Every five years, as Bangladesh approaches a national election, something shifts deep inside Bangladesh’s public procurement system. Contract values climb sharply, some suppliers suddenly win dozens – sometimes more than a hundred – contracts in a single year, and ministries appear to fall back on familiar, long-trusted vendors rather than opening competition to new bidders.
The data reveals a simple but powerful pattern: election years turbocharge existing procurement relationships. Ministries spend more, and their preferred firms receive a dramatically larger share of the contracts. – hinting at a familiar pattern: spend hard before the vote, slow down afterwards.
A System That Breathes With the Election Cycle
Across 219,766 tenders from 2015–2025, the standardized awarded value totals ৳157,858.00 crore. Two election-period surges stand out in the time series: in 2018, total value reaches ৳11,470.467 crore (+31.4% vs 2017’s ৳8,729.399 crore), and in 2024 it reaches ৳24,315.131 crore (+27.8% vs 2022’s ৳19,029.947 crore; -4.9% vs 2023’s ৳25,557.623 crore). These are aggregate signals—not proof of wrongdoing—but they help identify years and actors worth closer, contract-level scrutiny.
Election cycle 1 – From 2017 to 2018, ministries lock in their favourites
To see how elections reshape market power, we zoom into the first cycle: the years 2017 (pre-election baseline) and 2018 (election year). For each year, we look at the top 10 firms by number of contracts won, and how many distinct ministries or divisions they served.
2017: A relatively open, mixed market of repeat winners
In 2017, repeat winners exist, but their dominance is moderate. The most active firm wins 43 contracts in the year, and some suppliers work across multiple ministries.
In 2017, some firms (like Flora Limited and Contemporary) serve multiple ministries, while most others are tied to one or two.
2018: Election-year procurement tightens into ministry loops
In 2018, the election year, the picture changes sharply. Top firms more than double their award counts – and almost all of them work with just one ministry or division.
The top firm in 2017 wins 43 contracts. In 2018, the top firm wins 104 contracts. Meanwhile, 2 out of these 10 election-year winners work with just a single ministry or division.
Election spending here looks like a ministry-level patronage surge: each ministry leans heavily on its preferred circle of contractors, awarding them dozens of contracts in the election year.
Election cycle 2 – From 2022 to 2023, repeat winners meet cross-ministry giants
The second cycle — 2022 and 2023 – tells a slightly different story. The system still has many single-ministry repeat winners, but now a set of large, cross-ministry suppliers emerges, especially in ICT and infrastructure.
2022: Rising concentration and the first major cross-ministry ICT player
By 2022, you already see very high repetition. Some firms win more than a hundred contracts a year, often within a single ministry. But a major outlier appears: a national-level ICT supplier working across nine ministries.
Global Brand pvt ltd – 78 contracts, 103.16 Cr BDT, 24 ministries/divisions
This ICT firm behaves very differently from the rest: it is not attached to a single ministry, but appears across the government. The seeds of centralized digital procurement are visible even before the 2024 election.
2023: Election-year repetition plus cross-ministry modernization
In 2023, the year leading into the January 2024 election, repetition gets even stronger – and the cross-ministry trend deepens, especially in ICT and office infrastructure.
Most leading firms in 2023 are still single-ministry specialists, winning dozens of contracts from the same ministry. But a few players stand out:
M/S. Maa Enterprise – 67 contracts, 16.86 Cr BDT, 7 ministries/divisions
M/S. Khandakar Enterprise – 63 contracts, 8.58 Cr BDT, 4 ministries/divisions
These firms behave like cross-government utilities, embedded across many ministries. When election-year budgets swell, they are perfectly positioned to capture ICT, furniture, and equipment spending on a national scale.
Comparing the two election cycles
Looking across both cycles, election years do not simply increase spending – they reshape how much power sits in the hands of repeat winners, and how tightly those winners are tied to ministries.
The 2018 election cycle looks like a classic case of patron-client procurement: ministries dramatically ramp up awards to their established suppliers, without bringing in many cross-ministry players.
The 2024 election cycle (via 2022-2023) still has the same tight ministry-contractor loops, but now layered on top is a wave of centralized modernization spending, channeled through firms like Smart Technologies and Hatil that supply dozens of agencies at once.
