CONTRIBUTED PAPER FOR THE 2007 CONFERENCE ON COR- PORATE R&D (CONCORD) Drivers of corporate R&D investments, Parallel Session 3B

Transcription

1 Knowledge for Growth Industrial Research & Innovation (IRI) The Impact of R&D Tax Incentives on R&D costs and Income Tax Burden CONTRIBUTED PAPER FOR THE 2007 CONFERENCE ON COR- PORATE R&D (CONCORD) Drivers of corporate R&D investments, Parallel Session 3B File name: Sevilla_ernst.doc Author: Christina Elschner, Christof Ernst, Georg Licht Status: <Draft> Last updated: 25 September 2007 Organisation: Centre for European Economic Research ZEW, Mannheim Page 1 of 12

2 TABLE OF CONTENTS 1 Introduction Measuring the Impact of Tax Incentives Impact of R&D tax Incentives on the Effective Corporate Tax Burden Comparison to B Index Conclusions...10 Annex...11 References...12 Page 2 of 12

3 1 Introduction The structure of public support to corporate R&D has been rapidly changing over the last ten years. Two major developments can be observed in a significant number of countries: On the one hand, public funds for basic R&D performed by universities are increasing and linked via additional stimuli for technology transfer (e.g. grants for collaborative R&D projects) to corporate R&D activities. On the other hand, government R&D policies are increasingly oriented towards market signals and competition in order to allocate funds and to minimise the potential for allocative distortions in the R&D decisions of firms. Within this two sided multilateral strategy, tax based R&D incentives are introduced by a large number of countries. In these countries, governments finance a significant share of private R&D via R&D tax incentives. In some countries, reductions in taxes due to R&D tax incentives are even larger than project based direct R&D subsidies. A recent development in funding research and development therefore is the shift towards a higher share of indirect funding trough tax incentives rather than direct subsidies. Significant tax incentives for R&D in most of the OECD member states have been introduced within the past 20 years. R&D tax incentives either affect the taxable base, the tax rate, or the tax due. They reduce firm taxes such as corporate tax, business taxes, or wage taxes withheld by the employer. R&D tax incentives which influence the taxable base of a company lower the taxable income by exempting parts of the taxable income (exemption), by allowing extra amounts to be deducted over current business expenses from the taxable income (allowance), or by accelerated depreciation of assets. Special tax rates or even zero tax rates are sometimes granted to firms, which fulfil certain criteria, e.g. young innovative corporations. Tax credits lower the tax due. The amount of credit depends either on the total amount of R&D expenditures or on the increase in R&D expenditures over a certain earlier period. A fourth form of R&D tax incentives is to reduce wage taxes for R&D personnel and thus reduce employment costs. The design of R&D tax incentives is very diverse. Since the incentives are regulated via tax law and since theoretically all tax payers can claim the incentives, measuring the amount of subsidies via tax and the impact of incentives on total R&D costs is difficult. The B Index by Warda (2001) currently is the most often used indicator that tries to compare the different forms of R&D tax incentives internationally. This indicator is well accepted but is less useful in what concerns tax incentives with certain limitation amounts or regulations for companies with losses. In this paper, our objective is to analyse R&D tax incentives in a more detailed way and measure their impact on the total tax payments of a firm and its R&D costs by means of a simulation model. The remainder of the paper is as follows. The following section 2 describes our simulation model and the measurement of the impact of R&D tax incentives. In section 3, we apply the model to all EU member countries granting R&D tax incentives. Section 4 compares our results with the B Index and section 5 concludes. Page 3 of 12

