Introduction of Value Added Tax in Indian States: How Has the Revenue Performance Been?

Introduction of Value Added Tax in Indian States: How has the revenue performance been?

Mona Prasad[1]

January 2008

Summary

A majority of Indian statesreplaced the existing sales tax system with Value Added Taxes (VAT) with effect from April 2005. This policy paper analyzes the revenue performance of states which implemented VAT from April 2005 in comparison with those which implemented it later. It considers all the major general category states with the exception of Delhi and Goa. The study uses a first differenced regression analysis to analyze the differences between the VAT and non-VAT states. Against the express objective of increasing state government revenues, the study shows that there was actually a reduction in the rate of growth of revenues of VAT states compared with the non-VAT states. There were also several variations among states which implemented VAT. This paper then explains possible reasons for variations which includes factors like number of prevailing VAT rates, exemptions, inclusion of inputs in the concessional rate category, and differences in definitions of commodities and state tax legislations.

In order to improve VAT revenues of states and better achieve the objectives of VAT introduction, this paper recommends a freeze on the list of VAT exempted commodities, precise and uniform definition, across states, for key items like ‘capital goods’ and ‘industrial inputs’, removal of inputs from the concessional rate of 4%, and careful monitoring of refunds as well as the impact of the low threshold limit for VAT registration.

1. Introduction

Introduction of VAT by Indian states has been hailed as one of the biggest tax-related reforms in several decades. The Empowered Committee (EC) of state Finance Ministers was the central body which coordinated the design and implementation of VAT. Currently, all Indian states have implemented VAT and the transition to VAT has been fairly smooth.

One of the objectives of VAT implementation was the expected increase in revenues of state governments. This paper attempts to analyze the impact of VAT on revenues of state governments by comparing the group of states which implemented VAT in 2005 with those that implemented it later. Differences among the VAT states are also analyzed.

Section 2 traces the developments on VAT in India over the years, including the political factors that were at play at the time of VAT introduction in 2005. Section 3 gives a justification for the study while section 4 details the main assumptions and methodology. Section 5 analyzes the revenue performance based on various parameters and also explains the differences between VAT and non-VAT states. Based on the analysis, Section 6 suggests a few policy recommendations and Section 7 concludes.

2. The Indian Context

The White Paper[2] on VAT mentions that ‘VAT is a state subject derived from Entry 54 of the State List, for which states are sovereign in taking decisions’. The Empowered Committee (EC) of State Finance Ministers (created by the Ministry of Finance, Government of India) is the body which drafted the details of VAT through several rounds of consultations and also tried to get the assent of all states for implementation. After missing two[3] previous deadlines of April 1, 2002 and April 1, 2003 (Haryana implemented VAT from April 1, 2003),VAT was finally implemented with effect from April 1, 2005 by all states apart from the five[4] Bharatiya Janata Party (BJP – opposition party at the centre) ruled states of Gujarat, Rajasthan, Madhya Pradesh, Chhatisgarh and Jharkhand and three other states – Uttar Pradesh (UP), Uttaranchal and Tamil Nadu. The decision of the BJP ruled states was political while severe resistance by traders was cited as the reason for non-implementation by UP. Uttaranchal’s decision was dictated by that of UP as the former has recently been carved out of the latter. In the case of Tamil Nadu, the reasons included opposition by traders and issues on VAT compensation though the real reason seemed to be the impending state elections in 2006. As of January 1, 2008, VAT has been implemented by all states and union territories[5].

The primary justification for this reform was based on the inherent advantages[6] of VAT over the existing sales tax regime. This included rationalization of tax burden, reduction in prices, simplification of tax structure, greater transparency, improvement in tax compliance, reduction in inter-state ‘tax war’ and increase in state government revenues. Hence, the beneficiaries would include consumers, the business community as well as the government. The main disadvantages[7] of the existing sales tax regime were the multiplicity of taxes and the double taxation of commodities resulting into a cascading effect on prices.

While designing the main structure of VAT, the EC kept in view the need to have common features across states but at the same time allowed for some flexibility within states. However, states have not strictly adhered to the EC recommended guidelines on VAT and hence there are other differences among them.