Election-year repeat winners vs a decade of repeated winners
To see whether election behavior is exceptional or just an amplified version of normal procurement, we compare the election-cycle patterns with the top repeat winners across the entire 2015-2025 dataset.
Across 219,766 contracts in the dataset, there are 20,799 distinct firms:
5,937 firms win exactly one contract 14,862 firms win multiple contracts
A small club of companies dominate tender wins over the decade – appearing year after year, and often across multiple ministries.
Top overall repeat winners, 2015–2025
These long-term repeat winners fall into two structural groups:
-
Ministry-tied specialists – firms like NUPUR, Bismillah, Sikder, and Anik Consodiam, which win hundreds of contracts but remain anchored in one or a few ministries.
-
Cross-ministry giants – especially Smart Technologies (Bd) Ltd., which appears in contracts with 55 different ministries and divisions over the decade.
How election cycles compare to this decade-long pattern
The 2017–2018 election cycle intensifies the ministry-tied pattern:
-
Top firms in 2018 almost all work with just one ministry.
-
Award counts jump sharply compared to 2017, but cross-ministry players are rare.
This looks like the system’s ministry-contractor relationships being pushed into overdrive.
The 2022–2023 election cycle looks more like the full decade:
Single-ministry loops dominate among small traders.
Cross-ministry giants like Global Brand, Smart Technologies, and Hatil emerge around election time.
This mirrors the 2015-2025 structure, but with heightened intensity.
Election years do not create new power dynamics from scratch – they amplify existing procurement hierarchies. Firms that are already embedded in the system reap the biggest election-year gains, whether inside a single ministry or across dozens.
Why these patterns matter?
Election-year spending surges are not automatically corrupt. Governments may rush to finish overdue projects or fast-track essential infrastructure. But when the same firms keep winning hundreds of contracts, often from the same ministry, crucial questions arise:
Are tenders being rushed with weaker scrutiny in pre-election months?
Do ministry-tied suppliers enjoy an unfair inside track over new competitors?
Are cross-ministry ICT and furniture contracts genuinely competitive – or centrally steered?
How many of these repeat winners have links to politically exposed persons?
The Election Fever investigation does not attempt to answer those questions by itself – but it shows clearly where to look. Election-year spikes in contract values, combined with concentrated repeat winners and tight ministry loops, mark the hotspots where public money and political incentives are most deeply intertwined.
Methodology & Limitations
This investigation is based on analysis of public procurement award notices published through Bangladesh’s electronic Government Procurement (e-GP) system. The unit of analysis is an awarded contract record as it appears in the dataset: who awarded it, who received it, when it was awarded, and the awarded value captured in the record.
Before analysis, supplier names were standardised to reduce duplicate spellings that can fragment a single firm into multiple entries. This standardisation removed common prefixes (for example, M/S or Messrs), trimmed inconsistent punctuation and spacing, and unified common legal forms (for example, Limited vs Ltd; Private vs Pvt). These steps aim to make firm-level counts more reliable, but they cannot fully resolve deeper identity issues such as ownership changes, subsidiaries, or firms with genuinely similar names.
Joint ventures and consortium-style awards were excluded from the calculations shown here. In award-to fields, multi-party arrangements often appear as “JV”, “Joint Venture”, or as paired names separated by characters such as “/”, “&”, “and”, or hyphenated partner listings. Rather than attempting to split these awards across partners – an approach that can introduce new errors – this story removes them so repeat-winner and concentration patterns reflect awards to single suppliers only.
Contract values were analysed using the cleaned value field in the dataset and summarised in crore BDT. Records with missing or non-numeric value entries were excluded from value totals but could still contribute to simple counts where appropriate. All year-by-year totals, firm totals, and ministry totals in the narrative and charts are computed from the same cleaned dataset so that numbers remain internally consistent across the page.
These results describe structural patterns in the award record – repetition, concentration, and timing – but they do not prove illegality, collusion, or political influence. Higher spending around election periods may also reflect project completion cycles, budget utilisation deadlines, disaster response, or policy priorities. The findings should be treated as signals for deeper reporting: checking competition in specific tenders, reviewing documentation quality and timelines, examining contractor performance, and mapping beneficial ownership or politically exposed persons where relevant.
Because the story relies on administrative award data, it cannot directly measure the competitiveness of bidding (for example, how many firms submitted bids, how evaluation scores were assigned, or whether specifications were restrictive) unless those fields are available and linked.