Lesson 2
Testing the model in Ho Chi Minh City
Lesson 1 predicts that land prices and population density fall with distance from the centre. This lesson measures both gradients for Ho Chi Minh City, locates the edge of the city with population data rather than with its administrative boundary, compares the density gradient with 192 cities, and identifies where the city departs from the model.
Learning objectives
After this lesson you should be able to:
- estimate a price or density gradient with a semi-logarithmic regression, and interpret its coefficient;
- explain why an administratively set land price is weak evidence about a market bid-rent curve;
- assess how choices of measurement change an estimated gradient;
- distinguish the administrative boundary of a city from its urban extent measured with gridded population data, and relate both to the edge of the city in the model;
- compare Ho Chi Minh City with other cities, and identify the differences in measurement that limit the comparison.
2.1Measuring a gradient
The model predicts the direction of the gradients but not their functional form. Empirical studies usually summarise a gradient with a semi-logarithmic regression, a form that goes back to Clark (1951):
Here \(y_i\) is the price or density of observation \(i\), \(x_i\) its distance in kilometres from the centre, and \(b\) the gradient. One kilometre farther out, \(y\) is multiplied by \(e^{b}\), a change of \(100\,(e^{b}-1)\) percent; for small values of \(b\) this is close to \(100\,b\) percent.
Three choices must be made before equation (2.1) can be estimated: the location of the centre, the variable that is measured, and the observations that enter the regression. This lesson places the centre at Ben Thanh station on Metro Line 1, in the historic core of the city, and discusses the other two choices in each section.
2.2Land prices
Ho Chi Minh City sets a land price for each street segment in a land price table. The table applied from 1 January 2026 contains street segments. Map 2.1 shows, for each of the wards and communes that can be mapped, the median price of residential land across its segments.
Median price across the street segments of each ward, thousand VND per m²
Source: Ho Chi Minh City People’s Council, Resolution 87/2025/NQ-HĐND (land price table from 1 January 2026); ward boundaries © OpenStreetMap contributors (ODbL); base map OpenFreeMap.
Note: dashed circles are 10, 20, 40 and 60 km from Ben Thanh; the blue outline is the urban centre of 2020 (Map 2.3). commune without a boundary in OpenStreetMap is not shown; Con Dao special zone lies outside the map. The table sets prices for taxation, fees and compensation; it is not a record of market transactions.
Median residential land price, thousand VND per m², logarithmic scale
Source: as Map 2.1; regression by kich_ban/l2_tphcm.mjs.
Note: each dot is a ward or commune, placed at the distance from Ben Thanh to the centroid of its boundary. The fitted line is equation (2.1) estimated on wards within 40 km.
Within 40 km of Ben Thanh the price falls by per kilometre (\(b =\) , \(R^2 =\) , wards). Estimated on all mapped wards except the island of Con Dao, the gradient is and the fit is weaker (\(R^2 =\) ), because prices stop falling beyond about 40 km. Table 2.1 shows why: several towns far from Ben Thanh have prices well above those of wards at a similar distance.
| Ward | Distance (km) | Price | Median of wards within ±5 km | Ratio |
|---|---|---|---|---|
Source: as Map 2.1; kich_ban/l2_tphcm.mjs. Prices in thousand VND per m².
Note: before the 2025 merger, Thu Dau Mot was the capital of Binh Duong province and Ba Ria the capital of Ba Ria–Vung Tau province. Di An is included for comparison: it lies inside the continuous urban area and its price is close to that of its neighbours. The script stops if these statements no longer match the data.
Map 2.2 returns the prices of Figure 2.1 to the map. Each ward’s price stands as a column over the ward, on the logarithmic scale of the figure. Turning the fitted line of Figure 2.1 around Ben Thanh gives a surface on which price falls at the same rate in every direction, as the monocentric model assumes. The four steps set the table against that surface.
Residential land price by ward and the fitted surface of equation (2.1), thousand VND per m²
Prices by ward. Each column stands on the centroid of a ward, and its height is the ward’s median residential price on a logarithmic scale: every tenfold rise in price adds the same height. The tallest column, , reaches thousand VND per m². Columns are tallest in the centre, fall towards the edge of the map and rise again around the towns of Table 2.1.
The fitted surface. Turned around Ben Thanh, the fitted line of Figure 2.1 becomes a surface on which price falls by per kilometre in every direction. On a logarithmic scale the surface is a cone. It is drawn only to 40 km, the distance used in the fit; within it, wards stand above the surface (red) and below it (blue). The rings on the surface mark the prices at which the colours on the ground change.