3. Justification for the study

VAT has been referred to as an ‘efficient and fair[8]’ system of taxation. In order to judge the efficiency of VAT systems worldwide, the IMF conducted a study[9] to ascertain whether countries with VAT systems had higher tax revenues to GDP ratios. The idea behind the study was the fact that if VAT systems did result into higher efficiencies and a consequent lower cost of collection, then countries with VAT systems should be able to collect higher revenues. The study found that countries following the VAT system did have higher general government revenues and grants to GDP, holding other things constant.

One of the stated objectives of VAT introduction in India, as outlined by the EC in its White Paper on VAT, was the advantages that would accrue to state governments in the form of higher tax revenues. The EC expected that the greater transparency and wider tax net of VAT would result into plugging of taxation loopholes and hence would result into higher revenues for state governments.

One of the conclusions of a study[10] done by the Confederation of Indian Industry (CII) after the first 3 months of VAT was ‘One thing is clear that VAT will ultimately result in increase in revenue due to widening of tax base and better compliance’. The study showed a buoyant increase in VAT revenues of states in the first three months of fiscal year 2005-06 compared with the corresponding period in the previous year.

Now that VAT has been under implementation for more than a year, this study aims to find out the impact of VAT introduction on the revenues of state governments. A comparison has been done between the group of states which implemented VAT from April 2005 and those which implemented it later.

4. Assumptions and Methodology

This study evaluates the sales tax/VAT revenue performance of all the major general category statesexcept the National Capital Territory (NCT) of Delhiand Goa[11](total of 16 states). It excludes the 11 special category states[12] of Arunachal Pradesh, Assam, Himachal Pradesh, Jammu and Kashmir, Manipur, Meghalaya, Mizoram, Nagaland, Sikkim, Tripura and Uttaranchal. The decision not to include special category states was because they are primarily dependent on the centre for their revenues and have limited own revenue generating capacity.

The period under consideration is 2001-02 to 2006-07. Sales tax figures upto 2004-05[13]areaudited. However, revised estimates are used for 2005-06 (based on actuals for 9 months and projections for the next 3 months) and budget estimates for 2006-07. The period before 2001-02 has not been considered because 3 new states were formed in 2000-01 and the full year data for them is available only from 2001-02 onwards.

For population[14], the 2001 census has been used as the base and the average annual growth rates for the previous five years have been used to extrapolate the population figures for the rest of the years. For the newly created states of Chhattisgarh and Jharkhand, the population growth rates of Madhya Pradesh and Bihar(respectively) have been used.

The Gross State Domestic Product (GSDP)[15] figures are available up to 2004-05. For the next two years, projected figures are used based on the average growth rate in the previous three years.

The table below gives the per capita VAT/sales tax collection for VAT states (those in which VAT was introduced with effect from April 1, 2005 or before)and non-VAT states (those in which VAT was introduced from April 1, 2006 or later)for the fiscal year 2005-06.

Table 1: VAT and non-VAT states as of April 2005

VAT states / Per capita sales tax collection 2004-05 (INR) / Non VAT States / Per capita sales tax collection 2004-05 (INR)
1 / Andhra Pradesh / 1396 / 1 / Chhatisgarh / 749
2 / Bihar / 212 / 2 / Gujarat / 1547
3 / Haryana / 2119 / 3 / Jharkhand / 617
4 / Karnataka / 1567 / 4 / Madhya Pradesh / 603
5 / Kerala / 2031 / 5 / Rajasthan / 791
6 / Maharashtra / 1854 / 6 / Tamil Nadu / 2019
7 / Orissa / 650 / 7 / Uttar Pradesh / 491
8 / Punjab / 1498
9 / West Bengal / 674
AVERAGE / 1333 / AVERAGE / 974

Source: RBI Bulletin on State Finances

The treatment as well as control groups have a random geographical mix. However, it is interesting to note that the average per capita sales tax collection in the treatment group was higher than that in the control group. Nearly three fourths of the states in the control group have a per capita sales tax collection of under INR 1000 while the situation is just the reverse in the VAT states. However, the reasons why certain states did not go in for VAT implementation with effect from April 1, 2005 were varied.

