MENA Bridge · Market Research
Jordan Construction & Development Pipeline Tracker
Building licences are a census of all licensing authorities in Jordan, not a sample, and they lead completions by roughly two to three years — the earliest read the market has on future supply.
Where the pipeline stands
Licences issued · rolling 12 months
27,087
To June 2026
▲ 6.6% vs a year earlier
Licensed area · rolling 12 months
10.35M
Square metres, to June 2026
▲ 3.3% vs a year earlier
New buildings · share of licensed area
62.2%
2026Q2
▲ 4.0 pp vs 2025Q2
New buildings · share of licence count
22.4%
2026Q2 · 1,307 licences
▼ 0.1 pp vs 2025Q2
Residential share of licensed area
82.7%
H1 2026, all licence types
▲ 6.1 pp vs H1 2025
Residential units licensed
6,904
2026Q2 · residential buildings, new and extended
▲ 10.5% vs 2025Q2
Units per residential licence
5.20
2026Q2
▲ 0.83 vs 2025Q2
Average unit licensed
198
Square metres, gross · 2026Q2
▲ 6 m² vs 2025Q2
Residential units per 100 apartments sold
81
2026Q2
▲ 8 vs 2025Q2
The headline, and what is under it
The two series the Department of Statistics leads with, monthly. The period buttons above drive both.
Period Labels
Licences and licensed area, monthly
Bars are licences issued; the line is the area they cover, on its own axis. June 2026 is drawn hollow on a dashed segment because it is derived — see the note below.
The same two series, rolling twelve months
Jordanian licensing is strongly seasonal; a twelve-month total removes that and shows the trend the monthly bars hide.
How the latest month is obtained. The Department of Statistics publishes a monthly release about
two months after the month it reports. Its June 2026 release prints no monthly table and its own chart
carries no values, so June 2026 here is that release's published half-year total less the five months
already published: 1,966 licences and 886 thousand m². The
half-year area is published to three significant figures, so that month's area carries a band of roughly
five thousand m². It is drawn hollow on a dashed segment wherever it appears.
What the licences are actually for
A licence is not a building. Most of them are issued for buildings that already exist, and how much that matters depends on whether you count licences or measure area.
Licences by what they permit, quarterly
Counted as licences: new buildings, extensions, and permits issued for buildings that already exist. The label above each bar is the total. The gap is 2024, which is absent from this source at the Department's end.
The same quarters, measured as area
Counted as square metres. The mix looks materially different from the chart beside it, and the reason is in the note below.
New buildings as a share of all licensing — counted two ways
Solid line: share of licensed area. Dashed line: share of the number of licences. Same underlying data, two bases.
What one licence covers, on average
The gap between these two lines is exactly why the two shares above disagree.
The pipeline counted in residential units
A licence is a permit, not a dwelling. The Department also counts the residential units each permit covers, and that is the number that matters for supply. Units are counted in residential buildings only — a non-residential licence contributes area, never units.
Residential units licensed, and how many each licence covers
Bars are residential units licensed in the quarter; the line is how many units one residential licence covers on average, on its own axis.
Reading this chart
Why a permit count understates the pipeline
A licence is a permission, not a dwelling. In 2022Q1 one residential licence covered 3.07 units on average. In 2026Q2 it covers 5.20.
That is the missing half of the story the two shares above tell. New buildings have been getting larger — but the extra floor area is not going into bigger units. The average unit licensed moved only from 172 to 198 square metres over the same span. The space is going into more units per building.
These are gross figures. The Department reports unit area at essentially the whole residential building — 99.5% to 100% of it in every year since 2004 — so common areas, stairs and walls are not deducted. Net sellable area is materially smaller, which is why an average unit of 198 m² here sits above the average apartment recorded in registered sales — the same housing, measured gross rather than net.
What that means: the falling licence count overstates how much the pipeline has actually shrunk. Residential units licensed in 2026Q2 were 6,904, against 6,250 in 2025Q2 — the permits fell, the units did not.
What to watch: if dwellings per licence keeps rising while the licence count falls, formal supply can hold up even as permits decline, and the headline number keeps getting further from what is being built.
Reading these four charts. The Department of Statistics separates three kinds of licence, and defines
the third itself: an existing building is “a building that has been constructed for years without a
permit and received a permit during the month in which the data was collected”. Those are
retroactive legalisations of buildings that are already standing — they are not future supply.
The two shares are on different bases, and the gap is the point. By licensed area, new buildings were 63.2% of everything licensed in 2022 and 56.9% in 2025. By number of licences they were 33.3% and 21.8%. Both fall, but the count basis falls roughly twice as far, because it gives a small house the same weight as a large new building. The average new building licensed in 2022Q1 covered 585 m²; the average existing building handed a retroactive permit covered 189 m². Read the area line as the supply signal and the count line as the sharper cut.
