Summary of Dissertation Agricultural Economics: Attitudes toward risk and economic efficency in maize production of farmers in the Mekong delta

MINNISTRY OF EDUCATION AND TRAINING

CAN THO UNIVERSITY

LE VAN DE

ATTITUDES TOWARD RISK AND ECONOMIC EFFICENCY IN MAIZE PRODUCTION OF FARMERS IN THE MEKONG DELTA

SUMMARY OF DISSERTATION Major: Agricultural Economics Major code: 9 62 01 15

Can Tho, 2021

The research has been finished at Can Tho University, Can Tho City, Vietnam

Supervisors: Assoc. Prof. Pham Le Thong, PhD

Discussant 1:…………………………………………………….

Discussant 2:…………………………………………………….

The dissertation will be defended at the council of the school level at:

…………………………………………………………………………….

Citing of this the dissertation is available at following the libraries:

- Learning Resource Center-Can Tho University, Can Tho City.

- VietNam National Library, HCM City.

On: …............. hour………….date …………month………year……..

LIST OF PUBLISHED PAPERS RELATED TO DISSERTATION

1. Le Van De, Pham Le Thong, (2019). Technical efficiency of hybrid maize production in the Mekong Delta. Journal of Science Ho Chi Minh City Open University, 14(2), pp 72-85.

2. Le Van De, Pham Le Thong, (2019). Economic efficiency of maize producing households in the Mekong Delta. Economic studies, volume 4(491), pp 52- 62.

3. Le Van De, Pham Le Thong, (2020). Attitudes toward risk in maize production of farmers in the Mekong Delta. Journal of Economics and Development, 278 (08/2020), pp 83-91.

CHAPTER 1 INTRODUCTION

1.1 The statement of problem

1.1.1 Scientific relevance

Agricultural production is usually known as activities with a wide range of risks compared to other production activities, due to the impact of natural conditions; market volatility and social uncertainty. Facing several kinds of risk forces farmers to make decisions on production in an uncertain environment (Ellis, 1993). Farmers who are risk averse are often less willing or slow to apply innovations than others, although agricultural innovations can improve their productivity and income (Antle & Crissman, 1990, Ellis, 1993, Liu, 2013). In addition, farmers with a risk-averse attitude are more likely to invest resources in production that are below the economically optimal level, which will not be able to maximize profits. Therefore understanding the farmer’s attitudes toward risk is very important risks in order to understand the farmer's behavior, thereby planning production management strategy, technology transfer and building supportive policies supporting products in agriculture (Young, 1979).

toward risk enhances the farmer's attitude

This study attempts to measure risk attitudes of hybrid maize farmers and analyze the influence of farmers' risk attitudes on the economic efficiency in hybrid maize cultivation in the Mekong Delta. Based on the research results, the author has proposed solutions to the reduction of risk aversion and the enhancement of economic efficiency in production, contributing to improving the income of farm households. This paper contributes to literature and practice in respects. Firstly, we are the first to conduct empirical research using the method of Eckel & Grossman (2002) in order to measure attitudes towards risks of farmers in the Mekong Delta. We designed a game with real payoff in which farmers were asked to select risk options whose expected value linearly increased with risks due to increased variance to show their risk attitudes. This method can be extended to measure farmers' attitudes toward risks with different production activities in the region. Secondly, the analysis of factors the farmer's influencing understanding of the farmer's behavior. Thirdly, analyzing the effects of risk attitudes, household characteristics and production activities on economic efficiency in production to test the relationship of farmers' risk attitudes and optimal use of inputs in production. Finally, the study was conducted on hybrid maize farmers. Hybrid maize is encouraged to develop by the State to meet

domestic demand, and then, reduce dependence on imports. Nevertheless, hybrid maize production in the Mekong Delta involves many potential risks

1.1.2 Practical relevance

Maize is one of the important sources of raw materials in the industries of animal feed processing, food production and some other industries (fuel, pharmaceuticals, ...). Currently, the maize production area of Vietnam is about 1.12 million ha and the output is nearly 5.15 million tons/year, but the area and output tend to decrease, with an average rate of reduction of about 3%./year in terms of area and nearly 1.5% / year in output (GSO, 2019). In particular, the currently produced domestically only meets about 40% of domestic demand. Therefore, Vietnam imports with an average output of over 6-7 million tons / year, with a value of about $ 1.4 billion (average data for the period 2014-2018) and the import volume tends to is increasing due to the high demand for raw materials for processing animal feed (AMIS, 2018; FAO, 201. Therefore, the Government and the agricultural sector have policies to develop maize production to meet domestic demand and reduce dependence on imports. The Ministry of Agriculture and Rural Development has also issued a policy on converting to maize production on inefficient rice areas (Decision No. 3367 / QD-BNN, Ministry of Agriculture and Rural Development, issued on July 31, 2014). Accordingly, Vietnam will convert 236,000 hectares of inefficient rice land to hybrid maize cultivation in the period from 2014 to 2020, with 83,000 hectares in the Mekong Delta alone.

However, the production of hybrid maize in the Mekong Delta is also facing many difficulties and obstacles. First, the technical level of hybrid maize production of most farmers is still limited, the application of production mechanization is low. Secondly, production is very fragmented, small, so it is difficult to call for businesses to link to consuming products, and the price is very volatile. Third, the infrastructure and logistics system is inadequate, and there is no macro policy for the maize industry from planning production areas to consumption markets (Ho Cao Viet et al., 2015). Fourthly, the production efficiency achieved by the households has fluctuation, the number of farmers having losses is quite high (Ho Cao Viet et al., 2015). Therefore, the product cost is high, the product is less competitive. Which leads to many animal feed processing enterprises choosing imported products.

Therefore, the study of the attitude towards risk and the relationship of attitude to risk and economic efficiency in hybrid maize production of farmers in the Mekong Delta is very necessary. The research results will be the scientific

basis for proposing a number of solutions to reduce risk-averse attitude and improve economic efficiency, contributing to increase income for farmers.

1.2 Research objective

1.2.1 General objective

This study attempts to measure attitudes toward risk and analyze the influence of farmers' attitudes toward risk on the economic efficiency in hybrid maize cultivation offarm households in the Mekong Delta. Based on the research results, the author has proposed solutions to the reduction of risk aversion and the enhancement of economic efficiency in produoction, contributing to improving the income of farm households.

1.2.2 Specific objectives

(1) To investigate the current situation of maize production in the Mekong

Delta.

(2) To measure the attitudes toward risk of hybrid maize farmers in the

study area.

(3) To analyze the effects of farmers' attitudes toward risk on economic

efficiency in hybrid maize production of farmers in the area.

(4) To propose solutions to reducing the risk aversion of farmers and improving economic efficiency in production, and hence, increasing income for farmers.

1.3 Research Hypotheses

This thesis will test the following hypotheses:

Hypothesis 1: Farmer’s attitudes toward risks are influenced by their

socio-economic characteristics.

Hypothesis 2: The risk aversion causes farmers to use less inputs than their

optimal level in production.

Hypothesis 3: The economic efficiency of farm households is lower when

they are risk averse.

