یک رویکرد محاسباتی عمومی تعادل برای تجزیه و تحلیل بخشی برای مالیات بالقوه: برنامه ای برای پاکستان

This study develops a dynamic general equilibrium model, applied to Pakistani data, in which optimizing agents evade taxes by operating in the underground economy. The cost to firms of evading taxes is that they find themselves subject to credit rationing from banks. Our model simulations show that in the absence of budgetary flexibility to adjust expenditures, raising tax rates too high drives firms into the underground economy, thereby reducing the tax base. Aggregate investment in the economy is lowered because of credit rationing. Taxes that are too low eliminate the underground economy, but result in unsustainable budget and trade deficits. Thus, the optimal rate of taxation, from a macroeconomic point of view, may lead to some underground activity. We note, in particular, that incorporating a VAT without any other tax reductions greatly reduces the tax compliance of the service sector. We have applied our model to Pakistan, and have calibrated our model to an 8 year period from 2004 to 2011. We note that it gives a reasonable approximation of Pakistani macro data. We then use a sectoral breakdown of tax data generated by the model to estimate tax gaps on a sector by sector basis. We note that certain sectors are currently paying taxes below their potential, while others may be above their tax potential. These sectoral gap estimates may be used as indicators of where greater tax enforcement efforts should be directed.

In many developing and transition countries, economic activity in the underground economy is estimated in excess of 40% of GDP (Friedman et al., 2000 and Schneider and Enste, 2000).1 This diversion into unofficial activity undermines the tax base and can significantly affect public finances and the quality of public administration (Dessy and Pallage, 2003, Johnson et al., 1997 and Loayza, 1996). The illegal nature of underground activity also constrains private investment and growth. One important cost imposed by the inability to enforce legal contracts is the limited access to formal credit markets. We develop a simple intertemporal general equilibrium model with heterogeneous agents, multiple production activities and credit rationing to explain the prevalence of a large underground economy and corresponding gap between potential and actual taxes collected. Our model is then applied to Pakistan. In particular, we explore the link between tax rates, access to credit, and the extent of tax evasion, and examine the consequences of the underground economy for public finances and aggregate economic performance.2 Entry and exit into the underground economy is derived as part of optimizing behavior that depends on taxes and interest rates. Firms operating in the underground economy are subject to credit rationing by banks, which reduces loans in relation to the firm's nonpayment of taxes. Since the size of the underground economy in the paper depends upon both endogenous and exogenous variables, our framework has scope for analyzing policy changes. In particular, we address the issue of policy responses in response to tax-avoiding activity, and we emphasize the ambiguous effects of taxation by means of numerical simulations of a computational general equilibrium (CGE) model for Pakistan. Economic reform will depend upon policies that reduce the various forms of tax evasion. Pakistan is a useful example of a country that has severe problems with tax evasion. In particular, apart from having a poor tax collection effort, Pakistan has had a steadily declining tax to GDP ratio. In recent years, tax collection has come down from 10.9% of GDP in 2003 to around 9% of GDP in 2012 (Cyan & Martinez-Vazquez, in press). Widespread evasion is reported; for example, only 2% of working age population pays personal income taxes (Kleven & Waseem, 2013). Additionally, a study in 2011 found that out of the 46,000 registered firms, only around 24,000 filed corporate income tax returns.3 Thus Pakistan provides a case study of a country in which there are both high degrees of tax evasion, as well as a steadily declining tax effort. As we shall see in Section 4, certain available Pakistani data is quite old.4 Hence our analysis should be viewed as both an example of how this modeling approach can be applied to any country suffering from tax evasion. Additionally the study provides certain useful insights on Pakistan despite the age of some of the underlying data. We would thus claim that our results, in particular the tax gap analysis, offers qualitative insights for Pakistan, although we would be hesitant to claim that the simulation outcomes represent “correct” absolute values. Section 2 provides a brief overview of our modeling of the underground economy. Section 3 presents our dynamic CGE model. Section 4 discusses the parameterization of the model and presents an initial calibration exercise. Section 5 discusses extensions to further disaggregation of the tax gap analysis.

We have constructed a dynamic general equilibrium model that incorporates endogenous tax evasion as part of intertemporal optimizing behavior by firms. We have used parameters from Pakistan to calibrate our model to an 8 year period from 2004 to 2011, and note that it gives a reasonable approximation of Pakistani macro data. We note that the service sector, at equilibrium, consistently evades some portion of its taxes. We then incorporate a value added tax of 10% in a series of counterfactual simulations, as the introduction of such a tax is currently under consideration by Pakistan. We note that incorporating a VAT without any other tax reductions greatly reduces the tax compliance of the service sector. If, on the other hand, the VAT introduction is accompanied by abolishing the GST, then tax compliance rises, but revenues from the non-service sectors decline, interest rates rise, and real GDP growth slows. We also use the model to carry out a sectoral estimate of the tax gap.16 Here the predicted outcomes of the general equilibrium model for 2010, assuming full compliance, are compared with actual tax revenues collected. That is, we set the “honesty” parameter for each sector at 0. The general equilibrium model then generates a path for tax collections for the 8 years of the simulation, and we choose the predicted collections for 2010. These are then compared with actual tax collections for selected sectors, as well as for the aggregate economy and the manufacturing sector. They indicate that, on the level of the overall economy, there is a tax gap of about 58%, while in the manufacturing sector the gap is approximately 53%. As might be expected from the general equilibrium model, capital intensive sectors such as iron and steel, and oil and gas, have smaller gaps than do less capital intensive sectors such as finance and insurance, or hotels and restaurants. These calculations should help in the measurement of the overall problem, as well as to identify those sectors where improvement is most needed. Our approach demonstrates a method for gap analysis which can be further developed to incorporate additional information on sectoral structures and use sectoral estimates of honesty parameters.