Procurement data makes public purchasing easier to examine. The strongest comparisons explain what is covered, what is being measured and which conclusions the records cannot support.
- Observation coverage
- 2011 to 2026-09-22
- Source build date
- 2026-09-22
- Website refresh policy
- Every Monday and Thursday (3–4 days)
Published snapshot. A scheduled check does not imply that a new source observation is available.
Loading the published dataset…
Reading the evidence
01
An award is a commitment, not a completed service
A notice of award records a contracting decision and a stated value. It does not certify payment, delivery, construction quality or final cost. Award totals can be used to describe the purchasing programme, but evaluating service delivery requires joining them to implementation and expenditure records. The charts consistently label values as nominal awarded amounts and keep them separate from tender-list statuses.
02
Growth can reflect the reporting system
The earliest years of an electronic procurement archive can capture only a small portion of purchasing activity. Later growth may combine broader portal use, changes in the mix of participating offices, price changes and more procurement. The latest year is also incomplete. Comparing raw annual totals alone cannot isolate a change in public investment or administrative performance. A robust follow-up would hold office coverage constant and express values in comparable prices.
03
Competition needs a market and a denominator
The procurement method describes a procedure; it does not report the number of qualified bids or the independence of bidders. Concentration across the entire country may be low while a particular office or product market relies on very few suppliers. Conversely, repeated awards can reflect specialisation or framework arrangements. Method shares and vendor concentration are useful screening measures when combined with market definitions, bidder participation and contract documentation.
04
Matching plans to awards introduces selection
The plan comparison covers packages with usable estimates and an identifiable award match. It is not a random sample of all procurement. Text references can be reused across offices or entered inconsistently; source filters exclude generic references and implausible ratios, but residual matching error remains possible. Award-to-estimate clusters should prompt document checks and comparison with applicable procurement rules. They are not proof of collusion, favouritism or a pricing violation.
05
The place on a contract is not the place of benefit
A centrally procured item may be used in several districts, while an office address can be mistaken for a delivery location. The map presents the district attached to the archived record. It should not be interpreted as a ranking of public resources received by residents. Establishing geographic equity would require project-level service locations, beneficiary populations and a consistent treatment of national projects.
Methods, sources and reuse
The Bun pipeline reads already-normalised e-GP products from one sparse source checkout and requires record counts to reconcile across analysis, geography, flags, office profiles and plan matching. National and district summaries stay in the primary payload; the office index, office-profile shards and enforcement cross-matches load separately on demand, each below a two-megabyte validation limit. BDT amounts are divided by 10,000,000 for crore display and are not inflation-adjusted. Political-spending tables preserve standard errors, placebo periods, event studies, alternative denominators, division-by-month controls, category checks and a vendor-concentration outcome. Vendor and owner links inherit source name normalisation and remain screening evidence rather than verified beneficial ownership.
- 1. Bangladesh e-GP: notices of award
Primary award records.
- 2. Bangladesh e-GP portal
Tender notices, procurement plans, awards and supporting records.
- 3. e-GP source pipeline and matching limitations
Archive coverage, entity matching, exclusions and correction history.
Download and cite this snapshot
Institute of Policy Dynamics, “Bangladesh Public Procurement: Awards, Methods and Geography”. Coverage: 2011 to 2026-09-22. Cite the original data providers alongside IPD.
Download the complete published dataset (JSON)Source fingerprint (SHA-256): a1df203041e4d527670b53b568dcbf619edbd30d2f3ab7755565f3331f30e9fc
Chart CSV files contain the selected observations. Original data remain subject to their providers' terms; public access alone does not establish an open licence.