Selected Issues

Abstract

Selected Issues

Revenue Mobilization and Inequality in Senegal1

This paper quantitatively assesses the macroeconomic and distributional impacts of fiscal consolidation in Senegal through value added tax (VAT), personal income tax (PIT), and corporate income tax (CIT). We analyze the trade-offs between growth and equity for each tax instrument. We find that VAT has the least efficiency cost in output and consumption but expands the rural-urban inequality gap because significant VAT tax incidence falls on the rural area. PIT is the most detrimental in terms of growth and inequality. CIT on the other hand, despite causing large efficiency loss, has better distributional implications by distributing the tax burden more evenly across regions. Much of the output and distributional costs can be mitigated by using the additional revenue for infrastructure investment and cash transfer.

A. Background

1. Senegal has maintained economic growth rates above 6 percent over the past four years and has improved the living standards of its citizens. Notwithstanding this achievement, according to the World Bank, about 47 percent of the population still live in poverty. Those living in urban areas have far better access to resources than those in rural areas, raising the need to sustain the high growth and tackle inequality.

2. Revenue increases are needed to finance investment in a sustainable manner. For Senegal to sustain robust growth and meeting its development objectives, it will need to increase revenues to finance its ambitious public investment agenda. At the median level, tax revenue as a share of GDP is around 15 percent in LIDCs, and around 20 percent in emerging market economies (EMEs). While the appropriate level of taxation depends on country characteristics, increasing tax capacity is an important element for improving living standards and attaining the sustainable development goals. In this context, increasing the tax-to-GDP ratio by 5 percentage points of GDP in the next decade is a reasonable target in many countries (Gaspar and others, 2018). Senegal has room to improve its tax revenue. At present, Senegal tax-to-GDP ratio is less than 16 percent, while the regional WAEMU’s target is set at 20 percent. Increasing the country’s tax-to-GDP ratio by 4 percentage points, in line with the WAEMU’s 20 percent, is an achievable target. Over the medium term, reforms will be needed to achieve that goal.

3. Sustainable development and inclusive growth also requires tackling inequality. Cash transfer programs in Senegal have been shown to have a positive impact on the beneficiaries. For example, the evaluation of a program offering a lump sum cash transfer of about US$200 to small farmers in Senegal reveals that after one year, agricultural production and livestock ownership was significantly higher in the transfer group compared to the control group, where the increase in productivity came mostly through increased investments in intermediate agricultural inputs like chemical fertilizer (Ambler, Brauw, and Godlonton, 2017; Donovan, 2018).

4. This paper simulates the macroeconomic and distributional impact of fiscal consolidation aimed at increasing the revenue to GDP ratio by 4 percentage points using a package of reforms on VAT, CIT, and PIT. The analysis also examines the impact of using part of the additional revenue from the fiscal consolidation for investment and cash transfer to rural areas. We show that the tax reform alone without efficient use of the additional revenue collected can cause a decrease in output as high as 4 percent and a slight increase in the total income Gini of 0.73 percent. However, the output and distributional costs are mitigated when the additional revenue is used for cash transfer programs and infrastructure investment, with output gain as high as 10 percent, depending on the efficacy of public investment.

B. Impact of Revenue and Expenditure Measures on Inequality

5. The literature has firmly established that tax reforms redistribute tax burden across households which leads to profound macroeconomic and distributional impacts that are usually not Pareto improving.2 The design of the policy thus must balance the efficiency and distributional costs of different tax instruments available to the government. Moreover, the economic structure of Senegal differs significantly from that of advanced economies which the literature studies extensively. The large agricultural sector, substantial informality, underdeveloped financial markets and a sharp rural-urban distinction imply that commonly used tax instruments have very different trade-offs compared to those found in previous studies. As a result, the potential impact of revenue mobilization is better evaluated quantitatively using a heterogeneous agents general equilibrium model that captures salient features of the Senegalese economy.

6. The general equilibrium model with heterogeneous agents and incomplete markets developed in Peralta-Alva and others (2018a) is used to explore quantitatively the impact of revenue mobilization using VAT, CIT, and PIT. The model is a modified version of the Aiyagari (1994) model. It contains four sectors of different productivity levels—food, manufacturing, services, and exporting cash crops—two regions with segmented labor markets—rural and urban—and a unified capital market. Each region is populated by a continuum of households who consume food, manufacturing goods, and services. Each household also faces persistent idiosyncratic productivity shocks that can only be partially insured against using a one period risk-free bond, which provides a parsimonious way of modeling the limited development of financial markets. Based on their comparative advantage, households divide their total hours between the formal and informal labor markets in their dwelling region. The formal and informal labor markets in each region host different sectors. The agricultural sector is hosted exclusively in the rural area, where workers in the formal and informal markets are hired to produce respectively cash crops and food. On the other hand, manufacturing goods and services are provided respectively by urban households who work in the formal and informal market. The government can tax food and manufacturing goods consumption by VAT, formal wage income by PIT, and revenues of manufacturing firms and profits from cash crops production by CIT.

