Determinants of Highland Bamboo ( Yushania alpine ) Culm Market Supply in Semen Ari District, South Omo Zone of Southern Ethiopia

,


Introduction
Forest resources yield both wood and nonwood forest products (NWFPs) in Ethiopia. NWFPs are important and valued resources in the country since they signifcantly improve the livelihoods of people by boosting household income and food security [1]. Te most important nontimber forest products (NTFPs) found in Ethiopia are cofee, honey, wax, bamboo, reeds, natural gum, and resin [2]; among those, the most important NTFPs such as forest cofee, honey, beeswax, spices, bamboo, herbal cures, gums, and resins are the most commercial ones in the country [3]. One of the key NTFPs being recognized as a potential species for environmental preservation and poverty alleviation is bamboo [4,5]. estimated to contribute an annual income of ETB 56,250,000 to the national GDP [3]. It has recently shown great commercialization potential in Ethiopia [13], and small-scale modern bamboo product processing workshops are expanding in big cities of the country [14].
Te bamboo items that were given to the market were prepared and sold individually by each producer from the origin of production to the destination of sale. However, they were distributed through various marketing intermediaries from producers to consumers. Te major actors in the bamboo value chain are producer farmers, village-level traders, town and city wholesalers, small-scale and medium-scale bamboo processing and marketing frms, and consumers [15]. Te development of the bamboo value chain is at its infant stage, and it is practiced in a traditional way in which the farmers have the least marketing margin, the marketing system is relatively simple, and there is no large commercial need for the increased supply [5].
Production and use of bamboo in Ethiopia is underdeveloped, in contrast to other nations, and its socioeconomic and ecological potentials have not yet been fully realized [14,15]. Despite few investigations, there is still a paucity of factual proof about bamboo's economic impact on the local, regional, and governmental levels. Te majority of bamboo's current usage in Ethiopia is restricted to its traditional applications, which include hut construction, fencing, beehive construction, and the manufacture of home and ofce furniture [4,5,16].
Tere are persistent issues with the expansion of the bamboo industry in Ethiopia, including a lack of understanding of the socioeconomic signifcance of the country's bamboo resources among stakeholders and a lack of connections between policymakers and other organizations [12,17,18]. Likewise, the country has a weak bamboo value chain [9,13], and the ones that are already in place have poor vertical and horizontal integration [15]. Te main obstacles to the Ethiopian bamboo value chain are the absence of market information, uneven laws and regulations, and the location of craft stores near the resource [3]. Only one-third of Ethiopia's entire production of bamboo is made accessible to the market; the remainder is used by households [5,15,16,19]. Te socioeconomic and exogenous elements at the household level, such as the size of the leased land area, the number of household members, market access, knowledge, and livestock ownership, had the greatest impact on the contribution of bamboo to household income [15,20].
According to the Woreda agricultural development offce report in 2021, more than 900 hectares of land is covered by highland bamboo in Semen Ari district, and most of the people in the area use highland bamboo for diferent purposes such as house construction, fence, and fuel wood. Despite being important to many farmers' livelihoods and serving as a fexible crop and a source of income at the research site, highland bamboo has not received the attention it deserves. Additionally, there is little to no empirical data on the variables that afect the availability of bamboo items on the market. Tus, this study was carried out to provide empirical evidence regarding the factors that afect the supply of bamboo products to the market.

Description of the Research Site.
Te study was conducted in Semen Ari district of South Omo, Southern Ethiopia. Semen Ari district is one of the 11 districts of South Omo zone. Te total land area of Semen Ari district is 28,123 hectares. Altitude ranges between 1,470 meters to 3,300 meters above sea level (m.a.s.l). Te administrative seat of the District is Gelila Town located 602 kilometer southwest of the national capital and 356 kilometer from the capital of Southern Ethiopia. In terms of traditional agro-ecology classifcation, the district is classifed into four agro-ecology zones such as dega (50.3%), woina dega (8.3%), kolla (37.1%), and wurch (4.3%). Average annual rainfall and temperature in the district varies between 876 mm and 1763 mm and 11°C and 22°C, respectively. According to Central Statistical Agency (CSA) [21], the total population of the Semen Ari district is estimated to be 85,483 (41,884 male and 43,889 female). More than 75% of the overall population lives in rural regions, and crop-livestock farming is the predominant mode of subsistence in the research site. Te map of the research site is shown in Figure 1.

