Taxes on production against productivity and competitiveness

18The taxation of an input (capital or intermediate factor) raises its relative price and will lead companies to make bad choices in their production process. This distortion can thus encourage them to use other inputs that are less productive or of lower quality but less taxed. This is therefore a misallocation that can affect the productivity of the entire economy. This may seem secondary, but recent literature on productivity and its slowdown has shown that this is not the case. These distortions at the level of individual companies can be amplified at the level of the economy as a whole as a result of the spread along the chain of production, as will be discussed in more detail below in the case of C3S. Thus, Baqaee and Farhi (2017) [16] use a general equilibrium model to assess the macroeconomic impact of these poor allocations –potentially from taxes on production– at the microeconomic level. They estimate that for the US economy, the reallocation of inputs accounts for half of the increase in aggregate factor productivity between 1997 and 2015, as much as technological progress. Introducing distortions in the choice of factors of production and intermediate consumption is therefore not a minor issue.

2

Production taxes paid by companies in 2016

Production taxes paid by companies in 2016

Reading: Heading D29 of the national accounts, excluding taxes on the wage bill paid by non-financial corporations and financial corporations.

19Among the taxes on production, some tax transactions between companies and households. This is the case for labour services (e.g. wage tax), or land services (CFE if the owner of the land or building used for productive purposes is a household). The degree of harmfulness of these taxes is less important than taxes on the exchange of factors between companies. This second part of Diamond and Mirrlees’ result led us to exclude taxes on the wage bill from the scope of the study. A caveat, however, is that the result of Diamond and Mirrlees assumes full use of the factors of production. [17] We keep this condition in mind when formulating our proposals.

20Finally, the same authors indicate that it is necessary to prohibit the taxation of trade between the domestic and foreign productive sectors. Indeed, the degree of harmfulness is comparable to taxing trade within the domestic productive sector. With the C3S, our country has achieved the feat of imposing a customs duty on its own production, as we detail in this Note (see below). A company that incurs a tax on a factor used in its production process sees its costs increase. It may decide to defer this additional cost in different ways, either by reducing other costs (e.g. salaries) or, for example, by increasing its selling price. In the latter case, this undermines the company’s price competitiveness, both in exports and on the domestic market. From this point of view, taxes on production also act as an export tax. In addition, companies needing to purchase intermediate goods and services are encouraged to source their supplies abroad rather than in France, because of this impact on prices.

Rather tax final goods and final income

21In contrast to input taxation, the optimal taxation literature recommends taxing only final goods, so as not to introduce distortions into companies’ production decisions. This form of taxation is the one operated by VAT. It is always the same very powerful result of Diamond and Mirrlees that inspires economists here. It is better to tax transactions between companies and households than to tax transactions between companies. This certainly distorts consumer choices, but it is a lesser evil compared to distorting producer choice.

22The taxation of profits as operated by the corporate income tax (CIT) is generally considered –in terms of efficiency– to be a better tax than taxes on production. It does not increase, unlike taxes on production, the company’s breakeven point and thus has less impact on the probability of survival of companies. It does not lead, at least directly, to an increase in production costs and prices and thus to a deterioration in competitiveness. However, competitiveness can also be affected by the CIT through its negative impact on innovation and investment. In addition, the CIT reduces the attractiveness of France as a production site. Finally, the taxation of corporate profits is particularly exposed to tax optimisation practices, exploiting differences or inconsistencies between different tax jurisdictions, aimed at artificially transferring profits to countries where they are taxed at a lower rate, thereby eroding the tax base. This problem is not unique to France. The growth of multinational companies and the development of the digitalisation of the economy invite us to explore ways of reforming corporate taxation so as to be better adapted to this context and to limit the possibilities of profit relocation. This subject will be the subject of a future CAE Note.

Who pays the CFE: the owners or users of the buildings?

23The CFE is based on the rental value of buildings used by companies. It taxes the use of the real estate factor, unlike the property tax that taxes real estate property. Although it is the occupying companies that pay this tax, the economic analysis has shown that it is not systematically the entity subject to the tax that suffers the burden. Thus, if the supply of commercial property is less responsive to changes in rents (supply is inelastic) than demand, then the tax will weigh on owners, and on user companies if it is not.