7. The model is calibrated to the Senegalese economy to reflect broadly its macroeconomic and distributional features. Specifically, we require the model to match the share of agriculture, manufacturing, and informal services in consumption and output, the share of tax revenues from VAT, PIT, and CIT, and the rural and urban income Gini coefficients. We defer the details of the calibration strategy to the Annex. It is worth mentioning that although we do not directly target the overall Gini coefficient, the value of 0.63 generated by the model is very close to the observed 0.65 in the data, lending support to the validity of the results of the model.

8. All the quantitative results are from steady state comparisons which are deemed to capture long-run effects of tax reforms. However, in the context of Senegal (and low-income countries more generally), steady-state comparisons also provide close approximation to short-run impacts when transitional dynamics are considered. The reason is that the length of transition is determined mostly by how fast the capital stock changes to the level in the new steady state. The transition is short because the capital stock is low and hence the adjustment is fast (Peralta-Alva and others, 2018b).

9. The design of a revenue mobilization package calls for thorough understanding on the equity-efficiency trade-offs embedded in each of the tax instruments. For this purpose, we simulate the policy scenarios of raising tax revenue equal to 1 percent GDP by VAT, PIT, and CIT separately (Figure 1). To isolate the effect of the tax measures, the tax revenues are assumed to be spent on manufacturing goods not directly valued by households. Several insights stand out First, VAT in the model has the least efficiency cost in growth but expands the rural-urban gap significantly (Figure 1). The low efficiency cost of VAT is in line with the non-distortive nature of consumption tax found in the literature (Anagnostopoulos and Li, 2013). Second, PIT in the model has both large efficiency cost and worsens the rural-urban gap, making it the least desirable from the perspective of the economy as a whole. However, the progressivity of PIT does reduce the inequality in the urban areas. This is because in the urban area, PIT is only imposed on the formal sector, that consists mostly of more productive and, hence, wealthier households. Third, while CIT causes a moderate efficiency output decline, it is the only tax instrument that shrinks the rural-urban gap.

Figure 1.
Figure 1.

Senegal: Comparison of Tax Instruments Non-Productive Government Expenditure

Citation: IMF Staff Country Reports 2019, 028; 10.5089/9781484396292.002.A003

10. Due to the significant rural population in Senegal, the distributional implications of the taxes are mostly dominated by the redistribution of tax burden between regions. The uneven distribution of the tax burden between regions can be seen both from the total Gini coefficient and regional decrease in consumption. VAT and PIT yield large between regional redistribution costs because while tax revenue is collected from the whole economy, only the richer urban population receives the “rebate” implicitly from government purchase. Such discrepancy in the tax incidence and government expenditure identities is much less of a concern for CIT because it is mostly collected from urban manufacturing firms. The results suggest that large redistribution between regions happens whenever tax incidence differs from recipients of government expenditure. The multi-sector nature of the model is key to generate this insight.

11. Because each tax has its own benefit and cost, the need to mobilize an additional 4 percent of GDP in revenues will require broad contribution from all tax instruments. Based on the previous results which found the VAT to have the least efficiency cost, the analysis suggests that raising 50 percent of the tax revenue from VAT and 25 percent each from PIT and CIT would provide a good balance given trade-offs. In the proposed package, VAT is relied on heavily to minimize the overall impact of fiscal consolidation on aggregate output, consumption, and investment. Additional tax revenue is collected through PIT and CIT to balance the within and between region redistribution concerns.

12. Under the assumption that the additional tax revenue collected is used solely for current and non-productive government expenditure on manufacturing goods, the package of measures causes a decrease in aggregate output, reduces within region inequality, and slightly enlarges the rural-urban gap (Figure 2). The reform has a more significant adverse impact on the rural area than the urban area by causing a larger drop in rural consumption. The assumption of non-productive government expenditure is made to isolate the impact of the taxes. The effects of the combined tax package are consistent with the aggregation of those from individual taxes.

Figure 2.
Figure 2.