Data Types and Source.
Primary and secondary data sources were used to gather both qualitative and quantitative data types. Household surveys, focus group discussion, key informant interviews, and feld observations were used to gather primary data from bamboo farmers. Te main information gathered from bamboo farmers includes details on household demographics, socioeconomics, land characteristics, relative income dependence on highland bamboo, income from other forest products, actors in the highland bamboo production-to-consumption, and variables infuencing bamboo production, and market supply of its products. Secondary data sources used were journals, pertinent textbooks, bulletins, and reports from various governmental organizations, zonal and district ofces of environmental protection, and forest management. All the secondary data gathered were examined and used to supplement the primary data.

Sampling Technique.
Te study site was purposively selected based on the bamboo production and availability. Multistage stage sampling techniques were employed to draw sample household heads. In the frst stage, 15 potential kebeles in bamboo production were identifed based on district information and, of those, three kebeles were randomly selected, where a kebele is the smallest administrative unit in Ethiopia. In the second stage, the number of sample households from each sample kebele was determined from the recent lists of households using a probability proportional to the size sampling technique. Given the relative homogeneity of households in terms of their socioeconomic characteristics and livelihood styles, bamboo-growing sample households were drawn using a simple random sampling method from each kebele.
where the sample size for a limited population is n. P is the population reliability (or frequency calculated for a sample of size n), where p is 0.5 which is taken for all developing countries population, and p + q � 1, and e is the margin of error considered which is 7% for this study. N is the size of the population, which is the total number of households. Z α/2 is at the 0.05 level of signifcance, and the normal reduced variable Z is equal to 1.96. Te total households in the 15 bamboo potential kebeles are 3022, and from this, 2775 household produce bamboo.

Methods of Data Collection.
All the pertinent primary data were collected using formal and informal procedures. Observing the market for bamboo products in the research site through a transact walk, conducting focus group discussions with members of predetermined social groups (elders, women, producers of bamboo, nonproducers of bamboo, traders, processors, consumers, and people with various levels of expertise) and conducting key informant interviews with residents of the research site who have lived there for a long time (the key informants were chosen using the snowball method) were among the informal data collection tools used. For household interviews, semistructured questions and checklists were employed. After completing a pretest with fve sample subjects from chosen sample kebeles (kebele means the smallest administrative unit in Ethiopia), International Journal of Forestry Research the questionnaire was modifed as needed. Based on their educational background and knowledge of the local language, four enumerators were chosen and provided by the researcher on the contents of the questionnaire for enumerators. Ten enumerators managed data collection with the intensive supervision of the researcher.

Data Analysis.
Te collected data were analyzed using descriptive statistics and multiple linear regression analysis by using STATA software version 14.2.

Descriptive Analysis.
Descriptive statistics such as mean, standard deviation, frequency, percentage, minimum, and maximum values of variables were computed to describe bamboo production in the research area.

Econometric Analysis.
Multiple linear regression models were used to identify the factors that afect the supply of bamboo products to the market.
(1) Model Specifcation. Te multiple linear regression (MLR) model was used to investigate the variables that infuence supply to market. Because the quantity of bamboo culms supplied to the market by growers were the dependent variable and continuous variables, an OLS model was applied [23]. Te details of the OLS regression are as follows: 6 Annual income β 7 Extension contact + β 8 land covered by bamboo + β 9 Market information + β 10 Market di stance + β 11 Price of Bamboo + Ui, (2) where Y is the quantity of bamboo culm supplied to the market, β o is the intercept, β 1−11 is a vector of estimated coefcients of the explanatory variables, and U i is the disturbance term.

