STRUCTURE, CONDUCT AND PRICE INTEGRATION
CHARACTERISTICS OF URBAN FOOD CROP MARKETS IN CAMEROON
Carlton G. Davis and Jean Ongla
Staff Paper 164 October 1980
Staff Papers are circulated without formal review by the Food and Resource Economics Department. Content is the sole responsibility of the authors.
Food and Resource Economics Department
Institute of Food and Agricultuial Sciences
University of Florida
Gainesville, Florida 32611
STRUCTURE, CONDUCT AND PRICE INTEGRATION
CHARACTERISTICS OF URBAN FOOD CROP MARKETS IN CAMEROON
Carlton G. Davis and Jean Ongla
In recent years Cameroon has experienced rapid rates of growth in the nonfarm demand for a wide range of staple food crops. This phenomenon, is to a large extent, associated with the rapid rate of off-farm migration and the growth of the non-agricultural urbanbased sectors of the economy. These two factors have impacted the income and population levels of urban areas, which in turn, have generated upward shifts in urban aggregate food demand. On the other hand, the historical skewness in infrastructural investment in favor of the traditional export crops ( cocoa and coffee), resulted in productivity and.distributional bottlenecks in the food crop sector; to the extent that urban areas have rapidly become food deficit areas .
One immediate impact of urban area food deficit is the rapid escalation of food crop prices in the two major urban consuming centers of Douala (400,000 inhabitants)and Yaound6, the capital (300,000 inhabitants). In 1976, approximately 10 percent of the country's population was located in these two consuming centers . The basic food crops are cereals and starchy roots and plantains. Millet, sorghum and maize are the primary crops in the cereal group, while cocoyams, cassava (manioc), and yams are the primary crops in the starchy roots category. These crops constitute the basic staple food items in rural areas, as well as a large proportion of the food' items consumed in urban areas. There is indication,
Carlton G. Davis is Professor of Food and Resource Economics, University of Florida and Jean Ongla is Secretary General at the University Center of Agriculture, Dschang, Cameroon, West Africa.
however, that the urban population is experiencing certain shifts in the food commodity bundle, with major shifts coming from increased consumption of pulses and imported food items . Table 1 shows the average retail price per unit for selected food crops among "Medium Income Consumers," in the city of Yaound4 during the month of February
1969 and 1975. The generally upward trend in the average price of food crops in the Yaound6 market, particularly since the 1971-1972
TABLE l.--Retail Price Paid by Consumers (Medium Income Families)
for Selected Staple Food Products, Yaound4 Markets, February 1969-1975*
Commodities 1969 1970 1971 1972 1973 1974 1975
Maize 38 30 40 40 43 35 42
Plantain 15 17 27 20 27 28 30
Cocoyams 16 14 14 18 23 24 24
Cassava 10 9 14 15 17 19 23
Yams 26 33 36 52 45 46 54
Beans 59 70 89 91 93 114 159
Onions 125 120 140 100 150 160 112
Rice 57 55 80 60 96 100 143
The National Price Reporting Board publishes two sets of
price data, one for "European Families," which is collected mainly at places where expatriates and high income Cameroonians shop, and the other for "Medium Income Families," which is collected in public markets and other places where the majority of population shop. The prices are derived from .
period, is a general indication of the upward trend in food crop prices in urban markets. Official government statistics indicate increases of the following magnitude among food crops consumed by medium income families in the Yaounde and Douala marketsybetween the first quarter of 1971 and the first quarter of 1972: cocoyams,
33.3 percent; yams, 25.6 percent; beans, 18.0 percent and palm oil, 27.3 percent.2/
Given the rapidly spiraling price levels in urban food deficit areas, it is of interest to gauge the relative price levels in urban centers and the food surplus rural areas of the country. Unfortunately, there are no comparable price series for rural areas that would facilitate assessment of the trend component. However, crude estimates of relative price levels in rural and urban markets were made possible by Ongla's February 1972 survey of cost of production and general price levels in the Zone of Intervention of the Yemessoa area . The zone is a major food crop producing area, which marketed approximately 35 percent of the total tonnage of the seven staple food crops produced in 1972 . It is located approximately 65 kilometers (40 miles), northwest of the capital, Yaounde. Table 2 shows a comparison of average unit prices for related food crops in the Yaound6 urban markets and two of the major farm markets in the Yemessoa area. The table shows that in February 1972 the average retail price of cassava,
TABLE 2.--Average Retail Prices of Selected Food Crops in Yaound6
Markets and Farm Markets in Yemessoa Area, February 1972 Commodities Yaound6 Marketsa Ngomo and Et ka Percent of
Market:b Yaound6 Prices
Average CFA Francs/Kilogram
Cassava 15.0 10.0 66.7
Cocoyams 18.0 10.0 55.6
Plantain 20.0 7.5 37.5
a/Derived from Table 1 and .
b/Derived from .
cocoyams and plantains ranged from a low 38 percent to.67 percent of Yaound6 retail prices.
Objectives and Rationale
This paper has a two-fold objective. First, it evaluates some of the important structural and conduct characteristics of Cameroon urban food crop marketing channels to provide some insights into how these characteristics impact the overall efficiency of the food distribution system. Second, it empirically evaluates the degree of price integration among selected food crop markets in a specific urban area, to arrive at some measure of the level of economic integration among specific commodity markets.
