Does destroying ivory stockpiles lead to a decline or an increase in the illegal ivory trade? Have the two highly regulated sales of ivory – the first to Japan in 2000 and the second to Japan and China in 2008 – had an impact on the illegal ivory trade?
In this post I look at what we need to think about to assess the evidence to answer these kinds of questions.
An approach that is often seen when trying to understand the impact of the sales of ivory is to look at the best available estimates of trends in the illegal ivory trade (for example by using the Transactions or Weights Index) over time. If the amount of, or trend in the, trade appears to change at times when specific interventions occur this is used as evidence that these interventions have had an impact on the trade. For example it is clear that there is an increase in illicit ivory activity from 2008 onwards. The fact that the sale of ivory to China was also in 2008 is often used as evidence that the regulated sale is the main driver of the illegal ivory trade.
But, just because these two things happen at the same time doesn’t mean that one causes the other – this is that well known truth that “correlation does not prove causation”. For a start, you would need to be able to describe the mechanisms by which the sale would lead to an increase – in this case there may be many plausible mechanisms to consider.
More importantly however, is that to gain an understanding of whether or how the sales have had an impact it is also important to understand how other factors (I will call them potential drivers) might influence the trade and how these potential drivers may interact with each other. To do this you need to look at all parts of the trade chain. That is drivers that may lead to increases or decreases in the supply of illicit ivory, drivers that determine how and where illicit ivory is transported and drivers that determine the sale and demand for illicit ivory would all need to be considered.
Why is it important to consider all the drivers? Well, without doing this you could draw the wrong conclusions.
For example you could wrongly attribute the increase in illicit trade to an intervention (for example the regulated sales) when actually something else has happened that also explains the change or most of the change (for example increasing wealth in potential consumer populations leading to an increase in demand). Alternatively you could wrongly assume that an intervention has not had an effect because its effect is masked by something else. For example suppose that there is no change in global illegal activity after an intervention such as the destruction of a country’s stockpile. Without considering other drivers the inference would be that this is because destroying the stockpile doesn’t have an effect on trade. In fact it could be that the destruction of the stockpile reduced illicit trade by some mechanism but other drivers of the trade have led to an increase in activity so that the overall effect is that there is no change in activity.
Considering all the drivers is, of course, possible without looking at any data at all. That is what rational and well thought out debate and enquiry requires.
Wanting quantitative evidence takes more steps. It requires identifying potential data sources that measure these different drivers, or if data is not directly available it requires identifying proxies for the drivers. It requires mechanisms for drawing together, and combining the data to represent how the different drivers interact on the appropriate spatial (national park, country, region) and temporal (one off events, annual cycles, long-term trends) scale. This would then enable an examination of the relative importance and effect of different drivers and provide the quantitative evidence for and against specific interventions.
This is not, by any means, a trivial exercise. It requires the cooperation of many different groups who understand different aspects of the trade – ecology, criminology, economics – and working with those who know how to combine, evaluate and assess data. It is particularly non-trivial because we are talking about an illicit and therefore mainly unobservable trade that operates on an international scale. This makes it challenging (but not necessarily impossible) to disentangle the effect of different interventions because we can’t watch it unfold in many different places independently from each other – it is all mixed up together. (I will write about this in more detail in a future post).
But without making an attempt to do this it is not clear how we assess the success or not of different interventions for both elephants and other species under threat now and in the future.
What do you think about this approach?