Gross distributive effects of the carbon tax

9When examining the impacts on purchasing power of the carbon tax, it is necessary to distinguish between the overall impact, considering all economic agents as a whole, and the impact on the different categories of households. The overall impact corresponds to the cost of emission reduction efforts that are needed to achieve a given objective. These costs are known as “abatement costs”. In this respect, a tax that makes carbon emissions costly for economic agents is similar to a “carbon price signal” It provides an incentive to mobilize abatement efforts by focusing on the least costly. The impact of the policy implemented on the different types of household results from the way in which the abatement efforts are distributed among the population. This in turn raises the question of the transfers associated with the tax, i.e. how its revenue is used.

10To address this issue, we first quantify the gross distributive impacts of the carbon tax, before behavior changes and before transfers associated to the the revenue of the tax. We are carrying out this exercise in the case of the French carbon tax, the contribution climat-énergie (CCE, climate-energy contribution). [4] The CCE is an excise duty applied to the CO₂ content of energy. [5] It was introduced in 2014 at the initial rate of €7 per ton of CO₂ and applies to the price of goods before VAT. It is part of the existing domestic consumption taxes on energy products: TICPE and TICGN. [6] With a share of around 60% in households and 40% in businesses, the CCE generated revenue of €6.4 billion in 2017. In 2017, the government announced an increase in the tax rate in order to reach 86.2 euros/tCO₂ in 2022. For 2018, the level of the CCE has thus increased from 39 to 44.6 euros/tCO₂, the level currently in force. This increase to 55 euros and the continuation of the catch-up of diesel-gasoline were stopped by the government in November 2018 following the “yellow vests” movement.

11Let us now examine the distributive impacts of the reform consisting of an increase in the CCE from its current level (44.6 euros/tCO₂) to the level expected to be reached in 2022 (86.2 euros/tCO₂), accompanied by the catching-up of the diesel tax by 7.8 cents per liter. This corresponds to the sum of the increases initially planned for January 2019, 2020 and 2021. This reform is first considered without any revenue redistribution mechanism, in order to establish a solid basis for proposing a more equitable system.

12The impacts of the reform on social equity are evaluated for each standard of living decile. [7] Figure 1 illustrates the burden of the reform in household disposable income (the supplement of tax to income level of households), broken down into three parts: additional expenditure on housing due to the CCE, transport due to the CCE and transport due to diesel-fuel catch-up. The poorer the households, the higher their burden: nearly 1% of disposable income for the first decile compared to 0.3% for the top decile. Thus, the reform is regressive ex ante. The effect caused by housing is more regressive than that caused by transport. The share in the cost ratio of diesel-gasoline catching up is far from negligible, as this catching up also has a regressive effect. This regressivity, which reflects the structure of emissions and the weight of energy expenditure in the respective household budgets, requires accompanying mechanisms to ensure the equity of the reform and its political acceptability. This is what has been lacking in the establishment of the CCE.

1

Effort rate of households associated with the reform before revenue use, in %

Effort rate of households associated with the reform before revenue use, in %

13Simulations carried out with two different methodologies and data [8] show that the unequal impacts of the taxation is not only due to the level of income (vertical heterogeneity).Within each income decile, the heterogeneity (called horizontal) of the taxon purchasing power is significant. Among the households belonging to the first decile, 10% of households lose more than 220 euros in purchasing power per year and per consumption unit with the implementation of the reform, while about 10% of households are not affected at all. Horizontal heterogeneity increases with income, so that in the 10th decile, 10% of households lose more than 480 euros in purchasing power per year and per consumption unit, while 10% of households lose less than 25 euros. This intra-decile heterogeneity is due in particular to the location of households and the type of equipment they own: oil or gas heating vs. electric heating, diesel engine vs. petrol. For example, for a given income, the relative loss of rural households compared to households in the Paris agglomeration is significant, around 130 euros on average per consumption unit. But a more in-depth analysis shows that horizontal heterogeneity is better explained by the equipment than by the geographical location alone. Once the differences in equipment are taken into account, the differences in losses between rural and urban households vanish: with equal income and equipment, a rural household loses 20 euros more than a Parisian one, per consumption unit. On the other hand, for the same income and type of location, a household running on diesel loses 230 euros more per consumption unit over one year than a household without a car, while a household heating on domestic fuel loses 157 euros more than an household heating on electricity. [9]

