Value Added Tax and Informality: Determinants of Registration of Enterprises under State VAT in India

1. Introduction

The first step towards participation in most tax regimes is registration in the regime. The existence of an informal sector in most economies, perhaps larger informal economies in developing countries, where the agents do not participate in the tax regime, raises concerns for tax departments. From the taxpayers’ perspective, being in the tax regime would be associated with certain costs and certain benefits. If the costs exceed the benefits, it is expected that the taxpayer would choose to remain outside the realm of taxation. Some studies have explored whether formalisation helps a firm to increase its value added or profits (refer, Demenet, Razafindrakoto, & Roubaud, 2016; McKenzie & Sakho, 2010). Participation in the tax system is important for tax administration as well since it can crucially influence tax revenue as well as tax morale in the country (Hofmann, Hoelzl, & Kirchler, 2008). The present article is an attempt to identify the factors that might influence the decision of the enterprises to register for taxes, specifically for the value added tax (VAT).

There are many studies which build theoretical models to understand the effects of a VAT on informality (Araujo & Rodrigues, 2016; Boadway & Sato, 2009; de Paula & Scheinkman, 2010; Emran & Stiglitz, 2005; Joshi, Prichard, & Heady, 2014; Kanbur & Keen, 2014; Keen & Mintz, 2004; Keen, 2008; Piggott & Whalley, 2001; Tumen, 2017). There is also an emerging body of studies which empirically explore factors influencing the registration for VAT (Hansford, Hasseldine, & Howorth, 2003; Harju, Matikka, & Rauhanen, 2015; Klahr, Joyce, Donaldson, Keilloh, & Salmon, 2017; Webley, Adams, & Elffers, 2002). Since there is no such literature specific to Indian unincorporated enterprises, according to our understanding, in this article we attempt to understand the factors influencing decision of unincorporated enterprises to take VAT registration. Taking cue from earlier studies conducted elsewhere, we explore factors which influence decision of the enterprises to take VAT registration across Indian states.

For India, the study acquires added importance in the context of the recent introduction of the goods and services tax (GST) regime from 1 July 2017. With a reduction in the extent of cascading in the tax regime, it is argued by some that the move to a GST would result in the expansion of economic activity. Since this new tax regime works through more integrated and redefined supply chains, for units to benefit from the new regime and for its success it is important that more and more firms find it useful to be a part of the regime. While firms and enterprises in the organised sector do participate in the GST regime, those in the unorganised sector may not be as well integrated. This poses a problem both for the units and the tax administration. For the former, apart from being unable to benefit from the growth-enhancing processes in the economy, they may be subject to irregular visits by various authorities potentially associated with the payment of bribes. For the tax department, non-participation by a segment of the economy can induce lower confidence in the tax regime resulting in higher non-compliance even among segments which would normally pay taxes.

It is the purpose of this article to examine within the space of the unin- corporated sector, the extent of participation in the tax regime among enterprises which are required by law to participate and to identify the characteristics of enterprises which could be associated with non-participation. This analysis is undertaken with a focus on state value added tax (VAT) regimes in India. The state VAT replaces sales tax system and was adopted across the states since April 2003. ¹

The focus is not on the amount of taxes paid but on whether the enterprise is registered with the tax department or not. This exercise can provide some inputs for designing policies to bring these enterprises into the mainstream. It may be noted that the results of this article are relevant even for the GST regime, since VAT and GST differ only in coverage.

The article is organised as follows: Section 2 explores definitions of the informal economy or unorganised economy in a country like India. In Section 3, we provide a brief description of the database and restrictions imposed to select the sample enterprises. We explore the factors that could influence the VAT registration of enterprises in Section 4. We provide detailed methodology to econometrically understand the differences in behaviour across enterprises—more specifically, binary choice probit models are estimated to explore factors that influence the decision of enterprises to register with the VAT department—and discuss the results in Section 5. Section 6 provides some concluding observations.

