The Use and Misuse of LCAs

This article is the third of a triptych of pro bono pieces published in Apparel Insider Issue 14, July 2020. The other two are “Was it Polyester all along?” and “Silk Revisited”

Sustainability is substantive - it can be defined and measured. In general terms “sustainable “ means something able to continue at the same level for a significant period of time without damaging future production potential; or production using methods so that natural resources are still available in the same mass and quality in the future. Exactly what it means in the apparel sector nobody knows because there is no agreed definition, no agreed method of measuring it, and no agreed system for verifying claims.

1. LCA and sustainability are not synonymous

This confusion is compounded by the misconception that all you need to demonstrate comparative fabric sustainability is a couple of LCAs, one for each of the fibers concerned - et voila. 

For farmed fibers this is categorically not the case.

You absolutely must consider the socio-economic impact on farmers, their families, and their workers. This is especially true if you are paying lip service to the SDGs but it seems to be largely ignored by fashion’s sustainability experts. I cannot overemphasise its importance. You cannot say whether a production method is more sustainable or not without knowing how it impacts the farmers. You need a social and economic impact assessment, Randomised Controlled Trial (RCT), or some equivalent. 

Are farmers going to be losing money, or engaging in riskier and more dangerous work? Will their income, security, and environment be enhanced? Will infant diarrhea be reduced for example, by switching to artificial fertilisers and sharply curtailing animal waste in the surroundings? Or will cancer risk be augmented by the use of toxic chemicals? What are the comparative QALYS (quality-adjusted life-years) involved? 

It is perfectly pointless for the apparel sector to spend tens of millions of dollars on pushing farmers to adopt organic cotton, if for the farmers, it does not make sound economic sense. We are repeatedly told by C&A, Laudes, Cotton Connect and many others that x thousand farmers have been converted to organic production methods, but nobody talks about attrition.

In fact, C&A’s own 2018 SEIA of Madhya Pradesh Cotton farmers found organic farmers had 1.6 times the debt of their conventional colleagues, and a lower net income. As a result, some 30% of the 1200 organic farmers were not actually producing organic cotton any more at the time of the 2018 study, and the study concluded: “A potential area for future research would be to identify the reasons that farmers adopt, but then quit organic cotton farming" (p.30).

Quite.

Similarly, a 2008 follow up study on an organic cotton initiative promoted by Shell Oil in Gujarat (a precursor to their creation of CottonConnect) found an extremely high rate of attrition once the enthusiasm of the initial years had passed. In the Surendranagar Area, after training 800 farmers, at the end of the 6th year only 460 remained - an attrition rate of 42%! For the  Kutch Area, from the table shown in the study, attrition appears much lower - after training 579 farmers, 560 remained at year end 2006/7 - but later we learn this is only because the entire new cohort from Mandvi, Kutch, have apparently been eliminated:

"The number of farmers participating in the Mandvi Taluka of Kutch has reduced drastically, due to increase in salinity of water in the region and large scale conversion of farmers to BT Cotton cultivation for immediate economic benefits." (p.19)

Hmm. So the Gujarat farmers too, quit organic farming because, after initial enthusiasm, they found their conventional fellows were actually making more money.

I asked CottonConnect for their attrition rates. They did not reply. CottonConnect’s actual staff appears to be small, and their activities are subcontracted to a variety of organisations. Given the complexities of growing cotton, I am sceptical that subcontracting to an Ashram, no matter how worthy their spiritual merits, will result in the conversion of 8,000 farmers to “more sustainable cotton” - whatever that might be.

I think we can safely assume that of the tens of thousands of farmers the brands and their funded initiatives claim to have “converted” thus far, a significant percentage no longer actually produce organic cotton. And whilst the billionaire owners of the brands that finance and promote these schemes have enhanced their sustainability credentials and bottom line as a result, some of the poorest people in the world have actually ended up worse off.

2. There is more than one type of LCA

Which one should we be looking at? A “consequential LCA” or an “attributional LCA”?

