Hanseatic League · beneath the chronicle #1
Did the Hanseatic League actually decline?
Lübeck’s own toll ledger says taxable trade was up 21 percent across the century the League is supposed to have declined in. Getting from that number to what actually happened takes two deflations and a count of ships.
The +21% lie
In the toll year that ran from 18 March 1368 to 10 March 1369, the clerks of Lübeck's pound toll recorded 546,000 marks lübisch of taxable trade moving through their port. The pound toll was a war levy: a small percentage charged on the declared value of cargo, imposed by the Hanseatic towns to pay for fleets, and its books are the closest thing the medieval Baltic has to a customs return. In 1492, a century deeper into the decline every standard account describes, the same city's pound toll recorded 660,000 marks. Up 21 percent.
Both figures are real. Neither one means what it looks like it means.
The suspect is the unit. A mark lübisch was not a coin; it was an accounting promise about silver, and the promise was kept differently in 1492 than in 1369. A money of account is itself a historical actor, with its own biography of debasements. The modern name for the correction is the nominal–real distinction, the reason nobody compares a 1970s salary to a modern one without deflating first. The medieval version is done in grams, and it has to be done twice: once for the silver the mark promised, and again for what that silver would buy.
Two divisions, in order. For the metal, multiply each toll total by the fine silver its own mark contained: the late-fourteenth-century mark held roughly 50–62 g depending on whose reconstruction you take, the mark of 1461 about 20 g. For the goods, take the nominal ratio of the two totals, 660,000 ÷ 546,000 = 1.209, and divide it by the ratio of rye prices between the two dates. Whatever that second division returns is the answer in things rather than in money.
If rye costs twice as much when you take the second measurement, the same pile of money buys half as much rye. So to compare two piles of money a century apart, divide out whatever happened to prices in between. Here that has to be done twice over, because two separate things moved: how much silver the coin held, and how much bread the silver fetched.
Take the ledger's side seriously first, because it has a real case. The pound-toll books are contemporary records, kept by men with money at stake, in the very years they describe. The decline story is retrospective: assembled centuries later, organized around an ending its authors already knew. Between a clerk counting cargoes on the quay and a synthesis written with hindsight, the burden of proof ought to sit with the synthesis. If the best surviving totals from the League's own ledgers rise by a fifth, perhaps the famous decline is the artifact, a story projected backward from 1669, when the last diet met, onto centuries that did not know they were declining.
The Lübeck silver flip
Two toll benchmarks; three ways of asking what they mean. The anchors never move. Change what the y-axis measures — marks, then the silver behind the marks, then the rye that silver would buy — and watch the century change its answer.
Two series, named rather than reconciled. They agree from 1400 on and disagree about the base year, where it matters most: Volckart puts the mark of 1365–74 at 61.53 g against the Riksbank reconstruction's 50–55 g for the late fourteenth century. Volckart's is the internally consistent choice here, it comes from the same Hamburg accounts as the rye prices, and it is what puts the silver fall at −57% to −60% rather than the shallower −52% the Riksbank base would give. Splice note: the Riksbank points through 1461 are the accounting-mark reconstruction (Vol. I ch. 3, verbatim-confirmed); the ~1506 point is the Staatsmark coin standard bridged on, a depreciation trajectory, not one homogeneous series.
Silver content ≠ purchasing power, which is what the rye reading is for. And the two books may not count quite the same toll universe, a limit no re-basing can fix.
The rye reading converts both toll totals at the rye price of their own decade, from Oliver Volckart, “Prices in Mark of Lübeck (14th–16th century)”, IISH Historical Prices & Wages (hdl:10622/VY7UY3), compiled from the Hamburg Kämmereirechnungen, the Lüneburg city archive and the published Lübeck editions. Two substitutions are built in: rye stands in for a toll basket that was mostly not grain (cloth, wax, furs, salt, herring, metals), and Hamburg and Lüneburg prices stand in for Lübeck's. Both are argued below the chart.
Static view: all three readings shown one below the other, and the 1492 column carries both dearth conventions. With scripts enabled, one control re-bases the axis between them.
What you just did: re-based the axis, twice. The anchors never moved. The unit under them did, first to the silver it promised, then to the rye that silver would buy.
Re-base those totals into the silver the marks actually promised and grew-by-a-fifth becomes shrank-by-half. The rise was the unit rotting.
