Gender and Taxation: Incidence of Indirect Taxes

Pinaki Chakraborty

Lekha Chakraborty

Krishanu Karmakar

May 2009


1

Preface

This study has been undertaken at NIPFP in collaboration with University of Kwazulu-Natal, South Africa and American University, Washington DC. The research team for the study was led by Pinaki Chakraborty; & Lekha Chakraborty.

The Governing Body of the Institute does not take any responsibility for the contents of this study; it belongs to the authors only.

M. Govinda Rao

Director

1

Gender and Taxation: Incidence of Indirect Taxes

  1. Introduction
The standard principle of public finance suggests that in the process of development, the tax structure should move from indirect to direct taxes and within direct taxes it should move from corporate tax to income tax and, in case of indirect taxes, it should move from foreign trade taxation to domestic indirect taxation (see Ahmed and Stern: 1991, Newbery and Stern: 1987). However, historically, it has been noted that “dependence on indirect taxes, particularly taxes on domestic trade, is however, greater in India than in many other countries at similar levels of development and, contrary to the trend observed in other countries, the transition of tax structure associated with economic development (Hinrichs: 1966) has not come about even after significant increase in the levels of income and shifts in the structure of the economy over the past four decades.” (Bagchi: 1997, pp-104-105). It has also been noted that “the fundamentals of Indian tax structure have changed little since the Government of India Act in 1935 set out the basic assignments of revenues and responsibilities to the centre and the states. The assignments were subsequently embodied in the Indian Constitution of 1947. However, in terms of the relative importance of the sources of revenue, there have been important changes. For example, land revenue has declined almost to the point of insignificance and the salt tax has been abolished, whilst excises, sales taxes and customs duties have increased greatly” (Burgess and Stern: 1994, pp-44). Also, in the post reforms period since 1990s, the share of direct taxes in total tax has been increasing steadily.
India has a very complicated tax structure. In a three-tier federal fiscal system, Seventh Schedule of the Constitution has divided the tax power between centre and states. As per the assignment of taxes, different taxes collected by the Union government are commonly known as the Central taxes that can be further classified into the Direct and Indirect taxes. The main components of the Direct Central Taxes are the Personal Income Tax and the Corporate Income Tax[1]. The Customs duties and Excise duties are the major indirect taxes collected by the Union government. On the other hand, the States collect taxes mainly in the form of Sales tax (now VAT), State Excise duty on liquor, Stamps and Registration Fee, Taxes on Transport and other minor taxes. The central government also has the power of taxing services. But the service tax is introduced on a standalone basis and is not integrated with the goods tax. The Central Government and the all the state governments together are contemplating introduction of integrated Goods and Services Tax (GST) from the fiscal year 2010-11, which should change the structure of taxation in India on a permanent basis. Apart from the central and state level taxes, there are few revenue sources that are with the local governments, mainly property taxes, entertainment tax and other minor taxes. However, the share of local taxes in the combined revenues of all levels of governments is negligible.
If we look at the recent history of taxation in India, the Central government undertook major tax reform programme since 1991 and states have also followed subsequently by introducing rapid reforms in indirect taxes by replacing the origin based sales tax system by destination based value added tax in 2005. These reform measures are structural in nature and have considerably influenced the tax structure and revenue mobilization at the central level, and have the potential to do so at the state level. Initial indications suggest that rapid reform in state level consumption taxes have resulted in higher revenue mobilization and helped removing certain distortions in the consumption tax system.
In this backdrop we evaluate the level of taxation in relation to GDP. The long run trend in tax-GDP ratio shown in figure 1 taking both central and all the state governments together revealed that the combined tax ratio, which was little more than 5 per cent of GDP during 1st half 1950s, increased to more than 16 per cent of GDP during the mid 1980s. With the introduction of rapid tax reforms during the economic reforms during the 1990s, though resulted in a decline in the tax to GDP ratio, it started showing an increasing trend from the early 2000 reaching to the mid 1980 level.


It is also to be noted that the decline in tax-to GDP ratio was primarily due to the fall in central revenuesdue to the fall in both the customs duty and Union excise duty. However, during the 1990’s, the direct tax effort increased sharply. It should be noted that as Indian tax structure is heavily dependent on the indirect taxes[2], the increase in the direct tax revenue effort could not compensate the resultant revenue loss arising out of the decline in the tax effort of the indirect taxes. While commenting on government’s excessive dependence on indirect taxes to mobilise resources, Burger and Stern (1994) noted that excise, sales and customs ‘taken the strain’ and in the process an adhoc structure has arisen with new goods being brought into the tax net and rates on certain goods being increased in response to the pressure on revenues[3].