4 2 Measuring the Impact of Tax Incentives The most common indicator evaluating the impact of R&D tax incentives on the cost of R&D is the B Index (Warda 1996) that is used in OECD studies. It measures the net present value of 1 monetary unit expenditure for R&D compared to 1 monetary unit spent in an alternative investment. The B Index is a useful indicator to generally show the impact of R&D tax incentives on R&D costs. However, it is not possible to integrate any ceilings of incentives, loss reliefs, or progressive tax rates. We therefore aim at overcoming these disadvantages of the B Index by using the simulation model European Tax Analyzer. In contrast to the B Index, our starting point is to compare identical investments, with and without allowing for R&D tax incentives. This enables to directly evaluate the R&D incentive: A lower tax due compared to non existence of the incentive indicates a positive impact of the incentive compared to the same investment without incentive, whereas the B Index compares expenditures of two different investments. The European Tax Analyzer is a firm model that measures the overall tax level of a typical company (Spengel 1995; Jacobs/Spengel 1996; Jacobs et al. 2005). In order to measure the impact of R&D tax incentives, we put the after tax income including the tax incentive in perspective to the after tax income excluding the tax incentive: The European Tax Analyzer is constructed as a model firm defined by an industry specific mix of assets and liabilities and by estimates of the development of production costs, sales projections, and profit distribution over a period of ten years. Based on the (generally existing) capital stock, the future pre tax profits are derived from estimates for future cash receipts and cash expenses associated with the initial capital. The model firm is based on firm level data from the Deutsche Bundesbank (2003) and is typical for a German mediumsized manufacturing company. The financial ratios of our model firm are shown in Table A 1. The economic framework is identical for each simulation, i.e., any differences between pre tax and post tax data in the model are caused only by taxation in the jurisdictions considered. In order to calculate post tax profits, the tax liabilities are derived by taking into account national tax bases and then applying the national tax rates. The effective tax burden is the difference between the pre tax and post tax accumulated value of the corporation at the end of the simulation period, i.e. in period ten. The assessment takes into account all relevant taxes for the corporation. With regard to the tax bases, the rules for profit computation cover depreciation, inventory (stock) valuation, development costs, employee pension schemes, elimination and mitigation of double taxation on foreign source income and loss relief. With regard to tax rates, the simulation model considers statutory linear and progressive tax rate structures, as well as special tax rates. By accounting for more details of tax incentives and for firm specific features, our results however are only valid for the particular firm which is considered in the simulations. Therefore, one has to approve the impact of R&D incentives by varying the underlying economic framework, such as considering various industries and firm sizes. In contrast to that, the B Index has a more general view on R&D tax incentives. But, for analysing and understanding the impact of a tax incentive on a company s tax level, our analysis seems to be more preferable. Besides measuring the impact of incentives on the total accumulated value of Page 4 of 12

5 the firm we also show the impact of the tax reduction on R&D costs when considering various industries and firm sizes. 3 Impact of R&D Tax Incentives on the Effective Corporate Tax Burden In order to analyse the impact of R&D tax incentives on the effective tax burden, we apply the European Tax Analyzer to all EU member states offering R&D tax incentives: Austria, Belgium, the Czech Republic, Finland, France, Greece, Hungary, Ireland, Italy, Malta, the Netherlands, Poland, Portugal, Slovenia, Spain, and the UK. At first, we determine the effective tax burden of our model firm in 25 EU countries and in addition the tax burden for 16 countries, when applying the specific R&D tax incentive to show how much R&D tax incentives can reduce the effective tax burden and country ranking. In a second step, we measure the impact of R&D tax incentives on the firm s accumulated value by simulating tax assessments in EU member states, which have R&D tax incentives in place, and analyse tax drivers. In a third step, we do variations by simulating other model firms under the same conditions. We use these variations to find correlated influences on the impact of R&D incentives. We close this section with a short summary of our findings. Table 3 1: firm value Effective tax burden in EU member states and impact of tax incentives on EU Member State Effective Tax Burden over 10 periods increase of firm value in % without incentive with incentive in 1,000 Euro rk in 1,000 Euro rk Estonia 29 1 no incentive 1 Czech Republic 1, Portugal 1, Ireland United Kingdom 1, Slovenia 1, Latvia no incentive 7 Malta 1, Netherlands 1, Cyprus no incentive 9 Slovakia no incentive 11 Poland Lithuania no incentive 13 Greece 1, , Hungary 1, , Spain 1, , Finland 1, , Sweden 1, no incentive 18 Italy 1, , Luxembourg 1, no incentive 20 Austria 1, , Denmark 1, no incentive 22 Belgium 1, , Germany 1, no incentive 24 France 2, , Table 3 1 shows the effective tax burden in 25 EU countries without R&D and the effective tax burden after applying the specific R&D incentive (R&D). E.g., a Czech model firm without any R&D shows a tax burden of 1,182,000 after the ten year simulation period. The Page 5 of 12

