Million Dollar Question: Does Fairtrade Work?

June 06, 2014

You may have recently seen articles in one of several media outlets about a new study that purportedly shows that Fairtrade has failed African farmworkers. Does that mean you should stop buying Fairtrade coffee and other products?

Not so fast. We hate to shake your faith in the mainstream media, but sometimes journalists misinterpret or exaggerate the conclusions of academic studies, so don’t regret that latte just yet.

Basically, the study doesn’t give us enough evidence to know whether Fairtrade works or not.

If you want to know why this study doesn’t answer this important question, sit down, pour yourself a cup of coffee (Fairtrade or not!) and let us explain.

The UK’s Department for International Development wanted to know if Fairtrade certification actually improves growers' lives, so they paid the Fair Trade, Employment and Poverty Reduction Project at London’s School of Oriental and African Studies (SOAS) more than $1 million to investigate.

The researchers identified twelve agricultural production sites across Ethiopia and Uganda, some with Fairtrade-certified farms and some without. They then gathered detailed data on employment, wages, and working conditions for agricultural workers living at these sites.

So they looked at what happened to Fairtrade farmers?

They actually studied farmworkers, not farmers. These are the people who get hired by farmers to do the heavy lifting, usually on a seasonal basis. They exist at the very periphery of the agricultural value chain and so are particularly difficult for Fairtrade benefits to reach.

And the study found that Fairtrade didn't help, right?

The headline result is that people working for wages at Fairtrade-certified sites earn less, on average, than those working on regular farms. They also appear to have worse work environments and fewer days of actual work. This has certainly upset fans of Fairtrade.

So the answer to the million-dollar question is that Fairtrade hurts workers?

Well, maybe. The problem is that this research wasn’t designed in a way that can tell us whether working on a Fairtrade-certified farm makes people better (or worse) off.

When we ask “does Fairtrade make farmworkers better off?” we actually mean “… compared to how well-off they would have been without Fairtrade”. But it is actually quite difficult to know how well-off a worker would have been if s/he hadn’t been working on a Fairtrade-certified farm.

Surely comparing workers in Fairtrade areas to those in non-Fairtrade areas answers this question?

Sadly not!  Suppose you were a farmer who lived in an area with a high risk of weather shocks such as droughts or floods. Then you might try harder to get Fairtrade certification to give yourself a more stable income.  It may be that such farms do less well than non-Fairtrade farms, but better than they would have done without Fairtrade.

The only way to learn what difference Fairtrade makes is to compare non-Fairtrade farms to Fairtrade farms that were otherwise identical before Fairtrade was introduced. The best way would be to randomly select a sub-set of very similar farms to become Fairtrade, which could be quite difficult. At a minimum you would want to compare sites which were similar prior to the intervention, to minimize the risk that the areas are inherently different for reasons other than their Fairtrade certification.

But the SOAS report doesn’t do this – it focuses on a small number of sites which the authors argue look similar after some of them adopted FT certification. In this case there is no counterfactual: there’s no way of knowing if these sites were truly comparable at the start.

That means we don’t know if the differences in wages in the two are because of Fairtrade or because of other differences between them such as transport links, soil quality, norms governing labour relationships, or dozens of other factors we could imagine. (Our CGD colleague Kim Elliott has written a comprehensive primer on Fairtrade markets and detailing some of the trade-offs faced by farmers seeking the certification).

So we cannot work out from this study whether Fairtrade improves workers’ lives?

Unfortunately not. The full report contains an important and careful descriptive analysis of the welfare effects of agricultural production across the twelve sites, but the design of the research means that it does not, and never could answer the million dollar causal question of what wages would look like without Fairtrade certification.

Also, given that Fairtrade certification differs across sites and not within them, even if you believed that the differences between the two groups were caused by Fairtrade, once you make the right statistical correction for the small number sites (or `clusters'), it would be very difficult to uncover a statistically significant result.

But didn’t the study employ some fancy statistical techniques to solve these problems?

The authors do recognize that the differences in wages might be due to `unobserved’ characteristics, so they use propensity score matching (PSM) to get around this problem, a method which essentially compares workers that are very similar along a host of dimensions. In order for PSM to work, you need to match on characteristics that are unaffected by Fairtrade certification. We don’t know for certain as the data haven’t been published yet, but it appears that the authors match on several characteristics which would be affected by certification, such as poverty levels and days worked, which would invalidate the results.

Sadly, given the structure of the study, even doing the PSM procedure correctly wouldn’t allow us to establish whether Fairtrade works. As measured, certification only differs across sites, but none of the controls available to the researchers accounts for other site-level differences. Once again, we’re left with the worry that wages differ because these locations and labor markets are inherently different, not because of their certification status. 

Great, so the Fairtrade movement is vindicated?

Not so fast. Just because this report doesn’t really answer the question does not mean that Fairtrade is successful. We think the burden of proof is not on the rest of the world to prove that Fairtrade doesn't work, but for the proponents of certification and the brand itself to prove it ‘does what it says on the tin’: that it benefits the poor compared to what their lives would be like in its absence.

While the report doesn’t actually furnish us with a clear answer, the commentariat jumped on the study to criticize Fairtrade: but that is either ignorance or wilful sensationalism. The recent debate has also ignored previous evidence suggesting there are modest gains from certification schemes, although many of those studies are also susceptible to the critiques we make above. 

While the SOAS report cannot provide clear evidence on the causal impact of certification, it does provide valuable insights into the lives of those working for certified farms. It takes care to evaluate the twelve survey sites with a combination of qualitative and quantitative surveys, generating a rich source of non-typical information about labour market dynamics.

The million dollar question, however, still goes unanswered.   


CGD blog posts reflect the views of the authors, drawing on prior research and experience in their areas of expertise. CGD is a nonpartisan, independent organization and does not take institutional positions.