Most people guess too high

Surveys in rich countries find people overestimate the share of migrants around them, usually by more than a factor of two.

Where migrants live

Figure 1

Mexico looks untouched

So what?

• Migration is small globally, concentrated locally
• One map only shows the countries people arrive in
• A country can be shaped by migration without receiving anyone

The wage gap is enormous

The same worker, doing the same job, earns several times more in a rich country than in a poor one.

What stops them

%%{init: {"flowchart": {"useMaxWidth": true, "htmlLabels": true, "padding": 26,
                        "nodeSpacing": 20, "rankSpacing": 80},
          "themeVariables": {"fontSize": "19px", "fontFamily": "Arial, Helvetica, sans-serif"}} }%%
flowchart LR
  A["Wage gap<br/>abroad"] --> D{"Decide"}
  B["Cost of<br/>travel"] --> D
  C["Visas and<br/>borders"] --> D
  E["Someone<br/>already there"] --> D
  D --> F["Migration"]

It is not the poorest who leave

• Moving costs money: transport, documents, months without earnings
• The very poorest cannot pay those costs
• So emigration is low in the poorest countries

Who gains and who loses

Figure 3

What that graph cannot show

• Net migration is arrivals minus departures
• A country can send many people abroad and still gain people overall
• So we cannot read emigration off it

Why migration clusters

• The first migrant faces the highest cost and the greatest risk
• The second follows a known route, to a known contact
• Each arrival lowers the cost for the next

Exercise 1 — Who leaves?

Two villages, same country. Going abroad costs 50,000 pesos, paid before you leave.

• Village A: a family earns 20,000 pesos a year
• Village B: a family earns 60,000 pesos a year

Quick first: which village sends more people abroad?

So what?

• Wage gaps create the incentive; costs decide who can act
• Development first raises emigration, then lowers it
• Networks make migration self-reinforcing once it starts

Remittances: the second map

Figure 4

Almost the opposite map

• Map 1 was dark in the Gulf, North America, Australia
• Map 2 is dark in Central America, the Caucasus, West Africa
• Same phenomenon, seen from each end

Mexico, 1990 to 2023

Figure 5

What that money does

• It goes to households, not governments
• It rises when the sending household is in trouble
• It is more stable than foreign investment during crises

The other side of the money

Remember who leaves: not the poorest. Leaving costs money.

• So the money comes back to households that could already afford to send someone
• Inside a village, remittances can widen the gap they are meant to close

So what?

• Migration binds two economies together, not one
• The receiving country gains workers; the sending country gains income
• Neither map alone describes what happened

The question everyone asks

Do migrants lower wages for workers already there?

A natural experiment

• April 1980: Cuba briefly allowed emigration from the port of Mariel
• About 125,000 people left for Florida in six months
• Miami’s labour force grew roughly 7% almost overnight
• Miami’s labour supply jumped for reasons unrelated to Miami’s economy

Mariel, 1980

Cuban refugees arriving in Florida during the Mariel boatlift, 1980. Source: Wikimedia Commons, public domain.

Two economists, one event

• Card (1990): Miami against four similar cities → no measurable effect on wages
• Borjas (2017): only non-Hispanic men, 25–59, without a high-school diploma → a large fall
• Peri & Yasenov (2019); Clemens & Hunt (2019): same data, re-examined → the fall does not survive

Why they disagree

%%{init: {"flowchart": {"useMaxWidth": true, "htmlLabels": true, "padding": 26,
                        "nodeSpacing": 34, "rankSpacing": 80},
          "themeVariables": {"fontSize": "19px", "fontFamily": "Arial, Helvetica, sans-serif"}} }%%
flowchart LR
  A["Same event:<br/>Mariel 1980"] --> B["Which<br/>workers do<br/>you count?"]
  A --> C["Which<br/>cities do you<br/>compare?"]
  B --> D["Different<br/>answer"]
  C --> D
  D --> E["The answer<br/>depends on<br/>those choices"]

What we can say

• Average effects on wages are small in most studies
• Effects concentrate on workers most similar to arrivals
• The answer depends heavily on how you define the comparison group

Exercise 2 — Design the test

125,000 people arrive in Miami in four months, and wages change. Did they cause it? You need a second city — one that shows what Miami would have done anyway.

1. Which city would you use? What goes wrong with it?

2. Whose wages would you look at? all workers · no high-school diploma · construction workers · Cubans already in Miami

Exercise 2 — Responses

The city. Card used four averaged together — Atlanta, Houston, Los Angeles, Tampa — so no single city’s quirks decide the answer.

Takeaways

1. Migration is 3.6% of humanity, very unevenly placed
2. Development raises emigration before it lowers it
3. Networks make migration self-reinforcing
4. Remittances bind sending and receiving economies together
5. Wage effects are small on average, contested at the margins