Domestic Revenue Mobilization in the Democratic Republic of the Congo1
This paper analyzes domestic revenue mobilization in the DRC and offers options to strengthen it. Domestic revenue mobilization (DRM) in the DRC has improved during the ECF program, standing at 13.7 percent over GDP in 2023, though it remains persistently low relative to peer countries. The recent improvements in revenue mobilization have been driven by stronger corporate income taxation (particularly stemming from the extractive sector). A comparison between DRC’s and peer countries’ tax structure points to significant room for boosting domestic revenues with stronger mobilization of personal income taxes, taxes on international trade and transactions and goods and services. In addition, the country’s tax potential (estimated on the basis of its structural characteristics and a stochastic frontier model) points to significant scope for improving tax-to-GDP ratio, by about 10 percentage points under more efficient tax policy and tax collection. Finally, tax administration reforms based on recommendations from the recently published TADAT report can significantly contribute to boosting DRM, with particular focus on tax-avoidance in the mining sector.
A. Domestic Revenue Mobilization in the DRC—Recent Dynamics and Developments
1. Revenue to GDP ratio has increased in program years, starting from a very low level. Since the beginning of the current program, general government revenue (excluding grants) to GDP ratio has improved significantly, reaching its peak at 15.4 percent in 2022 (linked to the commodity price boom) and declining to 13.7 percent in 2023 (Figure 1). Prior to the program, revenue ratios were weaker, on average 11.3 percent during 2010–2019.
2. However, revenue to GDP remains consistently below that of peer countries. Despite the encouraging recent developments, revenue mobilization remains consistently below comparable countries. Revenue to GDP has been the lowest among peer countries (resource-rich, non-oil SSA countries) and significantly lower than the Sub-Saharan Africa (SSA), low-income country (LIC) and fragile and conflict-afflicted states (FCS) averages (Figure 1). As an illustration, in 2023, the difference to SSA average was at 5.1 percentage points, while the 10-year average difference stood at 6.8 percentage points.
Revenues to GDP Ratios for DRC and Peer Countries
(Percent)
Citation: IMF Staff Country Reports 2024, 227; 10.5089/9798400283451.002.A001
Sources: World Economic Outlook (April-24 vintage) and IMF staff calculations.Note: SSA = sub-Saharan Africa; LIC = low-income countries; FCS = fragile and conflict affected states.3. Revenue mobilization in recent years has been driven primarily by increases in corporate income taxes. In line with the growing significance of mining in government revenues, corporate income taxes have been the main contributor to the growth of revenue over GDP ratios during the program (Figure 2, right panel). In 2023, corporate income taxes stood at 3.6 percent of GDP followed by taxes on goods and services and non-tax revenues (at 3.5 and 3.4 percent respectively). Revenues from corporate income taxes remained substantial in 2023, despite the weak copper and cobalt market conditions, pointing to the structurally important role of mining in government revenues.
Tax Structure for DRC and Peer Countries
(Percent of GDP)
Citation: IMF Staff Country Reports 2024, 227; 10.5089/9798400283451.002.A001
Sources: World Economic Outlook (April-24 vintage), country authorities, and IMF staff calculations.Note: SSA = sub-Saharan Africa; LIC = low-income countries; FCS = fragile and conflict affected states.4. Comparison with peer countries points to personal income, VAT and international trade and transactions taxation as potential sources for boosting domestic revenue mobilization. DRC underperforms in terms of these three tax categories relative to peer countries: personal income taxes have stagnated at 1.3 percent of GDP in the past three years, while taxes on international trade and transactions stood at 1 percent of GDP, both reflecting marginal improvement from the pre-program three-year averages of 1.2 and 0.8 percent respectively. Comparatively, the SSA average for 2023 for these categories reached 2.9 and 2.6 percent respectively (Figure 2, left panel). Regarding goods and services taxes, they have been stable in DRC at about 3.5 percent of GDP, substantially lower than the SSA average which reached 5.6 percent in 2023. Given the ease of VAT enforcement relative to personal and corporate income taxation, and evidence of its positive impact on reducing informality,2 these international comparisons point to significant scope for improving domestic revenues via stepping up goods and services tax collection efforts. Significant reforms are required, including addressing the large degree of informality and tax administration capacity bottlenecks, to allow the DRC to converge to average SSA levels in these tax categories.
