## _wp16182

## Source details

**Canonical URL:** [_wp16182](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16182.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16182.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16182.pdf.json)

---

### I. Revenue gap and potential
- Pakistan’s tax-to-GDP ratio: 12.4 percent in 2016.
- Tax-to-GDP ratio increased by over 60 percent in nominal terms (or 2.4 percentage points of GDP) over the past three years to 12.4 percent in 2016.
- Estimated tax capacity: 22.3 percent of GDP (Fenochietto and Pessino (2013) estimate).
- Tax revenue gap (difference between actual and potential revenue): about 10 percent of GDP as of 2016.
- Tax effort: improved from 0.43 in 2011 to 0.56 in 2016; comparator average: 0.64; high-income countries: 0.76.
- Medium-term realistic tax ratio by bringing tax effort toward average developing countries: 14-15 percent of GDP.
- Empirical elasticity findings (1960–2015): short-run and long-run elasticities of tax revenue are slightly above 1 over the sample period.
- Disaggregated elasticities: corporate income tax (CIT) and general sales tax (GST) show elasticity coefficients well above 1; personal income tax (PIT) appears significantly less elastic (partly reflecting reliance on withholding taxes treated as final tax).

### II. Tax regime structure and administrative fragmentation
- Federal government collects over 92 percent of total tax revenues; provincial governments’ own revenues contribute 7.7 percent.
- Provincial tax revenues increased by about 0.3 percentage points over the past three years to 1 percent of GDP in 2016.
- Constitutional assignment: provinces responsible for taxation of agriculture, services and immovable property, but provinces show inadequate administrative capacity and limited incentive to boost own revenues.
- High informality, gaps in information gathering and sharing, bureaucratic red tape and corruption undermine tax morale and compliance.

### III. Composition of tax revenues and taxpayer coverage
- Indirect taxes account for about two-thirds (~65 percent of federal tax revenue) of total tax revenue.
- Direct taxes share increase: from an average of 18.5 percent in the first half of the 1990s to 29 percent in 2000 and about 38 percent in recent years; nevertheless, income taxes remain about 4 percent of GDP.
- Statutory rates: CIT and PIT declined from 45 percent in 1990 to 31 percent and 20 percent, respectively, by 2016.
- PIT registration and filing:
  - Registered PIT payers increased from 752,695 in 2000 to over 3.6 million in 2014.
  - Active PIT filers: 1,000,790 (up by 27 percent over the past two years).
  - Employed persons: 56.5 million (32 million employed outside agriculture).
  - Reportedly 5.7 million people earning above the income tax threshold set at PRs 400,000.
- CIT registration and filing:
  - Active CIT filers: 27,334 out of more than 60,000 companies registered for CIT.
  - Active CIT filers represent 0.8 percent of the number of commercial/industrial electricity users (an illustrative pool of potential taxpayers).
- GST registration:
  - Registered GST entities: 178,190 out of about 1.4 million retailers and 3.4 million commercial and industrial electricity users.
- Cross-country filing context: share of population filing for income tax in Pakistan is 0.5 percent (comparative figures cited in source).
- NTN coverage: 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.

### IV. Agriculture, property, services and provincial taxation
- Agriculture’s economic weight and taxation:
  - Agriculture accounts for about 20 percent of GDP and employs 45 percent of the workforce.
  - Agricultural income subject to a minimum tax based on land holdings; farmers with land less than 12½ acres are exempt.
  - Over 90 percent of farmers appear to have land holdings less than 12½ acres; agricultural income largely untaxed.
  - Annual agricultural tax revenue: PRs 270 per acre (or less than US$3) per acre on average.
  - Agricultural tax revenue under provincial purview: less than 0.1 percent of total tax revenues.
- Property and services taxation under provinces generate a mere 0.04 percent and 0.6 percent of GDP, respectively.
- Services share of GDP: over 52 percent; industry: about 22 percent; agriculture: about 25 percent.