14Based on these initial analyses, the redistribution of carbon tax revenues seems necessary to better share the efforts of households serving climate objectives, but its implementation raises a series of difficulties. It is relatively easy to redistribute revenues according to income, but much more difficult to take into account the dimensions of horizontal heterogeneity in order to properly identify losers without creating overly complex arrangements, significant windfall effects or a weakening of the price signal. The previous exercise suggested calibrating the redistribution of revenues according to equipment. But it would be both complex regarding the information to be collected and environmentally counterproductive: the loss of transfers when an equipment is changed would indeed constitute a strong disincentive to make this change. Keeping transfers when equipment is changed would solve the incentive problem but would create a new form of injustice: why should a household that already made an effort for the environment before the introduction of a tax, by choosing less polluting equipment, receive less transfers than its alter ego who cared less about the environment and has just brought about the change? Another –still imperfect– option, would be to subsidize equipment changes. This makes it possible to target some losers of the reform (those with polluting equipment) and does not decrease the incentive. But this mechanism also has several disadvantages: it can create significant windfall effects, it can be regressive if it is poorly calibrated and, above all, it does not compensate all the losers. A last possibility is to base the redistribution of revenues on geographical location, since the previous year showed us that it is correlated to equipment, although not perfectly.

15Our conviction is that to restore an increasing carbon tax trajectory, this tax must be fair and therefore coupled with a redistribution of the revenue. The political economy constraint we impose on ourselves in the following exercise is that the combination of carbon tax and revenue redistribution should reduce to a minimum the number of losing households in the first five deciles. We explore different scenarios by taking into account two crucial dimensions: household income and location.

Application to the project frozen in fall 2018

16Five scenarios for the tax redistribution are under study:

• a neutral flat-rate transfer per consumption unit (for all households);
• Terra Nova’s [10] proposal to transfer 500 euros to households in the first income decile, then 400 euros in the second, 300 euros in the third, etc. until the fifth;
• an increased energy voucher (three times the amounts allocated in 2019) benefiting the 5.6 million currently eligible households;
• a redistribution of 30% of revenues through lump-sum transfers, and 70% through transfers proportional to income per consumption unit;
• a redistribution with transfers based both on income and geographical criteria.

17By simulating these scenarios, we try to identify which one would minimize losses for households in the first five deciles.

18Scenario 1. A neutral flat-rate redistribution per consumption unit makes it possible to eliminate the regressiveness of the reform because the poorest households consume less carbon than richer households. Households in the 1st decile earn on average 60 euros as a result of the reform; it is more or less neutral for households in the middle of the distribution (deciles 5, 6 and 7); finally, households in the 10th decile lose on average 80 euros. But there is still a significant horizontal heterogeneity: among the first deciles, 10% of households receive more than 150 euros of transfer per consumption unit, while 10% lose more than 80 euros per consumption unit.

19Scenario 2. Terra Nova’s proposal does not exhaust all revenues (remaining €2.3 billion). Its structure obviously makes it more progressive than the pure flat-rate redistribution of scenario 1. On average, households in the 1st decile earn 250 euros as a result of the reform. The effect is neutral for households in the 4th decile, while losses appear for the 5^th decile. Households in the 6th decile lose 150 euros per year on average. 10% of households in the 1st decile earn more than 380 euros per consumption unit but more than 25% of the households loose in the 3^rd decile (Figure 2a).

2

Distribution of net transfers, by standard of living decile, in euros

Distribution of net transfers, by standard of living decile, in euros

20Scenario 3. An increased energy voucher fails to target mobility problems and there are still many losers, especially among the first deciles. [11] This is mainly due to the current restrictive eligibility conditions for the energy voucher.

21Scenario 4. A redistribution of 30% of income in flat-rate transfers, and 70% in transfers proportional to income per consumption unit (semi-proportional redistribution) leads to a zero average impact according to income (there is therefore neither progressivity nor regressivity), except for the 10^th decile which earns on average 60 euros per year, this being mainly due to the gains provided to the very high incomes by a transfer proportional to income. The heterogeneity within each decile remains important.

