also considered some cases of macroeconomic impacts on cereal productions (i.e. weather extremes and changes in commodity prices).
The remainder of this chapter is organised as follows. Section 2 provides the formal definition of food security. Section 3 provides both the theoretical and empirical analysis of cereal production. Section 4 analyses the causes of food loss and waste. Section 5 interprets the additional issues of cereal production (i.e. the macroeconomic impacts) and Section 6 concludes the chapter.

Food security
According to the Food and Agriculture Organisation of the United Nations (FAO), food security exists when all people-at all times-have physical, social and economic access to sufficient, safe and nutritious food that meets their dietary needs and food preferences for an active and healthy life [1]. The issue of food security has been raised as a major agenda at the United Nations (for example, Goal 2 of the Sustainable Development Goals: End Hunger, Achieve Food Security and Improved Nutrition and Promote Sustainable Agriculture) [2].
The above agenda is mainly due to the expected increase in the world population potentially make it difficult to ensure food security with the current level of food production. Overcoming this situation will require increasing the production capacity of nutritious foods for human consumption (especially wheat, rice and maize). However, the cereal production capacity varies across countries. In particular, there is more room for improvement in the crop yield in developing countries than in developed countries. Moreover, a significant amount of edible food is being discarded without being consumed in many developed countries whereas some developing countries concern food shortages. Eliminating such consumption waste can contribute to food security as much as increasing production capacity.

Theoretical analysis
In this section, we used a simple economic model to illustrate the problems of cereal production in developing countries and the methods for its improvement 1 . In order to highlight the problem of cereal production in developing countries, we comparatively analysed developing and developed countries. A farmer produces a product (i.e. cereals) by using labour and capital as demonstrated in Eq. (1). Eq.
(1) takes the form of a Cobb-Douglas production function, where Y i is an output of cereals. i is a subscript that informs on whether the farmer lives in a developed or developing country (i.e. i ¼ R, i ¼ P for developed and developing countries, respectively); L i denotes the labour in country i and K i denotes the capital used to produce cereals (e.g. machinery). The degree of contribution of L i in producing cereals is measured by α i . We assume that α R > α P . Because capital is often exported from industrialised countries to the rest of the world, we assume that there is no significant difference in performance of capital between developing and developed countries. In this case, the degree of α i is determined by the production capacity of L i relative to the performance of K i . By considering its relationship to K i , labourers in developing countries are often under trained and less productive than in developed countries.
Next, we investigated the relationship between labour and capital in production for further details. Specifically, by fixing the production of cereals at a certain level (i.e.Y i ), the marginal rate of technical substitution of i (MRTS i ) is derived as Eq. (2). 2 According to Eq. (2), MRTS P is larger than MRTS R . This indicates that, as compared to developed countries, introducing one unit of capital can replace a larger number of labourers to produce a given amount of cereal in developing countries. This is because, as compared to when there is enough farming machinery, productivity is more likely be improved by introducing one machine when there is a shortage of farming machinery.
Next, we set the farmer's budget constraints. To use labour and capital, wages and purchase/maintenance costs for capital resources were incurred, and the farmers are supposed to cover them out of their budgets. Eq. (3) shows the farmer's budget constraints.
B i is a farmer's budget in country i, I i is an income of L i and C is the cost of purchasing and maintaining K i . We assumed that B R > B P and I R > I P because, in most cases, the income level in developed countries is higher than in developing countries. In contrast, we did not distinguish C in terms of subscript i. This is because capital is often exported from industrialised countries around the world, and hence its prices do not differ largely between countries. By using Eqs.
(1)-(3), we demonstrate the farmer's cereal production regarding to its first-order conditions. The first-order condition is derived by solving the problem with a Lagrange multiplier (Eq. (4)).
where LF i is the dependant variable of the Lagrange function in country i and λ i is a Lagrange multiplier of the problem in country i, respectively.
To solve the problem, we first differentiate LF i with respect to L i , K i and λ i shown in Eqs. (5)-(7), respectively. 2 The MRTS i measures the number of L i to be replaced by increasing one unit of K i to maintain a certain level of production (i.e. Y i ).
