Missing Revenues in Bangladesh Tax Potential, Gaps, and Distributional Consequences

Missing Revenues in Bangladesh Tax Potential, Gaps, and Distributional Consequences Mohammad Abdur Razzaque Md. Deen Islam Tahsin Rifat April 2026

Table of Contents Executive Summary .................................................................................................................................................. 1 1. Introduction ............................................................................................................................................................ 5 2. Tax Capacity, Incidence, and Compliance: Evidence from the Literature .............................................. 7 3. Data and Estimation Framework ...................................................................................................................11 4. Findings .................................................................................................................................................................16 4.1 Overview of the current tax system of Bangladesh...............................................................................16 4.2. Assessment of tax revenue performance................................................................................................22 4.3. Factors influencing low tax compliance..................................................................................................26 4.4. Potential of wealth tax for government revenue and social justice...................................................30 5. Policy Implications and Concluding Observations .....................................................................................35 References ................................................................................................................................................................. 39 Appendix .................................................................................................................................................................... 44 List of Tables Table 4.1: The estimated index of relative tax burden......................................................................................18 Table 4.2: Summary of tax revenue performance in FY22................................................................................23 Table 4.3: Active taxpayers on the PIT register as a percentage of the labour force and the tax-to-GDP ratio of various countries of Asia and the Pacific.......................................................24 Table 4.4: Income tax compliance rates...............................................................................................................25 Table 4.5: Summary of models explaining factors influencing people’s decision to pay taxes.................27 Table 4.6: Simulation of tax revenue by applying a wealth tax.......................................................................33 List of Figures Figure 4.1: Cumulative share of household income, income tax, and VAT held/to be paid by the cumulative share of households........................................................................................................16 Figure 4.2: Distribution of income tax and VAT relative to the distribution of household income...........17 Figure 4.3: Distribution of tax burdens as shares of household income as well as expenditure...............19 Figure 4.4: Weighted average household income levels, and shares of income tax and VAT by (a) rural household deciles and(b) urban household deciles.....................................................20 Figure 4.5: Weighted average household expenditure, and VAT as a share of total expenditure by (a) rural household deciles and(b) urban household deciles......................................................21 Figure 4.6: Tax-to-GDP ratios in Asia and Pacific Countries, 2023.................................................................23 Figure 4.7: Cumulative share of total wealth held by the cumulative share of household........................31 Figure 4.8: Percentage of wealth possessed by the percentiles of households in total wealth.................32

Acknowledgements This paper has been prepared with support from the Friedrich-Ebert-Stiftung(FES) Bangladesh Office. The authors express their sincere gratitude to Dr Felix Gerdes, Resident Representative, FES Bangladesh Office, for his keen interest in this work, his helpful guidance in shaping the initial concept note, and his comments on an earlier draft of the paper. Special thanks are due to Md. Khademul Islam Chowdhury, PhD intern at RAPID and PhD candidate in Economics at the Research School of Economics, The Australian National University, for his valuable inputs. The authors are also grateful to Mr. Shadhan Kumar Das of the FES Bangladesh Office for his overall support. Any remaining shortcomings or views expressed in the paper are solely those of the authors.

Missing Revenues in Bangladesh Tax Potential, Gaps, and Distributional Consequences Executive Summary Introduction: Bangladesh’s persistently low domestic revenue mobilisation has become a major macrofiscal constraint, limiting the state’s ability to finance public goods, social protection, and broader development priorities. Despite sustained economic growth, tax collection remains far below what the country’s legal tax framework and economic structure would suggest is achievable. This shortfall reflects not only weaknesses in tax policy and administration, but also deeper institutional problems related to governance, fairness, and compliance. The current tax structure is marked by heavy reliance on indirect taxation, especially VAT, alongside a narrow and weakly enforced personal income tax base, raising concerns about both efficiency and equity. Against this backdrop, the paper examines Bangladesh’s unrealised revenue potential and the distributional consequences of its tax system using household-level data from the Household Income and Expenditure Survey(HIES) 2022. It estimates the legally implied liabilities for personal income tax and VAT under full compliance and compares these with actual collections reported by the National Board of Revenue to derive tax gap estimates. At the same time, it analyses how the burden of taxation is distributed across income groups and explores the institutional and behavioural drivers of low compliance. By linking micro-level evidence on tax gaps and incidence with broader insights from the political economy of taxation, the study sheds light on why Bangladesh’s fiscal system remains both underperforming and inequitable, and what this implies for future reform. Tax Capacity, Incidence, and Compliance: Evidence from the Literature: The literature points to a persistent, structural revenue weakness in Bangladesh. Despite relatively strong economic growth, the country has remained one of the world’s weakest performers in tax collection, with the tax-to-GDP ratio hovering at around 7 to 9 per cent for many years. Existing studies suggest that this underperformance reflects not only low income levels or structural features of the economy, but also substantial unrealised tax bases, weak enforcement, and shortcomings in tax administration. The problem is compounded by a tax structure that relies heavily on indirect taxation while direct taxes remain underdeveloped, raising both efficiency and equity concerns. The evidence also shows that the personal income tax system is formally progressive in terms of statutory rates but operates on an extremely narrow base, with only a small fraction of individuals with taxable income appearing in the effective tax net. At the same time, the VAT system places a relatively heavier burden on lower-income households and suffers from significant compliance failures, especially among firms that collect VAT but do not fully remit it. The literature further highlights the role of widespread informality, administrative weakness, corruption, and discretionary tax exemptions in undermining compliance and narrowing the effective tax base. Taken together, these findings suggest that Bangladesh’s weak tax performance is rooted not only in structural economic constraints but also in institutional and governance failures that weaken both fairness and fiscal capacity. Data and Estimation Framework : The paper combines household-level microdata from the Household Income and Expenditure Survey(HIES) 2022 with secondary literature to assess Bangladesh’s unrealised tax potential, the distribution of tax burdens, and the institutional drivers of low compliance. Potential 1

personal income tax liabilities were estimated by applying the statutory FY2021–22 tax schedule to household income data, while potential VAT burdens were derived from household consumption expenditure by identifying VAT-liable goods and services and applying the relevant tax rates. These estimated liabilities were then compared with actual National Board of Revenue collections to derive indicative tax gaps under the existing legal framework. To assess distributional incidence, households were ranked by equivalised income using the OECDmodified equivalence scale, allowing more meaningful welfare comparisons across households of different sizes and compositions. The analysis then examined tax burdens across income deciles, urban and rural groups, and broader regional patterns. It also incorporated a simulation of wealth-related taxation using household asset data to gauge the potential revenue and equity implications of taxing highly concentrated wealth. Overall, the study adopts a mixed-methods framework that combines quantitative tax incidence and tax-gap estimation with qualitative analysis of governance, compliance, and fairness. Findings : The analysis shows that Bangladesh’s tax system is both underperforming and inequitable. Personal income tax is formally progressive, but its effective base is extremely narrow: the richest 10 per cent of households account for more than 40 per cent of total income and are liable for over 93 per cent of potential PIT, while the bottom four deciles have almost no taxable earners. VAT displays the opposite pattern. Although the richest 10 per cent bear only around 23 per cent of the total VAT burden, the poorest 10 per cent spend about 20.37 per cent of their income on VAT, compared with just 2.8 per cent for the richest 10 per cent, indicating strong regressivity when assessed relative to income. The VAT burden appears to be particularly severe in rural areas, where the poorest decile pays more than 23 per cent of income as VAT. However, the estimation of the VAT-to-income ratio, especially of the poorest decile, may be overstated due to the underreporting of income in the survey. One of the pieces of evidence of such underreporting in the survey data is that the average income-to-consumption ratio for the full sample is 0.85. Furthermore, among the poorest decile, total consumption expenditure is almost five times higher than their reported total income. On the other hand, VAT appears broadly proportional relative to expenditure, suggesting that regressivity arises largely because poorer households consume a much larger share of their income. The paper also finds very large unrealised revenues. In FY2021–22, actual PIT collection amounted to only 0.98 per cent of GDP, compared with an estimated potential of 4.85 per cent, implying a PIT gap of around 3.87 percentage points of GDP. For domestic VAT, actual collection was 1.8 per cent of GDP against an estimated potential of 3.82 per cent, implying a minimum VAT gap of 2.02 percentage points. Combined unrealised PIT and VAT revenue thus amounts to at least 5.89 per cent of GDP. Compliance indicators reinforce this diagnosis: only 3.57 per cent of the labour force appears as active taxpayers on the PIT register, while filing compliance among e-TIN holders is 43.06 per cent for individuals and 23.61 per cent for companies. More broadly, the evidence suggests that Bangladesh’s weak revenue performance reflects not only informality and administrative weakness, but also low audit intensity, governance failures, and perceptions of unfairness that erode tax morale. The findings further indicate that wealth is far more concentrated than income and remains largely outside the effective tax net. Using the HIES 2022 data at the household level, it is estimated that the wealthiest 10 per cent of households hold more than 76 per cent of total private wealth, while the top 1 per cent alone hold more than 47 per cent, rising to over 53 per cent in urban areas. Simulations suggest 2

that, under full compliance, even a 1 per cent annual wealth tax on the top 10 per cent of households could yield revenue equivalent to 4.5 per cent of GDP. However, given Bangladesh’s current administrative and institutional limitations, the paper concludes that an annual wealth tax is unlikely to be feasible in the near term. Inheritance taxation, combined with stronger asset registries and better taxation of highvalue secondary property, is likely to be a more realistic entry point for bringing concentrated wealth within the fiscal framework. Low tax compliance in Bangladesh can be understood through five interrelated lenses discussed in the paper. Under the Economic Deterrent Model, evasion remains attractive because the probability of detection is extremely low, making the expected cost of non-compliance small relative to its gains. Social-Psychology Models help explain how widespread perceptions of evasion, especially by influential groups, weaken tax morale and normalise non-compliance. The Fiscal Exchange perspective suggests that the willingness to pay taxes is further undermined when citizens do not see tax revenues translated into effective, visible public services. Comparative Treatment is also highly relevant, as perceptions that similarly placed taxpayers are not treated equally and that politically connected groups benefit from exemptions or escape scrutiny erode confidence in the fairness of the system. Finally, Political Legitimacy and Trust in Government bring these strands together by showing how weak institutions, corruption, poor governance, and limited administrative credibility reduce voluntary compliance. In Bangladesh, therefore, low tax compliance is not simply an enforcement problem. It is the outcome of a broader lowtrust equilibrium in which weak deterrence, low tax morale, fragile fiscal exchange, perceived unfairness, and limited institutional legitimacy reinforce one another. Policy Implications : The major policy implications from the analyses are that Bangladesh requires a broad-based reform agenda spanning tax policy, tax administration, and the wider institutional foundations of compliance. First and foremost, the personal income tax system needs to be broadened and modernised so that it no longer depends so heavily on a narrow pool of visible taxpayers. This calls for stronger use of third-party information, wider withholding arrangements, closer integration of financial, property, and administrative records, and a careful rationalisation of exemptions, allowances, and credits that erode the base. VAT reform is another critical area, with the emphasis placed on improving compliance and reducing regressivity without weakening its revenue role. Priority should be given to electronic invoicing, stronger audit trails, better filing enforcement, more transparent input tax credit systems, and a more coherent treatment of exemptions, while complementing these reforms with targeted transfers or similar expenditure-side measures to protect lower-income households. Corporate taxation also requires a more transparent and rules-based framework, with fewer discretionary incentives, regular publication of tax expenditure statements, and stronger monitoring of profit shifting and base erosion. In relation to wealth, the policy direction should be to bring concentrated assets more meaningfully within the fiscal system through instruments that are both administratively feasible and distributionally well justified. In this regard, inheritance taxation deserves particular priority. Compared with an annual net wealth tax, it is easier to administer because it arises at a specific legal moment when ownership must be established, and assets must be disclosed, making valuation and enforcement more manageable. It is also more defensible on equity grounds, as it targets large transfers of unearned advantage and 3

helps address the intergenerational transmission of wealth and privilege. If designed with sufficiently high thresholds and progressive rates, inheritance taxation can protect ordinary family transfers while ensuring only large assets or estates contribute more fairly to the public purse. Complementary measures should include better taxation of high-value secondary property, stronger asset declaration requirements, and improved registries of land and financial assets. More fundamentally, the reforms need to be anchored in a stronger conception of fairness and state credibility. Tax reform should be designed not only to raise revenue but also to strengthen both vertical and horizontal equity, so that those with greater ability to pay contribute more and similarly placed taxpayers are treated more equally. This, in turn, requires a modern data-driven tax administration, deeper inter-agency coordination, professionalised revenue institutions, and credible anti-corruption measures. It also requires a clearer link between revenue mobilisation and visible public spending outcomes in areas such as health, education, social protection, and local public services. In that sense, the policy challenge is not merely to improve collection efficiency. It is to rebalance the tax system in a way that enhances compliance, restores legitimacy, and helps build a fairer and more credible fiscal contract. 4

1. Introduction Domestic revenue mobilisation has emerged as one of the most pressing macro-fiscal challenges confronting Bangladesh. Despite maintaining relatively strong economic growth over the past decade, the country continues to collect markedly less tax revenue than what its economic structure and legal tax framework would suggest is achievable. This persistent gap between the legally implied tax liabilities embedded in existing tax laws and the revenue actually collected reflects deep inefficiencies in both tax policy design and tax administration. More broadly, it signals structural weaknesses in the fiscal system that constrain the state’s capacity to finance public goods, social protection programmes, and development priorities. A central feature of Bangladesh’s fiscal structure is the heavy reliance on indirect taxation, particularly the Value Added Tax(VAT). Indirect taxes constitute a large share of total tax revenue, while direct taxes, especially personal income taxation, account for a comparatively small proportion. This composition raises important concerns regarding both efficiency and equity. Because indirect taxes are levied on consumption rather than income, they tend to impose a relatively larger burden on households with lower incomes, who devote a greater proportion of their earnings to consumption expenditures. In the absence of sufficiently strong direct taxation and redistributive transfers, such a structure can lead to a fiscal system in which lower- and middle-income households bear a disproportionate share of the effective tax burden relative to their ability to pay. The narrow base of personal income taxation further amplifies this imbalance. Bangladesh’s personal income tax system generates very limited revenue relative to the size of the economy, reflecting both extensive exemptions and weak compliance. As noted in earlier research, the yield from personal income taxation remains strikingly low(Ahmed, 2015). Although the number of individuals required to obtain tax identification numbers has increased over time, the proportion of the population that actually files tax returns remains extremely small. Available estimates suggest that fewer than one per cent of the population submit income tax returns regularly(Khan et al., 2019). This situation reflects a combination of structural factors, including the large informal economy, limited administrative capacity, and weak enforcement mechanisms. At the same time, broader issues related to institutional trust and perceptions of fairness also appear to influence voluntary compliance behaviour. Against this backdrop, this study seeks to quantify the extent of Bangladesh’s unrealised revenue potential and to examine the distributional implications of the current tax structure. Moving beyond the conventional reliance on macro-level indicators, the paper uses micro-level data from the Household Income and Expenditure Survey(HIES) 2022 to estimate the legally implied liabilities for both personal income tax(PIT) and VAT under full compliance with the statutory tax framework. These estimates are then compared with actual tax collections reported in the National Board of Revenue’s(2023) Annual Report for fiscal year 2021–22 to derive estimates of the tax gap for these two major revenue sources. The use of household-level survey data also makes it possible to examine how the burden of these taxes is distributed across different income groups, thereby providing insights into the equity implications of the existing tax system. Much of the traditional literature on tax performance has focused primarily on structural or supply-side determinants of revenue mobilisation, such as per capita income, the share of agriculture in GDP, trade openness, and the size of the informal sector(Ahmed, 2024). While these factors undoubtedly shape 5