The practical implication of this type of analysis is likely to be derived from the implementation of informed market policy instruments that are likely to guide urban food prices to levels consistent with the overall economics of production and distribution. As such, inter- and intra-market price levels in urban areas would be less responsive to a number of non-economic forces operating in the food marketing channel. The importance of such policy instruments are likely to be great, in light of the projected high rate of increase in urban demand for foodstuffs over the next decade. For example, the population of Yaound4 is projected to increase to 650,000 by the year 1985/86. At that time the total quantity of seven selected staple food commodities in demand by the city is projected to be 2.8 times that demanded in 1972/73. Of these seven food products, the highest annual rate of increase in demand will be for rice (13.6 percent), yams (11.8 percent), maize (9.8 percent) and groundnuts (9.5 percent). The lowest will
be for plantain (7.5 percent), cassava (7.9 percent) and cocoyams (7.2 percent) .
Methodological Framework and Data Base
Market Structure and Conduct Analysis
Market structure refers to certain physical dimensions of the marketing system, while market conduct refers to the patterns of behavior of market participants . Empirical analysis of market structure and conduct parameters is based on objective characterization and quantification, when possible, of the institutional and behavioral components of the study area's marketing chain. Food crop marketing is dominated by the traditional marketing system in Cameroon. Often described as inefficient, costly, and disorganized, this system nevertheless handles most of the local foodstuffs marketed in the country. The principal characteristic of the system is that there are numerous intermediaries between the producer and the final consumer. These middlemen represents an aggressive group of entrepreneurs who, under conditions of rudimentary marketing facilities, informal market intelligence, and the virtual absence of official grades and standards for commodities, are the major integrating mechanism between rural/urban and intra-urban markets. One study , reports that, for all food products, there is an average of 2.1 middlemen operating between the producer and the final consumer.
In light of the important market integrating role played by food crop marketing middlemen, certain structural characteristics of the group will be evaluated to gleen insights as to the extent (if any) to which such characteristics might impact on the distributional efficiency of urban food crop markets. Specifically, the analysis
will concentrate on selected characteristics of these agents within the three major categories of markets (urban central, neighborhood, and street retail) identified by Ongla , in the Yaound6 urban trade area.3/ Ongla , also identified three major categories of urban marketing agents: (a) wholesalers (b) market retailers and (c) street retailers. Wholesalers provide the link between producers and urban central activities. They are primarily responsible for the collection, assortment and transportation of food crops from rural areas to urban centers. Wholesalers sell their products to other groups of middlemen, restaurants, institutions or in some instances, directly to consumers. Market retailers generally remain in the same principal markets or transport their products to neighborhood markets. They generally obtain their products from wholesalers or to a lesser extent from rural markets and sell to other retailers or to consumers. Street retailers, on the other hand, sell their products directly to consumers. They can be located in all categories of urban markets and at major intersections of the city.
Six selected characteTistics are used to evaluate elements of market structure and conduct for the three categories of marketing agents operating in the three types of urban markets. These characteristics can be grouped as (1) demographic (2) elements of market concentration (3) supply source (4) transportation utilization
(5) market operation financing and credit utilization and (6) market communication, trade association and price determination. Empirical evaluation of these six components is based on the results of a sample survey questionnaire administered between January and February of 1977. A random sample of 31 wholesalers was drawn from four of the major Yaounde market sites frequented by this group. Two of these market
sites are classified as Urban Central Markets (Central and Mokolo), and the remaining two are classified as Neighborhood Markets (Madagascar and Mvog-Ada) . Response was received from a total of 73 market retailers, and the distribution was similar to the wholesaler category in terms of market categories. A random sample of 35 street retailers was obtained from eight different sites throughout the Yaounde area.
Price Integration Analysis
A marketing system performs its coordinating role efficiently when market information is transmitted so that market participants know where to get satisfaction for their products or their expenditures. The well coordinated commodity market implies, among other things, that price levels will tend to move together in competitive markets. This relationship is based on the fact that since the markets are economically integrated, under a competitive structure, price signals will be transmitted with some degree of accuracy between commodity markets. In our analysis of price integration among food commodity markets in the Yaounde area, it is hypothesized that for those commodity markets that are more economically integrated, the price levels would tend to move together, while the converse or at least a lack of similar directional movements would be the case for commodity markets that are less integrated.
Price integration analysis used in this study is similar to the approach used by Jones , in his study of price integration among specific commodities in Nigerian markets. A similar approach has been used by Lele (4], in her study of rice marketing in India. This approach consists primarily of estimation of bivariate correlation coefficients of reported prices, for each commodity, in every pair
of markets for which usable price series was available. The underlying assumption of this approach is that the larger the coefficient, the more integrated are the marketing systems for individual commodities. Following Jones, and with 1975 price survey data as the relevant price series, bivariate correlation analysis is applied to market pairs for ten commodities in three major markets (Central, Mokolo and Mvog Mbi), and six neighborhood markets (Madagascar, Melen, NgoaEkele, Abattoir, Mvog Ada and Nlongkak), in the Yaound6 area.
Letting (X,Y) be the bivariate normal distribution, and (X 1 Y), (X2, 2)* * *I (X12,Y12) be the random sample drawn from the bivariate normal distribution, a test can be made for independence between the two distributions by calculating the coefficient of correlation, p, and testing that it is equal to zero. The estimator of p is 0 which is calculated by the formula:
0= E (X i (i Y
E= (x1 -X (Y1-) 1
12 12 12
n X i i- Xi i
i=1 i=1 =12
12 2 n 22 2 22
E X Q i)Jn E 2 (2 Yi
1= X=1 i= i=1i
Given 0 estimator of p, a test of hypothesis regarding 0 is
possible. Even though the probability distribution of 0 is difficult to obtain, an appropriate test for small samples can still be made. This is related to the fact that approximately, p is normally
distributed, with the standard deviation Sr, defined as:
Sr = (-p2) / (n-2) (3)
where, p is the coefficient found in the sample of size n. Therefore, the test can be applied for:
Ho: p po 0, using a test, in which, t = p-po Sr
or t p-0 or t (4)
(1-p 2) / (n-2)
The null hypothesis states that X and Y are independent, i.e., p=0 for the bivariate normal distribution (X,Y). If the calculated value of t lies in the rejection region,
to > t., n-2
where, a is the significance level and n-2 the number of degrees of freedom, then the evidence will be enough to support the contention that the series of prices X and Y in the two markets are dependent. A high correlation coefficient then means high consistency of the price movements in the two markets, while a weak correlation would mean a weak consistency.