Fewer new buildings, but bigger ones. The average new-build licence has grown from 667 m² in 2022 to 1,005 m² in 2025, while the average retroactive permit has barely moved (189 to 210 m²). Today a new-build licence covers about 5.9 times what a retroactive one does. That growth is what keeps the area share from falling as fast as the count.
And the newest quarter has turned up. In 2026Q2 the area share rose to 62.2% from 47.3% in 2026Q1, and the count share to 22.4% from 16.6%. One quarter is not a trend, but it is the first move in this direction in the series.
What to watch: a sustained recovery in the solid line, not in the total licence count. A rising headline made of retroactive permits tells you about enforcement and formalisation, not about construction.
The two shares are on different bases, and the gap is the point. By licensed area, new buildings were 63.2% of everything licensed in 2022 and 56.9% in 2025. By number of licences they were 33.3% and 21.8%. Both fall, but the count basis falls roughly twice as far, because it gives a small house the same weight as a large new building. The average new building licensed in 2022Q1 covered 585 m²; the average existing building handed a retroactive permit covered 189 m². Read the area line as the supply signal and the count line as the sharper cut.
Fewer new buildings, but bigger ones. The average new-build licence has grown from 667 m² in 2022 to 1,005 m² in 2025, while the average retroactive permit has barely moved (189 to 210 m²). Today a new-build licence covers about 5.9 times what a retroactive one does. That growth is what keeps the area share from falling as fast as the count.
And the newest quarter has turned up. In 2026Q2 the area share rose to 62.2% from 47.3% in 2026Q1, and the count share to 22.4% from 16.6%. One quarter is not a trend, but it is the first move in this direction in the series.
What to watch: a sustained recovery in the solid line, not in the total licence count. A rising headline made of retroactive permits tells you about enforcement and formalisation, not about construction.
Residential, and everything else
Residential against non-residential, at the windows the Department publishes the split for.
Licensed area by purpose
Residential against non-residential, at the windows the Department of Statistics publishes the split for.
Residential share of licensed area
Drawn on a fixed 70–90% axis so a movement of a point or two is not rendered as a cliff.
In H1 2026 residential purposes took 82.7% of all licensed area, 4,027,923 m² against
842,077 m² for everything else. Against the same half of last year residential area
rose 10.2% while non-residential
fell 24.8%. The average licence now covers
451 m² against 312 m² in January 2022 — larger
by 44.6%.
Where it is being licensed
Housing licensed per resident, which is where the pipeline is intense rather than where it is large.
Residential building licensed per resident
Area licensed for residential buildings in January–June 2026, per 1,000 residents of each governorate. Population is the Department's own estimate, printed in the same table.
The same figures ranked, against a year earlier
Per 1,000 residents removes the effect of size: Amman licenses the most in absolute terms every month.
Per resident the pipeline is not where the population is. In January–June 2026 Amman licensed
490 m² of housing per 1,000 residents against 83 in Zarqa — a gap of
5.9 times. Amman ranks 1st
on this measure at 490 m², while still taking 60.9% of all licensed housing area in
the kingdom. Read this as relative intensity, not as an absolute market size.
One source defect in the newest release. In the release covering January–June 2026 the licensed-area column repeats the population figure instead of the area. Every governorate's area for that window is therefore reconstructed as the per-capita figure the release prints multiplied by the population it prints beside it; the twelve reconstructed values sum to within 0.1% of the residential total stated in the same release.
Supply against absorption
One office licenses the supply, another registers the sales. Putting them side by side is the closest read available on whether building keeps pace with buying.
Residential units licensed against apartments sold
Residential units licensed by the Department of Statistics; registered apartment sales from the Department of Lands and Survey. Both count dwellings, so both are drawn on one axis. Shown only for the 6 quarters both offices cover — see the note below.
Residential units licensed per 100 apartments sold
The two series above as one ratio.
Reading this chart. Each point is the number of residential units licensed in a quarter for every
100 apartments that changed hands in the same quarter. Both sides count dwellings, so this is a like-for-like
read on whether supply is being started at the pace demand is clearing it.
It stands at 81 in 2026Q2 — 6,904 residential units licensed against 8,488 apartments sold — against a high of 81 in 2026Q2 and a low of 53 in 2025Q3. Over the whole overlap it is 70: 36,411 units licensed against 52,382 sold. Sales running above new supply is what a market with an active resale stock should show.
Why the chart is short. It runs from 2025Q1 to 2026Q2, 6 quarters, because that is the whole overlap between the two offices: the licence database holds no 2024 at all, and the transaction series begins in 2024. Read the level with that in mind — 6 quarters is a direction, not a cycle.
What else would change this reading. The two sides are not the same universe. The licensed figure counts residential units of every kind across Jordan; the sales figure counts apartments only, and only those registered. And licensing is not building: a permit can lapse, and this page's own evidence is that a large share of Jordanian housing is built without one at all, so the licensed side is formal supply rather than all supply.
