Questions & explanations
1. A study uses distance to a job training center as an instrument for program participation. Explain how IV would estimate the effect of training on earnings.
First, check that distance predicts participation: people far away are less likely to join. Also, distance should not affect earnings except through training. The IV method uses this variation: because distance is random in a sense, the part of participation explained by distance is like a natural experiment. It compares earnings of people who participated because they were close vs. those who did not because they were far. The estimate is the ratio of the effect of distance on earnings to the effect of distance on participation. This gives the causal effect of training on those who were induced to participate by distance (local average treatment effect). The key is that distance affects only participation, not earnings directly.
2. Give an example of a humanitarian cost tied to border policy.
One example is the Mediterranean Sea route, where thousands of migrants die each year trying to reach Europe from Africa or the Middle East. Because European countries operate patrols that intercept boats and sometimes push them back, migrants use smaller, more dangerous vessels. Many boats sink in bad weather, and people drown before rescue arrives. The cost includes not only the lives lost but also the trauma for families and communities. Search and rescue operations also cost many millions of euros per year. This humanitarian cost is directly linked to a policy of preventing arrivals rather than providing safe legal pathways. Many experts say that without safer routes, the death toll will continue.
3. What is a 'pushback' and how can it lead to deaths?
A 'pushback' is when border authorities forcibly send migrants back across the border without letting them apply for asylum. Instead of processing their claim, officials simply turn them away, often at a different point. Pushbacks can be violent and can happen at sea, where migrants are put back into unsafe boats and pushed out to international waters. This can lead to drowning if the boat sinks or migrants are left without enough food or water. Even on land, pushbacks may leave people in remote areas where they cannot get help. Many governments deny using pushbacks, but human rights groups and news reports document them. The practice is widely criticized as violating international law.
4. Compare the assumptions and strengths of D-in-D, RDD, PSM, and IV for causal inference.
D-in-D assumes parallel trends between treatment and control groups. It is good for policy changes over time. RDD assumes no other changes at the cutoff and gives local average effect. It is strong when treatment is assigned by a rule. PSM assumes no unobserved confounders conditional on observed variables. It is flexible but sensitive to model choice. IV assumes the instrument affects only through treatment and is not correlated with unobservables. It can handle bias from unobserved factors but is hard to justify. All rely on different assumptions. The best design depends on the context. When possible, RCT is still the strongest, but these are valuable alternatives.
5. Give an example of an ethical rule for immigration detention.
An ethical rule is that detention should only be used when necessary and for the shortest time possible. For example, people should not be held just because they cannot pay bail or lack identification. Another rule is that conditions must be humane: clean, safe, with access to healthcare, lawyers, and communication. Detainees have the right to challenge their detention in court. Special protections are needed for vulnerable groups like children, pregnant women, and survivors of torture. International guidelines, such as from the UN, say detention should be a last resort. When these rules are followed, detention is less harmful, but many countries fail to meet them.
6. What is propensity score matching (PSM) and how does it work?
Propensity score matching (PSM) creates a control group by matching treated units to similar untreated ones based on observed characteristics. First, calculate the probability (propensity score) of receiving treatment given variables like age, income, and education. Then match each treated unit with a control unit that has a similar score. This tries to balance observable factors. After matching, compare outcomes between the groups. PSM assumes that selection into treatment is based only on observed variables (conditional independence). It is useful when you have many measured confounders. But it cannot account for unobserved differences, so it is weaker than RCT.
7. Compare the risks faced by migrants crossing by sea vs. by land.
Migrants crossing by sea face drowning, hypothermia, and attacks at sea. Boats are often overcrowded, poorly maintained, and not suited for open water. Many die before reaching shore, and their bodies are rarely recovered. Migrants crossing by land face dehydration, heatstroke, or freezing depending on the terrain. They may walk for days through deserts or mountains with little food. Both routes have risks of robbery, assault, or arrest. Sea crossings often involve more immediate life-or-death emergencies, while land routes can be longer and expose migrants to extreme weather. Both types of journeys are deadly because of the measures taken to avoid detection.
8. Why do some countries have separate agencies for visas and for border control?
Some countries split immigration functions into separate agencies to specialize tasks. One agency focuses on visa processing and policy, often located in embassies abroad. Another agency handles border security and patrols inside the country. This separation allows each group to develop specific skills and use different technology. For example, visa officers review applications online, while border officers interact with people face-to-face. However, coordination between agencies is necessary to share information and prevent fraud. Some countries keep everything under one department for simplicity. The choice reflects each country's administrative structure.
9. Compare RCTs and quasi-experiments in terms of internal validity and practical constraints.
RCTs have higher internal validity because random assignment removes selection bias. They are the gold standard for causal claims. But they can be costly, unethical (e.g., denying treatment), or impractical (e.g., whole-country policy). Quasi-experiments use non-random methods to mimic randomization. They have lower internal validity because unobserved differences may exist. However, they are often cheaper and possible where RCTs are not. For example, you cannot randomly assign a new law to some cities. Quasi-experiments use clever designs to estimate effects. Both are useful; RCTs are best when feasible, but quasi-experiments are valuable when they aren't.
10. How do restrictive border policies contribute to migration tragedies?
Restrictive border policies often force migrants to use more dangerous routes. When governments seal official crossing points with walls, guards, and technology, people avoid those and go through isolated areas. For example, building a fence may push migrants into a river or desert where many drown or die of thirst. Policies that criminalize help from citizens also make it risky for anyone to rescue distressed migrants. Increased patrols can make smugglers charge more and take greater risks. The result is that the same policies meant to stop migration actually cause more deaths. Humanitarian organizations argue that these policies violate the right to life.
11. What is a health outcome difference between immigrants and citizens?
Immigrants often have different health outcomes, both better and worse, compared to citizens. For example, new immigrants might be healthier than the local population, a fact called the 'healthy immigrant effect.' Over time, this advantage disappears because of stress, poor living conditions, and limited healthcare. Immigrants are more likely to die from preventable diseases because they delay seeking care. They also face higher rates of job injuries and mental health problems from migration trauma. Children of immigrants may lack proper vaccinations if their parents cannot access services. These differences show how policy choices affect real human lives.
12. How do the policy approaches of Middle Eastern oil-exporting countries differ from those of East Asian countries?
Oil-exporting countries like Saudi Arabia rely heavily on revenue from natural resources, so fiscal policy is tied to oil prices. They often use state-owned enterprises to control the oil sector and distribute wealth. East Asian countries lack such resources, so they focus on manufacturing and exports. Oil exporters may invest oil revenue in sovereign wealth funds to save for the future. Their economic diversification policies aim to reduce dependence on oil, sometimes by subsidizing other industries. The two models face different constraints: oil exporters manage resource abundance, while resource-poor East Asian countries compete through productivity.