Portfolio — How far brokerage reaches

91.5%

Of 513,213 brokerage legs on apartment sales, 91.5% were handled by an office in the same district as the property. The median district is at 92.5%. The 8.5% that does cross a boundary has a median distance of 9.3 km and 83.2% of it stays within 20 km — not flows, the district next door.

One dot per district

Not a histogram. The shares run from 18.9% to 100% and the 5th percentile of the 186 districts is already 83% — nineteen in twenty sit above it. Any bin scheme wide enough to make that mass legible collects everything below into one bucket — and this page’s only real exception lives in that bucket. A dot per district has nowhere to hide it. The axis runs the full 0 to 100 rather than cropping to where the data is: cropping would spread the hump out and make it look like a distribution with a spread. The empty left half is the finding.

Same-district share, 186 districts with 200 legs or more

18.9%70.9%71.9%75.7%78.7%80.3%81.1%81.4%82.2%83%83.1%83.1%83.2%83.3%83.6%83.6%84%84.1%84.6%85%85%85.1%85.7%85.9%86.6%86.7%86.9%86.9%86.9%87%87%87.1%87.3%87.3%87.3%87.4%87.4%87.8%87.9%88%88%88.1%88.1%88.2%88.2%88.2%88.4%88.4%88.4%88.6%88.6%88.7%89%89.3%89.3%89.5%89.5%89.7%89.8%89.8%89.8%89.9%89.9%90%90%90%90.1%90.1%90.1%90.2%90.2%90.2%90.4%90.4%90.5%90.6%90.6%90.6%90.8%90.9%91.3%91.3%91.3%91.4%91.5%91.5%91.6%91.6%92.1%92.2%92.2%92.2%92.4%92.5%92.6%92.6%92.6%92.6%92.7%93%93%93.2%93.2%93.3%93.3%93.4%93.5%93.6%93.7%93.7%93.7%93.8%93.8%93.8%93.9%94.1%94.1%94.1%94.2%94.5%94.5%94.6%94.6%94.6%94.7%94.8%94.8%94.8%94.8%95.2%95.2%95.3%95.4%95.4%95.4%95.5%95.8%95.9%96%96.1%96.2%96.3%96.5%96.8%96.8%96.8%97%97.2%97.3%97.3%97.4%97.5%97.6%97.6%97.7%97.7%97.9%98.2%98.2%98.3%98.4%98.4%98.4%98.5%98.5%98.5%98.5%98.6%98.6%98.6%98.6%98.8%98.8%98.8%99%99%99%99%99.1%99.2%99.2%99.3%99.4%99.8%99.9%100%median 92.5%020406080100
MOLIT apartment sale filings, brokerage office district, {from} to {to}

Crossing a line is not going far

The distance of the 43,262 legs that crossed a district boundary. Median 9.3 km, 90th percentile 28 km, longest 522 km. Distance is the great-circle span between district centroids, and a centroid is the mean of that district’s sub-district centroids.

Distance of the 43,262 external legs, cumulative
DistanceCumulative
within 10 km54.5%
within 20 km83.2%
within 50 km94.9%
within 100 km97.7%
within 200 km98.8%

The exceptions are adjacency, with one exception

The twelve lowest of the 186 measured districts. Read the inflow column and it is almost always a district that touches: another ward of the same city, or the city across the boundary. One does not fit. 강진군 sits at 18.9% and 80.2% of the brokerage that came in came from 여수시, 85 km away. It is the only row distance does not explain, and this table exists to make that one row visible.

The twelve lowest same-district shares, with the largest source of inbound brokerage
DistrictSame-districtLegsLargest inflowkm
강진군 · 전남광주통합특별시18.9%243여수시 · 80.2%85
완주군 · 전북특별자치도70.9%1,355전주시덕진구 · 22.4%11
영암군 · 전남광주통합특별시71.9%292목포시 · 23.3%23
중구 · 서울특별시75.7%1,220성동구 · 10%4
중구 · 대구광역시78.7%2,393수성구 · 5.3%7
동구 · 전남광주통합특별시80.3%1,307북구 · 7.5%7
의왕시 · 경기도81.1%2,622안양시동안구 · 12.4%5
성남시수정구 · 경기도81.4%1,831성남시중원구 · 5.1%3
목포시 · 전남광주통합특별시82.2%2,637무안군 · 17%19
남구 · 대구광역시83%893수성구 · 6%6
용산구 · 서울특별시83.1%1,252강남구 · 2.9%9
동작구 · 서울특별시83.1%3,079서초구 · 3.8%8

What this page did not do

No flow map
Every sale filing names the district its brokerage sits in, so the origin-destination matrix falls out of the data and the figure that suggests itself is arcs across the country. It was measured and then not drawn: there are 5,032 distinct external pairs and the ten thickest carry 7.9% of the external volume, with the rest spread thin. Five thousand lines drawn to show 8.5% of the whole would leave a reader believing they had watched brokerage move around the country. A beautiful figure answering a question the data cannot answer is worse than no figure.
Not a co-brokerage rate
30,329 filings name two districts in the office field, 6.3% of the total. Of those, 30,317 name two DIFFERENT districts and 12 name the same one twice. A field recording offices would show same-district pairs in bulk — two offices on one street co-brokering a sale is ordinary. So what it records is the set of districts involved, and 6.3% is not the co-brokerage rate but the rate at which co-brokerage crosses a district line; the same-district kind cannot be counted from this source. These are counts rather than a share for a reason: 12/30,329 rounds to 100.0% at one decimal place, and that 100 was once read as "not one of them". Counts cannot round.
Legs, not deals
The unit is 513,213 legs, not 481,694 filings. Where the office field is comma separated, each entry is one leg. Counting deals would credit a co-brokered sale to whichever office was listed first, which understates how often brokerage crosses a line.
{belowFloor} not plotted
57 districts with fewer than 200 legs, and 2 with no brokered sale at all, get no dot. A share resting on forty legs and one resting on three thousand draw the same dot, so the table carries the leg count beside the rate.
Centroids are approximate
A district centroid here is the unweighted mean of its sub-district centroids. Without area weighting, a city with dense wards downtown and large rural townships around them pulls toward downtown. That is where the offices are, so it suits this page — but the km column is the distance between districts, not between offices.

District cartogram →

← Portfolio