2. Searching for a Definition of Informality

Reducing informality is often seen as one of the central objectives of tax reforms (Kanbur & Keen, 2015), and the proposed GST system for India envisages that informality would go down. There are many definitions of the informal economy or unorganised economy in a country like India. The definition of informality set out by the International Labour Organization (ILO) (2013) is based on labour and enterprise regulation rather than on tax considerations. In fact, all the Indian acts and rules consider the size of employment and/or capital investment to classify enterprises as registered and unregistered (or organised versus unorganised). To provide an overview of the differences in definitions adopted by different regulators in India, we summarise some of the important regulations which deal with the informal/small/unorganised economy in India.

The Indian manufacturing sector is categorised into organised and unorganised activities based on the status of registration under the Indian Factories Act, 1948, which (under Section 2[m]) defines ‘factories’ (organised activities) as those:

whereon 10 or more workers are working, or were working on any day of the preceding 12 months, and in any part of which a manufacturing process is being carried on with the aid of power, ² or is ordinarily so carried on, or

Therefore, the act defines organised and unorganised manufacturing on the basis of the number of workers and the source of energy. In this definition, no consideration is given to registration with the tax authority to define organised manufacturing.

The Micro, Small and Medium Enterprises Development (MSMED) Act, 2006, is applicable for enterprises engaged in manufacturing/production of goods (pertaining to an industry specified in the Industries Development and Regulation Act, 1951) or engaged in providing or rendering of services. The MSMED Act defines enterprises as shown in Table 1.

Enterprises falling under this definition are required to register under the MSMED Act and file an information memorandum with the authorities. In this definition, the size of investment in the plant and machinery (for manufacturing enterprises) or equipment (for service providers) is the only factor considered, and there is no requirement for registration with any tax authority.

Based on their registration under the Indian Companies Act, 1956, businesses are classified as incorporated or unincorporated entities. There are a few options for entrepreneurs to organise their business, and registering under the Indian Companies Act is one such option. Other options are firms (partnership), association of persons (AOPs), body of individuals (BOIs), proprietary enterprise, charitable entities, self-help groups, trusts and others. Tax treatments (under direct tax) for these business organisations vary where corporates and firms are required to pay corporate income tax; AOPs, BOIs and proprietors face income tax provisions similar to individuals, while cooperative societies face different tax rates.

In sampling the unorganised enterprises in India, the National Sample Survey Office (NSSO) further divides proprietary enterprises into two categories, such as own-account enterprises (OAE) and establishment, which is further divided into non-directory establishments and directory establishments. The definitions of these are as follows:

Own-account enterprises (OAEs) are those which are run without any hired worker employed on a fairly regular basis. ⁴

In addition to the above, businesses are also required to register under different acts, such as the Shops and Establishment Act, ⁵ Trade License (Municipal Corporation/Panchayats/Local Body), ⁶ Value Added Tax (VAT)/Sales Tax Act, Provident Fund Act, ⁷ Employees State Insurance Corporation Act and so on, which in turn have their own definitions of which enterprises are required to register. ⁸ Of these, the tax regime uses the turnover of an enterprise as the basis for determining whether it needs to be registered or not.

The above discussion shows that there are multiple definitions under different acts to define/categorise businesses in India. Enterprises need to register under different acts if they fulfil the criteria for registration. The existence of different definitions under different acts may be due to the purposes for which the acts were established, and these are not always mutually exclusive. However, broadly the number of workers and investment in plant and machinery (or equipment) are the major indicators chosen to classify activities. A tax policy, however, tends to focus on income or turnover rather than on investment or employment. In other words, while the other regulations focus on the processes involved in the economic activity, the tax department focuses on the outcome of the activity. While there can be some overlaps in these different definitions, there is no reason to expect them to converge completely. In studying informality with respect to taxes, therefore, it is useful to focus on the variable of interest in defining tax liability that is income or turnover as the case may be.

If one considers informality with respect to taxes, then there can be two types of agents in the informal sector. There are instances when businesses legally obliged to pay taxes do not pay them, and alternatively there are cases of businesses that are not legally obliged to pay taxes as their annual turnover falls below the threshold. ⁹ For the state VAT regimes in India, the rules vary across states. In most states, all businesses are not legally required to be registered under the VAT—any business with a turnover below a certain threshold which does not undertake interstate or international trade is not required to be registered. Further, even if they are registered, if their annual turnover is below the prescribed threshold for VAT registration, they do not have to pay taxes. The obligation to file a tax return also varies across states. For some states, filing a tax return under the VAT/sales tax is mandatory for all registered businesses even if they have zero annual turnovers (e.g., Maharashtra); whereas in others, filing is mandatory for registered businesses only if their annual turnover crosses the VAT threshold. However, if enterprises are not filing tax return they cannot raise VAT invoices against sales.

