Immigration is among the most politically contested policy areas and among the better studied, and the research findings are more consistent than the debate suggests.
What policy controls well
Legal migration categories, numbers and criteria are directly determined by policy.
Which means work, study and family migration respond to rule changes, and the responses are measurable.
Changes to student post-study work rights, for example, have produced rapid and large changes in application volumes.
What it controls less well
Irregular migration responds to enforcement and to conditions in origin and destination countries.
Which means enforcement affects routes and methods more than volumes in several studies, with flows displacing rather than stopping.
Asylum claims are governed by international obligations that constrain policy responses.
The labour market evidence
Extensively studied and reasonably consistent in aggregate.
Effects on average wages and employment of existing residents are generally found to be small.
Effects on specific groups can be larger, with the most consistent findings concerning previous immigrants competing in similar segments.
The debate among economists concerns methodology and the size of effects on particular groups rather than the general picture.
Fiscal effects
Depend heavily on the migrant population's age, skills and access to services.
Which means results differ by country and by category, and studies generally find working-age migrants contribute more in taxes than they receive in services during their working lives.
The longer-term calculation depends on assumptions about retirement, dependants and settlement, which is where estimates diverge.
Public services
Effects on service capacity depend on whether funding follows population.
Which is a policy design question rather than a consequence of migration, and lags in funding allocation produce local pressure.
Migrants are also disproportionately employed in health and care sectors in many countries, which affects service capacity in the other direction.
Integration
Language acquisition, employment and social contact are the factors most consistently associated with outcomes.
Which means policy on language provision, on the right to work while claims are processed, and on qualification recognition affects outcomes substantially.
Prohibiting work during asylum processing, which several countries do, is associated with worse long-term employment outcomes.
Public opinion
Survey research finds attitudes are affected more by perceived cultural change and by perceived control than by economic factors.
Which is a consistent finding, and it means economic arguments address a concern that is not the primary driver.
Perceptions of migrant numbers are systematically higher than actual numbers in most surveyed countries.
Where the evidence is weaker
Long-term social and cultural effects are harder to study and the evidence is correspondingly thinner.
Which is where much of the political argument concentrates, and where careful research is scarcest.
Acknowledging that is more honest than claiming the evidence settles it.
Processing capacity
Administrative capacity determines outcomes as much as policy does.
Which means backlogs function as de facto policy, producing long waits that affect people's lives regardless of what the rules say.
Asylum backlogs in several countries have grown to the point where processing times are measured in years, with substantial costs to individuals and to public finances through accommodation.
Detention
Practices vary enormously, from routine to exceptional, and time limits exist in some systems and not others.
Which is a significant human rights question, with evidence on the effects of indefinite detention being substantial.
Alternatives to detention have been evaluated and generally found to achieve comparable compliance at lower cost.
Data quality
Migration statistics are difficult to produce accurately, since departures are measured less well than arrivals in many systems.
Which means net migration figures carry substantial uncertainty and are revised frequently.
Statistical agencies publish the methodology and the uncertainty, and coverage generally reports point estimates as facts.
Enforcement
Interior enforcement, border enforcement and removals each have different costs and effects.
Which are studied with the general finding that enforcement affects methods and routes more than aggregate flows.
Deterrence effects have been found to be smaller than assumed in several evaluations, since decisions to migrate are driven substantially by conditions at origin.
Regularisation
Programmes granting status to people already present without it.
Which have been implemented in many countries, and evaluations generally find fiscal and labour market benefits.
The argument against concerns whether they encourage future irregular migration, and evidence on this is mixed.
Refugees and asylum distinguished
Refugee resettlement involves selection from abroad. Asylum involves claims made after arrival.
Which are governed by different processes, and conflating them produces confusion about what policy can control.
Public communication
Perceptions of migration numbers and composition differ substantially from measured reality in most surveyed countries.
Which means informational interventions have been tested, generally finding that correcting numerical misperceptions has limited effect on attitudes.
This is consistent with attitudes being driven by factors other than the numbers themselves.
Where to find evidence
Migration observatories and research centres publish evidence summaries aimed at general readers.
Which distinguish what is established from what is contested, and they are more useful than either advocacy from either direction.