Content uploaded by Stanley L. Engerman
Author content
All content in this area was uploaded by Stanley L. Engerman on Aug 20, 2015
Content may be subject to copyright.
15
IMF Staff Papers
Vol. 52, Special Issue
©
2005 International Monetary Fund
On the Accuracy of Some Past and
Present Forecasts
STANLEY ENGERMAN*
The topic of this paper has been inspired by something of particular interest to
Michael Mussa, something to which he made major and important contribu-
tions while at the IMF. The biannual
World Economic Outlook, published since the
1980s under the auspices of the Research Department, has, as its purpose, the pro-
vision of “analysis and pr
ojections .
. . [as] inte
gral elements of the IMF’s surveil-
lance of economic developments and policies in its member countries, developments
in international financial markets, and the global economic system.” In policy-
oriented institutions (and in most businesses and individual decision making), poli
-
cymaking decisions are often guided by projections and forward-looking indicators.
[JEL
B10, B20]
W
hen forecasts are accurate, rarely does someone comment, “That is what we
should have expected to happen.” However, when forecasts are inaccurate,
they attract attention, usually critical, often in conjunction with an attack not only on
the forecaster but also on the entire profession to which he or she belongs.
To understand the nature of forecasting, and what can go right or wrong, I exam-
ine forecasts in dif
ferent times and places.
This is not a complete survey; forecasts
and forecasting have an exceptionally long history
, so for the sake of brevity
, I will
focus on various forecasts relevant to the history of economics and economic his-
tory, to see what we can learn about this type of exercise.
*Stanley Engerman is John H. Munro Professor of Economics and Professor of History at the University
of Rochester. This paper was presented at the IMF conference Mussa Fest, held in honor of Michael Mussa.
I benefited from the comments of conference participants, particularly Michael Bordo and James Boughton,
as well as Seymour Drescher and David Eltis.
Stanley Engerman
16
Before giving examples I want to briefly describe several different characteris-
tics of forecasts that can go wrong, since presumably, to paraphrase the famous first
line of
Anna Karenina, there are many different reasons forecasts go wrong, but
accurate forecasts are presumably all based on one single feature.
First, would we expect long- or short-run forecasts to be more accurate? Short-
run forecasts require fewer projections and less information, but they may be sub-
ject to random, reversible events. Long-run forecasts can eliminate some sources
of the difficulty found with short-run forecasts, but they do allow more time for
shocks to occur, and, although this does not necessarily contribute to poor fore-
casting, more time is allowed for the undertaking of policies to offset trends and
move things in a different direction.
Why are forecasts sometimes wrong? Are errors the result of endogenous or
exogenous factors? (For a related discussion, see McCloskey, 1992.)
(1) We may use the wrong model or an incomplete model. This can be referred
to as the law of unintended consequences, since even if the main prediction
is accurate, there are often side effects that have been ignored or overlooked
that can alter the evaluation of what has happened. To firm believers in chaos
theory
, however, no forecasting model is acceptable.
(2)
We may fail to allow for shocks that will influence the system. This, of course,
leads to two questions: (i) to what extent can shocks and their impact be fore-
cast? and (ii) can we allow for expected adjustments to these unexpected
changes?
(3) The basic observations used in the forecast may be incorrect (that is, the
underlying data may be wrong or misinterpreted), so that the forecast, while
consistent with the presumed data, is wrong.
(4)
Forecasts made as part of a policy debate may be clouded or incorrect because
of ideological beliefs. Problems arise because in some cases people actually
believe the proposition; in other cases, these propositions reflect mainly the
rhetoric of debate and are used to convince others. An example of this is
the question of the relative productivity of slave versus free labor, a central
issue in the eighteenth and nineteenth century antislavery debate. Adam Smith
was the key figure, on the basis of his reputation, with his ar
gument on dif-
ferent incentives for free and slave labor. His claims featured in later debates
despite the fact that every slave owner knew about and tried to solve the
incentive problem, and the ancient Greeks had worked out clever incentive
schemes for differential rewards to encourage more output from slaves.
When discussing slavery on British and French Caribbean islands, Smith did
allow for the ef
fects of owner behavior on a slave’
s performance (Smith,
1776, Vol. I, pp. 97–99 and 387–89; Vol. II, pp. 587–88; pages refer to 1976
edition). Moreover, in his antislavery argument, Smith did not note, as had
others of the time, that hunger was a superb form of coercion, or as worded
more subtly by his friend David Hume (1752, p. 267 in 1987 edition) that
“necessity . . . is the great spur to industry and invention.” Claiming that end-
ing slavery would raise output and lower goods prices was, of course, the bet-
ter forecast with which to attract antislavery opinion. It must be noted that
nearly every reform advocated is supported by both the belief that it is the
right thing to do and the expectation that it will raise GNP. One of the few
exceptions comes from the early-twentieth-century debates in Australia to
end Pacific Islander indentured servitude used to produce sugar, replacing
it with European workers—but with large subsidies and tariff protection. In
Parliament it was noted that although the policy of keeping out the Pacific
Islanders would raise the price of sugar (a cost estimated equal to about
1
⁄
2
of
1 percent of GNP), white Australians would consider this price hike worth it
(Parliament of the Commonwealth of Australia, 1912, pp. xx–xxi, 553). For
the most part, however, ideology and the economic forecast seem to go in the
same direction.
(5) There is another problem with forecasts that merits more attention than it is
usually given. Most forecasts take the characteristic of comparing equilib-
rium positions: the focus is on the present position and where the system will
be in the future, in the absence of further shocks, with less attention given to
how long it will take to arrive at the equilibrium. Considering the time it will
take to reach equilibrium can be crucial for planning and policymaking. To
demonstrate this point, I will now highlight a few cases; I will discuss others
later
.