However to participate in the tax regime, it is essential to register with the tax department. It might be argued that any business that registers with the tax department would either file a return or expect to file a return sometime in the near future. Considering registration therefore as an expression of intent to participate in the tax regime, the present study seeks to analyse this decision of unincorporated enterprises.

If we define informality in terms of status regarding tax registration, there are some common perceptions about the characteristics of the informal sector. We attempt to test some of these characteristics for unincorporated enterprises in India. Before dealing with characteristics and their support from literature, it would be worthwhile to provide a brief description of the database.

The National Sample Survey Office (NSSO) conducts quinquennial surveys on unincorporated enterprises. ¹⁰ This article is based on unit-level data of the 67th round of the NSSO (National Sample Survey Office [NSSO], 2012a, 2012b, 2013) conducted from July 2010 to June 2011 in which the NSSO captures characteristics of the bottom strata of enterprises engaged in manufacturing, trading and services. A detailed description of the data is provided in Mukherjee and Rao (2015). In this article, we have restricted our analysis to proprietary and partnership enterprises, as the tax status of the other categories (e.g., self-help groups, trusts) are not clear.

We have restricted our sample to proprietary enterprises and partnership firms engaged in manufacturing and/or trading activities. The rationale for excluding services is that the majority of service providers are not required to be registered under the state VAT/Sales Tax Acts (including activities such as transportation and storage, postal and courier, information and communications, finance and insurance, real estate, education and human health and social work). Further, since we would like to identify state-specific features as well, we have excluded states with very few observations that is those with less than 100 observations. Given that the focus of the present exercise is on businesses required to be registered, we have excluded all enterprises with turnovers below the VAT-exemption threshold for the state in which they operate. At present the VAT registration threshold varies across states and some states have different thresholds for manufacturers and for traders—for example, Chhattisgarh, Gujarat and Himachal Pradesh (Table A1). Since registration of enterprises having annual turnover lower than the prescribed threshold for VAT registration is voluntary, we have considered only those enterprises which are legally required to be registered under the VAT/Sales Tax Act of the respective state of their operation. ¹¹

With all the restrictions (cleaning up of data), we are left with 47,528 observations, of which 11,729 enterprises (24.7%) are registered, 19,675 enterprises (41.4%) are unregistered and 16,124 enterprises (33.9%) did not reveal their registration status during the survey. All 47,528 sample enterprises are supposed to be registered under VAT, as their annual turnover is above the prescribed threshold for VAT registration, however only one-fourth of them are currently registered (Table A2). We have excluded enterprises which did not reveal their registration status during survey from our sample. Therefore, effectively our sample size becomes 31,404 (Table A2).

In order to capture state-specific factors influencing registration, we have included state dummies. However, out of 20 states we have selected for our analysis, one state is to serve as the base state. The selection of the base state is based on average per capita gross state domestic product (GSDP) from 2009–2010 to 2011–2012. ¹² The state with a per capita GSDP close to the median (Quartile 2) could be the base state. We found that Karnataka had an average per capita GSDP of ₹45,227 which is approximately close to Quartile 2 (₹43,721), and so we chose it as the base state (Table A1).

4. Factors that Could influence VAT Registration

We use a bivariate heteroskedastic probit model to estimate response probabilities for the registration of enterprises under the state VAT/Sales Tax Act (regvatact). ²⁰ ‘Hetprobit’ fits a maximum-likelihood probit model and is a generalisation of the probit model.

The estimated model and variables are described below:

y_i, i = 1, …, n, is a binary outcome variable taking on the value 1 (success) or 0 (failure). In the probit model, the probability that y_i takes on the value 1 is modelled as a nonlinear function of a liner combination of the k independent variables x_i = (x₁, x₂, …, x_k),

Pr ( ​ y i = 1 | x ) = Φ ( x ′ i β ) ​

where, β is constant k—dimensional vector and Φ() is a standard normal cumulative distribution function with variance one.

