Effective altruism is not only about looking for actions that do the most good, but also about avoiding ineffective actions. Here I will give four examples of campaigns or actions that can backfire in the sense that they can do more harm than good. I cover three areas – environmental pollution, animal suffering and social injustice – and one general strategy – fundraising. The objective of this article is to let us think more critically about helping others and become more effective in doing good.
A recent ineffective campaign from the environmental movement was the campaign to ban the herbicide glyphosate, which is primarily based on the World Health Organization International Agency for Research on Cancer (WHO IARC) 2015 evaluation that in terms of hazard (i.e. whether the substance is capable of causing an effect), glyphosate is probably carcinogenic. First of all, there does not seem to be a scientific consensus on the carcinogenicity of glyphosate. For example, the European Chemical Agency (ECHA) says available scientific evidence did not meet the criteria to classify glyphosate as a carcinogen. Second, even if glyphosate may be a carcinogenic hazard, there seems to be a scientific consensus that it poses no carcinogenic risk (i.e. the actual exposure to the substance – for example through diet – is too low to show a significant effect). The European Food Safety Authority (EFSA), the UN Food and Agriculture Organization (FAO) and the WHO agree that glyphosate is unlikely to be genotoxic at anticipated dietary exposures and unlikely to pose a carcinogenic risk to humans from exposure through the diet.
It happens that glyphosate is one of the safest herbicides. The Environmental Impact Quotient (EIQ) of glyphosate is 15, lower than many other herbicides that were used instead of glyphosate, such as imazethapyr (EIQ 20), trifluralin (EIQ 19) and pendimethalin (EIQ 30). The human toxicity potential of glyphosate is 0,7 gram 1,4-DCB-equivalent per kg substance, almost 40.000 times lower than the average herbicide. The soil ecotoxicity and freshwater ecotoxicity potentials of glyphosate (per kg substance) are respectively 800 and 300 times lower than the average herbicide. Acetic acid, salt or liquid bleach (sodium hypochlorite), or other common herbicides such as dicamba, are probably more polluting than glyphosate. And they are more toxic for mammals. The LD50 lethal dose of glyphosate (for rats) is 5 grams per kg body weight, compared to 3 g/kg for acetic acid and salt, 1 g/kg for dicamba and 0,2 g/kg for sodium hypochlorite.
So what would happen if glyphosate were banned? That might easily decrease global welfare, according to agricultural economics research. First, if other herbicides are not banned, farmers might switch to those more toxic herbicides, increasing pollution. Second, if those other herbicides are banned as well, farmers might switch to mechanical methods for weed control, such as tilling and ploughing. These methods can decrease soil quality and increase fuel consumption, CO2-emissions and soil erosion. Third, if those other methods for weed control are not used, one can try to do weed control through manual labor, but this would drastically increase labor costs and hence food prices. Finally, if no weed control method is applied, it will result in lower crop yields, which means more agricultural land required, which means more deforestation and loss of natural habitat. Evidence that herbicide-free organic farming is better for the environment, is lacking.
Given this lack of evidence that alternatives to glyphosate use are substantially better for human and environmental health, urging for a ban on glyphosate might be too early. We first need to know the counterfactuals: what would be used if glyphosate was banned? And we need more scientific studies about the environmental impacts of those alternatives.
What are more effective campaigns at this moment? A first, more straightforward campaign would be the ban on the most dangerous pesticides on the market and the promotion of integrated pest management. But above all, the most effective campaign for the environment, is a decrease of animal production (e.g. a campaign for a high tax on animal products or for the promotion of animal-free products). The production of meat requires on average 4 times more cereals and soy (used for animal feed) than the production of plant-based protein-rich products. This means meat requires 4 times more herbicides than vegan alternatives. Animal-free, vegan food offers many other human, environmental and animal benefits as well. Besides, according to the IARC, red meat is like glyphosate probably carcinogenic (category 2A) and processed meat such as bacon is carcinogenic (category 1). However, pointing this out, results in another problem, as we will see next.
Red meat has a higher environmental and health impact than other protein-rich foods. So a focus on the environmental and health consequences of diet might result in people eating more chicken meat and eggs as a replacement for red meat. The same goes for people who are concerned about animal welfare and are in favor of mammals. Campaigns that focus on the suffering of pigs and cows might increase the consumption of chicken meat, and vegetarians might increase their consumption of eggs and egg-containing products.
However, in terms of number of animals used and killed or hours of animal suffering per kilogram product, chicken meat and eggs are about 20 times worse than red meat. (Concerning eggs: male chicks and less productive layer hens are killed.)
