Apparel Brands Turned Disclosure into a Performance of Responsibility After Rana Plaza

The global apparel industry's sustainability reporting expanded substantially in the years following the 2013 Rana Plaza disaster. A new study examining 322 reports from 69 companies over seven years jhas found that the expansion was real and the accountability was not. Environmental content consistently outranked labour conditions, and the sector's reporting tone became progressively more optimistic regardless of what was happening in the supply chains those reports described.

Long Story, Cut Short
  • Apparel companies subject to post-Rana Plaza scrutiny consistently reported more on environmental topics than on the labour conditions that caused the disaster.
  • Across seven years of sustainability reports, positive framing intensified over time regardless of whether underlying supply-chain performance had actually improved.
  • The sector's reporting architecture became more sophisticated while its capacity to reveal consequential labour risks declined relative to its environmental coverage.
The industry-level corpus of 322 reports from 69 companies replicates the topic-priority and positive-framing patterns found among Accord and Alliance members, indicating that environmental emphasis and rhetorical optimism are sector-wide disclosure norms rather than crisis-specific responses.
Sector Pattern The industry-level corpus of 322 reports from 69 companies replicates the topic-priority and positive-framing patterns found among Accord and Alliance members, indicating that environmental emphasis and rhetorical optimism are sector-wide disclosure norms rather than crisis-specific responses. AI-Generated / magnific

This article is based on the paper 'Reporting What Matters, or Reporting What Looks Good? Materiality and Balance in Post–Rana Plaza Apparel Sustainability Disclosure' by Mahsa Mohammadrezaei and José Carlos Marques of the Telfer School of Management, University of Ottawa, Canada.

The 2013 Rana Plaza disaster forced apparel brands into a disclosure test they could not avoid. Regulators, trade unions, consumers, and NGOs demanded transparency on supply-chain safety and labour conditions — the precise failures the disaster had exposed. Two accountability mechanisms followed: the Bangladesh Accord on Fire and Building Safety, a legally binding agreement between brands and trade unions, and the Alliance for Bangladesh Worker Safety, a voluntary programme driven by North American retailers. Both required member companies to engage in substantially enhanced reporting.

What followed over the next seven years was a significant expansion of sustainability disclosure across the apparel sector. Reports grew in length, frameworks were more consistently cited, and alignment with recognised materiality standards generally improved. The case for reading this as progress has obvious surface logic. Companies were disclosing more. The pressure to do so was real.

A research study published in Business Strategy and the Environment in 2026 disrupts that logic. Examining 101 reports from 15 Accord and Alliance member firms and a broader corpus of 322 reports from 69 apparel companies between 2014 and 2021, the study applied a pipeline combining SBERT-based semantic analysis against SASB material topic benchmarks with VADER sentiment scoring to measure not only what companies reported but how they framed it.

Two findings define the study's analytical stakes. Among the firms most directly implicated in the disaster, environmental topics consistently received greater discursive attention than labour conditions across nearly the entire study period, with 2014 as the sole exception. The disaster was primarily a labour catastrophe. The Accord and Alliance were established specifically to address labour and safety failures. Yet companies whose reporting obligations were anchored to that failure emphasised the less contested terrain instead.

The second finding compounds the first. Across all four material topics and across both analytical levels, the tone of sustainability reporting was systematically and increasingly positive. Balance scores approaching zero, which would indicate genuine candour about both achievements and difficulties, were rare. Highly positive scores became the norm, intensifying over time.

These two patterns together constitute the study's central argument: that sustainability reporting in post-Rana Plaza apparel developed less as a transparent reckoning with material risk than as a managed language system, one in which disclosure could be extensive, formally compliant, and structurally evasive at the same time.

When Reporting Avoids Labour

The central contradiction in the post-Rana Plaza reporting record is not silence. Among the 15 Accord and Alliance member firms examined across 101 sustainability reports, labour conditions appeared consistently as a disclosed topic. The problem is ordering. Across nearly the entire 2014–2021 period, environmental impacts in the supply chain attracted higher alignment percentages than labour conditions, a pattern that held across most companies in both groups and remained stable over time. Companies subject to the most direct labour-accountability pressure in the global apparel sector were giving greater discursive weight to a different category of risk.

The sample makes this finding harder to dismiss. These 15 firms, including Walmart, H&M, Gap, PVH, and Next, were verified members of the Accord or Alliance, sourced from Bangladesh during the study period, and subject to the governance requirements those memberships imposed. Several had been implicated in documented supply-chain controversies — labour rights violations, wage disputes, factory safety failures — as recorded in investor ESG assessments and NGO monitoring reports. The Accord and Alliance were not peripheral to their reporting context. They were the institutional frame within which that reporting was produced.

