Where does rich-country plastic go? Evidence on pollution havens

In January 2020 Malaysia announced that it had sent 42 containers of illegally imported plastic waste back to the UK (Greenpeace UK, 2021, p. 9). Its environment minister, Yeo Bee Yin, said, in Greenpeace's quotation: "If people want to see us as the rubbish dump of the world, you dream on." (Greenpeace UK, 2021, p. 10) If your waste contractor reports that your plastic was "recycled", what do you rely on to know where it went? This post sets a report by Greenpeace UK, a campaigning organisation, beside a 1992 study of dirty-industry migration. The full case is in the UK plastic-waste system.

What does the Greenpeace 2021 report count?

It counts tonnes of exported plastic waste from official data, adds field investigation, and ends with demands. The four layers are not equally solid.

What does the report not tell you?

It does not say what share of exports is dumped illegally, compare treatment costs at home and abroad, or give data on why any exporter chose a destination. Exporting for recycling is legal (Greenpeace UK, 2021, p. 9), so a shipment to Turkey is not automatically a dumped shipment. One Turkish recycler even told journalists the industry was "desperate for plastic scrap" (Greenpeace UK, 2021, p. 18).

Two more cautions. The figure of 39 per cent for 209,642 tonnes implies about 540,000 tonnes of total exports in the trade data, below the 688,000 tonnes from the Environment Agency database; the report does not explain the gap. And multiples: from 11,626 to 209,642 tonnes is eighteen times, an extra 198,000 tonnes or so (Greenpeace UK, 2021, p. 13). "Eighteen times" sounds bigger because the starting point was small; the post on relative and absolute risk deals with the same trap.

How did Low and Yeats (1992) test the pollution-haven hypothesis?

They did not follow individual factories. They treated trade flows as a sign of where production sits (Low and Yeats, 1992, p. 102). The underlying question: how far do stricter standards in rich countries induce investors to move production to countries with lower standards (Low and Yeats, 1992, p. 89)?

"Dirty" meant 40 product groups whose US producers spent about 1 per cent of sales or more on pollution control; the US industry average was 0.54 per cent and the highest, cement, just over 3 per cent (Low and Yeats, 1992, p. 91). The yardstick was revealed comparative advantage (RCA): a country has one in an industry when the industry's share of its exports exceeds its share of world trade (Low and Yeats, 1992, p. 90).

Dirty goods fell from 18.9 per cent of world trade in 1965 to 15.7 per cent in 1988 (Low and Yeats, 1992, p. 92), yet South-East Asia's share of world exports of them rose from 3.4 to 8.4 per cent (Low and Yeats, 1992, p. 93). Across 109 countries, the average number of developing countries with an RCA above 1 in a dirty industry rose from 10 to 14, or 40 per cent, against 8 per cent in other industries (Low and Yeats, 1992, p. 98). In absolute terms that is four more countries per industry. The chapter says that "some factor is exerting a locational pull" on developing countries towards the polluting industries (Low and Yeats, 1992, p. 98), and does not name it.

Does the 1992 study rule out pollution havens?

No, and the authors say so. Without the rising trend in developing countries it would have been straightforward to conclude that policy differences did not influence where capital went; as it is, "this possibility cannot be dismissed" (Low and Yeats, 1992, p. 103). I read "possibility" as the chance that policy did matter. They also concede that their "aggregate data cannot capture particular instances of environmentally motivated investment decisions" (Low and Yeats, 1992, p. 103).

On the other side, the chapter points to modest pollution-control costs in the US (Low and Yeats, 1992, p. 89) and to another explanation: about 33 per cent of the dirty industries lean towards natural resources (Low and Yeats, 1992, p. 98). The authors lean towards "unlikely to be adequately explained by environmental policy", but on the strength of US cost data from other studies, not their own trade analysis (Low and Yeats, 1992, p. 103). A short summary such as "not mainly due to environmental policy" is therefore stronger than the chapter's wording.

Do the two sources contradict each other?

No: they answer different questions, so one result cannot count against the other.

The 1992 cost argument, that pollution control is about 1 per cent of sales, suits factories. For waste I suspect compliance cost is the treatment cost itself, so the argument may not carry over; neither source gives the price of accepting waste abroad against the cost of treating it at home. Greenpeace's own figures also complicate the picture of rich countries pushing waste on poor ones: in 2020, 80 per cent of UK plastic waste exports went to members of the Organisation for Economic Co-operation and Development (OECD), Turkey among them (Greenpeace UK, 2021, p. 21).

What fits the hypothesis is a pattern: when one destination closes, waste moves to the next. Turkey banned many types of plastic waste from 1 January 2021, yet in February 2021 the UK sent 30,300 tonnes, more than double the 14,000 tonnes of February 2020 (Greenpeace UK, 2021, p. 18). Greenpeace writes that UK companies "responded" to the growing number of import bans by sending more to Turkey (Greenpeace UK, 2021, p. 10); that is the report's reading of a series, not data on why each company chose. Demand from Turkish recyclers would explain the same series.

What should you ask a plastic-waste contractor?

Ask for the final destination of each polymer, not just the word "recycled". Greenpeace says the UK counts all exported plastic waste as recycled (Greenpeace UK, 2021, p. 9); in 2020, 688,000 of the 1,174,000 tonnes of plastic packaging in that category were exported, about 59 per cent by my arithmetic from two figures on page 8 (Greenpeace UK, 2021, p. 8).

1. What is my waste made of? Nearly half of UK plastic waste exports are mixed plastic, styrene or PVC, materials not easily recycled (Greenpeace UK, 2021, p. 8).
2. Where is the first receiving site and where is the final processor, what paperwork exists, and does the contract let me inspect?
3. If the receiving country bans or limits imports, where does the load go next? Turkey cut the share of capacity recyclers could use for imports from 80 to 50 per cent in September 2020 (Greenpeace UK, 2021, p. 18); according to the 2021 report, Vietnam is among the countries that have restricted plastic waste imports (Greenpeace UK, 2021, p. 10).
4. Is the price of taking my waste unusually low? According to reported claims, some refuse companies in Turkey were paid to take containers of contaminated waste (Greenpeace UK, 2021, p. 13); to me that is a prompt for questions, not proof.

This suits you if you contract directly with the contractor and ship enough for them to answer. It does not suit you if you pass waste to a small collector and do not know who takes the final load; ask who the final contractor is first. The four questions are my own inference from the report, not tested with any contractor in Vietnam.

Going deeper: what did I read, and what could I not check?

I read the whole Greenpeace UK (2021) report Trashed, 28 pages, as the public PDF; page numbers are the printed ones. I did not check the underlying data in the endnotes, so every figure is as the report gives it, from 2016 to February 2021.

I read the Low and Yeats (1992) chapter in the World Bank's open PDF, Discussion Paper no. 159, pages 89 to 103. The PDF's text layer garbles many table digits, so I used only figures stated in the prose and Table 6-1. The chapter itself says forty industries (p. 91) and Table 6-4 has 40 rows, while the overview of the same volume says 43; I follow 40. The overview also says the chapter is inconclusive on why dirty industries spread. The data end in 1988, and I have read no post-1992 work on the hypothesis, so this post says nothing about where the field stands today.

The full text of Clapp (1994) in Third World Quarterly is behind a paywall and I could not open it, so this post does not use it.

The post does not show that UK plastic goes to Turkey because rules are looser there, nor the opposite.