By using the first-order conditions of Eqs. (5)-(7), we can derive the optimal level of L i (L * i ) and K i (K * i ) as follows: Eq. (8) specifies the optimal number of labourers to produce cereals in country i. By assuming that the difference between B i and I i in developing and developed countries is the same (i.e. B R I R ¼ B P I P ), the optimal number of labourers in developing countries is smaller than in developed countries.
Eq. (9) specifies the optimal amount of capital resources to produce cereals in country i. Due to the comparison of B i and α i between developing and developed countries, we are unable to comparatively conclude the size of K * R and K * P . The optimal set of inputs L * R , K * R À Á and L * P , K * P À Á are illustrated in Figures 1 and 2, respectively.
The figures are two-dimensional, with the number of labourers in country i (L i ) on the vertical axis and the quantity of capital in country i (K i ) on the horizontal axis. The negative slope straight line in each figure illustrates the farmers' budget constraints in country i (B i ). 3 The curves illustrate the farmers' production in country i (Y i ). 4 In this situation, the farmers' best option is to produce a certain amount of cereals by minimising the cost, which is reflected by the points where the production curve touches the budget constraint.
As illustrated in Figures 1 and 2, due to the small budget, the optimal amount of both inputs in developing countries is less than in their developed counterparts. 5 However, closer inspection reveals that the optimal amount of capital is relatively larger than the number of labourers in developing countries. This is due to the lack of skilled labourers and the fact that there is more room for productivity improvements through capital in the case of developing countries. 6 Even though this is the optimum input situation based on the model, the situation in reality differs. In many developing countries, development support has not sufficiently progressed, and agriculture is often carried out manually. 3 The farmers are only able to purchase the set of inputs allocated inside (lower left) of the budget constraints. 4 The curve of Y i illustrates the sets of L i and K i to maintain the fixed amount of the farmer's production in country i. Moreover, slope of the curve is reflected by size of MRTS i (without the negative sign). Because MRTS P is larger than MRTS R , slope of the curve in Figure 2 is steeper than Figure 1. 5 In this case, it is assumed that the difference in B i is larger than the difference in α i between the countries. 6 However, as the country develops, the optimal input of labour and capital changes. As the country develops, labourers acquire higher skills. Moreover, capital accumulation leaves less room for marginal productivity gains. Hence, the developing country's optimal condition converges toward that of developed countries.

Empirical analysis
Although we theoretically demonstrate the optimal input of cereal production in developing and developed countries, it does not reflect the actual situation. In this  section, we investigate the empirical data to identify the discrepancies between the optimality condition (L * i and K * i ) demonstrated in 3.1 and reality.

Data source
The data employed for the empirical analysis was extracted from FAOSTAT. 7 Specifically, the production quantity (Element code: 5511) and domestic supply quantity (Element code: 5301) of wheat (Item code:

Data description
Tables 1 and 2 in the Appendix show the domestic production quantity and the production quantity of wheat, rice and maize, the number of tractors used for agriculture and the population in the world's least developed countries and the G20 countries. Close inspection of the tables reveals the problems related to cereal production and food security in developing countries.
First, in many developing countries, both the quantity of production and the domestic production of cereals are far below those of the G20, which indicates that developing countries lack cereal production capacity. Moreover, many G20 countries' production exceeds its domestic supply, indicating that those countries import cereals to cover shortages in their national consumptions. However, the opposite is true in the case of the least developed countries. A low level of production is indicative of threatened food security.
Second, although based on scarce information, the number of tractors in the G20 countries is significantly higher than in developing countries. As demonstrated in the theoretical analysis, abundant machinery contributes to the production of cereals. The lack of machinery likely causes the difference in the productivity between developed and developing countries.