### V. Tax expenditures and concessions
- Cost of tax expenditures:
  - Increased from 0.2 percent of GDP (about 2 percent of tax revenues) in 2000 to a peak of 1.9 percent of GDP (almost 20 percent of tax revenues) in 2014.
  - Government eliminated tax concessions and exemptions amounting to about 0.6 percent of GDP since 2014.
- Many concessions granted via Statutory Regulatory Orders (SROs) with limited transparency and parliamentary oversight; legislation passed to limit concessional SRO authorization temporarily in exceptional circumstances.
- Reported total tax expenditures (values in billion rupees) for 2010/11–2015/16: 166.7, 205.9, 239.5, 477.1, 411.4, 394.6 respectively.

### VI. Tax buoyancy and revenue volatility
- Tax buoyancy concept: change in tax revenue (real) divided by change in tax base (real); buoyancy of 1 implies neutral impact on tax-to-GDP ratio.
- Pakistan’s tax revenue buoyancy has exceeded unity in recent years but remains highly volatile.
- Tax buoyancy has exhibited a boom-bust pattern with high volatility reflecting tax policy changes, administrative challenges, political crises and economic difficulties on taxpayer behavior.
- In recent years, tax buoyancy has improved to well above unity from an average of 0.2 in the 1990s.
- The five-year moving average of aggregate tax buoyancy remains below the recent peak.
- Disaggregated dynamics:
  - GST buoyancy has recovered in recent years with the elimination of exemptions, but remains significantly below the peak when provinces agreed to the collection of GST on services by the FBR.
  - Buoyancy of direct taxes shows no sustained improvement, except over the past few years with the elimination of concessions and exemptions.
  - Buoyancy of excises (generally structured as non-ad valorem) is a significant cause of limited progress, holding back overall tax buoyancy because of loss of revenue caused by trade liberalization.
- Cross-country context:
  - Developing countries tend to struggle to mobilize domestic tax revenues to compensate for revenue losses caused by trade liberalization and tariff rationalization.
  - 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.

### VII. Tax efficiency and comparative performance
- Tax efficiency measure: tax revenue as a percentage of GDP, divided by the standard tax rate.
- GST efficiency:
  - Improved from an average of 0.11 in the 1990s to an average of 0.23 over the past three years.
  - Pakistan’s GST efficiency remains significantly below the unweighted average of 0.28 in Africa and 0.44 in Asia Pacific.
  - Low GST efficiency partly reflects pervasive use of exemptions and limited progress in bringing retailers into the tax base.
- CIT and PIT productivity:
  - Productivity rates of the CIT and PIT regimes show sustained improvement over the past decade with greater use of withholding taxes.
  - Pakistan’s performance in CIT and PIT remains below those of other developing countries due to overly generous tax concessions and exemptions, difficulties in tax administration, and low taxpayer compliance through informal activity and underreporting.