Hypothesis 4: The economic efficiency in hybrid maize cultivation

depends on the factors of socio-economic characteristics of the households.

1.4 Research subjects

The thesis is to measure the farm households' attitude toward risks and economic efficiency in hybrid maize production, as well as the influence of that

on economic efficiency in production. On that basis, the author proposes appropriate solutions, which are to reduce risk aversion attitude and improve the economic efficiency in production.

1.5 Scope of research area

The study area are in the Mekong Delta, specifically in three provinces: An Giang, Dong Thap and Tra Vinh because the provices have large production areas in the region. In addition, the selected provinces represent two different ecological zones. While Dong Thap and An Giang are located at the upper part of the Mekong River and Tra Vinh is at the lower part.

CHAPTER 2 THEORETICAL BACKGROUND AND RESEARCH METHOD

2.1 Theoretical background

2.1.1 Theoretical background and methods to measure attitude toward

risk

2.1.1.1 Overview of risk

Risk refers to a situation where probabilies can be attached to the occurrence of events which produce different outcomes of a decision making process (Ellis, 1993). Risk refers to the difference between received results and expected results in uncertain situations (Le Khuong Ninh, 2016).

In agricultural production, risks are pervasive and varied, so the farmers’ income is dependent on the risk. Musser and Patrick (2002), Hardaker et al (2004), Drollete (2009) and Aimin (2010) all stated that there are five types of risks, which are production risks, market risks, institutional risks, contractual risks and financial risks.

2.1.1.2 Overview of attitude toward risk

An individual's attitude toward risk is also considered as how individual acts in response to a risky situation (Dave et al., 2007; Pennings and Garcia, 2001; Weber and Milliman 1997). Walker and Jodha (1986) refers to attitude towards risk is expressed by the cautious attitude in making investment decisions, as well as the use of resources for individual production activities.

Because individual has different characteristics, depending on his psychological characteristics, social environment, and natural conditions. Hence, individual perceptions of risk are different (Slovic et al., 1982). Therefore, individual often has different attitudes toward risks such as aversion, liking, or indifference to risks.

When confronting with choices in the case of risk, individuals choose the one that produces maximum expected utility (von Neumann & Morgenstern, 1944). That is, the Decision Maker (DM) will compare his/her expected utility with the total utility generated from the results multiplied by the corresponding probability.

2.1.1.3 Measuring the attitude toward risk

The first studies used mean statistical indicators and random variance. However, many authors found that this method was not always a good measure of risk. Rothschild and Stiglitz (1970) developed a more general useful models and concepts of risk.

On reviewing previous studies related to measurement attitude toward risk, the author found that there are two main groups of methods used by many researchers: econometric model method and experimental method.

In regard of econometric method: the researchers used models of econometrics to estimate the parameters of the distribution of attitudes toward risk from production in general, with the assumption of maximizing expected utility. However, this approach is also considered to be limited because it can distort the attitude toward risk of individuals. They may reveal greater risk aversion than it is real to be under the resource-constrained condition they are facing (Eswaran and Kotwal, 1990; Singh et al., 1986; de Janvry et al., 1991; Sadoulet and de Janvry, 1995).

Regardinh experimental method, it derived from the research background of experimental psychologists. This method was originally formulated and reinforced by Luce and Suppes (1965). Charness and Gneezy (2012) believed that there were many experimental methods in measuring attitudes toward risk attitudes such as: Questionnaires, The Balloon modeling, The Gneezy and Potters Method, The Eckel and Grossman Method, and The Multi-price List Method.

On reviewing the methods of measuring attitude toward risk and analyzing the actual conditions in the research area, the author selected the Eckel and Grossman as the appropriate experimental method to measure the attitude toward risk of hybrid maize cultivation of farm households in the Mekong Delta. Charness (2013) believed that the Eckel and Grossman method was clearly and simply designed to explore an individual’s attitude toward risk. Dave et al (2010) also proved that this method was fairly reliable in measuring preferences to risk and less complicated than the others, especially for participants who have limited computing ability.

2.1.2 Economic efficiency and methods of measuring

2.1.2.1 Overview of economic efficiency

Farrell (1957) believed that production efficiency was constituted of three components: technical efficiency, distributional efficiency (or price efficiency) and economic efficiency

Technical efficiency is the ability to produce a given amount of output from the smallest amount of inputs or the ability to produce a maximum amount of output from a given amount of inputs, in accordance with a given technological level.

Allocative efficiency is the ability to select an optimal amount of input where the marginal revenue product of the final input unit is equal to the price of that input.

The economic or total efficiency is the technical efficiency multiplied with the allocative efficiency. EEi = TEi*AEi. In which: EEi, TEi and AEi are respectively the level of economic, technical and distributional efficiency of the i th producer.

2.1.2.2 Measuring economic efficiency

Estimating the economic efficiency in production is usually worked out using the two most common methods: The first one is Data Envelopment Analysis - DEA, which is non-parametric estimation method. The second one is parametric estimation method.

However, the parametric method has the advantage that it is possible to estimate the marginal effect of each input and exogenous factors on the output, which does not need to use the auxiliary regression model as for the non- parametric method (Chen et al., 2015). In particular, the random marginal method can separate random error from inefficient error. Therefore, this method does not consider all the distances from the production to the margin as being totally due to inefficiency, which helps to estimate the efficiency more accurately (Bezat, 2009). In addition, this method also allows statistical testing of hypotheses related to production structure and the level of inefficiency (Sharma et al., 1999). Coelli (1995) also argued that the random marginal method is encouraged to be used in efficiency studies in the field of agriculture, because measurement errors, lack of variables, and especially random factors such as weather, market, ... are common and have a great influence on agricultural production.

From all the above, the author found that using the parameter estimation method would be more suitable in estimating the efficiency in hybrid maize production.

2.1.3 Attitude towards risks, decision-making on the use of inputs and

production efficiency

2.1.3.1 Attitude towards risks and decision-making on the use of inputs

Picazo-Tadeo et al (2011) claimed that the input selection would be affected by the attitude toward risk of the production. In particular, risk attitudes did have an impact on efficient resource allocation (Wolgin, 1975). Kumbhakar (2002) identified that attitude towards risk had an important influence on the distribution of the use of inputs, as well as the supply of outputs. Many other studies also identified the nature of risk aversion of farmers influenced the use of inputs (Antle, 1987; Ghadim et al., 2005; Paulson and Babcock 2010; Dercon and Christiaensen 2012). Loehman and Nelson (1992) believed that optimal use of inputs will be influenced by the differences in attitude toward risk of individual.

Nmadu et al. (2012) stated that farmers with risk aversion would use such inputs as lower amounts of fertilizers and seeds than the required level to get expected maximum profit. They would exchange less inputs for higher income to lower the income gap, which is to maximize the usefulness of profit in risky conditions (Sandmo, 1971; Anderson và cộng sự, 1977).

