Rates of interest in 18th century England

Rates of interest in 18th century England

EXPLORATIONS IN ECONOMiC HISTORY 27, I-28 (19%) Rates of Interest in 18th Century England* KENNETH J. WEILLER? Citibank, N.A. AND PHILIP MIROWSK...

2MB Sizes 2 Downloads 20 Views

EXPLORATIONS

IN ECONOMiC

HISTORY

27, I-28 (19%)

Rates of Interest in 18th Century England* KENNETH

J. WEILLER?

Citibank, N.A. AND PHILIP MIROWSKI Tufts University We present a consistent, quarterly time series of English short- and long-term interest rates for the entire 18th century. Instability in the term structure of interest rates cannot be explained solely by wars or government intervention. Applying Shiller variance bounds tests, we find the rational expectations (RE) hypothesis rejected for the century and each subperiod. Using the modifications suggested by M. Flavin however, we reject the RE hypothesis less than half the time. On the other hand, regression tests on the relationships of short-term rates with holding period returns and long-term rates tend against the RE theory of the term structure. The evidence is also interpreted in terms of the 18th century debate between Adam Smith and Henry Thornton over the nature of interest rates. 0 199oh8dfkc press, 1~. Our island has been preserved, through the favour of Providence, from those violent convulsions which have been felt on the continent. We have, however, been exposed to many smaller evils, and in particular, to the interruption of our mercantile credit. Thornton, 1802, p. 319 I

The birth and development of the separate discipline of political economy in the 18th century owed much to the contemporary fascinations with banks, “funds of credit,” and interest rates. The respective fates * The authors thank Barry Eichengreen, Julian Hoppit, John James, Joel Mokyr, Larry Neal, Robert Shiller, Mark Watson, David Weiman, Michael Weinstein, and participants in seminars at Harvard University and the Social Science History 1983 meetings for helpful comments on this and earlier drafts. $ The views expressed herein do not necessarily reflect those of Citibank, Citicorp. or their staff.

0014-4983190 $3.00 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

2

WEILLER

AND

MIROWSKI

of the Banque Royale and the Bank of England seemed to contemporaries and their descendants to be fraught with implications for the political and economic fates of their nations. The course of interest rates were thought to be a barometer of national health, both in peace and in times of war. The need to explain those movements called forth efforts which then widened into the larger inquiry of political economy. In Great Britain, two broadly opposing views of the importance of money and finance were developed over the course of the century. One view, associated with David Hume and Adam Smith, portrayed money as a “great wheel of circulation,” in the long-run neutral with respect to the operation of the economy. Credit, in this scheme of things, was merely a different form of money which masked the real transactions which it represented. As Smith wrote, “By means of the loan, the lender, as it were, assigns to the borrower his right to a certain portion of the annual produce of the land and labor of the country to be employed as the borrower pleases” (Smith, 1976, i, p. 373). In general, credit expansion, and especially the expansion of public debt, was frowned upon; but the nature of the danger inherent in credit expansion was left to the reader’s imagination. Banks were viewed with suspicion because they encouraged “over-trading,” although this term was never defined. Movements in the rate of interest were posited as being controlled by movements in the real rate of profit. Above all, the prudent were exhorted to eschew the attractions of “the Daedalian wings of paper money,” and to remain upon “the solid ground of gold and silver” (Smith, 1976, i, p. 341). The alternative view, expressed by the early Henry Thornton, drawing upon strains of thought in John Law and James Steuart, denied the neutrality of money, and viewed the various forms of paper credit as important and useful tools which the State should employ in furthering economic development. These authors suggested that the level of real activity would be substantially altered by the availability of credit, but that the institutions of credit were themselves fragile and required the intervention of the state bank, often as lender of last resort. The state, by encouraging credit institutions, could then call upon them in times of need, such as wars: the state could create the conditions whereby a public debt could be a stabilizing influence in the market. Interest, for this school, was largely a monetary phenomenon, not necessarily linked to any given “real” rate of profit. For our present purposes, focusing on the question of the levels of interest rates and then on fluctuations will serve to illustrate the substantial theoretical differences between those two schools. The Hume/Smith school insisted that low and falling levels of interest were infallible indices of prosperity; but Hume admitted in an aside that “a sudden and great check to commerce,” would produce the same effect