22Scenario 5. The previous results show that the most favorable scenario for the poorest households is the one proposed by Terra Nova. However, it does not exhaust all revenues and does not sufficiently compensate households in the middle of the distribution. We are now studying a final scenario that attempts to address these issues. This scenario uses the principle of redistributive transfers, the amount of transfers being chosen in order to ensure progressiveness while avoiding excessive losses for households in the middle of the distribution (310 euros for the first three deciles, 300 for the 4^th, 255 for the 5^th, 240 for the 6^th, 150 for the 7^th and 60 for the 8^th).Transfers are both decreasing in income and based on geographical criteria. [12] Although imperfect, the geographical tag makes it possible, in principle, to take partly into account the possibilities of substitution available to individuals: urban areas are, for example, more equipped with public transport than rural areas which are more constrained in their choices. This scenario (compared to the others) is the one that minimizes losses for the first five deciles (Figure 2b), but mainly because it gradually redistributes the entire revenue.

23Concretely, transfers linked to carbon tax revenues would be a new simplified chèque énergie, whose eligibility conditions would be determined as described in scenario 5. In addition, there would be no mandatory allocation of the chèque anymore, like it is now as a voucher to energy expenditure.

24Four main lessons must be learned from these simulations:

• Building an equitable reform requires a comprehensive approach, first dealing with the regressiveness of the gross reform. It is therefore different from redesigning existing systems (energy vouchers, vehicle conversion subsidies, tax credits);
• it is possible, through an appropriate redistribution mechanism, to ensure that more than 90% of households in the first 5 or 6 deciles benefit from the implementation of the reform;
• it is very difficult, however, to erase horizontal heterogeneity. A differentiation of transfers by geographical criteria is a first approach to correct this horizontal heterogeneity. However, it is very imperfect, so that more detailed criteria accounting for local and public transport constraints would be desirable, but more complex to set up: [13]
• conversion aids are therefore necessary. However, they must be means-tested (see below).

25Recommendation 1. Redistribute all new carbon tax revenues with decreasing transfers based on income and geographical criteria. In addition, set up mechanisms to help the change of the most polluting equipment, targeted on the most vulnerable households.

Sharing the decarbonation effort

29Considering that the instruments put in place should absolutely be fair and perceived as such, we do not support the the strategy of the double dividend. A first reason is that this option was more or less explicitly chosen by the French government since the establishment of the CCE in 2014, and that it clearly did not convince (probably because of a lack of explanation). The second reason is that the possible gain to be expected from this option, in relation to the return of revenues, seems to us to be a second order benefit. In addition, equity and social justice issues would remain, and could even be exacerbated, unless a tax to be lowered is found that is paid more by the losers of the carbon tax. Finally, we do not think it is appropriate to wait for a thorough review of our taxation system to strengthen our climate policy.

30In contrast, we believe that broadening the tax base is advisable. In order to take into account equity concerns in the implementation of a green tax, our preferred option is to reactivate carbon pricing with a clear upward trend in the tax, but on a larger base to ensure that households are not the only ones making efforts. Indeed, what would be the point of taxing an ever-increasing narrow base of emissions when a significant proportion, with significant CO₂ emissions, would not be mobilized? It seems preferable to include all fossil products leaving refineries in the scope of the tax before “colouring” fuels according to their end use. This should go together with mechanisms (similar to the ones detailed above for households) that would redistribute revenue (in the form of refunds of revenue in proportion to activity) to the newly concerned sectors in order not to penalize their competitiveness.

31Currently, TICPE exemptions mainly concern the transport sector, with two main measures: the TICPE exemption for fuels used in commercial aviation (€3.6 billion in 2019), and its partial reimbursement to road hauliers (€1.5 billion in 2019). Excluding transport, the main measure is the reduced rate for farmers (€0.9 billion in 2019). [17] The point here is not to remove all TICPE exemptions since some TICPE exemptions are justified by the fact that neither agriculture nor the building industry use road infrastructure (contrary to road freight transport). We instead propose to shift the generating event of the carbon component, to target the pollution that will be induced whatever the reasons.

32Since about 15% of energy use emissions (excluding the sector covered by the European allowance market) are exempt, this effective widening of the tax base could allow a slightly lower growth rate of the tax than initially planned, while maintaining the same overall emission reductions.