Descriptions of Variables and Hypothesis
(1) Dependent Variable. Marketed supply of bamboo culms which is a continuous variable refects the quantity of bamboo culms that each household really sells on the market. Table 1 describes variables, measurement, and expected signs in the multiple linear regression (MLR) model.

Household and Socioeconomic Characteristics of
Respondents. It is crucial to understand the essential characteristics of the sample bamboo growers. Consequently, the analysis considered factors such as marital status, market information sources, extension services, family size, experience with bamboo production, land ownership, age, sex, family size, and educational background of the respondents.
Out of the total sample households, 86.3% and 13.7% of respondents were male and female-headed, respectively. Te majority of the respondents (85.7%) were married, followed by 8.7%, 4.9%, and 0.5% of widowed, divorced, and single household heads, respectively ( Table 2). One of the crucial aspects of the household characteristics is age as it afects the general state of health in the community; it has an impact on the productivity of the people.
Te respondents' maximum age was 69, while their minimum age was 21. Te average family size of the entire sample of respondents was 7, ranging from 2 to 15, and this might help the families participate more efectively in the production and sale of bamboo (Table 2). Te respondents are 44 years old on average, with a 12 year standard deviation.
Te other characteristic of the sample households is the education level that the respondents attended in a number of years in school following the country curriculum. Te educational background of the household heads is believed to be an important feature because as the number of grades to be attended increases, the household head may enable to accept new ideas and innovations such as various valueadding activities. Teir average education level was grade 3 with a minimum of illiterate (i.e., unable to follow formal education) and a maximum of grade 7. Education and training increase work efciency, which boosts income and food security. Additionally, education is crucial for managing a business because it afects one's capacity for decisionmaking [24].

Land Allocation.
In the research area, households use their land mostly for food and cash crop production, animal grazing, bamboo production, and tree plantations. Te respondents' average landholding size was 3.94 ha (i.e., between 0.625 and 10 ha), which is greater than the 1.25 ha national average for holding size per household [21]. Averaging 2.645 hectares out of the total holding, the respondents allocated the bulk of their plots for the production of agricultural crops. Bamboo plantations (0.447 ± 0.277) and eucalyptus plantations (0.651 ± 0.293) are spread across the remaining land holding (Table 3). Tis suggests that in comparison to the size of the ordinary landholding, the average quantity of land set aside for the cultivation of bamboo is quite tiny.

Extension Service and Market Information.
Te provision of adequate extension services for the communities improves the socioeconomic development of the communities generally and the wellbeing of the people in the communities specifcally. Agricultural and natural resource ofces at the zonal and district levels are in charge of ofering extension services. In fact, Table 4 shows that of the total sample of bamboo producers, 14.8% of the farmers reported having extension services throughout the 2021 growing season, while 85.2% of the farmers said they did not. All producers and traders are able to openly communicate with one another while determining pricing with the aid of a well-organized market information system. Families that have access to trustworthy market data can choose from a variety of transaction methods that will facilitate them to sell their excess bamboo and gain a variety of benefts [7]. In order to systematically gather, analyze, and distribute information pertinent to the needs of various actors, there is no structure in place, as is evident from a closer examination of market information. Despite this, more than half (56.8%) of sample respondents said they had market data from several sources (Table 4). Te major sources of information were provided by traders, friends, and other market participants, with proportions of 53.6%, 6.6%, and 39.9%, respectively. Tis indicates that the use of alternative sources such as radio and television is lacking.

Highland Bamboo Supply Chain in the Semen Ari
District. Te market chain of bamboo products in Semen Ari highlighted the involvement of diverse actors who participated directly in the chain. According to Faida [25], the direct actors are those involved in commercial activities in the chain (i.e., input suppliers, producers, traders, and consumers). Bamboo value chain actors in the study area include bamboo producers, traders, bamboo product processors, and product consumers and as well as considered to show the relationships and integrations of the processes and activities performed along the chain. Tus, the highland bamboo supply chain is responsible for the production and distribution of bamboo culms and products in the area as presented in (Figure 2).