What the data reveals is a question about materiality as a practice rather than a principle. Environmental reporting offers companies a workable language of progress: energy efficiency targets, renewable procurement, water conservation metrics, supplier collaboration frameworks, external validation through schemes such as the Carbon Disclosure Project. Improvements in these areas are measurable, communicable, and verifiable at a level of abstraction that keeps the most difficult supply-chain realities at arm's length. Labour conditions in complex, multitier supply chains concentrated in Bangladesh, Vietnam, and Cambodia operate on different terrain. Wages, safety enforcement, audit reliability, and supplier control are politically contentious, harder to resolve, and harder to narrate positively. The gap between stated commitment and documented outcome in this domain was, by 2014, already extensively recorded.

The result is a structural condition the research identifies as topic displacement. Labour conditions remain present in these reports; they are not omitted. But they are persistently outranked by content that is easier to frame as evidence of corporate responsiveness — content that satisfies materiality frameworks while shifting discursive attention toward territory where progress is more achievable and more communicable. Materiality, in principle, is the mechanism that should ensure companies report on what is most consequential. In practice, it is also a mechanism that can be selectively applied.

The one exception in the data anchors the interpretation. In 2014, the year immediately following the disaster, labour conditions received relatively stronger emphasis. The subsequent drift toward environmental dominance was not the product of stable corporate priorities. It was a trajectory — one that moved away from the crisis's defining issue as the immediate pressure of accountability receded and the conventions of sustainability reporting reasserted themselves. That drift is the finding. Not absence. Displacement.

The Accord and Alliance
  • The Bangladesh Accord on Fire and Building Safety, signed in 2013, was a legally binding agreement between global brands and trade unions covering factory inspections and remediation.
  • The Alliance for Bangladesh Worker Safety was a parallel voluntary initiative led by North American retailers, operating between 2013 and 2018 with its own audit and reporting requirements.
  • Both initiatives required member companies to publish substantially enhanced sustainability disclosures covering supply-chain safety and labour conditions in Bangladesh.
  • The study's firm-level subsample covers 101 sustainability reports from 15 companies verified as Accord or Alliance members, including H&M, Walmart, Gap, PVH, and Next.
Mrasuring Reporting Quality
  • Researchers applied SBERT-based semantic analysis to classify individual sentences in each report against SASB material topic benchmarks for the apparel industry.
  • A cosine similarity threshold of 0.60 was used to determine whether a sentence aligned with a given sustainability topic, balancing precision against coverage across a large corpus.
  • VADER sentiment scoring was applied within each classified topic to measure whether companies framed sustainability content positively, negatively, or neutrally.
  • The resulting normalised balance score ranges from negative one to positive one, where zero indicates equal representation of achievements and challenges in a given topic.

The Grammar of Positive Framing

Knowing which topics companies chose to emphasise is one analytical move. Knowing what they did with those topics is another. The study operationalises the second question through a normalised balance score, computed by decomposing each report's aligned sentences into positive, negative, and neutral sentiment categories using VADER, then calculating the ratio of positive to negative framing relative to each topic's share of total report content. The metric is designed to prevent distortion from uneven topic coverage. Applied across the full firm-level subsample, it produces a heatmap that runs predominantly green. Positive framing, across companies, across topics, across years.

Several patterns within that finding carry specific analytical weight. Labour conditions and environmental impacts, the two topics attracting the most content, also show the highest positive framing ratios across the widest range of companies. The topics generating the greatest volume of reporting are also the most positively framed. Management of chemicals and raw material sourcing, which attract less discursive attention, show lower and more variable scores. Companies engage somewhat more candidly with topics that face less stakeholder scrutiny. Negative scores do appear. Canadian Tire and Carter's record them in certain years; Puma shows notable fluctuation across reporting cycles. But they are uncommon.

The temporal pattern is the most structurally significant. The earliest years of the study period show greater variability, with more instances of neutral or candidly negative framing. By the later years, highly positive scores predominate across nearly all companies and topics. Next moved from near-neutral or negative scores in the early years to consistently high positive scores by 2021. PVH, H&M, and Sainsbury's follow comparable trajectories. Environmental impacts and labour conditions saw the sharpest shift. As companies became more practised at sustainability reporting and frameworks became more standardised, the language of their reports became progressively more aspirational and achievement-focused. The architecture formalised. The candour receded.

What that trajectory does not establish is deliberate misrepresentation. The study is careful on this point. Organisations that have genuinely improved their practices should report improvement. But consider what the data actually shows: uniformity of positive framing across companies with different governance structures, different sourcing profiles, and different exposure to the Rana Plaza accountability crisis — and intensification of that framing over time, independent of any documented convergence in underlying performance. Genuine and uniform performance gains cannot explain a pattern that consistent. Something structural is producing it.

