STV
7? COPY 2
BEB
FACULTY WORKING
PAPER NO. 1177
The Integration and Efficiency of the
London and Amsterdam Stock Markets in
The Eighteenth Century
Larry Neal
College of Commerce and Business Administration
Bureau of Economic and Business Research
University of Illinois, Urbana-Champaign
BEBR
FACULTY WORKING PAPER NO. 1177
College of Commerce and Business Administration
University of Illinois at Urbana-Champaign
September, 1985
The Integration and Efficiency of the
London and Amsterdam Stock Markets
in the Eighteenth Century
Larry Neal, Professor
Department of Economics
The author acknowledges with gratitude helpful comments on earlier
drafts by George Alter, Jeremy Atack, Robert Eagly, Barry Eichengreen,
John James, Donald McCloskey, John McCusker, and James Riley as well as
discussants in seminars at the University of Chicago, University of
Illinois and Indiana University. Research support was given by the
Research Board and the Bureau of Economic Research of the University of
Illinois and the National Science Foundation, (SES 83-09211).
Digitized by the Internet Archive
in 2011 with funding from
University of Illinois Urbana-Champaign
http://www.archive.org/details/integrationeffic1177neal
ABSTRACT
In the early eighteenth century, a special form of economic
integration occurred between the two leading mercantile cities of
Europe, Amsterdam and London. An international capital market
developed, the significance of which has been largely overlooked in
the more recent historical and the economic literature. This paper
lays the basis for a greater appreciation of the role of this early
international capital market in particular, and such markets in
general, by explaining its operation in theoretical terms and
analyzing its results in quantitative terms.
It concludes that these two markets were efficient and well-
integrated from the second quarter of the eighteenth century on.
The spot prices quoted in the London market followed a random
walk process consistent with efficient markets. The forward prices
in the Amsterdam market were highly correlated with them. The
semi-annual dividend payments in London then produced regular
patterns in the Amsterdam prices both from the Le Bachelier effect
and the practice of London quoting prices ex dividend during the
preparation of dividend payments while Amsterdam quoted them
with dividend.
In the early eighteenth century, a special form of economic
integration occurred between the two leading mercantile cities of Europe,
Amsterdam and London. An international capital market developed which
led to ever wider-ranging capital markets for each center over the
succeeding centuries. It may also have facilitated the progress of economic
integration for northwestern Europe in terms of markets in goods and labor.
The significance of this, however, seems to have been largely overlooked in
the more recent historical and the economic literature. This paper lays the
basis for a greater appreciation of the role of this early international
capital market in particular, and such markets in general, by explaining its
operation in theoretical terms and analyzing its results in quantitative
terms.
Shares of the great chartered joint stock corporations in England
were traded simultaneously on the stock exchanges of London and
Amsterdam at the end of 1723 and continued to be traded through peace,
war, and revolution into the nineteenth century. Little evidence remains of
this activity save for the prices of the shares but these exist in great
quantity. Using modern theory and quantitative techniques, a great deal of
interest can be inferred from them. Section I explains how this unique
form of capital market integration came into existence and how the two
markets operated separately and jointly. Section II summarizes what the
evidence available to us tells about the operation of the respective capital
markets. It shows that, at least for the multiply listed stocks, the two stock
markets were economically efficient by modern standards throughout the
eighteenth century, although some doubt may remain about the operation of
the Amsterdam market. Section III then tackles the challenging task of
explaining the small, but persistent, differences in prices that existed
between the two markets. The regression results show technical differences
in the operations of the two markets largely account for both the
persistence with which Amsterdam prices were above London prices, and
for the existence of apparently unexploited profit opportunities in the
Amsterdam market. It turns out that exchange rate fluctuations were also
often a major influence on price differences, sometimes reflecting large
movements in the capital markets and sometimes causing adjustments in
them. The concluding section draws some of the implications of this
chapter of history for our understanding of the integration of international
capital markets especially in terms of information flows and government
interventions.
It is traditional to date the beginning of the English National Debt,
referring to long-term funded debt, in 1693. In that year, Act 4 William
and Mary c.3 imposed a special duty on beer, ale, and other liquors to
guarantee payment of the interest on a million pound loan which was to be
floated at 10% interest.1 By this action, William simply applied to his new
domain the same techniques for raising credit that he had employed as
Stadhouder in the Netherlands. A crucial element in the set of financial
practices brought to England in William's train was the resale of shares in
joint stock corporations, i.e., the initiation of our modern stock exchanges.
While chartered joint stock companies existed in England prior to William,
it appears that trade in their shares increased considerably in the early
1690s and certainly the number of companies increased markedly in that
decade. This followed a very active period of stock trading in the
Amsterdam Beurs in the 1680s.
To aid him in raising money for his War of the League of Augsburg
against Catholic France, William brought with him numerous financial
advisors and military contractors from Holland. These men included a high
proportion of Jews and Huguenots who were eager to apply in a relatively
backward England the financial techniques and institutions that had been
Alice Clare Carter, The English Public Debt in the Eighteenth Century,
(1968), p. 5.
developed over the past century in Amsterdam.2 Their activities eventually
came to concentrate on shares of the Bank of England (founded in 1694),
the New East India Company (1698), the United East India Company (a
consolidation of the Old and New East India Companies which occurred in
1702), and the South Sea Company (1711). The charters of each permitted
foreigners to own shares and this right was upheld by the Crown and the
companies despite occasional challenges from members of Parliament.3 The
shares of these three great companies were liquid assets for both English
and foreign owners because an active resale market existed for them. The
trading activity occurred in the London Stock Market and in the Amsterdam
Beurs.
Van Dillen gives a few of the more noteworthy examples of the Jewish
financial investors. Moses Machado went with the king to England in 1688
and became his prime contractor for the campaign in Ireland; Joseph de
Medina had a large contract as military supplier in 1713; while Sir Solomon
de Medina was the greatest army contractor of his day, financing in
particular the campaigns of the Duke of Marlborough. (J. G. Van Dillen,
"De economische positie en beteknis der Joden in de Republiek en in de
Nederlandse koloniale wereld," in H. Brugmans and A. Frank, eds.,
Geschiednis der Joden in Nederland, (1940), p. 584.)
Further, our earliest description of the operation of the Amsterdam stock
exchange, Josef Penso de la Vega's Confusion de Confusiones, appears to be a
highly florid elaboration of an earlier technical manual that he had
prepared on the various techniques and regulations employed in the
Effectenbeurs. The purpose of this manual most likely was to inform his
countrymen who had gone to London and who wished to participate in the
speculation that was beginning there. Hermann Kellenbenz, introd. to Josef
de la Vega, Confusion de Confusiones, (1957), p. xiv.