24To our knowledge, there is no empirical estimate of the impact of taxation on commercial real estate in our country. Several analyses on residential real estate show that housing subsidies are in fact at least partly to the benefit of landlords and not tenants. [18] Assuming that these empirical results could be transposed to the CFE despite the differences (commercial and industrial real estate and non-residential sector, on the one hand, and tax and non-subsidy, on the other), the CFE could have relatively little impact on taxable companies.

Does CFE influence corporate behaviour?

25Faced with CFE, companies could choose technologies that require less real estate, but that are potentially less productive. However, we find no empirical evidence on this subject. On the one hand, an econometric analysis based on company data [19] reveals a weak negative relationship between the CFE rate and fixed assets such as land and buildings for companies in the manufacturing sector. On the other hand, no significant statistical link is detected between the CFE rate and rental expenses. Moreover, since real estate is necessary for the installation of machinery and personnel, a lower consumption of this factor could change the company’s production process. However, there is no link between the CFE rate and machinery equipment or employment, so that the production process of companies seems to be relatively unaffected by the CFE. Finally, companies could pass on the cost of the CFE to their selling prices. Yet, no statistical link emerges between the tax and the turnover or exports, whether focusing on all firms or restricting the sample to the manufacturing sector. The CFE therefore does not seem to cause any major distortions on the production of French companies or affect their sales.

26It is also possible that companies may substitute between the potential territories where companies are located. Rathelot and Sillard (2008a) [20] analysed the impact of TP –a significant part of which was the rental value base of the property used, such as the current CFE– and showed that if company relocations existed, they were very low. Continuing their study for greater tax impacts (free zones allow exemption from local taxes, but also corporate income tax and social security contributions), it was shown that these relocations were geographically close or even within the same municipality, [21] without any major effect on the overall number of establishments. [22] In addition, another study on Switzerland indicates that these effects decrease with agglomeration economies, [23] with companies in metropolitan areas being less highly responsive in terms of location. It therefore seems that companies’ behavioural responses to this tax are limited, generating few economic distortions.

A tax to encourage local authorities to reserve land for businesses

27In addition, the CFE, which provides revenues for municipalities and inter-municipalities, can be seen as a subsidy to encourage them to free up land for companies. Indeed, in their urban planning decisions, local authorities may wish to limit professional real estate, whose beneficial effects in terms of employment and activity are shared with the employment area, whereas it may present negative externalities in the immediate surroundings of the place where the company is located. The taxation of this property makes it possible to internalize at the level of the decisionmaker of local urban planning the benefit generated by the creation of business parks. In addition, it gives municipalities room for manœuvre. For these various reasons, we do not recommend a major reform of the CFE at this stage.

A tax on the tax and other “cascading effects”

29With a turnover tax, each good produced is taxed again if it enters the production process of another company: in this sense, C3S acts at each stage of production as a tax on the tax. Companies pass on the tax at least partially to their customers by increasing their prices: among these customers are companies that buy intermediate consumption necessary for their production, and which must therefore not only pay the tax, but also bear the increase in their production costs caused by the price increases of previous production stages. The latter companies therefore also tend to reflect both taxes and increases in production costs in their prices. This mechanism creates a “cascading effect” that particularly affects the prices of products at the end of the production chain, whose production requires many steps. Firms producing in France are therefore likely to be disadvantaged compared to their international competitors. In addition to encouraging vertical integration, the tax also encourages companies to review their procurement strategies by switching to untaxed foreign suppliers (or encouraging their suppliers to relocate), which allows them to reduce the impact of C3S on their competitiveness.

30The importance of “cascading effects” is largely determined by the degree and nature of competition in the different markets. A company facing a very high level of competition in its end market will not be able to pass through taxes to price increases, which will tend to limit the cascading effect on downstream sectors. However, if this company no longer generates a profit, it may be forced to cease its activities. The mechanism acts symmetrically on the input market: a company with little bargaining power with its suppliers will be forced to accept price increases.

