Enough Room to Return?
Where 7.2 million Afghan returnees settled, and whether public health services, rentals, and wages can accommodate them
Ahmad Zia Wahdat, PhD | Data for Afghanistan | November 2026
Between 2021 and October 2026, about 7.2 million Afghans returned from Iran, Pakistan and elsewhere. This report places every one of them in one of 401 districts, and measures that number against the people, land, health workers, wages and prices already there.
This is an independent analysis by Data for Afghanistan. It combines open and publicly accessible data from IOM, UNHCR, the Ministry of Public Health, WFP and the National Statistics and Information Authority, and the Data for Afghanistan population grid built for the ARMO 2026 report. All figures are descriptive. They show where pressure is building, not why.
Explore the district map
Key findings
- 1
Return is concentrated, and 2026 alone added a fifth of it. The 1.39 million arrivals recorded in the first ten months of 2026 make up 19% of the six-year total. Ten districts hold 21% of all returnees, led by Kabul city, Herat city, Mazar-e-Sharif, Jalalabad and Injil.
- 2
The heaviest load relative to population is in the north and west, not Kabul. Returnees are close to 1 in 4 residents in Kunduz, Daykundi and Herat. In Kabul, which receives the largest absolute number, returnees are 1 in 14 residents. In 58 districts with at least 20,000 residents each, returnees are a fifth or more of the population.
- 3
Concentration of returnees and spatial crowding are different pressures, and 45% of returnees live where both are high. Comparing returnees per resident with returnees per square kilometer of inhabited land sorts districts into four types. The 90 "saturated" districts hold 3.3 million returnees.
- 4
Health capacity has not kept up with returnees. Nationally there are 252 returnees for every public health worker. In Kunduz, Takhar, Balkh and Herat the figure is 469 to 563. New public health facilities did not follow returnees: Kunduz and Daykundi, two of the three provinces with the highest returnee share, opened 8 and 10 in the last five years, against 80 in Helmand.
- 5
One unskilled wage cannot keep a returnee family above the poverty line anywhere. Even working 26 days a month, an unskilled daily wage covers between 37% and 76% of the August 2026 poverty line for a household of 6.1 members. Rent for a modest four-bedroom house alone takes 10 to 33 days of unskilled wage a month.
- 6
Prices do not yet track returnee concentration. Provincial inflation in August 2026 ranged from 0.1% in Parwan to 22.3% in Sar-e-Pul, with no relationship to the returnee share. Nationally, health (19.0%) and housing (12.1%) costs rose faster than prices overall (7.8%).
- 7
A further wave of one million returnees would land mostly on the same places. If it follows the 2026 pattern, Nangarhar, Herat and Kabul would receive 46% of it, and Kunduz's returnee share would pass 25%. Holding today's health ratios would need about 537 more health workers and 65 more public health facilities; 380,000 more children of school age would need education places.
Two kinds of pressure, four types of district
A district's services can get stretched because returnees make up a large share of its people, or the district can get crowded because many returnees press onto little inhabited land. The two call for different responses. Each district is compared with the national average on both measures: a returnee share of 13.6% and 60 returnees per km² of inhabited land.
Saturated 90 districts
High returnee share and high spatial crowding, both above the national average. This implies pressure on jobs, rents and social services at once. There are 90 saturated districts with 3.3 million returnees.
Stretched 80 districts
High returnee share, but low spatial crowding. Returnees are probably spread through rural settlements, but livelihoods and services are under strain. There are 80 stretched districts with 1.2 million returnees.
Crowded 62 districts
Low returnee share, but high spatial crowding. Crowded districts are mostly large towns where returnees add to an already dense population. There are 62 crowded districts with 1.5 million returnees.
Buffered 143 districts
Below the national level on both returnee share and spatial crowding. Buffered districts are not free of needs, but they have relatively more room. There are 143 buffered districts with 1.1 million returnees.
Counts are for the 377 districts with at least 20,000 residents; smaller districts are typed and mapped in the same way. The types are relative to the national average, which is itself high: one resident in seven is a returnee. "Buffered" does not mean a district has no returnee needs.
What the evidence suggests
- Plan by district type, not only by province totals. Saturated districts need help on several fronts at once: jobs, rent, health and schools. Stretched rural districts need livelihoods and outreach services more than buildings. Crowded towns need housing and land solutions. A single provincial figure hides these differences.
- Look north and west as well as at Kabul. Kunduz, Takhar, Balkh and Herat combine high returnee loads with the heaviest pressure on health workers. New public health facilities over the last five years went mostly elsewhere.
- Treat earning capacity as the binding constraint. No province offers an unskilled wage that keeps a returnee household above the poverty line on one income. Cost is also the main reason returnee children are out of school.
- Watch rents and prices, especially health and housing. Provincial inflation does not yet follow returnee concentration, but national health and housing costs are rising fastest, and rent alone takes up to 33 days of unskilled wage a month in Nangarhar and 31 in Kabul.
- Close the data gaps. Two openly available datasets would sharpen this analysis considerably: a district list of schools with enrollment, and facility coordinates in the health registry so that travel time can be measured.
- Prepare for the next wave in the same places. The stress test shows that a further million arrivals following the 2026 pattern would deepen the same concentrations. Preparation in Nangarhar, Herat, Kabul and Kunduz would cover about half of the likely load.
Province profiles
One page for each of the 34 provinces: returnee numbers, district types, health capacity, earning capacity, rent and prices.
- 01 Badakhshan بدخشان Buffered 218,325 returnees
- 02 Badghis بادغیس Stretched 176,671 returnees
- 03 Baghlan بغلان Saturated 393,221 returnees
- 04 Balkh بلخ Saturated 480,696 returnees
- 05 Bamyan بامیان Buffered 80,251 returnees
- 06 Daykundi دایکندی Stretched 153,025 returnees
- 07 Farah فراه Stretched 169,601 returnees
- 08 Faryab فاریاب Crowded 277,858 returnees
- 09 Ghazni غزنی Buffered 188,640 returnees
- 10 Ghor غور Stretched 228,023 returnees
- 11 Helmand هلمند Buffered 301,554 returnees
- 12 Herat هرات Saturated 851,664 returnees
- 13 Jawzjan جوزجان Buffered 87,031 returnees
- 14 Kabul کابل Crowded 486,124 returnees
- 15 Kandahar کندهار Buffered 228,638 returnees
- 16 Kapisa کاپیسا Saturated 128,412 returnees
- 17 Khost خوست Buffered 101,505 returnees
- 18 Kunar کنرها Crowded 143,621 returnees
- 19 Kunduz کندز Saturated 389,074 returnees
- 20 Laghman لغمان Crowded 116,929 returnees
- 21 Logar لوگر Saturated 111,709 returnees
- 22 Nangarhar ننگرهار Saturated 596,781 returnees
- 23 Nimroz نیمروز Stretched 57,519 returnees
- 24 Nuristan نورستان Buffered 26,012 returnees
- 25 Paktika پکتیکا Buffered 84,340 returnees
- 26 Paktya پکتیا Buffered 86,766 returnees
- 27 Panjsher پنجشیر Stretched 21,320 returnees
- 28 Parwan پروان Crowded 135,700 returnees
- 29 Samangan سمنگان Buffered 97,103 returnees
- 30 Sar-e-Pul سرپل Saturated 179,211 returnees
- 31 Takhar تخار Saturated 378,632 returnees
- 32 Uruzgan ارزگان Buffered 51,798 returnees
- 33 Wardak میدان وردک Buffered 76,164 returnees
- 34 Zabul زابل Buffered 84,017 returnees