Heteroskedastic probit generalises the probit model by assuming Φ() to be a normal cumulative distribution function with a variance (σ²) that is a function of independent variables. Following Harvey (1976) ‘hetprobit’ models, the variance as a multiplicative function of m variables z_i = (z₁, z₂, …, z_m) with n observations each.

σ i 2 = { exp ( z ′ i γ ) } ​ 2

Therefore, the probability of success as a function of all independent variables is as follows:

Pr ( ​ y i = 1 | x , z ) = Φ ( x ′ i β exp ( z ′ i γ ) )

(1)

For identification of the Model (1), unlike in the index x ′ i β ​ , no constant term can be present in z ′ i γ ​ .

For convenience, let y be the n × 1 vector of all observations of y_i, let X be the n × k matrix whose ith row is x ′ i ​ and let Z be the n × m matrix with ith row of z ′ i ​ . Then the log likelihood function can be written as follows:

ln L ( ​ β , γ | y , X , Z ​ ) = ∑ i = 1 n ( ​ y i ln Φ ( x ′ i β exp ( z ′ i γ ) ) + ( 1 − y i ) ln ( 1 − Φ ( x ′ i β exp ( z ′ i γ ) ) )

Given a set of n observed values of the random vectors y, x, z, in order to obtain the maximum likelihood estimate, one maximises this function over the space of possible choices of (β,γ) ϵ R^k+m.

Using this methodology, we run the following binary choice heteroskedastic probit model:

Pr ( r e g v a t a c t = 1 | x , z ) = Φ ( x ′ i β exp ( z ′ i γ ) ) ​

(2)

Model 1 specification:

x ′ β = β 0 + β 1 l t u r n o v e r + β 2 l t o t a l w o r ker + β 3 l y e a r o o p + β 4 l o c a t i o n o u t * u r b a n + β 5 g o v t a s s i s t + β 7 m f g + β 7 p r o p * o a e + β 8 s t a t e d u m m i e s ​ z ′ γ = γ 0 l t u r n o v e r + γ 1 l t o t a l w o r ker

Model 2 specification:

x ′ β = β 0 + β 1 l t o t a l w o r ker + β 2 l m k t v a l t o t a s s t + β 3 l t o t a l c r e d i t + β 4 c r e d i t f o r m a l + β 5 g v a t u r n o v e r + β 6 l y e a r o o p + β 7 l o c a t i o n o u t * u r b a n + β 8 g o v t a s s i s t + β 9 m f g + β 10 p r o p * o a e + β 11 s t a t e d u m m i e s ​ z ′ γ = γ 0 l t o t a l w o r ker

Dependent variable:

regvatact: one if the enterprise is registered under the VAT/Sales Tax Act, zero otherwise.

Scale indicators (push factors):

lturnover: log of annual receipts/turnover (in ₹)

ltotworker: log of total workers (full-time and part-time, male and female)

lmktvaltotasst: log of the market value of total (own and hired) assets (other than land and buildings) (in ₹)

ltotalcredit: log of the total outstanding loan from all sources (in ₹)

Pull factors:

creditformal: one if there is outstanding loan from formal sources, zero otherwise

Enterprise-specific factors:

gvaturnover: ratio of annual gross value added and turnover

lyearoop: log of years of operation (as on 2011)

locationout: one if location of the enterprise is outside the household premises (permanent location), zero otherwise

govtassist: one if the enterprise received government assistance, zero otherwise ²¹

mfg: one if the enterprise is engaged in manufacturing only, zero otherwise

urban: one if the enterprise is located in an urban area, zero otherwise

prop: one if it is a proprietary enterprises, zero otherwise

oae: one if the enterprise is an own-account enterprise, zero otherwise ²²

State-specific factor:

State Dummy: one for the relevant state, zero otherwise ²³

The basic statistics for the data are summarised in Table A3. The results of the estimation exercise are presented in Table 6.