### VIII. Empirical approach, elasticity estimates and robustness
- Data and methodology:
  - Annual data from 1960 to 2015 used.
  - Regression estimated in first differences because GDP and tax series contain unit roots and become stationary upon first differencing.
  - Specification includes a weighted average of tax rates to control for tax policy changes over time.
  - Instrumental variable (IV) approach: IV via two-stage least squares (2SLS); IV-GMM used for robustness with similar results.
  - Considered instruments: rainfall shocks, international commodity prices (real terms), and trade-weighted average real GDP of Pakistan’s trading partners.
- IV-2SLS short-run elasticity estimates (first differences, robust standard errors in parentheses):
  - Total: ∆GDP 1.159*** (0.191); ∆Tax Rate -0.005 (0.005); Number of observations 55; F-stat 23.08; R2 0.50.
  - CIT: ∆GDP 1.372*** (0.501); ∆Tax Rate -0.006 (0.008); Number of observations 55; F-stat 5.31; R2 0.22.
  - PIT: ∆GDP 0.955* (0.558); ∆Tax Rate -0.002 (0.006); Number of observations 55; F-stat 12.06; R2 0.10.
  - GST: ∆GDP 1.095** (0.508); ∆Tax Rate 0.001 (0.010); Number of observations 55; F-stat 2.57; R2 0.13.
- Short-run interpretations:
  - Total short-run elasticity: 1.16 (statistically significant at the 1 percent level).
  - CIT short-run elasticity: 1.37.
  - PIT short-run elasticity: 0.96 (caveat: withholding taxes treated as final tax often not reflected in PIT returns).
  - GST short-run elasticity: 1.1.
- Error Correction Model (ECM) long-run elasticities and dynamics (Engle and Granger two-step, 1960–2015):
  - Long-run elasticities (Table 4):
    - Total GDP 1.260*** (0.036); Tax Rate 0.028*** (0.004).
    - CIT GDP 1.922*** (0.166); Tax Rate 0.009 (0.014).
    - PIT GDP 0.511*** (0.113); Tax Rate -0.018** (0.007).
    - GST GDP 1.470*** (0.068); Tax Rate -0.041 (0.049).
  - Short-run ECM coefficients (Table 4):
    - Total ∆GDP 0.718 (0.484); ∆Tax Rate 0.004 (0.005); ECT -0.196* (0.078).
    - CIT ∆GDP 0.514 (1.106); ∆Tax Rate -0.003 (0.009); ECT -0.083 (0.063).
    - PIT ∆GDP -1.524 (1.326); ∆Tax Rate -0.001 (0.006); ECT -0.086 (0.073).
    - GST ∆GDP 1.695* (1.005); ∆Tax Rate 0.025 (0.018); ECT -0.121* (0.053).
  - Long-run elasticity interpretations:
    - Aggregate long-run tax revenue elasticity with respect to GDP = 1.26.
    - Disaggregated long-run elasticities: CIT 1.92; PIT 0.51; GST 1.47.
  - Error-correction term (ECT) indicates low speed of adjustment: only about 8-20 percent of disequilibrium is corrected in a given year.
- Cross-country panel IV-2SLS (124 developing countries, 1980–2014) results:
  - Total: ∆GDP 1.892*** (0.125); ∆Tax Rate 0.001 (0.003); Number of observations 6891; Number of groups 62; R2 0.31.
  - CIT: ∆GDP 1.257*** (0.213); ∆Tax Rate 0.004 (0.004); Number of observations 11019; Number of groups 70; R2 0.10.
  - PIT: ∆GDP 1.836* (0.270); ∆Tax Rate 0.002 (0.003); Number of observations 2513; Number of groups 68; R2 0.12.
  - VAT: ∆GDP 1.724** (0.129); ∆Tax Rate 0.009* (0.005); Number of observations 1303; Number of groups 92; R2 0.17.
- Cross-country comparisons:
  - Estimated tax revenue elasticity across developing countries = 1.89 with respect to real GDP (instrumented).
  - Pakistan’s tax revenue elasticity is about 40 percent below the average of developing countries.
  - Disaggregated comparison: Pakistan’s CIT elasticity marginally higher than average (1.37 versus 1.26), while PIT and GST elasticities in Pakistan are significantly lower than in other emerging market economies.
- Robustness: IV-GMM approach yields similar results.

### IX. Policy implications and recommendations
- Overarching objective: unlock tax revenue potential via comprehensive reforms at federal and provincial levels to broaden tax bases, strengthen revenue administration and taxpayer compliance, eliminate distortionary concessions, and rationalize tax policy to be efficient and equitable.
- Administrative and institutional reforms:
  - Reorganize tax administrations along functional lines; integrate databases and information technology.
  - Require a tax identity number in all financial and immovable property transactions.
  - Fully implement a risk-based auditing system focusing on taxpayer noncompliance risks.
  - Increase tax fraud penalties and make tax evasion a criminal offense.
  - Fight tax evasion by compiling a comprehensive list of high-wealth individuals and corporate entities they control and prohibit “benami” transactions.
  - Enhance information gathering and sharing across layers of government; build provincial administrative capacity and incentives for own-source revenue mobilization.
  - Ministry of Finance to establish a tax policy research and analysis unit—outside the FBR—to improve revenue forecasting and fiscal policymaking.
- Tax policy reforms:
  - PIT: reduce the tax exempt income threshold, widen tax brackets, adopt more progressive and lower tax rates, rationalize concessions and exemptions.
  - CIT: simplify the regime and reduce concessions and exemptions to allow lower rates while enhancing revenue yields and competitiveness.
  - Agriculture: introduce presumptive taxes on agricultural turnover and land-based tax rates adjusted according to productivity characteristics with thresholds to protect low-income farm households.
  - Property taxes: modernize recurrent property taxes via a central fiscal cadastre and central valuation agency using market-based valuation.
  - GST: integrate regime with a single statutory rate under one collection agent; eliminate GST exemptions, zero-ratings, and special schemes for greater efficiency.
  - Excises: change structure of federal and provincial excises to ad valorem rates in a unified manner for domestically produced and imported goods to maximize revenue yield and address negative externalities.
- Reform caveat: avoid increasing revenue by further taxing already compliant taxpayers, which would worsen inequalities, undermine tax morale and cause distortions in economic activity.

