The Rise of Europe: Atlantic Trade, Institutional Change, and Economic Growth
American Economic Review, 2005 · paper · Economics
Argues that Western Europe's rise after 1500 was largely driven by growth in Atlantic-facing trading nations. Atlantic trade strengthened merchants where early political institutions constrained the monarchy, bringing institutional changes that protected property rights and supported growth.
Checked: PDF from economics.mit.edu, fetched 5 Oct 2026. Page numbers are those of that PDF.
13 findings
Reference or label Figures 2A and 2B are captioned "GDP PER CAPITA, WEIGHTED BY POPULATION, 1500–1870" and plotted on a 0–2500 scale, but the y-axis title on both panels reads "Urbanization rate".
- p. 3 Figures 2A and 2B, y-axis
FIGURE 2A. WESTERN EUROPE, EASTERN EUROPE, AND ASIA: GDP PER CAPITA, WEIGHTED BY POPULATION, 1500–1870
- p. 3 Figure 2A and 2B, y-axis label
Urbanization rate
- p. 3 Figure 2B caption
GDP PER CAPITA, WEIGHTED BY POPULATION, 1500–1870
Judge's note: Page 3 image: both panels have the y-axis title 'Urbanization rate'. The scale runs 0–2500, the captions say GDP per capita, and the plotted values match the Table 1 GDP figures (e.g. Eastern Europe about 500 in 1500). The axis title looks carried over from Figure 1 on page 2.
Second check: another model agreed
Report a problem with this finding
Reference or label The notes to Tables 1, 2 and 7 cite "Bairoch (1998)" for Asian urbanization. The reference list has only Bairoch 1988, Bairoch 1993 and Bairoch et al. 1988, and the text (p. 7) and Appendix Table 1 cite "Bairoch (1988)" for the same Asian data.
- p. 4 Table 1 notes
comparable data for Asia are from Bairoch (1998)
- p. 8 Table 2 notes
urbanization in Asia is from Bairoch (1998)
- p. 7 Section I.A text
qualitative assessments of Bairoch (1988)
- p. 26 Table 7 notes
urbanization in Asia is from Bairoch (1998)
Judge's note: Confirmed '(1998)' in the notes to Table 1 (p4), Table 2 (p8) and Table 7 (p26). Page 7 cites 'Bairoch (1988)' for the Asian urbanization estimates. The reference list on page 30 has no 1998 Bairoch entry. The likely fix is 1988.
Second check: another model agreed
Report a problem with this finding
Reference or label Footnote 8 says "The analysis of city-level growth in Section I B is informative...", but Section I B is "Economic Growth in Europe" (country-level). The city-level analysis is in Section I D, "Urban Expansion and Atlantic Ports" (Table 5).
- p. 7 Footnote 8
The analysis of city-level growth in Section I B is
- p. 7 Footnote 8
The analysis of city-level growth in Section I B is informative
- p. 7 footnote 8
The analysis of city-level growth in Section I B is informative on differential growth across historical political boundaries.
Judge's note: Footnote text checked on the page 7 image. The section headings are A Data, B Economic Growth in Europe, C Other Determinants, D Urban Expansion and Atlantic Ports. The city-level material (Table 5, Figure 4) is in I.D.
Second check: another model agreed
Report a problem with this finding
Reference or label Appendix Table 1 cites "Chandler (1996) for Asian city population". The text on p.15 cites "Tertius Chandler (1987)", and the reference list has only Chandler 1987.
- p. 29 Appendix Table 1, Urban population source
Chandler (1996) for Asian city population
- p. 15 Section I.D
Column 7 adds Asian cities from Tertius Chandler (1987)
- p. 15 Section I.D text
Tertius Chan- dler (1987)
- p. 15 Section I.D
Asian cities from Tertius Chandler (1987)
Judge's note: I saw the Appendix Table 1 wording on the p.29 image and the Section I.D text on the p.15 crop. The reference list entry (text layer) is 'Chandler, Tertius. Four thousand years of urban growth... 1987'. There is no Chandler 1996 entry.
