An interactive companion to a district-level study of land fragmentation, irrigation access, and agricultural production sufficiency — built from the official tables of the National Sample Census of Agriculture (NSCA) 2021/22.
For the majority of Nepalese holdings, own production does not cover household consumption. This study asks what kinds of districts those holdings sit in — and which structural feature, water access or parcel fragmentation, best predicts the gap.
A district-level dataset is built by merging the NSCA 2021/22 tables on parcel structure, irrigation, sufficiency, food-insufficiency duration, and six controls. Because the outcome is a bounded proportion in [0,1], the analysis moves from descriptive and weighted-correlation evidence to a weighted-OLS benchmark and a fractional logit with province fixed effects and mean-centred interactions.
Headline result: irrigation coverage is the dominant correlate of lower insufficiency — a roughly 20-percentage-point predicted gap between low- and high-irrigation district profiles, with bootstrap intervals that exclude zero. Fragmentation's bivariate disadvantage is absorbed by geography once fixed effects and controls are included, and the fragmentation–irrigation interaction is statistically null.
Holdings-weighted means across all 77 districts of Nepal (Thesis Table 4.1). The 52.5% irrigation figure weights district area shares by holdings; the national area aggregate cited in Chapter I is 54.5%. They use different weights.
True district-level shading from NSCA 2021/22 (Thesis Annex Table). Toggle between insufficiency, irrigation, fragmentation, and the Mountain/Hill/Terai belt — hover, click to pin, or search a district to read its full profile.
Hover any district for a quick read; click to pin it open in the side panel. Use the search box to jump to a specific district. Data: NSCA 2021/22.
| # | District | Province | Belt | Insuff. | Irrig. | Parcels | Severe |
|---|
At provincial resolution the inverse pattern is already plain: irrigation coverage rises, insufficiency falls. The Mountains-to-Terai gradient does most of the work.
Madhesh combines the highest weighted irrigation coverage (74%) with the lowest insufficiency (43%); Karnali combines the lowest irrigation (29%) with the highest insufficiency (70%). Fragmentation runs the opposite way: highest in Karnali (3.35 parcels per holding), lowest in Koshi (2.20).
All five highest-insufficiency districts are mountain or hill. All five lowest are in the Terai plains.
Holdings-weighted correlations with the share of holdings reporting insufficient own production (Thesis Table 4.5). Green bars show negative associations; terracotta bars show positive associations.
The irrigated-area share correlates at −0.86 with insufficiency. Fragmentation correlates positively at +0.40, but most of that signal is absorbed by geography in the multivariate models. Fast market access enters with a counter-intuitive positive sign — a compositional story about urban, part-time holdings, examined in §4.4.1 of the thesis.
Predicted insufficiency at the lower- and upper-quartile irrigation shares (23% and 67%) for two fragmentation levels. Bootstrap 95% CIs on the irrigation gap exclude zero.
~68–70% predicted insufficiency. Fragmentation adds about 2 pp on top.
~48% predicted insufficiency, with the fragmentation gap collapsing to ~0.1 pp — descriptively consistent with irrigation neutralising the fragmentation disadvantage, though the interaction is statistically null at n = 77.
A reproducible workflow that goes from the published NSCA 2021/22 tables to every number on this page.
The 10–12 month deprivation measure (NSCA Table 36.2) does not track overall insufficiency. Dolpa, the worst overall, has a severe share of 0.001. Severe peaks in Kaski (0.163) and Kathmandu (0.161) — urbanised districts of near-landless, part-time holdings.
Every bar shows a district's own overall and severe (10–12 month) insufficiency shares from the NSCA data; the eight districts are chosen to illustrate that the two measures do not track each other. Dolpa, Kaski and Kathmandu are the contrasts highlighted in §4.4.4 of the thesis, which draws no sufficiency conclusions from the severe measure — it is a distinct dimension of deprivation.
The page shows the patterns the thesis found. It does not — and the thesis does not — claim that pulling one lever would, on its own, fix the gap. Before you read the policy implications, read these.
The evidence is associational. Recommendations follow §5.4 of the thesis; the limitations are set out in §1.6 and revisited in §5.3.
The full 20-slide presentation — the question, the data, the method, the findings, and the honest caveats — embedded below. Open it fullscreen for the room, or download the PDF.
If a term here is unfamiliar, look it up. If a finding made you curious, the references below are the starting points cited in the thesis.
A full bibliography is included in the thesis PDF.
The full thesis PDF, the presentation deck, and the complete code-and-data repository are all one click away.
The complete document — five chapters, tables, robustness annex, and references.
The defense presentation — plain-language walkthrough with speaker-ready framing.
The full project — the NSCA 2021/22 source tables, the assembled district dataset, and the scripts that reproduce every table and figure in the thesis.