If you want to know what the average impact is for current production, then you need the latter. But, if we want to analyse the impacts of substituting one fabric for another, say silk for cotton, in our product mix we need to compare the cotton producer who would cease to produce with the silk producer who would respond to an increase in demand. That is called a "consequential LCA", and it has some differences in both methodology and results from "attributional LCA", as it is considering the marginal rather than the average supplier.

Personally, I would argue that you also need to consider the opportunity cost, in the sense that farmers are tied to their land. If our marginal farmer can no longer produce cotton, he will plant something else. Peanuts? Rice? Whatever it would be, we know that the switch has harmed the farmer economically, or he would have produced peanuts/rice/whatever in the first place. And if the environmental impact of those crops is greater than that of cotton, not only have we reduced global welfare, we have actually increased climate change as well

3. Sustainable means for life not just for production

If not all fabrics behave the same way when used, and when disposed of, then it is meaningless to look at sustainability from cradle to factory gate - which incidentally, is how the Sustainable Apparel Coalition's (SAC) Higg Materials Sustainability Index (MSI) is evaluated.

And, they don’t all behave the same way.

I think we all suspect that garments made of some fibers - like polyester - are more likely to be disposed of after a handful of wears, than garments made of more expensive fibers like silk or cashmere. If the original owner doesn’t want them, garments made of noble fibers can generally find a ready market on Vestiaire, Ebay or at the local charity shop.

If a silk shirt has a cradle to gate impact of 200 and is worn 100 times (by one or more owners), it has an impact per use of 2, and as long as the fibers have not been treated with toxic chemicals, it will cause no further harm through fiber shedding and end of use. A polyester shirt with a cradle to gate impact of 20, worn 10 times, also has an impact per use of 2, but once we have factored in the lifetime impact of fiber shedding and end of life polluting landfill, the impact is much higher than that. But simply comparing cradle to gate LCA’s, by definition does not tell you this. It tells you that the polyester shirt with an initial impact of 20, is 10 times more sustainable than the silk shirt with a production impact of 200.

In just four sentences, with the most rudimentary calculations,we have established that this is nonsense, and so following the Higg index may well be resulting in a huge misallocation of resources.

4. LCA results from different studies should be compared with caution, as not all follow the same methodology or have similar scope.

This is a direct quote from the 2014 Oxford LCA of Indian silk production used by the Higg. Sadly, this exhortation is ignored by the Higg itself.

We cannot see the underlying LCAs employed for each of the fibers analysed by the Higg. Indeed, it would appear that the only people who can see all of them are the Higg/SAC themselves.

Everyone else must take it on trust.

Are they all measuring Global Warming (GWP) in the same manner? Do they all have the same methodologies and system boundaries?

Given that they all appear to have been produced by different organisations, with different vested interests, at different times, it seems highly unlikely.

I have tried to address this question to Tom Gloria of Industrial Ecology Consultants and Harvard Extension School, who the Higg claim is their gatekeeper, but he does not reply to my emails. Handily, closer examination of the 2014 Oxford silk study provides clarification. 

This study actually computes two different sets of impact values - one for production according to Recommended Practices - the other obtained from the 2006 practices of 100 bivoltine silk farmers in Tamil Nadu.

As stated in that study;” farm practices differ from recommendations... farmers tend to overuse FYM and underuse synthetic fertilizer, substituting locally available non-traded inputs for those requiring expenditure. Consequently leaf yields are also lower” 

(FYM=Farmyard Manure)

The 100 farmers concerned were applying roughly one third of the recommended amount of synthetic fertilizers, and 2.5 times the recommended amount of manure. As a result their GWP/kilo of silk was more than 50% higher than that obtained under recommended practices. Eutrophication (FE) was almost 1.5 times as high. 

Based on this study,  Higg set the MSI for raw silk production in Global Warming Potential (GWP) at 83.7per kilo - double the total impact of polyester - and in eutrophication at 138.2 - more than three times the total impact of polyester.

For silk, the Higg claims using manure instead of chemical fertilisers greatly increases environmental impact.

When it comes to cotton however, Higg states the GWP for organic cotton is virtually half that of conventional production, whilst FE is stated to be only one third of conventional cotton’s impact.

