Tax legislators, product legislators and service legislators are sometimes confronted with the political wish to integrate environmental or social goals into regulation. This is an understandable wish, well intended. However, the devil is in the detail. We will see in this article where exactly, and how, the devil could possibly be outplayed.

I. The main difficulty: multiple causality chains, none of them easy to be calculated

1. One of the main issues of environmental policies is the interaction between various environmental parameters. Even within the same type of legislation, these interactions can cause headaches for lawmakers. E.g., there was, at least in the past, a devilish interaction between the immediately damaging NOx emissions of cars and their CO2 emissions. Optimisation in view of NOx made CO2 emissions sub-optimal and vice-versa. As more parameters are taken into account, then more optimisation, in view of one parameter, will trigger anything but ideal states for others. The substances which can easily be recycled are not necessarily those which are energy-efficient in their production and those optimised in view of one, or the other, are not necessarily those which are harmless for use in food-packaging. And so-called “ecological” fuel based on sugar cane, destroys the rain forest in Brazil indirectly because Brazil is unable or unwilling to control deforestation which aims at producing sugar cane. The harm caused by this deforestation both for the climate and biodiversity is said to be far bigger for the climate than what we gain by E10 fuel instead of normal fuel. Not to speak of the social effecton the indigenous population and the poor who are pushed aside for sugar cane, losing their subsistence livelihood. Going a level deeper, we can analyse that the possibility to substitute, to some extent, one energy form with the other and the fuel-crop bridge between the energy market and the food and feed market have created a huge energy/food-feed market. This huge market is on one side linked to the market of energy saving technologies and on the other, to biotechnology and the market for land; possibly it is even linked to more markets. Predicting consequences of any legislative or other action on one of these markets would require tremendous data and simulation models. The author is not sure that the necessary data and artificial intelligence for such an undertaking is currently available or in reach. Ironically, such an amount of data and corresponding simulation models will probably first be available in totalitarian jurisdictions with a high technological level. Guess which.

2. To conclude: complex causality chains or even causality nets across the world make it virtually impossible to calculate in advance the effects of legislative action requiring e.g. a certain minimum standard for whatever environmental or social parameter. Even less so the causality chains for various parameters can be calculated. Not to speak of the mentioned interaction between the various parameters. Hence any undertaking imposing one or several environmental or social requirement in product legislation might easily cause more harm than good.

II. First element for a solution: complex rule systems

3. But wait a second. How do we operate in other complex ethical contexts where the consequences are not foreseeable? Do we refrain from deciding based on ethical criteria! Not at all. As we have to decide in one way or the other, we mostly decide based on rules. The rules system must be complex enough to cover complex situations. Here an example of how such rules systems can be constructed, by exemptions to the exemptions of the basic rule:

Level 1: You shall save lives.

Level 2: You shall save lives unless, by doing so, you miss the opportunity to save more lives elsewhere.

Level 3: You shall save lives unless, by doing so, you miss the opportunity to save more lives elsewhere. If, however, the minority group of persons you can rescue is innocent whereas the majority has caused the ethical dilemma by criminal action, you may or even shall rescue the minority.

And so on …

4. In the same way, environmental or social rules could be established:

Level 1: Products produced by minors below 14 years shall be banned.

Level 2: Products produced by minors below 14 years shall be banned, unless the minors live in an area where they have de facto no chance to go to school or obtain better professional possibilities than working in the factory.

Level 3: Products produced by minors below 14 years shall be banned, unless the minors live in an area where they have de facto no chance to go to school or better professional possibilities than working in the factory. If, however, the production causes high risks for the physical or mental development of the minors, the products shall be banned nonetheless.

Level 4: Products produced by minors below 14 years shall be banned, unless the minors live in an area where they have de facto no chance to go to school or better professional possibilities than working in the factory. If, however, the production causes high risks for the physical or mental development of the minors, the products shall be banned nonetheless. The last sentence shall not apply where its application would lead to prostitution, starving or other misery in so far as there is no alternative lawful activity accessible to the minors.

5. Even when applying the already quite complex 4-level-rules of the example just given, there will be situations where the rules are insufficient. E.g., certain jurisdictions might, if the social standard was applied to products, refrain from creating schools or other alternatives. The complex 4-level-rules are not complex enough to cover all situations. And even a 9-level-rules system will fail. But on average a 4-level-rules system will be more appropriate than a 2-level-rules system and thus is preferable. We should strive thus for more levels and, if needed, several exemption strains, but stop short of the degree where the system becomes unmanageable. The application of such a relatively complex system itself is ethically justified because, on average, we better apply such a system than not applying any system or rules at all.