Determinants of Bamboo Culm Market Supply.
Analyses of factors afecting the quantity of bamboo culm supply were found to be important for policy implications  International Journal of Forestry Research 5 on bamboo commercialization in the research site. In the research site, highland bamboo is produced mainly for the market and home consumption. It is one of the highest cashearning NTFPs for producer households. According to the study fndings, all the sample households supplied bamboo culms to the market during the survey period. Before running the OLS regression model, all the hypothesized explanatory variables were checked for the existence of multicollinearity and heteroscedasticity problems. Te variance infation factor (VIF) and contingency coefcients (CC) were used to check the existence of multicollinearity problems among continuous explanatory variables and dummy variables, respectively. Te results for all VIF values ranged between 1.04 and 6.82, which is less than 10. Likewise, the values of CC ranged between 0.165 and 0.328, which is less than 0.75, and hence, multicollinearity was not a serious problem both among the continuous and discrete variables. Te existence of heteroscedasticity was detected by the Breusch-Pagan-Godfrey test. Because the Prob > chi2 � 0.0033 or (p � 0.0033), i.e., the small p value (the lowest signifcance level) rejects the null of homoscedasticity, there is heteroscedasticity in the explanatory variables. Te problem of heteroscedasticity was solved by undertaking robust ordinary least square (OLS), and hence, all the explanatory variables were included for the model analysis of factors afecting the market supply of bamboo culm.
Te model was statistically signifcant at a 1% level of probability, as presented in Table 5, demonstrating the model's suitability for explaining the correlations between the predicted variables. Te coefcient of determination evaluates the general goodness of ft of the regression model (adjusted R 2 ). It tells what proportion of the variation in the dependent variable, or regresses, is described by the explanatory variable. Te number of explanatory variables is not very modest in comparison to the number of observations; adjusted R 2 has a tendency to present an unduly optimistic view of the regression's ft [26]. Te overall model goodness of ft, represented by model count R-square, is good (76.27%). Tis result indicates that about 76.27% of the variation in the market supply of bamboo culm was attributed to the hypothesized variables.
Since all of the respondents included in this study delivered their bamboo culms to the market, factors afecting the supply of bamboo culms to the market were estimated using the OLS model. Only 8 out of the 11 variables (market information, extension services, allocated land for bamboo production, experience selling bamboo products, proximity to the market center, family size, sex, and age of household head) were found to be signifcantly afecting the household marketable supply of bamboo culms at the household level, as presented in Table 5. Tese fndings are discussed in the sections that follows.

Age of Household Head (AGEHH).
At the 5% level of probability, the age of the respondent had a positive and statistically signifcant relationship with the amount of bamboo culm delivered to the market. Tis means that older people contribute more than younger people do. Te amount of bamboo culm ofered for sale raises by three culms as the age of the respondent increases by one year while holding other variables constant. Tis is in line with the fndings of Dessie et al. [27], who revealed a positive relationship between age of the household head and the commercialization of eucalyptus in Wogera district,

Actors Growers Traders Processors Consumers
Flow bamboo and its products  Northern Ethiopia. Tey concluded that due to the laborintensive nature of the forestry industry, older households have a strong propensity to participate more in the production and selling of forest products. In contrast to this study, Fayera et al. [18] revealed that in North Western Ethiopia, as age of household head increases, the number of bamboo culms supplied decreases by 0.091 culms, and they indicated that as age of household heads increases, fewer bamboo culms are being ofered for sale. In the Illu Ababora zone, western Ethiopia, Asfaw and Lema [28] found that the possibility of households participating in NTFPs marketing was inversely connected with the age of household heads engaged in NTFP activities. Te variable was predicted to have a detrimental impact on the amount of bamboo culm delivered to the market.

Sex of the Respondent (SEHH).
Te sex of the respondent was discovered to have a positive and signifcant impact on the market supply of bamboo culms at a 1% level of signifcance. Te good news is that when all other factors are held constant, having a male head of household greatly improves the amount of bamboo delivered to the market by 69.8 culms. Traveling a distance to sell culms is the common practice in the research site, and this may encourage maleheaded households to be more engaged in bamboo marketing. Tis supports Chukwuone [29] result that the gender of the respondent had a positive and signifcant impact on the market supply of NTFPs. Te authors stated that as females often do not own land property, as a result of this, males are more likely to cultivate NTFPs than females. Tis will afect the amount of NTFPs that they supply to the market compared to male-headed households. In the research site, male-headed households have better economic power, land size, extension contacts, and market information compared to female-headed households. Te sex of the household head had a positive and considerable impact on the quantities of papaya sold into the market [30].

