The Impact of Migration on Political and Social Beliefs
Final Report Abstract
Our original project planned to conduct research with university students from Pakistan who left to study abroad. However, just a few months after receiving the grant, the COVID-19 pandemic began. In light of this massive, unanticipated disruption, we discussed with the DFG our desire to move forward with the project answering a similar question in a different setting. Below, we explain the reasons for the change and then explain our adapted plan. Context: Changes to Research Plan As development economists, we expect challenges when working in developing countries as part and parcel of the work. The very reason why we focus on countries is to understand their low state capacity and instability, which can directly potentially jeopardize our work. Hence, we had taken the necessary precautions to keep into account potential delays, potential changes in experimental design, and multiple channels for collaboration. However, the nature of the shocks was unprecedented. Together, the COVID-19 crisis and an IMF program (which led to substantial budget cuts for education) directly impacted the work of the Higher Education Commission Pakistan – our main counterpart. The pandemic brought life in general to a standstill, but it brought unprecedented economic collapse in developing countries, including Pakistan, in particular. Due to these challenges, our partner’s attention shifted to crisis mode to ensure the viability of the public higher education and ensure that remote learning could take place in a poor country with patchy internet connectivity. All travel to Europe came to a standstill, and visa services remained closed for over two years. In-person contact was limited and it was impossible to plan ahead with regular, unanticipated lockdowns due to new waves of the pandemic. The lockdowns, disruptions in global trade, and a slump in domestic demand led to an economic contraction. This further reduced the space for our partner to continue its work with substantially strained budgetary space as the government diverted resources towards healthcare and relief efforts. In addition to this, Pakistan entered into a challenging IMF program, seeking financial assistance to stabilize its economy. While the program aimed to address longstanding structural issues, it demanded stringent fiscal and monetary measures. Unfortunately, these reforms, including austerity measures and subsidy cuts, created a ripple effect, contributing to economic instability. The prescribed devaluation of the currency led to inflationary pressures, eroding the purchasing power of the common citizen, which led to social unrest. Around 18 million people were forced into poverty, as the number of poor increased from 69 million in 2018 to 87 million in 2020 in a country of around 200 million. New Plan In view of the above, instead of working with students going abroad, we decided to work with students inside Pakistan. We looked at various options for collaboration and settled with the Government College University (GCU), Lahore. GCU is one of Pakistan’s oldest public-sector universities. It attracts students from low-to-middle income households, coming from small villages and far off, economically underdeveloped areas. This socio-economic segment was our original target. Together with GCU, we created two courses that were developed by the research team to teach cognitive and non-cognitive skills. The goal of the cognitive skills courses was to expose the students to knowledge from economics, philosophy and psychology. Thus, the students were exposed to content that is generally not taught explicit at the university, except for the economics content. Hence, they were exposed to ideas surrounding logic and rationality. The courses were taught mostly by faculty trained at universities in Europe and the US. The noncognitive skills course taught student professional communication, grit, and planning. This new plan stayed as close as possible to the original idea. It shifted the focus from foreign countries to Pakistan and taught them subject matter focusing on ideas that the students would usually not be taught. These courses were randomly offered to over 300 students. We solicited interest in the courses from the student body of GCU (4th semester undergraduate students as the study was done in the Spring semester). After opt-in, every student was equally randomized into one of the three groups: cognitive skills course, non-cognitive skills course and a pure control group. We measured various outcomes at baseline and endline at the end of the semester. At baseline, we asked about political beliefs and preferences with regard to voting intentions, knowledge of economic affairs, belief in free press and utilization of different media sources and trust in them. In addition to this, we took the opportunity to ask about different actors related to financial decision-making in health. At endline, we repeated the same questionnaire. Under this revamped research plan, we used the funds for the doctoral researchers who ensured this collaboration functioned, designed the survey instruments, collected and analyzed the data. Further, the funds were used for fieldwork in Pakistan: for research assistants who managed the work with GCU on a daily basis, and enumerators for the data collection. Unfortunately, our analysis did not yield the hypothesized results. While we were able to confirm a successful randomization with balancing checks from our baseline enumeration and attrition between baseline and endline was low and equal across groups, we did not find robust significant effects of participation in the cognitive skills or non-cognitive skills course on any of the measured outcomes. We ran cross-section OLS regressions (for outcomes only measured at endline) and Difference-in-Differences regressions (for outcomes measured at baseline and endline), where we regressed various outcome variables on the assignment to either of the two courses. To determine significant effects, we did interpret a) the coefficients for the assignment to either course in comparison to the control group as well as b) a test for significant differences of the two coefficients on the two course assignments. While the former test was meant to test the effect of any Western-offered course on students’ political beliefs, the latter was to test whether a cognitive skills course in particular would affect students’ political beliefs. We hypothesized that teaching cognitive skills would help students consume news and tell fake news from actual facts, e.g., by distinguishing correlation from causation. We did however find that neither of the tests yielded significant results. This is, neither did offering Western courses change any of the outcome variables, nor did teaching cognitive skills exhibit a particular effect. Our main outcome variables were students’ intention to vote, how well they follow news about the IMF program as well as political affairs in particular, and whether they think that in general, all news can be trusted. Neither was significantly increased or decreased by either of the two courses. We trace these non-results to the course being not sufficient for changing students’ cognitive skills. We measured students’ cognitive skills with Raven’s Progressive Matrices, a common cognitive skills test in psychology. We did not find an increase in cognitive skills in students that took the cognitive skills course, neither in absolute value nor relative to the two control groups. Apparently, one course is not enough to improve students’ cognitive skills. Instead, universities (or better schools) would need to change their entire curriculum to favor cognitive skills over learning by heart from early on already. Return of Unused Funds We conducted this research over one semester and did not find statistically significant effects. This can be due to multiple factors: a weak treatment, hard to change beliefs, or a small sample. We had initially planned to continue the experiment for more semesters. However, given these weak results, we do no longer expect that the intervention would move outcomes if continued for longer. The research team is also not of the belief that a bigger sample would change effect sizes. Instead of keeping the remaining funds tied up for a longer period to explore alternative settings, we see it best to return the unspent part of the grant intended to conduct the experiment. As a consequence of the above-explained unforeseen problems, the two PhD students funded with the grant used part of their time to pursue alternative research projects, which due to the Corona-induced setting in part involved non-experimental research. In spite of the described setbacks, both of them very successfully finished their PhD studies. Tobias Korn is now PostDoc at the University of Hannover; Zain Chaudhry is economist at the Mind, Behavior and Development Unit (eMBeD) at the World Bank.
Publications
-
Your Pain, My Gain. Estimating the Trade Relocation Effects from Civil Conflict. Hannover Economic Papers, DP-698
Korn, Tobias & Henry Stemmler
-
Fickle Groups: A Field Experiment on Time Preferences
Chaudhry, Zain & Karrar Hussain
-
Mobilizing women voters: experimental evidence from Pakistan, Oxford Economic Papers 75, 2: 444-459.
Chaudhry, Zain, Karrar Hussain & Attique Ur Rehman
-
The Persistent Consequences of Civil Conflict. Evidence from a New Measure for Subnational Conflict Exposure.
Korn, Tobias