III. Second element for a solution: rating in accordance with as many parameters as possible

6. We have seen in I. that applying just one criterion / parameter can easily cause more damage than do good. The best approach to avoid such negative interaction is to include all criteria / parameters on which the first criterion / parameter has an influence or vice-versa. Some of these criteria / parameters have a further interaction with other criteria / parameters not yet mentioned; hence the latter should be integrated as well to the extent possible and so forth. Once the bundle of criteria / parameters is pretty complete, it is sufficient to measure or rate in accordance with all the criteria / parameters. As a further step, it is necessary to correlate or scale, certainly to some extent in an arbitrary way, the various criteria / parameters. The criteria / parameters will not all have the same value. E.g. child slavery might be regarded as twice to 10 times as immoral than child labour. The scaling or establishment of a correlation as such is also ethically justified: better a somehow arbitrary scaling than none, because a scaled system will permit better decision-making on overall environmental and social acceptability e.g. of a product or service. And it must not necessarily be so arbitrary if a high number of persons of different cultures are requested to give their personal scaling / correlation preference. Taking the average of a high number of persons would make the scaling universal. It goes without saying that data science could tremendously help both for rating and for appropriate scaling.

7. To sum-up: the complexity of the many causality chains described in I. is somehow “ironed out” by applying many different criteria / parameters at the same time. With such a multitude, it will become virtually impossible to optimise in view of certain criteria / parameters and to do more harm in fields which are not covered by criteria / parameters. Anything less ambitious would have the system failures outlined in I.

IV. Estimating the overall environmental and social acceptability

8. At paragraph 6 above, we dropped the key word “overall environmental and social acceptability”. But how could the overall environmental and social acceptability be estimated in practical terms? Simply by combining the elements of II. and III. with auditing. There already exists social and environmental audits, and is a burgeoning market led by the world’s top consulting companies. However, such audits are currently still incomplete. But once they have become more complete and deep enough to reflect considerations of Level 3 to 4 listed under Part II., they should be able to provide a rating which is fair enough to be referred to in legislation.

9. The auditing is evidently to be undertaken by accredited organisations which are themselves assessed, for their auditing capacity, according to the same standards in a fair way. Hence there would be a need for a kind of comprehensive apparatus behind the auditing organisations, an inter-jurisdiction cooperation that sets out the qualification criteria for auditing organisations and the methods for their assessment. The need for ensuring comparability is due to the fact that auditing organisations will need to have the possibility to build their audit on the audits undertaken by other auditing organisations. They cannot themselves audit all suppliers and service providers or the upstream products or provided services. Evidently, here it becomes very tricky.

10. Given the size of the undertaking, it is not imaginable to introduce auditing in one strike to an entire national economy. Hence one would need to start with certain branches and extend from there. Within the same branch, companies cannot all be assessed at the same time either. Hence it would be commendable to start with voluntary auditing, linked to some privileges like positive labelling. Once a large part of the companies have been audited, mandatory auditing can be announced with a reasonable lead-time.

V. How can auditing the overall environmental and social acceptability be referred to in regulation?

11. There are many ways in which regulation could refer to the overall environmental and social acceptability? As stated above, it should first not be referred to in a non-mandatory way, permitting instead a slow phasing-in. Hence we could foresee the following order:

• Facultative labelling starting from certain benchmarks;

• Fiscal privileges;

• Mandatory labelling, with a bad labelling by default unless auditing has provided evidence for a better one;

• Tax based on rating (an example of such use is provided in a concept developed 10 years ago at Annex A); and

• Minimum benchmarks as a condition for providing services or placing products on the market.

12. In particular the minimum benchmarks can become too sharp an instrument that it causes collateral damage. Auditing is expensive and operators of poorer countries might struggle to afford it. Hence there is a danger of protectionism by the backdoor. Ethical arguments may thus easily be abused. At the end of the day, how can one explain to a South-Asian textile worker that s/he will lose his/her job because his/her company cannot fulfil certain environmental or social standards or is not able to prove compliance due to the costs linked to such proof? Regulators wishing to use sharp instruments should keep this question in mind.

VI. Summary and outlook

13. We have seen at the beginning that referring to one or a few environmental or social criteria can easily do more harm than good. In a second step, we have identified three elements which, when combined, could nonetheless overcome this difficulty. However, they bring along the need to establish a complex auditing machinery which covers a wide range of criteria / parameters. Such auditing machinery will not be easy to establish, the more so as we have complex global supply chains. Transparency of all involved in these supply chains is therefore key to providing fairness. Once transparency has been ensured, data science methodology might help tremendously to aggregate key information and to bundle it in a way that at least creates a good basis for rating. Hence the size of the undertaking might subjectively dwindle with the development of transparency and data science. And here comes the chance for regulators to prepare a slow phasing-in at various levels outlined in Section V and from one sector to the next and from the biggest operators downwards. During this process, regulators can gain experience including to avoid distorting effects and punishing weak economic operators who have natural difficulties to adapt.

