Measuring the pre-and post-GST tax cost of investment

by Gaurav S. Ghosh and Jack Mintz.

The most wide-ranging change to the Indian tax system in decades is the introduction of the goods and service tax (“GST”). But is this tax beneficial or harmful for investors, especially for new investment? Is the GST impact uniform across sectors, or does it favour in some sectors while harming investment in others?

India’s GST reform has integrated state level sales taxes and central excise and service taxes into the GST, which is a centralized value added tax (“VAT”) and enables businesses to claim more refunds of taxes paid on purchases from other businesses. It removes a significant amount of taxes on business inputs that are cascaded into business costs and passed on to consumers or businesses purchasing goods and services from other businesses. Some non-refundability of input taxes remain for exempt sectors such as agriculture, petroleum and alcohol. Generally, though, the Indian VAT reform will result in a signficant reduction in VAT on business input costs as we show below.

We report results from our recent study where we evaluate the cost burden placed by India’s tax code upon potential investors, both before and after the introduction of the GST. We do this by developing an economic model that is calibrated to represent the Indian tax system, and then using this model to simulate GST impacts. The model and its implementation are described below. This is followed by discussion of our results.

The METR model

Our tool for evaluating the impact of the Indian tax system on investment incentives is the Marginal Effective Tax Rate (“METR”), which measures the tax wedge imposed upon investment. The METR is an analytical framework developed in the 1980s to evaluate tax systems in their aggregate and to facilitate cross-country and cross-sectoral comparisons. Seminal METR studies from this era are Auerbach (1983), Boadway et al. (1984) and King & Fullerton (1984). The METR has since been used by academics and governments to evaluate tax competitiveness, gauge the economic impacts of changes to the tax code, and design investment-friendly tax policies. The METR analytical framework is useful because it provides a strong empirical basis to tax policy debates. It has proved itself over the years by being used by policy makers worldwide. Ours is the first in-depth implementation of the METR framework to the Indian economy.

The tax wedge, $\omega$, estimated under the METR framework, is the difference between the pre-tax rate of return earned by a marginal project, $r_g$, and the post-tax rate of return that accrues to the marginal project’s investors, $r_n$. The METR itself is the tax wedge divided by the pre-tax rate of return.

\[ METR = \frac{\omega}{r_g} = \frac{r_g - r_n}{r_g}\]

The sizes of the tax wedge (and the related METR) is affected by direct taxes, sales taxes on capital purchases and other capital-related taxes like stamp duties. The METR also accounts for tax incentives including accelerated depreciation, initial allowances, and tax credits. High METRs imply high tax loads and low returns to investors and vice versa. High METRs therefore indicate that the associated tax systems are less competitive when it comes to attracting capital investment.

Estimating the METR for a marginal project requires the estimation of $r_g$ and $r_n$ for that project. In the METR framework, both variables are functions of a wide array of tax rates and incentives; the functional forms are derived through recourse to standard microeconomic principles and assumptions. We do not provide technical details here. Interested readers can refer to Ghosh & Mintz (2017). We only note that the post-tax rate of return $r_n$ is equal to the inflation-adjusted market interest rate for issuing bonds and equity finance, which is the same across all businesses net of risk. The pre-tax rate of return, $r_g$, is estimated by estimating the user cost of capital model (Jorgenson, 1963) net of depreciation and risk. The Indian tax system therefore affects the estimation of $r_g$ and $r_n$ – and by extension, the estimation of the METR – because of its impact on the real market rate of return and the user cost of capital within the structure of our model.

Aggregation principles

METR estimation requires consideration of the marginal investment that earns a pre-tax rate of return sufficient to cover taxes and the market rate of return. The marginal investment depends on the characteristics of its industry and the choice of production technology. Every sector, in effect, has a multitude of marginal investments varying across characteristics like industry, asset class and financing structure. Each marginal investment has a different METR because it faces a different tax treatment once all relevant taxes, exemptions and their interactions are considered. Debt-financed transportation in the power sector will, for example, have a different METR than equity-financed machinery in the agriculture sector. There are multiple reasons for this: the differential treatment of debt and equity in the Indian tax code where interest payments are deductible, but dividends are not; the different incentives available to investors in the power and agriculture sectors; and the differential tax treatments of asset classes.

