The Welfare Cost of Autarky: Evidence from the
Jeffersonian Trade Embargo, 1807–09
Douglas A. Irwin*
Abstract
The United States came close to complete autarky in 1808 as a result of a self-imposed embargo on international
shipping from December 1807 to March 1809. Monthly prices of exported and imported goods
reveal the embargo’s striking effect on commodity markets and allow a calculation of its welfare effects. A
simple general-equilibrium calculation suggests that the embargo cost about 5 percent of America’s 1807
GNP, at a time when the trade share was about 13 percent (domestic exports and shipping earnings). The
welfare cost was lower than the trade share because the embargo did not completely eliminate trade and
because domestic producers successfully shifted production toward previously imported manufactured
goods.
1. Introduction
In theory, the gains from international trade are represented by comparing welfare at
the free-trade equilibrium with welfare at the autarky equilibrium. In practice, such a
comparison is almost never feasible because the autarky equilibrium is almost never
observed, except in unique cases. One such case is Japan’s dramatic opening to the
world economy in the 1850s, in which domestic and world prices on comparable
commodities exist for the period of Japanese autarky and later free trade.1
Another
exceptional instance is the United States during the period from December 1807 to
March 1809,when Congress imposed a nearly complete embargo on international commerce
at the request of President Thomas Jefferson. Unlike many trade embargos, this
one almost completely eliminated trade and was not compromised by widespread
smuggling.
The Jeffersonian trade embargo provides a rare opportunity—a “natural experiment”
if you will—to observe the effects of a nearly complete (albeit short-lived)
elimination of international trade.The main objective of this paper is to investigate the
impact of the embargo on the US economy, such as its effect on the prices of traded
goods and the degree of import substitution, as well as provide a rough calculation of
its welfare cost. Section 2 provides some historical background on the embargo and
describes its effect on trade. Section 3 sets out the simple general-equilibrium framework
for evaluating the welfare effects of the embargo and describes the data used to
provide an estimate of the overall cost of approaching autarky.The results suggest that
the welfare cost of the embargo was about 5 percent of GNP in 1807, a time when the
ratio of exports to GNP was about 13 percent. Along with the embargo’s immediate
failure to achieve its objective of softening Britain’s policies toward neutral shipping,
the high economic cost of the measure helped persuade Congress to repeal it in March
1809, just 14 months after its imposition. Section 4 concludes the paper.
Review of International Economics, 13(4), 631–645, 2005
© Blackwell Publishing Ltd 2005, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA
*Irwin: Department of Economics, Dartmouth College, Hanover, NH 03755, USA. E-mail:
douglas.irwin@dartmouth.edu. I wish to thank Earl Grinols, Arvind Panagariya, and seminar participants at
Clark, Princeton, UCLA, and the NBER Summer Institute for helpful comments. The National Science
Foundation provided partial research support.
2. Historical Background
The first decade of the nineteenth century was marred by near constant military conflict
between Britain and France in the Napoleonic Wars. As part of the war, Britain
sought to blockade the continent and to enforce this blockade by stopping the shipping
of neutral countries, especially the United States. The British navy patrolled the
eastern US coast and regularly intercepted American vessels, conducted searches and
seizures, confiscated ships, cargo, and other property, and even impressed sailors (said
to be British subjects) who were evading military service. (France engaged in similar
activities, but less intensely than Britain.) American frustration over these actions
steadily grew over time, but the United States had few alternatives to deal with the situation.
The main options were continued submission to the British actions, declaring
war on Britain in an attempt to force them to respect American shipping, or enacting
an embargo to safeguard American property and impose an economic penalty on
Britain.2
After several provocations by Britain, the Jefferson administration opted for an
embargo, not only to protect domestic vessels and sailors from possible seizure, but to
inflict economic costs on Britain in the hope of forcing them to change their policy. In
December 1807, President Jefferson requested that Congress enact a general embargo,
and Congress quickly prohibited all American ships from sailing to foreign ports and
foreign ships from taking on any cargo in the United States.3
Almost all accounts of
the embargo considered it to have been effective, at least initially. Very few American
vessels appeared in European ports in 1808. Figure 1 presents the tonnage of
American ships entering into British ports during the period from 1801 to 1810. While
the tonnage entering in 1808 is not zero, that volume fell 80 percent from the
previous year.
According to Heaton (1941, p. 189), there was “little effort to indulge in large-scale
defiance of the Embargo, at least until the end of 1808, when patience was becoming
exhausted and repeal seemed inevitable.”4
Although the data underlying Figure 1 are
632 Douglas A. Irwin
© Blackwell Publishing Ltd 2005
1800 1802 1804 1806 1808 1810
Tons
0
40,000
80,000
120,000
160,000
200,000
Figure 1. Tonnage of American Ships Entering Great Britain, 1801–10
Source: Great Britain,“An Account Relating to Imports and Exports,” 18 February 1812,
Parliamentary Papers, Vol. 10, 1812, p. 5.
not available on a monthly basis, the American tonnage entering Britain is likely to
have been concentrated late in 1808. Late in that year, frustration with the embargo
was high, its termination was widely anticipated, and merchants were increasingly
tempted to ignore it. After carefully comparing US and British price data and considering
other quantity measures of direct and indirect trade between the two countries,
Frankel (1982, p. 308) concludes that, although smuggling existed, it was small: “the
Embargo was well-enforced, so each country [the United States and Great Britain]
was reduced practically to autarky vis-a-vis the other.”There are many reports of smuggling
activity across Lake Champlain into Canada,but it is not believed that the volume
of such traffic was very significant.