6 effective tax burden falls 307,000 if the model firm can claim the Czech R&D tax incentive. All analysed countries with R&D tax incentive, except France, improve their ranking for at least one position. Those countries become more attractive to companies running R&D activities eligible for the analysed tax reliefs. Table 3 1 shows the impact of R&D tax incentives on the accumulated firm value in the 16 member states of the EU with R&D tax incentives. Most of the countries in our survey give significant tax incentives for R&D which can heavily decrease the tax bill of an investor and his corporation. However, countries like Finland or Poland have smaller incentives in place. Thus, there is a large gradient in between these European countries. In order to explain that strong gradient, we split between tax regulation related factors and nonregulation related factors. Regulation related factors stem from specific features of R&D tax incentives in the analysed countries, like the extent of an extra allowance or tax credit and the eligible R&D expenditures (see Table A 2 for more details about the R&D). The Czech Republic shows the largest impact in our analyses. It achieves the largest decrease in the effective tax burden with its extra allowance of 100% for current R&D expenses (so that in effect 200% of the expenditures are deductible for tax purposes). Countries which show large increases in the firm value and thus significant reductions in the tax burden as well are Malta (50% volume, extra allowance) and Portugal (20% volume and 50% incremental, tax credit). In the Netherlands, the accumulated firm value increases by 16.6%, caused by their degressive tax credit of 42%/14% on personnel expenditures. The firm value in Hungary, having a huge extra allowance of 100% which is similar to Czech Republic, increases by 7.1%. Ireland applies a tax credit only for increases in R&D expenditures. The increase of R&D expenditures in our simulation is quite low at 2.3% a year. Therefore, its impact on the firm value is only 0.7%. The lowest impact is found in Finland and Poland; these countries allow accelerated depreciation on certain capital expenditures. That incentive shows only little impact because capital expenditures only account for approx. 11% of R&D expenditures and also accelerated depreciation causes only small interest effects in our model. One reason for the different impact of incentives can be seen in the applied kind of R&D tax incentives. The relief of extra allowances depends on the volume of the extra allowance and on the applicable corporate tax rate. The higher the tax rate, the higher the reduction by the incentive. Corporate tax rates play an important role. The Czech Republic and Hungary, both have got the same allowance of 100% of current expenditures, but different corporate tax rates of respectively 24% and 16%. Thus the Czech incentive is more effective than the Hungarian one. In addition, the Hungarian incentive is limited to a maximum of approx. 200,000 of eligible expenses which limits the impact as well, especially for corporations featuring high levels of R&D expenditures. The UK applies a progressive corporate tax rate. Accordingly, we found that the allowance causes progression (tax brackets) effects. A company having a lower corporate tax base and therefore a lower corporate tax rate experiences a lower tax reduction than a company, which has to apply the full corporate tax rate. The impact of tax credits depends only a little on the tax rate. Portugal, for example, having a moderate corporate tax rate, is first among countries with R&D tax credits, due to the high percentage for its tax credit. If the impact of a tax incentive should become equal for all legal forms of companies, this quality can become important. Usually, there are different Page 6 of 12