In the absence of a formal market intelligence system in the Yaound6 area, these comparative correlation coefficients are used to evaluate the degree of market integration resulting from 'formal and traditional information systems. Specifically, these correlation coefficients are expected to provide answers to questions such as:
(a) for areas covered by the city, and in the absence of an organized system of market news, information collection and dissemination, do
prices in one market have definitive relationships with prices in other markets within the area? and, (b) are price relationships the same for most crops, and if not, what are the factors accounting for the differences?
In the absence of official and/or other reliable price series data for food crops in the Yaound6 markets, the analysis is based, as indicated earlier, on a.1975 sample of food crop prices in nine Yaounde markets. Three enumerators were assigned to the survey. They carried portable scales and a set of weights for measuring the weight of products in the markets. The nine markets were surveyed on a rotating basis, as per an established schedule, between the months of January and December 1975. Sellers of the ten products used in this analysis were selected randomly by the enumerators, and up to three sellers of a particular product could be interviewed the same day. The ten commodities studied were selected on the basis of their current and projected importance in the daily diets of Yaound4 consumers. The ten commodities studied are plantain, cassava (tubers), cassava (flour), cocoyams,.groundnuts, maize (grain), beans, tomato, onion and rice.4/
Structure and Conduct Characteristics
Survey results shows significant differences in the sex composition of food wholesalers, market retailers and street retailers (Table 3). The table shows that 79 percent of the wholesalers were male, compared to 36.6 percent and 11.6 percent, of the market retailers and street retailers, respectively. There were also significant differences between the three groups of middlemen in terms of schooling, number of dependents, and years of business experience. While only
TABLE 3.--Demographic Characteristics of Marketing Agents,
Yaound6 Markets, 1977
Characteristics Wholesalers retailers retailers
(n=31) (n=73) (n=35)
---------------Percent----------------------Males 79.0 36.6 11.6
Primary or no
schooling' 21.7 49.0 61.9
CEPEb! 43.4 42.9 38.1
BEPCSI d/ 26.1 8.2 -Probatoire- 8.7 -- -Number of dependents
0-4 25.0 67.4 66.7
5-8 58.3 24.5 26.7
9-over 16.7 8.2 6.6
Number of years in
Less than 1 year 11.1 16.4 26.9
1-5 55.5 64.4 61.5
6-10 18.5 12.3 11.5
11-15 11.1 4.1 -16-over 3.7 2.7 -a-The respondents included. in this category declared to interviewers that they did not attend school or that they attended Primary School but dropped out and never completed Primary School.
b/The respondents included in this category declared that they completed Primary School and obtained the Primary School Leaving Certificate, the "Certificat d'Etudes Primaires Elementaires" (CEPE).
- The "Brevet d'Etudes du Premier Cycle" is a diploma awarded students who successfully complete the fourth grade of High School.
-iThe Probatoire" corresponds to the British General Certificate of Education, Ordinary Level*
21.7 percent of the wholesalers had some primary school education or no formal schooling, 49.0 percent of the market retailers, and 61.9 percent of the street retailers fell into this category. It is interesting to note that, for the three categories, the percentage of those obtaining the primary school diplomas (CEPE) were roughly similar (43, 43 and 38 percent, respectively). The CEPE level was the highest educational level achieved by street retailers. On the other hand, as much as 26 percent of the wholesalers received their BEPC, which corresponds to the fourth year high school level. In contrast, only 8.2 percent of the market retailers achieved this level. The BEPC was the highest educational level attained by market retailers, while 8.7 percent of the wholesalers were found to have obtained their "Probatoire", the equivalent to the British General Certificate of Education (Ordinary Level).
On the average, wholesalers tended to have larger families than market and street retailers. The percentage of wholesalers reporting having 5 or more dependents was significantly higher than those of the other groups (Table 3),. Also, a relatively higher proportion of street retailers entered the marketing business each year, compared to wholesalers and market retailers. Approximately 27 percent of the street retailers indicated that they had been in the business less than a year. In contrast, the percentages were 11 and 16 for wholesalers and market retailers, respectively. More than 50 percent of each group had been in the business 1 to 5 years. Most street retailers had been in the business less than 10 years, while almost 15 and 7 percent, respectively, of the wholesalers and market retailers were there between 11 and over 16 years.(Table 3).
TABLE 4.--Business Ownership and Labor Use Characteristics of Marketing
Agents, Yaound6 Markets, 1977
Characteristics Wholesalers Retailers Retailers
(n=31) (n=73) (n=35)
Individual 70.0 89.5 96.4
Partnership 30.0 10.5 3.6
Family 27.6 15.8 22.2
Hired 31.0 -- 3.7
No Help 41.4 84.2. 74.1
Some rough indication of market concentration is suggested by Tables 4 and 5. Table 4 shows business ownership pattern and labor input characteristics of the three groups of marketing agents. The table shows that most wholesale units (70 percent) were owned by individuals, but that this percentage was significantly lower than those owned by market retail (89.5), and street retail units (96.4). However, a significantly larger percentage of wholesale units were owned by some form of partnership. All three categories of middlemen employed family labor in their operations. Wholesalers, however, reported a relatively high usage of hired labor in their operations, in contrast to little or no usage among market and street retailers. In general, however, both market and street retailers' operations were characterized by larger owner-operator labor inputs.