The highly controversial embargo was in effect for just 14 months. Growing domestic
opposition to the trade restrictions, particularly in New England, forced Congress
to repeal the measure in March 1809.5
The consensus among historians is that the
embargo failed to achieve its objective because Britain and France refused to change
their policies regarding American shipping. This was not due to the failure to eliminate
trade, but the failure of the trade measures to weaken the political resolve in
Britain to suppress neutral shipping in its effort to strangle the French economy.6
Although it did not completely eliminate US trade, the embargo succeeded in having
a dramatic effect on commerce. Figure 2 shows the value of total US merchandise trade
(merchandise exports of domestic produce plus re-exports, and imports) from 1800 to
1815. Table 1 presents more detailed trade data for the years around 1807. The value
of merchandise exports of domestic produce fell from $49 million in 1807 to $9 million
in 1808, a decline of over 80 percent. Net shipping earnings fell considerably less, by
45 percent, such that total export earnings dropped 64 percent. Meanwhile, the value
of imports for domestic consumption fell from $85 million in 1807 to $45 million in
1808,a decline of nearly 50 percent.These figures understate the impact of the embargo,
THE WELFARE COST OF AUTARKY 633
© Blackwell Publishing Ltd 2005
0
20
40
60
80
100
120
140
160
1800 1802 1804 1806 1808 1810 1812 1814
Exports Imports
Millionsofdollars
Figure 2. US Merchandise Trade, 1800–15
Source: North (1960, p. 600).
however, because the government’s trade statistics were collected for the fiscal year,
not for the calendar year. Therefore, the figures for 1808 refer to the period from
1 October 1807 to 30 September 1808, and thus include three months in which the
embargo was not in effect.
Exports fell more sharply than imports in 1808 because of the timing of the embargo
and its effect on trans-Atlantic shipping (Heaton, 1941, p. 190). After December 1807,
no American ship was allowed to leave the United States for a foreign destination.
However, many American vessels were then in British ports, where they would typically
spend the winter months. To encourage these ships to return home, Congress
allowed merchants to bring into the country previously purchased foreign goods.Thus,
American ships returning from Europe in the spring and summer of 1808 were permitted
to unload their cargoes, but were not then permitted to return to Europe. (Most
ships involved in the trans-Atlantic trade made two round trips during the sailing
season.) As a result, exports fell immediately upon the imposition of the embargo
(because there could be no spring exports) and remained low throughout the year. By
late 1808 and early 1809, as disregard for the embargo increased, some ships left US
ports in violation of the law. On the other hand, imports did not fall as much because
of the spring and summer ship arrivals from Europe. But after that, few ships with
foreign cargo arrived in the United States.
The embargo had a dramatic impact on prices in the United States, driving down
the domestic prices of exported goods and driving up the domestic prices of imported
goods. Figure 3 presents the domestic prices of the four leading US commodity exports
on a monthly basis from 1807 through 1809. Together, these four commodities—raw
cotton, flour, tobacco, and rice—accounted for about two-thirds of US exports of
domestic produce in 1807. Their prices fell sharply in early 1808, consistent with the
previous discussion of how the embargo affected exports immediately by prohibiting
American ships from departing for foreign ports. The export-weighted average of the
domestic prices of these commodities fell by about 27 percent between December 1807
and June 1808.7
Prices bottomed out in the summer of 1808 and even began to recover
toward the end of the year. This is consistent with the evidence that merchants were
willing to violate the embargo as time went on.
Unlike exports, US imports were highly diversified and were not concentrated in
just a few key commodities. Figure 4 presents two price indices for imported commodities
in Boston and Philadelphia during this period. Both indices show that prices
634 Douglas A. Irwin
© Blackwell Publishing Ltd 2005
Table 1. Selected US International Transactions, 1807–09 (figures in millions of dollars)
Exports
Imports
Paying Free or Total
Net ad paying (for
Domestic shipping valorem specific Re- domestic Customs
Year merchandise earnings Total duties duties exported consumption) receipts
1807 48.7 42.1 90.8 57.8 86.9 59.6 85.1 17.0
1808 9.4 23.0 32.4 21.2 36.9 13.0 45.1 7.7
1809 31.4 26.2 57.6 28.5 32.5 20.8 40.2 7.4
Source: North (1960, p. 600); Pitkin (1816, p. 36, 252); Irwin (2003, Table 1).