Family Size Household (FAMSZ).
Te result demonstrates that family size has a positive relationship with the quantity of bamboo culm supplied for sale at a 5% signifcance level. An increase in family size by one individual increases the quantity of bamboo culm supplied for sale by 20.95 culms, keeping the other variables constant. Tis is because families with more household members tend to have more labor, which in turn increases bamboo production and then increases bamboo culm market supply. Tis outcome is consistent with Consult [6], who indicated that household participation in bamboo commercialization is positively correlated with family size. Te study indicated that large family size is probably a factor owing to the presence of sufcient labor to transport bamboo culms from distant forests, which is required in excess for household consumption, and such families may be motivated to sell what is left over from their consumption. Te outcome is consistent with research by Fikir et al. [31], who suggested that family size afects households' gross revenue from forest products positively.

Land Covered with Bamboo (LACWB).
Te result confrms the expected positive trend by demonstrating that land set aside for bamboo has a substantial impact on bamboo sales at a 1% signifcant level. Te increase in area allotted to bamboo production is implied by the positive coefcient for this factor, which means that more culms will be available for sale. By increasing the bamboo production area by one hectare while holding all other variables constant, it was possible to raise the farm-level marketable supply of culms by 316. Te idea is that if the farmers have a bigger area, they will have a better chance of allocating bigger areas for bamboo production and, as a result, will be able to produce a lot of goods that are sold on the market. Tis supports Toma Dilebo [32] results that cofee-growing land allocation had a positive impact on the amount of cofee provided to the market. Additionally, Kassahun and Emana [33] presented a comparable fnding, demonstrating that a 0.1-hectare increase in bamboo production area led to an increase in the farm-level marketable supply of culms by 285.

Experience of Household Head (EXPHH).
Te result demonstrated that bamboo product selling experience has a signifcant efect at a 1% signifcant level on the number of bamboo culms to be sold with an expected positive sign. Te result suggests that as producers have high bamboo culm production and selling experience, the number of bamboo culms supplied to the market increased as expected. One year of experience in bamboo culm selling and production increases the number of bamboo culms supplied for sale by 15.17 culms, citrus paribus. As a result of having access to more marketing networks and information, producers with more expertise in the sale of bamboo culms have a tendency to sell more bamboo culms on the market than producers with less experience. Marketing skills and experience are essential for buyers and sellers of NTFPs, and experience is an important asset for traders who remain surprisingly faithful to particular products [34]. Bamboo culm producing and selling experience signifcantly and positively afects the number of bamboo culms supplied to the market and households engaged in bamboo marketing for a longer time supplied more culms than new harvesters who recently started bamboo marketing [18,33]. Te study by Dessie et al. [27] shows that a household experience with Eucalyptus woodlot production gets more commercial benefts from it. Tis result was confrmed by the studies by Shumeta et al. [35] and Ayelech [36], as selling and production experience afected the amount of forest cofee and avocado delivered to the market positively.

Market Distance (DSNMRK).
It was found to have a negative and signifcant efect on the amount of bamboo culm supplied to the market at a 5% level of probability. As the market distance increased by one walking minute, it resulted in a decrease in the amount of bamboo supplied to the market by 1.5 culms, citrus paribus. Tis is because as the market distance increases, transportation costs also increase. Te longer the market distance, the higher the transportation charges increase for walking kilometers, and other International Journal of Forestry Research marketing costs. In line with this result, Kassahun and Emana [33] demonstrated that market distance in the Dawuro zone signifcantly and negatively afected the amount of bamboo culm delivered to the market. A similar issue was studied by Erifo and Amanuel [17] on the amount of bamboo delivered to the market in the case study of Gedeo Zone, which indicated a negative relationship between the market distances from the respondent's home to the bamboo market and the amount of bamboo culm supplied. Te results verifed that an increase in distance from the market place resulted in a decrease of 44 culms per year per household. Market distance showed a signifcant and negative correlation with the degree of bamboo commercialization in Awi, Sidama, and Sheka [6]. Again, Ayelech [36] indicated that market distance caused a market surplus of avocados to decline in the Gomma district.