Above and below the surface. The ground is now shaded by the ratio of each ward’s price to the price on the surface at the same distance, for wards within 40 km. Of the five wards furthest above the surface, four lie 20–40 km to the north and north-west, led by at times the price on the surface; the fifth is in the centre ( times). The five furthest below lie 10–25 km to the north, west and south-west, down to ( times). Apart from the centre itself, the largest departures all lie 10–40 km from Ben Thanh. Section 2.6 asks whether a city with more than one centre accounts for them better.
Farmland and the edge. In the model the city ends where land rent falls to agricultural rent, equation (1.5). The same table prices farmland in four zones (Table 2.2 in Section 2.3); the green discs show position 1 for annual crops, from thousand VND per m² in zone I to in zone IV. Extended beyond the 40 km on which it was fitted, the surface falls to these prices between and km from Ben Thanh (green band). The extension is not an estimate, and the residential prices of the table do not follow it: as Section 2.3 shows, they never fall to the price of farmland.
Source: as Map 2.1 and Figure 2.1; farmland prices as Table 2.2; urban centre as Map 2.3; base map OpenFreeMap, © OpenStreetMap contributors.
Note: the surface is equation (2.1) fitted to the wards within 40 km and turned through a full circle, although part of that circle is river or sea. Both prices are set by the city, not observed in the market. Switch the height to the ordinary scale to see the surface as the curved tent of Figure 1.2: on it, the price 10 km from Ben Thanh is of the price at the centre.
How far do these prices tell us about the bid-rent curve of Lesson 1? The land price table is set by the city government. The World Bank’s Vietnam Urbanization Review describes the system as follows:
The table therefore shows how the city government values locations, not what buyers and sellers pay. That the table falls with distance is consistent with the model, but it is not a test of it: a test requires market prices by location, which this course has not been able to obtain for the whole city. There is a second reason for care. By Section 1.4, the price of land should fall faster than the price of floor space, because buildings are also shorter farther out. A land price gradient cannot be compared directly with a gradient in dwelling rents, such as the one reported for 191 cities by Liotta, Viguié and Lepetit (2022):
2.3Where does Ho Chi Minh City end?
Equation (1.5) closes the model of Lesson 1: the city ends at the distance \(\bar{x}\) where land rent falls to the rent that land earns in agriculture. The edge is an outcome of the land market. Nothing in the model makes it coincide with the boundary that the law draws around a city, and a gradient measured inside that boundary may include land beyond \(\bar{x}\) and leave out land within it.
How, then, should a city be drawn from data? Researchers take two broad approaches, which de Bellefon, Combes, Duranton, Gobillon and Gorin describe as follows:
The most widely used morphological definition is the urban centre of the Degree of Urbanisation, endorsed by the UN Statistical Commission in 2020 for international comparisons and applied to every city in the world by the Global Human Settlement Layer (GHSL) of the European Commission:
The definition counts residents, not buildings, and its authors chose population on purpose:
Population density alone, however, does not locate the edge of the model’s city. In Lesson 1 the city ends where land stops being used for housing, and every household in it commutes to the centre. Map 2.3 therefore draws Ho Chi Minh City in three ways at the same scale: by population density, by built-up land seen from satellites, and by commuting. Scroll through the eight steps, or use the buttons or the ← and → keys.
Administrative boundary; urban area by population density, built-up land and commuting; growth 1985–2015; land prices
1The legal boundary
The black line is the administrative boundary of Ho Chi Minh City after the merger of 1 July 2025, which joined the former provinces of Binh Duong and Ba Ria–Vung Tau to the city. It encloses km².
The line marks where the authority of the city government ends. It was not drawn by the land market, so it is not \(\bar{x}\).
2Where people live
Each cell of about 100 m is coloured by its estimated number of residents per km² in 2020. The grid ignores the boundary. Dense settlement continues across the line to the north-east, into Bien Hoa in Dong Nai province, while much of the land inside the line, such as the mangrove forest of Can Gio in the south, has few residents.
Two peaks stand out away from Ben Thanh. Near Thu Dau Mot the densest 500 m cell has residents per km², against a median of for cells at the same distance from Ben Thanh; near Bien Hoa, against .
3Edge 1: population density
The blue outline is the GHSL urban centre of 2020: the 1 km cells with at least residents per km², joined into one cluster of at least people. It covered km² and held million people.
km² of it, , lies outside the administrative boundary; only of the administrative area lies inside it.