Accurate estimation of the METR, whether at the sectoral or all-India level, requires consideration of this heterogeneity across different types of marginal investment. Consistent with the literature (King & Fullerton, 1984; Mintz, et al., 2016), we estimated METRs by using a bottom-up approach.

First, we identified four key tax and economic characteristics of a marginal investment in India. These were sectors, asset types, investment sizes, and whether the marginal firm was paying the corporate income tax (“CIT”) or the minimal alternate tax (“MAT”). Each marginal investment would have some level of each of these characteristics. Second, we identified the number of levels for each characteristic. There were nine sectors, six asset types, two firm sizes and two types of tax payers, as shown in Table 1, leading to 216 unique marginal investments. The identified sectors (except “Others”) were the main investment destinations in India, accounting for 93 percent of investment in 2015 among firms in the Prowess database. The asset types were those identified for differential treatment under the Indian tax code. Firm size was selected on the basis of the investment threshold for initial allowances: these are only allowed for investments exceeding INR 250 million.

Table 1: Characteristics and levels of marginal investments in India
Industry / Sector	Asset type	Firm Size	Tax payer type
Agriculture, forestry & fishing	Buildings	Small firms, with investment < INR 250 million in 2015	CIT payer
Construction	Furniture & fittings
Electricity, steam, gas & AC supply
Finance & insurance	Inventory
Information & communication (“Infocom”)	Land	Large firms, with investment > INR 250 million in 2015	MAT payer
Manufacturing	Machinery
Wholesale & retail trade, repair of motor vehicles & motorcycles	Transport
Transportation & storage
Other

Third, METRs were estimated for each of the 216 marginal investment types. This required collection of tax and economic data for each type, and then the incorporation of this data into the formal METR model. Finally, METRs were calculated at different levels of aggregation as weighted averages of subsets of the 216 types. For example, the all-India METR was a weighted average of all 216 METRs, while a sectoral METR was a weighted average of the 24 METRs relevant to that sector. The weights used were the capital shares associated with each marginal investment type. The capital share was the proportion of all new investment capital in a given year that was allocated to a given marginal investment type. The capital share data and certain other data were obtained from the Prowess database. Other data sources were the Indian Income Tax Act, tax guides published by EY and other accounting firms, official Indian government communications, the Thomson Reuters EIKON financial database, and publications by the Central Statistical Office.

Results

Some results from our analysis are presented below. We begin with a comparison of METRs before and after the implementation of the GST, which is shown in Figure 1. We compare METRs both at the all-India level and at the sectoral level. The blue (red) bars represent pre-GST METRs (post-GST METRs).

We see that the GST leads to a fairly large drop in METRs at the all-India level, from 28% to 22%. This is because the GST removes pre-GST blockage of many input tax credits. Many of the blockages arose because different indirect taxes – such as state VAT and central excise and service taxes – could not be set off against each other. Consider the plight of a services firm, which paid state VAT on some inputs, but collected central service tax from its customers. Since state taxes were not creditable against central service taxes they were blocked, leading to tax cascading. Unable to claim credit for state taxes paid, the firm’s input costs would be higher by the amount of the blocked taxes. This higher tax burden would lead to a higher METR. Post-GST, the distinction between central and state taxes vanished and therefore many blockages and cascades also vanished.

Other blockages arose because of exemptions granted under the pre-GST system. It is a common misconception that tax exemptions are business-friendly. Nothing could be further from the truth. Since the exemption recipient does not pay the output tax, it cannot set off its input taxes. All input taxes in exempt sectors are therefore blocked and lead to a higher METR on investment. The GST has retained some previous exemptions, such as in the agriculture sector, but has removed others. This has contributed to the overall drop in the METR.