Much effort has been made to examine the effects of the farmers’ risk perferences on the economically optimal use of inputs. Ellis (1993) built a theoretical model representing the relationship between the use of input factors in the uncertain conditions with the risk attitudes, as shown in Figure 2.1

Figure 2.1 shows three response curves of output (in value) to a single variable input, say, units of nitrogen fertilizer. TVP1 represents the total value product response to the input use in a “good weather”. In contrast, TVP2 represents the total value product with respect to the input use in “bad weather”. E(TVP) represents the expected value of the product corresponding to the farmers' subjective probabilities of a “good” and “bad” season. P1 and P2 are the subjective probabilities associated with the occurrence of these events. TVP1 and TVP2 are described as the outcomes of events or states of nature. The TC curve represents the total cost of input use, linearly increasing with the input use as input price is constant.

TVP

TVP1

TFC

E (TVP)

TVP2

Figure 2.1 Production decisions under risk Source: Ellis (1993)

The input use at X2 represents the allocative efficiency on TVP2 as the bad weather occurs. The individual who decides to use this input level obtains a profit of ce if the “good” season occurs. On the contrary, if the season is “bad”, the farmer obtains a small profit of de. Whichever season may occur, the individual is able to make profit. The individual who makes the choice at X2 is described as risk-averse because this person prefers the safety of acting as if the worst possible outcome will happen, even though the probability, in her/his opinion, of the bad season is small.

On the contrary, a individual with a risk-loving attitude is likely to use the input at X1 that corresponds to the allocative efficiency on TVP1. If the good season occurs, he/she gets a big profit of ab. If the season is bad, the individual will suffer a substantial loss of bj. This choice shows that the individual prefers a chance at the largest possible profit, even though he/she knows that it is uncertain. Otherwise, the risk-neutral individual will decide to use the input at XE, that is consistent with the allocative efficiency with a balanced assessment of the average outcome of “good” and “bad” seasons.

This theoretical framework describes the risk in terms of the “income variance” approach. Under uncertainty, income from production varies with possible events. Risk-averse farmers are likely to use the input levels that ensure the safety of their income. Input uses with safe outcomes are typically less than those with maximum expected outcomes and then, well below the ones that yield the highest possible outcome.

2.1.3.2 Attitude towards risks and other decisions in production

In addition to studying the influence of attitude toward risk on the use of inputs, many other researchers have also studied the impact of it on other decisions in production, like:

Oparinde et al. (2018) stated that farmers afraid of risk were less willing to adopt technology investments on a crop with a high risk of failure (which means high risk), even though they knew that the result might be higher than that using traditional technology. Heltberg and Tarp (2002) believed that risk also influences farmers' choice of market. Gine and Yang (2009) stated that risk attitudes affected farmers' credit and insurance needs. Wale and Yalew (2007) also found that they had an impact on farmers' choice of crop varieties.

With the theoretical model of the relationship of risk attitudes and the optimal use of inputs by Ellis (1993); particularly, with a review on experimental studies carried out on the influence of attitude toward risk on the use of inputs and other behaviors having indirect impact on production efficiency, the author have a basis for hypothesis about the effect of attitude toward risk on economic efficiency in production.

Through empirical studies and theoretical models on the relationship of risk attitudes and optimal input use and other behaviors in production. This helps the author have a basis for hypothesizing about the effects of risky attitudes on economic efficiency in production.

2.1.4 Test optimal use of inputs in production

On the basis of Ellis' theoretical model (1993) and related studies on the relationship of risk attitudes and input decisions used in production. Verification of optimal input usage can be performed. Accordingly, the optimal input level determined at the marginal productivity of each input is equal to the exchange rate between the inputs of each of those inputs and the output prices.The Cobb- From the function of the Cobb-Douglas production function logarized on two sides, with the following form:

(2.1)

equivalent to

Partial derivative of an input, for example X1, gets:

Becomes (2.2)

). (2.2) can be written as: is the marginal productivity of the input X1 (

(2.3)

, can be The conditions for maximizing profit on input X1 are

written as:

=1 (2.4)

Substituting (2.3) into (2.4), this condition becomes:

(2.5)

Ellis (1993) referred to as the allocative efficiency ratio

(k) for a single input the conditions for maximizing profit on input is k=1. If input X1 has a ratio k > 1 household uses the input X1 which is lower than the optimal input; however, if input X1 has a ratio k <1, household uses the input X1 which is more than the optimal input. Thus, the test of farmer’s optimal input can be performed based on the comparison of k and 1.

2.1.5 Theoretical model of the relationship between attitude towards risk

and economic inefficiency

On the basis of theory and empirical studies on the relationship between attitude toward risk and optimal input level, the author establishes a theoretical model of the relationship between attitude toward risk and economic inefficiency.

is the output price,

is the amount of input with

Consider a profitable production activity , with

represents the fixed factor in the production function.

Output is the form of production function in which the price being ,

is the two-component composite error (including and ).

The profit function can be written as:

(2.6)

The utility maximization model is used for risk research, because it is assumed that the individual maximizes the expected utility with the received value.

Therefore, the expected utility function corresponding to the above return

function has the form:

(2.7)

Suppose r represents an individual's aversion to risk and is assumed . This means that the more risk-averse attitude becomes, the greater

the profit earned will decrease due to the fear that the risk of using inputs is lower than the economically optimal level. (Ellis, 1993). So, function according to and it has features , .

This individual's utility l function, then, is rewritten to:

(2.8)

is the personal profit received under the influence of the

where,

individual's aversion to risk attitudes.

or:

Under terms of profit maximization.

(2.9)

is the normalized profit,

is the normalized

where:

represents the fixed element.

input price and

Then, the individual's normalized utility function, has form:

(2.10)

von Neumann và Morgenstern (1944) has demonstrated that an individual

wants to maximize the usefulness of expectations

(2.11)

Risk aversion farmer will be have utility with an certainty equal (U(CE))

equal the expected utility in the risk condition, means:

= (2.12)

Return to the stochastc frontier profit function with the following form:

(2.13)

is the composite error (model residue) that can take some form,

depending on the analytical framework of a study (Jaenicke and Larson, 2001).

(2.14)

where, is the symmetry error, representing the effects of the stochastic

factors, having a normal distribution with an expected value is 0 and variance is

. is the one-tailed error, representing the ineffectiveness

calculated from the difference between the real profit ( ) and the maximum

profit possible given stochastc frontier profit function above. If > 0,

household production activities are below the frontier and the middle difference

and is the ineffective part (Coelli ,et al, 2005).

To determine the factors affecting economic inefficiency, is regression

with its explanatory factors.

The economic inefficiency function has a form.

(2.15)

where: represents the factors affecting the economic inefficiency of the

group of factors of characteristics: age; gender; education; experience; ...) and represents the group of factors that characterize the attitude towards risk

(including: risk-loving, average risk neutral and risk aversion) of a farmer. Therefore, the variable of attitude toward risk of farmers is considered as exogenous variable in the inefficiency function model. (Jirgi, 2013; Yeager và

is the error value

Langemeier, 2017; Haneishi và cộng sự, 2014).

representing factors outside the model or random noise.

2.3 Methods research

2.3.1 The method of data collection

2.3.1.1 Methods of collecting secondary da

Secondary data used in the thesis are collected from statistical yearbooks, reports of professional management agencies, etc. In addition, the thesis also uses data from relevant previous studies which were publicized.