For details, see Larry Neal, "Efficient Markets in the Eighteenth
Century? The Amsterdam and London Stock Exchanges," in Jeremy Atack,
ed., Proceedings of the Business History Conference, (1982).
The prices at which shares changed hands on the London market are
available to us on a daily basis in John Castaing's Course of the Exchange.4
This remarkable data source began publication in March 1697 and continued
to appear twice-weekly, on Tuesdays and Fridays, through the entire
eighteenth century under a variety of publishers.5 Each issue gave price
quotations for the past three trading days on each of the major securities
traded by the brokers in Exchange Alley. The list was headed by the shares
in the three leading joint stock companies but included the other chartered
companies as well as quotations on various issues of government debts and
annuities.
For the prices on the Amsterdam Beurs, we have available a series of
price quotations taken every two weeks for the shares of the Bank of
England, the East India Company and the South Sea Company. These were
The first regular publication of stock prices on the London Stock
Exchange appears to have been John Houghton's Collection for Improvement
of Husbandry and Trade, which began publication March 30 (O.S.), 1692 and
each week provided the Wednesday prices for the shares of eight chartered
corporations until it ceased publication in June; when it resumed in January
20, 1692/3 it continued to give Wednesday prices. With the issue of January
22, 1696/7 changed format and the dates were no longer for the preceding
Wednesday of that week, but rather for a preceding Tuesday or Friday.
This makes it appear that they were merely copied from Castaing's
definitive Course of the Exchange which began publication erratically in
1697 and then continuously from the beginning of 1698 on Tuesdays and
Fridays.
John J. McCusker, Money and Exchange in Europe and America, 1600-
1775; A Handbook, (1978), p. 31. A full history of the publication can be
found in Larry Neal, "The Rise of a Financial Press in Western Europe,
1677-1796," Proceedings of the Business History Conference, 13 (1985),
forthcoming. In 1811, when it was published by Wetenhall, it became the
official price list of the London Stock Exchange.
published by the Dutch economic historian, J.G. Van Dillen in 193 1.6 Van
Dillen drew these data from the Amsterdamsche Courant. a Dutch
newspaper which appeared thrice-weekly. Starting July 14, 1723 it began to
give price quotes for shares of the Dutch East India and Dutch West India
Companies with the agio rate for the Bank of Amsterdam. Beginning with
the issue for August 9, 1723 (N.S.), the Courant began giving in addition
quotes for the three English joint-stock companies. These continued to
appear regularly in each issue, with quotes for other Dutch and English
securities coming in for a time and then leaving, until December 19, 1794
with the French occupation of Amsterdam and the founding of the Batavian
Republic. The official price list for stocks on the Amsterdam Beurs began
publication the following year.7
This fortnightly series by Van Dillen reduced the number of joint
observations for each of the English companies from over 30,000 for just
the London market to only 1,676. For each date in the Amsterdam series, I
took the London quotation on the same trading day for each of the three
J.G. van Dillen, "Effectenkoersen aan de Amsterdamsche beurs 1723-
1794," Economisch-historisch Jaarboek, dl. 17, (1931), pp. 1-34.
•t
Johannes de Vries, Een Eeuw vol Effecten, Historische schets van de
Veremging voor de Effectenhandel en de Amsterdamse Effectenbeurs 1876-
1976, (1976), p. 19.
stocks.8 Graphing both the levels and first differences of the prices in each
market against each other for each company makes it evident that the two
markets were very closely correlated from the beginning of the series.
These graphs are far too long to reproduce here but the visual effect of
nearly perfect congruence is indicated perhaps in Table 1. This gives the
correlation coefficients between the levels and the first differences of the
price series in Amsterdam and London for the Bank of England, the East
India Company, and the South Sea Company. These correlation coefficients
are quite consistent for the levels across the four peacetime periods that
occurred between 1723 and 1794.9 And they are consistently high for each
of the three companies. The correlations for the first differences are
naturally much lower but also show much more variation by company and
by time period. One's first question upon finding two distinct price series
for the same financial asset -- were the two markets in which prices were
struck closely integrated? — is answered here with a resounding affirmative
This exercise was complicated by two features: 1) the Dutch had been
on the Gregorian calendar since the middle/end of December 1582 while the
British did not shift until September 2/13, 1752; and 2) the Amsterdam
market traded Sunday through Friday while the London market traded
Monday through Saturday. To deal with the first feature, I counted back
eleven days to find the corresponding London quotes before September 13,
1752; the second feature was dealt with by matching the Saturday quote in
London to a Sunday quote in Amsterdam whenever one appeared.
From Keller's Dictionary of Dates (1934), I chose October 19, 1739 as the
start of the first war period (War of Jenkins Ear) and October 1748 as the
end (Treaty of Aix-la-Chapelle). Hostilities began for the Seven Years War
in August 1756 while for financial purposes in the capitals they ended with
the Treaty of Paris signed February 10, 1763. I took March 13, 1778 as the
effective date of hostilities in Europe arising from the American War for
Independence since this was when the Treaty of Alliance of France and the
United States was communicated to England. This ends with the
Preliminarv Treaty with the United States signed in Paris on November 30,
1782.
if one looks only at the levels. The first differences, however, raise more
interesting questions about the sources of the varied patterns that occurred.
The striking thing is the absence of any trend in any of the three stocks
toward closer integration over the course of the century.
-- TABLE 1 ABOUT HERE --
The apparently close and stable integration of these two capital
markets will not be surprising to 18th century historians, enamored as they
are by the leisurely modernity of the Age of Enlightenment. Mail packet
boats left London regularly for Amsterdam on a twice-weekly basis and
dispatches from London four to seven days old appeared each week in the
Amsterdam and other Dutch newspapers. Dutch investors were represented
among major holders of Bank of England stock from the beginning. Their
holdings grew until 1751 with a "Dutch rush" occurring between 1721 and
1726. They continued to receive dividends and capital bonuses even when
the Netherlands had become the Republic of Batavia in 1795 and then
departments within the Napoleonic empire in the first decade of the 19th
century 10
How large were these investments by the Dutch in the English public
debt? In the most recent summary of the available studies, Wilson relies on
John Clapham, The Bank of England, A History, v. I, "1694-1797"
London: Macmillan, 1945, pp. 278-289.