31An exercise to quantify the price effect of the C3S was carried out (box): it reveals that in all sectors, the price effects are higher than the effective tax rate because of the cascading effect, which amplifies the harmful effects of the tax throughout the production chain. In the industry, the price increase is about twice the effective tax rate, or about 0.19%. This price effect induced by the C3S could only be even more significant before 2015 and the introduction of successive lump-sum allowances for the calculation of this tax. The fact that the C3S currently only directly affects companies with a turnover above €19 million does not mean that SMEs are not indirectly affected by the cascading effect. This is all the more the case when they buy intermediate goods from large companies (paying C3S) which can more easily pass through the tax to their prices.

32It could be argued that, even if the multiplier effect of C3S is proven, the total impact should not be too high, as the tax rate is low. We shouldn’t be deluded by this argument. In a sector such as mass market retailing, margins are low and a price difference of around 0.3 to 0.5% is already very significant, which encourages distributors to develop their own brands or to use imports to escape the multiplying effect of the C3S.

C3S against business productivity, competitiveness and firm survival

33When the production process requires the use of intermediate goods, it is therefore in the interest of companies to replace their purchases from other companies with internal production, even if this choice is not optimal in terms of productivity. A turnover tax therefore encourages companies to integrate vertically in order to limit additional production costs but at the expense of the productivity of the economy as a whole: [25] companies do not optimise the use of factors and production methods, which makes them less efficient. Using the multi-sector model taking into account the inputoutput structure of Osotimehin and Popov (2019), [26] a rough estimate of the productivity loss due to the misallocation of production factors due to the C3S was made. The resulting permanent loss of productivity is about 0.01 to 0.02%. Taking into account the impact of lower productivity on capital accumulation, the loss of GDP is about 0.016 to 0.032%, or about 360 to 720 million euros. The cost may seem low, but it is related to a tax that brings in 3.8 billion euros, i.e. a dry loss for the economy of 10 to 20% of tax revenue, when considering productivity only.

Estimation of the “cascading effect” on the prices of the C3S

This box is based on a study by Claire Lelarge that quantifies the mechanism of price amplification by “cascading effect” of the Taxe sur le chiffre d’affaires (turnover tax) such as contribution sociale de solidarité des sociétés (C3S, corporate social solidarity contribution). [a] The analytical framework corresponds to a situation in which all companies operate in monopoly competitive markets and are able to pass through taxes and increases in production costs to their customers (the impact of taxes is fully borne by the buyer, whether a company or a final consumer). The quantification exercise is calibrated on the national accounts branch data for the year 2016, in particular the Input-Output table describing the intersectoral relationships of the economy. [b] This makes it possible to take into account the cascading effects by integrating into the analysis the production chains of the French economy. We obtain the cumulative impact of C3S on prices in the French industrial sectors (graph).
Several lessons can be drawn from this quantification exercise. First, recent changes in the tax scale, in particular the reduction in the tax base to 19 million euros of turnover since 2016, have led to a reduction in the effective tax rate well below the nominal rate of 0.16% in most sectors of activity. For the industrial sectors, only the energy, transport equipment (including automotive) and pharmaceutical sectors bear tax rates close to 0.16% because the companies operating there are very large. For manufacturing industry as a whole, the effective rate of C3S is estimated at 0.11%.
In all sectors, price effects are higher than the effective tax rate due to the cascading effect: the ratio between the two quantities is often close to 2 but varies from 1.2 to nearly 3 depending on the sector of activity and its inclusion in the input-output matrix of the French economy. In 2016, the average price effect in manufacturing industry was estimated at 0.19%. This price effect induced by the C3S was naturally even greater before the increase in the amount of the reduction on turnover for each company.
Sectors whose companies source their supplies largely from international markets have lower price effects. This is a direct consequence of the incentive to import intermediate goods described above. This is particularly the case in the coking and refining sectors. Conversely, other manufacturing industries, or the water and waste management sector, import fewer intermediate goods or services and are not able to limit the impact of the tax by increasing their imports. This is illustrated by comparing the total price effect of C3S in different sectors in the current situation with a hypothetical situation in which companies could not import their intermediate consumption: the difference between these two estimates roughly indicates the incentive given to each sector to import these intermediate goods (graph). Naturally, this illustration makes little sense when a large proportion of imported intermediate goods cannot be produced in France (as in the case of the extractive industries sector, for example), but it provides interesting insights in the other case, i.e. when domestic producers of intermediate goods and services for companies compete with foreign companies and suffer in a way from a “negative customs duty” via the C3S.

