Another regular challenge we face concerns the way our sources characterise “returnees”. Sometimes the term refers to IDPs who are said to have returned to their place of origin or habitual residence, but it is also used to refer to people who have returned to their country of origin having fled across a border. In both cases, the people concerned may or not have returned voluntarily, and they are still defined – if not necessarily counted – as displaced until they are able to achieve a durable solution. They may also be displaced again if they find themselves facing renewed violence or another hazard (see part 2). To make things even more confusing, sometimes data on returnees is needed to measure the number of displacements, rather than the number of people displaced at a particular moment in time. In El Salvador and Iraq, for example, our sources surveyed several hundred thousand people who said they had been displaced by conflict and violence during 2016, but many of whom were no longer displaced at the time they were interviewed. While these people are not included in our year-end figures, we nevertheless need to account for these incidents of displacement. Why more investment is needed in monitoring over time One of the key gaps in our data and knowledge concerns what happens to people once they have become displaced. This information is absolutely vital for measuring the extent to which the global target of reducing internal displacement by 50 per cent by 2030 is being achieved. Most importantly, without longitudinal data it is impossible to know how many IDPs there are a given moment in time. This explains why we have struggled to produce stock figures for displacement associated with disasters. More importantly, halving the number of IDPs is not simply about bringing the numbers down. It’s about ensuring that the people counted achieve durable solutions. Without multiple comparable data sets on different situations and how they have evolved over time, there is relatively little empirical evidence that indicates what works and what doesn’t. Having this information would remove some of the guesswork involved in humanitarian and development financing. It is needed to monitor and evaluate programmes intended to benefit IDPs, to hold those responsible to account and to ensure that limited funding is allocated effectively. Every time we compile our annual figures for displacement associated with conflict, we find ourselves asking the same questions. Should we use data we believe to be out of date? And if so, how? As we attempted to produce a global stock figure for people displaced by disasters for the first time this year, we confronted the same problem in a different form. What do we do when data stops being collected before the number “goes back to zero”? Do we keep counting these IDPs? And if so, for how long before we feel that the data no longer accurately reflects the situation on the ground? Outdated information on certain conflict caseloads Last year, we introduced the concept of “decaying” data and presented our estimates for displacement associated with conflict based on the age of the source data. We have done so again this year. Around 93 per cent of the people displaced by conflict as of 31 December 2016 are accounted for by data that was last updated during the year (see figure 3.6). That said, a small number of our estimates are based on old data sources, some of them more than a decade old. Figure 3.6: People internally displaced by conflict and violence as of 31 December 2016, by year of latest data update Source: IDMC INSIDE THE GRID: Overcoming data shortfalls 77