Taxation and Development from the WIDER Perspective

Transcription

1 Taxation and Development from the WIDER Perspective Jukka Pirttilä (UNU-WIDER) UNU-WIDER 30th Anniversary Conference 1 / 29

2 Outline Introduction Modern public economics approach to tax analysis Taxes in developing countries UNU-WIDER research work on taxes (and social protection) Tax-benet microsimulation Tax analysis using register data Conclusion 2 / 29

3 This presentation Approach: how do the problems in and current research on taxation in developing countries look like from the perspective of someone who has worked in conventional public economics research? This conventional public economics includes normative analysis of how a good tax system should be designed positive empirical analysis on the causal impacts of taxes on economic behavior (such as employment, savings and investment) How combining this background as well as we could with the needs of developing countries has led to the current UNU-WIDER research program on taxation and development 3 / 29

4 Outline Introduction Modern public economics approach to tax analysis Taxes in developing countries UNU-WIDER research work on taxes (and social protection) Tax-benet microsimulation Tax analysis using register data Conclusion 4 / 29

5 Outline Introduction Modern public economics approach to tax analysis Taxes in developing countries UNU-WIDER research work on taxes (and social protection) Tax-benet microsimulation Tax analysis using register data Conclusion 5 / 29

6 On optimal taxation The basic goal of the government: collect revenues to nance public goods by minimizing the tax distortions achieving the distributional goals of the government This gives rise to the basic eciency-equity tradeo there would be no need for distortive taxes if there were no distributional goals in such circumstances, lump-sum taxes would be sucient No artical distinction between taxes and transfer (negative taxes): a system Note that taxes can also have corrective role (to achieve the rst best) 6 / 29

7 Empirical work on the impacts of taxation Current research (For a survey, see Saez, Slemrod, and Giertz 2012) emphasizes the notion of the elasticity of taxable income, dened as dti d(1 t) (1 t) TI where TI denotes taxable income and t is the marginal tax rate In addition to labor supply, it also captures other margins of response, such as eort and avoidance It can be shown that ETI is a sucient statistic of the deadweight loss caused by taxation For developed countries, acceptable ETI estimates combine large taxpayer register data credible identication of the causal impact (tax reforms create quasi-experimental variation in tax treatment) 7 / 29

8 Outline Introduction Modern public economics approach to tax analysis Taxes in developing countries UNU-WIDER research work on taxes (and social protection) Tax-benet microsimulation Tax analysis using register data Conclusion 8 / 29

9 Tax take is lower Total tax revenues incl. SSC Calendar year (nearest) Low income Upper middle income OECD high income Lower middle income Non-OECD high income Source: Own calculations based on the GRD of the ICTD (see Prichard, Cobham, and Goodall (2014) 9 / 29

10 Rich countries' tax collection was also small (Source: Besley and Persson 2013) 10 / 29

11 Some observations Perhaps the main question in taxation and development: How do countries go from raising 10% of GDP to around 40% of GDP as tax revenues? In other words, how to obtain more tax capacity An interesting contrast: in poor countries the emphasis is on how to raise tax rates, but in developed countries the worry is in the eciency cost taxes create How countries should design their tax and transfer system as a whole / system? 11 / 29

12 A remark on current academic research on tax and development Of course, there is a long tradition and a vast amount of research on taxes in developing countries, for surveys see Keen (2012) and Besley and Persson (2013) An interesting new development is the emergence of eld experiments researchers work together with tax authorities to develop interventions, often targeted to reduce evasion/avoidance randomly selected taxpayers face the intervention, others left to the control group This literature (surveyed by Mascagni 2014) is no doubt very policy relevant and exciting 12 / 29

13 Outline Introduction Modern public economics approach to tax analysis Taxes in developing countries UNU-WIDER research work on taxes (and social protection) Tax-benet microsimulation Tax analysis using register data Conclusion 13 / 29

14 Four main themes The need to see the tax and social protection arrangements as a system: building tax-benet microsimulation models We currently know too little on the actual impacts of taxes on economic behavior (or ETI): try to use tax authorities' register data for tax analysis Improving the data access also on macro tax variables: maintaining and conducting research using the GRD of the ICTD How can countries overcome some of the political hurdles in building up tax capacity and and social protection systems: a symposium on the political economy aspects, to be held early next year 14 / 29

15 In this talk: The need to see the tax and social protection arrangements as a system: building tax-benet microsimulation models We currently know too little on the actual impacts of taxes on economic behavior (or ETI): try to use tax authorities register data for tax analysis 15 / 29

16 But work will also be conducted on Theoretical aspects of designing cash transfer system based on the optimal tax approach (or in the Mirrlees (1971) tradition) International tax aspects clearly a recognition that tax avoidance by the multinationals is a problem not yet hard evidence on to what extent prot shifting by the multinationals reacts to relative changes in the tax dierential between parent and (developing host) country papers addressing this subject are in preparation 16 / 29

17 What is tax-benet microsimulation? Models that combine representative household-level data on incomes and expenditures detailed coding of the tax and benet legislation in a country Simulation because the user can design alternative social policies or tax structures and examine their impacts on income distribution government budget work incentives etc. These models in everyday use in developed countries by ministries but also by researchers econometric tax analysis often requires information on (eective) marginal tax rates. These can be obtained from a microsimulation model 17 / 29