Those who have read my previous work on claims made for organic cotton, will recall that the 2014 Thinkstep LCA employed by the Higg to evaluate that cotton fiber showed zero use of chemical fertilizers by organic farmers, and 10 tons of FYM per hectare as a probable norm. 

For cotton then, the Higg claims using manure instead of chemical fertilisers greatly reduces environmental impact.

In silk production, increased use of manure results in increased environmental impact. In organic cotton production, increased use of  manure results in reduced environmental impact. How is that possible? 

think we can safely conclude from the above that the studies concerned do not follow the same methodology, do not have the same scope, and so cannot be compared.

By the same token, we were told that the need to air condition silk worm factories was the main reason for silk’s harmful environmental impact. That assertion is clearly false but it raises an interesting question about boundaries - all large modern factories and offices are to some extent air conditioned. Do the LCAs for factory fibers include this? Or is only direct impact calculated? 

LCAs for farmed fibers (i.e. cotton, wool, silk...) appear to include every farm journey made in the crop’s production. Under Covid lockdown, we have seen first hand the emissions impact of eliminating most commuting. When calculating the impact of factory fibers, is worker commuting included within the impact boundary? Should it be? Certainly these trips would not be made if the polyester or elastane was not being produced. And if commuting isn't included, are the comparisons with farmed fibers misleading when it comes to total impact?

Italy’s Como silk manufacturers published an LCA in 2019. This is not publicly available and I have only seen a summary, but this not only claims that all the Chinese raw silk production used in Como is entirely rainfed, but also that each mulberry tree offsets 100 lbs of carbon per year. The 2014 Oxford study used by Higg assumes zero carbon offset. Which is correct is not for me to say but it is inconsistent for the fashion sector to claim endless plaudits for their tree planting carbon offsets, without explaining why mulberry trees offset no carbon at all, and why we should ignore so many recent studies in India, Brazil and Iran which have found that they do (mulberry, drip irrigated by wastewater from Mobarakeh Steel Company stored 10.0Mg/hm2)

5. Prices change, so apportionment by value may lead to very different results in different years.

The FAO have produced a  project proposal from the cradle of bivoltine silk - Huzhou. This, too claims that the mulberry trees are entirely rainfed, provide excellent carbon sequestration and air purification, that the entire system is completely circular “The chain of productions in the system does not depend on external inputs and achieves near-zero emission to the environment”, and that silk/mulberry cultivation provides a range of co-products from tourism to tea. The point here is the point also raised in the Oxford silk LCA used by Higg: the more co-products a system has, the less impact that needs to be apportioned to any one of them. 

The problem then becomes: how do you decide how much of the environmental impact to assign to each of the co-products? Generally, the answer is that you do it by relative financial value. For Huzhou silk, it appears tourism revenues might dwarf those of silk production itself, and indeed, by the time you factor in the value of flood prevention, the environmental impact of this particular method of silk production might be zero or negative? 

But I digress, apportionment of impacts by relative value is the method used by the Higg Index “ Economic allocation is used to assign the impacts across the products derived from cattle. Allocation to leather is 3.6%.” The problem is that relative prices change. The value of cowhides has fallen significantly in recent years (by 40%-60% between 2014-2019). When this happens the values in your LCA quickly become invalid. The Higg data is from 2014, so ceteris paribus, the impact of cow hides now is less than half that claimed by the Higg.  Indeed Stephen Sothmann  President of the Leather & Hide Council of America tells me that recently, the hide on a US steer has been worth 0.96-1.6% of the entire value of the animal on average, with about 15-20% of the lowest quality hides being landfilled or destroyed rather than utilized for leather.

Because nobody raises cattle for leather - hides are a by-product of the dairy and particularly beef industries - if we all stopped wearing leather tomorrow and demand for milk and beef continued unabated, there would be a large number of hides with nowhere else to go. These would have to be disposed of at considerable financial and environmental cost, so technically, a portion of the hides going into the leather industry have zero or negative environmental burden. 

By the same token, were the Higg to include the impact of manure in their evaluation of organic cotton - which by ISO standards,they should - unless the price of manure has also dropped, this would now also be higher, as less of the impact of methane and other emissions would be assigned to the cowhide and so by definition, more would be assigned to manure/milk/beef.