Among the major opponents of southern slavery, and one of the more active
politically, was the distinguished Irish economist John Elliot Cairnes. His 1862 book
The Slave Power is still used in discussions among scholars. Cairnes (1862, pp. 274
and 350 in 1969 edition) believed that slavery was doomed and that northern free
labor would prevail, but only “with the progress of time” and “by a gradual but
sure progress.” He quoted the North British Review (February 1862, p. 142), saying
even when slavery’s “doom is sealed . . . the execution of the sentence may seem to
be relegated to a very distant day.”
1
A more precise time frame was considered in the debates about abolishing the
British transatlantic slave trade in the 1790s. It was proposed that ending slavery by
parliamentary legislation would not be necessary
, since once the transatlantic slave
trade was abolished, on grounds of profitability, slavery would quickly end. Using
land-labor ratios for Jamaica, the abolitionist William Wilberforce (1807, pp. 291–92)
and Prime Minister William Pitt (see Coupland, 1923, pp. 270–74) estimated that
if nothing were done by legislative or other means, it would take between 220 and
250 years for slavery to end once the transatlantic slave trade was abolished. This,
clearly, was an unacceptable time frame.
Another interesting example of the importance of looking at the length of time
before the equilibrium point of decline is reached comes from perhaps the most
famous of the historical works on decline,
The History of the Decline and Fall of
the Roman Empir
e
,
by Edward Gibbon (Gibbon, 1776–1788, p. 177 in 1900 edi
-
ON THE ACCURACY OF SOME PAST AND PRESENT FORECASTS
17
1
Cairnes (1862, p. 276) claims that by the time the South would “feel the restraints of its spacious
prison, at least a quarter of a century would have elapsed.” His discussion of “ultimate extinction” points
out that adherents may “in the end be balked of its reward.” Michael Bordo, who has written on Cairnes’s
analysis of gold discoveries on mid-nineteenth-century money supplies has suggested that Cairnes did
better in this study (see Bordo, 1975). Is it a mere coincidence that the prediction regarding the more morally
char
ged issue was more problematic?
Stanley Engerman
18
tion). He notes, in presenting his plan for his last two volumes, that, not as a his-
torical explanation of an event that had occurred (not really a forecast) but as an
explanation that does raise the relevant question, “five centuries of the decline and
fall of the empire have already elapsed; but a period of more than eight hundred
years still separates me from the term of my labors, the taking of Constantinople by
the Turks.” It took 1,300 years, but presumably, someone got it right at some time.
A more recent discussion of a contemporary policy issue also places in per-
spective the importance of time frames for forecasts. Not surprisingly, the movie
The
Day After Tomorrow,
about a five-day cataclysm resulting from global warming,
attracted flak from those skeptical about the importance of global warming. What
was surprising was the strong criticism from those in the environmental movement
concerned with getting the forecast right. As described in Andrew C. Revkin’s
“NASA Curbs Comments on Ice Age Disaster Movie” (
New York Times, April 25,
2004, pp. 1–16), “Some leaders of nonprofit environmental groups are also dis-
tressed about the movie, though for dif
ferent reasons. In conference calls and
e-mail exchanges, they have said it so overstates the issue—turning a decades-long
or century-long threat into one that explodes over five days—that it might cause
people to simply laugh off the real questions.”
2
One can understand the concern about overstatement. Perhaps environmental-
ists were familiar with the problems of one of the major religious movements
in upstate New York in the 1840s (see Doan, 1987, pp. 4, 31–53, 93–112, and
202–206). The Millerites, with between 10,000 and 1,000,000 adherents, used com-
plex mathematical procedures to estimate the date of Jesus’s return to Earth: October
22, 1844. Alas, that date passed with nothing unusual happening, and after the Day
of Great Disappointment the number of active believers plummeted. The movie
The Day After Tomorrow, however, has led to new discussions about global warm-
ing that go beyond stating equilibrium positions to describing the question of when
this event might occur
.
I. Ev
aluating For
ecas
ts
How does one measure and evaluate the success of forecasts? Generally, with a large
enough population of forecasters considering any one issue, it is highly probable that
someone will get it right, but how often is that forecaster right? What is the appro-
priate success rate for a forecaster? In baseball anything above 3 hits in 10 at bats is
regarded as successful; in 3-point attempts in basketball, anything above 4 goals in
10 shots is excellent; and a quarterback who completes more than 5 in 10 forward
passes is considered good. For forecasters, however, we generally give much less
slack and anticipate much better performance. Perhaps we should focus on some
weighted averages of forecasts, weights based on amounts won or lost. Does winning
big only once outweigh numerous losses, or, in baseball parlance, do you want play-
ers who hit for a high average or those who hit many home runs?
2
The movie’s director later pointed to the difficulties in making “events that stretch over decades or
centuries scary enough for a movie” (Andrew C. Revkin, “When Manhattan Freezes Over,”
New York
Times,
May 23, 2004, Arts and Entertainment, pp. 19, 22).
ON THE ACCURACY OF SOME PAST AND PRESENT FORECASTS
19
After major stock market movements, the Business section of the New York
Times
generally features an interview with the guru who predicted it. He (or she)
often explains in depth the analysis that led to these shrewd and insightful predic-
tions. One point to note, however, is that the same guru very seldom appears in the
paper twice.
Discussions of the forecasting skills of the boy who cried wolf, who was incor-
rect all times except the last, raise the question of how many incorrect forecasts lead
us to disregarding the next forecast, an issue familiar to many Malthusians in the
past who explained away repeated failures while still confidently anticipating the
dreaded events. There is always the last refuge of a forecaster: based on Pascal’s
wager about whether to believe in God, an argument based on anticipations of asym-
metric returns, permitting past accuracy to be irrelevant for predicting major events.