7 effective tax rates for limited companies and unlimited partnerships, a tax credit can achieve the same reduction of the tax bill for the amount of R&D expenditures. Further variables for extra allowances and tax credits are the eligible expenditures. Although most countries use the Frascati Definition for R&D activities, some of the countries only accept certain R&D activities or expenditures for their incentives. E.g. the Netherlands and Belgium focus on R&D personnel expenditures. Accordingly, we found that the Dutch tax credit has a stronger impact for companies having more personnel expenditures. Belgium grants a 25% tax credit for R&D personnel expenditures; we thus found similar effects for model companies having more R&D personnel expenditures. In addition, there are countries like Spain which have an extended understanding of eligible R&D activities and expenditures. However, our model firm is not engaged in these R&D activities and this aspect does not show up in our results. In general, a wider understanding of R&D leads to more eligible R&D expenditures and causes other features remaining constant, a stronger impact. By varying our model firm in terms of financial ratios while applying the same national tax system, we found significantly different impacts of the R&D incentives for these model firms, compared to our standard model firm. ω thus can be quite different for different companies within the same country and using the same R&D incentive. Therefore nonregulation related influence factors are the second group of variables to influence the impact of an R&D tax incentive. These factors can explain why there are relevant differences in the impact of an R&D tax incentive between different model firms within the same country and tax legislation (for details see Table A 1). To get a better view for the reduction in the effective tax burden caused by R&D tax incentives, we put the reduction of the effective tax burden within 10 periods in perspective to the accumulated R&D expenditures (see Table 3 2). Table 3 2: Reduction of effective tax burden to accumulated R&D expenditures in % Benchmark case Model firm A Model firm B Model firm C Model firm D Model firm E Model firm F Czech Republic Portugal Malta Slovenia Spain Netherlands United Kingdom France Greece Italy Austria Hungary Belgium Ireland Poland Finland Again, we can see a strong gradient in between the analysed countries from 19.2% in the Czech Republic down to 0.0% in Finland. Comparing the analysed model firms, we found significant variations. These variations are significant enough to change the ranking of the countries for specific model firms (e.g. Netherlands / UK for model firm B and France / Page 7 of 12

8 Greece for model firm C). Taking one specific country and comparing the variations, it turns out, that within the same country, the reduction of tax burden to R&D expenditures can be between 15% and 3,7% (Hungary). Apart from the obvious direct effect of the amount of R&D expenditures on the value of a tax incentive, there is an important influence given by the R&D expenditure structure. As described above, variations in the share of personnel, used materials and capital expenditures change the impact of the incentive in countries having different rates for different expenditures. If there is a limit or a degressive rate for those different kinds of expenditures, there are additional size effects by the total amount of expenditures. The Netherlands have a tax credit for the amount of withheld income tax on salaries and wages for personnel in R&D at a rate of 42% on the first Euro and at 14% on the exceeding amount. As a result, the strongest impact for the Netherlands was found for companies staying below the 110,000 limit for the 42% bracket (model firm C, E). Companies exceeding that limit faced a significantly smaller impact due to the lower 14% rate for the credit (B, F). Countries featuring less generous incentives tend to be more constant for the analysed model firms. Therefore the profitability of the company seems to be an important variable. A higher profitability leads to a higher tax due, as the corporate tax rate must be applied on a higher corporate tax base. If a R&D tax incentive should be available in total for the corporation, the incentive needs a corporate tax base or tax liability which is high enough to allow a tax credit or extra allowances to be fully deductible. If the tax burden of the company is not high enough, only a share of the beneficial incentive can be used in that specific tax assessment year. In most cases, there exist options for carry forwards for some periods (assessment year). The efficiency of the incentive then depends on the profitability in the years to come. If the profitability rises in the future, the company is more likely to be able to take advantage of the new incentives in the recent year in addition to the unused carry forwards. If the profitability falls, it is likely that there are new carry forwards. In case the profitability stays low for some years, those carry forwards can either be lost by law or can be used only after a long period of time (due to regulations). We found that the following effects for a combination of companies featuring either poor profitability or having a high ratio of R&D expenditures to turnover or both and for countries giving significant incentives. The effect was strongest for the Czech Republic, Portugal, Greece, and France. If a company has got a huge amount of carry forwards which are in danger of being lost, there could be an incentive to even lower the recent R&D activity in order to first use those carry forwards which are about to deteriorate. Some countries therefore refund money for certain non useable R&D credits or allowances. However, there are significant discounts from the carry forwards (e.g. UK: refund of 24% of unuseable R&D allowances). Generous conditions for carry forwards or carry backwards can help to avoid problems with non useable tax incentives. Focusing on Hungary and Spain for different model firms, we found that there is a smaller relief for model firms featuring high R&D intensities (B, F) compared to other model firms. These two countries have limitations for their incentives. Hungary has a limit for its extra allowance of approx. 200,000 and Spain limits the R&D tax credit to 35%/50% of the corporate tax bill. The distribution of R&D expenditures over time can change the impact of R&D tax incentives significantly. We assume an increase of 2.3% each year in our calculations. Constant Page 8 of 12