6. What about Biodiversity

I am not an LCA expert, but I have yet to see an LCA that attempts to measure this. However, I see no reason why it couldn’t be included if the global community so desired. The question is how best to do this? Those of you who read my November 2019 article “Sustainable Cotton, Myths versus Reality” will recall that Textile Exchange (TE), in their right of reply, responded to none of the issues I had raised, but attached a list of studies, mostly obsolete (1981-2005), predominately on biodiversity, and not one of which actually covered cotton. 

They seem to have been trying to suggest that organic automatically equates with higher biodiversity, and that appears to be a common assumption. But even the studies TE linked to did not demonstrate that - finding, for example, disappointing outcomes for bees in conventional and organic wheat fields alike. It seems obvious that for nectar dependent insects a field without flowers is a desert and it makes no difference whether the crop is organic or not. 

Australian cotton producers - not one of whom is organic - seem to be the only people who regularly study biodiversity in their fields, and they seem to be achieving very satisfactory results, by rainwater harvesting in tanks, strategically planting plots of native species in and around fields and so on. USA cotton farmers - both conventional and organic - seem to be following suit. However, it is generally agreed that most of the biodiversity is lost when the habitat is transformed in the first place, so some argue that the real measure of biodiversity is yield, because higher yields frees up land that could be used for conservation. Obviously, these two means of promoting/measuring biodiversity might pull in opposite directions, so this is yet another aspect of the definition and measurement of sustainability that needs to be discussed and agreed upon by the global community.

7. Unlike Factory Fibers, Farming Depends on Nature

Whilst changes in relative prices, and so apportionment, can apply to all fibers, farmed fibers are subject to an additional variable which makes it particularly difficult to capture their impact globally, and over time - nature. 

Experts with showers, flushing toilets and swimming pools - not to mention transatlantic flights - decreeing that small farmers with none of the above, cannot use irrigation, pesticides or artificial fertilizers, as their produce is then “not sustainable”, when these inputs are also those that provide the desperately poor and vulnerable with insurance against failed rains, pest infestations, and paltry yields, infuriates me. But let’s stick to the science. 

Sometimes the rains fail. If farmers have irrigation, they may then use an inordinate amount of it. If they do not, the crop may fail, and the yield per input will be paltry. This does not mean that the whole thing is “unsustainable”, that we should throw out natural fibers and rush to embrace “circularity” and factory based everything. It does not mean that, as Laudes Foundation appears to believe, producing cotton in a greenhouse in Essex is automatically the solution.

Reducing global inequality and promoting the Sustainable Development Goals seems to be part of every brand’s CSR statement. Eliminating poverty and promoting decent work and economic growth in the global South, means helping farmers. To be of any worth, it is crucial that farmed fiber LCAs are representative, and that years and locations are not cherry picked.

By cherry picking, I mean selecting LCAs so that preferred crops can be shown in the best possible light, and those crops that are not favoured by the funding brands, in the worst possible light. An example of the former is  the Higg LCA of organic cotton which focuses on areas where the cotton is rainfed, and then excludes manure, so that organic cotton appears to consume much less water and have much lower impact than conventional cotton. An example of the latter is the Higg’s choice of LCA for silk which is based solely and entirely on silk that is fully irrigated, in the least efficient manner, and with minimum monetisation of co-products, and so appears to show that all silk production consumes huge amounts of water, and generates considerable emissions.

8.The LCA must be up to date

As noted, the LCA used by Higg to evaluate silk’s impact is also old, whilst published in 2014, the farm data is from 2006, and to the extent that Higg are using any Chinese data, it appears to be from 1998. In 1991, Australian cotton benchmarked its environmental performance by undertaking an independent environmental audit as a catalyst for transformation. Producing a bale of irrigated cotton in 2019, required 48% less water and 97% less insecticides than in 1992. In just the last 5 years 2014-2019, water use per bale fell 9.4% and pesticide use by 18.2%. Clearly using data that is even 5 years old could be misleading. Most of the agricultural data input to the Higg index however, is far older than that. 

