TuvaluRepresentation and Missing Evidence
The score is low because D5 (0.083) and D4 (0.145) are close to zero, and the geometric mean multiplies through both.
Tuvalu's provisional GEVI is 37.2 (90% interval 28.2 to 46.8), the lowest in the series. It has no women in parliament (0 of 16 seats) and none in ministerial positions, and eight of 13 inputs are imputed (blind-spot ratio 0.62), including all three Women, Business and the Law indicators, which are absent from the Tuvalu brief.
Source: the attached country paper. Animation reveals reported measures; it does not show change over time.
Abstract
Tuvalu's provisional GEVI is 37.2 (90% interval 28.2 to 46.8), the lowest in the series. It has no women in parliament (0 of 16 seats) and none in ministerial positions, and eight of 13 inputs are imputed (blind-spot ratio 0.62), including all three Women, Business and the Law indicators, which are absent from the Tuvalu brief.
The score is low because D5 (0.083) and D4 (0.145) are close to zero, and the geometric mean multiplies through both. The result is real for the political indicators, which are measured, and provisional for everything else.
Priorities include Treat political representation as an open policy question; Add Tuvalu to the WBL survey if it is not yet covered and publish the labour force tabulation by sex; Use administrative data.
Download the paper PDFHeadline result
Tuvalu's provisional GEVI is 37.2 (90% interval 28.2 to 46.8), the lowest in the series. It has no women in parliament (0 of 16 seats) and none in ministerial positions, and eight of 13 inputs are imputed (blind-spot ratio 0.62), including all three Women, Business and the Law indicators, which are absent from the Tuvalu brief.
90% interval 28.2–46.8.
Rank 15 of 15 · Blind-spot ratio 0.62.
versus D1 at 0.825
How the index is built
Dimensions are simple means of their indicators. The index uses a weighted geometric mean. Imputed inputs use the series medians; these are placeholders, not observed country outcomes.
| Dimension | Score D | Weight w | D^w |
|---|---|---|---|
| D1 Labour participation | 0.825 | 0.25 | 0.9531 |
| D2 Job quality and seniority | 0.810 | 0.20 | 0.9587 |
| D3 Enterprise and legal rights | 0.569 | 0.20 | 0.8932 |
| D4 Institutional visibility | 0.145 | 0.15 | 0.7485 |
| D5 Statistical and political voice | 0.083 | 0.20 | 0.6084 |
= 37.2
Product of the five terms = 0.3717, so GEVI = 100 × 0.3717 = 37.2, rank 15 of 15. Values are rounded as reported in the source.
Uncertainty and rank sensitivity
The 10,000-draw simulation samples each score with the standard deviation of its grade (5 grade A at 0.02, 8 grade C at 0.12) and draws dimension weights from a Dirichlet distribution centred on the base weights. Median 37.4, 5th to 95th percentile 28.2 to 46.8; blind-spot ratio 8/13 = 0.62. Under 5,000 random weight sets alone, the rank is 15 in every draw, so the last place does not depend on the weights.
Where the gap is
Each scenario raises one dimension to full parity, holding the others fixed.
| Dimension raised to 1.0 | GEVI | Gain |
|---|---|---|
| D1 | 39.0 | +1.8 |
| D2 | 38.8 | +1.6 |
| D3 | 41.6 | +4.4 |
| D4 | 49.7 | +12.5 |
| D5 | 61.1 | +23.9 |
Single-indicator scenarios
Single-indicator scenarios, each lifting one measured or coverage indicator to the upper quartile of countries with data (imputed indicators excluded):
| Indicator lifted | From | To | GEVI | Gain |
|---|---|---|---|---|
| I13 Data availability | 0.167 | 0.750 | 50.2 | +13.0 |
| I12 Women MPs | 0.000 | 0.382 | 47.2 | +10.0 |
| I11 Women ministers | 0.000 | 0.373 | 42.1 | +4.9 |
Scenario values and gains are reproduced as printed; rounding can affect the subtraction of displayed values. These are conditional model scenarios, not forecasts or causal estimates.
Reading the result
The score is low because D5 (0.083) and D4 (0.145) are close to zero, and the geometric mean multiplies through both. The result is real for the political indicators, which are measured, and provisional for everything else.
Two changes dominate. Reaching 15% of parliamentary seats and 15% of ministerial posts would add 13.6 points to 50.8, and raising data coverage to 9 of 12 rows would add 13.0 to 50.2. Either would move Tuvalu from last to the middle of the group.
The interval is among the widest in the series because so many inputs are imputed, so the rank of 15th should be read as "lowest on current evidence".
Priorities and practical use
The priorities connect the index to actions and named users.
Treat political representation as an open policy question. With one chamber of 16 seats, a single seat is 6.25% of the chamber; the data show the size of the gap, not its remedy.
Add Tuvalu to the WBL survey if it is not yet covered and publish the labour force tabulation by sex.
Use administrative data. With a very small formal sector, the government payroll and business register can give a complete count of women's formal work.
Who uses the data, and for what decision?
| Data product | Primary user | Decision it informs | Cadence |
|---|---|---|---|
| I13 coverage and imputed labour inputs (I1, I3, I4, I6, I7) | Statistics office, finance ministry | A defined data project: publish the labour and business rows by sex, worth +13.1 points | Once, then annual |
| I12 and I11: no women in parliament or cabinet | Parliament, island councils, women's organisations | Baseline for any representation discussion; a move to the series upper quartile would add +10.0 points for parliament alone | After each election |
| WBL indicators (imputed) | Justice ministry | Measured legal, enforcement and supportive scores | Next WBL round |
Use Tuvalu’s provisional score to prioritise missing evidence. Parliamentary and ministerial counts are measured. The source also grades youth NEET, management and coverage as measured, while eight inputs are imputed.
Series: GEVI Research Series · Paper 13 · DOI: 10.5281/zenodo.23209691