Dickson's benchmark figures for 1723-24 and 1 750.11 Dickson found that in
1723-24 the total foreign holdings of stock in the "Big Three" companies
amounted to 9.3 percent of the total capital while by 1750 the total of
foreign holdings in the same companies (by now South Sea annuities had
replaced the original stock) amounted to 19.2 percent.12 Looking more
closely at the growth of foreign holdings in Bank of England stock,
Dickson found that it was not until the South Sea Bubble that foreigners,
especially the Dutch, began to invest. He concluded that
As a result of the South Sea Bubble this trend was markedly
accentuated [towards holdings of government and company
stock]. By 1723-24 foreign holdings of English government
securities had reached -- for the first time -- a really
substantial size. They were to go on increasing in amount
until the massive foreign disinvestment of the last twenty
years of the eighteenth century.13
Even though the result of modern scholarship has been to reduce the
proportions of English Funds held by foreigners from the heights of 40% or
more reported to Lord North in 1776, 14 the new lower bounds that have
been established still exceed the proportions of foreign trade and foreign
Charles Wilson, "Dutch Investment in Britain in the 1 7th- 19th
Centuries," in Credit Communal de Belgique, Collection Histoire Pro Civitate,
no. 58, "La Dette Publique aux XVIIIe et XIXe Sieclcs." (1980), p. 201.
12
P.G.M. Dickson, The Financial Revolution in England: A Study in the
Development of the Public Credit. 1688-1756. London: Macmillan, 1967, p. 312
and p. 321.
13 Ibid., pp. 311-12.
14 Alice Carter, "The Dutch and the English Public Debt in 1777,"
Economica, 20 (1953), pp. 159-161.
labor in the English economy of the eighteenth century.15 The evidence is
persuasive that economic integration between these two great mercantile
powers occurred first through the movements of capital.
II
15 N.F.R. Crafts, "British Economic Growth, 1700-1813: A Review of the
Evidence," Economic History Review, 36 (1983), p. 197 puts exports as a
proportion of national income for Great Britain between 8 and 12 % during
the eighteenth century and not over 15% until 1801. Neil Tranter, "The
labour supply, 1780-1860," in D. N. McCloskey and R. Floud, eds., The
Economic History of Britain since 1700, v. I (1981), p. 211, puts the maximum
share of foreign immigrants, mainly Irish, in the labor force at much lower
levels.
The more interesting questions of the ebb and flow of Dutch
investments in the English securities, particularly whether they were
destabilizing and speculative as contemporary English opinion had it, or
whether they were passive and on the whole stabilizing, remain unanswered.
Evidence to deal with these questions lies primarily in the price data for
the two markets described above. It should be noted, however, that the
Bank of England was responsible for recording transfers of ownership in
"government stock" which included shares of the Bank itself, the East India
Company, annuities, and Consols. In principle, these records could be used
to link foreign movements of capital to sustained rises or falls in the market
price of British company shares. To date the only use of these records has
been by Carter and then for only a three month period at the beginning of
11
1755. 16 The best that can be done at present is to search for evidence that
either of the markets for English securities, the Amsterdam and the London
stock exchanges, were less than efficient in setting prices.
Table 2 gives the initial results of time series analysis on the four
main time series of interest -- the cash prices for Bank of England and East
India Company stock quoted in London, and the forward prices for the
same two stocks quoted in Amsterdam. Basically, we are interested in
testing the proposition that the following equations are accurate descriptions
of price movements in each case:
1)Lit+l ■ Lit " ut+l ; i - 1. 3; t - time
2)Ait+l " Ait = ut+l ; i ■ 1» 3; t - time
where L is the London price and A is the Amsterdam price.
The standard technique is to estimate AutoRegressive, Integrated,
Moving Average (ARIMA) models for the changes in prices. If it can be
found that some combination of autogressive and moving average processes
yield consistently good descriptions of price changes, then presumably these
processes could be discovered by interested speculators and used to make
profits in the markets. For efficient markets to have existed, these models
should show (0,0) — i.e., that last period's price alone remains the best
Alice Clare Carter, "Transfer of Certain Public Debt Stocks in the
London Money Market from 1 January to 31 March 1755," Bulletin of the
Institute of Historical Research, 28 (November 1955), p. 203.
predictor of this period's price. The results of two different techniques for
estimating ARMA models are shown in Table 2.17
-- TABLE 2 ABOUT HERE --
For the period as a whole, both methods are consistent in showing
market efficiency in both markets for Bank of England stock and in the
London market for East India Company stock. Both methods indicate that a
(0,3), i.e., a third-order moving-average process, existed in the Amsterdam
prices of the East India Company. This appears to be in place only after
1737, however. The remainder of the table shows the estimated ARMA
models for each period of war and peace after Barnard's Act. Here the
results are mixed, although most cases are clear random walks, i.e., ARMA
(0,0). But it is only in the London market for East India stock that we find
consistent evidence in each sub-period for an efficient market. In the two
17
Both the "B-J" and "H-R" methods are strictly mechanical procedures
that estimate the autocorrelation coefficients and the partial correlation
coefficients for up to 10 lags over the time series. The "B-J", or standard
Box-Jenkins method, determines which of these coefficients are statistically
significant from zero and then the investigator selects the most plausible
model. (Box and Jenkins, Time Series Analysis: Forecasting and Control San
Francisco: Holden-Day, 1976.) The "H-R" method is a recursive process in
which each autoregressive process up to order 10 is estimated and then the
residuals of each estimate are used to calculate variances. The process
which minimizes the expression:
o2k + 2k/n
where k is the order of autoregressive process and n is the number of
observations. Then, using ordinary least squares regressions, ARMA models
are estimated up to order (p,5) where p is the order of AR model selected
above. Finally, the residuals of the estimated ARMA models are used to
calculate sample variances and (p,q) selected which minimizes:
log (o2p,q) + (log n/n)(p + q)
where q is the order of moving average process. (Hannan and Rissanen,
Biometnka, 1983.) The tests are repeated or various subperiods,
corresponding to the peacetime and wartime periods selected above.
13
markets for Bank of England stock we find sub-periods when some kind of
ARMA process seems to have been at work. It is interesting, however, that
these never occur in both markets for any given sub-period. The only sub-
period when a (0,3) ARMA appears for East India Company stock on the
Amsterdam exchange is 1739-1748, the War of the Austrian Succession. This
is also when an anomalous (3,0) ARMA is found for Bank of England stock
on the London market. Since only the coefficient on the third term was
found to be significantly different from zero in each case, some data error
was suspected. None was found, but another possible source of the anomaly
was detected.