Silver is not purchasing power either, and its own value moved: the fifteenth century was Europe's bullion famine, silver grew scarce and dear, and grain priced in silver fell by roughly a quarter between the 1370s and the 1490s. So deflate a second time, into something a merchant would have recognised. What could the trade taxed in each year actually buy?
In rye, about a third less. Set the two toll totals against grain prices kept in the same currency by the chambers of Hamburg and Lüneburg, and the taxed trade of 1368–69 is worth some 160,000 wispel of rye, a wispel being roughly 840 litres of grain, against about 92,000 to 117,000 in 1492. The spread is one honest fork rather than a failure of measurement: 1491 and 1492 were dearth years in that grain market, prices better than doubled, and the 1492 toll year sits inside the crisis. Count those years and the fall is 42 percent; average them out and it is 27. Lübeck's own scattered prices side with the smaller figure. And the check that makes it more than arithmetic: Antwerp's market prices, a separate archive in a separate currency, bridged only through the silver content of the mark, return the same two answers to within a single percentage point.
One limit survives both answers. The two books do not count quite the same world: the later one barely sees North Sea traffic, the earlier one falls in a year when Norway was under embargo, and compliance visibly decayed even inside the later window. That caps how precisely any of these percentages can be pushed. It does not restore the +21 percent, which needed the mark to hold its silver and it did not.
So Lübeck's ledger will tell you how much smaller its trade got. What it cannot tell you is where the trade went instead, and the chronicle rarely entertains the possibility that it went somewhere: that we have been watching the wrong city.
The wrong capital
The Gdańsk Grain Exchange still boasts about it: 118,000 lasts of grain shipped through Danzig in 1618, some 271,000 tonnes. Rewind to the 1490s, the very decades the chronicle has the Hansa fading, and Danzig was already moving about 10,000 lasts of rye a year, on its way to controlling three-quarters of the Baltic's grain exports.
Somebody forgot to tell the Vistula about the decline.
The wider Baltic grain story, the “mother trade” that would feed Amsterdam's rise, has its own chapter in the economic-history book; here it matters as a single fact: the grain kept coming.
The chronicle's geography is not stupid. Lübeck earned its billing: the diets convened there, Lübeck law governed appeals from a hundred towns, the war chests were subscribed there, and when the League spoke to kings it spoke through Lübeck's seal. If political weight tracks economic weight, and it usually does eventually, then the Queen of the Hansa was the right place to watch.
And once Danzig's rise is conceded, the tidy mind wants the rest: from 10,000 lasts in the 1490s to 118,000 in 1618, surely a smooth curve of growth connects the dots, the granary of Europe filling on schedule.
Danzig: the anchors, and the line you choose to draw
The hard anchors are two, a century and a quarter apart. Everything between them is an authorial choice: two different histories fit the same points. Link’s harbor books say the shocky one is likelier.
Marks follow the series legend: solid = recorded, open = estimate, hatch = assumption, dashed = hypothetical.
Path-choice space, not statistical uncertainty: both lines fit the same two anchors. The anchors don’t even share a commodity basket, rye-only (1490s) vs broader grain (1618), carried in the provenance cards, not resolved.
= different series: Dutch Baltic grain imports, all origins (Brand 2007), context only, not a Danzig anchor; the hypothetical paths never touch it.
✕ The “80,000 lasts in 1567” that wasn’t
The research dossiers behind this page carried a hard-looking anchor: “80,000 lasts of Danzig grain exports, 1567.” Source-checking exposed it as a misattribution: in the source (Brand 2007), 80,000 lasts is the loading capacity of Holland’s merchant fleet in 1567, not anyone’s exports. The corrected figure for that year — Dutch imports of Baltic grain, ≈60,000 lasts — is a different series (all Baltic origins, not Danzig exports), so it renders as a square context marker the paths never touch. Which means: between the 1490s and 1618 there is no direct Danzig export observation at all, only the derived ≈50k marker from the verified “quintupled over the 16th century” claim. The chart you are reading is more honest than the one the dossiers implied.
Static view: both hypothetical paths shown faint. With scripts enabled, the chart starts with anchors only, you must choose to draw a line.
What you just did: chose whether history gets a line at all, the published curves chose for you, in the same ink as the anchors.