Third, the production of wheat, rice and maize varies according to the country's region. For example, wheat production in many countries in Table 1 is lower than that in countries in Table 2. This is because many developed countries are Western countries whose staple food is wheat (i.e. bread). On the contrary, many countries 7 FAOSTAT (http://www.fao.org/faostat/en/#data). 8 The production quantity is a value that considers import / export and change in stocks from domestic production quantity. 9 Due to a small data availability, the author decided to show the number of tractors in 2006, which reports more data than other years. 10 The author identifies G20 countries in the 'countries' and the least developed countries in the 'special group: Least Developed Countries > List) in the domains of the items (i.e. new food balance, machinery and annual population). We excluded the European Union (EU) from G20 because the latter is not a single nation.
in Table 1 produce more rice. This is because rice is often consumed in Asia and Africa, and those regions include many countries in Table 1. Unlike the other two cereals, maize is actively produced in the countries listed in both Tables 1 and 2. This is because maize is often consumed in Africa as well as in Western countries.

Empirical analysis
To make the discussion in 3.2.2 more reliable, we performed regression analyses. We confirmed the first (as compared to developed countries, developing countries lack cereal production capacity) and third argument (i.e. the production of wheat, rice and maize varies according to the country's region). 11 In Eq. (10), the dependent variable Cereal cit represents the domestic supply quantity of cereal c in country i at year t. 12 Subscript c distinguishes three types of cereals (i.e. wheat, rice and maize). Subscript t specifies the year in the sample period. 13 The explanatory variable G 20i is a dummy variable that takes the value of 1 if country i is a member of G20, and 0 otherwise. Africa i , Asia i , America i and Europe i are dummies that take the value of 1 if country i is in that region and 0 otherwise. 14 β c0 , β c1 , β c2 , β c3 , β c4 and β c5 are a constant term, the coefficients for G 20i , Africa i , Asia i , America i and Europe i in the case of regressing for cereal c, respectively.
P 2018 t¼2015 Year t is the sum of the dummies between 2015 and 2018 for capturing the year fixed effect. 15 ε it is an error term for country i at year t: 16 Table 3 presents the descriptive statistics for the data employed for the regression analysis of Eq. (10). As mentioned earlier, the G20 production of all cereals far exceeds that of the developing countries. The difference is particularly noticeable in maize production and smallest of the three (wheat, rice and maize) in rice production. This is because there are many countries that have rice as their staple food in poor regions such as Southeast Asia. Table 4 shows the estimation results of Eq. (10). Close inspection of the estimation results reveals some features that are consistent with the argument in 3.2.2. First, the estimated coefficients of G 20iβc1 À Á in the case of all three cereal types are 11 Due to the small amount of available data, we did not perform the regression analysis on the second argument (i.e. the number of tractors in the G20 countries is significantly higher than that in developing countries). 12 We separately regressed Eq. (10) for the three cereals types (i.e. Cereal Wheat i t , CerealRice i t and CerealMaize i t).
Although i is used to distinguish developed or developing country (i.e. i ¼ R, i ¼ P) in subsection 3.1, in this subsection, i is used to distinguish the countries in the dataset (i.e. i ¼ the US, i ¼ the UK, etc … ). 13 Although Tables 1 and 2 only show the 2018 data for each country, we decided to use the 2014-2018 data for the estimation. This is because a larger data size helps to perform a more significant quantitative analysis. 14 To deal with collinearity, we dropped Oceania i from the regression equation. 15 To deal with collinearity, we dropped Year 2014 from the regression equation. 16 The data employed in subsection 3. positively estimated with strong significance, indicating that G20 countries produce larger amounts of all three cereals than the least developed countries. Second, by looking at the estimation result of the regional dummies, differences in cereal production by region can be found. For example, in case of Cereal Rice i t , the size of the estimated coefficient for Asia iβRice 3 À Á is significantly larger than those for other regions. This finding makes sense given that rice is the staple food for people in Asian countries. Similarly, in the case of Cereal Maize i t , the size of the estimated coefficient for America iβMaize 4 À Á is much larger than those for other regions. This is also understandable because people in the United States consume many foods made from maize. In case of Cereal Wheat i t , estimation of the regional dummies is different from our expectation. Given that wheat-based foods are mainly consumed in Europe, the estimated coefficient for Europe iβWheat 5 À Á was expected to be larger than for other regions. However,β Wheat 3 is larger than that ofβ Wheat 5 . This might be due to China's wheat production, that surpassed that in European countries. Finally, the low R 2 value in all three cases indicates that the explanatory ability of these estimation results is not high. This is because all the explanatory variables (except the constant) are dummy variables.