33A key monitoring indicator is the “carbon pricing gap indicator” proposed by the OECD, which measures how far the pricing of all emissions is from a uniform carbon reference price for all emissions. This indicator captures, for example, how the increase in the CCE has brought us closer to countries with the most ambitious carbon pricing policies, while pointing out the exemptions that remain. [18] Figure 3 measures the carbon pricing gap [19] compared to an objective of 60 euros/ tCO₂, the OECD benchmark, which is a reasonable level if the trajectory is regularly raised to be in line with a price level of 100 euros/t in 2030. [20]

3

Carbon pricing gap in France compared to the target of 60 euros per ton of CO₂

Carbon pricing gap in France compared to the target of 60 euros per ton of CO₂

Reading: For each year, the pricing deficit compared to the target of 60 euros/ton of CO₂ corresponds to the area between the curve and the target (dotted line). It is 52% in 2018 and 69% in 2017.

34There is thus a gradual reduction in this carbon pricing gap thanks to the increase in the carbon component and, at the end of the period, to the reform of the functioning of the European carbon market. However, in absolute terms, the price of the European quota remains low and the effective price per ton of carbon remains lower than the target set by the CCE. If the exempt sectors were to be effectively integrated into the carbon tax, the larger base would ease the upward trajectory of the price per ton of CO₂. By 2022, the target price would thus be 20% lower for the same average effective carbon price.

35Recommendation 2. Continue to increase the climate-energy contribution and explain its issues and mechanisms to the public. Broaden the base by uniformly applying the tax at the refinery exit level, with no exemption nor refund mechanism. Use the new revenues to support the affected sectors in order to preserve their competitiveness.

36Shifting the point of application of the tax in order to broaden its base is required so the households would not to be the only ones supporting the burden. A complementary measure would be the introduction of a European carbon floor price to reduce the distortion between the price of allowances on the EU-ETS market and the price of diffuse emissions. [21] It would also allow to remove conflicting injunctions on the electricity sector, that could in the short term call power plants in order of merit, and in the medium term choose new equipment. [22] Environmental diplomacy must work to implement this floor price.

37With regard to the legitimacy of the effort required on French households and businesses in an uncooperative world, it may be recalled that the objectives set for 2030 reflect the European desire to be a driving force in the construction of climate cooperation and its pivotal role in what has already been undertaken, without setting an excessive example. The (global) carbon budget compatible with the 2°C target being much stricter than fossil resources accessible in the subsoil, the implementation of an increasing carbon price, in addition to market prices, is necessary. Thus, environmental diplomacy must propose a convincing plan to strengthen the ambition of international efforts, particularly those of major emitting countries, to strengthen the legitimacy of national policies. [23]

38Recommendation 3. At European level, align the EU-ETS market price of allowances with the levels targeted for domestic emissions, complementing the existing scheme with a floor price.

39Beyond the new mechanisms to support the increase in the carbon price, it is necessary to build their legitimacy, by better informing the public about the effectiveness of climate policies and by gradually establishing a legal framework so as to give credibility to the respect of the announced redistribution rule. [24] To do this, it is important that this rule precedes the definition of the price path. Under these circumstances, the agenda for restarting the process would be, in parallel with the establishment of this new legal framework, the definition, after broad consultation, of the basic rule for the use of the revenue. [25]

40To make credible the environmental commitment of its tax and its only objective to provide incentive, Switzerland uses a mechanism whereby scheduled increases are automatically applied if emission targets are not met. On the other hand, increases are postponed if they are exceeded. In this way, the incentive nature of the tax is claimed and credibility is lent to it since the State will not have any additional revenue when its objectives are achieved. Beyond that, each citizen understands that, if everyone’s efforts are strengthened to achieve the desired objective, taxes will be reduced. One of the tasks of the new Haut Conseil du climat (High Climate Council) could be to give an opinion on the increases in CO₂ taxes (for the following year) according to the observed CO₂ emissions (for example, over the last two years by smoothing weather-related hazards).

41Recommendation 4. Establish the rules governing the use of the revenue before the price trajectory, accompanied by broad communication to the public to make it a shared project. Introduce a mechanism to freeze the trend if emission reduction targets are exceeded.