Draw both candidate histories and the record votes for neither comfort. Christina Link's reading of the surviving harbor books is blunt: no smooth breadbasket curve fits. Harvests, war, and Polish–Prussian politics made the real path lurch. The two anchors do not even count the same thing, rye alone in the 1490s, all grain in 1618, in a unit that changes weight by commodity and port. And the one mid-century number this research first carried, 80,000 lasts in 1567, dissolved on source-checking into a Dutch fleet-capacity figure that was never Danzig's exports at all; the chart wears that correction openly.
What survives every caveat is the direction. While the League's machinery rusted, the Baltic economy it supposedly embodied grew, and the growth pooled somewhere the chronicle wasn't standing. “Hanseatic decline” is partly a statement about where the observer chose to stand.
Two stages in, the suspicion is hard to avoid: how much of any Hanseatic trade curve you have ever seen was measurement at all?
What the record will bear
You have just accepted three of my assumptions. That rye can stand in for a cargo of cloth and wax and furs. That Hamburg's grain prices can speak for Lübeck's. That two toll books a century apart are counting comparable things, or near enough. Each one was stated where you could see it, and each one could be wrong. Now turn that same suspicion on every other number you have ever met about this trade, starting with the map: thick, confident arrows sweeping from Novgorod to Bruges, sized as though a customs computer had been running since 1250. Here is what actually survives behind arrows like those. One toll-year cross-section from Lübeck. A handful of Danzig harbor books. English customs rolls for the ports where Hansards traded. And then nothing continuous anywhere until 1497, when the Danish Sound registers begin, near-complete only from 1574, recording destinations only from 1668.
That is the whole inventory, and each piece has a ceiling. Cross-sections support comparisons, never trends. Toll values are declared values, never tonnages. The Sound registers see the Sound, never the whole Baltic. The appendix at the end of this page itemizes every number used here with its label and source, the same discipline the arrows never show.
None of which makes reconstruction a sin. Historians must bridge gaps, and the good ones build their bridges in the open. Watch a master do it. Norway's stockfish trade through Bergen: English customs, the one well-kept series, show England receiving 1,500 to 2,000 tonnes a year in 1303–11. Knut Helle's estimate of the total rests on a single stated assumption, that the continent took about as much as England, and out comes 3,000 to 4,000 tonnes. Every step is inspectable. Disagree with the assumption and you can recompute the answer yourself.
Bergen: the assumption slider
Helle’s reconstruction, made manipulable. England’s customs anchor the numerator; the share you assume for England sets the total. The published number is one slider position.
Marks follow the series legend: solid = recorded, open = estimate, hatch = assumption, dashed = hypothetical.
At 50%, the dossier’s assumption, the implied Bergen total is 3,000–4,000 t of stockfish per year.
The band is England’s anchor divided by the share you assume, parameter-choice space, not a confidence interval. The two context overlays are fixed: they never move with the slider.
Static view: the parameter sweep with the 50% assumption marked. With scripts enabled, the slider moves the assumption.
You just did what every historian of this trade must do, the difference is you saw yourself do it.
That is what the slider above hands you: Helle's footnote made manipulable. Set England's share at a third instead of half and the “total” swells by half again. Nothing about the past changed; only your assumption did. Reconstruction earns its keep exactly when it is exhibited like this. Helle exhibited his; the trouble starts downstream, when a number built on a stated assumption gets redrawn as an arrow with no assumption attached to it.
You now hold this series' one reflex, and it fits on a bumper sticker: solid marks are the past; hatched marks are us.
Carry it back to the question we started with, and ask that question properly this time.
So did it decline?
The classic statement is Philippe Dollinger's The German Hansa, still the synthesis everything else argues with. Its arc is the one you know even if you have never read it: a rise through the thirteenth century, a golden age sealed by victory over Denmark at Stralsund in 1370, then the long slide past Dutch competitors and territorial princes to the last, thinly attended diet of 1669. The documentaries and the popular histories descend from that arc.
And the slide was real. Whatever the ledgers say about trade, the institution measurably decayed. Ivan III shuttered the Novgorod Peterhof in 1494. The Bruges kontor withered with Bruges itself. In London the Crown revoked the Steelyard's privileges, and Hanse cloth exports collapsed from about 44,000 cloths in 1550 to under 14,000 by 1552, a natural experiment in what the privileges had been worth. Diets were summoned; towns stopped coming. No honest verdict waves this away as a matter of perspective.
