The relationship between refugee flows and the incidence of conflict has troubled security analysts and policymakers, as well as the humanitarian aid community, for some time. Although the connection has long been recognized, little systematic work has been done to pick it apart. We know that some refugee communities are conflict prone while some are not; some refugee movements appear to induce substantial regional insecurity while others do not. Why?

The author's argument (pp. 5, 20–34) in explaining this variation highlights the significance of the political context in which these movements occur. First is the character of the refugee population. Situational refugees leave in panic and are not politically organized. Persecuted refugees leave because they are directly persecuted, and may or may not have existing political structures. State-in-exile refugees move as a result of a decision by a political leadership as part of the latter's strategy to prevail in conflict. The three categories together form a low-to-high spectrum of violence proneness.

Second is the nature and response of the receiving state. The capacity of the state to control borders and to supervise refugees within those borders, and the interface between the political interests of the receiving state and the conflict in the state of origin, are particularly important. Weak states find it difficult—whatever their intent—to contain the propensity of organized refugee groups for violence. Capable states may control, tolerate, or promote the politico-military activities of groups that they host, depending on their larger objectives regarding the sending state.

Sarah Kenyon Lischer has chosen her cases (Afghanistan, Central Africa, and the Bihac Pocket) well. The Afghan war produced two major flows of refugees. In Pakistan, the flow of refugees into Afghanistan laid the basis for prolonged cross-border insurgency. In Iran, there was no significant cross-border military activity. Since both states were comparatively capable, the difference may be explained largely by variation in the policies of the two receiving states toward the conflict in Afghanistan and the consequent degree of restraint they imposed on refugees.

In the Central African case, Lischer's analysis of Zaire yields little that is new. The discussion of flows from Burundi and Rwanda in Tanzania provides considerable value-added. The flow of Rwandan refugees into Tanzania produced little subsequent cross-border violence. That from Burundi was associated with substantial cross-border violence and resulted in significant increases in tension between the two states. Both groups had grievances against their home governments. The host government was clearly capable of controlling cross-border activity in either case. Lischer argues convincingly that the variation was produced by the different nature of Tanzania's political relations with the sending governments.

In short, the author asks a good question, develops a sensible comparative methodology for addressing it, chooses her cases well, and produces a very credible analysis. Dangerous Sanctuaries makes a useful contribution in unpacking the ways in which political factors affect the propensity for insecurity associated with refugee movements. As she argues, the analysis here is more nuanced and therefore more useful both analytically and in policy formulation than more generic socioeconomic explanations of refugee-related conflict.

This leads to a few more critical observations. A considerable amount of time is spent in the book debunking the socioeconomic explanations (camp location and size, poverty levels, gender balances within camps, lack of productive activity) for violence. This is overdone. What Lischer has shown is that these factors do not appear to be determining in the cases she analyzes. It remains, however, intuitively plausible that, all other things being equal, these factors are positively related to a community's propensity to conflict. In other words, these factors may have causal significance at a structural level, but by themselves are insufficient to explain specific outcomes. If so, it makes sense as part of a strategy of preventing or mitigating refugee-related violence to address these factors. The instruments of humanitarian agencies are not well suited for addressing political problems; they are relevant in dealing with these socioeconomic conditions.

The critical analysis of socioeconomic explanations of refugee-related violence is linked to a wider critique of the role of humanitarian actors in fueling violence. This also seems overdone. After all, in the absence of politico-military support in disarming refugee warriors and separating them from noncombatants, and given the reluctance of humanitarians to let people starve on political grounds, there is little they can do about the unintended consequences of their actions. The critique is not new and is widely understood in the humanitarian community. It also confuses the argument. While the author expends considerable effort debunking socioeconomic factors, here she presents a socioeconomic factor (aid) as an important contributor to that violence.

This is linked to a third point. The book gives a monolithic impression of the humanitarian community as unwitting and naive contributors to refugee-related violence (see p.141). In so doing, its record of the humanitarian community takes insufficient account of the relatively high level of humanitarian awareness of the political and conflict contexts in which they work, and of their concern about the impact of their activities on conflict. Humanitarian agencies have come a long way in the past decade in their realization that humanitarian action does not proceed in an apolitical void and that they need to take account of the negative unintended consequences of their work. The author's own account establishes the point (e.g., the withdrawal of some agencies from the camps in eastern Zaire, and the Médeeins sans Frontières advocacy of military intervention in Kosovo). A number of the sources she cites (e.g., Fiona Terry, pp. 91–92, 160) in support of her criticism of humanitarian agencies are from humanitarian agencies.

The weakest section of the book is the conclusion. Oddly, given the fact that the core of its argument is about states, political communities, and political processes, the concluding chapter is essentially about what humanitarian organizations need to do about it. Of all the institutional players in these games, humanitarians are the least well equipped to address the politics of refugee-related violence. The onus here lies on the UN Security Council, the regional organizations, and the states themselves. The aid-fuels-conflict dynamic is essentially the result of abdication of responsibility by all of the above.

On the other hand, it is right that humanitarians have their own contribution to make. Many of the book's recommendations are sensible in the abstract, but hardly new and difficult in practice. It seems rather banal to suggest that aid organizations should pressure states and organizations to provide police and military support (p. 143). They do pressure … and they are seldom listened to. Or that they should publicize the need for security measures (p. 149). They do … and they are often ignored. Or that they should build security partnerships with local players (p. 151). They try … but, for reasons well explained in the book, the locals may be unwilling or unable to deliver. As for the withdrawal of humanitarian assistance to those in need (p. 143), humanitarian organizations do consider it and, on occasion, they do it. It is more difficult for them than for the author, because their commitment is humanitarian and they know that the people they are working with may die as a result of their decision.

These criticisms aside, the core argument and analysis are useful contributions to the growing literature on the relationship between forced migration and security.