Mallooki

Research blog / Kuwait public economics / August 2026

Kuwait's Procurement Portal Is a Lead, Not Yet a Dataset

A field-level audit of one indexed CAPT winning-bids page

By Abdulmalik Altararwa

At a glance

Question
Does one indexed public winning-bids page contain the fields needed to study spending concentration, bidder competition, or SME preferences?
Main finding
The page identifies tenders and most winners, but contains no numeric award values, losing bids, technical results, dated SME status, or award reasons.
Evidence limit
The ten rows are a complete page, not a representative sample or a portal-wide audit. Missing fields may exist in other page states or documents.
Browse all research

Before estimating who wins Kuwait's public contracts, a researcher needs a defensible award universe, numeric values, stable supplier identifiers, and the competing bids. I audit every row in one complete indexed winning-bids page. It is useful for discovery because tender identifiers and subjects are present, and nine of ten winners are named. It is not a competition dataset because every displayed value is missing and the choice set is absent.

Snapshot
10 rows, serials 11 to 20
Meeting dates
28 May to 17 June 2025
Named winners
9 of 10 rows
Numeric award values
0 of 10 rows

A ten-row field audit

The sample is one complete indexed page with ten winning-bid records. Before reading the rows, I defined ten fields that a procurement study would need and measured the share of rows where each field was visibly populated.

Horizontal bars showing field coverage across ten indexed CAPT winning-bid records.
Field coverage in one indexed winning-bids snapshot. A dash followed by KD is treated as missing text, not a zero-value contract. This page is not a representative sample of all tenders. Source: source-checked CAPT public rendering; author's audit.

View full-size figure (opens in a new tab)

Tender number, meeting date, organisation, and subject appear in all ten records. A winning supplier appears in nine. Every value field displays a dash followed by KD, so none contains a numeric award value.

A winner is not a competition dataset

Procurement is a choice among bids. Observing only the winner removes the choice set. Without losing bids, a researcher cannot measure participation or price dispersion. Without technical qualification, it is impossible to know which prices remained eligible. Without final values, supplier concentration cannot be weighted by spending.

The nine named winners happen to be distinct after light name normalization. That is not evidence of low concentration. Ten page-ordered rows do not define a procurement market, and a small office-supply award should not count the same as a large infrastructure contract.

The access boundary is part of the audit

On 12 August 2026, the live CAPT portal presented a Cloudflare verification screen, while the indexed pre-production host no longer resolved. The collection process did not bypass either control.

The audit therefore covers a preserved public rendering. A field missing there may exist in an opening record, a meeting-minute PDF, or another portal state. The page is a useful lead for locating tenders, not a complete statement about CAPT's full disclosure system.

What the audit does and does not establish

The ten rows show that one public-facing rendering aids discovery but lacks the quantities needed to measure spending, competitive intensity, technical selection, or SME preference. The note makes no portal-wide transparency ranking and no claim about procurement outcomes.

Citation and sources

Suggested citation: Abdulmalik Altararwa. 2026. "Kuwait's Procurement Portal Is a Lead, Not Yet a Dataset." Research blog.

Sources: The source is a preserved public search rendering of the Central Agency for Public Tenders winning-bids page. The live CAPT winning-bids route was access-checked without bypassing its verification controls.

Public-release boundary: This page publishes aggregate estimates, original figures, and source links. It does not publish raw records, restricted data, or local research files.