Commercialization of Papua New Guinea's vegetable sector: Identifying constraints using quantitative, qualitative, and large language model methods

Peixun Fang, Iga Anamo, Harry Gimiseve, Glen Hayoge, Rishabh Mukerjee, Emily Schmidt and Xiaobo Zhang

No 7, Other briefs from International Food Policy Research Institute (IFPRI)

Abstract: This paper evaluates the constraints within Papua New Guinea's vegetable sector, drawing on large scale household surveys, extensive qualitative interviews with key stakeholders, and Large Language Model (LLM) methods. Our survey data reveal that vegetable production is ubiquitous, with almost all households surveyed (91%) growing at least one vegetable. Indigenous varieties, such as leafy greens (96%) and fresh beans (69%), are widely cultivated across regions, while high-value vegetables like onion (17%) and tomato (19%) show more regional concentration. Over half (53%) of PNG vegetable farmers sell their produce, with farmers located in the nonseasonal highlands agro-ecological area leading in market participation (66%), contributing to an overall commercialization rate of 24% (defined as the share of pro duction sold). However, modern input use, particularly improved seed adoption, significantly lags behind output commercialization in all agroecological zones except the islands survey areas. The quantitative analysis, using the PNG Rural Household Survey 2023, and the qualitative analysis, drawing from both manual review and LLM-assisted processing of in-depth interview notes, consistently identify poor feeder roads as critical bottlenecks for every stakeholder across PNG's vegetable value chain. Beyond transportation, these interviews repeatedly highlighted persistent seed supply shortages and high seed costs as significant hurdles. Since PNG depends on imported vegetable seeds, several structural and procedural barriers contribute to these shortages. These include potentially arduous quarantine procedures, limited foreign exchange for seed imports, and cumbersome permit processes, often leaving major distributors with insufficient stock. These reported bottlenecks may contribute to the low improved seed adoption and use reflected in the household survey analysis.

Keywords: vegetables; large language models; commercialization; household surveys; artificial intelligence; Papua New Guinea; Melanesia; Oceania; Asia (search for similar items in EconPapers)
Date: 2025-07-08
References: Add references at CitEc
Citations:

Downloads: (external link)
https://hdl.handle.net/10568/175558

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:fpr:othbrf:175558

Access Statistics for this paper

More papers in Other briefs from International Food Policy Research Institute (IFPRI) Contact information at EDIRC.
Bibliographic data for series maintained by ().