### X. Expected outcomes, projections and constraints
- Projection assumption referenced: assumes an average real GDP growth rate of 5 percent and uses the estimated tax revenue elasticity with respect to GDP.
- Without continued reform efforts, Pakistan cannot realistically raise its tax revenue to around 15 percent of GDP in the foreseeable future (projection assumes average real GDP growth rate of 5 percent and uses estimated elasticity).
- Comprehensive reform efforts already led to significant increase in revenue mobilization and improvements in tax buoyancy and elasticity over the past three years.
- Long-term potential: Pakistan has the potential to double its tax revenue-to-GDP ratio over the long term, conditional on concerted federal and provincial reforms.

*Source: _wp16182*

### References .............................................................................................................

### References

### I. Introduction — revenue gap and potential
- Pakistan’s tax-to-GDP ratio: 12.4 percent in 2016.
- Tax-to-GDP ratio increased by over 60 percent in nominal terms (or 2.4 percentage points of GDP) over the past three years to 12.4 percent in 2016.
- Estimated tax capacity: 22.3 percent of GDP (Fenochietto and Pessino (2013) estimate).
- Tax revenue gap (difference between actual and potential revenue): about 10 percent of GDP as of 2016.
- Tax effort: improved from 0.43 in 2011 to 0.56 in 2016; comparator average: 0.64; high-income countries: 0.76.
- Medium-term realistic tax ratio by bringing tax effort toward average developing countries: 14-15 percent of GDP.
- Empirical elasticity findings (1960–2015): short-run and long-run elasticities of tax revenue are slightly above 1 over the sample period.
- Disaggregated elasticities: corporate income tax (CIT) and general sales tax (GST) show elasticity coefficients well above 1; personal income tax (PIT) appears significantly less elastic (partly reflecting reliance on withholding taxes treated as final tax).

### II. Tax regime structure and administrative fragmentation
- Federal government collects over 92 percent of total tax revenues; provincial governments’ own revenues contribute 7.7 percent.
- Provincial tax revenues increased by about 0.3 percentage points over the past three years to 1 percent of GDP in 2016.
- Constitutional assignment: provinces responsible for taxation of agriculture, services and immovable property, but provinces show inadequate administrative capacity and limited incentive to boost own revenues.
- High informality, gaps in information gathering and sharing, bureaucratic red tape and corruption undermine tax morale and compliance.

### III. Composition of tax revenues and taxpayer coverage
- Indirect taxes account for about two-thirds (~65 percent of federal tax revenue) of total tax revenue.
- Direct taxes share increase: from an average of 18.5 percent in the first half of the 1990s to 29 percent in 2000 and about 38 percent in recent years; nevertheless, income taxes remain about 4 percent of GDP.
- Statutory rates: CIT and PIT declined from 45 percent in 1990 to 31 percent and 20 percent, respectively, by 2016.
- PIT registration and filing:
  - Registered PIT payers increased from 752,695 in 2000 to over 3.6 million in 2014.
  - Active PIT filers: 1,000,790 (up by 27 percent over the past two years).
  - Employed persons: 56.5 million (32 million employed outside agriculture).
  - Reportedly 5.7 million people earning above the income tax threshold set at PRs 400,000.
- CIT registration and filing:
  - Active CIT filers: 27,334 out of more than 60,000 companies registered for CIT.
  - Active CIT filers represent 0.8 percent of the number of commercial/industrial electricity users (an illustrative pool of potential taxpayers).
- GST registration:
  - Registered GST entities: 178,190 out of about 1.4 million retailers and 3.4 million commercial and industrial electricity users.
- Cross-country filing context: share of population filing for income tax in Pakistan is 0.5 percent (comparative figures cited in source).