Dick Rosett, Michael’s colleague, chairman, and then dean, first at Rochester
and then at Chicago, reminded me of the introduction he used at the annual
University of Chicago School of Business forecasting luncheon. He claimed that
there were two characteristics necessary to be a good forecaster:
1. Be able to accurately and confidently predict the future.
2. Be able to accurately and confidently explain away the past.
This, of course, does not mean to suggest that only economists and economic
historians have a poor track record, if and when they do. Drawing on my knowledge
of sports and entertainment, there are some interesting cases to note—interesting
because they involve forecasts made with what might seem to be considerable infor-
mation. Professional basketball has some infamous cases: Portland choosing Sam
Bowie ahead of Michael Jordan. Several years earlier Portland had made possibly
the league’s worst number-one draft choice by selecting a player with minimal
accomplishments. In Rochester, the choice was made to draft a player named Sihugo
Green, a solid but limited player, over Bill Russell, an all-time great. Perhaps co-
incidentally the Rochester team left town several years later. In football, three of the
greatest quarterbacks of all time—John Unitas, Joe Montana, and Dan Marino—
were all relatively low draft choices. Unitas was the one hundred-second player
chosen and was released quickly by the team that drafted him; Montana was the
eighty-second player and fourth quarterback selected; and while Marino was drafted
in the first round, he was the twenty-seventh player and sixth quarterback selected
that year. Drafting baseball players is much more difficult, since it is usually done
earlier and with less information than in football and basketball, and so the success
rate is not at all high. Before switching subjects, there is another baseball-related
prediction worth mentioning: it involves the argument that baseball faces “universal
bankruptcy” unless player salaries drop, since they are “swelling to unreasonable
and ruinous proportions.” Bud Selig made this forecast in 2004 and in every year
that he has been commissioner of major league baseball, but it was also made by one
of the founders of the National Baseball League, A. G. Spalding, and other owners as
early as 1881, when salaries were at about one two-thousandths of the current level.
3
3
See Seymour (1960, p. 119) and Levine (1985, p. 61). In 1884 the editor of Sporting News claimed
that without the reserve rule “the exorbitant demands” of players would “eventually bankrupt the strongest
company in the professional arena” (Levine, 1985, p. 53).
Stanley Engerman
20
There are many examples of poor forecasting from the world of entertainment.
As a forecaster, would you prefer to be known as the studio executive who signed
Tab Hunter to a long-term contract with the expectation that he would soon become
a great matinee idol, or the studio executive who terminated the contract with Rita
Hayworth, since it was thought that she lacked star quality? How about one of the
twenty-odd publishers who turned down J. K. Rowling’s first Harry Potter book?
Rowling is now apparently the richest woman in Britain.
At some point in our professional lives, most of us are required to forecast the
performance of others. Consider the case of an individual with a Ph.D. He or she has
four years of college, four years of graduate school, evaluations from several teach-
ers, a job market paper, and, sometimes, a dissertation. In many cases we accurately
predict that the person with the Ph.D. will have a high level of professional success,
but that is not always the case. Some people achieve more success than anticipated;
others do not quite achieve what had been expected.
The case of Michael Mussa has obviously been one of great anticipation, and of
great accomplishment, and all the early optimistic forecasts about him have turned
out to be correct. The first opportunity I had to forecast his career came with a
long—very long (close to 100 pages)—paper he wrote for Bob Fogel’
s economic
history seminar at the University of Chicago. The subject of the paper was agrarian
decline and agrarian discontent in the late-nineteenth-century United States. The
combination of empirical material and analytic interpretations, and his ability to
place this work in a broad context, resulted in a first-rate piece of economics and of
economic history that unfortunately is still not published. Clearly, Michael was a
student who would go far—that correct forecast was relatively easy to make. His
paper also invited another correct forecast: it was obvious that the author of that
paper would not, in his professional writing and talking, suf
fer from the sin of excess
brevity.
II.
The Time-Dura
tion of For
ecasts
There is a famous story—probably the most famous and widely read story concern-
ing the perils of forecasting—of an individual who it is believed was led to make the
supreme sacrifice because of an incorrect forecast based on a faulty interpretation.
(This ending is, of course, not to be suggested for any IMF forecasters, or indeed for
any other forecasters, who are incorrect.) Chicken Little was sitting under a tree
when an acorn fell on his head. He thought that the sky was falling and, displaying
admirable concern for the public, went to tell the king. Along the way he picked up
several others—Henny Penny, Ducky Lucky, Goosey Loosey, and Turkey Lurkey—
all of whom he convinced of the accuracy and importance of this situation. (As an
aside, one can’t help but feel that this gathering of like-minded animals going to the
king is intended to suggest another aspect of the forecasting profession, the herd
instinct that often seems to develop among those making predictions.)
T
rouble hit
the group of forecasters when they ran into Foxy Loxy, who deceived them; took
them to his den, not to the king; and ate them all up. It has been suggested that
Chicken Little would have been eaten even if his interpretations and forecast were
correct, but I find the idea of paying for your incorrect forecasts more appealing.
ON THE ACCURACY OF SOME PAST AND PRESENT FORECASTS
21
Within the sphere of economics, undoubtedly the most famous and widely used
forecast—one, incidentally, that thus far has often been incorrect—is that based on
the Malthusian doctrine of the relationship between resources and population.
Malthus was not the first to make this basic prediction; one can find similar forecasts
about anticipated decline owing to insufficient resources as far back as several thou-
sand years, but these predictions were premature and were made by individuals who
no doubt lacked Malthus’s analytical and empirical ability, as well as his public rela-
tions skills.
4
Rather than discuss the many problems that Malthus’s forecast has pre-
sented, I describe how Malthus linked his theory of population and resources to his
anticipation of the specific realization of events in a nineteenth-century British case
(see Hollander, 1990). His basic question was the following: How long would it be
before a lack of available resources slowed British population growth? Writing in
1803, Malthus saw the deceleration in the growth of food supplies, and consequently
of population, as three or four decades ahead, a prediction left unchanged as late as
the 1817 edition of
An Essay on the Principle of Population (Hollander, 1990, p. 21).