2.3.1.2 Methods of collecting primary data

In the Mekong Delta, hybrid maize production is not widely distributed around the region, but concentrated in few areas. For this reason, the authors conducted the survey at three provinces with the leading production areas, namely An Giang, Dong Thap và Tra Vinh. The selected provinces represent two different ecological zones. While Dong Thap and An Giang are located at the upper part of the Mekong River and Tra Vinh is at the lower part. In each province, the authors selected districts with large and concentrated production areas. Then, 2-3 communes in each district were randomly selected for the survey. The surveyed households were randomly selected from the list of hybrid maize households provided by the Commune People's Committee. The survey collected information from 256 households in which the number of households in An Giang was 122 households, Dong Thap was 71 and in Tra Vinh was 63 households.

2.3.2 Methods of data analysis

2.3.2.1 Methods of measuring attitude toward risk

Charness and Gneezy (2012) believe that there are many experimental methods in measuring attitude towards risk such as Questionnaire method; the Steam ball modeling method; Method Gneezy and Potters; The Eckel and Grossman method and the multi-price list method. However, Charness (2013) argues that the Eckel and Grossman methods are fairly simple and clear professional methods for exploring individual attitudes towards risk. Dave et al. (2010) also demonstrate that this method is quite reliable in measuring attitudes to risk and is less complex than the other methods, particularly suitable for respondents who has low computational ability. Therefore, the authors apply the experimental method developed by the Eckel and Grossman (2002) to measure the attitude towards risk of hybrid maize farmers in the Mekong Delta.

To measure the risk attitude of each farmer, the researchers selected the respondents who took decisive role of and directly participated in the hybrid maize production for the survey. Investigating farmers' attitudes towards risks with experimental games was conducted on an individual basis. Each respondent was asked to play 3 lottery selection games (with the nature of chance). Of which, games 1 and 3 were based on hypothetical payoffs. For game 2, the respondents were given real payoffs according to their choice and the lottery result (in the form of tossing a coin). The combination of 3 games was aimed to check the robustness of the resulting player's attitude towards risk, by comparing results among games. In addition, to ensure the players understand the nature of the games, the interviewers would first explain the players the

nature of the games. The result of a player's final choice in each game was only recognized when the player had a reasonable interpretation of the reason for their choices (Barry, 2014). The organization of the games is presented in details below.

Game No.1:

The purpose of the game is to preliminarily identify the participants’ attitude towards risk. In addition, this game is also intended to test the player's understanding of the game and to detect the ones who make inconsiderate decision, affecting the elicit of attitude towards risk. Below is how to proceed the game:

Farm household heads are asked to take turn to answer all 7 questions. Then, they will have the right to choose either the hypothetical reward "A" or "B" corresponding to the questions. If choosing reward "A", the participant will receive the value of the award, which is certain (without risk). In contrast, the reward will be determined through the results of a coin toss (tail or head). If the head appears, the participant will receive 100,000 Vietnamese dongs; otherwise, there is no reward. This option is risky but may let them have a chance to get a higher value of reward. The game is continued by this way until the participant decides to choose the reward "B" at which the switch point is determined at that question and the game will stop.

The interviewer observes and identifies the one making the decision to change from option “A” to “B” in that question and marks “x” in the transfer point column. Then, they will move on to Game No. 2. Table 2.1 shows the questions and hypothetical rewards in Game No. 1.

Unit: VND thousand

Table 2.1 Game 1 to preliminarily determine risk attitude

Reward B Probability: 50 – 50

Questions

Reward A

Switch point

100 90

Tail (Low payoff) 0 0

1* 2

Head (High payoff) 100 100 100

100

Questions 1*: How old are you? Questions 2: what grade are you in? Questions 3: How many years of experience do you have production? Questions 4: How much is the production area? Questions 5: How much is the productivity? Questions 6: Who do you sell products for? Questions 7: What is the selling price?

25 15 05

100 100 100

0 0 0

5 6 7

Source: Authors’ design

Games 2:

Game 2 is the most important game since it is used to determine a player's risk attitude. It is the lottery game with real payoff according to the choice of the player. The player's attitude towards risk is manifested in their decision- making in the game that is designed to have different levels of risk. The way the game is proceeded is presented as follows: Participants are asked to choose one of the six options, from A to F in Table 2.2 Players will receive real payoff based on tossing a coin with a probability of 50% for the head or the tail, equally.

The base reward value is 50,000 VND, which is 2.2 times of the hourly wage of hired labor in the rural region. This design aims at attracting the attention and thinking of farmers in making decisions in the game. In addition, the game is also designed to prevent players from making losses (the smallest reward is zero), in order to avoid farmers thinking of choosing a safe option.

Table 2.2 Games 2 to determine risk attitude Unit: VND thousand

Options

Tail (High payoff)

Expected return

Standard deviation

50.0

0.0

Head (Low payoff) 50 45 35 20 10 00

50 60 90 125 140 150

52.5 62.5 72.5 75.0 75.0

7.5 27.5 52.5 65.0 75.0

A B C D E F Source: Authors’ design

Features of the game:

(2.6)

Option A is the safest option whose standard deviation is 0. Options from B to E are designed so that expected values linearly increase with risk, expressed as the increasing standard deviation. Option F is designed to have the same expected value as option E but with a higher standard deviation.

The risk-averse people tend to choose options from A to D, the risk-neutral people tend to choose option E and the risk-taking people tend to choose option F. The coefficient of risk attitude for each option is determined by the constant partial risk aversion coefficient (CPRA):

in which r is the coefficient of risk aversion and x is the income of certainty equivalent to expectation. The individual's risk attitude coefficient is determined by solving the equation for the expected utility of expectation between two adjacent options.

Game No. 3:

Game 3 is designed to consider whether a player's risk attitude changes with increased reward value. However, the game is proceeded with hypothetical pay-offs. Outcomes of Game 3 are used to compare with those of Game 2, to reinforce confidence about the eliciting method in the Game 2. The organization of the game is similar to game 2. The options in this game are shown in Table 2.3

Table 2.3 Games 3 to determine risk attitude

Options

Head (Low payoff) 100 80 70 30 10 00

Tail (High payoff) 100 120 140 200 240 250

Expected return 100 100 105 115 125 125

Unit: VND thousand Standard deviation 0.00 28.28 49.50 120.21 162.63 176.78

A B C D E F Source: Authors’ design

2.3.2.2 Testing the use of inputs in comparison with the optimal level

Testing the farmer's optimal input is based on the profit maximizing condition of using inputs. Accordingly, the marginal productivity of each input factor is equal to the rate between the inputs of each input factor and the output price or the ratio of allocative efficiency k = 1. To perform this test, first estimate the production function in the form of a logarized Cobb-Douglas function, the experimental production function of the form:

(2.7)

The variables in the model are defined as follows: Qi: total output to the area (ha) of hydrid maize. This dependent variable is measured in kgs.

: Intercept of the production function.

: are parameters of the model.