### IV. Agriculture, property, services and provincial taxation
- Agriculture’s economic weight and taxation:
  - Agriculture accounts for about 20 percent of GDP and employs 45 percent of the workforce.
  - Agricultural income subject to a minimum tax based on land holdings; farmers with land less than 12½ acres are exempt.
  - Over 90 percent of farmers appear to have land holdings less than 12½ acres; agricultural income largely untaxed.
  - Annual agricultural tax revenue: PRs 270 per acre (or less than US$3) per acre on average.
  - Agricultural tax revenue under provincial purview: less than 0.1 percent of total tax revenues.
- Property and services taxation under provinces generate a mere 0.04 percent and 0.6 percent of GDP, respectively.
- Services share of GDP: over 52 percent; industry: about 22 percent; agriculture: about 25 percent.

### V. Tax expenditures and concessions
- Cost of tax expenditures:
  - Increased from 0.2 percent of GDP (about 2 percent of tax revenues) in 2000 to a peak of 1.9 percent of GDP (almost 20 percent of tax revenues) in 2014.
  - Government eliminated tax concessions and exemptions amounting to about 0.6 percent of GDP since 2014.
- Many concessions granted via Statutory Regulatory Orders (SROs) with limited transparency and parliamentary oversight; legislation passed to limit concessional SRO authorization temporarily in exceptional circumstances.
- Tax expenditures measurement table (values in billion rupees) reported for components including Income tax, GST, Customs duty and Total, with total tax expenditures reported as 166.7, 205.9, 239.5, 477.1, 411.4, 394.6 for years 2010/11–2015/16 respectively (table entries preserved as presented in source).

### VI. Tax buoyancy and revenue volatility
- Tax buoyancy concept: change in tax revenue (real) divided by change in tax base (real); buoyancy of 1 implies neutral impact on tax-to-GDP ratio.
- Pakistan’s tax revenue buoyancy has exceeded unity in recent years but remains highly volatile.
- Over last two and a half decades, Pakistan’s tax revenue-to-GDP ratio oscillated between a low of 9 percent and a peak (table/figure referenced for detailed trends).

### VII. Empirical approach and cross-country perspective
- Estimation methods used:
  - Instrumental variable (IV) approach to address potential endogeneity of tax revenue and economic growth.
  - Multivariate cointegration analysis to estimate an Error Correction Model (ECM).
  - Panel data analysis to compare Pakistan with developing countries.
- Key empirical findings:
  - Short-run and long-run elasticities slightly above 1 for aggregate tax revenue.
  - Significant variation across tax types: CIT and GST more responsive to GDP; PIT less elastic.
  - Pakistan’s tax revenue elasticity is well below the average of developing countries in cross-country panel analysis.

### VIII. Policy implications and recommendations
- Unlocking tax revenue potential requires comprehensive reforms at federal and provincial levels to:
  - Broaden tax bases.
  - Strengthen revenue administration and taxpayer compliance.
  - Eliminate distortionary and overgenerous tax concessions and exemptions.
  - Rationalize tax policy to be efficient and equitable.
  - Improve perceived fairness of the tax system (address corruption and weak auditing procedures).
- Reform caveat emphasized in source: avoid increasing revenue by further taxing already compliant taxpayers, which would worsen inequalities, undermine tax morale and cause distortions in economic activity.
- Additional institutional measures highlighted: enhance information gathering and sharing across layers of government; build provincial administrative capacity and incentives for own-source revenue mobilization.

*Source: _wp16182 - References .........................................................................................................*

### 12.4 percent. Tax buoyancy, on the other hand, has exhibited a boom-bust pattern with high

### _wp16182 - 12.4 percent. Tax buoyancy, on the other hand, has exhibited a boom-bust pattern with high

### Tax buoyancy: patterns and drivers
- Tax buoyancy has exhibited a boom-bust pattern with high volatility reflecting tax policy changes, administrative challenges, political crises and economic difficulties on taxpayer behavior.
- In recent years, tax buoyancy has improved to well above unity from an average of 0.2 in the 1990s.
- The five-year moving average of aggregate tax buoyancy remains below the recent peak.
- Disaggregated federal-level features:
  - Limited synchronization of buoyancy across subcategories of tax revenue.
  - GST buoyancy has recovered in recent years with the elimination of exemptions, but remains significantly below the peak when provinces agreed to the collection of GST on services by the FBR.
  - Buoyancy of direct taxes shows no sustained improvement, except over the past few years with the elimination of concessions and exemptions.
  - Buoyancy of excises (generally structured as non-ad valorem) is a significant cause of limited progress, holding back overall tax buoyancy because of loss of revenue caused by trade liberalization.