However, after the 1821 census, and by focusing on the declining marriage rate and
not the slight increase in the population growth rate, Malthus put off the occurrence
of the necessary (on economic grounds) deceleration until a full century had passed
(Hollander, 1990, p. 24). Thus, his pessimism about real contemporary events was
not as extreme as the predictions of his model might suggest (see also Wrigley,
1988). His theoretical position was consistent, based on real-world conditions, with
quite different periods until its realization. Whether the minor early-nineteenth-
century economist Thomas Chalmers should be classed as a Malthusian is an ongo-
ing discussion among historians of economic doctrine (Hollander, 1985, p. 58).
The positive claim that he was a Malthusian rested on his statement that there
would be “ultimate ‘stationarity,’
” since “every country has its limits.”
The coun-
terargument notes his comment that “the time may be indefinitely distant, and
indeed may never come, when the absolute and impossible barrier shall at length be
arrived at” (Hollander
, 1985, p. 58).
III. For
ecasts of
American Growth
When we examine the beginnings of settlement in the New World the predictions
seem rather obvious (see Engerman and Sokoloff, 2002). Spain was the number one
country in Europe, or number two depending how France is considered. For several
centuries Spain was to be number one in the
Americas. Columbus, commenting on
the gold in Spanish territories, argued for its pecuniary benefits, since “he who pos-
sesses it, can do as he wishes in the world.” This in addition to the nonpecuniary
4
See, for example, the Carthaginian theologian
T
ertullian, writing in about
A.D. 200, commenting that,
“What most frequently meets our view (and occasions complaint) is our teeming population. Our numbers
are burdensome to the world, which can hardly support us” (quoted in
Crunch Time, “Overpopulation,”
www.crunchtime.org.). He is quoted as anticipating pestilence, hunger, war, and flood. Wilfred Beckerman
(1972, p. 332) points to a possibly even earlier example, stating that “a fifth-century B.C. Club of Rome
(?Athens) team would certainly have predicted the imminent end of the world by extrapolating the already
growing shortage of timber resources and pointing to the over
-population of Athens.”
Stanley Engerman
22
benefits that come, because “it can even drive souls into Paradise” (Pagden, 1993,
p. 17; see also Elliot, 1970, pp. 1–27). Spain (and Portugal) had a one-century lead
in coming to the New World compared with England and France, and used this
to good advantage, following the rule of American bank robber Willie Sutton: go
where the money (and the people) are. The Spanish went to areas that contained
more than 70 percent of the New World’s Native American population and that had
very well developed societies, economically and politically, with great amounts of
wealth. The 13 colonies and Canada were settled later, and these were considered
the dregs of the Americas, rather undesirable for settlement. Even when they began
to come to the New World, for the better part of a century more British migrants
went to the West Indies than to the mainland. Mainland areas had only about 5 per-
cent of the Native American population, and the Native Americans lived mostly in
small units, more like hunter-gatherers than the sophisticated societies of South
America. It was only after one-and-a-half centuries that the 13 colonies surpassed
Latin America and parts of the Caribbean in terms of income per capita, at roughly
the same time the British surpassed the Spanish on the European continent, though
even 350 years after Columbus, Spain’s largest remaining New World colony, Cuba,
was still wealthy
, having become the producer and exporter of about one-half of the
world’s cane sugar.
France had less success in the New World than did the British, even though its
colony of San Domingue (now Haiti) was in 1780 about the richest area in the world.
The French were not as successful in Canada as they were with San Domingue; it
seems that for a long time France did not regard Canada’s prospects highly
.
Thus,
the duc de Sully, finance minister at the start of the seventeenth century, argued
against plans to settle Canada since “great wealth was never derived from places
beyond forty leagues [of north latitude],” anticipating a familiar
World Bank map
relating climate to income level (Cole, 1939, p. 42). Much later, during the Seven
Years’ War between the British and the French, Voltaire presented an interesting
benefit-cost analysis, claiming that these “two nations are fighting over a few acres
of snow,” and he regarded both to be “mad” since “they spend more money on this
glorious war than the whole of Canada is worth,” but he does make this argument
without a statement of what the appropriate rate of interest should be (Voltaire,
1759, p. 110 in 1947 edition).
Of the many forecasts made for the new
American nation after it achieved inde
-
pendence, I will deal with only a few. One of the more accurate predictions was
made by the English writer John Lord Sheffield in 1784. He claimed that “the states
will suf
fer—they have lost much by separation,” but that, for the British “it is not
probable our Commerce will be much hurt.” He believed that English manufactur-
ing sales to the Americans would remain high, since “it will be a long time before
the Americans can manufacture for themselves.” Sheffield also argued “that the
commerce with the revolted colonies was of advantage to this country cannot be
doubted; nevertheless it may be easily shown, that it was not the most advanta
-
geous”
(Shef
field, 1784, pp. 135, 191, and 241 in 1970 edition).
This forecast of
continued American reliance on British manufacturing exports held for several
decades, as did the importance of the American agricultural labor force, and he
might have been right in the claim, made also by Edmund Burke, that England
ON THE ACCURACY OF SOME PAST AND PRESENT FORECASTS
23
would have been better off freeing the colonies earlier.
5
The economic strength of
Sheffield’s contention is seen in the fact that U.S. per capita incomes declined in
the three decades after the start of the Revolution, while, after losing the 13 American
colonies, Britain experienced a several-decades-long growth spurt that marked the
period of the Industrial Revolution (Shepherd and Walton, 1976, pp. 412–414;
Deane and Cole, 1962, pp. 75–82).
Two of the most intellectual and intelligent of America’s founding fathers made
some interesting predictions. In 1751, describing the economics of western settle-
ment, Benjamin Franklin commented that “so vast is the territory of North America,
that it will require many ages to settle it fully” (Franklin, 1751, p. 313 in 1836
edition). Fifty years later, discussing the Louisiana Purchase, Thomas Jefferson
claimed that it would make it “possible [for the United States] to remain a nation of
farmers for a thousand years”—that presumably being the time it would take to fully
occupy the land (McDonald, 2003).