Second check: another model agreed
Report a problem with this finding
Reference or label Stone is quoted with "(1973, p. 144)" but the only reference entry reads "Stone, Lawrence. The causes of the English Revolution, 1529–1642. New York: Routledge, 1972."
- p. 20 Section II.B, Britain
on the side of Parliament” (1973, p.
- p. 34 References
olution, 1529–1642. New York: Routledge, 1972.
- p. 20 Section II.B, Britain
(1973, p. 144)
- p. 34 References, Stone
Stone, Lawrence. The causes of the English Rev- olution, 1529–1642. New York: Routledge, 1972.
Judge's note: Page 20 image shows '(1973, p. 144)'. Page 34 shows a single Stone entry dated 1972. The year in the text and the year in the reference list differ.
Second check: another model agreed
Report a problem with this finding
Reference or label Table 7 Panel C columns 6–10 are headed "Panel, 1500–1850" but report 240 observations, more than the 192 reported for the longer 1300–1850 constraint panel in columns 1–5 and in Table 6.
- p. 26 Table 7 Panel C, column headers and N
Number of observations 192 192 192 192 192 240 240 240 240 240
- p. 26 Table 7, Panel C, columns 6–10
Panel, 1500–1850
- p. 26 Table 7, Panel C number of observations
240
- p. 26 Table 7, Panel C column headers (6)–(10) and observation counts
1500–1870 ... 1500–1850 ... 240
Only the authors can say what is right here.
Judge's note: Page 26 image: Panel C columns 1–5 are 1300–1850 with N=192, and columns 6–10 are 1500–1850 with N=240. Table 6 (constraint, 1300–1850) also has N=192. With 24 countries, 1500–1850 (6 dates) would give 144. The 240 matches Panel A's 1000–1850 columns, so the header or the N is wrong. Only the authors can say which.
Second check: another model agreed
Report a problem with this finding
Reference or label Table 7 Panel B (log GDP) row label reads "p-value for initial institutions × year (1600, 1700, 1750, 1800, 1850)", but the GDP dates are 1600, 1700, 1820 and 1870.
- p. 26 Table 7 Panel B row label
p-value for initial institutions year (1600, 1700, 1750, 1800, 1850)
Judge's note: The label appears to be carried over from Panel A. The GDP dates are given on p.10 and in Appendix Table 1.
Second check: another model agreed
Report a problem with this finding
Reference or label Table 7 notes say "urbanization in Asia is from Bairoch (1998)". The reference list has no Bairoch 1998 entry, and Appendix Table 1 gives this source as "We use Bairoch (1988) for urbanization in Asia".
- p. 26
urbanization in Asia is from Bairoch (1998)
Judge's note: I saw the wording on the p.26 image. The reference list (text layer, p.30 area) has Bairoch 1988 (Cities and economic development), Bairoch 1993, and Bairoch et al. 1988 only. I viewed Appendix Table 1 on the p.29 image, and it cites Bairoch (1988) for Asian urbanization. The text layer shows the same 'Bairoch (1998)' in the notes to Table 1 and Table 2, so the wrong year appears in several places.
Second check: another model agreed
Report a problem with this finding
Reference or label Appendix, 1751–1800: "Inikori's British America trade estimates of £19,545 for 1781–1800 implies annual profits of around £10m". The unit is missing or wrong; the figure is presumably £19.545m (£19,545 thousand).
- p. 29 Appendix, 1751–1800
estimates of £19,545 for 1781–1800 implies annual profits of around £10m
- p. 29 Appendix, 1751–1800
estimates of £19,545 for 1781–1800 implies
Judge's note: Checked on the p.29 image. The neighbouring periods quote Inikori in £m (for example £6.8m on p.28) and use a 50% profit share. Profits of about £10m require trade of about £19.5m, so '£19,545' as printed is off by a factor of 1,000.