All three middlemen groups handled more than one product (Table 5). In the wholesalers group, almost 38 percent indicated they sold
plantain, while only 7 percent indicated groundnuts and cassava. Plantain was the commodity handled most frequently by most retailers and wholesalers. Cocoyams appear to be second in importance for all groups. Although this table does not include all the foodstuffs sold by the different groups, it nevertheless provides some evidence that middlemen tend to simultaneously handle a wide variety of products. This finding is consistent with Muziol 1976 findings of product diversification among food crop marketing agents for the county as a whole [6, Table 4].
TABLE 5.--Product Distribution of Selected Commodities Handled
by Marketing Agents, Yaound4 Markets, 1977
Commodity Wholesalers Retailers Retailers
(n=31) (n=73) (n=35)
----------------- Percent-------------------------Plantain 37.9 59.3 53.5
Maize 31.0 33.3 23.3
Groundnuts 6.9 42.6 7.0
Yams 20.7 33.3 2.3
Cassava 6.9 46.3 2.3
Cocoyams 34.5 59.3 41.9
There appears to be significant differences between commodities,
in terms of the type of suppliers used (Table 6). Among market retailers, 30 to 50 percent of all commodities, except cassava, came from farm and/or roadside locations. Only 20 percent of cassava purchases came from these two sources. Among wholesalers, 53 to 92 percent of all commodities came from rural markets and/or other locations. Street
TABLE 6.--Supply Sources of Marketing Agents, Selected Commodities,
Yaound6 Markets, 1977
Place of Market Street
Commodity Purchase Wholesalers Retailers Retailers
--------------Percent------------------Plantain Farm 19.05 10.0
Roadside 19.05 20.0
Rural market 42.85 55.0 -Other-, 19.05 15.0 100.0
100.0 100.0 100.0
Maize, Farm -- 10.0 -Roadside 8.3 16.7 -Rural market 75.0 33.3
Other 16.7 50.0 100.0
100.0 100.0 100.0
Groundnuts Farm 25.0 10.0 -Roadside -- 30.0 -Rural market 75.0 60.0 100.0
Other -- -- -100.0 100.0 100.0
Yams Farm 12.5 16.7 -Roadside 12.5 33.3 -Rural market 75.0 50.0 -Other -- -- 100.0
100.0 100.0 100.0
Cassava Farm -- 20.0 -Roadside 33.3 -- -Rural market 66.0 40.0 -Other -- 40.0 100.0
100.0 100.0 100.0
Cocoyams Farm 17.6 10.5 -Roadside 29.4 15.8
Rural market 47.1 42.1 -Other 5.9 31.6 100.0
100.0 100.0 100.0
- Representing such outlets as small markets, the riverside, Food Development Authority (MIDEVIV), Yaounde markets, etc.
retailers, on the other hand, almost never purchase their commodities from farm and/or roadside locations. All their commodities tended to be purchased from "other" locations, with the exception of groundnuts, which are generally purchased in rural markets.
Table 6 illustrates the central role played by wholesalers in linking rural and urban markets. It supports the earlier statement that wholesaler middlemen are the key link in a system of collection, assembling and distribution of foodstuffs to urban markets. Table 7, however, shows some interesting differences among marketing agents in the responsibilities assumed for transporting their commodities. Among wholesalers, 85 percent assumed the major responsibility for transporting their commodities. The proportions were 94 percent and 95 percent among market and street retailers, respectively. This interesting difference probably can be explained by the considerable difference in transportation source among wholesalers and retailers. Table 7 also shows that wholesalers exclusively used trucks in transporting the six commodities studied. The proportion of agents transporting goods by trucks declined significantly as one move from the market retailer group to the street retailer group. Conversely, the proportion of agents transporting commodities by headload and "other" methods is highest among street retailers and market retailers. These characteristics are directly related to the observed responses of the agents regarding responsibility assumed for transporting their commodities to market. Wholesalers tended to cover a relatively wide geographical area in their food crop assembling activities. As such, they are forced to utilize certain types of transportation contractual arrangements to have their goods delivered to Yaound6 markets, since
TABLE 7.--Transportation Utilization of Marketing Agents, Selected Commodities, Yaound.
Wholesalers Market Retailers Street Retailers
Commodity -------------------------------------Source of Transportation--------------------------Truck Headload Other Truck Headload Other Truck Headload Other
-----------------------------------------Percent-------------------------------------Plantain 100.0 -- -- 60.0 33.3 6.7 36.8 31.6 31.6
Maize 100.0 -- -- 52.9 29.4 17.7 28.6 28.6 42.9
Groundnuts 100.0 -- -- 62.9 25.7 11.4 30.8 23.1 46.1
Yams 100.0 -- -- 68.2 25.0 6.8 20.0 20.0 60.0
Cassava 100.0 -- -- 67.7 22.6 9.7 -- -- 100.0
Cocoyam 100.0 -- -- 73.5 20.4 6.1 25.0 37.5 37.5
Percent of Marketing Agents Transporting Own Products Wholesalers Market Retailers Street Retailers
85.0 93.0 95.0
not all wholesalers are truck owners. These types of contractual arrangements would tend to explain the lower percentage of wholesalers who responded that they assumed the major responsibility for transporting their products.