a country’s tax capacity, an exclusive focus on them can obscure other equally important drivers of tax performance. As Bird et al.,(2005) emphasised, tax compliance is also strongly influenced by demandside factors, including institutional quality, perceptions of fairness, and citizens’ willingness to comply voluntarily with tax laws. These factors are closely linked to broader governance conditions, including control of corruption, government effectiveness, and the rule of law. Consequently, weak domestic revenue mobilisation cannot be understood solely as a technical problem of tax policy design or administrative capacity; it is also deeply connected to issues of institutional credibility and the perceived legitimacy of the fiscal system. The analytical framework of this study, therefore proceeds along two complementary dimensions. First, the paper assesses the distributional consequences of Bangladesh’s current tax structure. Using household survey data, the analysis examines the extent to which the heavy reliance on consumption taxation contributes to a regressive distribution of the effective tax burden. At the same time, it considers how the limited reach of personal income taxation, together with extensive exemptions and weak enforcement, restricts the progressivity of the overall tax system. Second, the study explores the institutional and behavioural factors that shape tax compliance. Drawing on established theoretical frameworks in the tax compliance literature, including the economic deterrence model, social-psychological approaches, fiscal exchange theory, comparative fairness perspectives, and political legitimacy models, the paper reviews the broader institutional environment within which tax compliance decisions are made. These frameworks highlight how weak governance, low perceived probabilities of detection, widespread informal economic activity, and the social normalisation of evasion can jointly erode tax morale and reduce voluntary compliance. Finally, the paper considers policy options for strengthening revenue mobilisation while improving the equity of the tax system. In this context, wealth assumes particular importance, not only because it is typically far more concentrated than income, but also because concentrated wealth can serve as a powerful mechanism through which inequality is reproduced across generations. In a country such as Bangladesh, where concerns over rising inequality have become increasingly salient, the concentration of wealth deserves closer analytical and policy attention, both from the perspective of distributive justice and from that of fiscal capacity. A tax system that relies heavily on consumption taxes while leaving concentrated wealth only largely untouched is difficult to reconcile with the principle of ability to pay and may do little to counteract the intergenerational transmission of economic advantage. To examine this issue more directly, the study uses the asset-related information available in the HIES 2022 to develop indicative estimates of household wealth and its distribution. On that basis, the paper explores the potential role of wealth-related taxation in addressing the extreme concentration of wealth in Bangladesh and examines the possible revenue implications of imposing a wealth tax on the wealthiest segments of the population. Given the administrative and valuation constraints associated with annual wealth taxation in developing economies, the paper also considers the potential merits of inheritance taxation as a more administratively feasible instrument for enhancing progressivity, mitigating intergenerational inequality, and broadening the fiscal base. Overall, the analysis aims to contribute to the policy debate on domestic revenue mobilisation in Bangladesh by linking micro-level estimates of unrealised tax revenues with broader institutional explanations of tax compliance behaviour. By combining distributional analysis with insights from the political economy of taxation, the paper seeks to shed light on both the magnitude of the country’s revenue shortfall and the structural factors that continue to limit the effectiveness and fairness of the tax system. 6

2. Tax Capacity, Incidence, and Compliance: Evidence from the Literature Bangladesh’s tax performance has long attracted attention in the fiscal policy literature because the country consistently records one of the lowest tax-to-GDP ratios in the world despite registering relatively strong economic growth. The tax-to-GDP ratio has, for many years now, fluctuated between roughly 7 and 9 per cent, placing Bangladesh among the weakest performers not only in South Asia but also globally(Ahmed, 2024; Razzaque& Khan, 2021). Cross-country analyses of tax effort further suggest that Bangladesh collects significantly less revenue than would be predicted by its economic characteristics and legal tax structure. For example, according to UNESCAP(2014), Bangladesh’s unrealised tax revenue, referred to as the tax gap, exceeds 7.5 per cent of GDP. Such cross-country analyses typically compare actual revenue mobilisation with a country’s estimated tax capacity based on structural variables such as income levels, economic composition, and trade openness. These studies indicate that Bangladesh’s weak revenue performance reflects not only its stage of development but also substantial unrealised tax bases and weaknesses in tax administration and enforcement(Bouzouita, 2025; Fenochietto& Pessino, 2013) 1 . The implications of this persistent revenue shortfall are substantial. The planned expansion of the government in public expenditure as a share of GDP on health, education, rural development, and social protection has consistently fallen short, reinforcing the case for structural reform of public finances that increases both the equity of tax collection and the adequacy of public spending(Bangladesh Planning Commission, 2020). Several strands of the literature emphasise that Bangladesh’s fiscal challenge is not merely one of revenue inadequacy but also one of structural imbalance. In particular, the tax system relies heavily on indirect taxation, which accounts for about 66 per cent of total tax revenue, while direct taxation remains underdeveloped(National Board of Revenue, 2025; Razzaque& Khan, 2021). In the absence of sufficiently strong direct taxes and redistributive transfers, such a structure can generate both inefficiencies and equity concerns. Other studies emphasise that a weak rule-of-law, corruption, and perceptions of unfairness in the tax system can discourage voluntary compliance and undermine revenue performance(Etzioni, 1986; Jimenez& Iyer, 2016; Torgler et al., 2008). As a result, research on tax systems in developing economies has increasingly moved beyond purely economic models of taxation and has incorporated insights from behavioural and institutional perspectives to explain compliance behaviour. Within the Bangladesh context, a substantial body of research has examined the distributional incidence and coverage of personal income taxation. These studies consistently show that although the statutory personal income tax(PIT) structure is formally progressive, the effective PIT base remains extremely narrow in practice. Analyses based on household survey data indicate that only a very small fraction of individuals with taxable income actually appear within the effective tax net(Khan et al., 2019). In other words, while marginal tax rates increase with income according to the statutory schedule, the tax system 1 UNESCAP(2014) estimated Bangladesh’s tax gap using a cross-country panel regression of tax-to-GDP ratios on three structural variables: agricultural value added as a share of GDP, GDP per capita, and trade openness across 144 countries over the period 1990-2012, with regional dummies included. The estimated coefficients are used to generate a predicted tax-to-GDP ratio for each country, and the ratio of actual to predicted tax revenue is averaged over time to derive a country-specific efficiency score. The tax potential is then obtained by dividing the most recent actual tax revenue by this average efficiency ratio, and the tax gap is calculated as the difference between this potential and actual collections. For Bangladesh, using the Government Finance Statistics database of the IMF and augmented by data from CEIC, this method yields a tax potential of 18 per cent of GDP against an actual ratio of 10.5 per cent, implying a tax gap of 7.5 percentage points of GDP. 7

relies on a very small number of high-income taxpayers for most PIT revenue. This limited coverage reflects both weak horizontal coverage across potential taxpayers and a broader erosion of the PIT base, partly owing to the proliferation of exemptions, allowances, and loopholes that allow capital gains and other income streams to escape the tax net(Bangladesh Planning Commission, 2020; Khan et al., 2019; Razzaque et al., 2023; Thomas, 2023). A frequently cited explanation for the narrow PIT base in Bangladesh is the large size of the informal economy. Estimates suggest that roughly 85 per cent of the workforce operates in informal employment, where earnings are rarely documented and are therefore difficult for tax authorities to monitor(Moazzem et al., 2023). As a result, a significant share of economically active individuals remains outside the formal tax net. In addition to these structural constraints, administrative weaknesses also play an important role. Previous studies note that weak audit capacity and the continued reliance on manual tax assessment procedures create opportunities for underreporting and informal negotiation between taxpayers and tax officials(Ahmed, 2024). In such an environment, individuals with greater economic and political influence may find it easier to avoid or minimise their tax liabilities. The combination of the extremely narrow tax net, corrupt practices, and a large informal economy has collectively contributed to Bangladesh’s chronically weak fiscal performance(Razzaque& Khan, 2021). The heavy reliance on indirect taxation, particularly the Value Added Tax(VAT), has also been widely discussed in the literature on Bangladesh’s tax system. VAT constitutes a major component of domestic tax revenue and is collected at multiple stages of production and distribution. From a theoretical perspective, consumption taxes such as VAT tend to place a larger burden on lower-income households when evaluated relative to income, since poorer households typically spend a higher proportion of their income on consumption. Consequently, many studies describe VAT as regressive when assessed against income distribution(Blasco et al., 2023; Maboshe& Woolard, 2018; Razzaque et al., 2023). Although reform of the VAT system was a stated priority in the 7th Five Year Plan(7FYP), most notably through the planned implementation of the VAT Law of 2012, the reform did not materialise, leaving the structural regressivity of the indirect tax system largely unaddressed(Bangladesh Planning Commission, 2020). 2 This persistent regressive pattern may contribute to widening inequality if it is not offset by sufficiently progressive direct taxation or targeted social transfers. In addition, higher consumption taxes can limit the ability of poorer households to save or invest in human capital, potentially reinforcing longer-term income disparities(Maboshe& Woolard, 2018). The institutional mechanisms through which VAT is administered also affect its revenue performance. In principle, VAT is paid by consumers through prices, while firms act as intermediaries that collect the tax and remit it to the authorities. In practice, however, the effectiveness of this system depends heavily on compliance among firms. Evidence from Bangladesh suggests that many small and medium-sized enterprises(SMEs) either fail to register properly for VAT or do not submit regular returns, even when they charge VAT to consumers in market transactions(Faridy et al., 2014). Empirical studies based on key informant interviews with SMEs indicate that compliant firms often cite the complexity of VAT legislation 2 Many of the key tax reforms envisioned under the 7FYP, such as full implementation of the VAT Act 2012, establishment of a comprehensive and progressive personal income tax system, rationalisation of income tax exemptions, and modernisation of tax administration, were not effectively implemented(Bangladesh Planning Commission, 2020). As a result, tax revenue performance remained weak despite strong economic growth. These same reforms have been reemphasised under the 8FYP, with renewed focus on broadening the tax base, improving compliance, and modernising both VAT and income tax systems to strengthen domestic resource mobilisation. 8

and the high administrative costs of compliance as major obstacles. By contrast, non-compliant firms frequently report that their behaviour is influenced more by their interactions with tax officials and the possibility of resolving disputes informally rather than by formal legal requirements or penalties(Faridy et al., 2014). International evidence also suggests that VAT compliance costs can be relatively high compared to other forms of taxation, and these costs tend to fall disproportionately on smaller firms that lack economies of scale in accounting and administrative capacity. Consequently, small businesses may remain informal or underreport their activities to avoid compliance burdens(Faridy et al., 2014). This dynamic creates a situation in which consumers may pay VAT embedded in prices, while a portion of the tax collected by firms fails to reach the tax authorities. Such leakages help explain why Bangladesh can simultaneously experience a relatively high consumption tax burden at the household level and yet collect comparatively low VAT revenues. A further constraint on revenue mobilisation arises from weaknesses in governance and tax administration. Several studies emphasise that corruption and administrative discretion can significantly erode the effectiveness of tax systems in developing economies. In Bangladesh, discretionary incentives and tax exemptions granted through instruments such as Special Regulatory Orders(SROs) have been widely discussed as a source of revenue loss and policy distortion(Ahmed, 2024). While such measures are often justified as tools for promoting investment or industrial development, their proliferation can reduce transparency, complicate tax administration, and narrow the effective tax base. 3 The International Monetary Fund(2025) has identified the rationalisation of such exemptions as a central reform priority, noting that the country’s low tax-to-GDP ratio underscores the urgent need for reforms to create a fairer and more transparent system by rationalising exemptions, enhancing compliance, and separating tax policy from administration. 4 According to the IMF staff estimates based on the data from the National Board of Revenue, removing tax exemptions through repealing or phasing out selected SROs and eliminating certain provisions in the Income Tax Act 2023 could yield more than 0.1 per cent of GDP in FY26. Moreover, the process through which exemptions and special tax treatments are granted may create opportunities for political lobbying and rent-seeking. When tax policies become fragmented through numerous discretionary interventions, both compliance and enforcement can be weakened. In this context, malpractices within tax administration can further undermine taxpayer confidence in the fairness of the system. The existing literature identifies political economy factors that prompt rentseeking activities, encourage tax evasion and corrupt practices as among the most important barriers to reform in Bangladesh, arguing that a strong commitment to tackling these factors is a prerequisite for any durable improvement in domestic resource mobilisation(Razzaque& Khan, 2021). 3 Effective tax base is the actual amount of income which is subject to taxation, after accounting for all legal deductions, exemptions, credits, and other tax provisions. 4 The IMF recommends a comprehensive reform agenda to enhance revenue mobilisation in Bangladesh, centred on broadening the tax base, rationalising exemptions, and strengthening tax policy across VAT, customs, and income taxes. It also emphasises administrative modernisation through digitalisation, including e-filing, e-invoicing, improved taxpayer data systems, and risk-based auditing to reduce evasion and improve compliance. Institutional reforms, such as separating tax policy from administration, strengthening governance, and enhancing transparency through tax expenditure reporting, are also highlighted as critical. These measures are embedded within a Medium- and Long-Term Revenue Strategy aimed at sustainably increasing the tax-to-GDP ratio and creating a more efficient, equitable, and transparent tax system. These reform measures also inform the policy implications discussed in Section 5 of this paper, aimed at addressing the domestic revenue mobilisation constraints highlighted in this study. 9

The existing literature highlights several interrelated explanations for Bangladesh’s weak tax performance: the narrow coverage of direct taxation, heavy reliance on indirect taxes, widespread informality, administrative weaknesses, and governance challenges that undermine voluntary compliance. These insights provide an important foundation for analysing the structure of Bangladesh’s tax system and for assessing the distributional and institutional factors that contribute to the country’s persistent revenue shortfall. 10

3. Data and Estimation Framework This study relies primarily on the Household Income and Expenditure Survey(HIES) 2022 conducted by the Bangladesh Bureau of Statistics(BBS). 5 The survey provides detailed household-level information on income, consumption expenditure, assets, and demographic characteristics, making it a useful dataset for examining both the potential tax liabilities of households and the distributional implications of the existing tax structure. Using the HIES microdata, we estimate the potential personal income tax(PIT) liabilities and the value-added tax(VAT) burden associated with household consumption in 2022. Unlike approaches that infer tax potential indirectly from cross-country relationships between tax performance and broad structural characteristics of the economy, the present study adopts a countryspecific micro-simulation approach grounded in Bangladesh’s own statutory tax framework. The advantage of this method is that it derives the revenue benchmark from the taxes that would be due under existing laws if households and firms complied fully, rather than from the average performance of structurally comparable economies. In this sense, the exercise undertaken here is better understood as an estimate of the legally implied revenue potential under the current tax regime. Such an approach is particularly useful for identifying the extent of unrealised revenue within specific tax instruments, locating the sources of leakage more precisely, and examining how the burden of taxation is distributed across income groups. It therefore provides a more policy-relevant basis for analysing compliance gaps, administrative weaknesses, and distributional consequences than benchmarks derived solely from crosscountry structural averages. For estimating the potential personal income tax(PIT) liability of all earners, first, the annual income of each earner in a household was calculated using the standard income aggregation methodology adopted by the Bangladesh Bureau of Statistics. This approach consolidates income from multiple sources, including wage and salary earnings, self-employment income, agricultural income, remittances, and other transfers. In HIES 2022, wage income is reported at the individual level. However, other income sources, such as agricultural income, business profits, and remittances, are reported at the household level. To estimate each individual’s total income, household-level non-wage income was allocated equally among all active earners in the household. This allocated share was then added to each earner’s reported wage income to approximate the total individual income of each earning member. Although this approach inevitably simplifies intra-household income distribution, it provides a tractable method for estimating potential individual tax liabilities using household-level survey data. The potential PIT liability of each earner, that is, the amount of tax payable under full compliance with the statutory tax rates imposed in Bangladesh, was calculated using the tax brackets specified in the National Board of Revenue(NBR) Income Tax Paripatra for Fiscal Year 2021–22. 6 For example, if a male 5 In the official household income estimates of the Bangladesh Bureau of Statistics(BBS), household incomes were defined as money inflows from various sources into the household occurring during the last 12 months. For instance, labour income includes wages and benefits of household members who are day laborers or employees with missing values imputed using medians by stratum, area, or industry. Business income equals net revenue after deducting expenditures. Agricultural income accounts for crop, livestock, and related production with outliers and missing data adjusted using median imputation, and non-labour income covers rents, dividends, interests, social transfers, and remittances received in cash or kind. See Ahmed et al.(2019) for an elaboration of the estimation method. 6 According to the Income Tax Paripatra for the fiscal year 2021-22 from the National Board of Revenue, no income tax was payable for the first BDT 3,00,000, 5 per cent for the next BDT 1,00,000, 10 per cent for the next BDT 3,00,000, 15 per cent for the next BDT 4,00,000, 20 per cent for the next BDT 5,00,000, and 25 per cent on any remaining income. For individuals who are females or older than 64 years old, the tax-free income threshold limit is BDT 3,50,000. These brackets were employed for the income tax liability estimation for each earner of the household. 11