Second check: another model agreed
Report a problem with this finding
Reference or label The O'Rourke and Williamson (2002) reference gives "Journal of Economic History, 2002, 62(2), pp. 471–56", a page range that ends before it starts.
- p. 33 References, Neal 2000
1648–815.
- p. 33 References, O'Rourke and Williamson
62(2), pp. 471–56.
- p. 33 References, O'Rourke and Williamson
pp. 471–56
- p. 33 References, O'Rourke and Williamson
62(2), pp. 471–56
Only the authors can say what is right here.
Judge's note: Seen on the p.33 image. The reference list abbreviates end pages (e.g. '117–40' for 117–140), so '471–56' would mean pages 471–456, which is impossible. Rests on outside knowledge: that the true range is probably 417–456, i.e. transposed start digits. The Neal '1648–815' defect also listed here is handled separately as C042.
Second check: another model agreed
Report a problem with this finding
Wording The text says GDP per capita in the rest of Western Europe "grew at approximately the same rate as in Eastern Europe, just under 30 percent". Table 1 gives about 29% for non-Atlantic Western Europe (850.73→1095.40) but about 42% for Eastern Europe (506.94→719.5) over 1500–1820.
- p. 3 Introduction, paragraph on Figure 2
same rate as in Eastern Europe, just under 30 percent
- p. 4 Table 1, GDP per capita rows
506.94
- p. 3 Introduction, Figure 2 discussion
it grew at approximately the same rate as in Eastern Europe, just under 30 percent
- p. 3 Introduction, para after Figure 2
in the rest of Western Europe it grew at approximately the same rate as in Eastern Europe, just under 30 percent
Judge's note: Checked the text on the page 3 image and Table 1 on the page 4 image. The non-Atlantic figure matches 'just under 30 percent', but Eastern Europe's growth in Table 1 is about 42%. Figure 2B shows the same gap (about 480 to about 660). In the same sentence, Atlantic traders' 'factor of almost two' (1.83) is consistent with the table.
Second check: another model agreed
Report a problem with this finding
Wording For the log GDP specification, the text says "countries with Roman heritage grew more rapidly between 1400 and 1600". The GDP panel (Table 4, panel B, column 7) covers 1500 to 1820, and the Roman heritage interactions start in 1600.
- p. 13 Section I.C, Roman heritage paragraph
Roman heritage grew more rapidly between 1400 and 1600
- p. 13 Section I.C
grew more rapidly between 1400 and 1600
- p. 10 Section I.B, Table 3 description
Maddison (2001) reports estimates of GDP per capita for 1500, 1600, 1700, 1820, and 1870
Only the authors can say what is right here.
Judge's note: Table 4 on the page 12 image heads column 7 'Panel, 1500 to 1820, controlling for Roman heritage', and its notes say Roman heritage is interacted with year dummies for 1600 and after. Page 10 says Maddison GDP starts in 1500. The 1400 start cannot come from this regression; it may have been meant as 1500.
Second check: another model agreed
Report a problem with this finding
Wording Footnote 18 says "cumulative growth between 1500 and any subsequent date is always higher for Atlantic ports than for potential Atlantic ports". Using the Table 5 Panel A coefficients (base 1300–1400), potential Atlantic ports show more growth from 1500 to 1750, 1800 and 1850 in the weighted pair (cols 1 vs 3) and from 1700 on in the unweighted pair (cols 2 vs 4).
- p. 14 Footnote 18
always higher for Atlantic ports than for potential Atlantic ports
- p. 15 Table 5 columns 1 and 3
0.70 0.71 0.79 0.89
- p. 14 footnote 18
cu- mulative growth between 1500 and any subsequent date is always higher for Atlantic ports than for potential Atlantic ports
- p. 15 Table 5 Panel A, columns 1 and 3
Atlantic port 1800 0.79 0.92 0.95 1.10
Only the authors can say what is right here.