Even a replacement of red meat by a vegetarian (non-vegan) alternative that contains egg-protein might be more harmful to animals. Suppose a vegetarian sausage contains 4% chicken egg protein. To produce 1 kg of egg protein, one needs 14 kg of eggs (i.e. 230 eggs). However, eggs not only contain protein but also other by-products with a market value, so we need to multiply the amount of eggs with an economic allocation factor of 0,45. That means 1 kg of vegetarian products containing egg-protein requires 0,25 kg eggs (i.e. 4 eggs). According to this calculation, egg-protein containing vegetarian products are 5 times worse than red meat in terms of animals killed and hours suffering. Reducetarians and flexitarians who replace red meat by such vegetarian alternatives, increase animal suffering.
What is a more effective strategy? First of all, we should avoid single issue campaigns that merely ask for a reduction of red meat consumption without mentioning chickens and eggs. Hence, we should promote vegan alternatives. In particular, red meat can be replaced with vegan protein-rich foods and chicken meat and eggs can be replaced with nuts, seeds and vegetables. Vegetables are better for the environment and human health than chicken meat and eggs, so this message is compatible with a concern for the environment and human health. Second, in terms of reducetarian or flexitarian campaigns, we should focus on reducing the consumption of chicken meat, eggs and farmed fish, because these animal products have the highest moral footprints. Finally, consumers can put pressure on producers of vegetarian products to eliminate eggs in their products.
A third example of an ineffective measure is a consumer boycott of products made with low wage labor or relatively bad working conditions (e.g. sweatshops). As long as there is no involuntary slavery involved, such boycotts might easily backfire: a boycott might result in those workers losing their jobs, and hence they often become worse off because they no longer generate an income or they move to another job that is less favored (i.e. with worse working conditions).
Buying Fairtrade is not always the answer either. The higher prices for Fairtrade gives the producers an incentive to produce more, which can result in overproduction and consequentially a decrease in price of the non-fairtrade products. The producers who were not able to get a Fairtrade certification can end up being worse off, with lower prices for their products and hence lower incomes.
So what is the answer? What is more effective to improve social justice? The general answer are campaigns against unearned income (income gained not through labor or entrepreneurship but through ownership of land and other monopoly), economic rent (a surplus profit above normal profit, received for non-produced inputs) or rent seeking (seeking to increase one’s share of existing wealth without creating new wealth). More specifically, a tax shift is possible: taxing economic rent (e.g. natural resources) instead of labor. Other related examples are a global resources dividend (an idea from Thomas Pogge) and a clean hands trust and clean trade in natural resources (an idea from Leif Wenar).
For individual consumers, a more effective alternative than boycotting sweatshops and buying fair trade is donating money to organizations that give unconditional cash transfers, such as GiveDirectly or Eight. In fact, one can argue that we have a duty to donate money to those charities.
If we measure the cost-effectiveness of measurable interventions (e.g. in terms of numbers of lives saved, loss of quality adjusted life years avoided, kilograms of toxics avoided, hours of animal suffering avoided, levels of income increased or levels of crime decreased per dollar invested), we see a very skewed (often log-normal or fat-tailed) distribution. A minority of interventions is far more effective than the vast majority, doing a lot more good per dollar. Most interventions have an effectiveness below the averages, because the small minority of highly effective interventions drives up this average. This is just like the global income or wealth distribution with a small number of very rich people. We can expect that the immeasurable interventions (whose cost-effectiveness we are not able to measure yet) have a similar skewed distribution.
This has important implications for fundraising. We can consider three types of organizations. First there are the big, multiple-issue organizations that do a lot of campaigns, projects and interventions (e.g. Greenpeace, Peta, Unicef). If their campaigns are randomly distributed, these large organizations have an average cost-effectiveness. Next, there are the single-issue organizations focusing on specific problems or specific interventions. Most of those single-issue organizations have an effectiveness below average, because most likely they focus on low cost-effective interventions. Examples are local environmental organizations, animal shelters and organizations that focus on minority groups, poverty and diseases in rich countries. A third group of organizations are the minority of highly effective single-issue organizations (e.g. the organizations recommended by GiveWell and Animal Charity Evaluators). They have an effectiveness above average.
Now suppose that you go fundraising for an organization. As a result, the number of donations and the amount of money that people donate might increase a bit. But we also see a shift between organizations: people start to donate more to your charity and less to other charities. This means there is a shift away from a group of charities with an average effectiveness. If your charity has an average effectiveness (e.g. it is a multiple-issue organization), this shift is neutral. But if your charity has an effectiveness below average, your fundraising might actually do more harm than good. The average effectiveness might be an order of magnitude (a power of ten) higher than the median effectiveness (i.e. than the effectiveness of most single-issue organizations). That means if your fundraising causes a shift towards a charity with below-average effectiveness, the amount of dollar donated would have to increase with an order of magnitude in order to compensate for the loss of effectiveness by the shift between charities. A world where you are not fundraising for that charity might be a world where more good is done. So even fundraising for a charity can sometimes be a harmful job.