We also see that METRs are heterogeneous across sectors. Pre-GST, METRs were lower in production-related industries than in service industries. Production sectors may have had lower METRs, but for different reasons. Manufacturing benefited from relatively low central and state indirect tax rates, as well as few input tax credit blockages. This is because most investment in the manufacturing sector is into machinery and equipment (“M&E”), which had relatively favourable tax treatment. Agriculture and electricity, on the other hand, were exempt sectors and therefore had blocked credits. However, these sectors received other benefits from the tax code that, together, brought down their METRs. Agriculture was exempt from the corporate income tax, while electricity benefited from large tax depreciation allowances and (like manufacturing) low indirect tax rates on M&E. The service industries had higher METRs because they faced higher taxes on their inputs as well as greater blockages on input tax credits. The asset mix for the service sectors consisted of comparatively less M&E (with the exception of finance) and more of other assets like land, buildings, transport and inventory. The pre-GST tax treatment of these other assets was relatively less favourable than for M&E.

When comparing pre- and post-GST METRs, two results stand out. First, METRs have reduced for all sectors with the sole exception of electricity (to be explained below). Second, the size of the METR reduction varies across sectors with the reduction being higher in the service sectors than in the production sectors. As a result, although variations in METR persist in the post-GST era, the size of these variations has reduced, with the specific result that the tax competitiveness gap between the production and service sectors has also reduced.

These results can be unpacked. Perhaps the most notable result in the figure above is that the GST raises the METR in the electricity sector, while reducing it in others. Pre-GST, the electricity sector faced two distortions: sector-specific indirect tax incentives and blocked input tax credits. The former was beneficial and the latter harmful from an investment (and METR) perspective. The two distortions cancelled each other out, leading to a relatively low METR of 29%. Post-GST, the beneficial sector-specific incentives were removed while the harmful blockages remained. As a consequence, the METR increased sharply in this sector to 38%. The services sectors benefited more from the GST because they were the ones for whom blocked credits were a greater problem in the pre-GST era. This led to a reduction in the tax competitiveness gap between the services and production sectors.

Summary

The merits and demerits of the GST have been debated vigorously in the press and among the policy community in recent months. We contribute to this debate by presenting the results of the first (to our knowledge) empirical investigation of the impact of the GST on the incentive to invest in India. We find that the GST does improve investment incentives at the all-India level by reducing the marginal effective tax rate or METR from 28% to 22%. This is achieved through a reduction in the blockages of input tax credits across value chains and a reduction in indirect tax exemptions. The all-India METR numbers mask heterogeneity at the sectoral level. Pre-GST, the Indian tax code incentivized investment in production sectors like manufacturing and electricity through lower METRs, while the tax cost of investment was relatively higher in service sectors like transport, information & communications and trade. Post-GST, this pattern of incentives has changed. While METRs for manufacturing remain low, METRs for the power sector have increased significantly. METRs for the service sectors have also come down sharply post-GST. METRs in some service sectors like finance and trade are now roughly equal to that for manufacturing.

In summary, the GST has reduced the overall tax cost of investment in India and reduced investment distortions in the tax code, somewhat levelling the playing field between the production and service sectors as destinations for investment. This is good news, but we caveat it by pointing out that the full potential of the GST as an incentive for investment has not been reached. We have unreported results that show that METRs would come down further – particularly in the electricity sector – if the remaining exemptions were removed.

References

Auerbach, A., (1983), Taxation, corporate financial policy and the cost of capital, Journal of Economic Literature, 21(3), pp. 905-940.

Boadway, R., Bruce, N. & Mintz, J., (1984), Taxation, inflation, and the effective marginal tax rate on capital in Canada, Canadian Journal of Economics, 17(1), pp. 62-79.

Ghosh, G. & Mintz, J., (2017), Investment and the Indian tax regime: Measuring tax impacts on the incentive to invest in India, Bangalore: EY.

Jorgenson, D., (1963), Capital theory and investment behavior, American Economic Review, Volume 53, pp. 247-259.

King, M. & Fullerton, D., (1984), The taxation of income from capital: A comparative study of the Unites States, the United Kingdom, Sweden and West Germany, Chicago: University of Chicago Press.

Mintz, J., Bazel, P. & Chen, D., (2016), Growing the Australian economy with a competitive company tax, Sydney: Minerals Council of Australia.

 

Gaurav S. Ghosh is Senior Manager, Ernst & Young, LLP and Jack Mintz is Palmer Chair in Public Policy and Director, School of Public Policy, University of Calgary.