9 or falling expenditures in those countries which give an incentive for increased R&D expenditures (e.g. Ireland, Greece, Portugal, France, and Spain) would lead to significantly lower impacts. One extreme would be Greece with its extra allowance of 50% on the volume in case there is an increase in R&D expenditures (in case of no increase or constant expenditures there is no incentive). Slightly increasing or heavily increasing R&D expenditures could change the picture as well. 4 Comparison to B Index Section 2 gave a brief discussion to what extent the methodological approaches of measuring the impact of R&D tax incentives by Warda, i.e. the B Index, and by means of the European Tax Analyzer differ. The following Table 4 1 compares the resulting indicators of both approaches for several European countries. The second column shows the reduction of the effective tax burden caused by R&D tax incentives and discussed in Section 3. The third column shows the B Index (OECD STI Outlook 2006). Table 4 1: Comparison of reduction of effective tax burden compared to accumulated R&D expenditures with B Index. Reduction of effective tax burden compared to accumulated R&D expenditures B Index (Large firms) for Benchmark case (in %) Czech Republic Portugal Malta 16.6 Slovenia 14.8 Spain Netherlands United Kingdom France Greece Italy Austria Hungary Belgium Ireland Poland Finland It turns out that our ranking of tax incentives widely differs from the B Index for some countries. There are several reasons for this finding. First in our simulation model, we compare identical economic situations when comparing the tax burden with and without allowing for R&D tax incentives. Our calculations show the effect on the effective tax burden of a model firm over 10 periods. Warda compares the situation of R&D expenditure to the possibility of a full write off of current expenditure (which would equal a B Index of 1). Thus Warda has a B Index of above 1, i.e. when investments in machinery have to be depreciated over time (compared to being fully written off in the period). The B Index would be below 1 if there is a better treatment of R&D expenditures because of an R&D tax incentive. Our indicator, the total accumulated value of the firm after considering taxes and R&D tax incentives, rises if there is any R&D incentive. Page 9 of 12

10 Second, our indicator accounts for the economic situation of the model firm and limitations of the incentive. Thus countries, that grant high incentives in terms of high rates but at the same time have low ceiling until which the incentive can be used, rank worse than in the B Index, like it is the case for Spain. In addition, the B Index might apply different proportions for current and capital expenditures and wages and salaries for R&D. Third, we also take into account other taxes than just the corporate tax. Hungary holds a good rank in the B Index but ranks 12 in our ranking. This is due to fairly high amount of taxes payable under the Hungarian business tax which is a non profit tax. Other relevant taxes are property taxes, payroll taxes and taxes on capital. Fourth, our indicator accounts for the deductibility of interest cost against taxable income at the corporate level and for the ability of internal financing. 5 Conclusions In this paper, we measured the impact of R&D tax incentives on the effective corporate tax burden. Our calculations are based on calculations for the effective corporate tax burden, which we run with the European Tax Analyzer. We showed that R&D tax incentives in the EU member states have a significant impact on the effective corporate tax burden. Rankings showing the effective tax burden of a specific firm changed when R&D tax incentives are taken into account. Comparing the reduction of the effective tax burden to the accumulated R&D expenditures, we found that there are significant differences in the impact of R&D incentives for the analysed model firms, even within the same country. We found variables and effects for the impact of the analysed tax incentives on the tax burden and on R&D costs by simulating a model firm including its underlying economic framework over 10 periods. Our different approach leads, compared to the common B Index, to different results for some countries. These differences can be explained with the different methodical approach and different assumptions. There is need for further work, especially in the field of analysing the impact of certain influence factors, like the volume and the distribution of R&D expenditures over time. Page 10 of 12