On baseline characteristics like population, GSDP, sales tax revenues and per capita sales tax collections, the differences between VAT and non-VAT states is not statistically significant, as the table below indicates.

Table 2: Baseline characteristics between Treatment and Control Groups

Variable / Difference in Means / t-statistic
Sales tax revenues / 10502.67 / 0.43
Population / -9.23 / -0.44
GSDP / 176747.1 / 0.38
Per capita sales tax / 359.5 / 1.13

Source: RBI Bulletin on State Finances, CSO, Indiastatonline.com

Hence, randomization seems to have worked well. Therefore, any other confounding variables which could be present would not unduly affect any one of the cohorts.

5. Revenue Performance Analysis

The figure below showsthe performance of the treatment and the control groups over the past five years.

Figure 1: Average per capita sales tax/VAT growth

Source: RBI Bulletin on State Finances and Indiastat online

The figure above shows results contrary to the expectations of the EC. The rate of growth of sales tax collections during 2002-03 and 2003-04 were broadly similar for both groups. The VAT states performed better than the non-VAT states during 2004-05 while the performance reversed during 2005-06, the first year after implementation of VAT.

There is reason to believe that the actual collections under VAT for the year 2005-06 will be lower than those given in Figure 1. This is because the tax collections for 2005-06 have not been fully adjusted for refunds. Some refunds were provided during 2005-06. However, comprehensive data on refunds is not available.

An IMF study[16]had estimated the average VAT refunds (as a percent of gross VAT collections) as given in the table below.

Table 3: Region wise average refunds

Country/Region / Average refund level*
Canada / 50.3%
European Union / 38.1%
Eastern Europe / 36.8%
New Zealand / 35.5%
Former Soviet Union countries / 29.6%
Latin America / 17.4%
Middle East / 16.2%
Asia (excluding Singapore) / 7.0%
Africa (excluding S. Africa) / 6.0%

* As a percentage of gross VAT collections over a four-year period (1998-2001)

Source: IMF survey responses, IMF staff estimates and World Economic Outlook

As can be seen from table 3, refunds are typically higher for advanced countries and countries in transition. The study calls refunds the ‘Achilles heel’ of the VAT system. This is because refunds usually presents a host of problems for the administration like fraud, corruption and delay or denial of refunds by some governments which are facing cash shortages. Refunds mostly accrue to exporters whose sales are zero rated[17]. As the study points out, one of the major reasons for delay in VAT refunds is because of extensive verifications by governments to prevent fraudulent claims.

If we assume a 7% refund level for India (the estimate for Asia), the difference between VAT and non-VAT states is even greater, as shown in figure 2.

Figure 2: Average per capita sales tax/VAT growth after refunds

Source: RBI Bulletin on State Finances and Indiastat online

The figure above shows that there was a significant decline in the rate of growth of per capita collections in VAT states in the first year of implementation.

5A. First Differenced Regression Analysis

In the regression analysis, the dependent variable is growth in per capita VAT (or sales tax) collection. The growth rate (as opposed to the level) has been chosen because the absolute amount of collections varies significantly across states. This indicator also controls for population. The base year on which VAT/sales tax growth for 2005-06 is calculated is 2004-05. The independent variables are the treatment dummy, that is whether the state went in for VAT or not, and the rate of growth of GSDP over the same time period.

A first differenced regression analysis is used to find out the differences between the VAT and non-VAT states between 2005-06 and 2004-05. This regression shows that without considering refunds, the difference between VAT and non-VAT states is not statistically significant. However, if refunds are taken in to account (at the rate of 7% for all states), this difference is statistically significant, as the table below indicates.

Table 4: Regression results with 2004-05 as the base year

Coefficient / t-statistic
Difference in growth rate without refunds / -0.04 / -1.31
Difference in growth rate with refunds / -0.11 / -3.85***

*** Significant at the 1% level

A sensitivity analysis on the percentage of refunds shows that the difference between VAT and non-VAT states continues to be statistically significant (at the five percent significance level) even if refunds are reduced to 2% of VAT revenues.