5. Beyond corporate income taxation, DRC’s extractive sector is also a source of non-tax revenues, including royalties, dividends from national resource companies and income from profit sharing agreements. Royalties from mineral production have been on an upward trend in recent years, peaking at 1.1 percent of GDP in 2022 before subsiding to 0.7 percent in 2023, and are anticipated to become an increasingly important source of revenue with the recent Sicomines deal (Annex VI). Dividends and other income from profit-sharing agreements represent minor components of revenues, potentially linked to reporting and governance challenges of the Gecamines state-owned company.
Non-Renewable Resource Revenues
(Percent of GDP)
Citation: IMF Staff Country Reports 2024, 227; 10.5089/9798400283451.002.A001
Sources: Country authorities and IMF staff calculations.6. Revenues from the mining sector account for roughly one-third of the government’s domestic revenues. They are mainly derived from corporate income tax (CIT), royalties and signing bonus instalments as well as import duties. These three revenues account for 67 percent of total revenues on average over the last five years. Revenues from the mining sector have increased significantly since the new Mining Code in 2018. This could be explained by increased production of the DRC’s main export mining products (copper and cobalt) and the government’s larger share of mining companies’ profits. Average revenues from the mining sector rose from USD 4,002 million to USD 5,541 million, respectively for the periods 2013–2017 and 2018–2022. However, there is an almost perfect correlation between domestic revenues and copper prices (Figure 4).
Domestic Revenues and Copper Prices
Citation: IMF Staff Country Reports 2024, 227; 10.5089/9798400283451.002.A001
Sources: Country authorities and IMF staff calculations.DRC Mining Sector Revenue
Citation: IMF Staff Country Reports 2024, 227; 10.5089/9798400283451.002.A001
Sources: Congolese authorities and IMF staff calculations.1/ DGI = Direction Générale des Impôts (Directorate General of Taxes); DGDA = Direction Générale des Douanes et Accises (Directorate General of Customs and Excises); DGRAD = Direction Générale des Recettes Administratives, Judiciaires, Domaniales et de Participations (Directorate General of Administrative, Judicial, State, and Participations Revenue).B. Estimating DRC’s Tax Potential
7. DRC’s tax potential depends on several structural, institutional and policy factors. The size and structure of the country’s economy is critical for determining tax potential, as is the authorities’ capacity to access different tax bases in an efficient and equitable manner. To assess DRC’s tax revenue potential (or else tax frontier) we employ stochastic frontier model analysis (see Box I for technical details). This is a regression-based method that allows to estimate the highest level of tax revenue that DRC could achieve at its best performance, based on key underlying country characteristics and on an empirically determined benchmark observed in other peer countries. Within this framework, deviations from the tax frontier are driven by unobserved country heterogeneity which can be interpreted as unobserved tax policy and tax administration inefficiency (such as country-specific inefficiencies in tax codes, rates, exemptions and collection capacity and efforts).