- Cross-country context:
  - Developing countries tend to struggle to mobilize domestic tax revenues to compensate for revenue losses caused by trade liberalization and tariff rationalization.

- Correlations and caveats:
  - 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.
  - The numerator in buoyancy measures reflects both tax policy and administration efforts, complicating exclusive policy interpretation.

### Tax efficiency: measures and comparative performance
- Tax efficiency is measured as tax revenue as a percentage of GDP, divided by the standard tax rate.
- GST efficiency:
  - Improved from an average of 0.11 in the 1990s to an average of 0.23 over the past three years.
  - The rise was partly due to the agreement between the provinces and the FBR to collect GST on services on their behalf during 2000-2010.
  - Pakistan’s GST efficiency remains significantly below the unweighted average of 0.28 in Africa and 0.44 in Asia Pacific.
  - Low GST efficiency partly reflects pervasive use of exemptions and limited progress in bringing retailers into the tax base.
- CIT and PIT productivity:
  - Productivity rates of the CIT and PIT regimes show sustained improvement over the past decade with greater use of withholding taxes.
  - Nonetheless, Pakistan’s performance in CIT and PIT remains below those of other developing countries due to overly generous tax concessions and exemptions, difficulties in tax administration, and low taxpayer compliance through informal activity and underreporting.

### Estimating tax revenue elasticity: definition and methodology
- Tax revenue elasticity defined as the percent change in tax revenue divided by the percent change in the tax base, controlled for tax policy changes.
- Difference from buoyancy: elasticity accounts for discretionary changes in tax policy.
- Data and specification:
  - Annual data from 1960 to 2015 used.
  - Regression estimated in first differences because GDP and tax series contain unit roots and become stationary upon first differencing.
  - Specification includes a weighted average of tax rates to control for tax policy changes over time.
  - Note: Using GDP as the tax base for disaggregated taxes (CIT, PIT) is a constraint due to data limitations.

- Instrumental variables (IV) strategy to address endogeneity and measurement error:
  - IV via two-stage least squares (2SLS) employed; IV-GMM used for robustness with similar results.
  - Considered instruments: rainfall shocks, international commodity prices (real terms), and trade-weighted average real GDP of Pakistan’s trading partners.
  - Trade-weighted average real GDP of trading partners used as an instrument for Pakistan’s real GDP following Acemoglu et al. (2008), since it is correlated with Pakistan’s real GDP but uncontaminated by Pakistani tax policy changes.

- Data sources for variables and instruments:
  - Tax data: FBR and Ministry of Finance.
  - GDP and deflator: State Bank of Pakistan and World Bank’s World Development Indicators.
  - Rainfall: Climatic Research Unit, University of East Anglia.
  - International commodity prices: World Bank’s Global Economic Monitor Commodities database.
  - Trade-weighted partner GDP: IMF’s World Economic Outlook database.

### Empirical elasticity results and interpretation
- Short-run elasticity estimates (IV regression, first differences):
  - Short-run elasticity of total tax revenue: 1.16 (statistically significant at the 1 percent level).
    - Interpretation: total tax revenue increases by about 1.16 percent in response to a 1 percent increase in GDP, controlling for tax-rate changes.
  - Short-run elasticity of CIT revenue: 1.37.
    - Interpretation: CIT revenue increases by about 1.37 percent for a 1 percent increase in GDP.
  - Short-run elasticity of PIT revenue: 0.96.
    - Interpretation: PIT revenue increases by about 0.96 percent for a 1 percent increase in GDP.
    - Caveat: PIT estimates do not capture withholding taxes treated as final tax and often not reflected in PIT returns.
  - Short-run elasticity of GST revenue: 1.1.
    - Interpretation: GST revenue increases by around 1.1 percent for a 1 percent increase in GDP.
- Findings on tax-rate changes:
  - A change in tax rates appears to have no significant effect on tax revenue in the short run in the IV estimates.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16182.pdf*