Table 3 shows the average length of time interval between each
Amsterdam quote from the Amsterdamsche Courant. as well as the variance
of the intervals and the number of observations in each wartime and
peacetime period. The mean ranges from 15 to 16 days, as one expects for a
sample taken fortnightly with occasional gaps for religious holidays. But
the variance is especially high for the War of the Austrian Succession and
the peacetime between the Seven Years War and the War of American
Independence. It is in these periods that three of the six exceptions to
random walks occur. This is reassuring for the validity of the remaining
(0,0) processes since if a regular time series process exists but is sampled at
irregular intervals, the bias will be to find a random walk process. If a
random walk exists in a series, on the other hand, but is sampled at
irregular intervals, the possibility of finding a deterministic process arises.18
To reach a preliminary conclusion, efficient markets for the leading
British financial securities appear to be in place in both Amsterdam and
London after the South Sea Bubble of 1720. Moreover, they seem to have
continued to operate efficiently up to the outbreak of the French
Revolutionary Wars near the end of the century. Various episodes of
market inefficiencies leading to speculative profit possibilities probably did
arise at times but they appear to have been confined to the Amsterdam
market. These periods merit closer examination, especially in terms of
possible differences in the way each market operated.
Ill
The key to understanding the (0,1) and (1,0) processes found above
lies in the finding that the London prices were spot, or money, prices while
the Amsterdam prices were forward, or time, prices. The London practice
likely arose as a matter of convenience since no fixed settlement days
among brokers existed originally and time contracts could vary widely. But
in 1734, Barnard's Act (7 Geo. II, cap. 8) forbade all dealings in options and
18 Time-series purists will object that any irregularity in the timing of the
observations violates the assumptions of the statistical technique and so the
irregular appearance of the Courant rules it out for time-series analysis
Given the frequency of religious holidays in the eighteenth century even in
Protestant Amsterdam and London, however, a case can be made that the
irregularities in its appearance reflect precisely irregularities in trading
activity on the Effectenbeurs. It is the market activity, after all that is the
underlying process, not the appearance of the newspaper
15
future deliveries of stocks, with a fine of 500 pounds to be levied on each
person party to such a contract.19
This latter act was persistently violated in fact, according to
Mortimer.20 Dickson, on the other hand, feels that the Act may have been
effective in transferring options business to Amsterdam and encouraging
London traders to deal on margins.21 Castaing's price lists was consistent in
showing prices at money although Cope speculates that the curious practice
of printing the names of the Bank of England and the East India Company
in capital letters may have developed to indicate those securities in which
dealings in time may have been possible.22 Since the transfers of stocks for
the chartered companies came to be handled through the Bank of England,
purchases could in principle be made for a forward date at which the
transfer could be made at the Bank. The illegality of dealings in futures
and options may not have eliminated the practice in the London market —
Malachy Postlethwayt, "Stock-Jobbing", in The Universal Dictionary of
Trade and Commerce, 2 vols., 4th edn. (1774), repr. 1971.
Thomas Mortimer, Every Man His Own Broker: Or a Guide to Exchange
Alley, 3rd ed. (1761).
Dickson (1967), p. 508, quotes a letter by an Amsterdam broker to a
Haarlem merchant in 1735 that states "...in London only cash purchases and
sales can be made."
S. R. Cope, "The Stock Exchange Revisited: A New Look at the Market
in Securities in London in the Eighteenth Century," Economica, 45 (February
1978), p. 18. On this point it is interesting that Houghton followed the same
practice in his early listings of stock but clearly stated that the companies
whose names were printed in "Great Letters" were those that had charters
while those that had asterisks in front of them were patent monopolies.
John Houghton, A Collection for Improvement of Husbandry and Trade, 9
vols., London: Randall Taylor et. al., 1692-1703, republished Westmead,
Farnborough, Hants.: Gregg, 1969. The list of stocks in no. 106 (Fri., August
10, 1694) has note: "Great Letters by Charter, (*) by Patent."
indeed, the introduction of stiff er bills in the House of Commons in 1745,
1756, 1771, and 1773 may indicate the continued prevalence of futures
trading — but it no doubt was effective in eliminating the printed quotation
of future prices for those contracts that were made.
In Amsterdam, by contrast, the practice was always to deal in time
contracts since legally binding possession of shares in the Dutch East India
Company was not possible until the actual transfer of the share or shares
was entered in the Company's books. This was not possible until the books
were opened for the payment of dividends. De la Vega's original
description of the Amsterdam Beurs in fact describes "putts" and "refuses" in
very modern terms for options trading. The extensive trading of dealers
with one another on both hedging and speculative contracts in the same
stock required regular "rescounter" settlement dates to settle the net
differences and straighten out the accounts among the various brokers.
These occurred quarterly, on the fifteenth of February, May, August, and
November.23 The quarterly rescounters may have been for the English
funds only, since de la Vega reports monthly rescounters, on the 20th of the
month for real stock with payment due the 25th, and on the first of the
month for "ducaton" shares.24 ["Ducaton" shares were small fractions of
actual Dutch East India Company shares that were devised to increase the
Isaac de Pinto, Traite de la Circulation et du Credit, (1771), p. 305.
24 de la Vega, introd. by H. Kellenbenz, p. xviii.
17
possibilities for trading by smaller investors since the cash value of each
original share had increased over the decades to very high levels indeed.]
Van Dillen notes the difficulties in deciding whether the figures in
the Amsterdamsche Courant were cash or time prices:
Until 1747 this is not mentioned, but in comparing them with
those found in brokers' notes preserved from 1725 to 1737 it
appears that in that period the quotations are cash prices. In
the year 1737 both prices are sometimes mentioned. After this
year we find generally the forward rates. From 1759 onwards
the quotations are often followed by the name of the next
settlement month, e.g., "all of February." The difference
between the cash price and the next paying month is,
however, not more than a few percent.2
If the Amsterdam prices quoted on the English securities were for
future delivery, then in general they should lie above the London cash
prices quoted on the same day. Chart 1 illustrates why.26 At regular
intervals, dividends are paid on each of our securities. If nothing else
happened to disturb the price of the shares from time O to time A on the
graph, the nominal value of each share would be fixed until the dividend
was paid, at which time the value would rise abruptly. Cash transactions in
the shares between time O and time A will take into account the
forthcoming dividend payment which the buyer of the share will receive.