Judge's note: Read Table 5 on the page 15 image. Col 1: 1500 −0.04, 1750 0.70, 1800 0.79, 1850 1.09. Col 3: 0.027, 0.79, 0.95, 1.19. Col 2: −0.05, 1700 0.62, 1850 1.00. Col 4: 0.048, 0.76, 1.23. Growth since 1500 is the date coefficient minus the 1500 coefficient. Atlantic ports are higher only for 1600–1700 (weighted) and 1600 (unweighted). These are point estimates with large standard errors, but the footnote's 'always' does not hold for the printed numbers. The reviewer remark about 'quantitatively smaller' is handled under C020.
Second check: another model agreed
Report a problem with this finding
16 minor (typos and the like; most checked by the judge only)
Wording The text says the Western Europe × year interactions 'are again insignificant', but in Table 2 panel B the p-values are [0.02] in column 9 and [0.00] in column 6, with [0.06] in column 4.
- p. 10 Section I.B, discussion of equation (2)
In all columns, the estimate of
- p. 8 Table 2 Panel B, p-value row
[0.35] [0.06] [0.83] [0.00] [0.11] [0.16] [0.02] [0.81]
- p. 10 Discussion of Table 2 panel B
are again insignificant
Judge's note: The column 6 result (with Asia) is explained earlier in the text; column 9 is not explained. The statement overstates the table.
Second check: another model agreed
Report a problem with this finding
Wording The text says that when Atlantic trade is omitted 'the estimates of αt's are significant, positive, and large' in both Table 2 and Table 3, but Table 3 column 1 (1500–1820) has p = [0.44].
- p. 11 Section I.B summary paragraph
Overall both Table 2 and Table 3 show an
- p. 11 Table 3 Panel A col 1
[0.44] [0.05]
Judge's note: Page 10 already calls the GDP pattern 'somewhat less pronounced', so this is an overstatement in the summary sentence rather than a hidden contradiction.
Second check: another model agreed
Report a problem with this finding
Wording Text says column 6 coefficients "after 1700 are about half those of column 1"; the printed ratios are about two-thirds for 1700–1800 and 0.42 for 1850.
- p. 14 Discussion of Table 5 column 6
tions after 1700 are about half those of column 1
- p. 15 Table 5 column 6
0.47
Judge's note: Table 5 col 6 vs col 1. By contrast, the column 5 'approximately halved' claim does hold. The column 6 wording is loose rather than wrong in direction.
Second check: another model agreed
Report a problem with this finding
Wording Column 2 (unweighted) is described as "similar, but quantitatively smaller" than column 1, yet several Panel A coefficients are larger: ×1600 0.46 vs 0.36, ×1750 0.71 vs 0.70, ×1800 0.92 vs 0.79.
- p. 14 Section I D, column 2 discussion
The results are similar, but quantitatively smaller
- p. 15 Table 5 Panel A, Atlantic port × 1800
0.92
Judge's note: Read Table 5 on the page 15 image. The unweighted coefficients are smaller only at 1700 (0.62 vs 0.71) and 1850 (1.00 vs 1.09), and in Panel B (0.16 vs 0.17). 'Quantitatively smaller' fits only part of the table, so it is loose wording rather than a clear error.
Second check: another model agreed
Report a problem with this finding
Wording Section IV says results with a different regressor are the same: "(results using the coastline-to-area ratio measure are identical, and are contained in Acemoglu et al., 2002b)". Elsewhere the paper calls such robustness results 'similar'.