Senegal: Main Experiment: Non-Productive Government Expenditure

Citation: IMF Staff Country Reports 2019, 028; 10.5089/9781484396292.002.A003

13. The detrimental growth and social effects of fiscal consolidation could be offset partially by using a portion of the increase in tax revenues to finance infrastructure investment and social transfers targeting the rural area. Because the objective of fiscal consolidation is to allow the government to have the fiscal space for increasing investment efficiency and expanding the social safety net, the assumption of non-productive government expenditure overstates the cost of the reform. Since the previous analysis shows that the urban-rural gap is the most important source of negative distributional impact from the tax reform, we evaluate quantitatively to what extent a cash transfer program targeting the rural area can mitigate the widening urban-rural gap. We also investigate how public investment that boosts aggregate TFP would modify the outcome of the tax reform. In particular, if 25 percent and 50 percent of the additional revenues are used instead, respectively, for cash transfers to rural area and infrastructure investment to boost the aggregate TFP of the economy, then the losses in output and consumption would be mitigated or completely overturned, depending on the efficiency of the public investment3 At the same time, both the overall and rural income Gini decline sharply (Figure 3). In addition, Figure 4 singles out the effects from changes on the expenditure side. Specifically, the cash transfer program substantially reduces the within and between region inequality, but leads to a slight extra output loss owing to the crowding out of private capital associated with the decreasing precautionary saving motive (Aiyagari, 1994). On the other hand, public investment improves the overall efficiency significantly while leaving inequality largely untouched.

Figure 3.
Figure 3.

Senegal: Comparison of Expenditure Schemes: Compared to No Reform, Alpha = 0.05

Citation: IMF Staff Country Reports 2019, 028; 10.5089/9781484396292.002.A003

Figure 4.
Figure 4.

Senegal: Comparison of Expenditure Schemes Compared to Non-Productive Expenditure Alpha = 0.05

Citation: IMF Staff Country Reports 2019, 028; 10.5089/9781484396292.002.A003

14. A conservative estimate of investment efficiency is enough to overturn the negative impact from the fiscal consolidation. In the simulation, a conservative estimate of the productivity enhancing effect of public investment is used (α = 0.05), with 2 percent of GDP of the additional tax revenue going into investment and leading only to a 4.65 percent increase in TFP. The results reveal that the negative impact from revenue mobilization is overturned (Figures 3 and 4).

15. Senegal has room to improve its investment efficiency to reap the full benefits of the fiscal consolidation. While Senegal’s recent Public Investment Management Assessment (PIMA) points to a relatively strong and fairly well implemented existing framework for public management investment, it also highlights some efficiency gaps. To facilitate the implementation of Plan Sénégal Emergent (PSE)—aimed at closing infrastructure gaps in roads, education, health and electricity sectors, while creating jobs—the authorities have undertaken several reforms to improve investment efficiency, some of which would be reinforced by implementation of the 2018 PIMA recommendations. These include: (i) streamlining the triennial investment plan; (ii) bringing implementation of ad hoc investment projects back under normal procedures and proper oversight; and (iii) updating the legal framework for PPPs.

16. Under improved investment efficiency, the benefits from the fiscal consolidation reform would be even higher. Under an assumption that the productivity enhancing effect from public investment is more optimistic (α = 0.10 which is the lowest bound estimated by Berg and others, 2013), TFP would increase by 9.52 percent and the benefits from the reform would be even larger (Figures 5 and 6). Specifically, output increases by about 10 percent, compared to about 2 percent when the efficiency is projected more conservatively.

Figure 5.
Figure 5.

Senegal: Comparison of Expenditure Schemes: Compared to No Reform, Alpha = 0.10

Citation: IMF Staff Country Reports 2019, 028; 10.5089/9781484396292.002.A003

Figure 6.
Figure 6.

Senegal: Comparison of Expenditure Schemes: Compared to Non-Productive Expenditure Alpha = 0.10

Citation: IMF Staff Country Reports 2019, 028; 10.5089/9781484396292.002.A003

C. Concluding Remarks

17. The model-based analysis of the macroeconomic and distributional impacts of fiscal consolidation highlights the trade-offs between growth and equity in various tax instruments. The analysis focuses on VAT, PIT, and CIT to model different tax bases parsimoniously. VAT is found to have the least efficiency cost in output and consumption, but expands the rural-urban inequality gap because most tax incidence falls on the rural area. While PIT is the most detrimental in terms of growth and inequality, CIT has better distributional implications by distributing the tax burden more evenly across regions. At the same time CIT exhibits large efficiency losses. The simulations reveal that much of the output and distributional costs can be mitigated by using the additional revenue for cash transfer programs and infrastructure investment.

18. The analysis, though highly stylized, reveals several insights into the design of an optimal fiscal consolidation package. First, all tax instruments should be included to balance the equity-efficiency costs. Second, special care should be devoted to the urban-rural gap. Third, any mismatch between the identity of tax incidence and government expenditure has important redistribution implications, and thus deserves extra caution. Fourth, carefully designed social programs and public investment can potentially overturn all the costs embedded with the increase in tax burden.