So the cash prices between time O and time A will show a gradual upward
trend along line OB. A contract made at time O for future delivery of the
25 Van Dillen (1931), p. 13.
26
Louis Bachelier, "Theory of Speculation," trans, in Paul Cootner, ed.,
The Random Character of Stock Market Prices, (1964).
share at time A, however, will require the buyer to pay a "contango" to the
seller, equal in the absence of disturbances in the price of the share to the
dividend. This arises since the seller will hold the share until the delivery
date but will then yield possession of the stock, and its dividend, to the
buyer who only then will make full payment. This means that the futures
price equivalent for the cash price that runs along line OB will be line CB,
which always lies above the cash price but gradually converges to it at
dividend payment dates.
- CHART 1 ABOUT HERE --
If Bachelier's exposition explains as well the differences between
Amsterdam and London prices, then the Amsterdam prices should be the
same as the London prices with only small random disturbances until
Barnard's Act in 1734 or until 1737 when Barnard's Act was made a
perpetual law. Tables 4 and 5 present regression results for linear
regressions of the difference between the Amsterdam and London price at a
given date on three variables: 1) DAYSDIVD, the number of days from the
date of the observation to payment of the next dividend; 2) PAYTIME, a
dummy variable set equal to one during the times London prices were
quoted ex dividend; and 3) AMEXPREM, the number of penningen banco
the English pound sterling was worth less its mint par ratio. Only the Bank
of England and East India Company stocks are analyzed since the South Sea
Company stock was essentially dormant for most of the period after 1730.
-- TABLES 4 AND 5 ABOUT HERE --
The variable DAYSDIVD is intended to capture the Le Bachelier
effect. Its coefficient should be positive since the dependent variable is the
Amsterdam price minus the London price. On average, it should be equal to
one-half the annual dividend since dividends were paid semi-annually. For
each sub-period, its effect is estimated separately in the third equation. The
second equation in each panel adds the effect of AMEXPREM while the
first equation has all three explanatory variables. Comparing the third
equation in each panel over the two tables, one notices that only in the pre-
Barnard Act period does the constant term in the regressions become
statistically different from zero, and it does so for both the Bank of
England and the East India Company. For the entire period 1738-1794 and
for the three peace-time periods within, the constant terms are
insignificantly different from zero. This implies both that the contango
rate was on average the same as the dividend rate, which we should expect
in the absence of persistent expectations for things to improve or to
deteriorate, and that no serious barriers existed to equalizing the rate of
return on the same financial assets in the two different countries.
The presence of a constant term that is negative and significantly
different from zero in the pre-Barnard Act period could imply segmented
capital markets or exuberant outlooks by speculators, if in fact the
Amsterdam prices were forward prices consistently. Inspecting the pattern
of residuals and estimating the regression for subperiods within the period
1723-1738 leads me to believe that this was the case. On average the
Amsterdam price was higher than the London price even in the period 1723-
37, although the difference was much less than it became after 1737. This
holds for both stocks. If this is true, then the negative constant term
implies optimistic expectations by investors during this period. The
evidence of the exchange rate variable presented below strengthens this
presumption.
There remain differences between the regression estimates for the
Bank of England stock and the East India Company stock. On average, the
price difference was 1.6 points for Bank stock and 2.5 points for East India
stock. This reflects the generally higher dividend rates paid by the East
Company stock. In sum, these regression results, combined with the
evidence of extremely tight market integration presented in the
introduction, demonstrate that what small, but persistent differences in
prices remained between the Amsterdam and London markets for the British
securities were due to the London prices being cash, or spot, prices while
the Amsterdam prices were forward prices.
When the semi-annual dividend payment dates approached, the
transfer books for the particular stock would be closed so that the sums due
to each owner could be calculated and made ready. During this period
which usually lasted two weeks, the stock would be quoted ex dividend.
Any deliveries of stock taking place during that period, then, would not
include the dividend about to be paid since the clerks would be in the
process of making the payment ready for the currently registered owner. So
the variable DAYSDIVD was calculated as the number of days from the
21
given date to the day the first ex dividend quote appeared in the Course of
the Exchange. In Amsterdam, on the other hand, it appears that the quoted
price was always with dividend. In the printed form shown by Dickson for
sales of stock made in Amsterdam for delivery (and payment) in London,
explicit provision is made that if the receiver of the stock does not get the
current dividend, then he deducts that dividend from the stated price he
has agreed to pay. This is stated clearly in a contract dated 4 April 1730
between Jacob Reynst and David Leeuw, both of Amsterdam, that Dickson
translates as follows:
I the undersigned acknowledge to have Bought from Heer
David Leeuw One Thousand Pounds Sterling Capital Shares of
the Bank of England at London, at a price of a Hundred and
Forty Five and a Quarter per Cent remaining after the
Dividend paid last October, for settlement on next 15 May, the
which 11000 I oblige myself to receive in London at the stated
Price. And in case in the interim any Dividend is paid, it
shall be to my profit and to reduction of the above Price.
Contrarywise all supplementations and Calls shall be at my
expense, in the usual way. All done in good faith at
Amsterdam the Fourth April Seventeen hundred and thirty.
To capture this difference in practice in quoting prices during the
period dividends were being calculated, a dummy variable was created,
PAYTIME, which was set to a value of one for the first observation after
the ex dividend quotes began in London. It proves to be positive, as
expected, and usually significant, especially in the earlier years. It also has
the felicitous effect of reducing substantially the serial correlation in the
regressions.
Dickson, (1967), p. 335. This was a pre-printed form except for those
items shown in italics, which were entered by hand.
Since transfers of stock had to take place in London; where the
actual stock had to be paid for in pounds sterling, we should expect some
effect in the Amsterdam market from fluctuations in the exchange rate
with London. Taking deviations in the observed sight rate from the mint
par ratio as the measure of changes in the exchange rate (AMEXPREM),
what effect should one expect on the Amsterdam price of an English stock
(in English pounds sterling) of, say, an increase in the value of the English
pound relative to the Dutch guilder? Recognizing that any effect will be
merely transitory until prices are equalized on the basis of the new
exchange rate, one might expect the demand for English stock in
Amsterdam to be shifted downward (any given price in pounds sterling is
now felt to be more expensive by a Dutch purchaser) while the supply of
the English stock in Amsterdam would be shifted outward (any given price
in pounds sterling is now more attractive to a Dutch supplier). The effect
of these shifts in both demand and supply is to reduce the price of the
English stock in Amsterdam relative to its price in London. Since the
dependent variable in the regressions is the Amsterdam price minus the
London price while the independent variable for the effect of the exchange
rate is the price of the pound sterling in terms of Dutch bank money less
the mint par ratio, the expected sign on the exchange rate variable is
negative. That is, the higher the value of the pound on the foreign
exchanges, the lower we expect the price of a British security to be in
Amsterdam relative to London.