It stands at 81 in 2026Q2 — 6,904 residential units licensed against 8,488 apartments sold — against a high of 81 in 2026Q2 and a low of 53 in 2025Q3. Over the whole overlap it is 70: 36,411 units licensed against 52,382 sold. Sales running above new supply is what a market with an active resale stock should show.
Why the chart is short. It runs from 2025Q1 to 2026Q2, 6 quarters, because that is the whole overlap between the two offices: the licence database holds no 2024 at all, and the transaction series begins in 2024. Read the level with that in mind — 6 quarters is a direction, not a cycle.
What else would change this reading. The two sides are not the same universe. The licensed figure counts residential units of every kind across Jordan; the sales figure counts apartments only, and only those registered. And licensing is not building: a permit can lapse, and this page's own evidence is that a large share of Jordanian housing is built without one at all, so the licensed side is formal supply rather than all supply.
What is financing it
A licence is a decision somebody has to fund. This section sits beside the supply charts because the credit series explains more of their shape than the interest rate does on its own.
The cost of debt fell. Construction’s share of bank credit did not follow.
Weighted average bank lending rate against credit to construction as a share of all bank credit, monthly from January 2020. Two axes — the series are separate readings, not a spread. Source: Central Bank of Jordan.
How much of all bank lending goes to construction
Credit facilities extended to construction as a share of all credit facilities extended by licensed banks, monthly from January 2020. One series, on its own axis. Source: Central Bank of Jordan.
The same thing in money: construction inside total bank credit
Credit facilities extended by licensed banks, JOD billion, monthly from January 2020. The
lower band is construction; the two bands together are all bank credit. Source: Central Bank of Jordan.
Reading these three charts. They answer different questions. The first asks whether debt has become
cheaper and whether more of it has gone to construction; the second asks only how large a slice of all bank
lending construction is, on an axis where the whole move is legible; the third asks the same question in
dinars rather than percentages, which is where the scale of it shows.
Since the lending rate peaked in August 2023 it has fallen 173 basis points, from 9.21% to 7.48%, and total bank credit has grown +11.8%. Over the same months credit to construction moved −0.4%. Its share of all credit facilities fell from 24.31% to 21.65%, and from its peak of 26.03% in June 2022 it is down 4.38 percentage points, the lowest being 21.65% in June 2026. Measured from January 2020, total credit is +36.4% against construction’s +15.5%.
The same thing in money. Total credit facilities have grown from JOD 27.2bn to JOD 37.1bn since January 2020 — JOD 9.9bn of new lending. Construction went from JOD 7.0bn to JOD 8.0bn, an increase of JOD 1.1bn. So 10.9% of all the new credit the banking system extended went to construction — about 11 dinars in every hundred, against the 21.65% of the existing book construction already held. The lower band is not shrinking; it is standing still while everything above it grows.
Cheaper money has not reached the people who build. On this evidence the constraint on development is the availability of credit rather than its price — which is the reading to carry into the licence and unit charts above.
What would change this. Credit to construction growing faster than total bank credit for two consecutive quarters. What these are not. A lending rate is a rate on a loan and the share is a proportion of a credit stock; the two are separate readings and the gap between them is not a number with a meaning. Neither is a return — an entry yield is asset-specific and is the reader’s to supply. And a share is not a level: construction credit can grow in absolute terms while its share falls, which is close to what has happened.
Since the lending rate peaked in August 2023 it has fallen 173 basis points, from 9.21% to 7.48%, and total bank credit has grown +11.8%. Over the same months credit to construction moved −0.4%. Its share of all credit facilities fell from 24.31% to 21.65%, and from its peak of 26.03% in June 2022 it is down 4.38 percentage points, the lowest being 21.65% in June 2026. Measured from January 2020, total credit is +36.4% against construction’s +15.5%.
The same thing in money. Total credit facilities have grown from JOD 27.2bn to JOD 37.1bn since January 2020 — JOD 9.9bn of new lending. Construction went from JOD 7.0bn to JOD 8.0bn, an increase of JOD 1.1bn. So 10.9% of all the new credit the banking system extended went to construction — about 11 dinars in every hundred, against the 21.65% of the existing book construction already held. The lower band is not shrinking; it is standing still while everything above it grows.
Cheaper money has not reached the people who build. On this evidence the constraint on development is the availability of credit rather than its price — which is the reading to carry into the licence and unit charts above.
What would change this. Credit to construction growing faster than total bank credit for two consecutive quarters. What these are not. A lending rate is a rate on a loan and the share is a proportion of a credit stock; the two are separate readings and the gap between them is not a number with a meaning. Neither is a return — an entry yield is asset-specific and is the reader’s to supply. And a share is not a level: construction credit can grow in absolute terms while its share falls, which is close to what has happened.