individual earned BDT 450,500 in a year, his tax liability under the Tax Paripatra would be calculated as follows: 0 per cent on the first BDT 300,000; 5 per cent on the next BDT 100,000; and 10 per cent on the remaining BDT 50,500. Finally, to estimate the potential PIT revenue achievable at the household level, total potential PIT liability was calculated at the household level by summing the liabilities of all earners within each household. Each household’s estimated potential tax liability was then multiplied by the corresponding household weights provided in HIES 2022 to obtain the population-level potential tax liability represented by each sample household. Finally, the population-level expanded tax liabilities of all sample households were aggregated to approximate the national potential PIT revenue. It is important to recognise that the rich and very rich are typically underrepresented and often underreport their income. This underrepresentation and income underreporting in household-level surveys may stem from intentional misreporting, privacy concerns, recall difficulties, and limited access to high-income individuals. For example, a study by the Inter-American Development Bank covering 18 Latin American countries finds that individuals in the top decile of survey-based income distributions report earnings comparable to well-educated professionals rather than to the truly highest-income households, suggesting that the upper tail of the distribution is not adequately captured(Székely& Hilgert, 1999). On the other hand, income is also frequently underestimated among lower-income households, particularly those relying on informal or agricultural activities. In such cases, recall error, irregular earnings, lack of written records, and the complexity of separating household and business finances contribute to substantial underreporting(Carletto et al., 2022; Meyer& Sullivan, 2003). Furthermore, HIES 2022 captures only around 33 per cent of GDP in terms of total household expenditure, whereas private consumption constituted more than 68 per cent of GDP in FY22. These distortions would imply that the simulated potential revenues estimated from HIES 2022 are likely to be conservative estimates, and that the true tax gap in fact should be considerably larger than the estimate. To partially address these limitations, a HIES-consistent GDP equivalent is constructed by dividing total household expenditure estimated from HIES 2022 by the actual private consumption-to-GDP ratio, so that potential tax revenues can be expressed as credible percentages of GDP. The potential PIT-to-GDP ratio is then estimated using this GDP equivalent as the denominator, and further adjusted upward to account for the systematic underrepresentation of high-income individuals in HIES relative to NBR’s administrative tax return data. The methodology underlying both adjustments is elaborated in detail in Appendix A1. The VAT burden of households was estimated using the consumption expenditure data reported in the HIES 2022 survey. The first step involved identifying VAT-liable consumption items by excluding goods and services that are exempt under The Value Added Tax and Supplementary Duty Act 2012. In Bangladesh, most VAT-liable goods and services are subject to the standard VAT rate of 15 per cent, although a number of items are taxed at reduced rates, including 1.5 per cent, 2 per cent, 2.4 per cent, 4.5 per cent, 5 per cent, 7.5 per cent, and 10 per cent(Price water house Coopers(PwC), 2025). Using the expenditure information in HIES 2022, each household’s consumption was classified according to the applicable VAT rate for the relevant goods and services. It was assumed that VAT is fully passed on to consumers through prices and that the expenditures reported in the survey are VAT-inclusive. Therefore, for each category of goods and services subject to the same VAT rate, VAT-inclusive household spending was converted into VAT-exclusive spending by dividing total spending by one plus the applicable VAT rate. The VAT amount embedded in the reported expenditure was then obtained by multiplying the VAT12

exclusive expenditure by the corresponding tax rate. This procedure was applied across all VAT-liable consumption categories, after which the VAT amounts were summed to obtain each household’s total annual VAT burden. National potential VAT revenue was then estimated by applying the HIES sampling weights to household-level VAT burdens and aggregating the weighted values across all households in the dataset. The potential VAT-to-GDP ratio is then estimated by the ratio of the national potential VAT revenue estimate from HIES to the GDP equivalent calculated in Appendix A1. For the potential VAT revenue estimations, the VAT rates applicable to domestically produced goods have been used, and the VAT rates applied to imported goods have not been considered because the survey does not distinguish between the consumption of imported and domestically produced goods; the estimated potential VAT revenue actually reflects the VAT revenue achievable from domestic goods and services under full compliance. 7 Furthermore, under the Value Added Tax and Supplementary Duty Act, 2012, a large number of goods are listed as VAT-exempt only when sold in unwrapped or uncanned form, such as milk, vegetables, raw meat, fish, and spices. As the survey does not indicate whether such goods were purchased in wrapped or unwrapped form, they have been assumed to be unwrapped and no VAT has been applied to them. In addition, many luxury purchases made by high-income households are unlikely to be fully captured in the survey, which may further contribute to the underestimation of potential VAT revenue. To assess the magnitude of unrealised tax revenues, the estimated potential revenues from PIT and VAT were compared with the actual collections reported by the National Board of Revenue for fiscal year 2021–22. The official data on PIT and VAT revenue were obtained from the NBR Annual Report for FY2021–22. The difference between estimated potential revenue and actual collections provides an approximate indication of the revenue gap under the existing legal framework. In addition to estimating aggregate revenue potential, the analysis examines how the burden of taxation varies across income groups. For this purpose, households were grouped into income deciles, where the lowest decile represents the poorest 10 per cent of households and the highest decile represents the richest 10 per cent. A key methodological consideration in such distributional analysis concerns the choice of welfare metric used to rank households. While total household income is often used for this purpose, relying solely on household income may distort welfare comparisons because it assigns equal weight to households of different sizes and compositions(Thomas, 2023). An alternative approach is to use per capita household income, which divides total household income by the number of household members. However, this method implicitly assumes that each additional household member requires a proportional increase in income to maintain the same standard of living. In reality, this assumption is unlikely to hold because households benefit from economies of scale in consumption. For example, a three-person household does not require three times the housing space or electricity consumption of a single-person household. Furthermore, the resource needs of household members vary across age groups and economic activity status. Children typically require fewer resources than adults, while working adults may incur additional costs related to transportation, clothing, and other employment-related expenditures(OECD, 2013). To address these issues, this study employs equivalised household income as the welfare metric for ranking households. Equivalised income adjusts total household income using an equivalence scale 7 Later in this study, the potential domestic VAT revenue is compared with the NBR’s domestic VAT revenue collections. 13

that accounts for differences in household size and composition while allowing for economies of scale in consumption. Two widely used equivalence scales are the square-root scale and the OECD-modified scale. In this study, the OECD-modified scale is applied, which assigns a weight of 1 to the first adult household member, 0.5 to each additional household member aged 14 or above, and 0.3 to each child under the age of 14(Thomas, 2023). This adjustment allows income levels to be expressed in terms of the income required by a single-person household to achieve the same standard of living as the household in question. Equivalised household income can therefore be interpreted as a proxy for the economic resources available to each individual within the household. The distributional incidence of taxation was then evaluated by examining the tax burden borne by each income decile. To assess the progressivity or regressivity of the tax system, we calculated the relative tax burden of each decile using the following measure: 8 Decile’s share of tax in total tax revenue Relative tax burden= Decile’s share of income in total income A value greater than 100 indicates that a particular income group contributes a larger share of total tax revenue than its share of national income. If the index increases as one moves from lower to higher income groups, the tax structure can be interpreted as progressive. Conversely, if the index declines with rising income groups, the tax structure is considered regressive. Using this framework, the study evaluates the distribution of both PIT and VAT burdens across income groups. In addition, household expenditure estimates derived from HIES 2022, following the official BBS methodology, are used to calculate the share of consumption expenditure allocated to VAT across different income deciles. The analysis also explores regional differences in tax incidence by comparing urban and rural households. This allows an examination of how variations in household income levels, consumption patterns, and economic activities translate into differences in effective tax burdens. Beyond household-level comparisons, the analysis also highlights broader regional patterns such as differences in the concentration of taxable capacity and the relative contribution of different tax instruments to inequality in urban and rural areas. In addition to the quantitative analysis, the study incorporates qualitative insights from the literature to examine how institutional and governance factors influence tax compliance behaviour in Bangladesh. This includes the role of administrative capacity, enforcement mechanisms, corruption, and perceptions of fairness in shaping compliance decisions. By integrating these insights with the empirical estimates derived from HIES data, the analysis seeks to provide a more comprehensive understanding of the structural drivers of Bangladesh’s weak revenue mobilisation. Finally, the study evaluates the potential role of wealth taxation as an instrument for enhancing equity and expanding the fiscal base. An attempt has been made to estimate the potential revenue that could be generated from applying different tax rates to the wealth held by the richest households. To identify these groups, households were ranked according to their equivalised wealth and grouped into wealth deciles. For this purpose, household wealth was estimated using information available in the HIES 2022 dataset. In this study, a household’s assets are referred to as the household’s wealth, which is the sum of financial assets(savings, stocks, bonds, etc.), real property, hard assets(e.g., automobiles, jewelry), business 8 This measure of relative tax burden is adopted from Faridy& Sarker(2011). 14

capital(machines, equipment), and durable goods, minus any outstanding personal debt(Sherraden, 1991). 9 The household asset base includes owner-occupied housing, land, business and agricultural assets, livestock, durable goods, insurance holdings, savings, jewellery, stocks, bonds, and other reported assets. Most of these values are directly available in HIES 2022 through respondents’ self-reported assessments. Since the survey reports the area but not the value of operational land, its monetary value is imputed using median unit land prices estimated at the upazila and, where necessary, district levels, separately for rural and urban areas, drawing on information reported for non-operational land. The resulting estimates provide an indicative measure of household wealth distribution in Bangladesh, although they should be interpreted with caution given data limitations, including incomplete coverage of additional residential properties and partial information on savings. Further details of the estimation procedure are provided in Appendix A2. Overall, the study adopts a mixed-methods approach that combines cross-sectional micro-level analysis using the HIES 2022 survey with qualitative insights drawn from the existing literature. This approach allows the analysis to capture both the quantitative dimensions of tax incidence and revenue potential, as well as the institutional and behavioural factors that shape tax compliance and perceptions of fairness within Bangladesh’s fiscal system. 9 Sherraden(1991) listed tangible assets as including(1) money savings;(2) stocks, bonds, and other financial securities; (3) real property;(4) hard assets other than real estate, such as automobiles, jewelry, art, and collectibles;(5) machines, equipment, tools, and other tangible components of production;(6) durable household goods;(7) natural resources; and (8) copyrights, patents, and other intellectual property. Intangible assets include(1) access to credit,(2) human capital, (3) cultural capital,(4) informal social capital,(5) formal social capital or organizational capital, and(6) political capital. This list is not definitive, but illustrates that the term assets(along with a related term, capital) has been broadly applied. 15

4. Findings 4.1 Overview of the current tax system of Bangladesh A well-functioning tax system is expected to do more than raise revenue. It should do so in a manner that is efficient, administratively workable, and reasonably consistent with accepted principles of fairness. In the public finance literature, fairness is typically understood through the lenses of horizontal and vertical equity. Horizontal equity requires that individuals with similar economic capacity be treated similarly, while vertical equity implies that those with greater ability to pay should contribute more, both in absolute terms and, in progressive systems, as a proportion of income. Progressive taxation, therefore, is generally seen not only as an instrument for resource mobilisation but also as a means of supporting a more equitable social contract and financing public goods, services, and social protection programmes (Alm, 2024; Razzaque et al., 2024). Against this backdrop, we begin by examining the extent to which Bangladesh’s current tax structure aligns with these principles. Figure 4.1: Cumulative share of household income, income tax, and VAT held/to be paid by the cumulative share of households Source: Authors’ analysis using the Household Income and Expenditure Surveys, 2022. Note: The distribution curve for income shows the cumulative distribution of national income(sum of weighted household income estimated from HIES 2022) across cumulative households ranked from poorest to richest based on their equivalized income. The line of equality represents an income distribution in which all the households possess the same level of equivalized income, that is, the ability to have the same standard of living. Figure 4.1 plots the cumulative distribution of weighted household income, potential personal income tax (PIT) liability, and VAT burden across households ranked by equivalised income. The figure is informative because it shows how the burdens of the two principal taxes examined in this study are distributed 16

relative to the underlying income distribution. If the system were proportional, in the sense that each group contributed a share of tax equal to its share of total income, the concentration curves for PIT and VAT would coincide with the Lorenz curve. 10 A progressive tax would lie below the Lorenz curve, indicating that higher-income groups contribute more than proportionately, while a regressive tax would lie above it. Figure 4.2: Distribution of income tax and VAT relative to the distribution of household income Source: Authors’ analysis using the Household Income and Expenditure Surveys, 2022. Note: The graph shows the cumulative distribution of potential VAT and PIT burden across percentiles of national income. If the VAT and PIT were imposed proportionally with income distribution, then the distribution curves for potential VAT and PIT would lie on the line of equality. In this graph, a large number of households are clustered in the lower segment of the income distribution(x-axis), while relatively few households represent a disproportionately large share of total income at the top. This pattern is consistent with substantial income concentration and also reflects the underrepresentation of high-income households in HIES 2022. The bottom 52 per cent of households account for only 20 per cent of total national income, while it takes 84 per cent of the households to constitute just half of the country’s aggregate earnings. That is, the top 16 per cent of households represent half of the total national income on the horizontal axis. The top 20 per cent of the total national income is represented by only 2 per cent of the households. The results point to a sharp contrast between PIT and VAT. The richest 10 per cent of households account for more than 40 per cent of total income and are estimated to bear over 93 per cent of total potential 10 The Lorenz curve for income shows the cumulative share of income received by the cumulative share of income units when those units are ranked from poorest to richest. The further the Lorenz curve lies from the line of equality, the greater the degree of income inequality. Traditionally, it has been used as a simple visual tool to describe inequality in income or wealth. Over time, the idea has been generalised to compare the distributions of other economic variables as well. These generalised Lorenz curves for other indicators are called concentration curves, and the Lorenz curve is just the concentration curve for income. In Figure 4.1, the income Lorenz curve is demonstrated to compare the distribution of household income with tax concentration curves for PIT and VAT. 17