Adjust to oil price volatility

42The yellow vest crisis began with a surge in fuel prices due to their market component, not taxes. [26] This points to a recurring problem: the lack of mechanisms to manage fuel price volatility on the purchasing power of some highly exposed households, leading to the same crises each time fuel prices soar. It led to the reignition of an idea that had already been experimented (without success) between 2000 and 2002: the floating carbon tax. The justification for this system is that oil demand does not depend solely on the carbon tax, but on the total price paid by the consumer, the sum of the world oil price and the tax. Under these circumstances, aiming for a total price trajectory that makes it possible to achieve the environmental objective by adjusting the tax to fluctuations in the price of oil may seem attractive. But such a strategy is not optimal because it is not cooperative at the international level. If it were adopted by all consumer countries, it would encourage producer countries to raise prices. At national level, it would probably give distributors an incentive to increase their margins. Finally, the experience of 2000-2002 shows that it is very easy politically to lower the tax when the producer price rises, but very difficult to carry out the symmetrical operation when the price falls. [27] This asymmetry is well documented by the post-mid-2014 sequence: the carbon tax did not compensate for the drop in oil prices at all, and its increase was problematic as soon as the price rose, even though the final price remained close to the 2012 level.

43To address the issue of pre-tax price volatility, mechanisms should be considered to reduce prices peaks in order to limit the difficulties of households that are indeed increasingly constrained by pre-committed expenditure and therefore vulnerable vis-à-vis unforeseen situations. In this respect, 2018 has combined the two problems: first, a substantial increase in the carbon component and the price of diesel fuel and second, a rapid rise in the price of oil. The implementation of the revenue restitution mechanisms discussed above is in itself a response to the first problem. Such an approach is, indeed, in line with the general idea that it is through the use of carbon tax revenues that the redistributive problems of the carbon tax must be addressed, not by weakening its rate, and that redistribution must be appropriately targeted. To respond to the second problem, two complementary insurance mechanisms for households can be considered.

44The first would legally authorize the government to adjust, by order, the TICPE rate collected during the year. For example, it would allow the increase in TICPE (carbon component and diesel/gasoline convergence) as stated in the law to be reduced by half for a maximum of three months. This modulation would be possible in the event of an increase in the price per barrel (in euros and excluding distribution costs) of more than 10% compared to the average of the previous four quarters. After three months, the legal rate would be restored. This mechanism could be activated at any time of the year, for example, after public notice from the Haut Conseil du climat (High Climate Council). The corresponding mechanism must operate symmetrically, in the event of an unexpected sudden drop, which would thus make it possible to advance half of the expected increase in TICPE for the following year. This modulation would not aim to erase all variations in the price of oil but to smooth them out.

45The second would be an optional subsidized mechanism to insure the most vulnerable households, due to their income and geographical location without access to public transport, against peaks of oil price. It could take the following form. In exchange for the payment of a premium partially paid by the State for vulnerable households, the “insurer” pays the price above a threshold of “current price + x %” for a certain number of months (for a fixed quota); the State also acts as a regulator to ensure the emergence of a supply of this type of product for households.

46Recommendation 5. To protect households from extreme oil price fluctuations, temporarily adjust the path of increase of the TICPE or propose a specific coverage mechanism for the most vulnerable households.

Reducing diffuse emissions from construction and transport

47Existing environmental policy uses many instruments other than carbon taxes: technical standards, including for vehicles and buildings, subsidies for clean technologies and renewable energy, energy efficiency programs… However, these instruments are less effective than carbon pricing. Indeed, standard setting freezes technological choices while competition should be stimulated to be “greener”. Subsidies also allow socially unproductive structures to be maintained when new sectors emerge. They are also subject to a very strong tension between two objectives. On the one hand, targeting to avoid deadweight effects and contain public expenditure. And on the other hand, avoiding distortions resulting from too narrow targeting that breaks the neutrality between equivalent technologies, which often generates undue rents for some agents. The cost per ton of carbon avoided by these policies can therefore be very high.

48French greenhouse gas emissions have decreased from 550 MtCO₂eq in 1990 to 470 today. Most of the reduction was achieved between 2005 and 2014, the most recent period being even marked by a reversal of the trend, in a context of low oil prices. This 80 MtCO₂eq reduction, although insufficient to meet the targets set in 2003, [28] was achieved without any carbon pricing other than the pre-existing fuel pricing. This suggests that substantial emission reductions are possible at moderate costs once the process is started, with the switch from coal to electricity production and the reform of the most polluting industrial processes. Then, this dynamic can continue by switching from different uses to electricity, whose production must be totally carbon-free.