The dashed circle has the same area, a radius of km: the city as Lesson 1 would draw it. Along the section of Figure 2.2 the urban centre extends km north of Ben Thanh but only km south.
4Edge 2: built-up land
Each dark cell, about 90 m across, was at least half built up in 2015 according to the World Settlement Footprint, which reads Landsat satellite images. It records buildings and paved land, not residents.
Only of the land inside the urban centre was built up. The rest has no buildings, such as rivers and the runways of Tan Son Nhat airport: the urban centre includes it because the outline is drawn on 1 km cells and its gaps are filled. Outside the urban centre, another km² was built up within the functional urban area of step 5.
5Edge 3: commuting
The purple outline is the functional urban area of 2015: the city together with its commuting zone, the surrounding places from which at least of employed residents commute to it. In concept it is the closest of the three to the city of the model, in which every household commutes to the centre.
It covered km² and million people, million of them in the commuting zone; km² of it lies outside today’s administrative boundary.
No commuting survey of Ho Chi Minh City was used: the GHSL predicts the commuting zone from travel time and population, learning from countries that do have commuting data.
6Growth seen from satellites, 1985–2015
The colours show when each cell became built up. Within the functional urban area, built-up land grew from km² in 1985 to km² in 2015, as much.
The red outline is the GHSL urban centre of 1990, a year observed from satellites. It had two parts: Ho Chi Minh City, and Bien Hoa ( km²), then a separate town that joined the main part in . Only of the land inside the 1990 outline was built up that year.
7Density as height
The population grid of step 2, summed into 500 m cells, with the height of each column proportional to its density (scale in the legend). The tallest column has residents per km².
The surface is not the single cone of Lesson 1. It falls away from the centre and rises again at Thu Dau Mot and Bien Hoa. Drag with the right mouse button, or with two fingers, to turn the view.
8Land prices and the urban centre
The land price table of Section 2.2 covers only the administrative city, so Bien Hoa has no price. The median ward whose centre lies inside the 2020 urban centre is priced at thousand VND per m² ( wards); the median ward or commune outside it at (); the ratio is .
A price table and a population grid, made independently, separate urban from rural land in a similar place, and neither follows the administrative boundary.
Source: urban centre: GHSL Urban Centre Database R2024A; population: GHS-POP R2023A, 2020; functional urban area: GHS-FUA R2019A (all European Commission, Joint Research Centre; the last with the OECD; CC BY 4.0). Built-up land: World Settlement Footprint Evolution v1, © 2024 DLR (CC BY 4.0). Administrative boundary: the ward and commune boundaries of Map 2.1 joined together, © OpenStreetMap contributors (ODbL). Land prices: as Map 2.1. Base map: OpenFreeMap.
Note: Con Dao special zone lies outside the map. A built-up cell groups three by three Landsat pixels and counts when at least five of the nine are built up; land built up before 1985 is shown as 1985. GHSL outlines exist for every fifth year, but only 1975, 1990 and 2000 among them are observed years; the others are not drawn. Wards are placed inside or outside the urban centre by a point inside each ward boundary. Prepared by kich_ban/l2_vung_do_thi.py, kich_ban/l2_luoi_dan_so.mjs and kich_ban/l2_ve_tinh.mjs; each stops without writing its output if a statement on this map no longer matches the data.
Which line is the edge?
None of the three lines is \(\bar{x}\) exactly, and each answers a different question. The urban centre is a statistical convention designed to compare cities everywhere. Built-up land comes closest to the land use that defines the edge in the model. The functional urban area comes closest to its commuting households, but is itself modelled. A population grid has a further limit near the centre:
Series over time need the same care. The GHSL publishes an outline for every fifth year, but most of those years were not observed:
A series built this way will always show a city that grew smoothly outward. The satellite series of step 6 has a similar limit, since it too works back from the map of 2015 and cannot show land that was built on and later cleared, but the built-up land of each year is read from images of that year. Where the rest of this lesson needs a single outline, it uses the urban centre of 2020, which reaches km from Ben Thanh at its farthest point, slightly beyond the 40 km circle used for density in Section 2.4.
The edge in land prices
The land price table also prices agricultural land, so the condition \(r(\bar{x}) = r_A\) can be examined from prices as well. The regulation attached to Resolution 87/2025 assigns each of the city’s wards, communes and special zone to one of four agricultural zones (Article 3), and prices agricultural land by zone and by position (Article 5). A plot is in position 1 if it lies within 200 m of a road named in the table, which makes position 1 comparable with residential prices, which are also set for plots on named roads. Table 2.2 gives the prices for annual crops, a category that includes rice.