PIT liability. On paper, therefore, the PIT structure appears highly progressive. Yet this progressivity is achieved through an extremely narrow base rather than through broad-based vertical equity. By contrast, the same top decile bears only around 23 per cent of the total VAT burden despite holding more than 40 per cent of the national income. This indicates that VAT is distributed much more heavily towards lower- and middle-income households when assessed relative to income, and is therefore regressive in incidence. This pattern becomes clearer in Figure 4.2 and in the estimated relative tax burden indices reported in Table 4.1. The relative PIT burden rises sharply towards the top of the distribution. Nationally, it increases from 31.84 in the ninth decile to 229.09 in the top decile, implying that the highest-income households bear a share of PIT liability far in excess of their income share. This confirms that PIT is progressive in statutory incidence. At the same time, the magnitude of the jump also reveals how thin the effective tax base is. Much of the middle and upper-middle distribution contributes relatively little, leaving the system heavily dependent on a narrow group at the very top. The VAT pattern runs in the opposite direction. Lower-income households pay a much larger share of total VAT relative to their share of income, while the relative burden declines steadily as one moves up the income distribution. The index exceeds 400 for the poorest decile and falls below 60 for the richest. This is strong evidence of regressivity. Even though exemptions and reduced rates are intended to soften the burden on essential consumption, the overall structure remains regressive, which is consistent with earlier findings for Bangladesh(Faridy& Sarker, 2011). Table 4.1: The estimated index of relative tax burden Decile 1 Decile 2 Decile 3 Decile 4 Decile 5 Decile 6 Decile 7 Decile 8 Decile 9 Decile 10 Relative Personal Income Tax Burden National Urban Rural 0 0 0 0 0.04 0 0.02 0.02 0 0.08 0.62 0 0.3 1.38 0.23 1.86 3.99 1.19 4.34 9.38 4.08 13.03 21.93 12.33 31.84 53.03 35.49 229.09 196.34 270.12 National 410.63 230.46 174.96 153.09 135.3 131.93 107.99 103.25 86.48 57.82 Relative VAT Burden Urban 365.45 211.02 185.54 161.3 144.21 125.92 120.34 104.61 101.09 59.16 Rural 446.7 229.39 173.71 151.38 130.14 121.04 111.25 95.01 85.9 47.56 Source: Authors’ analysis using the Household Income and Expenditure Surveys, 2022. Note: A value greater than 100 indicates that a decile contributes a larger share of total taxes than its share of total income. The results show that, at the national as well as regional level, only the richest decile has a relative PIT burden greater than 100. This implies that no other income group are supposed to contribute a higher share of potential personal income tax relative to its income share. In contrast, all lower and middle-income deciles are supposed to bear a higher share of total potential VAT burden than their respective income shares, while only the richest one or two deciles bear a tax share lower than their income share. Figure 4.3 provides a complementary view by expressing tax burdens as shares of household income and expenditure. Here too, the contrast is striking. The richest 10 per cent of households are found to be 18

liable to pay about 11.03 per cent of their aggregate income as PIT, whereas the second-richest decile is liable for only around 4.12 per cent. In the bottom four deciles, PIT liability is close to zero because very few earners cross the taxable threshold. This again shows that the statutory PIT structure is progressive, but also that the effective base is extraordinarily narrow. Figure 4.3: Distribution of tax burdens as shares of household income as well as expenditure Source: Authors’ analysis using the Household Income and Expenditure Surveys, 2022. Note: In the bottom two deciles, there are no earners with taxable income. In the next two deciles, there are so few households liable to pay personal income taxes that the total PIT liable to these groups as a share of the aggregate income is near 0. That narrowness is reinforced by the structure of exemptions and credits. The Sixth Schedule of the Income Tax Act, 2023 excludes a range of major income flows from the tax base. These include pensions, gratuity, provident and superannuation fund receipts, workers’ participation funds, and various forms of capital and financial income. Partial or full exemptions also apply to certain dividends, interest from selected savings instruments, repatriated foreign income, some agricultural income, and specific SME and export-related incomes. 11 In addition, tax credits for investments and donations are concentrated in instruments that are more accessible to formal, better-off households. The cumulative effect is to reduce the breadth of the PIT base and increase reliance on a relatively small segment of top earners. The PIT base is further squeezed through tax credits for investments and donations in eligible items. 11 PIT base erosion is deepened by a long list of fully exempt incomes under the Sixth Schedule of the Income Tax Act, 2023. Major retirement flows such as government-approved gratuity(up to Tk 2.5 crore), pensions, and provident fund and superannuation fund receipts are excluded from taxable income. Financial income exemptions include returns from wage earners development bonds and foreign currency bonds(US dollar, euro, and pound sterling), as well as interest from pensioners’ savings certificates up to Tk 5 lakh in accumulated investment. Additional exemptions cover repatriated foreign income, small agricultural income(up to Tk 2,00,000 for farmers with no other significant income), certain SME incomes, income from handicraft exports, and income of indigenous communities within the Chittagong Hill Tracts. Taken together, these provisions carve out a sizeable portion of government revenue achievable from personal income taxes, shrinking the effective PIT base relative to total household income. 19

Since these eligible items are mostly formal financial products like life insurance, provident or pension savings, savings certificates, listed securities and approved donations, the people who benefit most are richer, formal-sector households who can afford these investments and keep records documented. As a result, the same high-income groups who earn taxable incomes and can pay taxes at higher marginal rates make the tax system rely even more on a small number of top earners. The VAT results require more careful interpretation. As per the estimates from the HIES 2022, if all individuals complied fully with the legal taxation framework, then the poorest 10 per cent of households in Bangladesh would end up spending around 20.37 per cent of their aggregate income on VAT, whereas the richest 10 per cent would pay only 2.8 per cent of their aggregate income as VAT. These findings are demonstrated in Figure 4.3. This occurs due to the fact that low-income earners spend larger portions of their earnings on consumption; therefore, they pay a greater portion of their income in indirect taxes compared to their wealthier counterparts(Razzaque et al., 2024). However, this estimation of the VATto-income ratio may be overstated due to the underreporting of income in the survey. For example, the average income-to-consumption ratio for the full sample is 0.85, which implies that, on average, reported incomes are substantially lower than reported household expenditures in HIES 2022. This discrepancy is even more pronounced among the poorest decile, whose total consumption expenditure is almost five times higher than their reported total income. Given that poorer households typically have limited access to credit or assets to sustain persistent dissaving, such large gaps are widely interpreted as evidence of significant income underreporting in the household-level surveys rather than dissaving. 12 Figure 4.4: Weighted average household income levels, and shares of income tax and VAT by(a) rural household deciles and(b) urban household deciles (a) Rural(b) Urban Source: Authors’ analysis using the Household Income and Expenditure Surveys, 2022. Note: To generate these graphs, deciles are constructed separately for rural and urban households by first splitting the sample by region and then ranking and dividing households into ten income groups from the poorest to the 12 Using a large cross-country database of household income surveys in low- and middle-income countries, a study found that the extent of underreporting is closely associated with the source of income, with agricultural income suffering more from underreporting than any other component(Meyer& Sullivan, 2003; Carletto et al., 2022). Income data collected in household-level surveys are also highly vulnerable to recall error as respondents are often required to report earnings from multiple and sometimes irregular sources over a reference period. In contexts where households do not maintain written records and rely on memory, particularly for seasonal or self-employment activities, misreporting is very likely to take place. This issue is especially relevant for Bangladesh, where a wide majority of households depend directly or indirectly on agricultural activities. As a result, reported income is likely to be substantially distorted due to both recall problems and the inherent complexity of agricultural accounting. 20

richest based on their equivalised incomes within each region. The blue bars show the weighted average household income(in thousand taka) for each decile, using the HIES 2022 household weights, for rural areas in(a) and urban areas in(b). The orange and pink lines plot the ratio of PIT liability and VAT burdens to their expenditures for the corresponding deciles. Figure 4.5: Weighted average household expenditure, and VAT as a share of total expenditure by(a) rural household deciles and(b) urban household deciles (a) Rural(b) Urban Source: Authors’ analysis using the Household Income and Expenditure Surveys, 2022. Note: To generate these graphs, deciles are constructed separately for rural and urban households by first splitting the sample by region and then ranking and dividing households into ten income groups from the poorest to the richest based on their equivalised incomes within each region. The blue bars show the weighted average household expenditure(in thousand taka) for each decile, using the HIES 2022 household weights, for rural areas in(a) and urban areas in(b). The pink lines show the ratio of VAT burdens to expenditures for all income groups. When VAT is measured relative to expenditure instead of income, the burden appears broadly proportional across the income distribution. This pattern is expected, since VAT is fundamentally a tax on consumption. Lower-income households pay a larger share of income as VAT, not because the tax rate on commodities differs across people, but because they devote a much larger fraction of their income to consumption. This is precisely why consumption taxation often conflicts with the ability-to-pay principle, especially in settings where redistributive transfers are weak(Faridy& Sarker, 2011). Although exemptions, reduced rates, turnover thresholds, and differential treatment of essential items are intended to mitigate these effects(Bird, 2013), the evidence suggests that VAT in Bangladesh remains regressive overall. This is consistent with the finding that increases in indirect taxation are associated with higher poverty incidence in Bangladesh(Razzaque et al., 2024). The urban-rural breakdown adds another layer to the incidence story. Figures 4.4 and 4.5 show that income inequality is more pronounced in urban areas, where the upper deciles pull away much more sharply. Rural households have lower incomes across all deciles, and the rural PIT base is even narrower than the urban one. Indeed, the bottom six rural deciles contain virtually no households with taxable income. The richest rural decile faces an estimated PIT burden of 6.87 per cent of income, compared with 14.64 per cent for the richest urban decile. This suggests that, structurally, rural PIT revenue depends on a very small number of high-income households. 21

At the same time, estimated VAT burdens are considerably higher among poorer rural households(Figure 4.4). The poorest 10 per cent in rural areas pay more than 23 per cent of their income in VAT, compared with around 16.5 per cent among the poorest urban households. When measured relative to expenditure, however, VAT remains fairly stable across deciles in both regions, again indicating that the main source of regressivity lies in the underlying consumption-income relationship rather than in sharply differentiated effective rates across the spending distribution. The slight upward drift in VAT burden at the top of the urban distribution may reflect greater spending on higher-value goods and services subject to VAT. Although the core empirical analysis in this paper focuses on PIT and VAT, it is also useful to briefly consider the broader tax structure, including corporate income taxation(CIT). Statutory CIT rates in Bangladesh are relatively high by regional standards. In FY2021–22, according to the Tax Foundation’s data, the average CIT rate was about 32.5 per cent, exceeding that of several neighbouring South Asian economies. Yet the effective corporate tax base remains narrow. Bangladesh’s economy is characterised by a large informal sector, a dominant SME presence, and a concentration of formal corporate activity in a limited number of sectors, notably ready-made garments. With informal activities accounting for 84.9 per cent of employment and around 30.2 per cent of GDP(Bangladesh Bureau of Statistics, 2023; Moazzem et al., 2023), only a relatively small set of registered firms contributes meaningfully to CIT revenues. High exemption thresholds and widespread incentives, exemptions, tax holidays, and credits further erode the base, while the capacity of elite firms to evade liabilities through underreporting and informal dealings with tax authorities further narrows the corporate tax base. For instance, eight types of industrial companies, including jute, textile, knitwear, and poultry industries as well as private educational institutes, enjoy reduced tax rates ranging between 3 per cent and 15 per cent based on various criteria. In the interest of industrialisation and investment, tax holidays have been continuing in the tax regime of Bangladesh since its liberation in 1971. The benefits of these tax holidays are enjoyed mainly by the garment industry. However, it has been argued in the literature that the growth of the RMG sector in Bangladesh has been driven largely by external factors. The cumulative fiscal cost of these preferential treatments is substantial. The National Board of Revenue estimates that, for the 2021–2022 fiscal year, total corporate income tax(CIT) expenditure in Bangladesh was Taka 71,394 crore, equivalent to around 1.8 per cent of GDP(National Board of Revenue, 2024). The CIT expenditure refers to the revenue forgone due to deviations from the benchmark tax system, including reduced rates, exemptions, deductions, and tax holidays granted to specific sectors or firms. It underscores how deeply embedded preferential treatment is within Bangladesh’s corporate tax architecture and how significantly it constrains the effective revenue yield relative to what the statutory framework would otherwise imply. 4.2. Assessment of tax revenue performance Bangladesh’s weak revenue effort becomes even more evident when viewed in comparative perspective. As shown in Figure 4.6 the country’s tax-to-GDP ratio stood at 7.3 per cent in 2023, well below many regional comparators and far short of thresholds often cited in the literature as broadly consistent with financing sustainable development objectives(Choudhary et al., 2024). This underperformance matters not merely as a descriptive statistic. It reflects a chronic shortfall in domestic fiscal capacity. 22

Figure 4.6: Tax-to-GDP ratios in Asia and Pacific Countries, 2023 Source: Organisation for Economic Cooperation and Development(OECD), 2025; Revenue Statistics in Asia and the Pacific 2025, https://oe.cd/revstatsap2025. Note: For Australia and Japan, data for 2022 are derived, as for these countries, 2023 data were not available. To estimate the average tax-to-GDP ratio for South Asia, Nepal’s 2021 data were used, derived from World Bank data, and India’s 2022 data was derived from Khan et al.(2019). The picture sharpens further when actual collections are compared with the legally implied revenue that could, in principle, be obtained under full compliance. In FY2021–22, NBR’s reported PIT collection was equivalent to only 0.98 per cent of GDP. By contrast, our HIES-based simulation suggests that potential PIT revenue under full compliance with the statutory schedule could have reached at least 4.85 per cent of GDP. This implies an approximate PIT gap of 3.87 per cent of GDP. In other words, if all citizens earning taxable income had filed returns in accordance with the legal tax schedule, domestic tax revenue would have increased by about 3.87 per cent of GDP. The corresponding VAT gap is also large. The NBR’s domestic VAT collection in FY2022 was 1.8 per cent of GDP, whereas the estimated potential domestic VAT revenue amounts to around 3.82 per cent of GDP. Thus, the domestic VAT revenue foregone due to the lack of compliance was at least 2.02 per cent of GDP. These findings are summarised in Table 4.2. Table 4.2: Summary of tax revenue performance in FY22 Actual Collection Potential Collection Tax gap Personal Income Tax(PIT) (Domestic) Value Added Tax(VAT) Corporate Income Tax(CIT) % of total tax 12.90 23.70 19.32 % of GDP(a) 0.98 1.80 1.46 % of GDP (p) 4.85 3.82 % of GDP (p-a) 3.87 2.02 Source: Author’s estimation from HIES 2022, Annual Report 2021–22 of the National Board of Revenue(NBR), Bangladesh Bureau of Statistics Note: Taxes withheld at source are included in the personal income tax. In the values of actual collection of taxes as a percentage of total tax, the total tax amount is actually the total amount of tax collected by NBR in FY22, which is BDT 301,050.34 crore. The tax revenues and tax gaps are presented as percentages of GDP at current market 23

prices, which is BDT 3,971,716.40 crore. Potential corporate tax calculation is beyond the scope of this paper. In this table, actual revenue collections of PIT and VAT as percentages of GDP are referred to as“a” and potential revenue collections as percentages of GDP are referred to as“p”. Hence,“p-a” infers estimated tax gaps, the difference between the actual and potential collections. The higher(lower) the value of the tax gap, the higher(lower) the amount of domestic revenue loss due to factors such as non-compliance, tax evasion, and administrative inefficiencies. Some international comparison is useful here. Tax gaps exist in all countries, including advanced economies. Yet the magnitudes are generally much smaller. The tax gap in the United States, for example, has been estimated at around 3 per cent of GDP(Sarin, 2026). Bangladesh’s estimated tax gap reaches 5.89 per cent of GDP, only accounting for the PIT and domestic VAT gap. So, the unrealised PIT and VAT revenues resulting from non-compliance and weak enforcement, it can be said that Bangladesh’s tax-toGDP ratio could have reached at least 13.47 per cent in FY2022 under full compliance within the existing statutory framework, as demonstrated in Appendix A1. It should be noted that this does not imply that the two estimates of tax gaps are strictly comparable in a methodological sense, but it does underline the depth of Bangladesh’s compliance and enforcement problem. The coexistence of a very low tax-toGDP ratio and a very large legal tax gap indicates that the country suffers from both a narrow effective base and substantial leakage within that base. An alternative simulation further accentuates the magnitude of unrealised VAT revenue. In this exercise, the reduced VAT rates are not taken into account; instead, the standard 15 per cent VAT rate is uniformly applied to all non-exempt VAT-applicable goods and services. After adjusting for the discrepancy between total household expenditure estimated from HIES 2022 and private consumtpion at current market prices reported by the Bangladesh Bureau of Statistics(BBS), the potential domestic VAT revenue, without the reduced rates, is estimated at 5.16 per cent of GDP. Under this assumption, the overall tax-to-GDP ratio could have reached approximately 14.81 per cent in FY2022 under full compliance. This exercise illustrates that a significant portion of the Bangladesh government’s revenue shortfall can be attributable to reduced VAT rates. Table 4.3: Active taxpayers on the PIT register as a percentage of the labour force and the tax-toGDP ratio of various countries of Asia and the Pacific Country Active taxpayers on the PIT register as a percentage of the Labour Force Tax-to-GDP Ratio Bangladesh 3.57 7.3 Pakistan 7.04 10.5 Mongolia 10.41 24.7 Indonesia 12.87 12 India 13.43 11.7 Maldives 15.90 23.5 Thailand 26.74 17.6 Bhutan 29.98 11.6 Japan 33.16 19.8 Nepal 41.86 17.5 Malaysia 53.76 10.6 China 79.42 20.2 Source: International Survey on Revenue Administration(ISORA), IMF; OECD(2025); Revenue Statistics in Asia and the Pacific 2025, https://oe.cd/revstatsap2025. 24