18 Why essential for developing countries? Many developing countries are (or should be) moving away from scattered individual social protection programs to actual systems of protection At the same time, they need to nd ways to increase tax revenues Microsimulation is a tool that is needed to understand the economy-wide impacts of such changes Yet, very few developing countries have access to such models 18 / 29

19 Our work on microsimulation Joint undertaking with the EUROMOD team at the University of Essex and Saspri (Cape Town) EUROMOD both a microsimulation model for European countries a software for building new microsimulation models Saspri has succesfully built models for Namibia and South Africa based on the Euromod structure The three institutions together will build models for selected developing countries, educate local stakeholders to use the models, and prepare research papers utilising the new tools 19 / 29

20 Tax authorities' data in developing countries In principle, such information has been collected in developing countries But extremely little research work that would use such data has been conducted Notable exceptions include Kleven and Waseem (2013) and Best, Brockmeyer, Kleven, Spinnewijn, and Waseem (2015) for Pakistan Evidence on tax responsiveness urgently needed from elsewhere 20 / 29

21 Our experience from South Africa UNU-WIDER has been engaged in a data compiling project where all register data of the revenue service (SARS) and statistical oce are at researchers' disposal at the National Treasury A main part of the dataset is all the rm-level register data from SARS since 2009 Papers which use these data for the rst time for research purposes have been commissioned by us New calls for papers are both open at the moment and forthcoming 21 / 29

22 An example of this research In a joint project with Wian Boonzaaier, Jarkko Harju and Tuomas Matikka, we examine the rm-level responses to the discontinuities in the tax code These discontinuities include rms whose turnover does not exceed 1 million Rand do not need to register for the VAT (a notch) the corporate income tax is linearly progressive for small and medium sized enterprises (the MTR is rst 0%, then 7%, then 21% and nally the standard rate of 28%) (kink points) Firms have an incentive to locate just below these kink or notch points It can be shown (see Saez 2010 and related work) that the stronger the response, the greater is the elasticity of the income tax base. So with this method, one can estimate the distortions created by the corporate tax system 22 / 29

23 Bunching at a lower kink of the CIT schedule, Lower kink, Excess mass: (.559), Elasticity:.238 (.089) Frequency Distance from the kink (1000 rand) Observed Counterfactual 23 / 29

24 Bunching at the VAT notch, VAT notch Excess bunching:.505 (.095), Elasticity:.123 (0) Upper limit: 14 (1.082) Frequency Distance from the kink Observed Counterfactual 24 / 29

25 Some comments on these results The results indicate that rms react quite strongly to the incentives created by the tax system The elasticities are estimated locally and that is one reason for why they vary Ongoing work is the response real (are rms truly smaller) or do they for instance misreport earnings? does the system have implications on the growth of small enterprises? 25 / 29

26 Outline Introduction Modern public economics approach to tax analysis Taxes in developing countries UNU-WIDER research work on taxes (and social protection) Tax-benet microsimulation Tax analysis using register data Conclusion 26 / 29

27 Some concluding thoughts While we will be happy to also participate in eld experiments, much of our work on taxes and development will focus on understanding the overall (distributional) impacts of the tax and benet systems and designing improvements building evidence base on the impacts of taxes using register data not only from S-A but also from other countries This type of work is, I would argue, an essential part of tax research that is needed to support countries in their eors of building up equitable and ecient tax systems 27 / 29

28 References I Besley, T., and T. Persson (2013): Chapter 2 - Taxation and Development, in handbook of public economics, vol. 5, ed. by M. F. Alan J. Auerbach, Raj Chetty, and E. Saez, vol. 5 of Handbook of Public Economics, pp Elsevier. Best, M., A. Brockmeyer, H. Kleven, J. Spinnewijn, and M. Waseem (2015): Production vs Revenue Eciency With Limited Tax Capacity: Theory and Evidence From Pakistan, Journal of Political Economy, Forthcoming. Keen, M. (2012): Taxation and Development: Again, IMF Working Papers 12/220, International Monetary Fund. Kleven, H. J., and M. Waseem (2013): Using Notches to Uncover Optimization Frictions and Structural Elasticities: Theory and Evidence from Pakistan, Quarterly Journal of Economics, 128(2), / 29

29 References II Mascagni, G. (2014): A Review of Tax Experiments: from the Lab to the Field, Evidence report 97, Institute for Development Studies. Mirrlees, J. A. (1971): An Exploration in the Theory of Optimum Income Taxation, Review of Economic Studies, 38(114), Prichard, W., A. Cobham, and A. Goodall (2014): The ICTD Government Revenue Dataset, Working Paper 19, International Centre for Taxation and Development. Saez, E. (2010): Do Taxpayers Bunch at Kink Points?, American Economic Journal: Economic Policy, 2(3), Saez, E., J. Slemrod, and S. H. Giertz (2012): The Elasticity of Taxable Income with Respect to Marginal Tax Rates: A Critical Review, Journal of Economic Literature, 50(1), / 29