- p. 26 Section IV, first line
to-area ratio measure are identical
- p. 25 Section IV, before Table 7
(results using the coastline-
- p. 26 Section IV
to-area ratio measure are identical, and are
- p. 26 Section IV, first paragraph after Table 7
to-area ratio measure are identical, and are contained in Acemoglu et al., 2002b
Judge's note: Checked the p.25–26 images (crops b and d, and the full p.26). The coastline-to-area estimates for Table 7 are not shown in this paper. A continuous regressor cannot give coefficients identical to those from the 0/1 Atlantic trader dummy. The paper's other robustness statements use 'similar', and the next paragraph begins 'The results in all three panels are similar'. 'Identical' is probably meant qualitatively, so this is a wording overstatement, not a wrong number.
Second check: another model agreed
Report a problem with this finding
Wording Footnote 33 says "in three specifications in Table 7, the interactions between initial institutions and dates after 1600 are jointly significant". Table 7 prints four p-values below 0.05 for this test: Panel B cols 4 [0.004] and 9 [0.01], and Panel C cols 4 [0.008] and 9 [0.01]. Which three are meant?
- p. 27 Footnote 33 continuation
in three specifications in Table 7, the interactions between initial institutions and dates after 1600 are jointly significant
- p. 26 Table 7 p-value rows
[0.40] [0.31] [0.004] [0.08] [0.66] [0.64] [0.01] [0.58]
- p. 26 Table 7 p-value rows
[0.004]
Only the authors can say what is right here.
Judge's note: I read every p-value row in Table 7 on the p.26 image. Panel A: 0.61, 0.51, 0.71, 0.85, 0.12, 0.08, 0.42, 0.92. Panel B: 0.40, 0.31, 0.004, 0.08, 0.66, 0.64, 0.01, 0.58. Panel C: 0.27, 0.14, 0.008, 0.69, 0.43, 0.33, 0.01, 0.95. That gives four below 0.05 and six below 0.10, so no usual threshold gives three. The footnote (p.27 image) may count only specifications where the unreported coefficients are negative. That is not stated, so I am raising it as a query.
Second check: another model agreed
Report a problem with this finding
Wording Appendix, 1576–1600: "total profits from privateering in 1585–1603 were £700,000. Dividing by 25 years gives an average of £28,000 per year". The 1585–1603 span is 19 years, and the paper does not say why it divides by 25, the length of the 1576–1600 window.
- p. 28 Appendix, 1576–1600
1585–1603 were £700,000. Dividing by 25 years gives an average of £28,000 per year
- p. 28 Appendix, 1576–1600
profits from privateering in 1585–1603 were £700,000. Dividing by 25 years gives an average of £28,000 per year
- p. 28
1585–1603 were £700,000. Dividing by 25
Only the authors can say what is right here.
Judge's note: Checked on the p.28 image. The division is correct as printed (700,000/25 = 28,000). The divisor is the length of the target window, 1576–1600, not of the source period. The total also includes 1601–1603, which falls outside that window. Spreading the total over the whole window may be deliberate, but the text does not explain it, so I am raising it as a query rather than an error.
Second check: another model agreed
Report a problem with this finding
Wording The deflator is described as a "50 percent increase from 1600 to 1650, followed by rough stability through 1700", yet £600,000 is converted to "approximately £500,000" and £1.1m to "£900,000" in 1600 prices (factor about 1.2, not 1.5).
- p. 28 Appendix, 1651–1675
yielding profits of £600,000 per annum, or approximately £500,000 in 1600 prices
- p. 28 Appendix, deflator description
then a 50 percent increase from 1600 to 1650, followed by rough stability through 1700
Only the authors can say what is right here.
Judge's note: Appendix p.28. Other conversions in the paper (for example £800k→£500k for 1688) use about 1.6. Authors should confirm the conversion factors.
Second check: another model agreed
Report a problem with this finding
Typo The Table 1 note uses 'Atlantic coast-to-area', while the table row and the rest of the paper use 'Atlantic coastline-to-area'.
- p. 4 Table 1 notes
Atlantic coast-to-area includes those parts of Germany
Judge's note: Minor terminology inconsistency, visible in the Table 1 notes.