In case of indirect taxes, significant reform is the introduction of VAT in the fiscal year 2005-06 by 21 states. Now except, Uttar Pradesh, all other states have introduced Value Added Tax. The VAT is the only alternative to the current system of sales taxation, which apart from cascading of tax through production of inputs, creates economic distortions through multiple tax rates, selective exemptions and incentives. A proper design of VAT is critical to achieve the desired outcome of value added taxation. As per the consensus achieved among the states, recommended VAT rates are three, viz., 1 per cent, 4 per cent and 12.5 per cent. While, the 1 per cent rate is applicable to jewellary, the 4 per cent rate is applicable to inputs and some items of basic necessities, the 12.5 per cent rate is the general rate applied to residual commodity group. While exports are zero rated, each state is given the flexibility to exempt few goods of local importance from VAT. Four commodities, viz. Liquor, Petrol, Diesel and Aviation Fuel are kept outside the VAT and the floor rates for these commodities are 20 per cent. Also, Textile, Tobacco and Sugar are not taxed by the states but an additional excise duty of 4 per cent imposed on these items by the central government and revenues accrued are shared with the states under tax rental arrangement.

In this paper we examine the incidence of VAT on consumptionto find out the gender differential incidence of consumption taxes across households based on the data obtained from the 61st round of the National Sample This survey is conducted in India; at spatially disaggregated level in every five years. Gender based tax incidence analysis can be used to examine the distributional impact of taxes by taking into account intra-household relations. Gender incidence analysis, as with burden analysis,require assumptions based on limited data, but can nevertheless is a useful exercise[4]. We examine the VAT incidence for a specific state, viz., West Bengal.

2. Indirect Tax Incidence

As mentioned earlier, although there are a spectrum of indirect taxes imposed by both central and the state governments, we focus our attention on the incidence of consumption tax levied by the states[5]. Consumption tax or VAT constitutes the single largest source of revenues at the state level constituting more than two-third of the own tax revenues of the state. We have 28 state governments and 7 centrally administered union territory having an independent tax system and administration, and also consumption pattern and the level of taxation measured in terms of tax to GDP ratio differs widely across states. Give these differences; it would be ideal to examine the level of incidence of taxes in each of them separately. We focused on one of the representative states, namely, West Bengal for the purpose of our analysis of consumption tax incidence.

West Bengal is a middle income state with a very low tax to GSDP[6]ratio. As evident from the Figure 3, compared to many other states of similar levels of development, West Bengal’s tax ratio is lower at little more than 4 per cent of GSDP when the average tax ratio for all the states together is around 7 per cent (See Figure 3). The tax structure given in Figure 4 also reveals that VAT constitutes the single largest share of own tax revenues for the state (59 per cent). Prior to the introduction of VAT in 2005, consumption tax in the form of sales tax had 15 different rates ranging from 1 to 20 per cent. But with the introduction of VAT, these rates have been rationalised and consolidated into 3 basic rates, viz. 1 per cent, 4 per cent and 12.5 per cent and exports are zero rated. The character of the goods falling under different categories of VAT rate and used for the incidence analysis is given in Table 5. As evident from the Table, basic food items, viz., cereals and vegetable, basic clothing, domestic services, basis stationary and books are exempted from VAT. Food items, other than basic, are taxed at 4 per cent and the rest of the category is taxed at the general rate of 12.5 percent. As mentioned earlier, like other states in India, West Bengal also has a list of exempted commodities, which are treated, as goods of local importance and exempted from VAT. There are 170 items that are exempted from VAT including those that are goods of local importance.

Table 1: Nature of Commodities and the Applicable VAT Rate

Tax Rate / Number of Goods under VAT / Nature of Goods
Exempted / 170 / Basic Food Items like Cereals and Vegetables, Basic Clothing, Domestic Services, Basic Stationary and Books,
1% / 3 / Jewellery and Ornaments
4% / 79 / Food Items other than basic, Household Goods, Medicine etc.
12.50% / 81 / Some Beverages, Some Electronic Goods, Some Medical Equipments and Expenses, Household Goods which can be classified as luxuries

Source: Commercial Taxes Department(Government of West Bengal)

Another issue in VAT design needs mention here is that there are selected items currently under sales tax kept outside the VAT and taxed differently at a higher rate. These items are diesel, petrol and aviation turbine fuel (ATF), crude oil and liquor (See Table 6). Floor rates for all these taxes are 20 per cent for all the states. In West Bengal except for petrol and diesel which are taxed at 25 per cent, non-vatable commodities are taxed at the floor rate of 20 per cent.