6
And, in 1808, when reneging on a previous
commitment for federal aid to help states build canals, Jef
ferson described New York
State’s Erie Canal as “little short of madness,” given its length through wilderness
(more than 10 times longer than any existing U.S. canal). To Jefferson the canal was
“a century ahead of its time.” Even in 1822, when the new canal was nearing com-
pletion, he stated that “many
, I dare say, still think with me that New York has antic-
ipated, by a full century, the ordinary progress of improvement” (Shaw, 1966, p. 36).
7
IV. Forecasts About the Future of American Slavery
In the first half of the nineteenth century there were many discussions of the proba-
ble demise of slavery on economic grounds, presumably leading to a voluntary free
-
ing by slaveowners.
8
Those predictions were generally based on estimates made
from data relating to population, area, and resources, and all were based on one ana-
lytical principle drawn from basic economics—the law of diminishing returns.
One systematic example can be found in writings of the Virginia economist
George Tucker, who also came up with some quite sophisticated national income
estimates based on the 1840 census. Tucker actually presented some earlier decline
estimates when he was in Congress, during the debates concerning the Missouri
Compromise, in 1820. He discussed population density estimates, and he related pro-
ductivity to density to evaluate the argument that for about one century “there could
be no danger from the relative increase of slaves” (Tucker, 1820, p. 13). He was
offended when a believer in that danger claimed that politicians should not worry
5
Bentham (1995), at the start of the nineteenth century, used the British experience with the United
States to argue for Spain’s willingly freeing its American colonies rather than fighting against their desire
for independence. Here, however, wars were fought, and independence improved British trade with South
America.
6
Earlier
, in his first inaugural address (given in 1801), Jefferson commented that the United States was
“. . . a chosen country, with room enough for our descendents to the thousandth and thousandth generation.”
7
The Erie Canal was only one of many public works, including the Big Dig, the Louisiana Superdome,
and just about all preparations for the Olympics, for which there were rather large cost overruns relative
to the original projections. The initial projections were often deliberately low to make the projects politi-
cally acceptable (Engerman and Sokoloff, 2006).
8
Parts of the next paragraphs draw on Engerman (1993).
Stanley Engerman
24
about this danger since it was too far in the future. Tucker claimed, however, that
“an hundred years, sir, is a long period for human life, but is a short one in the life
of a nation” (Tucker, 1820). (Tucker’s political career, it might be noted, was rather
brief.) Two decades later, he provided a more systematic approach to the question.
He looked at the conditions of population density when villenage was abolished in
England, using the midpoint between the fourteenth- and seventeenth-century esti-
mates and adjusting for differential soil fertility and living standards. Using slave and
white population growth in the United States, as determined from the 1790 through
1840 censuses, he gave slavery’s future “a little upwards of eighty years” (Tucker,
1855) (although with many caveats and qualifications). Twelve years—and one
census—later, and with the addition of Texas, he noted that the addition of this
new state would clearly “tend to prolong the continuance of slavery”; otherwise,
he claimed that “the views presented by the author in 1843 remain unchanged, as
he has met with neither fact nor argument to affect their soundness” an attitude
clearly worthy of any academic reply or rejoinder today (Tucker, 1855, pp. 108–18
App. 28). I know one historian who thinks that Tucker’s extension of his prediction
of the demise of slavery from 80 to about 100 years reflects a political fear, in case
the time of slavery’
s ending would get too close and become politically relevant, but
Tucker’s revision is most likely an appropriate triumph of new empirical information
about the addition of Texas, rather than ideological belief.
Another example from the antebellum period relates to a position taken by
Abraham Lincoln in his 1858 debates with Stephen Douglas, a position based solidly
on the law of diminishing returns, and a belief in what could ultimately be called the
political, not the natural, limits of the expansion of slavery. Lincoln’s position in the
1858 debates was not to touch slavery where it existed but to prevent its expansion
into the territories—a position that, however sound politically, seemed to Douglas
to be ill-fitting with Lincoln’s moral attacks on slavery. Douglas paraphrased
Lincoln’s position as aiming to keep the slaves confined to their present limits while
they continued increasing until the soil on which they lived would no longer feed
them. According to Douglas, Lincoln would thus be able to put slavery
in a course
of ultimate extinction by “starving them to death.” Lincoln’s response was that
“while it [slavery] drives on in its state of progress as it is now driving, and as it has
driven for the last five years, I have ventured the opinion, and I say to-day, that we
will have no end to the slavery agitation until it takes one turn or the other.
[Applause.] I do not mean that when it takes a turn towards ultimate extinction it will
be in a day
, nor in a year
, nor in two years. I do not suppose that in the most peace-
ful way ultimate extinction would occur in less than a hundred years at the least; but
that it will occur in the best way for both races in God’s own good time, I have no
doubt [Applause].” Lincoln’
s concern with the founding fathers’
belief in “ultimate
extension” points to another pitfall in prediction, since he states that it was put off
by the unexpected invention of the cotton gin (Lincoln, 1989 edition, pp. 603–604,
677–678, 753, and 820).
9
9
The 100-year expectation was held by other northerners at this time, including a leading political fig-
ure, Horace Greely (Van Deusen, 1953, p. 92).
ON THE ACCURACY OF SOME PAST AND PRESENT FORECASTS
25
Finally, in the discussion of southern slavery, I want to turn to a source with
some contemporary resonance, the market newsletters of those slave traders in
the South, who issued weekly or monthly newsletters to customers detailing the
prices of slaves and providing relevant information on the expected behavior of
markets and of slave prices. I will quote from two at the time of Lincoln’s election,
where there is some new uncertainty about slave prices but no expectation of war,
something that must also have been the case at this time in the North, given the will-
ingness of northerners to lend money to southerners.