8. To identify those characteristics, we first establish an econometric relationship between tax-to-GDP and relevant factors for this ratio in a sample of comparator countries. The key explanatory variables, that have been found in previous literature to be significant predictors of tax ratios are: (i) GDP per capita which aims to capture the level of development and the size of potential tax base of the economy; (ii) the share of agriculture value added in GDP, which captures economic development and may be linked to higher degrees of informality for countries reliant on subsistence farming; (iii) openness to trade (measured as the sum of imports and exports over GDP), as trade taxes are an important source of revenue and trade provides the opportunity to collect taxes relatively easily on clearly identified transit points (roads, ports airports); (iv) size of international aid, which may create a disincentive for stepping up tax efforts given the availability of external sources and is linked to lower level of development; (v) the rural population share, as rural population is associated with lower income levels and higher collection costs; and (vi) debt service to GNI, which captures government’s need to raise domestic revenues in order to address binding constraints.3 Data used for the regression analysis is from the IMF’s World Economic Outlook (WEO), and the World Bank’s World Development Indicators (WDI); the panel covers more than 30 years (1990–2022) and almost all SSA countries (34 countries).4
Actual and Potential Tax-to-GDP Ratios for DRC
(Percent)
Citation: IMF Staff Country Reports 2024, 227; 10.5089/9798400283451.002.A001
Sources: World Economic Outlook (WEO), World Development Indicators (WDI) and IMF staff calculations.9. The results of the stochastic tax frontier suggest that DRC’s theoretical tax capacity is close to 19 percent of GDP, significantly above its current tax ratio. The estimated tax potential has remained steadily above 15 percent of GDP over the last two decades and has hovered around 19 percent of GDP in recent years (Figure 4). The gap between actual tax revenues to GDP the tax frontier has also remained stable and is estimated at 10.7 percentage points on average over the last decade, pointing to significant scope for improving tax policy and tax collection efficiency. For benchmarking these results, it is worth noting that FAD’s 2023 SDN estimates – using similar methods and time horizon – that the tax potential of Low-Income Developing Countries (LIDCs) was 19.9 percent of GDP in 2020, comparable to DRC’s estimates. However, their average tax effort, defined as the ratio of the observed level of tax collection over the tax frontier, was 0.67, substantially higher than DRC’s tax effort, estimated at 0.36 in 2020.5
C. Impediments to Effective Domestic Revenue Mobilization
Tax Administrations and Governance Issues
10. Several factors negatively affect revenue performance:
Complexity and lack of transparency in the tax system. The tax system is incredibly complex, with hundreds of tax categories managed by different levels of government6 (central, provincial, and decentralized territorial entities or ETDs), various revenue agencies (DGI, DGRAD, DGDA) and various special funds and entities having their own revenue sources. Moreover, the legal framework for taxation and revenue collection is fragmented and lacks accessibility, primarily due to the absence of a General Tax Code. This complexity not only hinders taxpayer compliance but also facilitates corruption by increasing interactions between civil servants and taxpayers.
Exemption and tax expenditure regimes characterized by their discretionary and even arbitrary nature. This leads to a significant estimated loss of earnings—5.7 percent from exemptions and 22.3 percent from tax expenditures of the central government’s revenue. Management of these regimes, especially for strategic partnerships, involves considerable discretion with negotiable benefits. Similarly, businesses holding permits under the Investment Code see their benefits managed by a commission, yet the discretion remains, contrasting with more standardized investment promotion mechanisms like tax credits or accelerated depreciation.
Revenue management and collection inefficiencies. The proliferation of revenue collection channels and the discretionary management of tax exemptions contribute to the dispersion and mismanagement of public funds. Relations between mining operators and revenue agencies are strained, with frequent disputes over tax assessments and collections. This situation is further complicated by poor information sharing and coordination among government ministries and agencies, affecting the accuracy of tax base estimations and revenue collection.
11. Revenue administrations in DRC are ranked weak in the last TADAT evaluation (2023), further highlighting the challenges from the country’s vulnerabilities. The tax administrations suffer from a multitude of weaknesses and blockages across a range of activities, with “weak” ratings (D scores) in 26 of the 32 dimensions. Compared to the 2016 evaluation, only three dimensions did improve and three other deteriorated further. Important room of improvements exist regarding the following:
Insufficient efforts to identify new taxpayers—these have been even more deficient since the last TADAT evaluation (2016) and could be explained by a focus of the administration on collecting taxes from the traditional taxpayers, e.g. mining companies, underlining the reliance of revenue mobilization on the mining sector.
The DGI lack of formal mechanism to detect frauds, errors, or omissions in taxpayer declaration—the audit capabilities are weak, and the use of third-party data is very limited. This lack of oversight can be easily exploited by taxpayers and especially mining companies with highly skilled tax experts.