OPEN IN VIEWER

Table 6

Regression Results

Variable Name	Model 1			Model 2
	Coefficient	SE	ME	Coefficient	SE	ME
lturnover	1.670***	0.244	0.192
ltotalworker	0.992***	0.156	0.118	0.493***	0.043	0.171
lmktvaltotasst				0.018*	0.010	0.006
ltotalcredit				0.317***	0.022	0.109
creditformal				0.155***	0.05	0.053
gvaturnover				−1.195***	0.137	−0.41
lyearoop	0.104***	0.035	0.011	0.092***	0.023	0.031
locationout*urban	0.506***	0.093	0.055	0.227***	0.043	0.078
govtassist	0.758***	0.195	0.085	−0.062	0.062	−0.021
mfg	−1.271***	0.190	−0.134	−0.510***	0.057	−0.173
prop*oae	−0.360***	0.098	−0.039	−0.157**	0.062	−0.054
Jammu & Kashmir	−0.930***	0.214	−0.096	−0.688***	0.119	−0.223
Himachal Pradesh	1.614***	0.313	0.183	0.461***	0.133	0.156
Punjab	0.313	0.193	0.035	−0.009	0.127	−0.003
Uttarakhand	−1.083***	0.264	−0.110	−0.602***	0.175	−0.196
Haryana	1.180***	0.266	0.133	0.573***	0.135	0.192
Delhi	1.322***	0.281	0.149	0.507***	0.175	0.17
Rajasthan	0.012	0.181	0.001	−0.142	0.132	−0.049
Uttar Pradesh	1.370***	0.253	0.154	0.624***	0.123	0.208
Bihar	1.011***	0.263	0.114	0.542**	0.217	0.181
Assam	−1.816***	0.320	−0.175	−0.318*	0.191	−0.107
West Bengal	−2.217***	0.338	−0.216	−0.726***	0.096	−0.237
Odisha	0.558*	0.290	0.062	0.126	0.166	0.043
Chhattisgarh	−1.370***	0.321	−0.136	−0.082	0.243	−0.028
Madhya Pradesh	−0.201	0.167	−0.022	0.052	0.169	0.018
Gujarat	0.114	0.162	0.012	−0.124	0.105	−0.042
Maharashtra	−1.269***	0.241	−0.129	−0.467***	0.099	−0.156
Andhra Pradesh	−0.921***	0.202	−0.096	−0.229**	0.091	−0.078
Kerala	−1.253***	0.251	−0.127	−0.379***	0.093	−0.128
Tamil Nadu	−1.605***	0.282	−0.159	−0.467***	0.094	−0.156
Constant	−25.859***	3.774		−4.477***	0.287
lturnover	0.080***	0.010
ltotalworker	0.081***	0.017		0.102***	0.026
Number of observations	31,321			6,733
Odd ratio	0.595			0.914
LR chi2	4404.52***	df: 26		1858.5***	df: 29
Log likelihood	−16322.2			−3559.0
Pseudo R2	0.368			0.484
chi2 for Ho:lnsigma2 = 0	172.33***	df: 2		15.69***	df: 1

Source: Provided by the author.

Notes: ***, ** and * imply that the estimated z-statistic is significant at the 0.01%, 0.05% and 0.10% levels, respectively; SE is standard error; ME is marginal effect; df is degree of freedom.

The results show that all the scale variables influence entrepreneurs’ decisions to take VAT registration positively and significantly. The results are as per our expectation, because as an enterprise grows, it aspires to get integrated into a larger chain of economic activities, and therefore taking VAT registration enables them to pass on the input taxes to the next level of economic agents. Access to credit from formal sources (creditformal) also works as an incentive (pull) factor which induces entrepreneurs to take VAT registration. In other words, while access to formal credit does not appear to be the sole determinant of the decision to remain informal or become formal, at the margin, it does seem to play a role in encouraging registration for VAT.

Turning to enterprise-specific factors, the results show that relatively old enterprises (operating for longer periods), located outside the household premises in urban areas and those that received assistance from government are more likely to take VAT registration. On the other hand, enterprises with a higher ratio of GVA in their turnover (gvaturnover), engaged in manufacturing and which are own-account proprietary enterprises are relatively less likely to take VAT registration. Enterprises with a higher gvaturnover will face a larger tax burden on output compared to their input tax credit, therefore, they might be relatively reluctant to take registration under VAT.