23
In fact, the estimated coefficient for the exchange rate variable does
not prove to be statistically different from zero in the earliest period and
when it does become significant it has a positive sign! After mid-century,
however, it always has the expected sign and is often significantly different
from zero as well, especially in the regressions from which outliers have
been removed. The anomaly that needs to be explained then is the positive
sign and significant effect before 1750, especially during the War of the
Austrian Succession. The most likely explanation stems from the fact that
the period from 1723 to 1750 was precisely the period when the Dutch built
up their holdings in the English joint stock companies and long term
government debt most rapidly. Dickson found that Dutch holdings of Bank
of England stock rose from 10.5% of the total capital in 1723-4 to 30.3% in
1750 while their share of East India Company stock rose from 13.4% to
21.4% over the same period.28 It would appear, then, that in this period the
massive inflows of Dutch capital to the English long-term securities market
were sufficient to drive up the value of the English pound whenever surges
of Dutch demand lifted the Amsterdam price above its predicted level,
based on the purely technical factors embodied in our variables PAYTIME
and DAYSDIVD.
The final step taken in the regression analysis was to remove outliers.
Analyzed the graphs of predicted vs. actual values of the Amsterdam-
London price differences revealed a very close clustering of the actual
28 Dickson, (1967). Calculated from Tables 47 and 48, pp. 312-13 for
1723-4 and from Tables 50 and 51, p. 321 and p. 324, for 1750.
1J
values to the regression plane with a few (less than 2%) of the observations
causing much of the unexplained variance. Setting these equal to the
predicted value (or the actual value in the few cases where it turned out
that an error had been made in the data entry), improved the goodness-of-
fit greatly, without altering the size or significance of the estimated
coefficients. (An exception is the AMEXPREM variable which did prove
responsive in some time periods to the removal of outliers.) Some of the
outliers may be due to errors of transcription in the original data source.
Some of the numbers in Van Dillen's table, for example, appear to be in
error — two digits are reversed or two columns are reversed. But most of
the anomalies in Van Dillen appear to occur in the Amsterdamsche Courant
as well.29
This leads us to confront the second source of transcription error —
that from the original source on the actual trading day in London or
Amsterdam to the printed source used now by historians. This must remain
a matter of speculation, but extensive use of the Course of the Exchange by
the author has not revealed other examples of typographical errors — e.g.,
inconsistencies in the date headings which had to be altered with each issue
or reversal of data entries. But it is true these would be easier to pick up
by the original typesetters as well.
29
The ten most dubious numbers in Van Dillen were checked against
xerox copies of the Amsterdamsche Courant obtained from the New York
Public Library. Only three, those for East India stock on November 16,
1733 and April 3, 1789 and that for South Sea stock on November 25, 1793
proved to be Van Dillen's mistake in transcribing.
25
Studying the pattern of residuals for the period 1763-1778, when the
largest number of outliers occurred, suggests another explanation. The
outliers in this period appeared as a result of sharp movements of very
short duration on the London Stock Exchange. If one had taken the
observation for London of three days previous, for example, in most cases
the difference from the Amsterdam quote would have been largely
eliminated. If the next observation was also an outlier, it was nearly always
of opposite sign. This suggests that information of great influence on the
price of English stocks was reaching one market well before it reached the
other. Since our observations are taken from the same day in both markets,
ephemeral information of the kind associated with panics (the panics of
1763 and 1772 occur in this period) that reaches one market before the
other will not be reflected in the price difference. In anything other than
panic situations, this is not a problem since the Amsterdam prices are taken
only every two weeks and most of the information flow that has occurred
in that time interval will also have reached the London market. It is not
surprising then that the worst goodness-of-fit occurs in the period 1763-1778
and again in the period 1790-1794.
IV
In sum, the regression results strengthen the conclusion that these
two markets were efficient and well-integrated from the second quarter of
the eighteenth century on. The spot prices quoted in the London market
followed a random walk process consistent with efficient markets. The
forward prices in the Amsterdam market were highly correlated with them.
The semi-annual dividend payments in London then produced regular
patterns in the Amsterdam prices both from the Le Bachelier effect and the
practice of London quoting prices ex dividend during the preparation of
dividend payments while Amsterdam quoted them with dividend. These
regular patterns, however, were sometimes masked by unusual expectations
(e.g., during the War of the Austrian Succession), fluctuations in exchange
rates, or panics. In those periods we find random walks in general,
although the War of the Austrian Succession remains an anomaly.
The significance of these findings will be seen differently by
economists and historians. For historians, they will simply confirm in large
part the authoritative work of P.G.M. Dickson on the operation of the
London capital markets in this period and his perceptive comments on the
Dutch influence. Dickson's work, in turn, relied heavily on the earlier work
done by Carter, van Dillen and Wilson.30 All are agreed on the importance
and effectiveness of the integration of these two capital markets at the
dawn of modern capitalims. After all, even Karl Marx raged at the
Alice Carter, "Dutch Foreign Investment, 1738-1800," Economica, 20
(1953), pp. 322-340; "Dutch Foreign Investment, 1738-1800, in the Light of
the Amsterdam 'Collateral Succesion' Inventories," Tijdschrift voor
Geschiednis, (1953), pp. 27-38; and "The Huguenot Contribution to the Early
Years of the Funded Debt, 1694-1714," Proceedings of the Huguenot Society
of London, 19 (1955), p. 21. All these are reprinted and summarized in her
book, Getting, Spending and Investing in Early Modern Times (1975). Cf.
Charles Wilson, Anglo-Dutch Commerce and Finance in the Eighteenth Century,
Cambridge: University Press, 1941, and his "Dutch Investment in Eighteenth
Century England," Economic History Review, 2nd series, 12 (1959), 434-439.
Van Dillen's works have been cited above.
27
establishment of the National Debt in England as the single most effective
device in the "primitive accumulation of capital".
For economists, the quantitative results should strengthen their
confidence in doing analytical work on the financial relations between
London and Amsterdam.32 The pioneer work in this regard was done by
Eagly and Smith who used interest rate and foreign exchange rate data in
the framework of an interest rate arbitrage model. Their results, according
to them, "support the general hypothesis of a trend towards increased
integration among money markets during the course of the century, but at
the same time they show that during individual sub-periods there was
considerable variation in this trend."33 This could as well be the conclusion
from the results above but my emphasis would be on the high level of
integration that existed at the beginning of the second quarter of the
eighteenth century and the continued efficiency of the operations of both
stock markets rather than a trend toward improved integration.