Note: The data for the Tax-to-GDP ratio for India is for 2022 and derived from PwC. Nepal’s tax-to-GDP ratio in this table is for 2022 and derived from World Bank data. The data for the ratio for all other countries is for 2023. Table 4.4: Income tax compliance rates Number of e-TIN accounts until 2022 Number of total taxpayers in 2022 Compliance rate Individual and others Company 7,519,261 157,846 3,237,848 37,268 43.06% 23.61% Source: Authors’ representation using data from the Annual Report 2021–22 of the National Board of Revenue(NBR). Note:“Individual and others” include sole proprietor, firms, etc. In the study, it is referred to as personal income tax. The compliance rate is measured as the percentage of the total number of taxpayers in the total number of e-TIN holders. Anyone eligible to pay tax according to the Income Tax Act, 2023, is required to obtain an e-TIN. Compliance indicators reinforce this diagnosis. IMF International Survey on Revenue Administration (ISORA) data suggest that only 3.57 per cent of the labour force in Bangladesh appears as active taxpayers on the PIT register(Table 4.3). NBR data show a higher filing rate among e-TIN holders, with a compliance rate of 43.06 per cent for individuals and others, but this metric is narrower because it excludes the large share of economically active individuals who remain outside the registration system altogether(Table 4.4). Bangladesh, therefore, appears to face a double weakness: a low proportion of potentially taxable individuals enter the system, and even among those who do, compliance is incomplete. Despite the overall low level of compliance, income tax collection in Bangladesh depends heavily on taxes withheld at source, with 62.8 per cent of total income tax revenue collected through this mechanism (National Board of Revenue, 2023). Withholding arrangements, often implemented through Advance Income Tax(AIT), undoubtedly facilitate revenue collection by targeting formal-sector income streams and recorded transactions that are relatively easy for the tax authority to observe. Yet this also reveals a deeper structural weakness in the tax system, namely its disproportionate dependence on income that is visible, documented, and administratively convenient to tax. In effect, employees and other formalsector participants, whose earnings are automatically deducted or otherwise traceable, shoulder a much greater share of the effective tax burden than many other income earners. By contrast, income derived from self-employment, small and medium business activity, professional services, and other less transparent sources is far more difficult to verify, assess, and enforce, leaving substantial room for evasion, underreporting, and tax minimisation. These income streams are often loosely associated with the broader informal economy, but treating them simply as part of an exogenously given informal sector risks missing an important point. The failure to bring such activities within the effective tax net does not merely reflect informality, it also helps reproduce and sustain it. When significant segments of economic activity can continue operating outside meaningful tax scrutiny, the incentives for formalisation remain weak, the culture of non-reporting becomes entrenched, and the tax base remains persistently narrow. In this sense, weak enforcement and informality are mutually reinforcing: limited tax coverage allows informality to persist, while the continued prevalence of informal and weakly documented activity further constrains the expansion of the tax base. The result is a system that captures revenue primarily from accessible tax handles rather than from the full distribution of taxable capacity, thereby weakening both the horizontal equity and the intended progressivity of the personal income tax regime. 25

The VAT story is equally revealing. The large VAT gap suggests that the problem lies less in statutory design alone and more in weak remittance and enforcement. Consumers pay VAT through prices, but firms serve as the collecting agents. 13 Where sellers fail to register, underreport sales, or do not remit what they collect, the state loses revenue even though households still bear the burden of the tax in the marketplace (Ebrill et al., 2001; Faridy et al., 2014). Thus, the final incidence falls on consumers, disproportionately on lower and middle-income households, while leakage occurs upstream among businesses. Corporate taxation exhibits a similar imbalance in the tax structure composition of Bangladesh. Only 19.32 per cent of total tax revenue comes from corporate income tax, despite Bangladesh having the highest average CIT rate of 32.5 per cent among the South Asian economies, followed by Pakistan at 29 per cent and Bhutan, India, and Nepal at 25 per cent(Moazzem et al., 2023). From the NBR data, we have found that the corporate compliance rate is only 23.61 per cent(Table 4.4). It has been suggested that Bangladesh’s tax-to-GDP ratio would be about 2 percentage points higher without various tax exemptions(Razzaque et al., 2023). According to the recent IMF staff estimates based on the data from the National Board of Revenue, rationalising tax exemptions by repealing or phasing out 12 VAT-related SROs, 4 income tax-related SROs, and exemptions under the Income Tax Act 2023 is projected to yield over 0.1 per cent of GDP in FY26. These exemptions are often random, not based on any cost-benefit analysis to evaluate the effectiveness of the proposed tax incentives, and influenced by elite stakeholders (Razzaque et al., 2023). It has been pointed out that influential and politically connected businesspeople are often able to evade taxation by failing to file returns or by submitting low declarations that avoid audit scrutiny(Ahmed, 2024). Businesses incur substantial costs to comply with the VAT laws, which require meticulous recordkeeping, complex calculations, regular filing, and frequent direct engagement with tax officials. It has been argued that these compliance costs for small and medium-sized enterprises amount to roughly 2–3 per cent of annual turnover, and are inherently regressive insofar as smaller firms lack the administrative economies of scale that allow larger enterprises to absorb such costs more efficiently(Faridy et al., 2017). In Bangladesh, these pressures are further compounded by a VAT compliance architecture that has remained extremely complex. The Eighth Five Year Plan(8FYP) explicitly acknowledges that the failure to implement the VAT Law of 2012 during the period of 7FYP, a fundamental reform designed to modernise the VAT system and reduce the VAT compliance costs, represents a significant institutional shortcoming. So, 8FYP lists full adoption of the VAT and Supplementary Duty Act 2012 as an urgent and overdue priority(Bangladesh Planning Commission, 2020). The persistence of this outdated legal and procedural framework means that small firms continue to operate under a compliance environment that is both burdensome and costly, reinforcing their incentive to remain in the informal sector or underreport their activities rather than engage fully with the formal tax system. 4.3. Factors influencing low tax compliance The weak compliance outcomes documented above are not merely technical failures of tax administration. They are rooted in deeper structural, institutional, and behavioural features of Bangladesh’s economy and governance system. One of the most immediate constraints is the large informal economy. A substantial number of individuals and firms earn taxable income while operating outside registration, filing, and audit systems. This directly narrows the observable tax base. Although cross-country evidence 13 It is assumed that VAT is fully shifted to the consumers via prices. 26

does not always show a simple unconditional relationship between informality and tax-to-GDP ratios, the informal economy has nevertheless been identified as a major source of domestic revenue loss in Bangladesh, with one estimate placing the loss at around BDT 842 billion, which is equivalent to 0.25 per cent of the total tax revenue collected in FY2023(Moazzem et al., 2023). However, informality alone may not explain the full pattern of non-compliance. The broader literature on tax behaviour has long moved beyond purely mechanical accounts based on tax rates and enforcement. Drawing on Ali et al., 2023; Bello& Danjuma, 2014; Fjeldstad et al., 2012; OECD, 2012, five broad frameworks are especially relevant for interpreting Bangladesh’s compliance challenge: the economic deterrence model, social-psychological models, fiscal exchange, comparative treatment, and political legitimacy or trust in government. These are summarised in Table 4.5 and together provide a useful framework for understanding why a low-compliance equilibrium can persist. Table 4.5: Summary of models explaining factors influencing people’s decision to pay taxes 14 15 Model The (Economic) Deterrent Model 14 Socialpsychology Models Fiscal Exchange Comparative Treatment Explanation Studies It explains tax evasion by assuming that taxpayers are rational and self-interested, and therefore choose whether to comply by comparing expected costs and benefits. In this school of thought, evasion is like gambling: taxpayers weigh the certain loss of paying taxes against the potential gain from evading and not being detected, and the potential large loss if detected and punished. 15 Allingham& Sandmo, 1972; Becker, 1968; Bernasconi, 1998; Webley et al., 1991 The theory of reasoned action implies that people make decisions based on personal factors and social influences. While explaining taxpayers’ behaviour, these factors are usually referred to as personal norms and social norms, and both are seen as important drivers of tax compliance. In taxation, personal norms include tax ethics, tax morale, and attitudes toward paying taxes. People may avoid evading because it causes guilt, anxiety, or a negative self-image. In fact, fear of guilt and social shame can deter evasion more strongly than the risk of audits or penalties. Social norms reflect the influence of society on behaviour. As people often discuss taxes with friends, family, and colleagues, their compliance decision may depend on what they believe others are doing. Evans& McKerchar, 2009; Franzoni, 1998; Lewis, 1982; OECD, 2012; Sour, 2001; Taylor, 2004 This theory builds on economic deterrence and social-psychological ideas, and views taxation as part of an informal contract between citizens and the state. It argues that people are more willing to pay taxes when they feel that they receive valuable public goods and services in return. In this sense, taxpayers“trade” some of their purchasing power for government services, and their compliance rises when they are satisfied with government performance, even without strong enforcement. Alm, 1999; Alm et al., 1992; Evans& McKerchar, 2009; Feld & Frey, 2007; Fjeldstad et al., 2012; Torgler, 2003 People are more likely to follow tax laws when they feel the tax system and its decision-making process are fair. This idea is central to this model; taxpayers compare their own treatment and tax burden with those of others with a similar economic background. When people perceive their tax burden to be broadly comparable to that of peers, they are more willing to comply. Perceptions of fairness also extend beyond one’s own judgments about how tax burdens are distributed across society and whether others are complying. If taxpayers believe evasion is widespread, non-compliance can become socially normalised, weakening the intrinsic motivation to pay taxes. Ali et al., 2023; Evans& McKerchar, 2009; Feld& Frey, 2007; GIZ Sector Programme Public Finance, Administrative Reform, 2010; OECD, 2012; Sour, 2001; Torgler& Schneider, 2005; Tyler, 2006; Walsh, 2012 14 The classic Allingham-Sandmo model predicts that evasion is mainly driven by the tax rate(which raises the reward from evasion), the probability of detection through audits, and the severity of penalties(Allingham& Sandmo, 1972; Cowell, 1985). As a result, higher audit probability or stricter penalties should, in theory, increase compliance, and tax authorities can curb evasion by raising either detection risk or penalty rates(Allingham& Sandmo, 1972). 15 The literature also highlights major limitations of the(Economic) Deterrent models. They are often criticized for being overly simplistic as they reduce people’s decision of compliance to a narrow cost-benefit calculation and ignore other important influences such as morality, social norms, trust in government, and the costs and complexity of complying; they also focus mainly on direct taxes and explain changes in compliance more than levels of compliance(Allingham& Sandmo, 1972; Tonry, 2011). 27

Model Explanation Studies Political Legitimacy and Trust in Government The more citizens trust their government and social institutions and believe them to be appropriate, just, and work for the common good, the more it is likely for the citizens to voluntarily abide by the rules set by the institutions. Many crosscountry studies support the fact that tax compliance is higher where corruption is controlled and bureaucracy is smaller, suggesting that commendable credibility and reputation of the government encourage taxpayers to cooperate. Ali et al., 2023; Bello& Danjuma, 2014; Fjeldstad et al., 2012; OECD, 2012; Picur & Riahi-Belkaoui, 2006; Tyler, 2006 Source: Authors’ compilation based on secondary literature. Note: The words“fair” and“fairness” appear multiple times in these models. So, to assess the factors influencing compliance, the OECD’s definition of fairness is used. The OECD identifies three dimensions of tax fairness. Distributive fairness concerns whether taxpayers view the government as a responsible and efficient user of tax revenues. Procedural fairness refers to whether tax authorities follow established rules and treat taxpayers impartially and respectfully. Retributive fairness indicates the perceived appropriateness and consistency of penalties when tax rules are violated. Distributive fairness aligns with the fiscal exchange perspective, that is, how efficiently the government is using the tax revenue to deliver public services. Procedural and retributive fairness depend on the quality of tax administration through transparent and equitable handling of tax matters. The traditional deterrent model, rooted in Allingham and Sandmo(1972), assumes that taxpayers are rational and self-interested, and therefore evade tax when the expected benefit of evasion outweighs the expected cost of being caught and penalised. Under the deterrence model, evasion becomes attractive when the expected probability of detection is low relative to the gains from non-compliance. Bangladesh appears to fit this pattern closely. According to the NBR Annual Report 2021–22, only 497 VAT-registered companies were audited out of 353,762 registered entities in that year. Even allowing for some reporting or definitional issues in the underlying data, the implied audit intensity is extremely low. Where detection is unlikely, statutory penalties, even if severe on paper, lose much of their practical force(Faridy et al., 2014). For many taxpayers, evasion then becomes a rational gamble. But deterrence alone does not explain the observed behaviour. Experimental and empirical studies have repeatedly shown that actual compliance tends to exceed what narrow neoclassical models would predict, even where enforcement is weak(Sikayu et al., 2022). This is where social norms and tax morale become important. When people believe that evasion is widespread, especially among those with economic or political influence, the moral cost of non-compliance declines. In Bangladesh, weak enforcement visibility, frequent informal transactions, and a widespread perception that the better connected can escape scrutiny all contribute to a social environment in which evasion becomes easier to justify(Bello & Danjuma, 2014; Jimenez& Iyer, 2016). Over time, this can normalise non-compliance. Since the 1990s, tax morale and social norms have received increasing attention in compliance research. A large part of taxpayers’ behaviour can be explained by personal norms such as ethics, attitudes, guilt, self-image, and social norms regarding what they believe others in society are doing(Bello and Danjuma, 2014). Tax compliance decisions are not determined solely by a taxpayer’s interaction with the revenue authorities and the tax structure; they are also reflected by the attitudes, perceptions, and moral judgements of individuals formed through their associations with family, friends, and reference groups(Lewis, 1982). In other words, the societal norms regarding paying taxes influence compliance by shaping the individuals’ personal tax morale(Jimenez& Iyer, 2016). In a country like Bangladesh, where corruption is prevalent and political or wealthy influence is very common, tax evasion or avoidance by the elites is common knowledge socially. When taxpayers believe that evasion is widespread, among businesses, wealthy professionals, or politically connected groups, non-compliance becomes easier to 28

justify socially. The weak visibility of credible enforcement, combined with daily exposure to informal transactions and underreporting of income and wealth, has gradually normalised evasion and weakened the internal moral cost of not paying taxes in Bangladesh. Even when individuals know that evasion is legally wrong, the social state of widespread non-compliance reduces the psychological pressure to comply. The fiscal exchange perspective points to another weakness. People are generally more willing to pay taxes when they believe that tax revenue is being converted into visible and useful public goods. Equity Theory similarly argues that people are more likely to obey rules if they feel that they are treated fairly in terms of benefits relative to contributions(Bobek, 1997). Where public spending is perceived as ineffective, captured, or poorly targeted, willingness to comply tends to weaken. Empirical work has found that better human development outcomes and more effective public service delivery are associated with lower evasion and stronger tax morale(Altaf et al., 2019). Bangladesh has remained in the“Medium Human Development” category since 2009 and ranked 130th among 196 countries in HDI in 2023, which suggests that progress in health, education, and living standards has been moderate. Similarly, in 2023, Bangladesh’s Government Effectiveness percentile rank in the World Governance Indicator(WGI) was only 26.89, which means that the country performs better than only about 27 per cent of countries and lags behind nearly three-quarters of the world in terms of the quality of public services and the effectiveness of policy implementation. In such a context, the informal fiscal contract between citizens and the state remains fragile, and this could be one contributing factor to the low tax compliance rate. Comparative treatment, or fairness in relative terms, is equally important. Taxpayers compare themselves not only with the state but also with one another. If ordinary citizens believe that others, particularly politically connected or economically powerful groups, can avoid taxes through exemptions, lobbying, or informal arrangements, their willingness to comply deteriorates. Bangladesh’s low score on control of corruption and the prevalence of discretionary interventions, including special tax treatments through SROs, reinforce these perceptions. Where the system appears selective in whom it burdens and whom it relieves, non-compliance may come to be seen not as deviant behaviour but as a rational response to unfair treatment. The above channels ultimately converge on a broader issue of institutional credibility. The deterrence model depends on enforceable rules. Social norms depend on the visibility and legitimacy of institutions. Fiscal exchange depends on state effectiveness. Comparative treatment depends on fair administration. In that sense, poor institutional quality sits at the centre of the compliance problem. Earlier studies have noted that manual assessment procedures and weak administrative systems create room for informal negotiation, selective enforcement, and underdeclaration(Ahmed, 2024). More generally, cross-country evidence shows that stronger institutions and better governance are associated with higher tax morale and better compliance outcomes(Baum et al., 2017; Sebele-Mpofu, 2020; Torgler, 2011). The Financial Secrecy Index(FSI), published by the Tax Justice Network, adds a related dimension to this discussion. The FSI measures the extent to which a country’s legal and institutional framework enables financial secrecy, which in turn can facilitate money laundering and tax abuse. Bangladesh’s relatively high secrecy score in the 2022 index suggests a greater perceived scope for shielding finances from regulatory scrutiny. While this should not be interpreted too mechanically as a direct measure of domestic compliance behaviour, it is broadly consistent with the concern that weak transparency and 29