Xki: are the amount of inputs per total production area, unit (kg)

Rji: are dummy variables represent "attitudes to risk for farmers"

Xki*Rji: are the interaction variables of the inputs and farmer’s attitudes towards risks.

From the results of production function estimation (2.7), the coefficients

estimation are determined . On the theoretical basis of profit

maximization, we have: = 1, it is the ratio of the

allocative eficiency of the k-th input. If k > 1, household uses the input Xk which is lower than the optimal input; however, if k < 1, household uses the input Xk which is more than the optimal input.

2.3.2.3 Analyzing factors that influence attitude towards risk

Ordered logistic regression model to analyze factors influencing farmers'

risk attitudes, following form:

(2.8)

where:

The dependent variable Y takes the unit value if the farmer is Risk- loving; 2 if the farmer is Risk neutral and 3 if the farmer is Risk aversion. X is a vector of independent variables explaining the farmer’s attitude toward risks including, farmer’s characteristics (gender and education of homeowner; number of labour in household, etc.), characterisitics of the production activities which affect to farmer’s attitude toward risks.  is a vector of unkown parameters of the model. (k = 0, ..., 6).

2.3.2.4 Estimating economic efficiency and factors that influence

To estimate economic efficiency and determine the factors affecting economic efficiency, the study performed a one-stage estimation from the simultaneous estimation of a Stochastic frontier profit function as in equation (2.9) and the economic inefficiency function as in equation (2.10), by the method of maximum likelihood estimation (MLE).

The Stochastic frontier profit function is as follows:

(2.9)

where:

: is UOP profit (variable profits) in dongs defined as gross revenues minus

*: are the normalized of each inputs, calculated by dividing each input price

variable costs of production all divided by the output price.

Pji

divided by the unit price output.

Zi: total fixed inputs per ha, defined as the total fixed cost, unit of calculation is

VND/1.000m2/crop.

). ei: is a error term which is composed of two components (

0, j, i: là các tham số cần ước lượng của mô hình (2.9) (j =1..7, i=1).

The economic inefficiency function has the following form:

(2.10)

where:

IEEi: is the economic inefficiency of the ith farmer households.

Wki: are variables representing the socio-economic characteristics of the ith

farmer and characteristics of the farming model.

Rji: are risk attitude classification of farmers.

CHAPTER 3 OVERVIEW OF STUDY AREAS

3.1 Maize production in Vietnam and Mekong Delta

3.2.1 Maize production in Vietnam

3.2.1.1 Area, yield, and productivity

a h

2000

n ì h g N

6000 Năng suất bắp đạt được của Việt Nam vẫn luôn thấp hơn năng suất bắp 4000 bình quân của thế giới. Đặc biệt, khoảng cách chênh lệch này không có xu hướng thu hẹp trong suốt giai đoạn 2014-2018 (Hình 3.1).

a h a / h n / â s T n o T

d n a s n o u h T

a h a h

During the period 2014 - 2018, the maize production area throughout the country tended to decrease, especially there has been a sharp decline from 2017 onwards (General Statistics Office, 2018). In particular, the Mekong Delta region is one of the regions with higher production area decrease than the national average rate (3.1% per year), with an average rate of reduction of 3.37% per year. Because the speed of productivity growth is not high, the yield has also tended to decrease.

2014

2015

2016

2017

2018

Quantity

Areas

Diện tích bắp Việt Nam Năng suất bắp Việt Nam

Sản lượng bắp Việt Nam Năng suất bắp thế giới

3.2.1.2 Tình hình cung - cầu và giá cả bắp ở Việt Nam

Do ngành chăn nuôi ngày càng tăng trưởng và phát triển, điều này sẽ kéo theo sự gia tăng nhu cầu sử dụng nguyên liệu phục vụ ngành sản xuất thức ăn chăn nuôi. Figure 3.1 Acreage, Yield and Productivity of maize in the world Source: Data collected from FAOSAT and AMIS, in 2019

The productivity of maize in Vietnam is always lower than the world’s average productivity. In particular, this gap did not tend to narrow during the period 2014 - 2018 (Figure 3.1). The reason is that most of the maize cultivation areas in Vietnam are distributed in the ethnic minority areas, where the technical level of production is still limited. In addition, the majority of maize cultivation is concentrated in mountainous and highland areas, where soil conditions are less fertile and often faced with a shortage of water for production. In these places, maize is seen as a substitute crop for other crops that cannot adapt to the natural conditions here so the yield is always low. Meanwhile, in the major maize-producing countries in the world, due to the large production scale, the degree of mechanical application and scientific and technical advances are high, so productivity is higher as well as production costs. and lower product costs. This is also one of the reasons affecting the competitive advantage of Vietnam's

maize products at present and in the future compared to imported maize products.

3.2.1.2 Supply - Demand and price of maize in Vietnam

As the livestock industry is growing and developing, this leads to an

increase in the demand for raw materials for the animal feed industry.

Unit: million tons

Table 3.1 Supply and demand of maize in Vietnam, period 2014 - 2018

Supply

Demand

Season

Production Import

Export Resere

Other consumption

Total

Previous reserve

Animal feed processing

2013/14

0.65

5.19

3.60

7.02

1.79

0.04

0.60

9.45

0.60

5.20

5.66

8.00

2.30

0.17

1.00

11.46

1.00

5.29

7.69

9.29

3.12

0.08

1.48

13.97

1.48

5.25

8.74

10.27

3.42

0.13

1.66

15.47

1.66

5.13

10.95

3.47

0.09

1.25

8.97 15.76

1.25

4.91

11.37

3.57

0.42

0.61

9.45 2014/15 11.46 2015/16 13.97 2016/17 15.47 2017/18 15.76 2018/19 17.97

9.82 15.97

Sources: Agricultural Market Information System (AMIS), 2019

According to the reported data shown in Table 3.1, the average domestic maize supply is about 6.27 million tons per year. Meanwhile, the average total demand in the same period is 13.68 million tons per year, so the supply is not enough to meet the demand. Therefore, our country has to import an average of 7.41 million tons per year. In particular, import demand has tended to increase strongly in recent years, with an average growth rate of 23.77% per year.

Although the domestic production capacity does not meet the domestic demand, the domestic maize products are difficult to consume at times and the price is very volatile. Most businesses still choose imported maize as the main material for production activities. The main reason is that the cost of maize products in the country is always high, but not so qualified as imported one. This result shows that the maize industry is encountering great challenges from the lack of competitive advantages with imported maize.

3.2.2 Maize production in the Mekong Delta

3.2.2.1 Productivity, area, and yield

The Mekong Delta is the region with high productivity, second to the Southeast only. During the period 2014 - 2018, the average yield reached 5.74 tons per hectare, which was 1.25 times higher than the national average maize yield (4.57 tons per hectare). This is because the Mekong Delta has natural conditions which are quite suitable for maize production.

However, domestic maize farming has been put at a disadvantage when imported maize has competitive prices and quality. This makes the domestic price of maize lower and volatile, farmer’s income is becoming decreased, even at a loss. Therefore, many farmers have switched to cultivating other crops.

