Quality parameters in relation to consumer's preference in

ricebean

Doreen Buergelt; Matthias von Oppen, Department of Agricultural Economics, CAU, Kiel, Germany

J.P. Yadavendra, Gramin Vikas Trust, Gujarat, India

1 Introduction

There are many neglected plants, like the ricebean (Vigna umbellata), which continue to be

grown particularly in rural areas of developing countries, and thus contributing to the

livelihood of the poor. These species are known as ‘orphan’ or ‘underutilized’ crops and they

are i) locally plentiful but globally rare, ii) scientific information and knowledge about them is

little and iii) their current use is limited, relative to their economic potential (GRUERE ET AL.

2007). Many of these orphans have a high potential to be improved by breeding as a mean of

generating food and sustainable income for the local producers and chain actors.

The ricebean as one orphan crop has received only little scientific attention thus, there are no

improved varieties, and it has fallen behind the major pulses in India and Nepal.

Disadvantages as low yield and high labour inputs in comparison to e.g., green gram (Vigna

radiata) and black gram (Vigna mungo) resulted in a decrease of area planted to ricebean.

The hypothesis behind this research is that plant breeding will reduce these disadvantages and

improve its position against competing crops in two ways. First, by increasing the yield and

second by varieties which meet the preferences of consumers and which would therefore gain

a higher price. Consumer preferences where captured from ricebean samples with their

respective prices from Nepalese and Indian markets. The samples were analysed in labs to

document the several parameters. Multivariate regression was used to estimate the influence

of characteristics significantly affecting prices. With these information breeders can assess the

expected price of an improved ricebean variety at an early stage in plant breeding as quantities

of 100 to 200 grams are sufficient. The terms ‘characteristics’, ‘traits’, and ‘attributes’ are

used interchangeably to designate quality attributes of commodities.

The remaining paper except the introduction is divided in further three sections. The

contributions of the ricebean for people in India and Nepal are described in the first section;

chapter two. This includes facts about ricebean marketing in rural markets, production and

consumption, research affords to improve the ricebean and how objectives of plant breeders

match with desires of consumers. The second section, chapter three gives a short literature

review of hedonic price analysis followed by a description of the selected variables included

in the analysis to identify the variables significantly influencing the ricebean price. Further,

data collection and the analysis with interpretation of the results are discussed. The last and

third section; chapter four summarizes research results.

2 Economic issues in ricebean development in India and Nepal

2.1 Ricebean: An orphan pulse crop in India and Nepal

The Ricebean (Vigna umbellata) has always been an established pulse in India and Nepal,

because of its nutritious value. Crops, like the ricebean are known as ‘orphan crops’ because

they have been largely overlooked by the research community despite their importance to

rural livelihoods on a local level - particularly in poor areas. In addition to their nutritional

value, these ‘orphan’ crops are well adapted to the low-input conditions typical of marginal

areas and so give yields in situations where modern crop varieties can barely survive.

However, there are no improved varieties, and so the ricebean has fallen behind other pulses

and only poor farmers in marginal areas grow it.

Ricebean is found in many parts of Nepal and in the Indian States of Uttarakhand, Orissa,

Madhya Pradesh and Chhattisgarh, often intercropped with maize. Areas where the ricebean

is grown today are characterized as remote regarding the access to markets and prevalence of

subsistence households. The ricebean grows well on many soils, and shows pest resistance

and potential for good yields of nutritious fodder and high quality grain. There are no

established marketing channels, and seed supply is limited or non-existent.

2.2 Ricebean marketing on rural markets, production and consumption in India and

Nepal

Since the ricebean is an orphan crop no official governmental data about area, yield or

production exist. Only the Nepalese government documents a group of ‘other’ pulses (Figure

1) in their statistics which include the ricebean, called masyng. One quarter of the included

group ‘Others’ is represented by the ricebean. The graph (Figure 1) also shows that the

production of the most important Nepalese pulse, the lentil, has increased in Nepal. The lines

of pigeon pea, chickpea, grass pea and soybean did not show any positive or negative changes

and were deleted from Figure 1 to increase the clarity.