In September 1860 Betts and Gregory of Richmond wrote, “The presidential
election is having considerable effect on the market. How it will go no man can tell.
But we would advice you not buy nothing but good negroes and buy them at prices
to sell immediately.” In January 1861 they wrote: “There is more activity in the mar-
ket with an upward tendency.”
Dickson Hill Co. of Richmond wrote in December 1860:
The year is about to close and it will be difficult for us to say what will be
done in the ensuing years. . . . We have no hope for any political change
which will give peace and some confidence in commercial matters.
The speculations and extravagances of the last three years has put up
prices of land, negroes, and crops, to such a height that it could not be
maintained, and the credit of the country has been extended to a degree
unparalleled in this country before.
We think for some years to come Negroes will not command over
$1,000 for best men and $800 for best women. During the incoming year
we ought not to expect more than $800 for men and $600 for women; and
we fear that even these prices will not be obtained for two or three
months to come (
Slave Traders, Duke University Archives).
V. For
ecasts About Natural Resources
After that of Malthus, probably the most famous prediction by a professional
economist, also seemingly incorrect thus far, was that by William Stanley Jevons
in 1865, concerning the problems to be created by the limited amount of coal in
England. This is only one of a very familiar series of forecasts about the
limita
-
tions of nonreproducible natural resources that have an extremely long history.
Jevons thought the limited amount of coal in England would ultimately retard
England’s economic progress, and he claimed that “the check to our progress
must become perceptible considerably within a century from the present time,”
while “the conclusion is inevitable, that our present happy progressive condition
is a thing of limited duration.”
This, he thought, would lead to “wholesale emi
-
gration” as well as possible “moral and intellectual retrogression.” He concluded
by of
fering the British “the momentous choice between brief greatness and
longer continued mediocrity” (W. S. Jevons, 1865, pp. lx, 215, 316, and 349). It
may be difficult to grade Jevons’s forecast, since even though he was close con-
cerning the precise timing of Britain’
s relative decline, his reasons for expecting
retardation were not correct.
In an interesting family follow-up, Jevons’
s son, H. Stanley Jevons, returned
to the problem about 50 years later, after the period including the one era of sharp
increases in relative coal prices during the late nineteenth and the twentieth cen-
Stanley Engerman
26
turies. He followed an analysis similar to his father’s, allowing some adjustments
for higher costs, and concluded that his father was quite right, but that “it is quite
possible, indeed, that we shall be suffering from this degree of exhaustion of our
mines in less than 200 years.” He attributed the change in timing to “the discov-
ery of far greater deposits of coal than could have been expected” (H. S. Jevons,
1915, pp. 756–57). This shift in timing from his father’s approximate 100 years to
his 200 years was based on the fact that more new reserves were being discovered
than had been used over the previous half-century. Actually, about three decades
before the elder Jevons’s writing, Charles Babbage, inventor of a computing
machine, had suggested that “the source of this power is not without limit, and the
coal-mines of the world may ultimately be exhausted.” He went on to argue, how-
ever, that because of the increased expense, “long before that period arrives, other
methods will probably have been invented for producing [heat]”—possibly by
drawing power from the sea (Babbage, 1835, p. 388 in 1963 edition).
Many other examples of predicted mineral and resource exhaustion are avail-
able for the United States and elsewhere, often with similar prolonged time periods,
under the quite-specific assumptions of no changes in key parameters. For example,
in England in 1624, Richard Eburne wrote that the supply of wood “fast decays” and
“that very want of it only
, within a few years is like to prove exceedingly hurtful to
our land, and can be no way repaired, but by transplanting the people” (Eburne,
1624, p. 22). Timber and other mineral shortages have been predicted in the United
States since at least the late nineteenth century. These predictions often were made
hastily
. Colonel Drake first found oil in Pennsylvania in 1859. By 1861 the fear of
an exhaustion of the oil supply, if nothing was done to prevent waste, was expressed,
and in 1882 the fear that oil would run out, apparently soon, was raised even when
about 99 percent of all
American oil came from New York and Pennsylvania
(Williamson, 1945, p. 99). Earlier, in 1862, Andrew Carnegie was so convinced that
oil would soon be in scarce supply that he stockpiled output, awaiting the rise in
prices (Sabin, 1999, p. 475; and Giddens, 1938, p. 82). At the start of the twentieth
century Andrew Carnegie forecast that coal would last for about 200 years (most
others predicted 80–100 years) and that the “best half of our iron ore” would be gone
in about 30 years (Williamson, 1945, p. 102). Again, it is easy to understand the
logic of expecting to run out of natural resources, but, as before, the key question
remains how far in the future is this expected to occur (see also Jones, 1991; and
United Nations, 1973, pp. 370–73).
VI. The Twentieth Century
I will write less about the twentieth century, since much more is known, and some
issues are covered by Bob Fogel earlier in this issue. Irving Fisher
’
s famous quote on
permanent prosperity was similar to remarks made by Herbert Hoover in 1929, as
chair of the Committee on Recent Economic Changes of the President’s Conference
on Unemployment. Hoover said, “The years 1922 to 1929 witnessed a marked
increase in the physical volume of production” owing to a “striking increase in pro-
ductivity per man-hour,” and that “we seem only to have touched the fringe of our
potentialities” (Hoover
, 1929, pp. xiv and xix). Nevertheless, Hoover commented that
the continuation of prosperity required both balanced change in the economy and suc
-
ON THE ACCURACY OF SOME PAST AND PRESENT FORECASTS
27
cessful leadership (Hoover, 1929, p. xxi). The New York World headline of May 15,
1929, however, was more optimistic: “Prosperity Only Dawning, Hoover Survey
Reports,” and repeated his comment that the “U.S. Has but ‘Touched the Fringe’ of
Potentialities After Seven ‘Splendid’ Years.” Less attention was given to the caveat
attributed to Hoover: that this could occur “only if the economic balance between
production and consumption of wealth can be maintained” (
New York World, 1929
p. 1). Yet even more concern was expressed in the last essay in the volumes by the
director of the National Bureau of Economic Research, Wesley Clair Mitchell.