To capture state-specific factors, we have used state dummies where we have taken Karnataka as a base state. The results show that compared to enterprises located in Karnataka enterprises located in Himachal Pradesh, Haryana, Delhi, Uttar Pradesh and Bihar are more likely to take VAT registration. Similarly, enterprises located in Jammu and Kashmir, Uttarakhand, Assam, West Bengal, Chhattisgarh, Maharashtra, Andhra Pradesh, Kerala and Tamil Nadu are less likely to take VAT registration; similarly, enterprises located in Punjab, Rajasthan, Odisha, Madhya Pradesh and Gujarat are equally likely to take VAT registration as enterprises in Karnataka.

Two interesting results come out from this study—first, on average, unregistered enterprises face a higher cost of capital from informal sources of credits. This could be understood as follows: a higher cost of capital could imply that the economic viability of the enterprise is lower, and hence the entrepreneur would have less interest in being registered for tax purposes. Further, a higher cost of capital is associated with borrowing from informal sources. Interest payments on such borrowings might have to be paid in cash requiring the need to keep transactions out of the books of accounts. ²⁴ If this direction of causation in decision-making is valid, it would suggest that increasing access to formal sources of credit could provide a windfall benefit to governments in the form of higher tax registration and perhaps a resultant increase in tax collections.

Second, an increase in assistance from the government is associated with a higher probability of registration with VAT departments. This result too supports greater intervention by the government in supporting unincorporated units, even from a tax department perspective.

A counter-intuitive result however follows from the finding that the coefficient of the dummy for manufacturing is negative in the regression models—suggesting that all other things remaining the same, the probability of a manufacturing unit being registered for VAT/sales tax is lower than that of a trading firm. This result is apparently counter-intuitive since compared to trading enterprises manufacturing units face a lower VAT registration threshold. This result suggests two things—first, it is possible that manufacturers are small units with their own marketing systems and not part of a supply chain. Since they are not integrated with the rest of the economy, they may not perceive any merit in registering for VAT. Second, the fact that manufacturing units are less likely to register suggests that tax departments are unable to monitor economic activity being undertaken in their jurisdiction.

Based on the turnover threshold set for VAT registration by states, different state tax administrations face different level of challenges in bringing unincorporated enterprises under the tax net. To integrate unincorporated enterprises with the rest of the economy, it is imperative to bring them under the tax system. While it is often argued that there are costs for enterprises in remaining outside the tax system, since a number of firms still choose to do so, it appears that the self-policing (or self-enforcing) dimension of the VAT regime does not provide adequate benefit (Keen & Smith, 2006). Even the existing tax compounding schemes do not seem to be attractive enough to bring small dealers into the system. It is therefore important to explore alternative measures which could change this scenario. From the results in the present study, it appears that facilitating access to formal credit might be one such instrument. The other can be a focus on expanding consumers’ incentives to ask for an invoice. If larger segments of the economy ask for invoices for their purchases, the incentive and the option to remain out of the tax regime would be correspondingly reduced.

The location of the enterprises also plays an important role in the incentive to register with state sales tax authorities. Enterprises located outside households (in fixed premises) and urban areas are easy to identify and could potentially attract inspections from state tax officers, and therefore should be more likely to take registration. Our data analysis shows that even in this category not all the enterprises are registered. This throws up a question on the efficiency of the state tax administrations. An efficient tax administration could potentially look for opportunities to expand the tax base by bringing more assees under the tax net.

It is not expected that mere registration with a state tax authority would result in sudden substantial tax revenue mobilisation for states, but a gradual increase in registration would result in the integration of unincorporated enterprises into integrated supply chains of the formal economy. In the long run, it is hoped that enterprises will reap the benefits of economic integration through backward and forward linkages, and tax administrations will get a cleaner system to deal with. Apart from interventions, such as improved access to formal credit and incentives from governments, tax departments also conduct regular surveys to identify errant taxpayers. This activity however is low on their list of priorities, given their limited manpower. In the GST regime, it is not clear which agency would take on this vital responsibility.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.