THE WELFARE COST OF AUTARKY 635
© Blackwell Publishing Ltd 2005
Embargo p eriod
1807
Cotton
Flour
Rice
Tobacco
1808 1809
40
50
60
70
80
90
100
110
120
1807:01=100
Figure 3. Domestic Wholesale Prices of Exported Commodities, Monthly 1807–09
Source: Cole (1938). Prices are of Georgia upland cotton at New York, superfine flour
at New York, tobacco at New York, and rice at Philadelphia.
E mbargo p eriod
140
130
120
110
100
90
1807 1808 1809
1807=100
Boston Philadelphia
Figure 4. Domestic Wholesale Prices of Imported Commodities, Monthly 1807–09
Sources: Smith and Cole (1935, p. 147), and Bezanson et al. (1936, p. 353). The Boston
index is a weighted average of prices on 18 imported commodities. The Philadelphia
index is a weighted average of prices of 59 imported commodities.
of imported goods failed to rise in the spring of 1808, probably because imports continued
to arrive in the port cities. Prices began to shoot up by the fall of 1808, as the
number of ships entering US ports fell to a trickle and imports became increasingly
scarce. The price of imported commodities rose by about a third as a result of the
embargo: the index of wholesale prices rose 33.9 percent in Boston and 31.1 percent
in Philadelphia between December 1807 and March 1809.8
Prices peaked around the
time the embargo was lifted in early 1809, and fell rapidly once trade was resumed.
In terms of its effect on import and export prices, therefore, the embargo was roughly
equivalent to at least a 60 percent import tariff, perhaps larger if the United States was
not a small country and could affect world prices. Although it did not reach the point
of complete autarky as a result of the embargo, the United States was by mid- to late-
1808 about as close to being fully shut off from international commerce as it has ever
been during peacetime.9
How much of the gains from trade were sacrificed in this
policy action?
3. Assessing the Gains from Trade
This section sets out the simple general-equilibrium framework that will be used to
calculate the welfare cost of the Jeffersonian trade embargo. The gains from trade can
be represented by the difference in the utility level associated with the free-trade consumption
bundle and the utility level associated with the autarky production bundle.
Following Grinols and Wong (1991), the compensating variation change in welfare
from an initial trading situation (subscripted 0) to another (subscripted 1) can be
expressed using the expenditure function:
(1)
By the definition of the expenditure function,
(2)
where x is the vector of consumption, y is the vector of production, m is the vector of
trade (mi ∫ xi - yi, such that imports are positive and exports are negative), p is the
vector of domestic prices, p* is the vector of world prices, B is any net transfers or borrowing
from foreigners, and “◊” denotes the inner product. Equation (2) states that the
value of consumption is equal to the value of production plus tariff revenue—represented
by (p1 - p1*)◊m1—plus transfers. As Grinols and Wong show, the change in
welfare can be rewritten as:
(3)
where the terms sC and sP represent the substitution effects in consumption and production
as a result of the change in relative prices. Specifically, sC = p1 ◊x0 - e(p1, u0)
and sP = p1 ◊(y1 - y0), and these substitution effects are non-negative (sC ≥ 0, sP ≥ 0) if
price-taking consumers optimize subject to a budget constraint and price-taking producers
maximize profits. Equation (3) can also be rewritten in terms of domestic prices
as:
(4)
In the case where there are no international transfers and where the embargo is
completely effective (such that m1 is zero and the country is in autarky), this expression
simplifies to three terms: p1 ◊(-m0), sC, and sP. Figure 5 illustrates these three
components graphically. The free-trade situation consists of production at the point
DW p p m p m B s sC P= -( )◊ + ◊ -( )+ + +1 1 1 1 0 1* .
DW p p m m p p m B B s sC P= -( )◊ -( ) + -( )◊ -( )+ -( )+ +1 1 1 0 1 0 0 1 0* * * ,
e p u p x p y p p m B1 1 1 1 1 1 1 1 1 1, * ,( ) ∫ ◊ = ◊ + -( )◊ +
DW e p u e p u= ( )- ( )1 1 1 0, , .
636 Douglas A. Irwin
© Blackwell Publishing Ltd 2005
where the initial relative price line p0 is tangent to the production possibility frontier,
and consumption at the point where p0 is tangent to the indifference curve U0. If all
trade is eliminated, domestic prices become p1, where the utility level U1 is tangent to
the production possibility frontier. The compensating variation welfare cost of eliminating
trade and moving to autarky is represented by the distance between the
indifference curves U0 - U1, evaluated at domestic prices p1.
Thus, the change in welfare is equal to the distance BC, i.e. DW = BC. This can be
decomposed as BC = AD - AB - CD, where AD represents the value p1 ◊m0, AB is
p1 ◊x0 - e(p1, u0) = sC, and CD is p1 ◊(y1 - y0) = sP. In other words, AD is the value of the
previous volume of trade at autarky prices,AB is the substitution in consumption, and
CD is the substitution in production as a result of the change in relative prices. AD
represents p1 ◊m0 because DE is the value of previous exports at autarky prices and
AE is the value of previous imports at autarky prices, so that AE - DE = AD (since
by definition imports are positive and exports are negative).