### 1.1 percent for every 1 percent increase in GDP. The estimation results at a disaggregated level

### 1.1 percent for every 1 percent increase in GDP. The estimation results at a disaggregated level

### Key estimation findings — IV-2SLS (short-run / instantaneous)
- IV-2SLS estimates are larger than OLS coefficients across all three subcategories of tax revenue.
- Table 3 (IV-2SLS Estimates) reported coefficients (robust standard errors in parentheses):
  - Total: ∆GDP 1.159*** (0.191); ∆Tax Rate -0.005 (0.005); Number of observations 55; F-stat 23.08 [p-value] 0.000; R2 0.50.
  - CIT: ∆GDP 1.372*** (0.501); ∆Tax Rate -0.006 (0.008); Number of observations 55; F-stat 5.31 [p-value] 0.008; R2 0.22.
  - PIT: ∆GDP 0.955* (0.558); ∆Tax Rate -0.002 (0.006); Number of observations 55; F-stat 12.06 [p-value] 0.137; R2 0.10.
  - GST: ∆GDP 1.095** (0.508); ∆Tax Rate 0.001 (0.010); Number of observations 55; F-stat 2.57 [p-value] 0.086; R2 0.13.
- At a disaggregated level, tax rates have no statistically significant effect; coefficient on tax rate comes out negative in the case of PIT.

### Error Correction Model (ECM) — long-run elasticities and adjustment dynamics
- ECM methodology: two-step Engle and Granger (1987) procedure applied for period 1960-2015. Short-run coefficients ߚଵ denote short-run tax revenue elasticity with respect to GDP; ߚଷ represents speed of adjustment (error-correction term).
- Long-run equation coefficients (Table 4; standard errors in parentheses):
  - Total GDP 1.260*** (0.036); Tax Rate 0.028*** (0.004).
  - CIT GDP 1.922*** (0.166); Tax Rate 0.009 (0.014).
  - PIT GDP 0.511*** (0.113); Tax Rate -0.018** (0.007).
  - GST GDP 1.470*** (0.068); Tax Rate -0.041 (0.049).
- Short-run equation coefficients (Table 4; standard errors in parentheses):
  - Total ∆GDP 0.718 (0.484); ∆Tax Rate 0.004 (0.005); ECT -0.196* (0.078).
  - CIT ∆GDP 0.514 (1.106); ∆Tax Rate -0.003 (0.009); ECT -0.083 (0.063).
  - PIT ∆GDP -1.524 (1.326); ∆Tax Rate -0.001 (0.006); ECT -0.086 (0.073).
  - GST ∆GDP 1.695* (1.005); ∆Tax Rate 0.025 (0.018); ECT -0.121* (0.053).
- Long-run elasticities interpretation:
  - Aggregate long-run tax revenue elasticity with respect to GDP = 1.26 (statistically significant at the 1 percent level) — implies tax revenue increases by about 1.3 percent over the long run in response to a 1 percent increase in GDP.
  - Disaggregated long-run elasticities: CIT 1.92; PIT 0.51; GST 1.47.
    - These indicate CIT and PIT increase by about 1.9 percent and 0.5 percent respectively for every 1 percent increase in GDP; GST increases by around 1.5 percent.
- Role of tax rate in long run:
  - Aggregate tax rate coefficient 0.03 (0.028 reported in table) is economically meaningful and statistically significant in the long run, but caution advised because tax rate is a weighted average of CIT, PIT and GST rates excluding other taxes (customs duties and excises).
- Error-correction (speed of adjustment):
  - ECT varies between -0.08 (CIT) and -0.2 (aggregate), indicating low speed of adjustment.
  - Only about 8-20 percent of disequilibrium is corrected in a given year following deviation from long-run equilibrium.
  - ECT statistically significant at the 10 percent level only for aggregate tax revenue and GST at disaggregated level.