In order to comparatively analyse cereal production in the developed and developing countries in greater detail, in addition to Eq. (10), we conducted another estimation procedure. The production of cereals in a country depends on its population. By using per capita productivity, we can exclude the impact of the population. In order to reveal the latter issue, we set Eq. (11) as follows: Eq. (11) is different from Eq. (10) in that it uses Cereal Per Capita cit as a dependant variable. This variable was obtained by dividing Cereal cit by population of country i. 17 Table 5 shows the descriptive statistics for Cereal Per Capita cit . Information in Table 5 is significantly different from Table 3. Specifically, as compared from the developed country, the mean value of Cereal Per Capita Rice i t is higher in the least developing countries. This is probably because the large populations in developed countries used as a denominator value of Cereal Per Capita Rice i t lowered their value of Cereal Per Capita Rice i t . Table 6 shows the estimation results of Eq. (11); the interpretation of these results requires careful consideration. In the case of wheat and maize, the coefficients of G 20i (β Wheat 7 andβ Maize 7 ) are positively significant, indicating that even if the effects of the population are eliminated, G20 countries are more productive than the least developed countries, which is consistent with our view. In contrast, in the case of rice, surprisingly, the estimation of G 20i (β Rice 7 ) is negatively significant. This is probably because G20 countries with rice food cultures include populous countries (i.e. China and Indonesia).
In the case of rice, looking at the estimated result of the regional dummies shows that only Asia i (β Rice 9 ) is positively significant. This confirms that Asia stands out in 17 The regression of Eq. (11) additionally uses population data of the countries in 3.  as β c6 , β c7 , β c8 , β c9 , β c10 and β c11 , respectively. terms of rice productivity. Similarly, in the case of maize, the estimated coefficient of America i (β Rice 10 Þ is significantly larger than those of other regions, indicating that maize productivity in the American region is higher than in other regions. One unexpected observation is that, in the case of wheat, all regional dummies are estimated to be negative. Because the estimated coefficient for Europe i (β Wheat11 Þ has the smallest negative value, we can argue that the per capita production of wheat in Europe is the largest; however, the reason for this is not clear.
Compared to Table 4, all the R 2 values are slightly higher in Table 6. Therefore, the explanatory power of the estimation results improved slightly in Table 6.

Cereal consumption
In Section 3, we describe both the theoretical and empirical analyses of the cereal supply. In this section, we identify the problems with and the methods for improving the consumption of cereals as food security issues. One might think that the only issue regarding food security is the improvement of production capacity; however, management of consumption is also important. Among many problems relating to cereal consumption, we focused on food loss and waste.
The food situation differs from country to country. While some countries suffer from poverty and lack of food, others have excessive food supply in their markets and dispose of consumable foods. Table 7 shows the amount of food lost in the least developed and G20 countries. 18 Not surprisingly, as compared to that in the least developed countries, the larger amount of food loss is reported in the G20 countries. Considering the larger production volume and the population, the loss in the G20 can be understood as a natural consequence of its food consumption. However, the developing countries are not completely devoid of food loss; although small comparatively, food loss is also reported in the least developed countries. However, reasons for food loss are different between the G20 and the least developed countries. In the case of the G20 countries, the cause of food loss is overstocking. For example, in Japan, consumers strongly demand food safety and quality [3]. As a result, foods that are not sold by the expiration date are discarded [3]. Moreover, consumable foods with slight scratches or incompatible size are discarded [3]. Another issue is that the main concern of business food suppliers is to make profit. When such economic agents compete with each other to form a market is a basis of capitalism. However, in order to make profits, they sometimes make decisions that are detrimental to society. For example, convenience stores in Japan often overstock foods, because they do not want to lose customers due to out of food stock [3].