31 Karl Marx, Capital, London: Lawrence & Wishart, 1970, v. I, Ch. 31,
pp. 754-6.
Two works should be mentioned. Brian Parsons, "The Behavior of
Prices on the London Stock Market in the Early Eighteenth Century,"
unpublished Ph.D. dissertation, University of Chicago, 1974 concentrates on
the daily course of prices during the South Sea Bubble and finds the market
operated efficiently although he finds unexplained differences in the price
quotes from different sources. Philip Mirowski, "The Rise (and Retreat) of
a Market: English Joint Stock Shares in the Eighteenth Century," Journal of
Economic History, 41 (September 1981), 559-577, compares internal accounts
of profitability for several joint stock companies with the pricing of their
equity on the stock market. He finds increasing discrepancies in the latter
part of the century.
Robert Eagly and V. Kerry Smith, "Domestic and International
Integration of the London Money Market, 1731-1789," Journal of Economic
History, 36 (March 1976), p. 210.
True, these operations were disturbed by repeated wars -- each
fought with different financial techniques and consequences, by trade
disturbances, and by occasional financial panics. Nevertheless, it must be
emphasized that these disturbances never changed the fundamental
economic characteristics of the two markets in the way that modern policy
measures of sovereign nation states manage to do. There were no attempts
at independent national monetary policy by either country in the period
1723-1794 — i.e., no controls on capital movements, no withholding taxes on
dividend or interest payments to foreigners, no changes in monetary
standards, no managed exchange rates, and no attempts to regulate Ml, M2,
or M3. There was an absence, therefore, of the modern impediments to
efficient operation of multiple listing markets. (Witness the current
difficulties in expanding multiple listings of stocks among the members of
the European Community.) In other words, a remarkably modern pair of
capital markets were permitted to interact in an unfettered (and hence
unmodern) fashion. This may be why economic integration occurred first in
these capital markets well before comparable degrees of integration could be
achieved in goods markets, much less in labor markets.
29
Table 1.
Correlation Coefficients between London and
Amsterdam Prices for Stock of Bank of
England, South Sea, and East India Company
for 1723-1794 and Various Sub-periods of War
and Peace, (levels and first differences of
actual prices)
Period Bank o
1723-1794
(changes)
Peace
8/09/23 - 10/19/39
(changes)
11/11/48 - 7/14/56
(changes)
2/18/63 - 3/04/78
(changes)
12/06/82 - 9/22/90
(changes)
War
10/21/39 - 10/23/48
(changes)
8/04/56 - 2/05/63
(changes)
3/02/78 - 11/20/82
(changes)
10/08/90 - 12/19/94
(changes)
England
East India
South Sea
0.994
0.993
0.989
0.589
0.624
0.394
0.978
0.990
0.951
0.381
0.544
0.505
0.983
0.988
0.983
0.327
0.370
0.361
0.993
0.997
0.974
0.656
0.716
0.389
0.996
0.988
0.969
0.570
0.578
0.153
0.988
0.977
0.954
0.536
0.593
0.305
0.976
0.963
0.979
0.655
0.610
0.408
0.823
0.943
0.907
0.465
0.536
-0.004
0.988
0.984
0.986
0.803
0.644
0.664
Table 2.
ESTIMATED ARMA MODELS FOR BANK OF ENGLAND AND EAST
INDIA COMPANY STOCK PRICE CHANGES IN LONDON AND
AMSTERDAM
Time Period
Bank of England
London
H-R
B-J
Amsterdam
H-R B-J
Entire Period
1723-1794
0,0 0,0
0,0 0,0
Pre- and Post-Barnard
1723-1737
0,0
0,0
1738-1794
0,0
0,0
Peace Periods
1748-1756
0,5
0,0
1763-1778
2,0
2,0
1782-1792
0,0
0,0
War Periods
1739-1748
3,0
3,0
1756-1763
0,0
0,0
1778-1782
0,0
0,0
0,0
0,0
0,1
0,0
0,0
0,0
0,0
0,1
0,0
0,0
0,1
0,0
0,0
0,0
0,0
1,0
31
Table 2 (cont.)
East India Company
Time Period London Amsterdam
H-R B-J H-R B-J
Entire Period
1723-1794 0,0 0,0 0,3 0,3
Pre- and Post-Barnard
1723-1737
0,0
0,0
1738-1794
0,0
0,0
Peace Periods
1748-1756
0,0
0,0
1763-1778
0,0
0,0
1782-1792
0,0
0,0
War Periods
1739-1748
0,0
0,0
1756-1763
0,0
0,0
1778-1782
0,0
0,0
2,0 0,0
3,0 3,0
0,0 0,0
0,0 0,0
0,0 0,1
0,3 0,3
0,0 0,0
0,1 0,1
Table 3.
Regularity of Observations from
the Aitisterdamsche Courant,
Sub-Periods of Peace and War, 1723-1792.
Period Mean Variance No. of obs
1723-39 15.1 28.55 391
1739-48 16.01 43.87 207
1748-56 15.35 19.05 185
1756-63 15.72 25.90 153
1763-78 15,94 97.34 357
1778-82 14.98 10.16 115
1782-92 15.17 21.94 189
33
Table 4.
Regression Results for the Bank of England,
Amsterdam-London Price Differences Explained
Summary of
1723-1794;
with:
1: Days to Next
2 : Changes in the
3 : Whether the
Dividend Payment (DAYSDIVD) ;
Exchange Rate (AMEXPREM) ; and
London Price was with or
ex dividend (PAYTIME) .
[Outliers removed and set to regression plane.]
Subperiods
1723-1739
1739-1748
1748-1756
1756-1763
1763-1778
1778-1783
1783-1790
1790-1794
[Peace, Pre-Barnard]
[War of Austrian Succession]
[Peace, no financial crises]
[Seven Years' War]
[Peace, Panics of 1763 and 1772]
[War of American Independence]
[Peace]
[French Revolution, start of war]
(A-L)DIF DAYSDIVD AMEXPREM PAYTIME CONST R2
DW
PANEL A: 1723-1739
OLS .005 .002 1.999 -.140 .30 1.78
(4.365) (.241) (8.839) (-.520) (1.083) 387
OLS
.010
(8.772)
.003
(.308)
-.397 .16 1.63
(-1.348) 388
OLS
.009
(8.780)
-.480 .16 1.63
(-4.006) 389
Panel B: 1739-1748
OLS .009 .056 1.141 1.284
(5.343) (4.952) (3.549) (4.016)
31
1.55
203
OLS
.012
(7.862)
.056
(4.843)
1.118
(3.436)
28
1.40
204
OLS
.011
(7.164)
-.233
(-1.32)
20
[T-statistics are in parentheses under respective
coefficients]
1.25
205
.010
.003
(7.15)
(.245)
.012
.002
(9.39)
(.159)
.012
(9.42)
Table 4. (cont.)