The level and types of credit utilization and investment practices are important components of marketing activities, in the sense that they forge linkages between distribution channels. Some indication of these components are provided by the responses reported in Table 8. The table suggests,among other things, that credit in the traditional sense, played a relatively minor role in food crop marketing operations. It has long been assumed in Cameroonian agricultural circles that marketing agents obtained the bulk of their investment capital from savings and/or loans from marketing agents associations, banks and informal moneylenders. This assumption was not substantiated by the survey data. None of the wholesalers received investment capital from such associations, and only 7 and 11 percents of the market and street retailers, respectively, received such funds. Both wholesaler and retailer agents appeared to have received the bulk of their capital from personal and/or family savings. In general, however, wholesalers tended to rely less on this source of financing than the other two categories of agents. This lesser reliance on personal and/or family savings might have been directly related to their utilization of credit (albeit low), from banks and informal moneylenders. Neither of the two retailer groups utilized loans from these two sources.
About 50 percent of the wholesaler agents who applied for bank loans received such loans. On the other hand, none of the retailers had.applied for bank loans. Reasons for relatively poor utilization of bank credit varied significantly between wholesalers and retailers.
TABLE 8.--Market Operation Financing and Credit Utilization Among Marketing Agents,
Yaound6 Markets, 1977
Characteristics Wholesalers Market Retailers Street Retailers
Personal or family savings 50.0 93.0 88.9
Savings through an association -- 7.0 11.1
Bank loan 25.0 -- -Loan from another moneylender 25.0 -- -Use of Bank Loans
Have applied for bank loan 50.0 -- -Received the loan 50.0 -- -Reasons for not Applying for Bank Loan
Lack of sufficient guarantee 91.7 41.7 36.8
Does not like to be indebted 41.7 -- -Too many complications at the bank 70.8 69.4 -Not well informed 91.7 69.4 47.4
Business too small -- 27.8 15.8
Always prefer to deal in cash 83.3 75.6 100.0
Receive credit from my suppliers 16.6 15.6 29.0
Give credit to my customers 25.0 18.4 52.9
Make advance payment to my suppliers 58.3 55.3 79.4
TABLE 9.--Market Information and Price Determination Mechanisms Among Marketing Agents,
Yaound6 Markets, 1977
Characteristics Wholesalers Market Retailers Street Retailers
-------------- Percent-------------------------Market Information
Possess some knowledge of what is potentially available in particular location prior to
going to that location 100.0 100.0 100.0
Lacks knowledge of potential quantities available for purchase in locations 66.7 82.4 91.2
Lacks knowledge of general unit cost of
commodities in locations 85.2 78.7 87.5
Possess membership in trader's association 73.9 5.6 3.1
Reasons for not joining an association Lack of confidence 40.9 22.4 23.5
Dislike for associations 27.3 6.9 29.4
Lack of partners 31.8 12.1 47.1
Lack of information -- 58.6 -Selling Price Determination
Price determined by taking into account the quality of product, quantity available in the
market, and purchase and transport costs 100.0 100.0 100.0
Normally receive unit price charged in market 25.9 30.0 30.1
Price list available in market where products sold 100.0 100.0 100.0
Have knowledge of prices charged by competitors 14.8 21.3 16.6
The majority of wholesalers (92 percent) cited the lack of sufficient guarantee or information. In contrast, the majority of market retailers (69 percent) cited complexity of loan application and lack of information, as reasons for poor utilization of bank loans. The lack of appropriate guarantee and credit information were the major reasons for poor bank credit utilization by street retailers. All three categories of marketing agents exhibited a strong preference for cash transaction in their business operations, although the preference was considerably stronger among street retailers (Table 8).
There are no official market instruments or network to guide
marketing agents in locating supplies, customers, or setting prices. In the absence of such mechanisms, attempts were made to identify the mechanisms used in making such decisions. Table 9 presents selected responses of marketing agents regarding market information and price determination mechanisms. Interestingly, all three groups of agents (100 percent), claimed that they had some prior knowledge of the area-specific supply characteristics of the various commodities that they handled, prior to purchasing those commodities for later sale. However, a high percentage of all three groups of agents indicated a lack of knowledge of available quantities of commodities in these supply areasas well as unit price of these commodities. This characteristic would raise serious doubts as to the soundness of the supply information that was being transmitted to the agents.
Price Integration Among Markets
The January-December 1975 average annual price of the ten food
commodities in the nine Yaound6 markets was calculated as a percentage of the average Yaound6 price level (Table 10). Results indicated that there were significant variability in commodity prices among the urban
markets. Cassava (tubers) prices were highest in the Ngoa-Ekbl' market (market 7), and lowest in the Mvog-Mbi market (market 9). However, cassava flour prices, although also highest in the Ngoa-Ek6l6 market, were lowest in the Mokolo market (market 8). Plantain prices, on the other hand, were highest in Madagascar (market 5), and lowest in Melen (market 1) and Mokolo. It was only in the case of rice, beans, and onions that several markets simultaneously indicated similar price levels. In the case of beans, these markets were Nlongkak (market 2) and Mvog-Mbi (market 9), while in the case of rice they were Melen (1), Mokolo (8) and Mvog-Mbi (9). For onions, the markets were Madagascar (5) and Mvog-Mbi (9). There is no evidence of a consistent pattern of high or low price markets among commodity groups in the different markets.
Table 11 shows estimates of the relative price variation among a number of selected markets over the survey period. The table uses as a proxy for the coefficient of variation, the percentage of the ratio of the range to the mean,of all observations in all of the relevant markets. Given such a measure of relative price variation, it appears that the price of rice was the least variable among market (6.4 percent), while the price of onions varied greatly among markets (118 percent). None of the remaining eight commodities, except cassava flour and groundnuts, registered price variability greater than 50 percent.