limited information exchange in Bangladesh facilitate the concealment of taxable assets and income. In such an environment, where influential actors are perceived to benefit from opacity in the legal and institutional framework governing financial disclosure, the perceived fairness of tax administration deteriorates further, weakening voluntary compliance more broadly(Onu et al., 2019). Cross-country analyses provide evidence on how governance shapes compliance. Studies show that the quality of political institutions, determined by indicators such as voice and accountability, rule of law, political stability, regulatory quality, and control of corruption, has a strong effect on tax morale(Torgler, 2003). It is also found that a one standard deviation increase in corruption perception is associated with about a 12 percentage-point lower tax revenue-to-GDP ratio(Baum et al., 2017). Bangladesh’s governance challenges are well-documented in global indices. Transparency International ranks Bangladesh among the bottom countries in corruption control, while World Bank governance indicators show Bangladesh’s position among the low percentiles across voice and accountability, political stability, government effectiveness, regulatory quality, rule of law, and control of corruption. It can therefore be inferred that such institutional weaknesses create a systemic environment in which tax evasion becomes easier, more rational from an individual perspective, and increasingly socially tolerated. In the literature, we find that improving governance and institutional quality raises tax morale and compliance(SebeleMpofu, 2020; Torgler, 2011), suggesting that Bangladesh’s compliance challenges cannot be alleviated without institutional reforms. 4.4. Potential of wealth tax for government revenue and social justice The preceding analysis has shown that Bangladesh’s existing tax structure relies heavily on indirect taxation while failing to capture a large part of high-income and high-wealth capacity. This naturally raises the question of whether wealth-based taxation could help improve both equity and revenue mobilisation. In principle, wealth taxes are among the most progressive fiscal instruments because net wealth is typically far more concentrated than income. Indeed, wealth taxation is often considered more progressive than property taxation alone, since real estate is only one component of wealth and is usually more evenly distributed than net wealth in its entirety(Saez& Zucman, 2016). The estimates of wealth using HIES(2022) data show that the wealthiest 10 per cent of households own more than 76 per cent of total private wealth, while the wealthiest 1 per cent alone holds more than 47 per cent. The distribution curve for wealth is demonstrated in Figure 4.7. The curve becomes extremely steep for households in the wealthiest deciles as their equivalised wealth increases, indicating extreme wealth concentration at the top. A comparison of wealth and income concentration clearly shows that wealth inequality is more severe than income inequality. 30

Figure 4.7: Cumulative share of total wealth held by the cumulative share of household Source: Authors’ analysis using the Household Income and Expenditure Surveys, 2022. Note: The wealth distribution curve shows the cumulative distribution of national wealth(sum of weighted household wealth estimated from HIES 2022) across cumulative households ranked from poorest to richest based on their equivalized income. The line of equality represents a wealth distribution in which all the households possess the same level of equivalized wealth, that is, the ability to have the same standard of living. Among the urban population, wealth is even more concentrated than the national average: the top 1 per cent of urban households possess more than 53 per cent of total private wealth held in urban areas. The regional wealth distribution in Bangladesh, along with the national wealth distribution, is illustrated in Figure 4.8. Wealth is measured based on respondents’ self-reported valuations of their assets at current prices, and recall errors or misreporting may therefore affect these estimates. In addition, because the richest households are underrepresented in HIES 2022, the true extent of wealth concentration is likely to be even more severe than what is suggested by the estimates presented here. Given such extreme concentration, even a modest wealth tax rate applied to a small fraction of households could generate significant revenue while directly targeting those who benefit most from wealth accumulation. 31

Figure 4.8: Percentage of wealth possessed by the percentiles of households in total wealth Source: Authors’ analysis using the Household Income and Expenditure Surveys, 2022. Note: The blue bars indicate the percentage of total private wealth possessed by each percentile of nationally representative households. The orange bars indicate the percentage of total urban private wealth possessed by each percentile of urban representative households. The grey bars indicate the percentage of total rural private wealth possessed by each percentile of rurally representative households. It is important to note that the World Inequality Report(2026), prepared by World Inequality Lab(WIL), reports that the richest 1 per cent of the population in Bangladesh holds 23.9 per cent of total wealth (Chancel et al., 2026). WIL’s“preferred approach” to estimating wealth distribution combines the income capitalisation method, starting from income tax data to infer asset holdings from taxable income flows, with imputations derived from household-level surveys. However, in the context of Bangladesh, relying on administrative income tax data to generate nationally representative wealth estimates raises concerns. Income tax records cover only individuals registered with Electronic Taxpayer Identification Numbers (e-TINs) who file tax returns, which represents only 4.5 per cent of the employed population. 16 Since administrative tax data excludes non-filers, the income tax microdata only covers the people belonging to the top income groups. As a result, estimates derived from income tax microdata do not adequately capture the nationally representative wealth distribution and instead represent only this small fragment of the population. In practical terms, the reported figure may be more appropriately interpreted as indicating that the richest 1 per cent within the tax-filing population accounts for 23.9 per cent of the wealth recorded among registered taxpayers, rather than representing the entire population of Bangladesh. 17 The problem is compounded by underreporting within that filing population itself. Even among return filers, 16 To estimate the share of Bangladesh’s labour force who file income tax returns, the total number of personal income tax filers reported in the National Board of Revenue’s Annual Report 2021–22, which is 3.23 million, is divided by the total employed population of 70.47 million as reported in the Labour Force Survey 2022 published by the Bangladesh Bureau of Statistics. 17 The fundamental problem is that the tax-return population is not representative of the population as a whole. Most people do not file tax returns, and those who do are disproportionately drawn from relatively better-off, formal, urban, and tax-visible groups. As a result, the top 1 per cent identified from tax records is not the top 1 per cent of all households or individuals in Bangladesh, but rather the top 1 per cent within a narrow and highly selective filing universe. 32

declared wealth often reflects only the visible and reported portion of total asset holdings. In Bangladesh, substantial wealth may be held in under-valued land and real estate, family businesses, jewellery, cashbased commercial activity, overseas assets, or benami arrangements that do not appear fully in tax records. Where wealth is inferred from taxable income flows, the scope for understatement becomes even greater, since many assets generate little reported taxable income or yield returns that are concealed or lightly declared. For these reasons, estimates based primarily on administrative tax data are likely to understate, potentially by a considerable margin, the true share of national wealth held by the richest 1 per cent. The World Inequality Lab may have used an alternative approach in countries where tax microdata is unavailable, combining household survey data with summary tax statistics(Distributional National Accounts Guidelines, 2024). However, the specific methodology used to generate the wealth distribution estimates for Bangladesh in the World Inequality Report(2026) is not clearly documented. An estimate based on administrative tax return data for FY2022–23 reported in the literature indicates that the wealthiest 1 per cent of tax filers in Bangladesh hold about 21 per cent of all declared wealth (Chowdhury, 2026). This estimate is quite close to the 23.9 per cent figure reported in the World Inequality Report. So, it can be inferred that WIL may have relied primarily on administrative income tax data for Bangladesh. Table 4.6: Simulation of tax revenue by applying a wealth tax Wealth tax rate Revenue from top 10%(crore BDT) Revenue as % of GDP Revenue as% of total tax Revenue from top 20%(crore BDT) Revenue as % of GDP Revenue as% of total tax 1% 178878.7 2% 357757.4 5% 894393.4 4.50 9.01 22.52 60.67 121.35 303.36 201797.4 403594.8 1008987 5.08 10.16 25.40 68.45 136.89 342.23 Source: Authors’ analysis using the Household Income and Expenditure Surveys, 2022. Note: The tax rates are hypothetically set to observe the potential revenue that could be collected if the respective tax rates were imposed on wealth and all the citizens complied fully with the tax laws. To derive the percentages of total tax and total GDP, the values of total tax revenue collected in FY22 by NBR were derived from the Annual Report 2021–22 of the National Board of Revenue(NBR) and the GDP at current prices of that year was collected from a report of the Bangladesh Bureau of Statistics. In FY22, the total tax collected by NBR was BDT 301,050.34 crore, and the total GDP at current prices was BDT 3971716.4 crore. The simulation exercise in Table 4.6 shows that an annual wealth tax applied only to the wealthiest 10 per cent households yields sizeable revenue, and expanding the base to the wealthiest 20 per cent increases the yield further. Even the most conservative option, a 1 per cent tax on only the top decile, would raise revenue equal to a staggering 4.5 per cent of GDP. A tax on wealth, therefore, directly supports the broader equity goal in the tax system by shifting back part of the tax burden toward households with the greatest capacity to pay. This raises an important normative question regarding the legitimacy of wealth taxation, particularly in contexts where wealth is accumulated from income that has already been subject to taxation. One line of argument holds that taxing wealth amounts to a form of double taxation. Under this view, individuals first pay taxes on income when it is earned and then face an additional levy on the stock of savings or assets created from that income. This is often seen as penalising thrift and long-term saving, especially in systems where capital income, inheritance, and property are already taxed in various forms. Critics also 33

argue that repeated taxation of accumulated assets may discourage investment, induce capital flight, and reduce incentives for formalisation, particularly in economies with weak administrative capacity. An alternative perspective challenges the characterisation of wealth taxation as double taxation. It emphasises that wealth is not simply a stored form of previously taxed income but a distinct indicator of economic capacity that reflects accumulated advantages over time, including capital gains, inheritances, market power, and access to preferential opportunities. In practice, many of these components are either lightly taxed or not taxed at all under existing systems. From this standpoint, wealth taxation is justified on equity grounds, as it targets the stock of economic power rather than the flow of income alone. It can therefore help correct disparities that arise from unequal asset accumulation, especially in settings where income taxation is narrow and enforcement is weak. A more pragmatic view most likely lies between these positions. Even if part of wealth originates from previously taxed income, the degree to which that income was effectively taxed is often limited in practice, particularly in countries with substantial evasion and exemptions. Moreover, the distributional implications of not taxing wealth can be significant when wealth concentration is high. The policy question, therefore, is not simply whether wealth taxation is theoretically justified, but whether it can be designed in a manner that balances equity objectives with administrative feasibility and economic incentives. Avoidance or evasion responses to wealth taxation can be large. In Switzerland, a 1 per cent increase in annual wealth tax has been associated with a 23 to 34 per cent reduction in reported wealth, reflecting valuation responses, asset rearrangement, and other forms of behavioural adjustment(Brülhart et al., 2016). Offshore concealment and changes in residence can also weaken the yield of wealth taxation, especially where information systems and enforcement are weak(Alstadsæter et al., 2018; Rose, 2017; Saez& Zucman, 2019; Zucman, 2013, 2016). The above considerations are particularly important in the Bangladesh context. A recurrent annual wealth tax requires reliable asset valuation, robust registries, strong third-party information systems, and high administrative credibility. At present, these conditions are only weakly met. The large informal economy, limited documentation of asset ownership, and governance weaknesses together suggest that a comprehensive annual wealth tax would be difficult to administer effectively. In such circumstances, the gap between simulated and realised revenue could be enormous. For that reason, inheritance taxation may offer a more feasible route for incorporating wealth more meaningfully into the tax system. Unlike annual wealth taxes, inheritance taxes are levied at the point of legal transfer, when there is a stronger incentive for heirs to document and establish ownership. This can make valuation and enforcement less demanding than in the case of recurring net wealth taxation (OECD, 2021). Inheritance taxation also has a clearer normative justification in many settings because it targets intergenerational transfers and unearned windfalls rather than ongoing returns to productive effort. If designed with sufficiently high thresholds and strong documentation requirements, it could enhance progressivity without the full administrative burden of an annual wealth tax. In light of the above discussion, the simulation results should not be interpreted as necessarily supportive of the immediate introduction of a broad wealth tax in Bangladesh. Rather, they underscore the scale of untaxed wealth concentration and the need to bring wealth more meaningfully within the fiscal framework. Given current institutional constraints, inheritance taxation, possibly complemented by better taxation of high-value secondary property and stronger asset registries, may represent a more realistic starting point for doing so. 34

5. Policy Implications and Concluding Observations The findings of this study point to a tax system in Bangladesh that is weak in revenue performance as well as deeply imbalanced in distributive terms. The personal income tax system is formally progressive in its statutory design; nevertheless, it rests on an exceptionally narrow base and depends heavily on a small group of visible and relatively compliant earners. The Value Added Tax(VAT) remains central to revenue mobilisation, but its burden falls disproportionately on lower- and middle-income households when assessed relative to income, while significant leakages prevent the state from realising its full yield. Corporate taxation, meanwhile, is constrained by exemptions, special treatments, and weak compliance, and highly concentrated wealth remains only lightly touched by the fiscal system. These features, on the whole, produce a tax structure that raises too little, relies too much on those easiest to tax, and does too little to align the burden of taxation with the distribution of economic capacity. These problems have implications that go well beyond technical questions of tax design. A system in which poorer households bear a relatively heavy burden through consumption taxes, while large portions of high-income, corporate profits, and accumulated wealth remain undertaxed, is difficult to defend on grounds of fairness. It weakens vertical equity by failing to ensure that those with greater ability to pay contribute proportionately more, but it also compromises horizontal equity, since individuals and firms with similar taxable capacity are not treated alike when many eligible taxpayers remain outside the tax net altogether, while others are compelled to comply. The result is a fiscal structure that narrows the redistributive role of taxation while at the same time eroding the perceived legitimacy of the system itself. In a context where trust in public institutions is already fragile, such asymmetries can reinforce the perception that the state extracts from those who are visible, organised, and easy to tax, while allowing others, often with greater economic or political influence, to remain beyond effective reach. That, in turn, has serious consequences for tax morale, voluntary compliance, and the durability of the broader fiscal contract. First and foremost, therefore, the reform agenda must begin with the personal income tax system. The central priority is not higher statutory rates, but a broader and more effectively enforced base. Bangladesh’s current PIT structure places too much weight on salaried and otherwise visible income, while many high-income professionals, self-employed individuals, and asset-owning households remain insufficiently captured. A key policy implication is the need to break the mutually reinforcing relationship between informality and weak tax enforcement by systematically expanding the effective tax net beyond the narrow set of visible and withholding-based income sources. This requires a shift from a predominantly passive and declaration-based system to one anchored in third-party information, risk-based assessment, and gradual formalisation incentives. Strengthening information linkages through third-party data and verifiable records across banks, mobile financial services, business registration systems, land registers, utility providers and other administrative sources, and professional licensing bodies can help identify economic activity that currently escapes scrutiny. At the same time, simplified tax regimes for small and self-employed businesses, combined with credible but proportionate enforcement, can lower the cost of entering the tax system while increasing the expected cost of remaining outside it. Crucially, enforcement should be sequenced and targeted so that expansion of the tax base does not disproportionately burden genuinely low-income informal operators but instead focuses on higher-income individuals and businesses operating under the cover of informality. Over time, such an approach can help shift incentives towards formalisation, broaden the 35