Extension Services (AEXSER).
At a 1% level of probability, extension services had a positive and signifcant efect on the market supply of bamboo culms. Tis indicates that producers who get more knowledge during contact with the extension agents for bamboo production methods contribute to increase the amount of bamboo production. Getachew [37] and Ayantu et al. [38] indicated that cultivators who have extension services for honey production afect the farm-level market supply of each honey signifcantly and positively. Development of infrastructure for linking resources, consumer centers, and extension education among producers may enhance the commercial engagement of producers and improve the accessibility of bamboo resources for commercial production [6]. According to Fikir et al. [31], households having an extension of contact are more likely to earn higher forest income and to be more dependent on forest income. Similar to this fnding, Gebrelibanos [39] indicated that the frequency of extension contact has a positive and signifcant efect on acquiring new information or adopting a certain technology.

Market Information (AMI).
As hypothesized initial, the market information afected marketable supply of bamboo culm positively. As bamboo producers get better market information, the number of bamboo culms supplied to the market increases by 113 culms, citrus paribus. Market information helps to acquire new ideas and information related to the bamboo market and increases bamboo culms supplied for sale.
According to Marshall [34], efective coordination and connectivity among producers and processors of nontimber forest products results in better product quality and quantity, more economical transportation, and improved negotiation power. Kassahun and Emana [33] and Erifo and Amanuel [17] reported that market information infuenced the marketable supply of bamboo culm in the Dawuro and Gedeo zones, respectively. Similarly, in Homosha district, when the harvesters have better market information, they have a better probability of getting better prices, which in turn increases the market supply of bamboo culms in the district [18]. In contrast to this fnding, Shumeta et al. [35] found a negative and signifcant relation between market information and the amount of forest cofee supplied to the market. Tey found when the farmers have market information, the amount of supplied forest cofee to the market decreases by 0.4 kg. According to Carr and Hartl [40], forest dwellers have no awareness of how much a consumer in a city or a developed country will pay for the fnished product because they lack market information. Tey also have limited or no ways of negotiating for a larger share of increasing earnings.

Conclusion and Implication
Te study revealed that most of the households engage in bamboo production and marketing as one of the main livelihood practices in Semen Ari district. Te main production location is the homestead and nearby farmlands. Moreover, bamboo has an important function as a coping strategy to fll income and subsistence gaps of diferent categories of people for cash and subsistence. Bamboo is perceived as highly valuable in the Ari culture above all because of its hand in domestic/traditional applications. Te major use of highland bamboo is for fencing, house construction, fodder, local grain storehouse "Gotera," basketry, foor mats, fuel wood, agricultural tools, beehives, furniture, granary, and household utensils. In addition to being used for subsistence, bamboo is sold on the market and is transferred from the producer to the customer through a variety of marketing intermediaries. Producers, traders, processors, and consumers were the main players in the marketing of bamboo. Estimation of the determinants of marketable supply of bamboo using robust OLS regression analysis was employed using eleven hypothesized variables. Only eight variables such as market information, extension service, allocated land for bamboo production, the experience of household in selling bamboo products, market distance, family size, sex, and age of household head were found to be signifcantly affected the supply of bamboo to the market. Tus, enhancing the production and productivity of bamboo products through extension, regular training for farmers concerning the processing of bamboo products, improving the relationships of value chain actors, and improving infrastructure could increase producers' income as well as the marketed supply of highland bamboo.

Data Availability
Te data used for the research was added to the manuscript and published in the paper.

Conflicts of Interest
All the authors declare that they have no conficts of interest. 8 International Journal of Forestry Research