Equation (4) is a general formulation for a change in welfare and does not assume
that all trade is eliminated. In principle, each component of the equation can be
calculated. In the present case, the subscript 0 will refer to the pre-embargo (1807)
situation and the subscript 1 will refer to the embargo (1808) period. Although data
on US trade in the early nineteenth century are scanty, let alone data on production
and consumption, sufficient data exist to generate an approximate calculation of most
components of equation (4). The calculation of each component deserves some
discussion.
THE WELFARE COST OF AUTARKY 637
© Blackwell Publishing Ltd 2005
P1
P0
A
B
C
D
Agriculture
Manufactures
U0
U1
0
E
Figure 5. The Gains from Trade
AD: -p1 ◊m0
This is the most readily calculated component of the welfare change. We observe
p0 ◊m0, which is simply the value of trade in 1807. The initial price vector is normalized
to unity, such that p0 = (1, 1, 1, 1) and m0 = (-48.7, -42.1, 34.0, 51.1), where the vector
elements are merchandise exports, shipping earnings, imports for domestic consumption
paying ad valorem duties,and imports for consumption paying specific duties.(This
distinction among imports will be important for reasons that will become clear in the
next calculation.) Re-exports have been excluded from the calculation because
the gains that accrued to the United States from such trade are reflected solely in
shipping earnings.
In the vector m0, the values for domestic exports and shipping earnings are taken
straight from Table 1. This table also presents the value of imports for consumption,
but government statistics only report the breakdown between ad valorem and specific
duties for total imports, not imports for consumption. The difference between total
imports and imports for domestic consumption is goods that were re-exported. The
working assumption will be that imports for consumption are divided in the same proportion
as total imports between those subject to specific duties (40 percent) and ad
valorem duties (60 percent). It might be supposed that imports subsequently reexported
consisted mainly of commodities from the West Indies (such as sugar, coffee,
etc.) that were subject to specific duties (and then rebated upon re-exportation), but
Pitkin (1816) indicates that re-exports also consisted of manufactures from Britain.10
As previously noted, the domestic price of imported commodities rose about 33
percent as a result of the embargo, and the domestic price of exported goods fell 27
percent. There is little data available on domestic shipping rates between 1807 and
1808, although (as Table 1 indicates) shipping receipts fell less than domestic exports
between 1807 and 1808.11
Therefore, in the absence of better information, the maintained
assumption will be that the price of shipping services fell the same 27 percent
as domestic exports.
Thus, the embargo price vector is p1 = (0.73, 0.73, 1.33, 1.33). The calculated value
of -p1 ◊m0, which is also the estimated distance AD, is therefore $47 million.
CD: sP = p1 ◊(y1 - y0)
This component is somewhat more difficult to estimate. This measures the shift along
the domestic production possibilities frontier—producing more of the goods previously
imported and producing less of the goods previously exported—as a result of the
change in domestic relative prices.
The scope for producers in the United States to start or to increase domestic production
of previously imported goods (import substitution) depends upon the type of
imported good. The ability of domestic firms to commence or increase production
of manufactured goods, such as cotton goods, glass, and iron, as a substitute for imports
was quite substantial. Conversely, prior to the embargo, nearly half of imports consisted
of consumption items such as sugar, coffee, tea, spices, wine and spirits, goods in
which East Coast producers could not easily initiate production.
Unfortunately, unlike the export statistics, the import statistics of this period do not
completely enumerate the types of goods imported, but they can help establish the
general range of potential import substitution by domestic producers. Imports subject
to specific duties are individually specified and consisted almost exclusively of the
consumption goods mentioned above (sugar, coffee, tea, wine, etc.) that domestic
638 Douglas A. Irwin
© Blackwell Publishing Ltd 2005
producers would have difficulty replicating.All other imports paying ad valorem duties
were not separately recorded, but included (among other items) manufactured goods,
where the possibility of import substitution was presumably high.As previously noted,
about 40 percent of imports were subject to ad valorem duties.
How much import substitution took place? Not much was possible, given the short
duration of the embargo; over a longer horizon, domestic producers would have been
able to establish much higher levels of output of previously imported goods.That said,
the stimulus of the embargo to domestic manufacturers (particularly around Philadelphia)
was widely and favorably noted at the time. The city of Philadelphia apparently
did well during the embargo, not just because income lost by merchants was shifted
partly to lawyers (who benefited from the proliferation of maritime-related law suits)
but because manufacturing production expanded in such goods as carpets, cloth (from
bedspreads to stockings), earthenware, glass, soap, lead and shot, and chemicals (Sears,
1921). Frankel (1982, p. 301) argues that “There is historical evidence that Americans
were remarkably successful in 1808 at switching into the production of manufactured
goods when they were cut off from their usual source of supply.”