Estimated price effects generated by C3S in the industrial sector, in %

Estimated price effects generated by C3S in the industrial sector, in %

34The second problematic aspect of C3S is that it reduces the competitiveness of companies acting as an export tax and import subsidy. No other tax succeeds in this “feat”. Indeed, it is not limited to taxing turnover for final consumers, it is not deductible on export. The impact on production costs is reinforced by the cascading effect, especially since the production chain is long, involving successively several intermediate companies in the production process. From the quantification of the price effect of C3S, we can estimate the impact on French exports and imports. With a price effect of 0.19% for the manufacturing sector alone, and using an elasticity of exports at prices of about 5, [27] we come to the conclusion that French exports could increase by just under 1% in the absence of C3S, or about €4.2 billion. As for manufacturing imports of intermediate goods, estimates from Claire Lelarge’s work suggest that C3S increases them by about 500 million euros. The manufacturing trade balance (deficit of 33 billion euros in 2018) would therefore be reduced by about 14% thanks to the elimination of the C3S. This figure, which is the result of a theoretical ex ante estimate, is certainly an upper bound and must be taken with caution, but it suggests that the effects of C3S on the competitiveness of the manufacturing sector are far from negligible.

35We also empirically estimated the effect of C3S on exports using very detailed data from nearly 80,000 manufacturing companies over the period 2011-2016. [28] To do so, we compare companies that have benefited from the elimination (in 2015 for companies with a turnover of 3.25 million euros lower and in 2016 for companies with a turnover of less than 19 million euros) or a reduction in the effective rate of C3S (due to successive deductions from the same thresholds in 2015 and 2016 with those that have not benefited or less). The estimated coefficient of the negative impact of C3S on exports is quite stable and statistically very significant in the various estimations that have been made. [29] Thus, the companies that benefited from the elimination of C3S successively in 2015 and 2016 saw their exports increase by about 1% more than those that did not. This effect is quantitatively important and above all consistent with the price elasticity of exports reported in the literature (see above). The positive effect on exports occurs quickly, suggesting that companies have incorporated the cost reduction induced by the elimination of C3S into their prices. This study is still preliminary but it suggests that C3S does act as a tax on French exports.

36Since C3S taxes companies at the top of the operating account, it increases the break-even point necessary for the company’s survival. [30] With the same identification strategy, we compared the probability of survival of companies just below and above the threshold before and after the reform. We find that the probability of a company disappearing within two years has been significantly increased due to C3S, especially in 2009, from around 9% to 10%. [31] This weakening of companies exists but less strongly in the years after the crisis. These results suggest that taxing turnover can negatively affect the probability of survival of companies, especially in times of crisis.

Completely remove the C3S

37We therefore consider that the C3S is nowadays the most harmful tax for companies. The removal of C3S in three stages was planned as part of the Responsibility Pact. While the first two stages were carried out in 2015 and 2016, leading to the current scale of 0.16% of annual turnover exceeding 19 million euros, the final abolition planned for 2017 did not take place. We propose to complete the reform and abolish this tax completely. The consequence of starting the phasing out of C3S gradually, starting with rebates, and thus completely eliminating the tax for companies with low turnover (excluding indirect effects passed on by companies that are still subject to it), is that the last phase of phasing out would directly benefit large SMEs, mid-cap companies or large companies. This should not be an obstacle to this abolition. Indeed, the cascading effects generated by the C3S mean that this tax spreads to the entire productive fabric, including small SMEs that are not liable for this tax. Smaller companies are all the more likely to bear the weight of C3S if they import little and face suppliers or service providers with greater market power (e.g. financial sector): in the end, all companies, regardless of their size, will benefit from the complete abolition of C3S.