So the verdict has to split the word “decline” into the three claims it smuggles together. The unit declined: the mark lost over half its silver, and with it went the legibility of every nominal comparison. Lübeck declined: its taxed trade fell by about a third in goods terms and its shipping by more than half. But the Baltic economy did not decline with it, which is the distinction the single word keeps hiding, the grain kept moving, the Sound filled with hulls, and the growth pooled in Danzig and in routes that had stopped calling at Lübeck. And the institution declined unambiguously: kontors, privileges, diets, the whole collective machinery genuinely died.
The evidence that separates those three cleanest is the one measurement on this page that needs no deflator, no silver content and no assumption at all. The same two toll books that produced the disputed values also counted hulls: about 1,800 ships in the toll year of 1368, about 760 in the 1490s. Scania herring, the cargo the city was built on, falls from some 70,000 barrels a year to 14,477. Salt rises. A port losing more than half its ships while its signature trade collapses and its bulk trade grows is not a port simply emptying. It is a port being reorganised around something else, by ships that increasingly had somewhere better to call.
The mechanism that best connects all of it is intermediation. Lübeck's position as the mandatory middleman eroded while the sea filled with ships that no longer needed one. The modern literature broadly agrees that the monolithic decline story is wrong; how much was margin loss, how much rerouting, how much institutional decay, remains open, because the record you have now seen cannot yet decompose them. The frame is settled; the numbers are still loose.
The questions this verdict opens are geographic. Which routes, bypassing what, starting when? That argument does not want an essay; it wants a map, and the map version of this story is being built next. The evidence runs thickest in two places this series will go next: Danzig's grain trade, and the year England cancelled the Steelyard's privileges and accidentally ran the experiment.
Appendix: every number on this page, and where it comes from
Labels: [CR] contemporary record · [M] modern scholarly tabulation of contemporary records · [E] scholarly estimate · [D] derived on this page. Sources were independently re-verified against the cited literature in July 2026; one number failed that check and was removed (see the note on the Danzig chart).
| Number | Label | Source | Status |
|---|---|---|---|
| Lübeck pound-toll total, 18 Mar 1368, 10 Mar 1369: ≈546,000 marks lübisch (≈339k in / ≈207k out) | [M] | Pfundzoll books, tabulated by Wendt (1902) and Lechner (1935); reported in Dollinger, The German Hansa | verified 2026-07 |
| Lübeck pound-toll total, 1492: ≈660,000 marks (≈442k in / ≈218k out) | [M/E] | Dollinger, from the Pfundzoll books of 1492–96 (Bruns' analysis; Vogtherr 1996 edition); seaborne traffic of 1492 | verified 2026-07 |
| Fine-silver content of the mark lübisch: 50–55 g (late 14th c.) → 46 (1401) → 40 (1411) → 28 (1424) → 25 (1433) → 20 (1461) | [E] | Edvinsson, Franzén & Söderberg, in Historical Monetary and Financial Statistics for Sweden, Vol. I ch. 3 (Riksbank) | verified 2026-07 |
| Competing silver series: 61.53 g (1365–74) → 51.81 (1375–78) → 47.24 (1400) → 40.47 (1411–30) → 20.2–20.3 (1461–1509). Disagrees with the row above chiefly in the base year, which alone moves the silver-terms fall from −57% to −60% | [E] | Volckart, "Prices in Mark of Lübeck (14th–16th c.)", IISH Historical Prices & Wages. Both series are shown on the conversion strip rather than one being chosen | verified 2026-07 |
| Rye prices in marks lübisch per wispel: Hamburg/Lüneburg mean 3.412 (1365–87) → 7.163 (1480–99), or 5.660 excluding the 1491–92 dearth; Lübeck's own 1366 reading = 5 and 6 shillings per Scheffel | [M] | Volckart (as above), from the Hamburg Kämmereirechnungen, Lüneburg city archive and published Lübeck editions; the 1366 Lübeck price independently matches Abel, Strukturen und Krisen der spätmittelalterlichen Wirtschaft (1980), p. 69 | verified 2026-07 |