Other evidence suggests that domestic manufacturers elsewhere did not necessarily
prosper during the short period in which the embargo was in effect. Perhaps the most
direct piece of evidence is from the Davis (2004) series on US industrial production
during this period. Davis’s overall industrial production index actually collapsed in
1808, falling 17 percent from 1807. However, the collapse was entirely due to the fall
in production by commercial and trade-dependent industries (such as merchant shipbuilding,
flour refining, fish curing, etc.) with virtually no change in the output of
domestic infant industries competing against imports (such as cotton and wool cloth
production, etc.).12
Although this suggests that there was little immediate import substitution
by domestic manufacturers, the index fails to include output from small
producers, such as local blacksmiths and glass blowers, household textile manufacturers,
etc., where an expansion in production might be found.
Thus, information on production substitution is incomplete and the existing evidence
is mixed. In terms of calculating the value of production substitution, one extreme
assumption is that the United States could replace one-for-one the quantity of the previously
imported manufactured goods (i.e. imports subject to ad valorem duties) with
domestic production of those goods at the higher autarky prices. We also assume
that domestic producers would not be able to produce any imported goods subject to
specific duties,including such commodities as sugar,coffee,tea,wine,etc.Symmetrically,
domestic production of the exportable goods and shipping services would be assumed
to fall one-for-one with the decline in the exports of those items. In this one-for-one
replacement case,then (y1 - y0) = (m1 - m0) with the non-ad valorem imports zeroed out.
Because m1 = (-9.4/0.73,-23.0/0.73,16.0/1.33,29.1/1.33),or (-12.9,-31.5,12.0,21.9),then
m1 - m0 = (35.8,10.6,-22.0,-29.2) and therefore (y1 - y0) = (35.8,10.6,-22.0,0).Then the
net gain from the production shift is p1 ◊(m1 - m0) = $4.7 million.
This one-for-one replacement assumption is extreme for assuming that domestic
production of manufactures could substitute perfectly for imports, and that domestic
production of exportables would fall as much as exports declined. An alternative
assumption is that domestic producers could only produce half of the quantity of goods
previously imported, and were forced to reduce by half their production of goods previously
exported. In this case, (y1 - y0) = 0.5◊(m1 - m0) = (17.9, 0, 5.3, -20.2, 0), and the
net gain from the production shift is p1 ◊(m1 - m0) = $2.3 million. This case will be
the central assumption, although the bounds on the welfare cost will be explored using
both full replacement and zero (substitution) replacement.
THE WELFARE COST OF AUTARKY 639
© Blackwell Publishing Ltd 2005
AB: sC = p1 ◊x0 - e(p1, u0)
The absence of any consumption data makes any estimate of the consumption substitution
virtually impossible.The scope for consumption substitution was probably small:
it is unlikely that domestic consumption of flour and tobacco would increase significantly
because their prices had fallen—and the prices of manufactures increased—due
to the embargo. It seems plausible to assume that the scope for substitution in consumption
is less than the scope for substitution in production, which means that it
would take a value of something less than $5 million. As an initial guess, we will take
the value of $2.3 million, the same value as the substitution assumed to be achieved
by domestic producers.
Alternatively, we can crudely simulate the value of consumption substitution with a
simple Cobb–Douglas utility function in which consumption expenditures are assumed
to be 80 percent of GNP and the expenditure share on imported manufactures is 30
percent. Under Cobb–Douglas, the elasticity of substitution is one, a higher value than
is commonly used in computable general-equilibrium models. This gives us an upperbound
estimate on consumption substitution. In this case, the change in relative prices
brings about consumption substitution valued at $3.2 million.
Other Components
The last two components of equation (4) are tariff revenue and net transfers in period
1. Tariff revenue—expressed as (p1 - p1*)◊m1—is not zero because some imports were
allowed into the country in early 1808. The government reported tariff revenue to be
$7.7 million in 1808, as noted on Table 1.13
Net transfers (B1) is derived from the constraint
in equation (2): p1(y1 - x1) + (p1 - p1*)◊m1 + B1 = 0. Solving for B1 in the year
1808 yields the figure of $1.5 million.This helps correct for trade imbalances that might
affect the previous calculations.
Table 2 presents the combined estimates of these various components of the welfare
cost. Column (1) presents the benchmark case using the figures calculated above. The
components sum to $33 million. This welfare cost may be more usefully expressed as
640 Douglas A. Irwin
© Blackwell Publishing Ltd 2005
Table 2. Welfare Effects of the 1808 Trade Embargo (figures in millions of dollars; negative
numbers in parentheses)
Central estimate Low estimate High estimate
Component (1) (2) (3)
p1 ◊m0 (46.9) (46.9) (46.9)
(p1 - p1*)◊m1 7.7 7.7 7.7
B1 1.5 1.5 1.5
sP = p1 ◊(y1 - y0) 2.3 4.7 0
sC = p1 ◊x0 - e(p1, u0) 2.3 4.7 0
Total (33.1) (28.3) (37.7)
As percent of GNP 4.9% 4.2% 5.5%
Source: See text for calculations. Nominal GNP in 1807 is assumed to be $680 million, from Gallman (2000,
p. 7). The share of domestic exports and shipping earnings in GNP was 13.4 percent in 1807; the share of
imports for domestic consumption in GNP was 12.5 percent in that year.
a percent of GNP. Gallman (2000, p. 7) reports a figure of $680 million for nominal
GNP in 1807, which is somewhat higher than other estimates but otherwise seems reasonable.