### Robustness check — cross-country panel perspective (IV-2SLS, 124 developing countries, 1980–2014)
- Panel specification: IV model of tax revenue elasticity for Total, CIT, PIT (VAT/GST in panel) using trade-weighted average real GDP of trading partners as instrument; fixed effects and clustered robust standard errors at country level.
- Table 5 (Cross-Country IV-2SLS Estimates; robust standard errors in parentheses):
  - Total: ∆GDP 1.892*** (0.125); ∆Tax Rate 0.001 (0.003); Fixed effects Yes; Number of observations 6891; Number of groups 62; F-stat 7.75 [p-value] 0.000; R2 0.31.
  - CIT: ∆GDP 1.257*** (0.213); ∆Tax Rate 0.004 (0.004); Number of observations 11019; Number of groups 70; F-stat 2.36 [p-value] 0.000; R2 0.10.
  - PIT: ∆GDP 1.836* (0.270); ∆Tax Rate 0.002 (0.003); Number of observations 2513; Number of groups 68; F-stat 3.17 [p-value] 0.000; R2 0.12.
  - VAT: ∆GDP 1.724** (0.129); ∆Tax Rate 0.009* (0.005); Number of observations 1303; Number of groups 92; F-stat 6.48 [p-value] 0.000; R2 0.17.
- Cross-country comparison conclusions:
  - Estimated tax revenue elasticity across developing countries = 1.89 with respect to real GDP (instrumented).
  - Pakistan’s tax revenue elasticity is about 40 percent below the average of developing countries.
  - Disaggregated comparison: Pakistan’s CIT elasticity marginally higher than average of other developing countries (1.37 versus 1.26), while PIT and GST elasticities in Pakistan are significantly lower than in other emerging market economies.
- Robustness: IV-GMM approach yields similar results, confirming robustness.

### Fiscal context and headline statistics for Pakistan
- Tax revenue-to-GDP ratio increased by 2.4 percentage points over the past three years to 12.4 percent of GDP in 2016.
- Pakistan’s tax effort improved from 0.43 in 2011 to 0.56 in 2016; comparator developing countries average 0.64; high-income countries average 0.76.
- Provincial own revenues contribute only about 8 percent of total tax revenues.
- Composition of tax revenues: indirect taxes account for about 63 percent of federal tax revenue.
- NTN coverage: 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.
- Projection assumption referenced: assumes an average real GDP growth rate of 5 percent and uses the estimated tax revenue elasticity with respect to GDP.

### Policy implications and recommendations — revenue mobilization and tax policy
- Strategic priorities:
  - Broadening tax bases, strengthening tax compliance across all sectors, eliminating distortionary tax concessions and exemptions, and addressing fragmented tax administrations at federal and provincial levels.
  - Strengthening tax administrations by reorganizing along functional lines, integrating databases and information technology, requiring a tax identity number in all financial and immovable property transactions.
  - Institutional reforms to reduce corruption and improve business climate to boost tax revenue collections.
- Specific administrative and compliance actions:
  - FBR and provincial revenue administrations should fully implement a risk-based auditing system focusing on taxpayer noncompliance risks, increase tax fraud penalties and make tax evasion a criminal offense.
  - Fight tax evasion focusing on a comprehensive list of high-wealth individuals and corporate entities they control and prohibit “benami” transactions.
  - Ministry of Finance to establish a tax policy research and analysis unit—outside the FBR—to improve revenue forecasting and fiscal policymaking.
- Tax policy reforms:
  - PIT: reduce the tax exempt income threshold, widen tax brackets, adopt more progressive and lower tax rates, rationalize concessions and exemptions.
  - CIT: simplify the regime and reduce concessions and exemptions to allow lower rates while enhancing revenue yields and competitiveness.
  - Agriculture: introduce presumptive taxes on agricultural turnover and land-based tax rates adjusted according to productivity characteristics with thresholds to protect low-income farm households.
  - Property taxes: modernize recurrent property taxes via a central fiscal cadastre and central valuation agency using market-based valuation.
  - GST: integrate regime with a single statutory rate under one collection agent; eliminate GST exemptions, zero-ratings, and special schemes for greater efficiency.
  - Excises: change structure of federal and provincial excises to ad valorem rates in a unified manner for domestically produced and imported goods to maximize revenue yield and address negative externalities.

### Expected outcomes and constraints
- Without continued reform efforts, Pakistan cannot realistically raise its tax revenue to around 15 percent of GDP in the foreseeable future (projection assumes average real GDP growth rate of 5 percent and uses estimated elasticity).
- Comprehensive reform efforts already led to significant increase in revenue mobilization and improvements in tax buoyancy and elasticity over the past three years.
- Long-term potential: Pakistan has the potential to double its tax revenue-to-GDP ratio over the long term, conditional on concerted federal and provincial reforms.

*Source: Author's calculations and analysis presented in the provided IMF working paper excerpt.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16182.pdf_