That structural condition is a sector-wide convention of rhetorical optimism: a communicative tendency, reinforced by reporting frameworks and professional norms, to foreground achievements, frame challenges as problems already being addressed, and translate the complexity of supply-chain risk into a narrative of managed, forward-looking progress. The pattern is consistent enough across companies and years that individual intent is the less useful analytical frame. What the data establishes is structural: a risk acknowledged is immediately paired with a management response; a safety incident followed by a commitment to enhanced monitoring. The negative content is present, but companies structure it to function as evidence of responsibility rather than as a candid account of ongoing difficulty.

Stakeholders receiving these reports encounter documents that may be materially populated but narratively softened. Distinguishing genuine progress from polished disclosure becomes harder than it should be.

The gap between the formal architecture of sustainability disclosure and its accountability function is the study's central finding: reports became more technically organised over time while their capacity to make labour risk visible to regulators, investors, and civil society declined relative to their environmental coverage.
The gap between the formal architecture of sustainability disclosure and its accountability function is the study's central finding: reports became more technically organised over time while their capacity to make labour risk visible to regulators, investors, and civil society declined relative to their environmental coverage. AI-Generated / magnific

A Sector-Wide Disclosure Norm

The sector-level data introduces the more demanding test. Across a broader corpus of 322 reports from 69 apparel companies, most of which had no direct membership in the Accord or Alliance, the question is whether the environmental-over-labour priority and the systematic positive framing were specific to crisis-linked firms managing acute reputational pressure — or whether they reflect something wider. The industry-level data does not support the first reading.

At the sector level, environmental impacts and labour conditions again dominate alignment percentages relative to chemical management and raw material sourcing, consistent with the firm-level pattern. Chemical management remains the most persistently under-represented topic across the full seven-year period. Companies' disclosures in this area ranged from 2.79% to 3.46% alignment. That is notably low for a topic with direct implications for both worker safety and product safety in apparel manufacturing. Raw material sourcing tells a different story. It shows higher variability and stronger growth over the study period, likely reflecting increasing consumer and regulatory attention to supply-chain transparency in the latter part of the 2014–2021 window. Higher volume, however, does not indicate greater candour about the challenges within it.

Labour disclosures at the industry level acknowledge risks more explicitly than in some other topics. References to forced labour, human trafficking, unsafe conditions, and audit findings appear in the corpus. But companies pair each acknowledgment with a corrective-action narrative, a policy commitment, or an assurance of enhanced monitoring, framing the negative content as a problem already being governed rather than as evidence of systemic failure. Environmental disclosures follow a comparable logic. Companies dominate their positive statements with references to energy efficiency investments, renewable energy adoption, water conservation programmes, greenhouse gas reduction targets, and alignment with frameworks such as the GRI and the Carbon Disclosure Project. In both cases, the corrective frame closes the negative disclosure before it can accumulate critical force.

The sector-level analysis confirms what the firm-level heatmaps suggest. Positive framing predominates across all four SASB topics at the industry level as well. Statements emphasising achievements, certifications, commitments, and forward-looking targets substantially outnumber statements acknowledging current failures, persistent risks, or unresolved challenges. The pattern is not confined to the firms most directly implicated in Rana Plaza. It is the sector's reporting norm.

Companies across the broader apparel industry became more technically organised in their reporting, more consistent in citing recognised frameworks, more aligned with established disclosure standards. The architecture improved. The accountability content it carried did not keep pace. Better reporting structure does not automatically produce better accountability. Without scrutiny of tone and topic priority, formal compliance with disclosure frameworks can strengthen a company's legitimacy position while leaving its most consequential risks underexamined. The reports become more organised. The risks become harder to see.

Accountability Versus Reputation

The unresolved stake is not whether sustainability reports should exist, but whether they can function as accountability documents rather than reputation documents. The evidence from post-Rana Plaza apparel is specific: a disaster rooted in labour failure produced a reporting culture in which environmental progress received more sustained attention, and in which the language of disclosure became progressively more optimistic regardless of what was happening in the supply chains those reports described. The centre of gravity shifted. It shifted without announcement, within the conventions of recognised frameworks, and it has not shifted back.

Richa Bansal

RICHA BANSAL has more than 30 years of media industry experience, of which the last 20 years have been with leading fashion magazines in both B2B and B2C domains. Her areas of interest are traditional textiles and fabrics, retail operations, case studies, branding stories, and interview-driven features.

 
 
 
Dated posted: 1 May 2026 Last modified: 1 May 2026