Taking this figure as a rough benchmark, the welfare cost of the embargo was
about 5 percent of GNP.
What is a plausible range for this welfare cost? The first component (p1m0) is a best
guess and does not change; this also applies to tariff revenue and to B1. The production
and substitution components are subject to greater uncertainty. Columns (2)
and (3) present lower- and upper-bound calculations based on assumptions about
substitution. Column (2) presents a lower-bound calculation in which domestic manufacturers
can fully produce what was previously imported and domestic farmers fully
reduce production by the amount that exports decline. This increases sP from $2.3 to
$4.7 million, decreasing the welfare cost to $28 million, or 4 percent of GNP. Column
(3) shows alternative, higher-bound calculations on the magnitude of the welfare loss
in which there is no production or consumption substitution at all. This increases the
welfare cost to $38 million, or 5.5 percent of GNP.
These alternative calculations provide some plausible bounds on the likely true
welfare cost of the embargo.The welfare cost appears to be centered around 5 percent
of GNP, but could be as low as 4 percent or as high as something closer to 6 percent.
In 1807, merchandise exports and net shipping receipts were about 13 percent of GNP.
This finding is remarkably close to Bernhofen and Brown’s (2005) result for Japan’s
emergence from self-imposed autarky after 1858.They place the static gains from trade
for Japan at no more than 8 percent of GDP.
In two classic papers on the cost of protection, Johnson (1960, 1965) presents some
different calculations that lend some support to these magnitudes. Johnson (1965) used
a constant elasticity of substitution utility function and a general specification for the
production transformation curve to arrive at the costs of self-sufficiency (autarky) in
a computable two-sector general-equilibrium model. In the case where the share of
national income spent on importables under free trade is 0.20 and the elasticity of substitution
in consumption between importables and exportables is 0.5, for example, then
a tariff of 76 percent would be sufficient to induce autarky at a welfare loss of 5.2
percent of free-trade utility level. If the elasticity of substitution is 1.0, then the selfsufficiency
tariff is 53 percent and the welfare loss is 4.0 percent. These figures are
remarkably close to the actual findings in this paper. Given that the calculation is performed
in a similar general-equilibrium setting, perhaps this should not be too
surprising.
Johnson (1960) also derived the commonly used partial-equilibrium measure of the
cost of protection as a percent of GDP as ah(t/(1 + t))2
, where a is the ratio of imports
to GDP in the (initial) equilibrium, h is the absolute value of the arc elasticity of import
demand, and t is the average ad valorem tariff rate. In the present case, the ratio of
imports to GDP is approximately 0.13, and if we assume that the absolute value of the
elasticity of import demand is 2 and that the tariff rate is 70 percent, then the cost of
protection is 2.2 percent of GDP. As is commonly pointed out, this partial-equilibrium
formula yields relatively small figures for the cost of protection.
Thus, the massive disruption of trade caused by the embargo imposed a substantial
one-year loss on the US economy, one that was not just a small fraction of a percent
of GNP.There is no doubt that 1808 was a difficult year for the US economy. One estimate
of real GNP showed a fall of 5.8 percent between 1807 and 1808 (Johnston and
Williamson, 2003). The year has been called a “depression” by Thorp (1926, p. 116),
who gives it the following description: “rigid embargo causes paralysis on coast,
1
2
THE WELFARE COST OF AUTARKY 641
© Blackwell Publishing Ltd 2005
gradually spreading inland; severe distress in New England; further sharp decline in
commodity prices to low point, third quarter; foreign trade completely checked.”
It is difficult to consider the policy of embargo a success, given the cost it imposed
on the economy while failing to achieve its objective of changing British and French
policy toward American shipping. One person who almost perfectly anticipated the
results of the embargo was Alexander Hamilton, the first Secretary of the Treasury.
Writing in 1794, Hamilton was prescient about the potential impact of any trade
embargo:
“The consequences of so great and so sudden a disturbance of our Trade
which must affect our exports as well as our Imports are not to be calculated.An
excessive rise in the price of foreign commodities—a proportional
decrease of price and demand for our own commodities—the derangement
of our revenue and credit—these circumstances united may occasion the
most dangerous dissatisfaction & disorders in the community and may drive
the governt. to a disgraceful retreat—independent of foreign causes.”