In the case of developing countries, due to a lack of sufficient capital for processing and preserving food, some foods are rarely delivered to consumers while fresh. For example, in developing countries, most of the postharvest grains are stored in traditional storage structures, which cannot prevent insect infestation and mould during storage [4]. Figure 3 illustrates an excess supply of food (caused by overstocking) in developed countries. 19 According to economic theory, if goods are traded at their market equilibrium (i.e. the intersection of the demand and supply curves), production and 18 Because the food loss data is reported from a limited number of countries, the countries with no data were deleted from Table 7. 19 S, D, P, Q , Q S and Q D in Figures 3 and 4 represent supply, demand, price, quantity, quantity supplied and quantity demanded for a product in the market. consumption match and no goods will be left unsold. However, as shown in Figure 3, if the price of goods deviates from its equilibrium, supply and demand of goods do not match, and some goods are not sold. Figure 4 illustrates a food shortage situation in poor countries. 20 Unlike in Figure 3, the supply curve in Figure 4 becomes vertical at a certain quantity (Q S ). This shows supplier's limit of cereal production capacity. Due to the limited production capacity in poor countries, once the supply reaches its limit, the quantity does not increase regardless of the price. Moreover, in this situation, the price of the goods deviates downward from its equilibrium. The discrepancy between supply and demand becomes the shortage.
In theory, eliminating oversupply of food and the assurance of food security in developing countries is mutually achievable. For example, transferring food surplus from developed countries to developing countries to combat the shortages would provide a solution to both problems.

Measures to ensure world food security
As discussed in Sections 3 and 4, there are the challenges to ensuring world food security both in food supply and consumption. On the supply side, the main reason for food insecurity is the lack of capital and training of human resources in developing countries. However, due to lack of funds, increasing capital and training human resources are hardly be achieved. Therefore, loans and assistance from international organisations and developed countries are required. Additionally, investment projects (i.e. foreign direct investment) are also effective in raising funds for developing countries. Furthermore, to gain an understanding of the outcome of such investments is larger in developing countries than in developed countries will be important 21 .
There are other points that need improvement. First, improvement in seed breeding is effective for sustainable agricultural production [5]. Second, in many cases, as compared to farmers with large operations, smallholder farmers are disadvantaged in accessing market; therefore, farming should be re-invented as an attractive local business opportunity for smallholder families [6]. Third, in the agri-food system, a few large international firms dominate the market share, and smallholders mainly trade on local short-supply chains [7]. Fourth, in recent years, production and consumption of rice in many parts of Asia has been steadily declining, whereas the opposite is true in Africa [8]. Fifth, due to economic development, the demand for maize to feed livestock has increased [9]. Six, the problem of food self-sufficiency does not apply only to developing countries. In Japan, due to the ageing population and the lack of young people willing to engage in agriculture, the self-sufficiency rate is low [10]. As a result, Japan imports a substantial amount of food from abroad (i.e. China) to cover their shortages [10].
Stabilisation of markets to protect suppliers from macroeconomic shocks is also important. For example, in 2020, the global economy was greatly affected by the spread of COVID-19. Because the main transmission route of COVID-19 is droplets from infected people, people stopped eating out. In Japan, food manufacturing and service businesses were severely damaged by people's restraints and many food manufacturers and restaurants went bankrupt [11]. This was mainly because in 21 This is because countries with small capital grow faster by capital investment than countries with large capital. However, on the other hand, we cannot ignore that investment in developing countries often involves higher risk than that in developing countries. Japan many restaurants are small and medium-sized enterprises that lack corporate strength to withstand severe economic adversities [12]. As a result, a significant amount of ingredients that were to be delivered to restaurants went left unsold. In addition, a sharp drop in the demand for restaurant food has spurred the overstock of food, leading to the collapse of the prices of some foods. If the situation does not improve, farmers will not be able to make sufficient profits and will be forced out of business. This will lead to a decline in food production capacity and will worsen food insecurity.