There are four observed marketing channels through which the ricebean gets in touch with

consumers in India and Nepal. The first channel ends on haats or mandis. A mandi is an

official open market place with fixed operating days and hours. There are about 7,200 mandis

in India which are mostly primary wholesale markets and controlled through APMC

(Agricultural Produce Marketing Committee) acts. Since the amended AMPC acts in 2003

farmers are allowed to sell direct on mandis but mandis are usually near to important hubs of

production, towns and district headquarters. Thus, farmers from remote areas have limited

access to mandis and depend on commission agents and wholesaler to sell their products. The

unregulated markets are called haats and there are about 47.000 in India (DEBROY 2005,

KUMAR AND PATWARI 2008). Comparable numbers for mandis and haats were not found for

Nepal. The second way to sell ricebean is through street traders. These traders are not selling

on official market grounds. They are sitting at the roadside during the operating time of

mandis. Fixed shops are the third observed channel to sell ricebean. These permanent shops in

buildings can be specialized on certain goods like pulses but most of them are selling

everything from food over hygiene products to toys. In India they are called ‘Kirana’ stores

too. The fourth marketing channel for ricebean is the emerging supermarkets (MINTEN 2008).

These were found to be sources for ricebean in Nepal but not in India. After the short

introduction into ricebean marketing the next part comprises facts about ricebean production

and consumption.

Figure 1: Pulse production in Nepal, 1993-2005

0

20

40

60

80

100

120

140

160

180

93-94 94-95 95-96 96-97 97-98 98-99 99-2000

2000-01

2001-02

2002-03

2003-04

2004-05

Year

1000

Ton

s

Lentil Black gram Horsegram Others*

* Masyng, Field pea, Cowpea, Broad bean, Phaseolus, Mungi etc.

Source: Agricultural Census Nepal

The work in India and Nepal revealed that the ricebean is cultivated in many ways. Often the

ricebean is intercropped and mixed cropped with maize at maize field borders. Further,

ricebean is grown on terraces of the rice fields so called bunds and together with pigeon pea.

Seldom is it grown as sole crop. The observed area which farmers use for ricebean varies

between 20 and 1500m² in Nepal and 1.5 to 1000m² in India. Nepalese farmer harvest

calculated 0.1-0.6 kg per m². Indian farmers harvest 0.1 to 1.0 kg per m². There are a lot of

farmers who use ricebean only for home consumption i.e., they do not sell it on markets

because they only have small amounts of ricebean and they will not get a good price for it.

The consumption of ricebean differs between India and Nepal because it is heavy connected

to religion and rituals in Nepal but not in India. The ricebean is consumed for many festivals

in Nepal. There it is a fixed part of their religious live. Nepalese offer ricebean at marriages

and festivals. Mainly it is eaten as dhal that is a kind of soup which is consumed together with

rice. It can also be roasted and offered as snack. Another form of preparation is to mix the

soaked ricebean with spices and let it dry in the sun. This is some kind of snack, too. Ricebean

is also used as fodder plat for animals.

2.3 Ricebean improvement research

The complex ricebean research is conducted in the context of ‘Food Security through

Ricebean research in India and Nepal’ (FOSRIN) a 6th Framework Programme of the

European Commission. It involves cooperation with seven partners (Universities, NGOs and

government) in Europe and Asia, including Agricultural Universities in Palampur, Himachal

Pradesh, and Jorhat, Assam both located in ricebean growing areas of India. The project aims

to make ricebean much more widely grown than it is now, by providing varieties that match

farmers’ and consumers’ needs. The interdisciplinary cooperation in the project permits the

identification of promising parents for breeding programmes, and varieties which also meets

consumers’ preferences.

2.4 Plant breeding objectives and the desires of consumers

In the case of the orphan crop ricebean there are two main objectives which should be

achieved. First, poor farmers in remote areas should benefit of an improved ricebean variety

with higher yields and good pest resistance to feed their families and animals. Thus, a new

ricebean variety has to fit in farmers agronomic and taste preferences. In a second step further,

farmers should be able to leave their subsistence level and sell the ricebean at markets to gain

income. Consumers, who buy on markets, will only purchase ricebean which meet their

preferences regardless of agronomic constrains. Thus, there has to be a compromise between

breeding for better yield and other agronomic aims and the preferences of consumers which

can be farmers and non-farmers. Consumer’s preferences were captured by estimating a

hedonic demand function which is explained in the next chapter.