(Syrett, 1961, Vol. 16, p. 275)
Hamilton proved correct not just about the trade and revenue effect of the embargo,
but about the domestic disorder it caused and the eventual retreat of the government:
New England grumbled about secession and the growing domestic discontent with the
642 Douglas A. Irwin
© Blackwell Publishing Ltd 2005
Table 3. Voting in the House of Representatives on the Embargo
Republicans Federalists Total
Region Yea Nay Yea Nay Yea Nay
A. The imposition of the embargo, 21 December 1807
New England 19 0 0 15 19 15
Mid-Atlantic 27 7 0 9 27 16
South 28 10 0 2 28 12
West 8 2 0 0 8 2
Total 82 19 0 26 82 45
B. The retention of the embargo, 24 February 1809
New England 1 13 0 15 1 28
Mid-Atlantic 10 27 0 6 10 33
South 18 16 0 2 18 18
West 4 3 0 0 4 3
Total 33 59 0 23 33 82
Note: This vote is on an amendment to eliminate all passages from a bill that would repeal the embargo. A
vote in favor is thus a vote to retain the embargo. The final bill that actually repealed the embargo replaced
it with non-intercourse restrictions on trade and was passed in a partisan vote, with Federalists opposing the
continuation of any restrictions on trade. Source: Voteview, from www.princeton.edu/~voteview/.
Classification: New England includes Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, and
Connecticut. Mid-Atlantic includes New York, Pennsylvania, New Jersey, Delaware, and Maryland. South
includes Virginia, North Carolina, South Carolina, and Georgia. West includes Ohio, Kentucky, and
Tennessee.
embargo, and the increasing willingness of merchants to violate it, led to its repeal
without it having accomplished any change in British policy. The political opposition
generated by the embargo can be illustrated by studying Congressional votes on the
embargo. Table 3 shows the pattern of voting across regions on the imposition and
retention of the embargo in the House of Representatives. The partisan division on
the issue is readily apparent: Federalists completely opposed the measure, while the
Jeffersonian Republicans initially supported it but later opposed it. The table shows
that Republican support for the embargo was completely lost in New England and
the Mid-Atlantic states. New England merchants were most vocal in opposing the
embargo, while in the Mid-Atlantic the losses to merchants and the farming community
(in terms of forgone income and lower land prices) surely exceeded the gains to
those starting manufacturing establishments.
Given that the major export crops came from the South, one might have expected
strong opposition to the embargo in that region. Yet in the February 1809 vote,
members from the South were evenly divided about whether to retain the embargo or
not. Jennings (1921) notes several factors that may have muted Southern opposition
to the embargo: the suspension of bankruptcy laws, the non-perishability of Southern
crops, and the deep-seated hatred of Great Britain. Despite these factors, there was
little support in Congress for continuing the embargo in early 1809.
4. Conclusions
This paper has used a major shock to trade—America’s embargo of 1807–09—to calculate
the total static losses from forgoing virtually all international trade. While the
United States did not achieve complete autarky in 1808, it came quite close to economic
isolation. The best-guess calculation of the static welfare cost of the embargo is
about 5 percent of GNP. The cost does not represent the total gains from trade,
however, because the initial trading equilibrium was one of restricted trade. Still, the
embargo inflicted substantial costs on the economy while it was in effect.
References
Bernhofen, Daniel M. and John C. Brown, “An Empirical Assessment of the Comparative
Advantage Gains from Trade: Evidence from Japan,” American Economic Review
95(2005):205–25.
Bezanson, Anne, Robert D. Gray, and Miriam Hussey, Wholesale Prices in Philadelphia,
1784–1861, Philadelphia: University of Pennsylvania Press (1936).
Cole, Arthur H., Wholesale Commodity Prices in the United States, 1790–1861: Statistical
Supplement, Cambridge, MA: Harvard University Press (1938).
Davis, Joseph H., “A Quantity-based Annual Index of U.S. Industrial Production, 1790–1915,”
Quarterly Journal of Economics 119 (November 2004):1177–215.
Davis, Joseph H. and Douglas A. Irwin, “Trade Disruptions and America’s Early Industrialization,”
unpublished working paper, Dartmouth College (June 2003).
Frankel, Jeffrey A., “The 1807–1809 Embargo Against Great Britain,” Journal of Economic
History 42 (June 1982):291–308.
Gallman, Robert E., “Economic Growth and Structural Change in the Long Nineteenth
Century,” in Stanley L. Engerman and Robert E. Gallman (eds), The Cambridge Economic
History of the United States. Vol. II: The Long Nineteenth Century. New York: Cambridge
University Press (2000).
THE WELFARE COST OF AUTARKY 643
© Blackwell Publishing Ltd 2005
Grinols, Earl L. and Kar-yiu Wong, “An Exact Measure of Welfare Change,” Canadian Journal
of Economics 24 (May 1991):428–49.
Heaton, Herbert, “Non-importation, 1806–1812,” Journal of Economic History 1 (November
1941):178–98.
Hickey, Donald R., “American Trade Restrictions during the War of 1812,” Journal of American
History 68 (December 1981):517–38.