In addition, inflation and instability of agricultural prices will put pressure mainly on the world's low-income group of people and an increase in labour productivity will be necessary to overcome the problem [13]. The effect of extreme weather is also a major issue for cereal production. Weather extremes in the United States in 2012 caused a sharp increase in the world maize price, and the poor countries with high maize import dependency were the most seriously affected [14].
On the consumption side, the main challenge is to reduce food loss in the developed countries. One measure to correct such market failure is government intervention in the market to internalise negative externalities. Additionally, transferring food surpluses from developed countries to developing countries to combat food shortages may improve the situation. Food banks are one of the potential measures for recycling food within a country [15]. If such food banks could be internationalised, food security in developed countries may be ensured in a way that is close to our idea. However, there are many difficulties with this potential solution. For example, some foods do not last long and are not suitable for longdistance transportation. Moreover, as discussed before, due to regional variances in food culture, some foods may be hard to replenish. 22 Many countries also impose tariffs on imported products because large amounts of imported food may adversely affect local farmers' businesses.
Additionally, food loss also occurs in developing countries, where food security is not ensured. The main cause of food loss here is lack of capital for storage; therefore, as in the supply side discussion, increasing capital through investment and assistance is the effective measure 23 .
In this section, we examined measures to improve food security. Although the authors' proposed methods have the potential to improve food security, there are many challenges. In order to achieve these improvements, in addition to international efforts (i.e. international organisations and governments), changing the mindset of each economic entity (i.e. supplier and consumer) will be necessary.

Conclusion
Cereals are essential for human nutrition. However, the ever-increasing world population makes it difficult to maintain food security. It is necessary to consider various factors related to production and consumption of cereals to ensure food security. This study considers the food security issues for cereals in terms of both supply and consumption. 22 South Asia, where rice is the staple food, is one of the most densely populated regions and the second poorest region in the world [16]. In order to guarantee its food security, self-sufficiency is important [16]. 23 Installation of evaporative coolers can be considered as one of the realistic measures to reduce food loss and waste [15].
On the supply side, we proposed that cereal food security can be improved by increasing production capacity in developing countries. The optimal condition of our theoretical model specifies that, as compared to the case of developed countries, farmers in developing countries should more actively utilise capital resources. However, reality largely differs from theory. As shown by the data, the use of farming machinery in developing countries is lower than in developing countries. Additionally, our regression result confirms that the production of cereals (i.e. wheat, rice and maize) in G20 countries surpasses that in the least developed countries. Furthermore, our empirical analysis also confirms that the type of cereals produced vary significantly regarding the region.
On the consumption side, we emphasised the problem of food loss. Although some countries in the world suffer from food insecurity, others discard large amounts of food. Reducing such waste can contribute to global food security. However, the causes of food loss differ between developing and developed countries. In the case of the former, the lack of capital means that food rots before it is consumed. In the case of the latter, foods are oversupplied to the market due to excessive competition between firms.
In order to mitigate these problems, we proposed the following measures. In order to increase production in developing countries, their capital resources need to increased; this requires significant investment and assistance. On the other hand, one way to prevent food loss in developed countries is to mitigate excessive competition among firms through government market intervention. It may also be effective to transport surplus food from developed countries to developing countries to fill shortages. However, there are many challenges to implementing these measures. To truly improve the global food situation, in addition to international efforts (i.e. international organisations and governments), each of us needs to change our consciousness. Note: Table 2 is provided by the author based on the data extracted from FAOSTAT. Product and domestic supply of wheat, rice and maize are in 1,000 tonnes. Population is in 1,000 people. Table 2.
Production of cereals, number of tractors and population in G20 countries. Note: Table 3 is provided by the author based on the data extracted from FAOSTAT. Cereal cit is in 1,000 tonnes.   Note: This table shows estimation results for Eq. (11). P-values are presented in the parentheses. The estimation is conducted with robust standard errors. Superscripts ***, ** and * indicate statistical significance at 1%, 5% and 10%, respectively. Note: Table 5 is provided by the author based on the data extracted from FAOSTAT.