Mitchell said, “All is not well. Americans have seen more uniformly fortunate
time; for example, in 1906. . . . The condition of agriculture, the volume of un-
employment, the textile trades, coal mining, the leather industries, present grave
problems. . . . How rapidly these conditions will mend, we do not know. Some
may grow worse.” Further, he predicted that prosperous industries would become
unstable, and since there had been no “severe depression since 1921 [there] is no
guarantee that we shall be equally prudent, skillful, and fortunate in the years to
come” (Mitchell, 1929, Vol. II, p. 909). Not an unequivocal prediction of an imme-
diate collapse by America’s leading analyst of business cycles, but not quite a repeat
presentation of Fisherian unbounded optimism.
I end this twentieth-century discussion by mentioning two major forecasts that
have had influence and attracted attention in the past half-century or so. First, various
predictions, some by leading economists, were made during World War II that the end
of the war would bring about a return of the depression (Sapir, 1949). Second, by my
count, the United States was expected to be economically overtaken by at least three
dif
ferent countries in the second half of the twentieth century: the Soviet Union,
Germany, and Japan. These are events that we are still waiting for, and while doing
so, we can no doubt add China and the EU to the list of predicted overtakers.
VII. Conclusion
To end this paper on forecasting, I searched for a more casual approach consistent
with Michael Mussa’
s spirit and his interests. I decided I would print what I regard
as the best popular song concerning the difficulty of forecasts, by George and Ira
Gershwin for the movie
Shall
We Dance.
(This song is also unusual in that it is one
of the few solos Ginger Rogers ever got to sing in the Astaire-Rogers movies.)
THEY ALL LAUGHED
By: Geor
ge Gershwin and Ira Gershwin
[Verse]
The odds were a hundred to one against me
The world thought the heights were too high to climb
But people from Missouri never incensed me
Oh, I wasn’
t a bit concerned for from history I had learned
How many, many times the world had turned
[Chorus]
They all laughed at Christopher Columbus
When he said the world was round
They all laughed when Edison recorded sound
Stanley Engerman
28
T
hey all laughed at Wilbur and his brother
When they said that man could fly
T
hey told Marconi
Wireless was a phony
It’s the same old cry
They laughed at me wanting you
Said I was reaching for the moon
But oh, you came through
Now they’ll have to change their tune
They all said we never could be happy
They laughed at us and how!
But ho, ho, ho!
Who’s got the last laugh now?
They all laughed at Rockefeller Center
Now they’re fighting to get in
They all laughed at Whitney and his cotton gin
They all laughed at Fulton and his steamboat
Hershey and his chocolate bar
Ford and his Lizzie
Kept the laughers busy
That’s how people are
They laughed at me wanting you
Said it would be, “Hello, Goodbye.”
But oh, you came through
Now they’re eating humble pie
They all said we’d never get together
Darling, let’s take a bow
For ho, ho, ho!
Who’s got the last laugh?
Hee, hee, hee!
Let’s at the past, laugh
Ha, ha, ha!
Who’s got the last laugh now?
THEY ALL LAUGHED
By: GEORGE GERSHWIN, and IRA GERSHWIN
©
1936 (Renewed)
GEORGE GERSHWIN MUSIC,
and IRA GERSHWIN MUSIC
All Rights Administered by WB MUSIC CORP. All Rights Reserved. Used by Permission.
ALFRED PUBLISHING CO., INC., Miami, Florida 33014
REFERENCES
Babbage, Charles, 1835, fourth edition enlar
ged,
On the Economy of Mac
hinery and Manu
-
factur
es,
reprinted 1963 (New
Y
ork: Augustus M. Kelley, 4th ed.).
Beckerman,
W
ilfred, 1972, “Economists, Scientists, and Environmental Catastrophe,”
Oxfor
d
Economic P
aper
s,
V
ol. 24 (November), pp. 327–44.
Bentham, Jeremy, 1995, Colonies, Commerce, and Constitutional Law: Rid Yourself of Ultra-
maria and Other Writings on Spain and Spanish America (Oxford: Clarendon Press).
ON THE ACCURACY OF SOME PAST AND PRESENT FORECASTS
29
B
ordo, Michael David, 1975, “John E. Cairnes on the Effects of the Australian Gold Discoveries,
1851–73: An Early Application of the Methodology of Positive Economics,” History of
Political Economy, Vol. 7 (Fall), pp. 337–59.
Cairnes, John Elliott, 1862, The Slave Power; Its Character, Career, & Probable Designs: Being
an Attempt to Explain the Real Issues Involved in the American Contest, second edition
reprinted 1969 (New York: Harper).
Cole, Charles Woolsey, 1939, Vol. I, Colbert and a Century of French Mercantilism (New York:
Columbia University Press).
Coupland, Reginald, 1923, Wilberforce (Oxford: Clarendon Press).
Deane, Phyllis, and W. A. Cole, 1962, British Economic Growth, 1688–1859: Trends and Struc-
ture (Cambridge: Cambridge University Press).
Doan, Ruth Alden, 1987, The Miller Heresy, Millennialism, and American Culture (Philadelphia:
Temple University Press).
Eburne, Richard, 1624, Plaine Pathway to Plantations (London: Goldsmiths’-Kress Library of
Economic Literature, No. 543.1).
Elliott, J. H., 1970, The Old World and the New, 1492–1650 (Cambridge: Cambridge University
Press).
Engerman, Stanley L., 1993, “Chicken Little; Anna Karenina, and the Economics of Slavery:
Two Reflections on Historical Analysis with Examples Drawn Mostly from the Study of
Slavery,” Social Science History, Vol. 17 (Summer), pp. 161–71.
———, and Kenneth Sokoloff, 2002, “Factor Endowments, Inequality, and Paths of Deve-
lopment,” Economiá, Vol. 3 (fall), pp. 41–109.