Given the evidence of considerable price variability among food commodities in urban markets (Tables 10 and 11), the bivariate correlation procedure outlined in the methodological framework was used to evaluate the degree of price integration among the ten commodities in the nine Yaounde markets. Tables 12 and 13 summarize the correlation coefficients (p) from this analysis. One characteristic that is
TABLE 10.--Average Annual Prices of Selected Commodities, Yaound6 Markets, as Percent of Yaound6
Average Prices, January to December 1975
MarketsCommodities 1 2 3 4 5 6 7 8 9
Cassava (Tubers) 108.6 100.0 90.4 100.0 105.1 113.0b/ 113.2 98.9 88.8
Cassava (Flour) 108.2 96.8 87.4 97.1 100.0 -- 124.3 84.9 87.8
Cocoyams 103.5 99.6 96.2 96.9 100.3 102.9b/ 113.1 99.6 96.5
Plantain 92.6 102.3 99.7 96.0 109.0 94.9b1/ 103.7 92.6 100.3
Maize (Grain) 96.1 110.4 96.5 100.3 98.4 75.4sJ 98.7 104.0 101.6
Tomatoes 96.4 107.9 104.8 100.0 99.9 -- 107.7 94.0 97.8
Groundnuts 116.4 108.1 95.6 87.4 84.1 -- 97.4 94.7 105.8
Beans 99.4 100.3 98.4 97.6 98.4 -- 89.4 99.3 100.3
Rice 100.0 92.5 101.6 99.1 100.3 -- 100.9 100.0 100.0
Onions 88.4 115.6 103.9 96.6 100.5 -- -- 99.9 100.5
a/The different markets can be identified as follows: 1=Melen 2=Nlongkak 3=Central 4=Mvog-Ada 5=Madagascar 6=Abattoir 7=Ngoa-Ek61 8=Mokolo 9=Mvog-Mbi.
b/Averages based on 11 months observations.
Averages based on 10 months observations.
TABLE 1l.--Relative Price Variation in Selected Yaoundg Markets, January-December, 1975
Monthly Average Absolute Value of Relative
Observations Annual Unit Price Range Price a/
Market in Each Market Unit Price VariationCommodities Observations
--------Number------------ ---------CFA Francs/Kg-------- Percent
Cassava (Tubers) 8 88 19.8 7 35.3
Cassava (Flour) 8 12 58.9 40 56.9
Cocoyams 8 12 28.9 8 27.7
Plantain 8 12 29.9 7 23.4
Maize (Grain) 8 12 62.2 17 27.3
Tomatoes 8 12 125.7 41 32.6
Groundnuts 7 12 136.2 70 51.2
Onions 7 12 147.2 174 118.2
Beans 7 12 112.8 55 48.8
Rice 8 12 141.0 9 6.4
- These relative values were derived by dividing-the range by the mean, and then multiplying by 100. It is used as a proxy for the commonly used measure of relative variability, the coefficient of variation. The coefficient of variation is commonly measured by S/R, where, S=standard deviation, and R=mean. It should be noted that Jones (footnote 12) used only the range as measured of relative variability in his Nigerian study.
TABLE 12.--Frequency and Percentage Distribution of Correlation Coefficients Bqtween Yaounde
Market Prices for Selected Commodities, January 1975-December 1975-'
coefficients (p) Rice Beans Onions Groundnuts Tomatoes
Freq. Percent Freq. Percent Freq. Percent Freq. Percent Freq. Percent
.95+ -- -- -- -- -- -- -- -- -- -.90 .94 2 7.14 -- -- 4 19.05 -- -- 1 3.57
.85 .89 3 10.72 3 14.29 4 19.05 1 4.76 -- -.80 .84 5 17.86 3 14.29 4 19.05 -- -- -- -.75 .79 2 7.14 2 9.52 3 14.29 1 4.76 2 7.14
.70 .74 1 3.57 2 9.52 -- -- -- -- 2 7.14
.65 .69 1 3.57 1 4.76 5 23.81 -- -- 2 7.14
.60 .64 1 3.57 3 14.29 -- -- 1 4.76 3 10.72
.55 .59 2 7.14 2 9.52 1 4.75 1 4.76 2 7.14
.50 .54 1 3.57 1 4.76 -- -- 1 4.76 -- -Less than .50 10 35.72 4 19.05 -- -- 16 76.20 16 57.15
Total 28 100.0 21 100.0 21 100.0 21 100.0 28 100.0
-Information is for nine markets.
TABLE 13.--Frequency and Percentage Distribution of Correlation Coefficients Between Yaoundg Market
Prices for Cocoyams, Cassava (Flour), Cassava (Tubers), Plantain, and Maize (Grains),
January 1975-December 19751
coefficients (p) Cocoyams Cassava(Flour) Cassava(Tubers) Plantains Maize (Grains)
Freq. Percent Freq. Percent Freq. Percent Freq. Percent Freq. Percent
.95+ -- -- -- -- -- -- -- -- -- -.90 .94 -- -- -- -- -- -- -- -- -- -.85 .89 -- -- -- -- -- -- -- -- -- -.80 .84 1 3.57 -- -- -- -- -- -- -- -.75 .79 1 3.57 -- -- -- -- -- -- -- -.70 .74 1 3.57 3 10.72 1 3.57 -- -- -- -.65 .69 2 7.15 2 7.14 -- -- -- 4 14.29
.60 .64 -- -- 3 10.72 1 3.57 2 7.14 1 3.57
.55 .59 1 3.57 1 3.57 2 7.14 2 7.14 1 3.57
.50 .54 -- -- 1 3.57 1 3.57 -- -- 1 3.57
. 0 .49 14 50.0 17 60.91 22 78.58 14 50.0 18 64.28
Less than 0 8 28.57 1 3.57 1 3.57 11 39.29 3 10.72
Total 28 100.0 28 100.0 28 100.0 28 100.0 28 100.0
-Information is for nine markets.