The area of maize production in the region tends to decrease over the years, with the average rate of reduction of 3.37% / year, higher than the average reduction rate of the whole country (3.1%). Regional productivity also tends to decrease, with an average decrease of 4.5% per year in the period 2014 - 2018.

CHAPTER 4 FARMER’S ATTITUDES TOWARD RISK AND ECONOMIC EFFICIENCY ECONOMICS OF HYBRID MAIZE PRODUCTION

4.1 Characteristics of farm households growing hybrid maize in the study area

4.1.1 Demographic characteristics

Results of the survey of 256 hybrid maize farmers in the research area showed that up to 90.23% of households have men as head of households. Households who are from the Khmer ethnic group account for 23.44%. Regarding the education level of the household head, most of them has primary education (48.05%), which is also a common characteristic of the population in the rural and agricultural areas of the Mekong Delta. This has an impact on the farmers' ability to absorb technical skills in production.

The average number of people is 4.34 people per household, the average number of labors in each household is nearly 03 people, accounting for 61% of the total number of people in the household. The number of labors in the household participating in the production of hybrid maize accounts for about 83.02% of the labor force in the farm households.

4.1.2 Characteristics of farmers

- Areas - Total income of household - Incom from hydrid maize production - Farming experience - Farmer’s participate in local organizations and unions - Farmer’s participated in the training

Table 4.1 Descriptive statistics characteristics of maize farmers Characteristics

Unit ha/household million dong/year million dong/year year % %

Mean 0.61 73.26 45.49 14,00 11.72 46.09

Source: Survey results in the study area, 2018

The average area of hybrid maize production per farmer in the province is 6,120m2, accounting for nearly 73% of their total area of agricultural land. Hybrid maize production currently contributes averagely 62.09% to the total annual income of farm households. The percentage of hybrid maize production households participating in organizations and unions in the area is about 11.72%. In particular, only about 46.09% of farmers have participated in the

training. The number of households borrowing capital for production accounts for 15.63%.

4.1.3 Other production characteristics and conditions

Most of the farm households (62.11%) focus on the form of crop rotation with different types of vegetables. Mechanical application in maize production is still very limited due to the small and fragmented production area. All farmers sell their products through traders, mainly in the form of dried seeds.

Actually, farm households doing hybrid maize farming are now facing lots of risks, which can be divided into groups such as: risks due to natural conditions, markets, finance, institutions and individuals.

Table 4.2 Risks in household production

Types of risks

Amounts (farmers) 183 51 Ratio (%) 71.48 19.92

Source: Survey results in the study area, 2018

Output price Responding ability of family and hired labores Disease on tree Others 127 49 49.61 19.14

The survey results in Table 4.2 shows that, most of the farmers perceive themselves to be exposed to the risk of fluctuating and decreasing output prices. This is also the reason why the profit is lower and lower, and even many of them make a loss. This leads to about 21.11% of farmers to shrink their farming area and 16.08% to switch to other crops.

4.1.4 Financial efficiency in production

The research results show that the average income of households in the study area is 2,025.47 thousand VND / 1,000m2. With the average amount of labor in farming households investing in production activities is 5.39 working days/1,000m2. If the average labor cost in the localities is 180 thousand VND/working day, the average family labor cost is 970.20 thousand VND/ 1,000m2. Therefore, the average profit is 1,055.27 thousand dong/1,000m2.

Unit

Farm's day labor costs are not included

Farm's day labor costs are included -

3,290.46

1,055.27 1.32 0.62 0.32 - 0.24 -

Revenue Cost Income Profit Revenue /Cost Income/Cost Profit /Cost Income / Revenue Profit / Revenue Income /farmer's day labor Profit /farmer's day labor

4,345.73 2,287.78 2,025.47 - 1.87 0.87 - 0.47 - 375.78 -

195.78

thousand dong/1.000m2 thousand dong/1.000m2 thousand dong/1.000m2 thousand dong/1.000m2 Lần Lần Lần Lần Lần thousand dong/day thousand dong/day Source: Survey results in the study area, 2018

Table 4.3 Financial results of hybrid maize production

4.2 Analyzing farmer’s attitudes toward risk

4.2.1 Distribution of farmer’s attitudes toward risk

Farm households' choice through Game No. 1 showed that most of them decided to choose at the transfer sites with safe characteristics or low level of risk. This initially shows that most of the farmers involved in hybrid maize production in the area have a common feature of being quite risk aversion.

Switch point 1* 2 3 4 5 6 7

Table 4.4 The results of games 1

Frequency 12 21 14 16 40 60 105

Proportion (%) 4,69 8,20 5,47 6,25 15,63 23,44 41,02

Source: Survey results in the study area, 2018

Results of the farm households participating in the experimental Game No. 2 are shown in Table 4.5. This is an important experimental game, which is used to determine the coefficient of attitude towards risk of each farmer.

Table 4.5 Risk coefficients and classification of risk attitudes by game 2

Head (Low payoff)

Choice

Risk coefficient

Tail (High payoff)

A B C D E F

50 45 35 20 10 00

50 60 90 125 140 150

Risk attitude classification Extremely averse Severely averse Intermediate Moderate Slight to Neutral Neutral to loving

>10 10 - 1,00 1,00 - 0,62 0,62 - 0,18 0,18 - 0 <0

Source: Survey results in the study area, 2018

Based on the results of Game No. 2 and useful functions CPRA, the risk

factor of each household in the study is defined, as shown in Table 4.5.

The results of the study through Game No. 2 also showed that most of the farmers involved in hybrid maize production in the study area had an attitude of fear of risk. However, the results also show that the number of households that are from less afraid to medium and risk-averse also accounts for a significant proportion, nearly 12% of the total number of households. This result shows that there will be more households willing to take risks to participate in new investment options with higher profit and corresponding risks.

In addition, the author also considers whether the attitude towards the risk

changes when the payment value changes (increases).

Games 2

Games 3

Choice

Table 4.6 The results of games 2 and 3

Risk attitude classification

Frequency

Extremely averse Severely averse Intermediate

A B C D Moderate E F

Slight to Neutral Neutral to loving

Proportion (%) 46,48 21,88 13,28 6,25 2,34 9,77

119 56 34 16 06 25

Proportion (%) 48,05 23,44 8,98 5,08 3,13 11,33

123 60 23 13 08 29

Source: Calculated from survey data, 2018

Although the experimental games were organized differently, the results of the farmer households’ distribution of attitudes to risk were almost similar among games 1, 2 and 3. These help reinforce the reliability for the values of the risk factors that have just been calculated through the experimental method in the second game.

4.2.2 Use the inputs and attitude towards risk

Working on the case of the whole area, the inputs: seed, potassium fertilizer and household labor all had a distributional efficiency ratio (k) greater than 1. This means that these factors used by farmers is even lower than the optimal level. Meanwhile, the remaining inputs such as nitrogen fertilizer, phosphate fertilizer, agricultural medicine and hired labor have a distribution ratio (k) of less than 1, which means that these factors are households use more than the optimal level. Although the behavior of using inputs is different, in general, the amount of inputs used is not at the optimal level (because only k> 1 or k <1).