Another set of first differenced regressions were done using the average per capita sales tax revenues of 2002-03 to 2004-05 as the base (instead of 2004-05 as the base) and the growth of per capita VAT (sales tax) revenues in 2005-06 was calculated on this new base. In the absence of any deductions on account of refunds, the results are not statistically significant. However, if we take into account refunds, at the standard 7% level, the difference between the two cohorts was statistically significant, as the result below indicates.

Table 5: Regression results with average of 2003-05 as the base

Coefficient / t-statistic
Difference in growth rate without refunds / -0.03 / -0.72
Difference in growth rate with refunds / -0.10 / -2.66***

*** Significant at the 1% level

Here again, a sensitivity analysis was carried out with different levels of refunds. Results indicate that the difference between VAT and non-VAT states continues to be statistically significant (at the five percent significance level) even if refunds are reduced to 5% of VAT revenues.

The numbers which have been used for the analysis of states’ performance in 2005-06 are revised estimates. However, it is not expected to have a significant impact on the end result. For example, actual audited sales tax collectionsin 2004-05 (for the 16 states we have considered) were 1.7% higher than revised estimates. This can vary on a year to year basis. However, this will not affect our analysis as the differences are spread across the treatment and control groups.

India’s experience, in terms of a decline in the rate of growth of revenues post VAT implementation is not unique. There have been other countries like Kenya and Philippines[18] which have undergone a similar experience.

5B. Further analysis of the two cohorts

In the Indian states, the VAT/GSDP ratiowas between 2.8-6.5% for 2005-06. However, some of the VAT statesshowed a decline in this ratio during 2005-06 RE. The tablebelow shows the sales tax (or VAT)/GSDP ratios for the treatment and control groups.

Table 6: Sales tax (or VAT) revenues as a percentage of GSDP

VAT states / 2001-02 / 2002-03 / 2003-04 / 2004-05 / 2005-06 RE / 2006-07BE
Andhra Pradesh / 5.1% / 5.1% / 5.0% / 5.5% / 5.7% / 6.3%
Bihar / 3.1% / 3.1% / 3.1% / 3.3% / 3.9% / 3.6%
Haryana / 4.9% / 5.0% / 5.2% / 5.7% / 6.0% / 6.2%
Karnataka / 4.9% / 4.6% / 5.1% / 5.9% / 6.5% / 6.8%
Kerala / 6.1% / 6.6% / 6.7% / 6.7% / 6.5% / 6.5%
Maharashtra / 4.5% / 4.5% / 4.7% / 5.1% / 5.4% / 5.7%
Orissa / 3.3% / 3.7% / 3.4% / 4.2% / 4.0% / 3.8%
Punjab / 3.8% / 4.2% / 4.1% / 4.3% / 4.7% / 5.3%
West Bengal / 2.5% / 2.5% / 2.6% / 2.8% / 2.8% / 3.0%

Source: RBI Bulletin on State Finances and CSO

Control States / 2001-02 / 2002-03 / 2003-04 / 2004-05 / 2005-06 RE / 2006-07BE
Chhatisgarh / 3.2% / 3.5% / 3.4% / 4.2% / 4.9% / 5.0%
Gujarat / 4.8% / 4.5% / 4.3% / 4.6% / 4.9% / 4.7%
Jharkhand / 4.6% / 4.5% / 3.9% / 4.1% / 4.5% / 4.6%
Madhya Pradesh / 2.8% / 3.6% / 3.4% / 3.8% / 4.1% / 4.5%
Rajasthan / 3.5% / 4.1% / 3.8% / 4.3% / 4.6% / 4.8%
Tamil Nadu / 5.8% / 6.2% / 6.5% / 6.9% / 7.2% / 8.0%
Uttar Pradesh / 3.4% / 3.6% / 3.6% / 3.8% / 4.7% / 5.2%

Source: RBI Bulletin on State Finances and CSO