| Rye-equivalent trade: ≈160,000 wispel (1368–69) → ≈92,000 (dearth years counted) or ≈117,000 (averaged out); change −42.4% or −27.1% | [D] | Derived on this page: nominal toll ratio 1.209 ÷ rye-price ratio (2.10 / 1.66). Rye stands in for a toll basket that was mostly not grain, so this measures purchasing power, not tonnage | derived |
| Independent corroboration: −41.8% / −28.6% under the same two conventions | [D] | Van der Wee's Antwerp rye series (1366–1603) via the Global Price and Income History Group, bridged into marks through the mark's silver content, a separate archive and currency reaching the same answers within one percentage point | derived |
| Grain priced in silver fell ≈25% from the 1370s–80s to the 1480s–90s across northern Europe (≈45% at the mid-century trough) | [M] | Measured across three independent series (Hamburg/Lüneburg, Antwerp, Munich); the monetary context is the fifteenth-century bullion famine, Day, Past & Present 79 (1978); Spufford, Money and its Use in Medieval Europe (1988) ch. 15 | verified 2026-07 |
| Physical units across the same two toll books: ships ≈1,800 (1368) → ≈760 (1492–96); Scania herring ≈70,000 → 14,477 barrels p.a.; salt ≈7,600 t → 9,000–10,000 t | [M] | Hammel-Kiesow (1993), in Jenks & North (eds.), Der hansische Sonderweg?, 77–93 | verified 2026-07 (print-only source) |
| Comparability limits on the two toll books: the 1492–96 books miss North Sea trade; 1368 falls under a Norway embargo; toll compliance decayed within the later window (income 3,560 marks in 1492 → 1,856 by 1496) | [M] | Burkhardt, "Business as Usual?" (Brill, 2013), 215–38 | verified 2026-07 |
| Staatsmark coin standard, early 1500s: ≈18 g fine silver | [E] | Numismatic literature (coin from 1502; "Staatsmark" 1506–1530). Bridged onto the accounting series as a trajectory, not one homogeneous series | verified 2026-07 |
| Silver-equivalent totals: 27–34 t (1368–69); 11.9–13.4 t (1492); change −57% to −60% | [D] | Derived on this page from the rows above; the range spans the two competing silver series rather than picking one | derived |
| Danzig rye exports, 1490s: ≈10,000 lasts/yr; "quintupled over the sixteenth century" | [E] | Brand, in Baltic Connections (Brill, 2007); firmness of the 1490s baseline qualified by Link (2014) | verified 2026-07 |
| Danzig grain exports, 1618: 118,000 Gdańsk lasts ≈ 271,000 t (łaszt ≈2.2–2.3 t; Sound-toll convention ≈2 t/last) | [E] | Gdańsk Grain Exchange historical summary; Polish scholarship reports 115,000+ | verified 2026-07 |
| Dutch Baltic grain imports, 1567: ≈60,000 lasts (context series on the Danzig chart) | [E] | Brand (2007). The "80,000 lasts" sometimes quoted is Holland's fleet loading capacity, not exports, the misattribution this page's research initially carried and removed on verification | verified 2026-07 |
| Danzig ≈75% of Baltic grain exports (16th c.) | [E] | Brand (2007) | verified 2026-07 |
| Link's Danzig harbor-book sample years: 1409, 1460, 1471, 1475, 1490–92; no smooth growth curve supported | [M] | Christina Link, Der preußische Getreidehandel im 15. Jahrhundert (2014) | verified 2026-07 |
| Danzig war of 1576–80: costs ≈1,296,700 marks; war revenue 582,058; debt 714,642 | [M] | Foltz, Geschichte des Danziger Stadthaushalts (1912), p. 77 | verified 2026-07 |
| Norwegian stockfish to England, 1303–11: 1,500–2,000 t/yr; equal-continental-demand assumption ⇒ total 3,000–4,000 t/yr; later-14th-c. fall to half or less; late-16th-c. total >4,500 t/yr (1597–99) | [E] | Helle (2019), resting on Nedkvitne, The German Hansa and Bergen 1100–1600 (2014) | verified 2026-07 |
| Hanse cloth exports from London: ≈44,300 cloths (1550) → ≈13,800 (1552), after the privilege revocation | [M] | Brenner, Merchants and Revolution (1993), p. 7 | verified 2026-07 |
| Sound registers: begin 1497; near-complete from 1574; destinations consistently from 1668 | [M] | Sound Toll Registers Online project documentation; Scheltjens et al., STRO 2.0 | verified 2026-07 |
| Chronicle dates: Stralsund 1370; Novgorod Peterhof closed 1494; last diet 1669 | [CR] | Standard chronology per Dollinger, The German Hansa | uncontested |
Silver content is not purchasing power, which is why the page carries the calculation through a second time into rye. What no deflator can fix is the coverage difference between the two toll books; that limit is stated in stage 1 and is the reason the physical counts (hulls, barrels) carry as much weight here as the values do.