Table 4.7 Allocative efficiency coefficient (k) of the inputs

Allocative efficiency coefficient (k)

Difference

(1)-(2)

(2)-(3)

(1)-(3)

Coeffic -ient (k) on sample

Risk- loving (1) 9.30 0.48 0.44 3.07 0.00 0.75 1.54

Risk neutral (2) 13.95 0.85 0.38 3.08 0.00 0.74 1.01

Risk aversion (3) 11.53 0.61 0.48 3.24 0.00 0.85 1.26

-4.65ns -0.37ns 0.06ns -0.01ns 0.00ns 0.02ns 0.53**

2.42s 0.24ns -0.10ns -0.16ns 0.00ns -0.11ns -0.25*

-2.24ns -0.13* -0.04ns -0.17ns 0.00ns -0.09s 0.27ns

11.74*** Seed Nitrogen fertilizers 0.64*** phosphate fertilizer 0.46*** Potassium fertilizer 3.19*** 0.00*** Pesticides 0.81*** Labor hired Farmer’s day labor 1.25***

Source: Survey results in the study area, 2018 Note: *** ,** , *, và ns indicate the level of significance at 1%, 5% and 10%, respectively

Inputs

When considering the behavior of using the input factors by different groups of households in terms of attitudes to risks, the input factors: seed, potassium fertilizer and family labor all have the coefficient k > 1. Most of the value of the distributional efficiency ratio (k) of these inputs did not have a statistically significant difference among groups of households with different risk attitudes. In particular, only the family labor factor of the group of households with the attitude "Risk neutral" has the value of the ratio k = 1. In addition, the group of households with the attitude of " Risk neutral " has a lower value of the distribution efficiency ratio (k) compared to the two groups of households with the "Risk-loving" and "Risk aversion".

The remaining inputs such as nitrogen fertilizer, phosphorus, agricultural medicine and hired labor all have a coefficient k <1 with all three groups of households having different risk attitudes. The value of k-ratio of most of these inputs also did not have a statistically significant difference between groups of households with different risk attitudes. In which, only the value of the ratio k

of the nitrogen fertilizer factor has a difference, statistically significant between the group of households with the attitude of "Risk-loving" and "Risk aversion".

Generally, the farmer's use of inputs in production does not depend much on the farmer's attitude towards risk. This shows that, farmers choose inputs mainly based on experience and have little adjustment to changes in prices, making it difficult to maximize profits in using inputs. On the other hand, prices are often variable, which is a factor that farmers cannot control.

4.2.3 Analyzing factors that influence attitude towards risk

The estimated results of the Ordered logit model to analyze factors affecting the farm attitude towards risk to risk are presented in Table 4.8 shows that the estimated coefficients of the variables: Education level of the household head; Farming experience; Farmer’s participated in the training; Farmer’s participate in local organizations and unions and Income off the farm income (other than income from maize production) is statistically significant at the 10% level of local mass organizations.

Table 4.8 Estimated results of the Ordered logit regression model

Variable

Estimated coefficient -0.017 ns -0.157 *** -0.078 ns -0.090 ns

Z value -1.17 -2.75 -0.59 -0.76

-0.979 **

-2.22

Age of the household head Education level of the household head Numbers of people in the household Numbers of labour participated hydrid maize production Farmer’s participate in local organizations and unions Farming experience logarit Farmer’s participated in the training logarithm of total average income per people Income off the farm Disease outbreak An Giang (1= An Giang, 0= Tra Vinh, Dong Thap) Tra Vinh (1= Tra Vinh, 0= An Giang, Dong Thap)

-0.075 *** 0.115 ns -1.074 *** -0.250 ns -0.846 *** -0.125 ns 0.248 ns -0.650 ns

-3.72 0.39 -3.33 -1.07 -2.76 -0.36 0.69 -1.52

256 76.28 0.000

(13)

Number of observations LR χ2 Pr > Cut 1 Cut 2

-7.222 -5.731

Source: Survey results in the study area, 2018

Note: *** ,** , *, và ns indicate the level of significance at 1%, 5% and 10%, respectively.

4.3 Economic efficiency in production

The economic efficiency in prodoction is measured from the simultaneous estimation of a Stochastic frontier profit function as in equation (2.9) and the economic inefficiency function as in equation (2.10), by the method of maximum likelihood estimation (MLE). Estimated results are shown in Table 4.9

Table 4.9 Estimated results of stochastic frontier profit and the economic inefficiency function

Variable

Estimated coefficient

Z value

Function profit stochastic frontier

logarithm normalized nitrogen fertilizer price logarithm normalized phosphate fertilizer price logarithm normalized potassium fertilizer price logarithm normalized pesticides price logarithm normalized seed price logarithm normalized hired labor price logarithm normalized fuel price logarithm total value fixed inputs Constant

0.218 ns -1.415 ** 0.202 ns 0.004 ns -0.017 ns 0.173 ns -0.160 *** -0.109 ** 9.564 ***

0.51 -2.26 0.55 0.18 -0.39 1.26 -2.99 -2.47 11.76

Economic inefficiency

-6.813 *** -6.017 *** -0.148 * 0.270 * -0.001 ns -0.868 * -0.271 ns

-2.97 -3.15 -1.65 1.70 -0.02 -1.81 -0.57

1.907 * 1,186 ns -1.787 *

1.85 1,20 -1.90

1,967 ***

1,50

An Giang (1= An Giang; 0= Dong Thap, Tra Vinh) Dong Thap (1= Dong Thap; 0= An Giang, Tra Vinh) Education of head of households (Years of schooling) Numbers of labours in the household (labours) Farming experience (Years) logarithm areas (1.000m2) Farmer’s participated in the training (1=Có tham gia; 0=Không tham gia) Risk-loving (1= Risk-loving, 0= Risk neutral, Risk aversion) Risk neutral (1= Risk neutral, 0= Risk-loving, Risk aversion) Proportion of income from hydrid maize production /Toatal income of household (%) Constant Number of observations

246 29.43

Wald

(8)

0.000

Pr >

6.046

1.413

0.039

1.452

0.973

Source: Survey results in the study area, 2018 Note: *** ,** , *, và ns indicate the level of significance at 1%, 5% and 10%, respectively.

4.4 Attitudes toward risk and economic efficiency in maize production of farmers

The estimated results in Table 4.9 shows that determinants of efficiency are production area, education level of the household head, number of labours in the household, cultivated area, ratio of maize income to total household income. In particular, the farmers' attitude towards risks also affects the economic efficiency of households, the estimated coefficient of the attitude variable "Risk aversion" is statistically significant at the 10% significance level and has positive value. This shows that the group of households with the attitude of "Risk aversion" has a higher level of economic inefficiency compared to the group "Risk-loving", that means the economic efficiency achieved in production of group "Risk aversion" is lower compared to the group of households with "Risk-loving" attitude, nearly 1.91%, under the same conditions of other factors.