49It is this type of strategy, mobilizing energy savings from installations generating significant emissions in relation to total emissions (point sources), that has been implemented in the past. Emission reductions were mainly achieved in the industrial and energy sectors (Figure 4), with the specificity, however, that French electricity production was already very carbon-free in 1990: 107 grams/kWh, now reduced to 52 for an average level in the European Union of 331, Germany being 485. In addition, industry emissions have been mechanically reduced by the evolution of the French productive system, but with the counterpart of an increase in emissions linked to imports. Thus, while the emissions physically emitted on French territory were decreasing, the carbon footprint, which measures emissions induced by domestic demand, continued to grow. [29]

4

Greenhouse gas emissions by sector in MtCO₂eq

Greenhouse gas emissions by sector in MtCO₂eq

50Diffuse emissions, from households or all companies, which now represent three quarters of French emissions, have remained stable since 1990, with slight decreases in the residential-tertiary sector and agriculture, being offset by increases in transport emissions and, more marginally, in the waste sector. It can only be noted that traditional instruments have not achieved the objectives that had been set for transport and building renovation.

51Yet, regarding their ability to achieve the objectives, the detailed targets by sub-sector, including the number of renovations or new emission standards for vehicles and buildings seem more reassuring than carbon pricing. Can these policies be made more effective than in the past and at what cost?

52First, it is important to identify more rigorously the conditions for the success of these traditional policies. The case of the “bonus-malus” policy in the car industry provides a good illustration of this. The monetary incentives introduced by this scheme have undoubtedly enabled a break in the trend in unit emissions from the vehicle fleet. But the system has favoured diesel over petrol, encouraging an increase in local pollution. And it has been shown that even by restricting the analysis on the environmental dimension, the system has not worked satisfactorily: the rebound effect and the resulting expansion of the fleet have ultimately led to an increase in carbon emissions. [30]

Which sacrifices for which emission reductions?

53Simulations carried out following the Quinet Commission with economic models incorporating demand closures (so-called “sectoral macroeconomic models”) show that our total emissions could be reduced to 300 MtCO₂eq in 2030 if a uniform carbon price of around 165 euros/tCO₂ were applied by 2030. [31] Insofar as this price sets the upper bound of the marginal cost of the efforts to be made, the average cost per ton of carbon avoided to achieve this objective would be in the order of 93 euros/tCO₂.

54The assessments carried out by the independent consulting company Carbone 4 [32] indirectly provide information on the cost of the alternative strategy. The level of reduction could be achieved by a few selected measures, such as: bunch of work combining thermal insulation of the envelope and substitution of end-of-life fossil fuel heating equipment with efficient low-carbon equipment in energy-intensive housing; electrification of vehicles. The choice of these measures is the result of an optimization, which emphasizes the priority to be given to renovation over the reinforcement of the performance of new buildings, the latter being twicemore costly. Despite this, their cost would be around 200 euros per ton of CO₂ avoided in the building and between 210 and 270 for electrification of vehicles, as long as the drop in battery costs remains a promise.

55We have a ratio of 1 to 2, or even much more, for the sacrifice of purchasing power to be made to ensure a certain level of decarbonation depending on whether the price signal is used or whether we rely on more traditional instruments. As mentioned in the Quinet Report, many existing decarbonation policies are in fact much more expensive than benchmarks that are considered difficult to accept when they are disclosed in full light to calibrate a carbon price signal.

What kind of instrument mix?

56However, the analysis cannot be limited to asserting the superiority of eco-taxes in terms of cost-effectiveness, as political economy considerations are crucial for the implementation of climate policies. The two strategies to consider are therefore not, on the one hand, “pure” pricing and, on the other hand, regulation or subsidies for green investments. Instead, we recommend a new mix of instruments, in which carbon pricing is central. [33]

57Indeed, the carbon tax stimulates green innovation and creates a business model for green investments. [34] However, supporting innovation and green investment using public funds from the general budget at the same time is justified and necessary for several reasons we discuss below. First, this policy is complementary to the taxation policy. There is no reason to link the financing of green investments to carbon tax revenues. Indeed, if a substantial share of carbon tax revenues were directed towards this financing, it would then become almost as much a return-oriented instrument as an instrument serving the sole purpose of guiding private behaviour. It might change its nature because there is no reason that the tax revenue and the financing needs of green investments should coincide. Government spending in this area should be guided primarily by the shadow value of carbon: [35] if the cost per ton of carbon avoided through investment is below a threshold defined by the shadow value, then the investment is socially profitable and must be made.