3 Estimating a hedonic demand function for ricebean

3.1 Brief review of hedonic demand estimation

The central idea of the hedonic price analysis is that goods are valued by consumers for their

utility deriving characteristics (LANCASTER 1966). Hence, consumers are not evaluating the

good itself but the characteristics of goods and demand is a function of product

characteristics, not of the product. But the demand for a special characteristic is revealed

through the demand for a product that contains the characteristic. Therefore, if the quantity of

the demanded characteristic changes the product prices are changing, too (BROCKMEIER

1993). Rosen (1974) proposed that characteristics have implicit prices in competitive markets

because consumers evaluate the characteristics by purchasing the product with the demanded

attribute. Further, the market price is a sum of all implicit prices of all quality characteristics.

Thus, the product price is a function of its characteristics (ROSEN 1974).

Several studies approve that the hedonic price analysis can be used to identify characteristics

of food products significantly influencing the price. (WAUGH 1928, LADD AND SUVANNUNT

1976, BROCKMEIER 1993, JIMÉNEZ PORTUGAL 2004.

3.2 Determinants of ricebean value to consumers

As described in the beginning of chapter three goods are valued by consumers because of

their utility-deriving characteristics which determinate the perceived value of a good in the

end. Thus, the quality or value of a whole good depends on its single characteristics. A good

quality which reveals through consumers acceptance and purchase is not determined by a few

visible characteristics. It is more a complex composition of several traits and a lot of them are

not visible (VON OPPEN 1978). Characteristics of beans which influence the quality and

therefore the value for consumers can roughly be separated into two groups, evident and

cryptic characteristics. Evident traits are visible for consumers like colour or shape and

cryptic ones are not visible and can only be judged after consumption. Examples for cryptic

traits are ingredients like fat or protein and culinary qualities like swelling capacity. Both

types of characteristics, evident and cryptic, are related to each other. Thus consumers infer

from evident about cryptic characteristics. We all know for instance, that a really green apple

will taste sour whereas red or yellow apples are sweet (JIMÉNEZ PORTUGAL, 2004).

The characteristics in Table 1 which were included into the hedonic price analysis were

chosen by reviewing articles about the nutritional composition of ricebean and about

characteristics affecting domestic processing and cooking (JIMÉNEZ PORTUGAL 2004,

(MALHORTA ET AL. 1988, THARANATHAN AND MAHADEVAMMA 2003, PATWARDHAN 1962).

After choosing the traits we selected chemical analysis from the AOAC (Association of

Official Analytical Chemists) catalogue. The analysis in Nepal was conducted in the lab of

one project partner, Li-Bird in Pokhara. In India the samples were analysed at the Agricultural

University in Anand, Gujarat.

Table 1: Selected characteristics for the hedonic price analysis

Ricebean characteristic

1. Fat 8. Swelling capacity

2. Protein 9. Water up-take capacity

3. Moisture 10. L/B ratio

4. Ash 11. Colour

5. Carbohydrates 12. Share foreign matter

6. Crude fibre 13. Colour diversity index

7. 100-seed weight

Table 2 gives an overview over the quality characteristics that were quantified in the labs. All

characteristics are explained in the following part in the order used in Table 2. The price as

dependent variable is shown to give an idea of the ricebean price variation. The weight of 100

seeds is an indicator for the seed size. The higher the weight of a sample of 100 seeds the

bigger are the seeds. The water uptake measures the weight increase by soaking. Ricebean

like other pulses is soaked over night before cooking; thus it is an important processing trait.

Instead of weight increase the swelling capacity documents the volume increase by soaking.