The Higg for Alpaca for example, states “Time Representativeness Good  Data are not older than 6 years with respect to the release date or latest review date”, but the data itself, for what it is worth, comes from a 2008 publication by the FAO “Wool and other animal fibers in South America”. 2008 was 12 years ago, not less than six. Similarly, for silk, the Higg states: “Time Representativeness Excellent  Data are not older than 4 years with respect to the release date or latest review date”. But, as I have already pointed out, the production data is from 100 farmers in Tamil Nadu, in 2006. That’s 14 years ago, not less than 4 (as for 1998, that’s 22 years ago).

9. No GIGO please…

As noted in my companion pieces on silk and polyester, every model is only as good as its base data. This is not an LCA manual, but in general terms, as for any other study, LCAs  must be large enough to capture whatever it is that is being measured. You can do an LCA for a single factory or small group of farmers but that tells you nothing about local production in general, let alone global output. If the LCA is comparative, farmers must be randomly selected, and matched, not just by regional climate but also by soil type, capitalistion, education etc. The data must be independently collected and evaluated and so on. It goes without saying that if robust data is sought, neither manufacturers nor initiatives can be allowed  to produce their own LCAs. 

10. The country weighting for global production data must match global consumption:

When calculating global averages In the LCAs that I have seen, it is common practice to weight the production data for different countries by their share in global production. But if this data is to be used by the global apparel sector to evaluate its impact, then this is correct only if the sector consumes fibers from different nations in direct proportion to their share in global output. This may very well not be the case.

 India produces some 22% of the global silk output, but as  the Oxford silk group states “gradable bivoltine silk, representative of internationally traded silk, constitutes only a small proportion of total Indian silk output.”  so to calculate a global average impact for the Higg - or anything else - using average data for India, assigning it a weighting of 0.2, and Chinese production data, a weight of 0.8, would be utterly misleading.

Similarly, India produced the most cotton in 2018/19 - 5.8 million tonnes. But India's population is 1.4 billion - so a lot of that cotton will have been consumed domestically, and will not have entered the global supply chain. The USA came second with 4 million tonnes. The US population is 0.3 billion. The United States is the world's leading cotton exporter, “providing more than a third of the world’s cotton exports" according to the USDA. whilst "Australia is one of the world’s largest suppliers of raw cotton with 99 percent of the domestic crop exported, mainly to China, Indonesia, and Thailand." Clearly, extrapolating directly from LCAs based on global production data, could be misleading.

Similarly, as noted, 94% of global polyester fiber is produced in Asia but The Higg uses 2011 data, from European production facilities, to estimate polyester impact. This is categorically unrepresentative of consumption by the global apparel sector

Summary and Conclusion:

Sustainable does not mean that you must produce with nothing but thin air. Some areas have been producing the same farmed fibers, on the same land for centuries, or even millenia. To the extent that the water table is not being depleted and the farming is regenerative - a very popular term, this actually describes practices many farmers have been using for generations - I would submit this is sustainable. 

It is also considerably more “circular” than the one way street of PET bottles to fabric that is currently euphemistically and misleadingly described as “recycled polyester” by the apparel industry . Polyester fabric still cannot be recycled at scale, nor composted, nor biodegraded. The mountain of fossil fibers shown in “Was it Polyester all along?” is still with us and, it seems always will be - albeit in ever smaller and potentially, more dangerous particles. 

If we are serious about halting climate change and attaining the Sustainable Development Goals, then the need for a definition of sustainability to be used in the apparel industry, a system of measurement for the same, and a robust method of oversight, is long overdue. This is not something for the big brands to unilaterally decide. Producers - including farmers - and consumers, with expert scientific input from leading academics, must agree on all of this , and as soon as possible.

If brands wish to make comparative fiber sustainability claims, they must cover cradle to grave, social and economic sustainability as well as environmental, and all the LCAs/SEIAs must have been produced using exactly the same methodology and boundaries - or you can’t compare them. 

Think of financial accounting - you cannot produce company accounts any old way you please. If companies wish to claim to be more sustainable, to parade their B-corp status and their Green Bonds -  with “science based targets”, they must first show us the science. Without this, absolutely all of it is just marketing.