International Taxation Issues
12. Significant evidence suggest that tax evasion is substantial in DRC. The top ten Multinational Enterprises in terms of copper exports represent approximately 63 percent of total DRC exports in 2023, with their countries of origin including China, Canada, Switzerland, Kazakhstan and UK. The cobalt sector in DRC is significantly more concentrated, with only 9 firms accounting for approximately 80 percent of total exports, their origin being primarily in China, Switzerland and Kazakhstan. However, more than 60 percent of FDI inflows comes from Mauritius, an important investment “hub”7 with light taxation which is conducive to profit shifting. Mauritius is in the top-5 inbound investment source for seven of the 15 resource-intensive economies of SSA and its share on FDI investment in DRC is the highest in Africa (see Figure 2).
13. The DRC has initiated measures to mitigate profit-shifting risks in its mining sector. These tax measures limit the deductibility of interests on loans between a company in DRC and its foreign shareholders, recognize undue benefits to foreign entities as indirect profit distribution, and enforce stringent criteria for the deductibility of payment for service from abroad suppliers. The Finance Law 2015 requires firms with notable foreign connections to provide transfer pricing documentation. Despite the absence of explicit thin capitalization rules, some guidelines are set by the Mining Code and the Organisation de l’HArmonisation du Droit des Affaires Treaty. However, the effectiveness of these initiatives is currently hampered by the tax administration’s limited capacity to audit complex transactions, suggesting that bolstering audit capabilities is crucial for addressing transfer pricing and profit shifting comprehensively.
Top 5 Sources of FDI Inflows to Extractive SSA Countries
Citation: IMF Staff Country Reports 2024, 227; 10.5089/9798400283451.002.A001
D. Policy Recommendations
Short-term revenue mobilization efforts (and generally fiscal adjustment) must be implemented in parallel with structural reforms that aim to improve tax administration. Otherwise, in a context of challenged tax agencies, higher tax targets and/or a broadened tax base could lead to higher risk of tax harassment which would in turn impede firm entry/dynamism and deteriorate business climate.
The tax system must be simplified by reducing the number of taxes and consolidating revenue collection under fewer agencies. Ongoing preparation of a comprehensive General Tax Code will provide clarity and accessibility and should speed up with the support of the World Bank and the IMF.
Stronger governance frameworks to improve revenue management and reduce corruption are critical. This includes enhancing transparency in tax exemptions, improving the credibility of the budget process, and adopting clear, non-discretionary processes for tax collection.
Inter-agency coordination should be enhanced, by improving information sharing and coordination among government ministries, revenue agencies, and other stakeholders. This could facilitate more accurate tax base estimations and improve overall revenue collection efficiency.
Increased public engagement and transparency about how tax revenues are used to improve public services and infrastructure will boost compliance. Demonstrating the benefits of tax compliance to taxpayers can encourage voluntary compliance and reduce tax avoidance and evasion.
The introduction of exemptions of all kinds should be limited as much as possible as their effects are uncertain, and they may incur considerable cost in terms of tax revenue losses.