11 Annex Table A 1: Financial ratios of model firm [Benchmark Case] (Period 6) Benchmark Case Model firm A Model firm B Model firm C Net profit 210, , , ,928 Balance sheet total 5,970,983 6,795,772 6,322,486 5,598,732 Turnover 8,073,091 9,840,933 9,249,152 9,082,228 Share of fixed assets 27.8% 31.86% 17.39% 30.75% Return on sales 2.6% 2.86% 2.55% 1.86% Return on equity 19.2% 16.16% 15.15% 16.09% Equity ratio 19.6% 27.99% 26.84% 20.09% Return on assets 5.0% 5.43% 5.25% 4.84% Inventories to total capital 25.8% 21.52% 29.63% 18.50% Costs for personnel to turnover 29.6% 22.88% 27.54% 17.40% R&D Expenditures to turnover 5.70% 5.00% 8.40% 0.60% Model firm D Model firm E Model firm F Net profit 211, , ,182 Balance sheet total 6,205,997 6,008,352 5,558,483 Turnover 8,285,231 9,431,544 8,887,880 Share of fixed assets 18.64% 29.02% 25.84% Return on sales 2.55% 2.71% 2.14% Return on equity 17.52% 22.60% 21.59% Equity ratio 21.43% 21.18% 17.65% Return on assets 5.04% 5.79% 4.61% Inventories to total capital 31.45% 24.49% 23.35% Costs for personnel to turnover 32.77% 25.55% 27.12% R&D Expenditures to turnover 3.70% 1.40% 6.50% Page 11 of 12

12 Table A 2: Extra allowance Tax credit Volumebased Incremental Volumebased Incremental Source: ZEW, IBFD Overview of R&D tax incentives. Personnel expenditures Other current expenditures Capital expenditures UK (150% SME/125%), UK (150% SME/125%), MT (150%), MT (150%), AT (125%), AT (125%), CZ (200%), CZ (200%), HU (200%, limitation 200T ) HU (200%, limitation 200T ) GR (150%, if increase), AT (135%) BE (25%, R&D personnel), FR (10%, limit 10Mio ), NL (42%/14%, R&Dpersonnel), IT (10%, SME), PT (20%), SL (20%), ES (20%, 35%/50% limit) FR (40%), IR (20%), PT (50%) GR (150% if increase), AT (135%) FR (10%, limit 10Mio ), IT (10%, SME), PT (20%), SL (20%), ES (30%, 35%/50% limit) FR (40%), IR (20%), PT (50%), ES (50%) BE (extra depreciation 14,5%), FI (immediate depreciation), UK (immediate depreciation), NL (immediate depreciation for certain list of assets), AT (25%), PL (extra depreciation 50%) PT (20%), SL (20%), ES (10%, 35%/50% limit) IR (20%), PT (50%) References Deutsche Bundesbank (2003) Verhältniszahlen aus Jahresabschlüssen deutscher Unternehmen von 1998 bis 2000, Statistische Sonderveröffentlichung, 6. IBFD (Corporate Taxation, 2006): Corporate Taxation in Europe, Amsterdam Jacobs, O.H. and C. Spengel (1996), European Tax Analyzer, Baden Baden. Jacobs, O.H., C. Spengel, T. Stetter, and C. Wendt (2005), EU Company Taxation in Case of a Common Tax Base: A Computer Based Calculation and Comparison Using the Enhanced Model of the European Tax Analyzer, Intertax 33, OECD (2002), Frascati Manual 2002: Proposed Standard Practice for Surveys on Research and Experimental Development, OECD Publications, Paris. OECD (2006), Tax Treatment of Business Investments in Intellectual Assets: an International Comparison, STI Working Paper DSTI/DOC (2006)4, Paris. OECD (2006), Science, Technology and industry Outlook 2006, OECD Publishing, Paris. Spengel, C. (1995), Europäische Steuerbelastungsvergleiche, Düsseldorf. Warda, J. (1996), Measuring the Value of R&D Tax Provisions, in: OECD, Fiscal Measures to Promote R&D and Innovation, OECD (96)165, Paris, Page 12 of 12