In his introduction to the review symposium of Why Muslim Integration Fails in Christian-Heritage Societies, Jeffrey Isaac, former editor of Perspective on Politics, characterizes our book as blending “empirical and normative judgments.” This was indeed the main objective we pursued in writing this book: we sought not only to offer an empirical analysis of the challenge of Muslim immigrant integration, but also to think hard about the implications of this analysis for policy and for France’s republican ideals. On the empirics, notwithstanding Norton’s sardonic put-down that what we accomplished empirically “passes for methodological rigor among its champions,” both Amaney Jamal and Abdulkader Sinno in their reviews celebrate our rigor. Both agree that our empirical strategy demonstrates that heretofore insufficiently identified anti-Muslim bias in France is sustained in a discriminatory equilibrium.

But, as was noted in the introduction, we did not stop with empirics. We offered prescriptions on what might be done to undermine this nefarious equilibrium. Alas, we did so without conducting randomized controlled trials or other policy evaluation tools that would have provided greater rigor in evaluating any policy treatments. Our goal was to infer what would be a positive equilibrium shift from the mechanisms we uncovered that sustain it. The discussion of all three reviewers focused – and in view of the PoP mission that political science ought better to speak to normative issues, properly so – on the one chapter of the book that offered remedies.

Both Sinno and Jamal excellently identified the problem we faced, in Sinno’s formation, that it is “tricky to formulate good policy recommendations, which often deal with messy and complicated realities, on the basis of rigorous research that clinically isolates a few suspected causal mechanisms.” Moreover, Sinno recognizes that we relied more on analogy (to remedies tried in other discriminatory domains) than direct evidence on policies towards Muslims in France; similarly for Jamal, who sees the empirical work as rock solid but notes that policy recommendations are not “an easy subject to tackle.” This leads to a series of serious critiques, from both of these reviewers. We address here the recommendations with which they took issue.

(1) Broadcasting of our findings – Sinno points out that this recommendation came from analogy (to professional basketball) rather than our own findings. While basketball referees became less discriminatory after seeing data showing that African Americans were penalized more harshly than Whites, the analogy in his judgment would probably not hold in France. It ignores the fact that the French government, at least up until now, “seemed more interested in marginalizing Muslims than helping them integrate.” If this is so, Sinno asks, from where would these broadcasts originate? While correct about the use of analogy, Sinno perhaps underestimates the power of republican thinking throughout France. There are many in French elite society who have the interest and power to publicize our results that would challenge the current “choice of ignorance,” as articulated by a senior researcher in the state statistical bureau (p. 198 of our book). Our hope was that the book would provide the evidence these elites needed to confront their own society.

(2) Muslims changing names – This was perhaps our most controversial recommendation. Sinno argues that our experiments shed no light on the returns to changed first names. He also worries, for Muslims with neutral first names, that after call back for an interview, discrimination could quickly reappear. Jamal also thinks it is “questionable at best” that the change in names could break the “cycle of discrimination” identified in the book. More broadly, although acknowledging that we demand much more from the French hosts in undermining the equilibrium than we do for the Muslim immigrants, she faults us for asking Muslims to shoulder the “burden…to constantly prove that they are not a) terrorists, b) fundamentalists, c) women haters, and/or d) religious fanatics.” The facts on the ground, she points out relying on Pew Survey data, undermine those very charges. She opines that “to make Muslim identity invisible privileges and reinforces, rather than challenges, the discriminatory equilibrium.”

These critiques are important, and we welcome future research to put our recommendation to direct test. However, a few responses are in order. First, it is not quite correct to say that we had no data supporting our recommendation. As we show in our experiments, Muslim Senegalese subjects with African (but not identifiably Muslim) first names receive more generous treatment from rooted French players than those with identifiable Muslim first names. Second, there is a growing - albeit not entirely conclusive - literature on the effect of name-changes on life outcomes based on the notion that names provide signals to employers about unobservable characteristics of any job applicant. In Sweden, immigrants from Asian, African, and Slavic countries experienced an increase in earnings after a name change (Arai and Thoursie 2013). In Israel, Rubinstein and Brenner (2014) find that Israeli Jewish males born to a Sephardic father and an Ashkenazi mother (who are therefore more likely to carry a Sephardic surname) earn significantly less than their counterparts born of an Ashkenazi father and a Sephardic mother (who are therefore more likely to carry an Ashkenazi surname). And in the United States, relying on a unique longitudinal dataset of naturalization record in the US, Biavaschi et al. (2013) find that migrants who Americanized their names at the moment of naturalization in 1930 experienced gains in occupation-based earnings.

To be sure, and this is in support of our critics’ skepticism on name changes, there are well-identified studies that obtain null results (Abramitzky, Boustan, and Eriksson Reference Abramitzky, Boustan and Eriksson2017; Fryer and Levitt Reference Fryer and Levitt2004). In accounting for their findings Fryer and Levitt (Reference Fryer and Levitt2004) suggest that minority job-seekers who are ignored by discriminatory firms get placed in institutions where they can better thrive, with the implication that in equilibrium the name should have no effect on economic success. Our data from the experimental games suggest that in France, a Muslim-sounding given name would have greater consequences than those found in Abramitzky et al (Reference Abramitzky, Boustan and Eriksson2017) and Freyer and Levitt’s (2004) American data. But further research is clearly needed.

Third, we recognize the emotional trade-off between sending signals that yield social and economic benefits and maintaining naming practices that are reflective of identities that are proudly held. But, as we note (p. 227 fn. 3) from Jhumpa Lahiri’s novel Namesake, among Bengalis in the US and UK there is a norm to maintain a different name for public and private consumption without any loss in dignity. Indeed, undermining equilibria demands the changing of expectations about others on both sides, and is not in any case without ethical dilemmas.

(3) State induced correspondence tests of firms as a regulatory device – We advocate a stricter enforcement of Article 225-1 of the Penal Code with penalties for religious discrimination administered through a public institution that would conduct audits of firms. Sinno points to a variety of difficulties in the implementation of this recommendation. This is undoubtedly true but it underestimates the range of data collection and expertise in surveys within the French administrative state that could be tasked with this important challenge.

Unlike Sinno and Jamal, several of Anne Norton’s charges are misguided. Her reaction to our suggested remedies, that they “would perpetuate, even intensify, discrimination against Muslims in France” is offered without any evidence. Her charge that we “ignored racial difference” is undermined by the care we took in controlling for racial difference. Her charge that we ignored discrimination outcomes beyond France is undermined by the chapter we wrote on Yugoslav refugees throughout Europe and Arab-Americans in the Detroit metropolitan area. The unfortunate charge that the structuring of the book is in line with the “work of LePens and other anti-Semites” enters into the unproductive territory of personal attacks.

The bigger picture, and one that merits continued concern, is that equilibria are stable social outcomes; undermining them when they are discriminatory is a challenging task. Yet ignoring the discriminatory equilibrium – i.e. the “choice of ignorance” – is a violation of human rights and France’s republican aspirations. If our prescriptions are misguided, likely to be ineffective, or inappropriate, it is incumbent on the policy community to do better. It is our hope that this symposium will induce the political science community, in cognizance of the clear empirical results on the fate of Muslims in Christian-heritage societies, to develop policy responses that will work to undermine the discriminatory equilibrium that blocks their success.