ANNEX A

Total Environmental Impact Taxation

The following concept was developed on request of an influential person looking for a policy approach which would reduce the environmental burden of industrial production (goal 1); above all with regard to products which are not covering basic needs (goal 2). The policy approach should not deteriorate the situation of disadvantaged or marginalised people (goal 3).

I. The idea in essence

1. The idea of a Total Environmental Impact Tax (TEIT) is to evaluate the overall environmental impact of long-living consumer goods and, in a second step, of other products and, even later, of services. To determine the Total Environmental Impact (TEI) as a basis for a VAT-like tax that is labelled, in shops, as part of the price, it is calculated on the basis of the average environmental resources consumed by the specific product type. The sub-criteria for calculating the TEI will include CO2 emissions caused, recyclability, toxicity, natural resources used, etc. For each of the sub-criteria negative points will be attributed to the product type in general. The concrete number of negative points might in addition depend on a few clear parameters like weight, size, energy consumption etc. E.g. a fridge might, on average cause X negative points, but the concrete number of points might be reduced by 10% multiplied by the weighing factor attributed to the volume, if the volume is 10% less than the average volume.

2. The TEIT should be introduced slowly as a complement to the VAT. At a later stage it might step-by-step substitute the VAT. The overall tax charge could remain the same or be lower or be higher, subject to the ordinary political decision making. Just the percentage of VAT versus TEIT would change. If, however, there was a need for increased state income, the TEIT could become the ideal means thereto.

II. Variants

3. The following variants can be added individually or cumulatively. Variant (a) is recommended:

(a) The TEIT should be labelled together with the price. Consumers interested in the protection of the environment will choose more environmental-friendly products.

(b) In case of step-by-step substitution of VAT by TEIT, there should not be a negative social impact. Nonetheless, it could be imagined to go further by reducing the anti-social effect equally attributed to the VAT. The VAT puts a higher burden on low-income people than those with a high income. To reduce this effect, some elements of the income tax could be introduced. The major element of the income tax is to keep the income tax-free in as much as it is necessary to survive. The income going beyond this limit is taxed progressively. The same effect could be achieved by attributing a certain amount of TEIT points for free to each citizen or household. Citizens can use these free points when buying products. The points attributed to a certain product that the citizen buys will be deduced from his or his household’s point credit. In order to economise his points, the citizen will look at the number of points labelled for each product. Therefore, he will choose rather environmental-friendly products. However, it could be argued that, if the tax is being fully implemented, the cost of all goods in financial terms will reflect its environmental costs. In this case, there would be no need to distribute points. To distribute cash can compensate any anti-social effect.

(c) Another option consists in taking into account to what extent the need covered by the product is basic (e.g. food, clothing, housing) or pure luxury (jewellery) by building in a factor distinguishing between different needs in the same way as some countries distinguish between 0%, a low percentage, a medium percentage and a high percentage for VAT. The criterion for classifying might either be the ones used for VAT or the question: To what degree does the product contribute to life prolongation or to the alleviating of pain? However, it could be argued that such a distinction, though based on a relatively clear criterion, complicates the system. Accordingly, we had different views with regard to this option.

(d) The mechanism outlined in Variant (b) might, in the long run, be extended in such a way that the TEIT is progressive. The first range of points going beyond the TEIT free points outlined in Variant (a) would be charged at a lower rate than the points of the second range, those of the second range lower than those of the third range etc.

(e) Give manufacturers a point discount on their products and service in as much as they can demonstrate, via eco-auditing, that the TEI is better as the average for this kind of product. The eco-auditors would have to calculate the TEI on an individual basis. This will give them a competitive advantage.

(f) Once the complex eco-auditing has taken place, it can also be combined with advantages with regard to direct taxation (company taxes, income taxes).

(g) The TEIT might be increased so as to be able to substitute other taxes as well. However, we partly fear that social equity cannot be assured if the TEIT was to replace a progressive income tax.

(h) The TEIT might cover governments’ spending. Otherwise governments will not have a sufficient incentive to spend in a green way.