19. The proposed analysis does not necessarily call for a hike in tax rates.4 The tax rates in the model are effective rather than statutory rates. As a result, the increase in the effective tax rates in the model could be achieved by closing loopholes, improving the efficiency of the revenue administration, and streamlining tax exemptions. In addition, while the model and simulations focus on VAT, CIT, and PIT, other tax instruments such as property tax which show great potential in Senegal could also be considered.

Annex I. Calibration Parameter Values and Model Fit

The parameter values and model fit of the benchmark calibration are shown in Table 1.

Annex Table 1.

Calibration Parameter Values and Model Fit

article image

For the definition of the model parameters in Table 1, please refer to Peralta-Alva and others (2018). There are several caveats related to the calibration exercise. First, we assign modern service industries like banking, telecommunication and airlines to manufacturing sector since economically they resemble more to the manufacturing sector in the model. This also means that the sectors should not be understood solely by their traditional statistical classification. As an example, the manufacturing goods in the model should indeed be thought of as goods and services provided by the formal sectors in the urban area. Second, in a somewhat related way, tax rates should also not be taken by their face value. The tax rates in the model are in fact the effective rates. As a result, an increase in the CIT for instance, does not necessarily mean that in reality the statutory tax rate needs to be increased, as elimination of taxation loopholes would work.

References

  • Aiyagari, S. Rao, 1994. “Uninsured Idiosyncratic Risk and Aggregate Saving.” Quarterly Journal of Economics, Vol. 109, No. 3, pp. 65984.

  • Ambler Kate, Alan de Brauw and Susan Godlonton, 2017, “Cash Transfers and Crop Production in Senegal,” International Food Policy Research Institute (IFRI) Discussion Paper 01659.

    • Search Google Scholar
    • Export Citation
  • Anagnostopoulos, Alexis and Qian Li, 2013, “Consumption Taxes and Precautionary Savings.” Economics Letters, Vol. 119, No. 3, pp. 23842.

    • Search Google Scholar
    • Export Citation
  • Berg, Andrew, Rafael Portillo, Shu-Chun S. Yang, and Luis-Felipe Zanna, 2013, “Public Investment in Resource-Abundant Developing Countries,” IMF Economic Review, Vol. 61, No. 1, pp. 92129.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Conesa, Juan Carlos, Sagiri Kitao, and Dirk Krueger, 2009, “Taxing Capital? Not a Bad Idea After All!American Economic Review, Vol. 99, No. 1, pp. 2548.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Domeij, David and Jonathan Heathcote. 2004. “On the Distributional Effects of Reducing Capital Taxes,” International Economic Review, Vol. 45, No. 2, pp. 52354.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Donovan, Kevin, 2018, “Agricultural Risk, Intermediate Inputs, and Cross-Country Productivity Differences,” Manuscript.

  • Gaspar, Vitor, Amaglobeli, D., Garcia-Escribano, M., Prady, D., and Soto, M., 2018, “Fiscal Policy and Development: Human, Social and Physical Investment for the SDGs,” IMF Staff Discussion Note (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Peralta-Alva, Adrian, Xuan Song Tam, Xin Tang, and Marina Mendes Tavares, 2018a, “The Macroeconomic and Distributional Impacts of Fiscal Consolidations in Low-Income Countries,” IMF Working Paper 18/146 (Washington: International Monetary Fund).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Peralta-Alva, Adrian, Xuan Song Tam, Xin Tang, and Marina Mendes Tavares, 2018b, “The Welfare Implications of Fiscal Consolidations in Low-Income Countries,” Manuscript.

    • Search Google Scholar
    • Export Citation
1

Prepared by Xin Tang, and Etienne B. Yehoue.

3

The remaining 25 percent can be assumed, for instance, to be used to pay down debt.

4

As noted in Box 1 of the accompanying staff report for 2018 Article IV Consultation and the Seventh Review Under the PSI, given already high tax rates in Senegal, a large portion of revenue gains are likely to come from reforms to improve tax administration and widen the base, including by reducing tax exemptions.

Senegal: Selected Issues
Author: International Monetary Fund. African Dept.
  • View in gallery

    Senegal: Comparison of Tax Instruments Non-Productive Government Expenditure

  • View in gallery

    Senegal: Main Experiment: Non-Productive Government Expenditure

  • View in gallery

    Senegal: Comparison of Expenditure Schemes: Compared to No Reform, Alpha = 0.05

  • View in gallery

    Senegal: Comparison of Expenditure Schemes Compared to Non-Productive Expenditure Alpha = 0.05

  • View in gallery

    Senegal: Comparison of Expenditure Schemes: Compared to No Reform, Alpha = 0.10

  • View in gallery

    Senegal: Comparison of Expenditure Schemes: Compared to Non-Productive Expenditure Alpha = 0.10