58What are the arguments in favour of subsidies for green equipment? First, it is rational to temporarily subsidize green research because it is easier to innovate in sectors where research and knowledge have already been accumulated. [36] Directing research and innovation towards breakthrough technologies with higher potential then requires additional support. Subsidies for green investment can also be justified by the existence of learning effects in the production of new technologies or infant industries. For example, prices of solar photovoltaic modules have fallen by about 80% since the end of 2009. This sharp drop in prices can be attributed to learning effects. However, the latter are difficult to anticipate: for example, there is no such decline in offshore wind energy. There may also be significant learning effects on the demand side in the use of new technologies. The use of a shared transport service, for example, can be made easier by the emergence of new business models and user feedback. For many new ideas, small-scale experimentation seems to be essential before triggering large-scale investments.

59Secondly, a given carbon price can avoid even more emissions if green substitutes are accessible and affordable, allowing households to change their behaviour. If these green substitutes are public goods (e.g. public transport infrastructure) or if there are network externalities (density of charging stations for electric vehicles), it is justified to subsidize them. From this perspective, green public investments and carbon taxes appear to be complementary, and in the timing of climate policy it would be justified to carry out the former before implementing the latter. [37]

60Moreover, if the announcements of future tax developments are not considered sufficiently credible by households, there is a risk of private under-investment in green equipment. It is also possible that households may overestimate the future benefits and costs of investment. [38]

61Finally, in addition to ecological taxation, subsidies or grants for equipment changes are justified for households that are too financially constrained and do not have access to credit. To avoid deadweight effects, it is important to target assistance to the most vulnerable households (see recommendation 1) and to direct it towards effective solutions. In this respect, one should not be misled by the fact that regulations and subsidies for green investments are generally better accepted than the carbon tax. Indeed, the link between these conventional policies and the generally delayed increase in electricity or energy bills is not always clear to the general public. However, the costs are very real, not unrelated to the tax bustle, with regressive effects even if little perceived by the public. [39] Indeed, conventional policies combining regulations and subsidies often benefit to the wealthier households: they have a greater interest in changing equipment (or can do it more easily), so they benefit more from subsidies (particularly those that do not depend on their income) and the net cost of upgrading is lower for them. Individual aid must therefore decrease with income, as is the case with some vehicle conversion premium schemes or boiler replacement premium schemes. On the other hand, the energy transition tax credit, by encouraging work to be carried out to standards higher than those of the regulations, equipment by equipment, but without any minimum requirement as to overall progress made or household income conditions, would deserve a serious revision. [40]

62More generally, it is important to carry out a systematic ex ante and ex post evaluation of the effectiveness of any grant and regulation, which should be independently certified in the same way as the counter-expertise requested by the Office of the Commissioner General for Investment in major projects. This requires experiments, which make it possible to compare the number of tons of carbon avoided with the cost of measurements.

63Finally, to reap the full benefits of climate policy, other policies (taxation, planning, technical regulations, etc.) must not thwart its effects and, beyond that, ensure their coherence. In this respect, land policies are a key issue. The structural trend towards urban sprawl must be reversed to reduce emissions from the transport sector, which requires financially accessible and attractive cities.

64Recommendation 6. In addition to carbon pricing, support innovation and green projects. Submit the choice of these projects, the corresponding grant programs and technical regulations to ex ante economic evaluation or small-scale experiments in order to compare their cost with the number of tons of carbon avoided.

65The price of carbon reflects the value we place on the sustainability of the planet for our children. If it is essential to effectively guide action, it must be based on governance that removes any suspicion of its purpose and be associated with a shared narrative vision of the issues of ecological transition. If its revenue is properly redistributed, it can be effective and fair. In addition, the ecological transition requires green investments, both public and private, for which the learning curve must be stimulated. The development of green finance and the definition of research or infrastructure programs are complementary to carbon pricing.