III. Technicalities

4. The cascade effect. VAT is a tax paid only by the final consumer, companies can reclaim VAT paid. This is a very bureaucratic and burdensome procedure and open to abuse. In theory, an environmental tax could avoid this. Each consumer would pay full tax on the product consumed. Companies would pay tax on the fuel they use and on the materials they purchase. Wholesalers would pay tax on their transport and storage costs and all of these costs would be passed on to the retailer and the final consumer. The final consumer may also pay a premium based on the intended use of the final product. However, it might not be so easy to tax imported products appropriately so as to not discriminate the domestic production. Equally, exporters might suffer from the TEIT if no reduction is given as long as other jurisdictions do not have the same TEIT.

5. Imports. Imports would be treated in the same way as products manufactured in the domestic jurisdiction. Unless the manufacturers demonstrate by eco-auditing a lower TEI for their products, their products will be subject to the full TEIT. If the TEI calculation is harmonised amongst the various jurisdictions applying the concept, the eco-auditing undertaken once can be used for several jurisdictions. Equally, auditing bodies might offer to apply diverging eco-auditing criteria of different jurisdictions in one auditing exercise. Bilateral and multilateral agreements could assure a fair treatment of imports and exports, avoiding unfair double taxation.

6. Balance between taxation of the product and of its use. It is needed to work out what part of the environmental impact is being taxed at what stage in the cycle. Cars are bad for the environment, but a car that is never used has only the negative environmental impact caused by its production process. It is when the car is used that it causes most environmental damage. The main tax would therefore need to be on fuel. If the full environmental cost of burning fuel were charged, then it would be an incentive to consumers to buy fuel-efficient cars and to use them less. In the case of fuel, it should be fairly easy to ensure that a single tax is levied calculated on the basis of both production impact and likely impact of use. Thus fuel that was extracted and refined at low environmental costs would be marginally cheaper, but the element of final impact would presumably be the same, regardless of origin. Only the quality of the fuel might play a role.

7. TEI to be estimated until sale. Basically, it is sufficient to tax the environmental consumption of a product until the product reaches the client. E.g., the energy consumption of a frozen meal in the hands of the consumer will be paid by him. Only if the energy consumption of consumers is not yet perfectly covered by the TEIT, the typical energy consumption of the product needs to be covered by the product TEIT.

8. Change of TEI in time. Every time a production process changes, or every time the weather changes the TEI changes. However, the difference is mostly marginal. To impose a renewed evaluation every 5 years will make sure that the rare cases of a deteriorating TEI will be covered. As to the substantial improvement of the TEI which might have occurred due to new technology, the manufacturer would be free to request a new evaluation earlier, at his convenience. Thereby he would take profit from the lower TEI very fast.

9. The creation of a self-regulating auditing market. The system would need to leave manufacturers the choice between paying the TEIT for normal products, and proving for one, more or all aspects that the product is more environmental friendly than the average. The manufacturer can prove a lower TEI either by an individual audit that analyses precisely the TEI. However, he might also just get audited with regard to the fulfilment of a production standard, e.g. a production respecting the rules of green chemistry. The fulfilment of production standards might lead to a TEIT reduction in as much as the fulfilment of the production standard creates a presumption of lower environmental costs. The presumption is based on estimation. Accordingly, the TEIT reduction must be slightly reduced to take into account the uncertainty of the estimation. These standards might cover all environmental aspects covered by the TEIT or, more likely, only some of them. The manufacturer might use a standard for low-energy transportation and not for green chemistry or vice versa or for a third purpose. The manufacturer can estimate best for which aspects an individual estimation of the TEI might give him most credit compared to the cost of evaluation. Responding to the needs of the manufacturers, a market of audit services of entrusted auditing bodies would be created. Auditing bodies would offer standard-based individual assessment of the TEI. As the TEIT would need to be introduced sector by sector and based on pilot projects, the new auditing industry would emerge at a natural growth rate.

10. Step-by-step introduction. Step-by-step introduction can be assured in two ways. First the TEIT might be introduced product by product, starting with those products which are using most environmental resources or for which estimates of environmental consumption can be easily made. Second there is also the possibility to introduce the TEIT parameter by parameter. E.g., the TEIT could be launched with energy consumption as single parameter, followed by a second parameter recyclability etc. When introducing a new parameter, states will be able to use the experience gained other states provided that statistical data and knowledge about assessment methodologies are exchanged. The OECD could be an appropriate forum for this type of exchange.

IV. Conclusion

There is a wide variety for future TEIT systems. TEIT could thus become a new policy field that can be adapted to the needs of the respective state, the traditions of its inhabitants, their environmental profile, the social structure etc. At the same time, the TEIT would increase the states’ income. Thereby, it would enlarge the possibilities for states (and politicians!) to shape policies. In times of shrinking fiscal revenues, in times in which more and more scientists claim that the state is losing its ability to act, the TEIT could become a cornerstone for an environmental shift of societies and, via the finances, for a more human design thereof.