Perils known and knowable: understanding illegal labor risk in fashion supply chains
Can brands claim ignorance of illegal supply chain subcontracting? Research suggests the answer is no
One of the most publicly distressing aspects of modern supply chains in many industries is the subcontracting of work by well-known brands to generally reputable suppliers who in turn send some of that work to suppliers with poor working conditions. The Rana Plaza collapse in Bangladesh in April 2013, which killed 1,129 people, revealed the horrible aspects of this practice in the fashion industry. Though reforms were implemented in Bangladesh after that event, problems persist in many locations, especially in the smaller factories that produce for local, non-Western brands. Moreover, dangerous workplaces persist to this day in many other nations. In China, for example, one factory sent production for a major Australian brand to a forced-labor facility in North Korea, and in Vietnam garments destined for suppliers to multinational companies have been produced under degrading conditions in detention centers.
This problem is not limited to apparel, of course. Apple, for example, has been in the headlines for outsourcing work to partners with poor working conditions. As one 2020 book noted, the conditions at its most well-known Chinese supplier were brutal:
Foxconn’s brutal work schedules, suicide nets outside buildings to catch falling bodies, corralling and abuse of interns, desolate atmosphere and indifferent safety practices are all part of a system that builds fortunes at extreme human cost. A commercial model in which a brand takes 50 per cent of the cost as profit is unusual, and so the wages and working conditions demanded by Foxconn are a direct consequence of Apple’s approach.
Western brands often note that they are unaware when their orders are passed to abusive subcontractors that may be buried two or three layers down in their supply chains. At first glance, this appears to be a reasonable defense. It would seem difficult for a procurement team sitting in Amsterdam to know the conditions of every small factory working for them on the other side of the world. However, recently updated paper by a team of international researchers (Felipe Caro, Leonard Lane and Anna Saez de Tejada Cuenca) examines subcontracting in the fashion industry to ascertain (a) the drivers of unauthorized outsourcing and (b) whether Western brands could predict with confidence the kinds of orders that were more likely to be contracted out illegally. Their research and conclusion make for compelling reading.
What drives illegal outsourcing?
For context, it helps to understand the typical order cycle in the apparel industry. In general, brands send their work to major suppliers who, in turn, often use a third-party to hire subcontractors to complete the work. In theory, these “middlemen” are supposed to alert the brands whenever this happens. In reality, though, when work goes to unauthorized subcontractors, brands are almost never informed. It’s this scenario the research team wanted to understand most: an instance of unauthorized subcontracting “when a supplier does not inform the buyer and the order is subcontracted to an unauthorized facility.”
To find the answers to their two major questions, the researchers acquired data from an apparel middleman covering eight months of operation. The data consisted of 32,000 orders from 226 factories delivered to 30 European fashion brands. Across the data set, the researchers observed major differences among factories:
Unauthorized subcontracting is reported in 36.4% of the orders in our sample. However, we find that there are big differences among factories. We observe three consistent patterns. For some factories, there is no incidence of unauthorized subcontracting; for others, it occurs in 100% of the orders; and, for a third group of factories, it happens occasionally.
With data in hand, the team tested six hypotheses:
Hypothesis 1: Unauthorized subcontracting is more likely when the previous order in the same factory was subcontracted.
This hypothesis was supported: “When a factory subcontracts an order, the next order this factory delivers is about 31 percentage points more likely to be subcontracted too than if the previous order had not been subcontracted.” In other words: “We find that the probability of unauthorized subcontracting increases by 87%, i.e., almost doubles, when the previous order at the same factory was also subcontracted without authorization.”
Hypothesis 2. Unauthorized subcontracting is more likely for orders that have a lower unit price.
This hypothesis was supported: “When a buyer offers a unit price for an order that is lower than the factory’s usual price for that product category, the probability of unauthorized subcontracting is higher. For instance, if the price pressure of an order is such that the unit price is 25% lower than the usual price, the chance of unauthorized subcontracting would increase by at least 9%.”
Hypothesis 3. Unauthorized subcontracting is more likely for orders that have a short lead time.
This hypothesis was not supported, as shorter lead times did not impact illegal subcontracting decisions. This makes intuitive sense, as outsourcing incurs time and delivery risk, not desirable, and the primary producer typically gets a, desirable, price premium for expedited work.
Hypothesis 4. Unauthorized subcontracting is more likely for orders that have a long lead time.
This hypothesis was supported: “When an order is placed earlier with respect to its scheduled due date, the probability that it is subcontracted is higher. Every extra day is associated with an increase in the probability of unauthorized subcontracting of around 0.09 percentage points, or 0.25%. This estimate supports the notion that orders of basic (i.e., less sophisticated) products are more likely to be subcontracted than fashion items because the former have markedly longer lead times than the latter.
Hypothesis 5. Unauthorized subcontracting is more likely for orders of categories other than the factory’s specialization.
This hypothesis was not supported, as the data showed no relationship between the nature of the order placed and the production specialization of the order taker.
Hypothesis 6. Unauthorized subcontracting is less likely when the buyer is a well-known brand.
This hypothesis was supported: “The evidence in practice has demonstrated that stronger brands face a higher cost and more reputation damage when disasters related to non-compliance occur. Second, when the final buyer is a well-known brand, the probability of unauthorized subcontracting is 22% lower.”
In their analysis of the six hypotheses, the research team made a critical finding, namely that “the propensity of each firm to engage in unauthorized subcontracting emerges as an inherent characteristic or strategic choice that can be (partially) inferred from past behavior.” Consequently, they constructed “a simple linear prediction model to detect unauthorized subcontracting at the order level.” Their model leveraged the findings on the drivers of unauthorized subcontracting, including order and factory characteristics.
By using their new model, the authors concluded that a theoretical middleman with a similar analytical tool can “correctly predict whether or not unauthorized subcontracting will occur for more than 82% of the orders, and can do so using information at his disposal.” Moreover, “if it were possible to observe the factory’s decision for every order it produced, then by adding a lagged dependent variable it would be possible to predict unauthorized subcontracting with more than 92% accuracy.”
Put simply: if a middleman, or even a brand itself, wanted to know if a given order was likely to be sent to an illegal contractor, it could do so with almost perfect certainty. In short, “unauthorized subcontracting is predictable, and therefore can be managed.” Indeed, as the authors note:
...some variation of our models can be embedded in a decision support system to help inform the decision on where to place orders or what suppliers to work with. Our results also show the need for more collaboration between buyers and suppliers. Though price is always important, we find that working closely with the supplier to avoid situations (or states) that push the factory into unauthorized subcontracting can also be effective. For instance, coordinated planning and scheduling can prevent periods of excessive workload. Implementing measures to reduce unauthorized subcontracting can help buyers reduce compliance risk and increase supply chain visibility.
The research presented in this paper should be analyzed carefully and considered thoughtfully by those responsible for sourcing of products from supply chains with high illegal outsourcing risk. If three researchers working with limited time and resources can produce such promising findings, one imagines what multi-billion dollar brands with sophisticated analytical teams and systems could create. This conclusion suggests that any claim that Western brands can’t know when their orders are likely to wind up in the wrong hands may not be as defensible a position.
Felipe Caro, Leonard Lane, Anna Sáez de Tejada Cuenca. Can Brands Claim Ignorance? Unauthorized Subcontracting in Apparel Supply Chains. Management Science 67 (4) 2010-2028 https://doi.org/10.1287/mnsc.2020.3679