Acemoglu, D., S. Johnson, J. Robinson, and P. Yared, 2008, “Income and Democracy,” American Economic Review, Vol. 98, pp. 808–842.
Baunsgaard, T., and M. Keen, 2010, “Tax Revenue and (or?) Trade Liberalization,” Journal of Public Economics, Vol. 94, pp. 563–577.
Belinga, V., D. Benedek, R. De Mooji, and J. Norregaard, 2014, “Tax Buoyancy in OECD Countries,” IMF Working Paper, No. 14/110 (Washington: International Monetary Fund).
Brückner, M., 2012, “An Instrumental Variables Approach to Estimating Tax Revenue Elasticities: Evidence from Sub-Saharan Africa,” Journal of Development Economics, Vol. 98, pp. 220–227.
Engle, R., and C. Granger, 1987, “Cointegration and Error Correction: Representation, Estimation, and Testing,” Econometrica, Vol. 55, pp. 251–276.
Fenochietto, R., and C. Pessino, 2013, “Understanding Countries’ Tax Effort,” IMF Working Paper, No. 13/244 (Washington: International Monetary Fund).
Im, K., M. Pesaran, and Y. Shin, 2003, “Testing for Unit Roots in Heterogeneous Panels,” Journal of Econometrics, Vol. 115, pp. 53–74.
International Monetary Fund, 2011, “Revenue Mobilization in Developing Countries,” IMF Policy Paper (Washington: International Monetary Fund).
International Monetary Fund, 2015, “Options for Low Income Countries’ Effective and Efficient Use of Tax Incentives for Investment,” IMF Background Paper (Washington: International Monetary Fund).
Johansen, S., and K. Juselius, 1990, “Maximum Likelihood Estimation and Inference on Cointegration with Application to the Demand for Money,” Oxford Bulletin of Economics and Statistics, Vol. 54, pp. 169–210.
Miguel, E., S. Satyanath, and S. Sergenti, 2004, “Economic Shocks and Civil Conflict: An Instrumental Variables Approach,” Journal of Political Economy, Vol. 112, pp. 725–753.
Talvi, E., and C. Vegh, 2005, “Tax Base Variability and Procyclical Fiscal Policy in Developing Countries,” Journal of Development Economics, Vol. 78, pp. 156–190.
The author would like to thank Harald Finger, Mario Mansour, Tigran Poghosyan, Alexander Tieman, participants at a seminar organized by the Fiscal Affairs Department of the International Monetary Fund, and representatives of the Federal Board of Revenue and the Ministry of Finance for insightful comments and suggestions, and Tasneem Alam and Hiba Zaidi for assistance in data collection. An earlier version of this paper was issued as a Selected Issues Paper and served as background material for the Executive Board Meeting on the 2015 Article IV Consultation for Pakistan.
The tax-to-GDP ratio is estimated to increase further by about 1.2 percent to 12.2 percent in 2016. Nevertheless, general government debt is still about 600 percent of tax revenue, and development spending remains significantly less than interest payments on government debt.
Expenditure rationalization aiming to change the composition in favor of growth-enhancing social and infrastructure spending is also critical, but beyond the scope of this paper.
To put it in a broader context, there are 15.6 million broadband internet subscribers and over 40 million individual bank accounts in Pakistan. From a cross-country perspective, the share of population filing for income tax in Pakistan is a mere 0.5 percent, compared to over 5 percent in India and 90 percent in Canada.
Services make up over 52 percent of GDP, while industry and agriculture account for about 22 percent and 25 percent, respectively.
A part of the increase in GST exemptions in FY2013/14 is a result of improvements in measurement.
The numerator reflects both tax policy and administration efforts, and therefore it may be difficult to interpret this concept exclusively for policymaking purposes.
The correlation between tax revenue growth and trend output growth is estimated to be 0.53 for developing countries and 0.38 for advanced economies.
Alternative measures of tax efficiency, such as the ratio of the change in tax revenues to the change in tax rates, confirm these trends.
While it is acceptable to utilize GDP as the tax base in estimating the elasticity of total tax revenue, it may cause ambiguity in estimations at a disaggregated level. For example, it is more appropriate to use corporate profits and household income instead of GDP as the tax base in estimating the elasticity rates of CIT and PIT, respectively. Due to data constraints, I use GDP as the tax base for all three subcategories of tax revenue.
For robustness check, I also estimate the models using the IV-Generalized Method of Moments (GMM) and reach similar results.
To be valid, the proposed IV must be strongly correlated with the endogenous variable (GDP in this case), but exogenous and unrelated with the error term in the regression equation.
Pakistan’s international trade (exports plus imports) amounts to only about 20 percent of GDP, allowing its trading partners’ average GDP per capita to function as an appropriate IV for the country’s own income per capita in the context of estimating tax revenue elasticity.
The unit root test results are available upon request.
This could be partly due to the fact that total tax revenues include other taxes while the weighted average tax rate is calculated by using only CIT, PIT and GST rates.
According to Johansen and Juselius (1990), the existence of a cointegrated linear combination of a set of timeseries variables indicates a long-term relationship among the variables.
To account for possible heteroskedasticity, robust standard errors are clustered at the country level.
I apply the Im-Pesaran-Shin (2003) procedure to conduct a panel unit root test. As expected, both GDP and tax series are highly persistent, and I cannot reject the hypothesis of a unit root in levels. Upon first differencing, however, the variables are found to be stationary. The panel unit root test results are available upon request.
The IV-GMM approach yields similar results, confirming the robustness of the baseline results presented in Table 5.
This assumes an average real GDP growth rate of 5 percent and uses the estimated tax revenue elasticity with respect to GDP.
The NTN system covers 3.6 million individuals (or less than 2 percent of population) compared to about 150 million people (or about 80 percent of population) covered in the CNIC database.
In “benami” transactions, assets are held by or transferred to a person, but have been provided for, or paid by, another person. The government has already submitted a draft bill to the National Assembly to disallow such transactions.