tax base, and reduce the system’s excessive reliance on easily observable income, thereby strengthening both revenue mobilisation and horizontal equity. The withholding net should also be widened to cover professional services, property income, and digital income streams. At the same time, it is important that withholding and Advance Income Tax (AIT) mechanisms operate as instruments of collection efficiency rather than as de facto final taxes disconnected from actual taxable capacity. Any tax collected at source should therefore be fully creditable against final income tax liability, with timely adjustment and settlement through the returnfiling process. Where such adjustment does not take place, withholding can become inequitable. This concern is especially relevant in Bangladesh, where around one-fifth of all personal income tax revenue is currently collected through AIT on import-related sources, yet much of this is not effectively reconciled with final tax liability. In practice, this can convert what is meant to be an advance payment into a nonadjustable presumptive levy, thereby penalising compliant taxpayers, distorting investment and business decisions, and weakening horizontal equity. A fairer and more coherent PIT system would require that all withholding-based collections be integrated into a transparent final assessment framework. Alongside this, a thorough review of exemptions, allowances, and tax credits is necessary to ensure that the PIT system does not reward those with better access to formal financial instruments and documentation, while leaving much of the wider tax base untouched. VAT reform is another critical area, focusing on strengthening compliance and reducing regressivity without undermining its revenue role. The discussion of this paper suggests that Bangladesh’s VAT problem is twofold. Households, especially poorer households, bear a substantial effective burden through consumption, yet the government does not receive the full revenue because of weak remittance, underreporting, and evasion by firms. This means VAT reform should not be limited to debates over rates and exemptions. Greater priority should be given to tightening compliance through mandatory electronic invoicing, stronger audit trails, improved monitoring of filing behaviour, and more transparent treatment of input tax credits. A more coherent VAT base, with fewer exemptions on non-essential items and less discretionary treatment, would also improve efficiency. Yet because VAT falls more heavily on those who consume a larger share of their income, reform in this area cannot be separated from social justice concerns. The state should therefore view VAT policy together with expenditure-side measures, particularly targeted cash transfers or other direct support for low-income households, so that the regressive effects of consumption taxation are not allowed to harden existing inequalities. Another important implication of the paper is that tax reform in Bangladesh should be judged not only by how much revenue it raises, but also by whether it shifts the burden of taxation closer to the principle of ability to pay. This is where the issue of wealth becomes especially important. The evidence presented here suggests that wealth is far more concentrated than income, while the current tax system does little to bring that concentration within the effective fiscal net. Given the administrative and valuation difficulties associated with an annual wealth tax, a more practical route could be to introduce a welldesigned inheritance tax focused on large estates and substantial transfers of wealth. Such an instrument would be more feasible in administrative terms and more defensible in distributive terms, as it would address the intergenerational transmission of economic advantage at the point where ownership must be formally established. It would also have the advantage of taxing unearned windfalls rather than current labour income or ongoing productive effort, thereby strengthening its equity rationale. Unlike a recurrent net wealth tax, which requires repeated valuation of assets and may impose a continuing burden on illiquid holdings, inheritance taxation arises at a specific legal event, making asset disclosure, valuation, 36

and ownership verification more manageable. It can therefore serve not only as a revenue instrument but also as a mechanism for improving documentation of wealth and limiting the persistence of hidden or informally held assets. If designed with sufficiently high thresholds and progressive rates, such a tax could protect ordinary family transfers while targeting the very large estates through which privilege and economic advantage are often reproduced across generations. Complementary measures, such as annual taxation of high-value secondary property and mandatory asset declarations for high-net-worth individuals, would also help. More broadly, bringing wealth into the tax conversation is not simply a revenue issue. It is central to whether the fiscal system is seen as sharing burdens fairly across social groups rather than concentrating adjustment on wage earners and consumers. Corporate tax reform also deserves much greater weight in the policy response. The paper suggests that the weakness of corporate tax performance in Bangladesh does not arise primarily from headline rates, but from the fragmentation of the base through exemptions, tax holidays, preferential treatments, and weak enforcement. A more rules-based and transparent corporate tax regime is needed. Annual tax expenditure statements should be published to make the fiscal cost of exemptions transperant. Incentives, where retained, should be tied to measurable outcomes such as employment creation, export performance, or technological upgrading rather than granted through broad and open-ended concessions. The capacity to monitor transfer pricing and other forms of base erosion should also be strengthened. These reforms are important not only for raising revenue but also for making the system directly linked to perceptions of fairness. When ordinary taxpayers believe that profitable firms and politically connected interests are treated more favourably than others, the legitimacy of the tax system is weakened for everyone. Equally important, this study points to the need to place distribution and social justice much more explicitly at the centre of tax policy. In Bangladesh, tax reform is too often discussed as though it were only a matter of collection efficiency. That is too narrow a view. The incidence results in this study show that the structure of taxation itself can amplify inequality if the state relies excessively on indirect taxes while failing to tax income, profit, and wealth more effectively. A more equitable fiscal framework would mean not only broadening direct taxes but also using the additional fiscal space to support social spending that is visible, targeted, and credibility-enhancing. The link between revenue mobilisation and public service delivery matters greatly. If citizens see that stronger direct taxation finances better education, health, social protection, and local public goods, the case for compliance becomes stronger both economically and morally. Underpinning all these reforms is the need to transform tax administration. Policy changes alone will not deliver results if the tax authorities continue to operate through fragmented systems, limited data integration, and administrative practices that create scope for discretion, opacity, and informal settlement. A modern, data-driven tax administration is therefore indispensable. This requires full implementation of integrated tax administration systems, stronger digital filing and payment architecture, and far more extensive data linkages across public agencies. It also requires institutional reform within the NBR, including professionalised human resource practices, stronger internal accountability, and credible anticorruption measures. Without such changes, even well-designed policy reforms will struggle to produce meaningful gains. Finally, there is a wider concluding observation that emerges from this study. Bangladesh’s tax challenge is not simply that the country collects too little revenue. It is the existing structure of taxation that risks 37

locking the country into a low-revenue, low-trust, and weakly redistributive equilibrium. Breaking out of that equilibrium requires more than better enforcement. It requires a rebalancing of the tax system so that those with greater economic capacity contribute more, those with less capacity are not asked to carry a disproportionate burden, and taxpayers can see a clearer connection between what they pay and what the state delivers. If such a shift can be achieved, tax reform can do more than finance development. It can help build a fairer and more credible social contract for Bangladesh’s next stage of economic transformation. 38

References Ahmed, F., Genoni, M. E., Roy, D.,& Latif, A.(2019). Official Methodology Used for Poverty Based on the Bangladesh Household Income and Expenditure Survey 2016/17: A Technical Note. The Bangladesh Development Studies, 42(2/3), 289–319. Ahmed, S.(2015). Growth with equity: Contemporary development challenges of Bangladesh. Bangladesh Institute of Bank Management. Ahmed, S.(2024). Institutional Dimensions of Tax Reforms in Bangladesh. In S. Raihan, F. Bourguignon, & U. Salam(Eds.), Is the Bangladesh Paradox Sustainable?(1st ed., pp. 185–217). Cambridge University Press. https://doi.org/10.1017/9781009284677.012 Ali, M., Fjeldstad, H.,& Sjursen, I. H.(2023). Factors affecting tax compliant attitude in Africa: Evidence from Kenya, Tanzania, Uganda and South Africa. Centre for the Study of African Economies (CSAE) Conference. Allingham, M. G.,& Sandmo, A.(1972). Income tax evasion: A theoretical analysis. Journal of Public Economics, Volume 1(1972), Pp. 323–338. Alm, J.(1999). Tax Compliance and Administration. In Handbook on Taxation. Routledge. Alm, J.(2024). Do we have the tools for achieving distributive tax justice? South African Journal of Accounting Research, 38(3), 211–228. https://doi.org/10.1080/10291954.2024.2306717 Alm, J., McClelland, G. H.,& Schulze, W. D.(1992). Why do people pay taxes? Journal of Public Economics, 48(1), 21–38. https://doi.org/10.1016/0047-2727(92)90040-M Alstadsæter, A., Johannesen, N.,& Zucman, G.(2018). Who owns the wealth in tax havens? Macro evidence and implications for global inequality. Journal of Public Economics, In Honor of Sir Tony Atkinson(1944-2017), 162, 89–100. https://doi.org/10.1016/j.jpubeco.2018.01.008 Altaf, F., Herani, G.,& Awan, N. W.(2019). An empirical analysis of determinants of tax evasion: Evidence from South Asia. New Horizons(1992-4399), 13(2), 93. Bangladesh Bureau of Statistics.(2023). Labour Force Survey 2022 Bangladesh. Bangladesh Bureau of Statistics(BBS). https://bbs.gov.bd/pages/static-pages/6922ddd1933eb65569e166e7 Bangladesh Planning Commission.(2020). 8th Five Year Plan, July 2020—June 2025: Promoting Prosperity and Fostering Inclusiveness. General Economics Division(GED), Bangladesh Planning Commission. Baum, M. A., Gupta, M. S., Kimani, E.,& Tapsoba, M. S. J.(2017). Corruption, Taxes and Compliance. International Monetary Fund. Becker, G. S.(1968). Crime and Punishment: An Economic Approach. Journal of Political Economy, 76(2), 169–217. https://doi.org/10.1086/259394 Bello, K. B.,& Danjuma, I.(2014). Review of Models/Theories Explaining Tax Compliance Behavior. Sains Humanika, 2(3). https://doi.org/10.11113/sh.v2n3.432 Bernasconi, M.(1998). Tax evasion and orders of risk aversion. Journal of Public Economics, 67(1), 123– 134. https://doi.org/10.1016/S0047-2727(97)00051-0 Bird, R. M.(2013). Taxation and Development: What Have We Learned from Fifty Years of Research?1 - Richard M. Bird, 2013. Public Finance and Management. https://journals.sagepub.com/doi/ abs/10.1177/152397211301300401 Bird, R. M., Martinez-Vazquez, J.,& Torgler, B.(2005). Societal Institutions and Tax Effort in Developing Countries. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.662081 39

Blasco, J., Guillaud, E.,& Zemmour, M.(2023). The inequality impact of consumption taxes: An international comparison. Journal of Public Economics, 222, 104897. https://doi.org/10.1016/j. jpubeco.2023.104897 Bobek, D. D.(1997). Tax fairness: How do individuals judge fairness and what effect does it have on their behavior? University of Florida. Bouzouita, N.(2025). Domestic Revenue Mobilization in Guinea- Bissau[Selected Issues Paper]. International Monetary Fund. Brülhart, M., Gruber, J., Krapf, M.,& Schmidheiny, K.(2016). Taxing Wealth: Evidence from Switzerland (Working Paper No. 22376). National Bureau of Economic Research. https://doi.org/10.3386/ w22376 Carletto, G., Tiberti, M.,& Zezza, A.(2022). Measure for Measure: Comparing Survey Based Estimates of Income and Consumption for Rural Households. The World Bank Research Observer, 37(1), 1–38. https://doi.org/10.1093/wbro/lkab009 Chancel, L., Gómez-Carrera, R., Moshrif, R.,& Piketty.(2026). World Inequality Report. https://wir2026. wid.world/ Choudhary, R., Ruch, F. U.,& Skrok, E.(2024). Taxing for Growth: Revisiting the 15 Percent Threshold. Washington, DC: World Bank. https://doi.org/10.1596/1813-9450-10943 Chowdhury, K. I.(2026). Effective Income Taxation and Fiscal Capacity in Developing Economies: Evidence from Bangladesh. Research and Policy Integration for Development(RAPID). https://www. rapidbd.org/ Cowell, F. A.(1985). THE ECONOMIC ANALYSIS OF TAX EVASION. Bulletin of Economic Research, 37(3), 163. https://doi.org/10.1111/j.1467-8586.1985.tb00192.x Distributional National Accounts Guidelines.(2024). World Inequality Lab. Ebrill, L., Keen, M., Bodin, J.-P.,& Summers, V.(2001). The Modern VAT. International Monetary Fund. Etzioni, A.(1986). Tax evasion and perceptions of tax fairness: A research note. Journal of Applied Behavioral Science, 22(2), 177–185. https://doi.org/10.1177/002188638602200209 Evans, C.,& McKerchar, M.(2009). Sustaining growth in developing economies through improved taxpayer compliance: Challenges for policy makers and revenue authorities. 21st Australasian Tax Teachers Annual Conference. Faridy, N., Copp, R., Freudenberg, B.,& Sarker, T.(2014). Complexity, Compliance Costs and NonCompliance with VAT by Small and Medium Enterprises(SMEs) in Bangladesh: Is there a Relationship? Australian Tax Forum, 29, 281–328. https://doi.org/10.2139/ssrn.2479146 Faridy, N., Freudenberg, B.,& Sarker, T.(2017). The Devil is in the Detail: An Analysis of VAT Compliance Costs for SMEs in a Developing Nation. New Zealand Journal of Taxation Law and Policy, 23, 176–203. Faridy, N.,& Sarker, T.(2011). Incidence of Value Added Tax(Vat) in the Developing Countries: A Case in Bangladesh. International Journal of Trade, Economics and Finance, 437–442. https://doi. org/10.7763/IJTEF.2011.V2.145 Feld, L. P.,& Frey, B. S.(2007). Tax Compliance as the Result of a Psychological Tax Contract: The Role of Incentives and Responsive Regulation. Law& Policy, 29(1), 102–120. https://doi.org/10.1111/ j.1467-9930.2007.00248.x Fenochietto, M. R.,& Pessino, M. C.(2013). Understanding Countries’ Tax Effort. International Monetary Fund. 40

Fjeldstad, O.-H., Schulz-Herzenberg, C.,& Sjursen, I. H.(2012). Peoples’ views of taxation in Africa: A review of research on determinants of tax compliance(Working Paper WP 2012:7). Chr. Michelsen Institute(CMI). https://www.cmi.no/publications/4577-peoples-views-of-taxation-in-africa Franzoni, L. A.(1998). Tax Evasion and Tax Compliance(SSRN Scholarly Paper No. 137430). Social Science Research Network. https://doi.org/10.2139/ssrn.137430 GIZ Sector Programme Public Finance, Administrative Reform.(2010). Addressing tax evasion and tax avoidance in developing countries. https://www.giz.de/public-finance Income Tax Act, 2023 Pub. L. No. Act No. XII of 2023(2023). https://nbr.gov.bd/uploads/acts/Income_ tax_act_2023.pdf International Monetary Fund.(2025). THIRD AND FOURTH REVIEWS UNDER THE EXTENDED CREDIT FACILITY ARRANGEMENT AND THE ARRANGEMENT UNDER EXTENDED FUND FACILITY, REQUESTS FOR EXTENSION OF THE ARRANGEMENTS, AUGMENTATION AND REPHASING OF ACCESS, MODIFICATION OF PERFORMANCE CRITERIA, WAIVER OF NONOBSERVANCE OF PERFORMANCE CRITERION, AND THIRD AND FOURTH REVIEWS UNDER THE RESILIENCE AND SUSTAINABILITY FACILITY ARRANGEMENT(Country Report No. 2025/150)[PRESS RELEASE; STAFF REPORT; AND STATEMENT BY THE EXECUTIVE DIRECTOR FOR BANGLADESH]. International Monetary Fund(IMF). https://doi.org/10.5089/9798229015714.002 Jimenez, P.,& Iyer, G. S.(2016). Tax compliance in a social setting: The influence of social norms, trust in government, and perceived fairness on taxpayer compliance. Advances in Accounting, 34(C), 17–26. Khan, T. I., Kamal, M.,& Talukdar, F.(2019). Potential of personal income tax in Bangladesh: An examination of survey data(Working Paper No. 126). Centre for Policy Dialogue(CPD). Lewis, A.(1982). The Psychology of Taxation. New York: St. Martin’s Press Maboshe, M.,& Woolard, I. D.(2018). Revisiting the impact of direct taxes and transfers on poverty and inequality in South Africa(WIDER Working Paper No. 2018/79). https://doi.org/doi:10.35188/ UNU-WIDER/2018/521-3 Meyer, B. D.,& Sullivan, J. X.(2003). Measuring the Well-Being of the Poor Using Income and Consumption (Working Paper No. 9760). National Bureau of Economic Research. https://doi.org/10.3386/ w9760 Moazzem, K. G., Habib, S. A.,& Fariha, C.(2023). Corporate Tax Transparency Issues and Concerns in Bangladesh. Centre for Policy Dialogue(CPD). National Board of Revenue.(2023). Income Tax Paripatra 2021-2022. National Board of Revenue. https:// nbr.gov.bd/taxtypes/income-tax/income-tax-paripatra/eng National Board of Revenue.(2023, October). Annual Report 2021-22. https://www.nbr.gov.bd/ National Board of Revenue.(2024). Tax Expenditure Estimation in the Direct Tax of Bangladesh: Financial Year 2021-22. National Board of Revenue.(2025, July). MEDIUM-AND LONG-TERM REVENUE STRATEGY FY2025-26 to FY2034-35. https://nbr.gov.bd/uploads/publications/MLTRS_E_Book_Version.pdf OECD.(2012). Understanding and Influencing Taxpayers’ Compliance Behaviour[Information Note]. OECD Publishing. https://doi.org/10.1787/47eadd78-en OECD.(2013). OECD Framework for Statistics on the Distribution of Household Income, Consumption and Wealth. OECD Publishing. https://doi.org/10.1787/9789264194830-en OECD.(2021). Inheritance Taxation in OECD Countries(OECD Tax Policy Studies No. 28). OECD Publishing. https://doi.org/10.1787/e2879a7d-en 41