Thousand VND per m²
| Zone | Wards, communes and special zone | Position 1 | Position 2 | Position 3 |
|---|---|---|---|---|
| I | ||||
| II | ||||
| III | ||||
| IV |
Source: Ho Chi Minh City People’s Council, Resolution 87/2025/NQ-HĐND, attached regulation, Article 3 (zones) and Article 5, Table 1; read from the text of the resolution by kich_ban/l2_nong_nghiep.mjs.
Note: zone I consists of central wards; zone IV of rural communes of the former Binh Duong and Ba Ria–Vung Tau provinces and the Con Dao special zone. Land for perennial crops (Article 5, Table 2) is priced about a fifth higher. The script checks that every ward with a mapped boundary belongs to exactly one zone.
Agricultural land prices vary little across the city: position 1 in zone I is priced at position 1 in zone IV. Figure 2.2 sets both prices along a section through the city, with the stretch that lies inside the 2020 urban centre shaded.
Thousand VND per m², logarithmic scale; distance along the section, south negative
Source: residential prices as Map 2.1; agricultural prices as Table 2.2 (position 1, annual crops); urban centre as Map 2.3.
Note: the section is the line through the centroids of Can Gio, Binh Khanh, Nha Be, Ben Thanh, Thu Dau Mot, Ben Cat and Bau Bang; it includes every ward or commune whose centroid lies within km of the line. The shaded band is where the line lies inside the urban centre of 2020. The fitted line is the regression of Figure 2.1, drawn only over the 40 km on which it was estimated.
Figure 2.2 can also be drawn on the map itself. In Map 2.4 the prices of each ward stand above its place on the section line, on the same logarithmic scale, so that the reader can see which wards the section passes through and where it crosses the outline of the urban centre.
Residential and agricultural land prices by ward, thousand VND per m², logarithmic scale, standing on the line from Can Gio through Ben Thanh and Thu Dau Mot to Bau Bang
Source: as Figure 2.2; urban centre as Map 2.3; base map OpenFreeMap, © OpenStreetMap contributors.
Note: each ward is placed at the point of the section line nearest to the centroid of its boundary; the line passes through the centroids of Can Gio, Binh Khanh, Nha Be, Ben Thanh station, Thu Dau Mot, Ben Cat and Bau Bang.
Nowhere in the city does the residential price fall to the price of agricultural land: the lowest ratio of the two, in Chau Duc commune, is . The price table therefore does not locate \(\bar{x}\) by itself; the population grid of Map 2.3 does so more directly.
2.4Population density
Population density does not suffer from the problem of administered prices, and it can be measured on a regular grid. GHS-POP, produced by the European Commission’s Joint Research Centre, estimates the population of each 100 m cell. Here the cells are added up into a grid of 1 km cells within 40 km of Ben Thanh, and density is population divided by the area of each cell (Map 2.5, which also shows the outline of the 2020 urban centre from Map 2.3). In total, the circle contains people in 2020.
Persons per km², cells within 40 km of Ben Thanh
Source: European Commission, Joint Research Centre, GHS-POP R2023A, epoch 2020 (CC BY 4.0); aggregation by kich_ban/l2_tphcm.mjs; base map OpenFreeMap, © OpenStreetMap contributors.
Note: GHS-POP distributes census population to cells using satellite measurements of built-up area; it is a modelled estimate, not a count. Cells with fewer than 50 persons per km² are not drawn. Dashed circles are 10, 20, 30 and 40 km from Ben Thanh. The blue outline is the urban centre of 2020 (Map 2.3).
Persons per km², logarithmic scale
Source: as Map 2.5.
Note: dots are 1 km cells with at least 300 persons per km². The solid line is the density of each 1 km ring computed on the same cells: their total population divided by their total area. The fitted line is equation (2.1) estimated on the dots.
Density in the first kilometre around Ben Thanh is persons per km². It remains above up to 5 km, falls to at 9–10 km and to at 19–20 km. Beyond 20 km the ring averages stop falling for several kilometres; Thu Dau Mot lies at this distance (Map 2.5).
Estimated on cells with at least 300 persons per km², the gradient is \(b =\) : density falls by per kilometre (\(R^2 =\) , cells).
The threshold of 300 persons per km² is a choice. Table 2.3 repeats the regression with two other thresholds and with a second population grid, WorldPop 2020.