Second check: none (judge only)
Report a problem with this finding
Typo Ungrammatical phrase: "constraint on the executive, which we will be described in greater detail and used in Section III".
- p. 7 Section I.A, last paragraph
utive, which we will be described in greater
- p. 7 Section I.A, last paragraph
which we will be described in greater detail
- p. 7 Section I.A, paragraph after Table 1 description
which we will be described in greater detail and used in Section III
Judge's note: Seen on the page 7 image. It should read 'which we will describe' or 'which will be described'.
Second check: none (judge only)
Report a problem with this finding
Typo "the differential growth of Western Europe from 1600 is closely linked to the extension of Atlantic trader." The word should be "trade".
- p. 11 Section I.B, first line on page
the extension of Atlantic trader.
Judge's note: Seen on the page 11 image. The sentence describes the structured specification, which uses the volume of Atlantic trade.
Second check: none (judge only)
Report a problem with this finding
Typo Table 5 mixes decimal precision: "0.297" (col 5, 1700), "0.799" (col 7, 1800) and "0.7" (col 7, 1750), where other cells use two decimals.
- p. 15 Table 5 panel A, columns 5 and 7
0.297
- p. 15 Table 5 panel A, column 7, 1800 and 1750 rows
0.799
Judge's note: Seen in the Table 5 Panel A crop. This is cosmetic formatting only.
Second check: none (judge only)
Report a problem with this finding
Typo Two misspellings. Appendix Table 1 has "Doring Kindersley (DK) Publishers (1997)", where the reference list has 'Dorling Kindersley'. The Rabb (1998) reference has "Sir Edwin Sandays", where the text has 'Sandys'.
- p. 29 Appendix Table 1, ports row
Doring Kindersley (DK) Publishers (1997)
- p. 33 References, Rabb 1998
Sir Edwin Sandays
- p. 29
Doring Kindersley (DK) Publishers
Judge's note: I saw 'Doring Kindersley' on the p.29 image. The reference list (text layer) spells it 'Dorling Kindersley (DK) Publishers'. The same table also cites 'DK Publishing (1997)', a minor naming inconsistency. I saw 'Sandays' in the Rabb 1998 entry on the p.33 image. The text spells the name 'Sir Edwin Sandys' twice.
Second check: none (judge only)
Report a problem with this finding
Typo The Bairoch, Batou and Chèvre (1988) reference title reads "La population des villes europeenees de 800 a 1850" and the publisher is given as "Librarie Droz".
- p. 30 References, Bairoch, Batou and Chèvre
La population des villes europeenees de 800 a
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the French spellings "européennes" and "Librairie Droz". The crop of page 30 shows "europeenees" and "Li-brarie Droz" as printed. The doubled "ee" is a misspelling, not an accent lost in extraction.
Second check: none (judge only)
Report a problem with this finding
Typo The Rabb (1998) reference title reads "Jacobean gentleman: Sir Edwin Sandays, 1561–1629", while the text spells the name "Sir Edwin Sandys" twice.
- p. 33 References, Rabb 1998
Jacobean gentleman: Sir Ed- win Sandays
- p. 33 References, Rabb 1998
Sir Ed- win Sandays
- p. 33 References, Rabb 1998
Sir Ed- win Sandays, 1561–1629
Judge's note: Page 33 shows "Sandays" in the reference entry. Pages 20 and 21 spell the name "Sandys", both in the text and inside the quotation. The same person is meant, so the spellings are inconsistent.
Second check: none (judge only)
Report a problem with this finding
Typo The title in the Neal (2000) reference reads "...During the First Global Capital Markets, 1648–815.", where the end year should be 1815.
- p. 33 References, Neal 2000
1648–815.
- p. 33 References
1648–815.”
Judge's note: Seen on the p.33 image. Other titles in the list print year ranges in full (e.g. 1450–1750, 1568–1648), so '815' is a dropped digit. This is distinct from the O'Rourke page range in C040.
Second check: none (judge only)