The analysis of traders in a developing country value chain: Pig traders in Uganda

Nadhem Mtimet and Derek Baker

23rd annual International Food and Agribusiness Management Association (IFAMA) forum and symposium

17-19 June 2013, Atlanta, GA

The analysis of traders in a developing country value chain: Pig traders in Uganda

Outline

1. Study of value chains, and trader functions

2. Sampling approaches

3. Experience and results from Kampala, Uganda 4. Next steps, including Symposium on trader sampling at African Association of

Agricultural Economists’ Conference, September 23-25, 2013, Hammamet, Tunisia

Background to research

Traders perform valuable value chain functions that are:

• often not recognised

• rarely quantified

• constrained by unknown factors

• conducted in isolation from

• officialdom

• large scale agribusiness

• collective actions

• the aid community

The research for which this is a preliminary presentation has sought to:

1. Better understand and quantify traders and their actions

2. Test approaches to sampling

3. Identify possible future sampling strata

Studying traders requires that:

• they can/want to be found

• they co-operate in divulging information

• they see a benefit in participation in research

• variation amongst traders is reflected by sampling

Trader sampling

Literature review: 3 common sample situations:

• No information at all about the sampling strategy and how respondents

have been selected (Ajala and Adesehinwa 2007, Jabbar et al. 2008, Loc

et al. 2010, Hap et al. 2012, MacFayden et al. 2012)

• Researchers report random trader selection but without much

explanation and detail (Bista and Webb 2006, Abdulai and Birachi 2009,

Kocho et al. 2011, Minten et al. 2013).

• Detailed information is provided about sample selection (Rab et al. 2006,

Wanyoike et al. 2010, Aoudji et al. 2012, Lagerkvist et al. 2013)

Traders are often surveyed

Sampling frames

Source of list for

sampling frame Advantage Disadvantage

1 Local producers Targets correct

commodities/products

Relies on another sample

Favours buyers

2 Local retailers Relies on another sample

Favours sellers

3 Local

processors/focal

processing facility

Targets correct

commodities/products

Can rely on population responses

Open to strategic response

May not correctly target products

May exclude non-locals

4 Local authorities Simplicity

Open to strategic response

Unlikely to correctly target

products

Excludes non-registered traders

Excludes non-locals

5 Word of mouth:

other traders

Simplicity

Allows “snowball” tracking Limits to knowledge

Open to strategic response 6

Word of mouth:

experts

Simplicity

Links to geography, infrastructure,

commercial interests

Ugandan study

Action:

• workshop-type survey activity

Situation:

• an unknown number traders around Kampala, apparently including

Mukono location

• active trading in both grown pigs and piglets

• some observed vertical integration of traders

• no information on transaction mechanisms, seasonality, margins,

price-quality incentives, services used, food safety and hygiene

practices, future plans,…, nor constraints faced

Sampling frames used: lists of traders

• from a sample of local producers

• from a sample of local retailers

• from local authorities.

Detail of traders’ sampling, by sampling source*

Sampling source Number of traders

contacted Number of

traders who

participated to

the workshop

Percentage

Farmers/producers 22 16 73%

Retailers 28 11 39%

Local authority 18 6 33%

*

14 traders belong to 2 different sampling sources.

None belong to all three lists

Summary statistics

Variables

Groups

Statistical

tests

Group 1

Sourced from

retailers or

producers

(n1=16)

Group 2

Sourced from

local

government

(n2=6)

“Young” “Experienced” Age (years) 28.19 42.17 10.691a*** Experience (years) 5.31 13.00 10.045a*** Piglets trading (%) 75% 17% 2.478b** Purchase from group of producers (%) 81% 33% 2.149b**

Taxes payment (%) 50% 100% 2.171b**

a t-test; b Z-test ***, ** : statistically significant respectively at 1% and 5% levels

Constraints reported by traders

In day-to-day buying operations

In day-to-day selling operations

Conclusions

Different approaches to sampling frame yielded different samples of traders

The samples generated exhibited:

• different characteristics, able to be assigned and named

• different functions

• different statements about constraints faced

Main constraints reported:

Few traders appeared on more than one sampling-frame-basis list

Concerning buying activities Young Experienced

Lack of working capital

Transport cost or quality

Seasonality

Low productivity

Poor animal health

Inappropriate animal feeds

Concerning selling activities Young Experienced

Lack of customers

Competition between traders

Unpredictable market conditions

Bad debts

Animal disease

Low prices

Symposium: September 2013

Sampling people that don’t stand still:

Targeting traders as key elements of value

chain function and performance, and how

they can be sampled

African Association of Agricultural Economists’ Conference

September 23-25, 2013, Hammamet, Tunisia

Sponsored by PIM

Contact:

Nadhem Mtimet [email protected]

Derek Baker [email protected]

International Livestock Research Institute www.ilri.org