Irwin, Douglas A., “New Estimates of the Average Tariff of the United States, 1790–1820,”
Journal of Economic History 63 (June 2003):506–13.
Jennings,Walter W., The American Embargo, 1807–1809, University of Iowa Studies in the Social
Sciences, Iowa City: University of Iowa (1921).
Johnson, Harry G., “The Cost of Protection and the Scientific Tariff,” Journal of Political
Economy 68 (August 1960):327–45.
———, “The Cost of Protection and Self-sufficiency,” Quarterly Journal of Economics 72
(August 1965):356–72.
Johnston, Louis and Samuel H.Williamson,“The Annual Real and Nominal GDP for the United
States, 1789–2002,” Economic History Services (March 2003). URL: http://www.eh.net/
hmit/gdp/.
North, Douglass C., “The Balance of Payments of the United States, 1790–1860,” in Trends in
the American Economy in the Nineteenth Century, Studies in Income and Wealth, Vol. 24.
Princeton, NJ: Princeton University Press for the NBER (1960).
Pitkin, Timothy, A Statistical View of the Commerce of the United States of America, Hartford,
CT: C. Hosmer (1816).
Sears, Louis M., “Philadelphia and the Embargo of 1808,” Quarterly Journal of Economics 35
(February 1921):354–9.
Smith, Walter B. and Arthur H. Cole, Fluctuations in American Business, 1790–1860, Cambridge,
MA: Harvard University Press (1935).
Syrett, Harold C. (ed.), The Papers of Alexander Hamilton, New York: Columbia University Press
(1961).
Thorp, Willard, Business Annals, New York: National Bureau of Economic Research (1926).
Notes
1. See Bernhofen and Brown (2005).
2. For background, see Hickey (1981).
3. Although foreign ships were permitted to bring goods to the United States, few did so during
the embargo because they would have had to return empty. In addition, the non-intercourse law
prohibited the importation of many British manufactured goods.
4. “The ships and brigs which carried on the greater part of the transatlantic trade were in
general tied up in obedience to the Embargo, or they went down the coast to load goods
in readiness for the day when the light would change from red to green, and then returned to
anchorage in their home ports . . . So far as the western hemisphere was concerned, smuggling
and embargo-breaking could not be prevented; but against Europe the Embargo was largely
effective. And if American ships did not go to Europe, British goods could not be brought back
in them” (Heaton, 1941, p. 189). Virtually all of US imports from Britain arrived on American
vessels during this era.
5. The embargo was replaced by a non-intercourse Act that continued to prohibit trade with
Britain, France, and their colonies and dependencies. Over the next four years, through the war
of 1812 until the Treaty of Ghent in 1813, trade was subject to varying degrees of formal restriction,
but none as draconian as the earlier embargo.
6. However, a later non-importation measure in 1811 coincided with an economic downturn in
Britain and apparently contributed to the British decision to lift some of its draconian measures
regarding American shipping.
7. The price data are in Cole (1938). The weights are 43 percent for cotton, 33 percent for flour,
17 percent for tobacco, and 7 percent for rice, and come from Pitkin (1816).
644 Douglas A. Irwin
© Blackwell Publishing Ltd 2005
8. In Philadelphia, “The peak of prices for imported commodities occurred in March 1809, the
month the embargo was repealed. At the time the index for imported commodities was
34.4 percent above the low of July 1807” (Bezanson et al., 1936, p. 143).
9. Figure 2 shows that the value of trade was lower in 1814 than in 1808, but this was due as
much to the war with Britain as to any formal policy measure. The calculations below will focus
on the shock between 1807 and 1808, two adjacent years in which the effects of policy can be
most clearly identified.
10. “The valuable articles of colonial produce, such as sugar, coffee, spirits, cocoa, pimento,
indigo, pepper and spices of all kinds, were carried, either directly to Europe, or were first
brought to the United States, and from thence exported in American vessels . . . The manufactures
of Europe, and particularly of Great Britain, as well as the manufactures and produce of
the East-Indies and China, have also been imported, and again exported, in large quantities,
to the West-Indies, to the Spanish Colonies in South America, and elsewhere. This trade, which
has been called the carrying trade, has, in some years, exceeded in value the trade of the United
States, in articles of domestic produce” (Pitkin, 1816, p. 136).
11. International shipping rates rose during the embargo, due to the added risk of undertaking
a voyage. But the relevant data would be the price of domestic shipping services, which presumably
fell during the embargo. Pitkin (1816, p. 391) reports that tonnage engaged in the
domestic coastal trade rose 23 percent in 1808 from the previous year, as ships normally used
for overseas trade were transferred to local trade.
12. This issue is explored much further in Davis and Irwin (2003).
13. Recall as well that the customs figures for 1808 refer to October 1807 to September 1808
and cover a period when there was no embargo and when imports were still allowed in despite
the embargo.The figures for 1809 (October 1808 to September 1809) cover six months when the
embargo was effective.
THE WELFARE COST OF AUTARKY 645
© Blackwell Publishing Ltd 2005