Both traits were quantified with a sample size of around 100 seeds not with single seeds. That

gives the advantage that both characteristics can reveal the problem of hard seeds. Hard seeds

are definitely disliked because their seed coat is impermeable for water. Hence, these seeds do

not swell and often they stay hard even after cooking (REYES-MORENO ET AL. 2000). The

ricebean form was measured by the length to breadth ratio. A value of one would reveal a

round form because length and breath are the same but the minimum value of the L/B ratio in

the sample (Table 2) is 1.36 which depicts an elongated form. The maximum value is at 2.05

and compared to the minimum value it shows the great differences in ricebean forms. The

cleanliness or better purity was determined by weighting the sample with and without dirt.

Dirt and foreign materials are small stones, other pulses and broken/ destroyed beans. The

foreign material corresponds to the share which women would remove before soaking or

cooking the ricebean.

Table 2: Outline Indian and Nepalese samples

Characteristic Unit Minimum Maximum MeanStandard deviation

Price Nepal Nepalese Rupees (NPR)/kg 25.00 65.00 43.38 9.94Price India Indian Rupees (INR)/kg 12.20 36.00 23.22 6.81Moisture % 8.10 15.51 10.26 1.16Protein % 14.53 32.16 22.53 4.14Fat % 0.12 0.75 0.39 0.16CrudeFibre % 3.55 6.98 4.71 0.76Carbohydrates % 58.15 71.99 65.08 3.49Ash % 2.44 5.10 4.03 0.49Seedweight g/100 seeds 4.07 19.42 9.20 3.33Foreignmatter % 1.10 19.31 6.81 3.53LBRatio Ratio length to breadth 1.36 2.05 1.66 0.16Water uptake ratio Ratio of weight increase 1.69 2.20 2.00 0.11Swelling capacity ratio Ratio of Volume increase 1.02 1.96 1.33 0.30Colour diversity Herfindahl-Index 0.27 1.00 0.59 0.22N = 162

Source: own data

The moisture content depicts the remaining water in the seeds. It can be disliked by

consumers because good quality ricebean are dried carefully in the sun and fresh harvested

seeds have high water content. The quantified characteristic ash denotes the remain after the

ricebean sample was burned in a funnel. Therefore ash does not contain any organic material

only minerals like iron, zinc or copper. When compared to related pulses like black gram and

green gram the ricebean has a higher average ash content. Analyses of the ricebean revealed

an ash content of 4.03 per cent (KAUR AND KAPOOR 1991) and the analysis for this article

unfold an average ash contents of 4.05 per cent. Whereas the ash content of black gram is

about 3.64 per cent (KAUR AND KAPOOR 1991) and for green gram 3.32 per cent (USDA

2008). The fat content in ricebean with about 0.4 per cent is quiet low but it is one of the basic

nutrition components and should be determinate. The more important component for poor

people in poor countries who cannot effort to buy meat is protein. The protein content of

ricebean is on average 22.5 per cent and was determined by the nitrogen content. The

digestible carbohydrates were calculated on base of the Weender analysis as difference. Due

to the fact that moisture, fat, protein, ash and crude fibre are determined in per cent they can

be subtracted from 100 per cent. The remaining share contains digestible carbohydrates.

Crude fibres belong to carbohydrates but they cannot be fully digested in the human intestine.

The characteristic colour is mentioned in Table 1 only; it is not shown in the outline of Table

2. The different colour shares in each sample were measured once to identify preferred

colours and to determine if consumers like mixed coloured lots of ricebean. The latter was

calculated by a colour diversity index. Usually ricebean is offered in sacks at the markets and

consumers will find, e.g. pure black, yellow and gray ricebean or they find all or some of

these colours mixed together in one sack or lot.

There are a lot of diversity indices from ecology and economy. The simplest way to measure

diversity is to count the different colours, but that does not consider the frequency i.e., the per

cent share of each colour is ignored (DRESCHER 2007). The Herfindahl-Index (HI) shown in F

1 is often used to determine industrial concentrations (PATIL AND TAILLIE 1982). It is

calculated as the sum of the quadratic colour shares of each sample (DRESCHER 2007). The

Herfindahl-Index (HI) is defined as follows:

F 1: ∑=

=

n

iisHI

1

2

where si is the relative abundance of the ith colour e.g., black in the sample containing S

colours (MOUILLOT AND LEPRÊTRE 1999). It ranges from 1/n to 1, whereas the maximum

value of 1 indicates total concentration i.e., the sample consists only of one single colour

(DRESCHER 2007).