Difference in economic efficiency (%)

(1)-(2)

(2)-(3)

(1)-(3)

Table 4.10 Economic efficiency according to differences in attitude towards risk

Economic efficiency (%) by attitude towards risk Risk neutral (2) 81.99 82.19 40.88 72.18

8.38*** 6.30* 2.46ns 7.75ns 16.74ns 12.39ns 9.20*

Risk aversion (3) 79.92 76.89 30.17 68.85

An Giang Dong Thap Tra Vinh Mean

2.08ns 5.29ns 10.71ns 3.33ns

5.87ns

Risk- loving (1) 88.30 84.65 42.56 78.05 Source: Survey results in the study area, 2018 Note: * , , *, và ns indicate the level of significance at 1%, 5% and 10%, respectively

The research results show that, the group of farm households with “Risk- loving” attitude has an average economic effect of 78.05%, while the group with the attitude of "Risk neutral" and the group having the attitude of "Risk aversion", the economic efficiency levels achieved are 72.18% and 68.85%, respectively. The results also show that only the difference in economic efficiency between the “Risk-loving” attitude group and the "Risk aversion" attitude group is statistically significant.

Due to the characteristics of the technical level in production, the risk- taking attitude of each farmer and the production conditions in each locality in the study area is different, so the level of economic efficiency that farmers achieve are different. The average economic efficiency that the farmers in the

area achieve is 70.65%. This result also shows the potential for improving economic efficiency and increasing profits.

Because economic efficiency is achieved in accordance with different attitudes to risks, the actual returns that farmers achieve in each group of households according to their attitudes toward risk are also different, specifically: “Risk-loving” has an average real profit of 1.37 million VND per 1,000m2, "Risk neutral" is 1.16 million VND per 1,000m2 and "Risk aversion" is VND 1.13 million per 1,000m2. In particular, the actual returns achieved by households with the “Risk-loving” attitude were statistically significantly different and higher than those with the “Risk aversion” attitude which is 0.24 million VND per 1,000m2.

The average profit lost due to economic inefficiency, specifically: "Risk- loving" is 0.18 million VND per 1,000m2; for "Risk neutral" is 0.30 million VND per 1,000m2 and 0.31 million VND per 1,000m2 for "Risk aversion". The average profit loss caused by inefficiency can be considered significant, as it accounts for nearly 25% of the net profit that the farmer earns. In particular, the less risk aversion households have, the more profits will be lost due to inefficient declining.

Unit: million VND / 1,000m2 Difference

Profit

Sample

(1)-(2)

(2)-(3)

(1)-(3)

Attitude towards risk Risk Risk- neutral loving (2) (1) 1.16 1.37 1.46 1.55 0.30 0.18

Risk aversion (3) 1.13 1.45 0.31

Reality Potential Loss

0.21ns 0.09ns -0.12**

0.03ns 0.02ns -0.02ns

0.24* 0.10ns -0.14**

1.17 1.46 0.29 Source: Survey results in the study area, 2018 Note: *** ,** , *, và ns indicate the level of significance at 1%, 5% and 10%, respectively

Table 4.11 Profits are lost due to economic inefficiency

Therefore, if we have appropriate solutions to help farmers reduce their fear of risk and improve economic efficiency in production, it will significantly increase profits from production activities of farmers.

4.5 Some solutions to limit risk aversion and improve economic efficiency in hybrid maize production

4.5.1 Solutions to limit risk aversion

Building and developing links in production and consumption of products: Farmers producing hybrid maize in the area need to actively participate in local mass organizations. It is necessary to promote building and developing linkage

contracts between producers and producers who supply inputs and consume products. This activity is aimed at helping farmers limit market risks. At the same time, this activity also helps step by step improve product quality and reduce production costs, contributing to improving product competitiveness.

Solutions to diversify income: Farmers need to boost the number of other sources of income from non-agricultural activities or other agriculture based on optimally exploiting the production area, family labor in hybrid maize production, participating in non-agricultural production activities. In addition, farmers can also diversify crops in the form of intercropping or rotation, to ensure compliance with technology and market trends.

In addition, Orientation to build a pilot agricultural insurance service model on maize: The State allows the State to research and build a pilot agricultural insurance service model on maize. This solution is to help stabilize development, to step by step improve the competitive advantage of the corn production industry, to more and more meet the domestic consumption demand as the state's orientation.

4.5.2 Solutions to improve economic efficiency

To improve economic efficiency, it is necessary to properly implement

measures to limit risk aversion of farmers as mentioned above.

Local authorities and agricultural sectors need to promote propaganda, in order to raise awareness of farmers in deciding to use input resources in production, as well as economic and markets knowledge for farmers.

Encouraging the expansion of production acreage and scale at the farm household, however, it is necessary to have the State's support in organizing the production development and orienting the market development for the hybrid maize production.

The State needs to plan and centralize production zoning in order to have a more effective investment strategy and product management organization due to the fact that not all localities in the production area achieve high efficiency. The State also needs to invest in and improve infrastructure for production, logistics infrastructure, and developing consumer markets.

CHAPTER 5 CONCLUSION AND SOLUTION

5.1 Conclusion

The results of measuring the risk attitude of farmers showed that most of the hybrid maize producers in the study area had risk aversion. Estimated result of the Ordered logit model shows that: education, mass participation, production experience, participation in production training of farmers and income diversification are the main factors affecting farmer's attitude towards risk. In particular, the results of testing the optimal use of inputs showed that the optimal input use of farmers did not change significantly according to attitude towards risks. The reason is that the decision to use inputs in production is mainly based on experience by farmers.

The results of estimating the economic efficiency in production of households show that the average economic efficiency achieved is 70.65%. In addition, the research results also showed that the factors affecting economic efficiency in production include the high risk aversion attitude of the farmers, the production area, the education level of the household head, the number of labors in the household, the production area and the ratio of income from hybrid maize production to the total income of the household. With the above average efficiency level, the average profit lost due to inefficiency is nearly 0.29 million VND / 1,000 m2.

Based on research results, the author has proposed several solutions: First, Solution to build and develop links in production and consumption of products. Second, farmers need to diversify their incomes through diversifying their production based on reallocating production resources, especially labor resources in the household. Third, the orientation to build a pilot agricultural insurance service model on corn. Fourth, the state continues to propagate, to raise awareness of farmers in deciding to use input resources in production, as well as economic knowledge, market for farmers. Fifth, encourage the expansion of acreage and scale of production at the farm household level. Finally, the state needs to plan and centralize production, to have an effective investment strategy and production organization.

These results have enriched the empirical literature on risk measurement

and served as a basis for research on other subjects in agriculture.

5.2 Proposal of next research directions

In the scope of research content set and research results achieved, this research still has some shortcomings and certain limitations. Therefore, the author recommends the next research directions.

Due to time and resource constraints, the size observant in this study is limited. Therefore, the author recommends that other studies be done on a larger basis. In particular, the experimental game of measuring attitude to risk is designed through more different payment levels, which can help to reflect the risk attitude characteristics of the farmer more accurately. At the same time, the author also recommends that research is done on the basis panel data to consider the evolution of attitude characteristics towards risks of farmers over time.

The author recommends that the study of farmers' attitudes towards risks for some other agricultural products in the area should be expanded in the coming time. The research results will be an important basis for policy development, planning and managing of agricultural production in the area.