TABLE 14.--Number and Percentage of Correlation Coefficients
Significant at the 5 and 1 Percent* evels, Selected
Commodities, Yaound6 Markets, 1975-'
Commodities 5 Percent 1 Percent
Number of Number of
Coefficients Percent Coefficients Percent
Cocoyams 6 21.46 3 10.70
Cassava (Tubers) 2 7.14 1 3.57
Cassava (Flour) 8 28.58 3 10.72
Plantain 2 7.14 -- -Maize (Grains) 5 17.86 -- -Rice 16 57.14 13 46.43
Beans 14 66.67 10 47.62
Onions 21 100.00 15, 71.43
Groundnuts 4 19.05 2 9.52
Tomatoes 10 35.71 5 23.81
Nine markets are included in the study.
immediately obvious from the data in these two tables is.that the correlation coefficients are generally low. None of the ten food commodities registered coefficients above 0.95, and only three had between one and four coefficients at or above 0.90. Onion was the only commodity with 19 percent of its coefficients at or above 0.90, and more than 57 percent at or above 0.80. Rice had more than 35 percent of its coefficients at or above 0.80, while almost 29 percent of the coefficients for beans were at or above 0.80. For groundnuts and tomatoes, 5 and 4 percent of the coefficients, respectively, were at or above 0.85 (Table 12). In the case of cassava (flour and tubers), plantains, and maize (grain), none of the coefficients were above the
0.75 level (Table 13).
Significant differences were registered among commodities in terms of the size of the bivariate correlation coefficients and their frequency distribution. Given our hypothesis that general price levels would tend to move together in economically integrated markets, the analysis suggests that the onion, rice and bean markets were the most economically integrated. Given this finding, how reasonable is it to believe that the true correlation coefficient is equal to zero (p=O)? Recall from our methodological framework that a test for independence among two distributions can be made by calculating the coefficient of correlation (p), and testing that it is equal to zero. Such a test was applied to the paired groups of commodities. For most of the commodities studied, 12 pairs of observations were used to compute the correlation coefficient, except in the case of cassava tubers, where 11 pairs were used.
Analysis indicated that the "critical values" of the correlation coefficients were 0.576 at the 5 percent level, and 0.708 at the 1 percent level, for all crops, except cassava (tubers). In the case of cassava, the critical values were 0.602 and 0.735, at the 5 and 1 percent levels, respectively. Thus, for an observed coefficient below these critical values, given the respective sample sizes, one would be led to conclude that the correlation between observed prices in the relevant markets were not significantly different from zero. Table 14 shows for each of the ten commodities, the number of correlation coefficients significant at the 5 and 1 percent levels, and their respective proportion, relative to the total number of coefficients computed for each commodity. Again, the table strongly suggeststhat the markets for onions were the most integrated, since prices in the various markets appeared to be moving in concert more
than any other crops. All coefficients for onions were significant at the 5 percent level, and more than 71 percent were significant at the 1 percent level (Table 14). Rice and beans had 57 and 67 percent, respectively of their coefficients significant at the 5 percent levels. These two commodities had 46 and 48 percent, respectively of their coefficients significant at the 1 percent levels. The commodities with the most disorderly price movements were cassava (tubers) and plantain. In either case, only 7 percent of the coefficients were significant at the 5 percent level. Only 4 percent of the cassava (tubers) coefficients were significant at the 1 percent level, and none of the plantain coefficients were significant at this level.
Summary and Conclusions
The two-fold objectives of the paper were to (a) evaluate selected structural and conduct parameters of Cameroonian urban food crop markets to provide insights as to how these parameters have impacted the overall efficiency of the urban marketing chain and
(b) evaluate the degree of price integration among selected urban food crop markets as a means of guaging the degree of economic integration among commodity markets. Structural analysis indicated that marketing agents fell into three major categories (a) wholesalers
(b) market retailers and (c) street retailers. These three types were significantly different from each other in terms of demographic characteristics and marketing behavior. Wholesalers provided the central link between rural and urban food crops markets. Furthermore, the wholesaler category was dominated by males, while the two retailer categories were dominated by females. Within each marketing group,
business ownership was primarily on a single-ownership basis, with the bulk of labor input being furnished by family labor. Marketing agents handled a wide variety of food commodities, suggesting product diversification rather than specialization within marketing channels. Given the highly perishable nature of most food crops, and the absence of storage facilities, product diversification might have been motivated by the desire to reduce the risk of major losses stemming from the handling of a single commodity. The wide variety of products handled and the large number of marketing middlemen would tend to produce conflicting price signals in the marketing channels, particularly in the absence of official grades and standards, market information and market regulations. It might very well be, that under the existing conditions of food crop marketing in Cameroon, a high degree of "cut-throat competition" is a natural characteristic of the market place.