Further diversity indices are the Simpson or Berry-Index (BI), which is closely related to the

Herfindahl-Index, the Entropie-Index (EI), the Shannon-Index (H) (PATIL AND TAILLIE 1982,

LEE AND BROWN 1989, MOUILLOT AND LEPRÊTRE 1999). Several indices like Gatlin’s-

Redundancy-Index, Camargo-Evenness-Index, Pielou-Regularity-Index or the Gini-

Coefficient are also used to determine diversity. Anyway, the Herfindahl-Index was chosen

because it cannot take the value zero. This is important since the metric variables are taken

into the regression in log form.

After all determined variables were explained some facts about the sample collection

circumstances are presented in the next chapter.

3.3 Ricebean sample collection and analysis

Ricebean is plated in June-July and the harvest season is from October to November. The

sample collection in India and Nepal had to be organized in a short term to have comparable

qualities and to assure the availability of ricebean in the markets.

All samples were collected between January 10th and March 9th. The investigated locations

were selected through a distribution report conducted by the Nepalese and Indian partners of

the project. The start was made in Nepal Ramechhap which is located in the east of Nepal, the

right end of the red line. All together 14 market places were visited to collect 114 ricebean

samples in Nepal. The route can be reconstructed by following the red line in Figure 2 from

right to left down.

Figure 2: Locations in Nepal

In India ricebean 56 samples were collected in the states Chhattisgarh, Orissa and

Uttarakhand. The research in the east of Madhya Pradesh was unsuccessful. Ricebean samples

were collected 16 markets.

Figure 3: Locations in India

The first samples were collected in Chhattisgarh. The main both areas in Chhattisgarh are

marked in the map (Figure 3) with red circles with a one inside. The work in Chhattisgarh was

hampered by riots of Naxalites a communist movement in India. Their presence avoided

access to remote areas and markets. The order of the remaining areas can be reconstructed by

the sequence of the numbers.

3.4 Estimates of hedonic demand functions

Multivariate regression was used to estimate the influence of characteristics significantly

affecting prices. A stepwise backward multiple regression with price as dependent variable

was conducted. In the beginning 29 variables were included in the model and all irrelevant

independent variables are eliminated step by step (BÜHL, 2002). The dependent variable price

and all metric independent variables were put into natural log form. The log form was chosen

to facilitate the interpretation of the variables coefficients. The calculated coefficients can be

read as elasticities of the characteristics. Further, dummy variables were used to differentiate

countries and regions within countries.

Table 3: Regression results

Variables Coefficient T-valueConstant 1.628 ** 3.424Country 0.899 ** 8.385Orissa 0.082 1.044Uttarakhand 0.196 ** 2.502Ramechhap -0.403 ** -8.365Dhading 0.228 ** 3.615Chitwan -0.156 ** -3.176Syanja 0.227 ** 2.867Nawalparasi 0.117 1.269ln_Protein 0.182 * 1.561ln_Fat 0.093 * 1.705ln_CrudeFibre 0.337 ** 2.644ln_Ash -0.168 -1.384ln_Seedweight 0.250 ** 4.031ln_Foreignmatter -0.072 ** -2.258ln_LBRatio 0.145 0.866ln_Black_group 0.024 ** 3.964ln_Gray_group -0.009 * -1.871ln_Olive_group -0.012 ** -3.054ln_Red_group -0.005 -1.088ln_Herfindahl Index -0.122 ** -2.194N= 162 adj.R²:0.84

* significant at α: 0.1, ** α: 0.05 Source: own data

Table 3 depicts the results of the regression. The origin dummies for country and regions have

an impact with a significance level of α = 0.05. The quality characteristics crude fibre, seed

weight and foreign matter have a significant price influence at a level of α = 0.05, too.