Estimating Tax Potential Using Stochastic Frontier Analysis
The Stochastic Frontier Analysis (SFA) model is a regression-based estimation method that assumes a onesided random error, differently to a traditional regression that assumes a two-sided normally distributed error term. The corollary of this assumption is that a country deviates from its highest observed level of tax revenue ratios only in one direction, namely only by underperforming relative to its frontier. Contrary to a traditional regression framework that minimizes both positive and negative residuals from observed values, thereby yielding estimates based on a country’s average performance, within the SFA framework we estimate the maxima of the dependent variable at given values of explanatory variables. Deviations between observed values and these maxima (that represent the full potential of the country) are due to a combination of random shocks and technical inefficiencies. The stochastic tax frontier model can be represented as follows:
where TR_it is the tax to GDP ratio in year t and country i, c is a constant, X_it is the set of explanatory variables (including GDP per capita in constant USD, the share of agriculture value added in GDP, openness to trade measured as the sum of imports and exports over GDP, the size of international aid measured as net official development assistance (ODA) received as a share of GNI, the rural population share, and debt service to GNI), β is the coefficient vector and there is a composite error term: v_it-u_it. The v_it is a zero-mean, symmetric error, due to which the frontier is stochastic, and it is independent of technical inefficiency. The non-negative error u_it represents unobserved inefficiency, and it reflects departures from the tax frontier which is the deterministic portion of the specification. If there are zero tax inefficiencies, then the country is collecting the maximum tax revenues possible for given values of the explanatory variables, namely it is at its tax frontier. We follow Benitez et al (2023) and assume a time-varying inefficiency SFA model for panel data, and we estimate the parameters of the stochastic frontier and the inefficiency simultaneously to avoid bias. All variables are in natural logarithms. The expected coefficient signs are the following: GDP per capita, a commonly used proxy for economic development and income levels, is expected to have a positive relationship with tax revenues to GDP; the share of agriculture value added to GDP, a proxy for informality and linked to a hard-to-tax sector is expected to have a negative relationship with tax ratios; the coefficient on trade openness is expected to be positive, as greater openness is associated with higher trade taxes and easier tax collection (given that trade occurs in specific locations); the size of international aid is expected to have a negative relationship with tax ratios as it is a proxy for lower economic development and may create disincentive for stepping up domestic revenue mobilization efforts; the coefficient on rural population share is expected to be negative, as rural population is associated with lower income levels and higher collection costs; and debt service to GNI is expected to be positively correlated with tax ratios as it reflects the need for higher revenue mobilization.
Estimates of tax potential are derived from the comparison of the observed experience of peer (SSA) countries from 1990 to 2021
References
Benitez Juan Carlos, Mansour Mario, Pecho Miguel, Vellutini Charles Vellutini, 2023, “Building Tax Capacity in Developing Countries”, Staff Discussion Notes No. 2023/006.
De Paula Aureo and Scheinkman Jose A. 2010, Value-Added Taxes, Chain Effects, and Informality, American Economic Journal: Macroeconomics.
Giorgia Albertin, Boriana Yontcheva, and Dan Devlin, 2021, “Tax Avoidance in Sub-Saharan Africa’s Mining Sector”. African and Fiscal Affairs Departments. DP/2021/022
Khwaja, Munawer Sultan & Iyer, Indira, 2014. “Revenue Potential, Tax Space, and Tax Gap: A Comparative Analysis,” Policy Research Working Paper Series 6868, The World Bank.
October 2017 SSA Regional Economic Outlook “The Impact of Fiscal Consolidation on Growth in Sub-Saharan Africa”.
April 2018 SSA Regional Economic Outlook “Domestic Revenue Mobilization in SSA: What Are the Possibilities?”.
Prepared by Laila Drissi Bourhanbour, Emmanuel Gbadi and Myrto Oikonomou
For instance, see De Paula and Scheinkman (2010) who show that higher enforcement of VAT collection via the credit method in the production stage increases formality downstream and upstream using Brazilian firm-level microdata.
See Khwaja and Iyer (2014) for an overview of tax potential literature developments, as well as October 2017 SSA regional Economic Outlook “The Impact of Fiscal Consolidation on Growth in Sub-Saharan Africa”, the April 2018 SSA Regional Economic Outlook ‘’Domestic Revenue Mobilization in SSA: What Are the Possibilities?’’ and Benitez et al (2023) that follow a similar estimation approach.
To avoid small sample bias, the sample includes resource rich as well as non-resource countries in SSA. Given that DRC is among the richest countries in the sample in terms of resource wealth (top copper producer in the region and top cobalt producer globally) we view these estimates as a lower bound of DRC’s tax frontier.
See Benitez et al (2023).
There are up to 137 possible types of taxes/charges by the 26 provinces, as well as 38, 34 and 38 possible types of taxes/charges by cities, communes and chieftaincies, respectively.
Mauritius has an FDI to GDP ratio of about 2,000 percent.