Table 2: Tax Rate on Commodities Outside the VAT

Item / Tax Rate*
Country liquor / 20.00
Beer / 20.00
Foreign liquor or refined liquor / 20.00
Petrol / 25.00
Diesel / 25.00

Source: Commercial Taxes Department

(Government of West Bengal)

Exclusion of services, from the VAT base is another major weakness of the evolved VAT design. Exclusion of services from the base, even if the states are empowered to tax selected services on a standalone basis, would not eliminate the problem of cascading from the tax system (Rao: 2004). It also discriminates goods against services and has given rise to serious problems in separating out the service component in sale of goods taking place in several instances, e.g., in the case of execution of works contracts, services of food in restaurants and so on (Bagchi: 1997).

Given this background, we examine the incidence of VAT on consumption expenditure of West Bengal to find gender differential tax incidence. Gender concerns in a limited way can be implemented through differential tax rates on commodities and exemptions of certain commodities (necessities) from VAT. This can help in neutralizing the implicit gender bias of commodity taxes. However, it should be taken note of that gender relation within the household and its decision making process also plays an important role in determining whether such a policy to neutralize gender bias will be successful. For example, a higher tax on a non-necessity like alcohol may not reduce the real consumption of alcohol rather lead to a reduction in the share of income allocated for the consumption of necessity items. Similarly, an exemption from tax of necessary commodities may lead to the allocation of a greater share of income for the consumption of non-necessities instead of hiking the consumption of necessities in case the gender relations in household decision making process is not very progressive.

As a starting point to this analysis, apart from analysing the tax incidence in aggregate, we have examined the gender differential consumption pattern across households and thereby the tax incidence across different categories of households in terms of headship and household composition and thereby the progressivity or regressivity of the tax system. We have also examined the rural and urban households separately. This has resulted in three different sets of estimates – One set for the aggregated for the state and the other two for the urban and rural sub-samples respectively. We think, a priori the rural and urban households will have different expenditure pattern and thereby tax incidence. This can only be analyzed if we consider the urban and rural sub-samples separately as separate entities. As Indian economy is pre-dominantly rural and more than two-third of the total population lives in rural India, with subsistence agriculture as the livelihood for majority of them, it is important that we analyse urban and rural sector separately. However, we have not been able to do the tax incidence according to the employment status of the households, as consumer expenditure survey does not give data on employment status of households. Employment survey is conducted separately and the consumption expenditure survey data cannot be mapped against the employment survey data.

We have used the data obtained from the 61st round National Sample Survey of consumer expenditure. For our purpose, we have taken sub-sample combined estimates based on all rounds. The survey collects data on household characteristics, demographic details of the household and household level expenditure on almost 400 items. For some frequently purchased items the expenditure data are collected for a 30 days reference period, for the non-so-frequently purchased items the data is collected for both a 30 days and a 365 days reference period. While for certain items data about both quantity and value are collected for others only values are collected. Finally, for consumer durables only the 365 days, reference period is used. Moreover, for certain items, especially food articles, both home production and total consumption data are collected.

A note of caution on household consumption expenditure data may be in order. The household consumption expenditure data have their limits for tax analysis, particularly for taxes on intermediate consumption (Rajemison and Younger: 2003) and existing studies have restricted themselves to taxes on final demand or make strong assumption about the incidence of taxes on inputs (Younger et. al; 1999, Sahn and Younger 1998, Bird & Miller, 1989). In this study also we consider taxes on final demand.Our sample consists of 7877 households from West Bengal - both urban and rural. In the tables we report the mean tax incidence figures for the households in each quintile.

An aggregate picture of tax incidence is important before one deals with the gendered classification of households and tax incidence thereon. The aggregate picture for rural, urban and combined is given in Table 3. It is evident from the table that tax incidence falls more on the poorest quintile vis-à-vis richer quintiles when we look at the combined incidence taking both rural and urban households together. However, the degree of regressivity in terms of incidence is much higher in urban households vis-à-vis rural households. In the rural households, the tax incidence, though falls more on the poorest expenditure quintile vis-à-vis other quintiles; the level of incidence is much lower than the urban households. Another point that emerges from the Table 3 is that VAT incidence is higher than the incidence of fuel levy and excise. Also, unlike excise, fuel levy appears to be progressive. The incidence of fuel levy goes up with the expenditure quintile. Fuel levy includes fuel levy for household fuel as well as well as fuel for transport. We would see in subsequent discussion, the incidence of household fuel is again regressive.