Regression of ln density on distance, 1 km cells within 40 km of Ben Thanh
| Population grid | Cells included | \(b\) | % per km | \(R^2\) | Cells |
|---|---|---|---|---|---|
| GHS-POP 2020 | all with population | ||||
| GHS-POP 2020 | ≥ persons/km² | ||||
| GHS-POP 2020 | ≥ persons/km² | ||||
| WorldPop 2020 | all with population | ||||
| WorldPop 2020 | ≥ persons/km² | ||||
| WorldPop 2020 | ≥ persons/km² |
Source: GHS-POP R2023A (European Commission, JRC); WorldPop (2018), unconstrained 100 m grid for 2020; kich_ban/l2_tphcm.mjs.
Note: the shaded row is the estimate used in the text. “% per km” is \(100(e^{b}-1)\).
The two grids give similar results for the same threshold. The threshold matters more: including sparsely populated cells at the edge of the circle steepens the gradient, and keeping only dense cells flattens it. Across the six rows, \(b\) ranges from to .
2.5Comparison with 192 cities
Liotta, Viguié and Lepetit (2022) assembled grids of population, rents and travel times for 192 cities on five continents, and estimated equation (2.1) for each of them.
Figure 2.4 places Ho Chi Minh City in the distribution of their density gradients. The city is not in their sample, so its estimate comes from this lesson.
Coefficient \(b\) from the regression of ln density on distance to the city centre, per km
Source: 192 cities: Liotta, Viguié and Lepetit (2022), Annex B, Figure B.6 (minimum, quartiles, maximum and mean, p. ; CC BY 4.0). Ho Chi Minh City: Table 2.3.
Note: box: first to third quartile; thick line: median; whiskers: minimum and maximum; open circle: mean. For Ho Chi Minh City, the dot is the estimate used in the text and the bar spans the six estimates of Table 2.3.
The estimate of is close to the median of . The comparison needs three qualifications. First, Liotta, Viguié and Lepetit locate the centre of each city with a rule of their own:
Second, they use an earlier version of the population grid:
Third, and most important, Table 2.3 shows that choices about which cells to include move the estimate for Ho Chi Minh City from below the first quartile of the 192 cities () to above the third (). A statement such as “Ho Chi Minh City has a typical density gradient” is only as good as the agreement between the two methods.
2.6Where the model does not fit
Two features of Ho Chi Minh City are not described by the monocentric model of Lesson 1.
A flat centre
The model predicts that density is highest at the CBD and falls most steeply near it (Figure 1.1). In Figure 2.3 density changes little over the first 5 km. The literature reports the same pattern in other cities:
More than one centre
Beyond 20 km, density stops falling near Thu Dau Mot (Figure 2.3), and land prices in Thu Dau Mot, Ba Ria and Vung Tau are several times those of wards at the same distance from Ben Thanh (Table 2.1). The merged city of 2025 contains several centres, not one. In 2011 the World Bank described the city before the merger as follows:
Liotta, Viguié and Lepetit find that such departures flatten gradients across their whole sample:
Extensions of the model that relax its assumptions are the subject of Brueckner (2011), chapter 3, “Modifications of the Urban Model”.
Exercise 2.1Interpreting a coefficient
A regression of ln density on distance gives \(b = -0.120\).
- By what percentage does density change per kilometre?
- What is the ratio of density at 15 km to density at 5 km?
Show answer
Exercise 2.2Set prices and market prices
Suppose that market prices near Ben Thanh are several times the prices in the land price table, while near the edge of the city the two are similar. Would the gradient of market prices be steeper or flatter than the gradient in Figure 2.1? What data would you need to check?
Show answer
Exercise 2.3Where does the city end?
Use Map 2.3 and Figure 2.2.
- Name one place that is part of the city in the sense of the model but not in the administrative sense, and one place for which the reverse is true.
- Along the section, the urban centre extends much farther north of Ben Thanh than south. Which assumptions of Lesson 1 does this contradict? Give two possible reasons.
- In Figure 2.2, compare residential prices just inside and just outside the shaded band, at its southern end and at its northern end. Does the edge of the urban centre look like the point where \(r(\bar{x}) = r_A\)?
- Would a density gradient estimated only on cells inside the administrative boundary be steeper or flatter than one estimated on a circle around Ben Thanh? Explain.
- In step 6, why is the urban centre of 1990 much larger than the built-up land inside it? What does this imply if you compare the area of a city over time?