———, 2004, “Digging the Dirt at Public Expense: Governance in the Building of the Erie Canal
and Other Public Works,” NBER Working Paper No. 10965 (Cambridge, Massachusetts:
National Bureau for Economic Research). Forthcoming in
Corruption and Reform, ed. by
Edward Glaesr and Claudia Goldin (Chicago: University of Chicago Press).
Franklin, Benjamin, 1751, “Observations Concerning the Increase of Mankind and the Peopling
of Countries,” reprinted 1836 in The Works of Benjamin Franklin, ed. by Jared Sparks (Boston:
Hilliard Gray), pp. 311–21.
Gibbon, Edward, 1776–1788, The History of the Decline and Fall of the Roman Empire, Vol. 4,
reprinted 1900 (New York: Thompson and Thomas).
Giddens, Paul H., 1938, The Birth of the Oil Industry (New York: Macmillan).
Hollander
, Samuel, 1985,
The Economics of J
ohn Stuart Mill
Vol. I (Toronto: University of
Toronto Press).
———, 1990, “Malthus’s Vision of the Population Problem in the Essay on Population,” Journal
of the History of Economic Thought, Vol. 12 (spring), pp. 1–26.
Hoover, Herbert (Chairman), ed., 1929,
Recent Economic Changes in the United States: Report of
the Committee on Recent Economic Changes of the President’s Conference on Unemployment,
V
ol. I (New
Y
ork: McGraw Hill).
Hume, David, 1752, “Of Commerce,” reprinted 1987 in
Essays Moral, Political and Literary
by David Hume,
ed. by Eugene F. Miller (Indianapolis: Literary Classics), pp. 253–67.
Jevons, Herbert Stanley, 1915, The British Coal Trade (London: Kegan Paul, Trench, Trübner).
Jevons, William Stanley, 1865, The Coal Question (London: Macmillan).
Jones, Eric L., 1991, “The History of Natural Resource Exploitation in the Western World,”
Research in Economic History, Supplement 6, pp. 235–52.
Levine, Peter
, 1985,
A.
G. Spalding and the Rise of Baseball
(New
Y
ork: Oxford University Press).
Lincoln,
Abraham, 1989,
Speec
hes and
Writings, 1832–1858
(New
Y
ork: Library of America).
Stanley Engerman
30
McCloskey, D. N., 1992, “The Art of Forecasting: From Ancient to Modern Times,” Cato
Journal,
Vol. 12 (spring/summer), pp. 23–43.
M
cDonald, Forrest, 2003, “Revising Jefferson,”
W
ashington Times,
O
ctober 25.
Mitchell, Wesley C., 1929, “A Review” in
Recent Economic Changes in the United States:
Report of the Committee on Recent Economic Changes of the President’s Conference on
Unemployment,
Vol. II (New York: McGraw-Hill), pp. 841–910.
New York World, 1929, May 15.
North British Review (American Edition), 1862, “Review of The American Union: Its Effect on
National Character and Policy
,” Vol. 36 (February), pp. 122–44.
Pagden, Anthony, 1993,
European Encounters with the New World (New Haven: Yale University
Press).
Parliament of the Commonwealth of Australia, 1912,
Report of the Royal Commission on the
Sugar Industry
(Canberra).
Sabin, Paul, 1999, “A Dive into Nature’s Great Grab Bag: Nature, Gender and Capitalism in
the Early Pennsylvania Oil Industry,”
Pennsylvania History, Vol. 66 (fall), pp. 472–505.
Sapir, Michael, 1949, “Review of Economic Forecasts for the Transition Period” in
Conference
on Research in Income and Wealth, Studies in Income and Wealth,
Vol. 11 (New York:
National Bureau of Economic Research), pp. 275–367.
Seymour, Harold, 1960,
Baseball: The Early Years (New York: Oxford University Press).
Shaw, Ronald E., 1966,
Erie Water West: A History of the Erie Canal, 1792–1854 (Lexington:
University of Kentucky Press).
Sheffield, John Lord, 1784,
Observations on the Commerce of the American States, reprinted
1970 (New York: Augustus M. Kelley).
Shepherd, James F., and Gary M. Walton, 1976, “Economic Change After the American
Revolution: Pre and Post War Comparisons of Maritime Shipping and Trade,”
Explorations
in Economic History,
Vol. 13 (October), pp. 397–422.
Smith, Adam, 1776,
The Wealth of Nations, Vols. I and II, reprinted 1976 (Oxford: Oxford
University Press).
Tucker, George, 1820, Speech on the restriction of slavery in Missouri given at the U.S. House
of Representatives, Washington, February 25.
———, 1855,
Progress of the United States in Population and Wealth in Fifty Years (New York:
Press of Hunt’s Merchant Magazine).
United Nations, 1973,
Department of Economic and Social Affairs Determinants and Con-
sequences of Population Trends: New Summary of Findings on Interaction of Demo-
graphic, Economic and Social Factors,
Volume 1 (Population Studies No. 50) (New York:
United Nations).
V
an Deusen, Glyndon G., 1953,
Hor
ace Greely: Nineteenth Century Crusader
(Philadelphia:
University of Pennsylvania Press).
Voltaire, Francois-Marie Arouet De, 1759,
Candide or Optimism, reprinted 1947 (Harmondsworth:
Penguin Books).
Williamson, Harold F., 1945, “Prophecies of Scarcity or Exhaustion of Natural Resources in the
United States,”
American Economic Review, Vol. 35 (May), pp. 97–109.
W
ilberforce,
W
illiam, 1807,
A Letter on the
Abolition of the Slave
T
r
ade
(London:
T
. Cadell &
W. Davies, J. Hatchard).
Wrigley, E. A., 1988, “Malthus on the Prospects for the Laboring Poor,”
Historical Journal,
V
ol. 31 (December), pp. 813–29.