Whereas crude fibre and seed weight have a positive influence and foreign matter has a

negative influence. Further the colour black is significantly preferred and olive is disliked at a

significance level of α = 0.05. According the diversity; consumers prefer mixed coloured

ricebean lots at a significance level of α = 0.05. The Herfindahl-Index ranges from 1/n to 1,

whereas the maximum value of 1 indicates total concentration i.e., the sample consists only of

one single colour. The Herfindahl index has a negative sign and by reducing this characteristic

colour diversity increases. The characteristics: fat, protein and gray are significant at a level of

α = 0.1. Fat is liked by consumers, which is reasonable because of its high caloric energy.

Protein is also preferred by consumers because pulses are an important source for proteins in

low income countries like India and Nepal. Other sources of protein like animal products

(meat, eggs, milk) are to costly for the major part of the people.

After the results have been explained and discussed in this chapter, the following chapter

refers to the interpretation of the coefficients obtained from the regression.

3.5 Interpretation of results

All following interpretations of the results are made under ceteris paribus assumption. If one

variable is changed all remaining variables remain constant. The coefficients for the colour

black and olive indicate that consumers in India and Nepal prefer black ricebean and do not

like olive ricebean. For the interpretation of the remaining characteristics Table 4 is used.

Table 4: Significant characteristics

Characteristic Mean Minimum Maximum Change %

Coefficient/

Elasticity

Price

effect %

Crude Fibre 4.71 3.55 6.98 2.27 48 0.34 16

Seed weight 9.20 4.07 19.42 10.22 111 0.250 28

Foreign matter 6.81 1.10 19.31 -5.71 -84 -0.07 6

Herfindahl-Index 0.59 0.27 1.00 -0.32 -54 -0.12 7

Source: own data

Table 4 includes all quality variables at a significance level of α = 0.05 with their respective

means, maximums and minimums to show how the trait differs in the sample. Since the

coefficient is the price elasticity of a characteristic a percentage change in the quantity of the

characteristic changes the price by the coefficient. Crude fibre has a positive coefficient of

0.34. That means consumers like crude fibre in ricebean and an increase of crude fibre would

increase the value of ricebean for consumers. An increase of the sample’s mean value to the

maximum value was chosen assess the limit value of crude fibre (values a shown in Table 2).

Thus, an increase from 4.71 g/100g to the maximum value of 6.98 g/100g reveals a positive of

48 per cent. A 48 per cent increase of crude fibre multiplied with the coefficient/ elasticity of

0.34 results in a price increase of 16 per cent, if all other characteristics remain constant. The

negative elasticity of the Herfindahl-Index could be interpreted as follows. Decreasing the

index implies an increase of colour diversity (see F 1). If the index is reduced from the mean

value of 0.59 to the minimum value of 0.27 the value of ricebean reflected by the price would

increase by 7 per cent because consumers like high colour diversity. The increase of seed

weight which represents the seed size would have the greatest price impact of 28 per cent. The

share of foreign matter in ricebean has a significant impact on the price. Consumers have a

willingness to pay for clean bean lots. Thus, a reduction of foreign matter from the mean

value to the minimum would increase the price by 6 per cent.

4 Conclusion and discussion

Breeding without having consumer’s needs in mind can lead to crop varieties which do not

help farmers or consumers. Because farmers cultivate an improved crop with high yields and

pest resistance but in the end they cannot sell it to create income and in the worst cases

farmers do not want to consume this crop themselves. There has to be a balance between

agronomic improvements and preferences of consumers.

The results show that there are potential characteristics which should be considered when

breeders choose the parents for further breeding. Increasing the seed size which is depicted by

seed weight, crude fibre and colour diversity would have a positive price effect which reflects

consumers preferences. Consumers do not like foreign material which can be considered by

decreasing these characteristics, whereas, the content of foreign matter is not a breeding

problem. Farmers or traders could solve this problem and clean the ricebean before selling.

But the result shows that there is a willingness to pay for cleaner ricebean lots. With these

elasticities breeders can assess the expected price of an improved ricebean variety at an early

stage in plant breeding as quantities of 100 to 200 grams are sufficient to calculate the

consumer preferences.