(A-L)DIF DAYSDIVD AMEXPREM PAYTIME CONST R2 DW
Panel C: 1748-1756
OLS .010 .003 .758 -.137 .34 1.51
(2.728) (-.468) 181
-.275 .32 1.32
(-.936) 182
-.316 .32 1.32
(-2.18) 183
Panel D: 1756-1763
OLS .011 -.014 .515 -.193 .19 1.40
(1.139) (-.495) 149
-.250 .19 1.38
(-.646) 150
.071 .19 1.36
(.317) 151
Panel E: 1763-1778
OLS .016 -.008 .243 -.537 .32 1.44
(.869) (-2.30) 353
-.577 .32 1.42
(-2.52) 354
-.353 .32 1.42
(-2.555) 355
.011
-.014
(4.69)
(-1.078)
.012
-.013
(5.95)
(-1.02)
.012
(5.99)
.016
-.008
(10.95)
(-1.206)
.017
-.009
(13.04)
(-1.222)
.017
(13.07)
35
Table 4. (cont.)
(A-L)DIF DAYSDIVD AMEXPREM PAYTIME CONST R2 DW
Panel F:
1778-1782
OLS
.010
-.025
.567
-.14
.21
1.60
(3.79)
(-2.683)
(1.102)
(-.428)
111
OLS
.011
-.025
-.223
.21
1.53
(4.95)
(-2.67)
(-.676)
112
OLS
.011
(4.82)
.364
(1.442)
.16
1.42
113
Panel G: 1782-1790
OLS
OLS
OLS
.017
-.033
.492
-.122
.35
1.72
(7.56)
(-4.31)
(1.105)
(-.520)
185
.018
-.032
-.196
.35
1.70
(9.08)
(-4.280)
(-.870)
186
.019
-.166
.29
1.55
(8.90)
(-.706)
187
Panel H: 1790-1794
OLS .009 -.002 1.761 .568 .20 1.58
(2.21) (-.179) (2.462) (1.14) 75
OLS .015 -.002 .253 .15 1.44
(3.90) (-.165) (.509) 76
OLS .015 .207 .16 1.44
(3.93) (.504) 77
TABLE 5.
Summary of Regression Results for the East India
Company explaining Amsterdam-London price differences
(A-LDIF) with:
1: Day to next dividend payment (DAYSDIVD) ;
2 : Changes in the exchange rate (AMEXPREM) ; and
3 : Whether the London price was with or ex
dividend (PAYTIME) .
[Outliers removed and set to regression plane.]
Subperiods:
1723-1739 [Peace, pre-Barnard]
1739-1748 [War of Austrian Succession]
1748-1756 [Peace, no financial crises]
1756-1763 [Seven Years War]
1763-1778 [Peace, panics of 1763 and 1772]
1778-1783 [War of American Independence]
1783-1790 [Peace]
1790-1794 [French Revolution, start of war]
(A-L)DIF DAYSDIVD AMEXPREM PAYTIME CONST R2 DW
PANEL A: 1723-1739
OLS
OLS
OLS
.006
-.001
2.884
-.477
.24
1.90
(3.214)
(-.106)
(7.930)
(-1.07)
387
.012
-.005
-1.00
.12
1.69
(7.28)
(-.342)
(-2.102)
388
.012
-.851
.12
1.69
(7.29)
(-4.43)
389
Panel B: 1739-1748
OLS .015 .068 1.383 1.251 .29 1.54
203
.27 1.38
204
.22 1.28
205
[T-statistics are in parentheses under respective
coefficients]
.015
.068
1.383
1.251
(5.74)
(3.95)
(2.751)
(2.52)
.018
.069
1.072
(7.93)
(3.945)
(2.14)
.018
-.633
(7.61)
(-2.44)
37
Table 5. (cont.)
(A-L)DIF DAYSDIVD AMEXPREM PAYTIME CONST R2 DW
Panel C: 1748-1756
OLS
OLS
OLS
.016
.028
2.032
.427
.41
1.69
(7.05)
(1.40)
(4.51)
(.927)
181
.016
2.024
-.137
.41
1.68
(7.07)
(4.48)
(-.62)
182
.021
-.416
.35
1.40
(9.91)
(-1.85)
183
Panel D: 1756-1763
OLS .013 -.044 .872 -.800 .27 1.43
(5.15) (-2.99) (1.72) (-1.78) 149
OLS .015 -.046 -.959 .26 1.38
(6.69) (-3.07) (-2.16) 150
OLS .015 .156 .22 1.29
(6.56) (.599) 151
Panel E: 1763-1778
OLS .021 -.063 .835 -1.42 .19 1.56
(6.57) (-3.828) (1.35) (-2.75) 353
OLS .023 -.063 -1.55 .19 1.55
(-3.03) 354
.023
-.063
(8.17)
(-3.84)
.023
(8.15)
OLS .023 .016 .16 1.49
(.053) 355
Table 5. (cont.)
(A-L)DIF DAYSDIVD AMEXPREM PAYTIME CONST R2 DW
Panel F: 1778-1783
OLS
OLS
OLS
.028
-.036
.684 -1.40
.41
1.93
(7.16)
(-2.71)
(.913) (-2.83)
111
.029
-.036
-1.51
.41
1.91
(8.64)
(-2.755)
(-3.177)
112
.029
-.66
.38
1.77
(8.38)
(-1.78)
113
Panel G: 1782-1790
OLS
OLS
OLS
.020
-.028
.839
.359
.33
1.79
(7.33)
(-2.987)
(1.552)
(1.34)
185
.022
-.028
•
.244
.32
1.73
(9.15)
(-3.005)
(.942)
186
.022
.320
.29
1.65
(8.90)
(1.215)
187
Panel H: 1790-1794
OLS .016 -.010 1.776 .709 .19 1.69
75
.17 1.66
76
.18 1.65
77
.016
-.010
1.776
.709
(2.82)
(-.501)
(0.707)
(1.11)
.021
-.012
.495
(4.24)
(-.600)
(.779)
.021
.267
(4.265)
(.527)
Chart 1
Equivalent Prices of Spot and Future Deliveries
Time
I