OECD.(2025). Revenue Statistics in Asia and the Pacific 2025: Personal Income Taxation in Asia and the Pacific. OECD Publishing, Paris. https://doi.org/10.1787/6c04402f-en Onu, D., Oats, L., Kirchler, E.,& Hartmann, A. J.(2019). Gaming the System: An Investigation of Small Business Owners’ Attitudes to Tax Avoidance, Tax Planning, and Tax Evasion. Games, 10(4), 1–17. Picur, R.,& Riahi-Belkaoui, A.(2006). The impact of bureaucracy, corruption and tax compliance. Review of Accounting and Finance, 5, 174–180. https://doi.org/10.1108/14757700610668985 PricewaterhouseCoopers(PwC).(2025). Bangladesh—Corporate—Other taxes. https://taxsummaries.pwc. com/bangladesh/corporate/other-taxes Razzaque, M. A., Islam, D., Rahman, J.,& Ali, S. R.(2024). Assessing the Impact of Indirect Taxation on Poverty and Inequality: A Pseudo-Panel and Cross-Country Analysis(BGD-23087). International Growth Centre(IGC). Razzaque, M. A.,& Khan, M. A.(2021). Resource Mobilization for Bangladesh’s Smooth Graduation from the Group of LDCs: Issues, Challenges, and Way Forward[Working Paper]. Research and Policy Integration for Development(RAPID). Razzaque, M. A., Rahman, J., Rabi, R. I., Islam, D.,& Eusuf, A.(2023). Using direct taxation to boost revenue and tackle inequality: A political economy analysis(Working Paper BGD No. 21181). Rose, M.(2017, October 3). Macron fights“president of the rich” tag after ending wealth tax. Reuters. https://www.reuters.com/article/world/macron-fights-president-of-the-rich-tag-after-endingwealth-tax-idUSKCN1C82DF/ Saez, E.,& Zucman, G.(2016). Wealth Inequality in the United States since 1913: Evidence from Capitalized Income Tax Data. The Quarterly Journal of Economics, 131(2), 519–578. https://doi. org/10.1093/qje/qjw004 Saez, E.,& Zucman, G.(2019). Progressive Wealth Taxation. Brookings Papers on Economic Activity, 2019(2), 437–533. Sarin, N.(2026, February 13). The Case for a Robust Attack on the Tax Gap. U.S. Department of the Treasury . https://home.treasury.gov/news/featured-stories/the-case-for-a-robust-attack-on-thetax-gap Sebele-Mpofu, F. Y.(2020). Governance quality and tax morale and compliance in Zimbabwe’s informal sector. Cogent Business& Management, 7(1), 1794662. https://doi.org/10.1080/23311975.2020 .1794662 Sherraden, M.(1991). Assets and Welfare. Society, 29(1), 64. https://doi.org/10.1007/BF02695365 Sikayu, S. H., Rahmat, M.,& Chan, A. N.(2022). Fairness, transparency and attitude towards tax evasion amongst owners of SMEs/ Susan Hydra Sikayu...[Et.al.]. International Journal of Service Management and Sustainability(IJSMS), 7(1), 185–206. Sour, L.(2001). An economic model of tax compliance with individual morality and group conformity. http://repositorio-digital.cide.edu/handle/11651/4136 Székely, M.,& Hilgert, M.(1999). What’s Behind the Inequality We Measure? An Investigation Using Latin American Data(Working Paper No. 234). Luxembourg Income Study(LIS). https://www.econstor. eu/handle/10419/160906 Taylor, N.(2004). Understanding Taxpayer Attitudes Through Understanding Taxpayer Identities. In Taxing Democracy, 71-92. The Value Added Tax and Supplementary Duty Act, 2012 Pub. L. No. Act No. 47 of 2012(2012). https:// nbr.gov.bd/uploads/acts/18.pdf 42

Thomas, A. G. A.(2023). Measuring Tax Progressivity in Low-Income Countries[Policy Research Working Paper]. Washington, DC: World Bank. https://doi.org/10.1596/1813-9450-10460 Tonry, M. H.(2011). The Oxford Handbook of Crime and Public Policy. Oxford University Press. Torgler, B.(2003). Beyond Punishment: A Tax Compliance Experiment with Taxpayers in Costa Rica(Revista de Análisis Económico Vol. 18, N o 1, pp. 27–56). https://papers.ssrn.com/abstract=1255342 Torgler, B.(2011). Tax Morale and Compliance: Review of Evidence and Case Studies for Europe(Policy Research Working Paper No. 5922). World Bank. https://papers.ssrn.com/abstract=1977173 Torgler, B., Demir, I. C., Macintyre, A.,& Schaffner, M.(2008). Causes and Consequences of Tax Morale: An Empirical Investigation. Economic Analysis and Policy, 38(2), 313–339. https://doi.org/10.1016/ S0313-5926(08)50023-3 Torgler, B.,& Schneider, F.(2005). Attitudes Towards Paying Taxes in Austria: An Empirical Analysis. Empirica, 32, 231–250. Tyler, T.(2006). Why do People Obey the Law. Princeton University Press. United Nations Economic and Social Commission for Asia and the Pacific(UNESCAP).(2014, September). Economic and Social Survey of Asia and the Pacific 2014. United Nations Publication. https://doi. org/10.18356/7189f5cd-en Walsh, K.(2012). Understanding Taxpayer Behaviour – New Opportunities for Tax Administration. The Economic and Social Review, 43(3), 451–475. Webley, P., Robben, H., Elffers, H.,& D.J.Hessing.(1991). Tax evasion: An experimental approach. Zucman, G.(2013). The Missing Wealth of Nations: Are Europe and the U.S. net Debtors or net Creditors?*. The Quarterly Journal of Economics, 128(3), 1321–1364. https://doi.org/10.1093/qje/qjt012 Zucman, G.(2016). The Hidden Wealth of Nations: The Scourge of Tax Havens(T. L. Fagan& with a F. by T. Piketty, Trans.). University of Chicago Press. https://press.uchicago.edu/ucp/books/book/ chicago/H/bo20159822.html 43

Appendix A1. Adjustment of Potential Tax Revenues Estimated from HIES 2022 Raw survey-based estimates of tax potential cannot be directly benchmarked against national accounts aggregates without appropriate scaling, given the well-documented tendency of household surveys to underrepresent both the top of the income distribution and the magnitude of private consumption compared to what is recorded in the national accounts. This section describes the methodological adjustments applied to the potential Personal Income Tax(PIT) and Value Added Tax(VAT) revenue estimates derived from the HIES 2022, so that they may be expressed as credible shares of GDP. The adjustments address two distinct sources of downward bias: the gap between HIES-based aggregates and national accounts, and the systematic underrepresentation of high-income individuals in survey data. In FY2021-22, • Total household income(from HIES)= BDT 1,370,000 crore • Total household expenditure(From HIES)= BDT 1,319,000 crore • Actual GDP at current market prices= BDT 3,971,716.40 crore • Actual private consumption at current market prices= BDT 2,743,526.90 crore • Actual total tax revenue= BDT 301,050.34 crore • Actual PIT revenue= BDT 38,827.43 crore(0.98 per cent of GDP) • Actual(domestic) VAT collection= BDT 71,350.35 crore(1.8 per cent of GDP) • Potential PIT revenue from HIES= BDT 71,231.86 crore • Potential(domestic) VAT from HIES= BDT 72,911.46 crore A fundamental comparability problem is immediately apparent. Actual private consumption at current market prices substantially exceeds both the total household income and total household expenditure figures derived from HIES 2022. Total household expenditure measured from the survey data captures only approximately 33 per cent of GDP for FY22. In the national accounts, by contrast, private consumption constituted 69.08 per cent of GDP. Consequently, expressing potential tax estimates derived from HIES directly as a percentage of actual GDP would produce severely understated ratios. To resolve this, a GDP-equivalent figure is constructed relative to which the HIES-based potential revenue estimates can be meaningfully expressed. For this adjustment, total household expenditure from HIES is treated as the survey’s implicit counterpart to national accounts private consumption. Applying the same consumptionto-GDP ratio, the GDP equivalent consistent with HIES-level expenditure is derived as follows: 44

As shown earlier in this study, the richest 10 per cent of earners would be liable for more than 93 per cent of total personal income tax under full compliance with the statutory tax framework. Therefore, the underrepresentation of high-income individuals in survey data can significantly affect the precision of the estimated potential PIT revenue. To address this issue, the potential PIT-to-GDP ratio is further adjusted to account for the underestimation of top incomes in HIES 2022. In this regard, income cutoffs across different percentiles of income tax filers, obtained from anonymised micro-level administrative PIT return data for the 2022–23 assessment year, are compared with the corresponding income cutoffs derived from HIES and the comparison is demonstrated in Table 4.6. To ensure comparability with NBR tax filer data, the HIES income distribution is adjusted so that the 10th percentile cutoff matches that of the tax filer distribution of the NBR’s data. Since the minimum taxable income threshold is BDT 300,000, the average percentage difference between the income cutoffs from the 60th(given that income levels below the 60th percentile fall below the taxable income threshold) to the 99th percentiles is used to further adjust the PIT-to-GDP ratio to account for the missing rich in HIES 2022. Table 4.6: Percentile-wise comparison between income cutoffs between NBR tax return data and HIES 2022 Income Percentile 10th 20th 30th 40th 50th 60th 70th 80th 90th 99th NBR’s Tax Return Data Income Cutoff(Full sample of income tax filers) 190,000 247,510 281,500 321,872 360,000 410,000 492,700 639,600 980,212 4,448,750 HIES 2022 Income cutoff with 10th percentile cutoff= 190,000 190,000 205,881 224,933 247,090 280,000 321,700 384,300 491,990 766,320 3,317,250 Percentage difference relative to the HIES cutoff 27 28 30 28 34 Average ≈ 30 Source: Chowdhury(2026); authors’ analysis using the Household Income and Expenditure Surveys, 2022. Note: The NBR’s tax return data include income tax filers whose reported incomes fall below the taxable threshold. Although these low-income filers are numerous, their contribution to revenue is negligible, with the bottom 60 per cent of the income distribution accounting for only 2.2 per cent of total PIT revenue(Chowdhury, 2026). The comparison reveals that taxable income reported in NBR’s administrative tax return data is, on average, approximately 30 per cent higher than the corresponding income cutoffs in HIES 2022. A 30 per cent upward adjustment is therefore applied to the unadjusted potential PIT-to-GDP ratio to account for the missing rich: 45

Since potential revenue achievable from other types of taxes like corporate income tax, supplementary duty, excise, customs duties, AIT, etc., under full compliance, are beyond the scope of this study, accounting only for the estimated PIT and VAT gaps: Thus, even limiting the simulation to VAT and PIT alone, the potential tax effort could reach approximately 13.47 per cent of GDP under full compliance. 46

A2. Estimation of Private Wealth Distribution The total net wealth of each household in the sample is estimated by adding the values of all reported assets and subtracting the total outstanding loans. Asset components include the value of the dwelling house, operational and non-operational land, other real and financial assets such as stocks, bonds, jewellery, non-agricultural businesses, agricultural assets, livestock, durable goods, savings, insurance, and lottery winnings. In HIES 2022, the values of most assets, such as owner-occupied housing, agricultural and non-agricultural assets, livestock, durable goods, insurance holdings, lottery earnings, and other non-land assets, are directly reported in the dataset. For example, the respondent of each household was asked,“If you want to buy or construct a dwelling just like this today, how much money would you have to pay?” This reported value of the dwelling house is included as an asset of the households that own the dwelling house. Such valuations are based on respondents’ self-assessments of their assets’ value. Savings data are partially available; that is, households report how much they saved from remittances received over the preceding two years, and this amount is included in the wealth estimate. The valuation of land requires additional assumptions. Although the survey reports the area of operating land owned by households, its monetary value is not provided. The area of operating land for each household is defined as the sum of total cultivable agricultural land owned, total dwelling-house or homestead land owned, and total non-cultivated land owned. For individuals engaged in agricultural activities and identified as day labourers whose data regarding owning any operational lands are missing, zero ownership of operational land is assumed, because landowners involved in agricultural work typically cultivate their own land and do not classify themselves as agricultural day labourers. For non-operating land, both area and value are reported. Households were asked whether any land or property was owned but not operated, and how much it would cost to purchase that land or property. Using this information, unit land prices were estimated for households owning non-operational land. By combining operating and non-operating land data, it was found that, out of 14,261 households, an observable unit land price could be directly derived for only 662 households owning non-operational lands. Therefore, median unit land prices were computed at the upazila(sub-district) and district levels, separately for rural and urban areas. The value of operational land was imputed by multiplying the reported operational land area by the corresponding median land price at the upazila level. For households owning both operational and nonoperational land, the unit price derived from the reported value of non-operational land was applied. For the households that did not have a corresponding upazila-level median unit land prices, district-level median prices were used instead. One of the limitations of this approach is that the survey does not fully capture ownership of additional residential properties, except for houses purchased within 12 months prior to the survey, which can lead to an underestimation of wealth. In addition, savings data are incomplete as only remittance-related savings are reported. Including these partial savings figures may introduce some bias into the net wealth estimates as well. 47

About the Authors Imprint Mohammad Abdur Razzaque is an economist specialising in international trade and development issues. Currently, he serves as Chairman of Research and Policy Integration for Development(RAPID). Md. Deen Islam is Research Director at RAPID and a Professor of Economics at the University of Dhaka. He obtained his PhD in economics from Boston University. Tahsin Rifat is a Research Associate at RAPID. She completed her BSS and MSS in Economics from the University of Dhaka. © 2026 Research and Policy Integration for Development(RAPID) Jointly Published by Research and Policy Integration for Development(RAPID) and Friedrich-Ebert-Stiftung(FES) Bangladesh Funded by Friedrich-Ebert-Stiftung(FES) Bangladesh Responsible Research and Policy Integration for Development(RAPID) Contact info@rapidbd.org bangladesh@fes.de Copy editing S.M. Asif Ehsan, Ph.D Printed by SR Printing and Supply, Dhaka About RAPID Research and Policy Integration for Development(RAPID) is a private, non-profit, and non-partisan policy research organisation based in Dhaka, Bangladesh. It engages in rigorous research and policy advocacy to inform national discourse, support the design of effective interventions, evaluate impacts, and contribute to informed decision-making. Committed to advancing evidence-based policymaking, RAPID employs both quantitative and qualitative methods across key thematic areas including trade, LDC graduation, social protection, the SDGs, human development, and climate change. It serves as a platform for expert engagement and facilitates inclusive policy dialogue through its convening power. RAPID's work is tailored to serve policymakers, development partners, researchers, and other stakeholders. About FES Friedrich-Ebert-Stiftung(FES) is a non-profit organization established in 1925 by social democrats in order to promote education of the less privileged. In the 100 years of its existence, it has developed into an internationally active think tank promoting political research, education and debate. It is named after Germany's first democratically elected president, Friedrich Ebert(1871-1925). Its mission is based on the values of social democracy: liberty, justice& solidarity. Disclaimer The views expressed in this publication are solely those of the authors and do not necessarily reflect the views of Research and Policy Integration for Development(RAPID) or Friedrich-Ebert-Stiftung(FES) Bangladesh. The commercial use of any media or materials published by FES Bangladesh or RAPID is strictly prohibited without prior written consent from them. Empowering Development Friedrich Ebert@ Stiftung Research and Policy Integration for Development(RAPID) https://www.rap idbd.o r g/ https://www.facebook.com/rapidonfb Friedrich-Ebert-Stiftung(FES) Bangladesh https://bangladesh.fes.de/ https://www.facebook.com/FESBangladesh/