Urban food supplies were obtained from widely dispersed geographic locations. Given the wide divergence in agroecological zones where food crops are produced, there is a tendency for wide divergence in unit supply price of commodities in the different areas. Here again, the absence of appropriate market coordinating mechanisms, grades and standards, market intelligence and regulations, resulted in a situation where conflicting price signals were simultaneously being transmitted
through the informal communication network. Price integration analysils of urban food crop markets in the city of Yaound6 revealed significant price variation among selected commodities among these markets. It was also evident that significant differences existed between average Yaound6 food crop prices and the average prices of similar commodities in rural markets. The magnitude of these differences were so large
that they could not be completely explained by the cost components of the services associated with delivering these products to urban consumers. Thus, significant intra-city price variation for food commodities must have been highly correlated to some of the structural and behavioral characteristics of the food marketing system.
Given the strong evidence of major structural and conduct bottlenecks to an efficient marketing system, and the associated impact on urban prices, attempts were made to assess the relative degree of price integration among urban markets with respect to selected commodities. It was hypothesized that the degree of concerted price movements (or the absence of such movements), among commodities in different markets, would reflect the relative degree of subsystem economic integration. This hypothesis was based on the argument that the higher the degree of economic integration the more efficiently would price signals be transmitted between marketing subsystems. Results of price integration analysis suggested that the marketing systems for onions, rice and beans were the most integrated, in the sense that the average monthly prices of these commodities in a number of urban markets tended to move in concert. None of the other commodities studied exhibited a similar characteristic. These findings raised questions regarding the possibility of structural and conduct differences in the marketing subsystems of these three commodities vis a vis those of the other commodities, which might have impacted the level of price integration.
A close scrutiny of the comparative marketing system for rice and plantain, another major food commodity, dramatize the impact of structural and behavioral characteristics on the level of price and economic integration among marketing subsystems. Rice, unlike plantains,
is produced in many diverse areas of the country. In general, however, the bulk of production occurs primarily on larger more commercially oriented farms, and usually under government production promotion schemes. Under government programs, large quantities of rice are produced in specialized areas, processed, packaged, stored and/or shipped to wholesalers, who in turn sell to retailers. It is official government policy to use scales and weights or standardized units of measurement in rice marketing in urban areas. Under such structural conditions, price signals are more efficiently transmitted in rice marketing subsystems, than in plantain and other food crop marketing subsystems where such a structure is absent. In the case of plantains and other food crops, production occurs under a diversified set of conditions, in terms of input and output characteristics. As such, the conditions conducive to uniform pricing are far less than those of rice; since there are no comparable market coordinating mechanism to guide prices to levels consistent with the economics of market conditions.
In all probability, the higher degree of price integration among
the onion and bean marketing subsystems is also a function of structural and conduct characteristics of these subsystems. In the face of severe shortages of vegetable commodities in the Yaound6 area over the 1974-1975 period, the national Food Development Authority (MIDEVIV), initiated a credit and input subsidy program for vegetable production in farm areas adjacent to the city of Yaound6. Financial support for this program was obtained through the National Fund for Rural Development (FONADER). However, Yaounde the primary market, became saturated at that time, largely as a result of a lack of adequate storage facilities and logistic bottlenecks in moving the surplus commodities to high
demand areas . The net result of these structural problems was that vegetable farmers were forced to absorb significant losses. The incident received extensive coverage by the newsmedia, which would suggest a relatively high level of market intelligence relating to these commodities. Since onions and beans represented a major portion of the commodities involved in the government production scheme, the observed level of price integration among the markets for these commodities could be a reflection of the higher level of information .flow in these subsystems.
-See footnote of Table 1 for definition of "Medium Income Families."
Percentage price increases for these selected food items in the
cities of Yaounde and Douala were computed from .
The Yaound6:trade area comprises the city itself with all its
markets and a belt extending 10 to 15 kilometers outside the city on
the main roads entering the city.
A/Each of the nine markets were visited an average of 2.5 times per month. For plantain, cassava (tubers) and cocoyams each marekt was
visited on the average of 30 times during the survey period, with an
average of 75 respondents permarkt. For cassava (flour), maize (grain)
and tomato each market was visited on the average of 30 times with an average of 74 respondents per market. For goundnuts, each market was
visited on the average of 29 times with an average number of 74
respondents per market. For beans the corresponding averages were
28 and 71. For onions and rice the corresponding averages were 29 and 73;
and 28 and 70, respectively.
 Joe S. Bain, Industrial Organization, John Wiley and Sons, Inc.,
New York, 1959.
 Groupe de Conseillers en D6veloppement d'Afrique Centrale,
Perspectives de D.veloppement Intdgrd de la Province de Centre
Sud-Syntheses et Annexes, Minister6 de 1'Economie et du Plan,
 William 0. Jones, The Structure of Staple Food Marketing in
Nigeria as Revealed by Price Analysis, Food Research Institute
Studies, Stanford, 1969.
 Uma J. Lele, Food Grain Marketing in India: Private Performance
and Public Policy, Cornell University, Ithaca, New York, 1971.
(5] Minister6 du Plan et de l'Am~nagement du Territoire, Note
Trimestielle de Statistique, Republique Unie du Cameroun,
Yaound6, various issues, 1970-1976.
 Winfried Muziol, R6sultats de 1'Engutte Sur la Commercialisation
Traditionnelle des Produits Vivriers, Food Development Authority
(MIDEVIV), Yaound6, 1976.
 Jean Ongla, An Economic Survey of Food Production Variables in
the Zone of Intervention: Yemessoa, Department of Rural Economy,
National Advanced School of Agriculture, University of Cameroon,
Yaound6, Research Report No. 1, 1973.
18] Jean Ongla, "Structure, Conduct and Performance of the Food Crop
Marketing System in Cameroon: A Case Study of Yaound4 and Adjacent
Areas," Ph.D. dissertation, University of Florida, 1978.