38In view of the severe constraints on our public finances, the implementation of this reform can be combined with another recommendation set out in a previous CAE Note which proposed to review contribution relief for higher salaries, targeting in particular the extension of the 1.8 point reduction in the Pacte de responsabilité (Responsibility Pact) in 2016 (PR16) [32] from 1.6 to 3.5 SMIC. The C3S now generates €3.8 billion (PLFSS 2019) while the PR16 costs between €4 and €4.5 billion: the PR16 could be reconfigured in such a way that the impact of the abolition of the C3S is neutral overall on public finances and for companies. Such a reform would make winners and losers in apparent terms, i.e. assuming that prices do not adjust to the new tax structure. However, we believe that all companies, regardless of their size and sector of activity, would benefit in terms of efficiency and lower input costs if we take into account the tax incidence phenomena: the impact of the abolition of C3S would spread to the entire productive sector by lowering input prices through intercompany trade. The advantage for large companies would be partly offset by the reduction in social security contributions, which would result in an increase in the cost of skilled labour. The reduction in social security contributions is interpreted as a reduction in a tax on trade between companies and households. It is justified in cases of clear underemployment of the labour force which is not proven for skilled work (the unemployment rate for professionals is 3.8%). To this extent, the result of Diamond and Mirrlees (1971, op. cit.), which assumes full use of the factors of production, then retains its full force, and we propose, therefore, to substitute a tax on trade within the productive bloc by a tax on trade between households and companies. Another approach in line with the fact that the least distorting taxes concern endusers would be to remove certain VAT tax loopholes such as VAT in the catering industry. [33] The increase in VAT rates is often criticised because of its anti-redistributive aspects, but as highlighted by a report by the Conseil des prélèvements obligatoires, [34] this is not the case for catering, which weighs more heavily on the budgets of wealthier households.

39Recommendation 1. Finalise the abolition of the corporate social solidarity contribution (C3S), which was interrupted in 2017.

A less neutral tax than it seems

41However, the fact that the rate applicable on VA is progressive according to the turnover, makes this tax de facto dependent on the turnover which exposes it to distorting effects, even if these effects are more complex and certainly less strong than those identified with the C3S. We can illustrate the distortions potentially generated by the great heterogeneity of the CVAEs actually paid as a function of the VA. For example, for companies in the 9^th decile of VA, 25% of companies have a rate below 0.7%, and 25% a rate above 1.4%. [36] Thus, at similar VA, the CVAE rate can be doubled. This heterogeneity results from the way in which the tax is calculated without deliberate action by the public authorities. It should also be noted that the methods used to calculate the CVAE have paved the way for tax optimization strategies and, consequently, for distortions, particularly because of the complexity of the legal concept of tax VA. For all these reasons, it would be simpler, more efficient and in accordance with the very principle of this tax, either to apply only a single rate or to introduce a progressive scale based on VA and not on turnover.

42A simple adjustment of the rules for calculating the CVAE would go in the right direction but would not solve the problem of input taxation insofar as the CVAE is based on gross VA: intermediate purchases are indeed deducted from the taxable base but this is not the case for other purchases from suppliers of capital goods. This again exposes taxable companies to the risk of cascading taxation, which could encourage them to produce their own capital (e.g. a patent) rather than buying it from another company. Above all, the CVAE has the major disadvantage of having as its base an operating balance well in advance of the company’s net income, unrelated to its profitability. It can be assimilated to two taxes applying, with the same rate, on the one hand, to the wage bill and, on the other hand, to the gross operating surplus (EBITDA). By taxing EBITDA and therefore depreciation, [37] this tax directly affects companies’ investment capacities, much more so than corporate income tax. Indeed, with the CVAE, companies cannot deduct from the taxable base expenses economically linked to production (consumption of fixed capital), which penalises in particular those who need to renew their production equipment regularly. The CVAE thus leads to distorting the profitability of investments according to sectors by concentrating its impact on the most capitalintensive sectors. Finally, the CVAE carries the risk of distorting companies’ productive allocation choices to the detriment of capital, with a negative effect on productivity. Our reasoning is always based on the same theoretical basis. The CVAE taxes two thirds of the labour factor and one third of the capital factor. In a situation of full employment –which seems to be the case for skilled jobs– taxation of the capital factor that taxes trade within the productive bloc is more harmful than taxation of the labour factor.