Show answer
Exercise 2.4Choosing a measure
You must report one density gradient for Ho Chi Minh City to compare it with the 192 cities in Figure 2.4. Which row of Table 2.3 would you choose, and why? How does your choice change the conclusion of Section 2.5?
Exercise 2.5Discussion: subcentres
Using Map 2.1, Map 2.5 and Table 2.1, describe how land prices and density would vary around a subcentre such as Thu Dau Mot if its residents worked there rather than in central Ho Chi Minh City. Which assumption of Lesson 1 would you change to model this?
2.7Data and code
| Data | Licence | Accessed |
|---|
Scripts: kich_ban/l2_tphcm.mjs (gradients, tables, map data) · kich_ban/l2_vung_do_thi.py (administrative boundary and urban centre) · kich_ban/l2_luoi_dan_so.mjs (100 m density image, 500 m columns) · kich_ban/l2_ve_tinh.mjs (built-up land from Landsat, 1985–2015) · kich_ban/l2_nong_nghiep.mjs (agricultural land prices, section) · kich_ban/trich_dan.mjs (quotations) · kich_ban/b1_gia_dat.mjs (extraction of the land price table). Each script stops without writing its output if a check fails.
2.8References
- Brueckner, J. K. (2011). Lectures on urban economics. MIT Press. Chapter 3, Modifications of the urban model.
- Clark, C. (1951). Urban population densities. Journal of the Royal Statistical Society, Series A, 114(4), 490–496. doi:10.2307/2981088
- Dijkstra, L., Florczyk, A. J., Freire, S., Kemper, T., Melchiorri, M., Pesaresi, M., & Schiavina, M. (2021). Applying the Degree of Urbanisation to the globe: A new harmonised definition reveals a different picture of global urbanisation. Journal of Urban Economics, 125, 103312. doi:10.1016/j.jue.2020.103312
- European Commission, Joint Research Centre, Melchiorri, M., Mari Rivero, I., Florio, P., Schiavina, M., et al. (2024). Stats in the City – the GHSL Urban Centre Database 2025. Publications Office of the European Union. doi:10.2760/3046391, CC BY 4.0.
- de Bellefon, M.-P., Combes, P.-P., Duranton, G., Gobillon, L., & Gorin, C. (2021). Delineating urban areas using building density. Journal of Urban Economics, 125. Quoted from NBER Working Paper 26482 (2019), © 2019 the authors.
- European Commission (2024). GHSL Data Package 2023. Publications Office of the European Union. doi:10.2760/098587, CC BY 4.0.
- European Commission, FAO, UN-Habitat, ILO, OECD & World Bank (2021). Applying the Degree of Urbanisation: A methodological manual to define cities, towns and rural areas for international comparisons. Publications Office of the European Union. doi:10.2785/706535
- Marconcini, M., Metz-Marconcini, A., Üreyen, S., et al. (2020). Outlining where humans live, the World Settlement Footprint 2015. Scientific Data, 7, 242. World Settlement Footprint Evolution v1, © 2024 DLR, CC BY 4.0.
- Mari Rivero, I., Melchiorri, M., Florio, P., Schiavina, M., et al. (2025). GHS Urban Centre Database 2025 (GHS-UCDB R2024A), version 1.2. European Commission, Joint Research Centre. doi:10.2905/1A338BE6-7EAF-480C-9664-3A8ADE88CBCD, CC BY 4.0.
- Moreno-Monroy, A. I., Schiavina, M., & Veneri, P. (2021). Metropolitan areas in the world. Delineation and population trends. Journal of Urban Economics, 125.
- Schiavina, M., Freire, S., Carioli, A., & MacManus, K. (2023). GHS-POP R2023A: GHS population grid multitemporal (1975–2030). European Commission, Joint Research Centre. doi:10.2905/2FF68A52-5B5B-4A22-8F40-C41DA8332CFE, CC BY 4.0.
- Ho Chi Minh City People’s Council (2025). Resolution 87/2025/NQ-HĐND on the land price table applied from 1 January 2026.
- Liotta, C., Viguié, V., & Lepetit, Q. (2022). Testing the monocentric standard urban model in a global sample of cities. Regional Science and Urban Economics, 97, 103832. doi:10.1016/j.regsciurbeco.2022.103832. Quoted from the accepted manuscript, arXiv:2111.02112, CC BY 4.0.
- World Bank (2011). Vietnam urbanization review: Technical assistance report. World Bank. Open Knowledge Repository, CC BY 3.0 IGO.
- WorldPop (2018). Global 100m population. University of Southampton. doi:10.5258/SOTON/WP00645