Future prospects are to combine farmers and consumers requirements. For example, if

breeders improve the yield of ricebean by 20 percent than the additional the improvement

through hedonic analysis adds even more improvement. Since all results can only be

interpreted under ceteris paribus condition, seed weight is chosen in the sample to represent

the quality improvements. If an Indian farmer has 100 kg of ricebean which he could sell on

the market for 20.00 Rs/kg he gains 2000.00 Rs. Possible yield enhancement by breeding

could give this farmer a 20 per cent increase to 120 kg of ricebean. An increased price

resulted from higher seed size (seed weight) could increase the price by 20 per cent to 24.00

Rs/kg. With these improvements the farmer could sell 120 kg for 24.00 Rs/kg which creates a

turnover of 2880.00 Rs. All together an upgrade of 44 per cent could be reached through yield

increases and higher prices which could pave the way for an innovation like the improved

ricebean.

5 Acknowledgements

The research was supported by various partners and without them it would not have been

possible to conduct the work. In India the surveys and sample collection was supported by

J.P. Yadavendra from Gramin Vikas Trust (GVT), Naveen Kumar from the Agricultural

Universities in Palampur, Himachal Pradesh, Seuji Neog from the Agricultural University in

Jorhat, Assam and Madhya Pradesh Rural Life Program (MPRLP). Li-Bird a NGO from

Pokhara in Nepal supported the local research. The authors wish to thank all of them for their

efforts by managing and improving the research.

6 References

Brockmeier, Martina, Ökonomische Analyse der Nahrungsmittelqualität, Kiel: Wissenschaftsverlag Vauk, 1993.

Drescher, Larissa, Healthy food diversity as a concept of dietary quality: measurement, determinants of consumer demand, and willingness to pay, Göttingen: Cuvillier, 2007.

Debroy, Bibek, About India's haats. http://www.rediff.com//money/2005/jul/13guest1.htm

Gruere, Guillaume, Nagarajan, Latha and King, E.D.I. Oliver, “Marketing underutilized plant species for the poor: a case study of minor millets in Kolli Hills, Tamil Nadu, India” Rome, Italy: 2007.

Jiménez Portugal, Luz Alicia, Relevant quality attributes of edible dry beans, Osnabrück: Der Andere Verlag, 2004.

Kaur, Deepinder and Kapoor, A.C., "Nutrient composition and antinutritional factors of rice bean (Vigna umbellata)" Food chemistry, 43, 1991, pp. 119-124.

Kumar, Vijay and Patwari, Yogesh, "Organised food retailing: a blessing or a curse?" Economic & Political Weekly, 17, 2008, pp. 67-75.

Ladd, George W. and Suvannunt, Veraphol, "A model of consumer goods characteristics," American journal of agricultural economics, 58, 3, 1976, pp. 504-510.

Lancaster, K.J., "A new approach to consumer theory" Journal of political economy, 74, 2, 1966, pp. 132-157.

Lee, Jonq-Ying and Brown, Mark G., "Consumer demand for food diversity" Southern Journal of Agricultural Economics, 21, 2, 1989, pp. 47-53.

Malhorta, Sarla, Malik, Darshan and Singh, Kulip, "Proximate composition and antinutritional factors in rice bean (Vigna umbellata)," Plant foods for human nutrition, 38, 1988, pp. 75-81.

Minten, Bart, "The food retail revolution in poor countries: Is it coming or is it over?" Economic development and cultural change, 56, 4, Univ. of Chicago Press: Chicago, 2008, 767-789.

Mouillot, David and Leprêtre, Alain, "A comparison of species diversity estimators," Researches on Population Ecology, 41, 2, 1999, pp. 203-215.

Patil, G.P. und Taillie, C., "Diversity as a concept and its measurement" Journal of the American Statistical Association, 77, 379, 1982, pp. 548-561.

Rosen, Sherwin, "Hedonic prices and implicit markets: product differentiation in pure competition," Journal of political economy, 82, 1, 1974, pp. 34-55.

Waugh, F.V., "Quality factors influencing vegetables prices," Journal of farm economics, 10, 1928, pp. 185-196.

Von Oppen, M., “A preference index of food grains to facilitate selection for consumer acceptance,” Economic program ICRISAT, 1978, Discussion paper 7, Hyderabad India.

USDA, National nutrient database for standard reference. http://www.nal.usda.gov/fnic/foodcomp/search/