Schedule the deletion of the CVAE

43The abolition of the CVAE, without being replaced by a new tax, would have two main advantages the elimination of the distortions described above and a major simplification of company taxation since companies would no longer have a specific declaration to pay the CVAE. This declaration is, as a matter of fact, far from simple for companies (many exemptions, rebates…). Another important advantage is that the abolition of the CVAE, more than a further reduction in the corporate tax rate beyond those already planned by 2022, would benefit to a large proportion of companies, both profitable and distressed, as well as companies in the expansion phase which, for example, invest heavily but still make little profit. We are aware of the negative impact of profit taxation on investment and innovation in particular as demonstrated by recent empirical studies, but we consider it preferable to shift taxes on production –especially when they affect EBITDA– to profits. We also have in mind the importance of the nominal CIT rate on the attractiveness of France. However, this point must be put into perspective: while the corporate income tax is the most well-known tax, investors are not unaware of the existence of other taxes weighing on companies in France, duly listed by most consulting firms. [38]

44Naturally, the abolition of the CVAE implies finding alternative revenues for local authorities. To this end, we are not advocating the creation of a new local business tax. While land can constitute a good territorialized base for a local tax, taxes on production are to be prohibited and corporate income tax –which is very sensitive to changes in the economic cycle– [39] generates revenues that would be far too volatile for local authorities. It should be noted here that the instability of CVAE’s revenues is already regularly pointed out by local elected officials, an instability that is very much due to the territorialized nature of the base. At this stage of the reasoning, it seems to us that two subjects of a different nature must in fact be decoupled: on the one hand, the reform of company taxation (and its financing) and, on the other hand, the distribution of tax revenues to local authorities.

45To allocate substitute revenues for local authorities deprived of CVAE, we recommend allocating a fraction of a national tax such as VAT, much less volatile than the CIT, with the same distribution key as for CVAE (two-thirds of the establishments’ staff and one-third of the land) but calculated in an aggregated manner at the municipal or intermunicipal level and no longer undertaken by company. With such a distribution method, local authorities would not be directly interested in the results of companies (or as currently in the VA produced by companies present on its territory) but a link with economic activity would be maintained, insofar as the arrival or departure of a company, or the fact that it hires or dismisses, would have an impact on the product actually received.

46Recommendation 2. Schedule the abolition of the contribution on the added value of companies (CVAE) and, in return, the allocation to local authorities of a fraction of national tax with the same distribution key (employment and local land) calculated at the level of the beneficiary local authorities.

47What avenues should be considered if the State’s financial margins are not sufficient to support the impact of the abolition of the CVAE? First, it should be pointed out that the disappearance of this tax would increase the tax base of the corporate income tax by the same amount and, assuming an average rate of around 20%, this would represent a return of CIT of €2.6 billion. [40] Second, the best guide to increasing revenues is to use the hierarchy of taxes that can be established by focusing on the least harmful sources of financing. The most neutral and therefore least negative tax on business productivity is VAT. However, in the current context, financing the abolition of a tax on businesses (without prejudging the impact of the CVAE on consumer prices) through an increase in a tax directly imposed on the final consumer seems difficult, especially since it is also important to control the redistributive aspects of such an operation. However, the current differentiation of VAT rates is not optimal and should encourage the questioning of certain inefficient tax loopholes (e.g. catering). With regard to the taxation of labour input, the charges on skilled labour are the least harmful because it is in a situation of full employment. In this respect, if the negative assessments of the effects of reductions in social security contributions above 1.6 SMIC were to be confirmed, both on employment and competitiveness, budgetary room for manœuvre could be created by refocusing the reductions in social security contributions resulting from the switchover of the CICE on low wages, in a degressive manner in order to avoid any threshold effect. Finally, an increase in CIT revenues could be achieved by broadening the base, in particular by more effective taxation of multinationals (a subject on which international negotiations are starting within the OECD framework and on which the CAE will soon make proposals) and by reducing the various tax loopholes.

48Our proposals are based on the observation that taxes on production are among the most harmful taxes to productivity and competitiveness. The elimination of the C3S must be the priority because of its particularly harmful “cascading effect”. Next, the removal of the CVAE must be programmed. These proposals are also based on a simple strategy: eliminating the most harmful taxes on companies, and financing this movement by reducing the least effective tax expenditures and taxing corporate profits less distortingly. Beyond the gains in productivity and competitiveness that we expect from such a reform, it would also help to reverse a trend that is deeply rooted in our country, which consists in giving with one hand and taking away with the other.