INTEREST

RATES IN 18TH CENTURY

ENGLAND

3

“by throwing so many stocks out of trade.” In the manner of other optimistic assessmentsof this school, he then proceeds to state that such interruptions would be temporary, and ignores the implied paradox (Hume, 1970, pp. 55-56). The Steuart/Thornton school, on the other hand, stressed monetary influences, and therefore did not expect relatively low interest rates to necessarily imply relative prosperity. They did expect that the manipulation of the outstanding amounts of paper money would move interest rates in the opposite direction (here abstracting from the effects on international exchanges), and that “clogs laid upon circulation” would result in high interest rates, essentially due to the frustrated need for liquidity (Sen, 1957, p. 37). More than one historian of economic thought has seen the 20th century “loanable funds vs liquidity preference” controversy mirrored in these early controversies (Hicks, 1967, p. 177). While this dispute has intrinsic interest for specialists in the history of economic thought and the evolution of monetary theory, we would like to suggest it also merits the attention of economic historians of 18th century Britain, primarily because roughly the same issues have made their reappearance in the history literature under the topics of growth accounting and financial crises. In Williamson (1984, the claim is made that the slow growth of Britain in.the later 18th and early 19th centuries could be directly attributed to British government war finance. The large debt issues, he asserted, drove up interest rates and crowded out private investment, along the lines of a conventional neoclassical aggregate model of saving and investment. (See also Barro, 1987.) The attractions and drawbacks of this argument have been discussed in detail in (Heim and Mirowski, 1987). While Williamson’s thesis was explicitly predicated upon neoclassical theory, it is worth noting that there are some pronounced similarities with the Smith/Hume position: Government credit expansion is only thought to be detrimental to the economy; and most monetary disturbances are blamed upon government finance. Other similar Smithian themes may be found in the recent work of Hoppit (1986) on British fmancial crises in the 18th century. In contrast to Williamson, Hoppit does not propose that the entire 18th century can be explained by a single neoclassical theory; he acknowledges that a “financial crisis” does not admit of a single simple definition. Nevertheless? he does agree with Williamson that at least prior to 1770 government finance was periodically disruptive. He adopts a position very reminiscent to Smith’s “overtrading” for the period after 1770: “‘What distinguishes financial crises after 1770 from those before is that they were in large part caused by economic growth. Growth encouraged speculative business expansion funded by trade credit” (Hoppit, 1986, p. 51). Note that the causality does not generally run in the other direction. More striking than the fact that some of these issues are still contro-

4

WEILLER

AND

MIROWSKI

versial is the fact that neither the 18th century protagonists nor the 20th century economic historians have made much use of the available data on interest rates to gain some insight into these issues. The 18th century is a particularly fertile source for empirical study of finance because many of the complications which bedevil the study of interest rates in later centuries were absent then: there was no differential taxation of capital gains and interest income; and with the exception of the first and last decades, the period experienced less inflation or deflation than later centuries (Mirowski, 1987). Moreover, for nearly the entire century there exists a uniform and reliable source of daily data on prices and yields in the London money market, The Course of the Exchange. One major obstacle to empirical discussion of this controversy is the fact that both the 18th century protagonists and 20th century economic historians referred to “the” rate of interest, as if the ensemble of rates moved together (cf. Ashton, 1955, p. 28; Pressnell, 1956). With the exception of Pressnell (1960) there is no discussion of the term structure of interest rates; nor do Smith or Thornton distinguish between shortand long-term lending. However, much of the differences between the respective schools can be rendered more comprehensible if one notes that many of their strictures concerning the fluctuations of rates implicitly refer to differing maturities. Compounding this problem is the fact that in the 18th century there was an independent controversy as to whether the rate of interest on government funds adequately represented the reigning interest rate for private borrowing (Dickson, 1967, 479-481). Because the usury laws were still in force over the century, private loans could not charge in excess of the legal rate, and lenders had to resort to hidden charges and other subterfuges in periods of credit stringencies, making existing evidence on private rates hard to interpret. Before any discussion of theoretical issues, the empirical evidence must be placed on a much more solid footing. Pressnell (1960, pp. 211214) provides examples of various interest rates from the 18th century, taken from Thorold Rogers’ History of Agriculture and Prices in England, however, Thorold Rogers’ sampling technique is nowhere described, and his sources are not uniform. Therefore, we propose to develop consistent quarterly time series of both short- and long-term effective interest rates over the course of the 18th century, calculated to two decimal places (i.e., basis points). With this quantitative evidence in hand, we shall then discuss the following issues: (a) we shall ask did the long and short rates have a relatively fixed relationship; (b) we shall test whether the long rate was a weighted average of expected future short rates, and thus discuss the efficiency of the money market; (c) we shall provide some evidence in support of the Steuart/Thornton school on liquidity preference and credit crises; and finally (d) we shah comment on the “Ashton

INTEREST

RATES IN 18TH CENTURY

ENGLAND

hypothesis” (Ashton, 1955, 1959) and recent work on crowding-out credit crises in the 18th century.

5

and

II What is an appropriate representative interest rate in 18th century England? First, because rates in the countryside and rates in London were not well integrated until late in the century, pragmatism suggests that the scope of this study be narrowed to the London money market, which was well developed by at least the second decade of the century (see Cope, 1978; Mirowski, 1981). Given that fully private loan rates are rendered problematic by the usury laws and by the lack of any consistent data source, the compromise here advocated is to draw the data from the three big semiprivate joint stock holding companies of the 18th century, the Bank of England, the East India Company, and the South Sea Company, and to use the consol rate when it became available in 1750. Bank stock quotations, as a proxy for infinitely lived consols, begin with the century thus providing us with a substantially longer sample. In many respects, after the renewal of the Bank of England’s charter in 1709, and until the suspension of cash payments in 1797 the Bank is less risky an institution than the British government for investors, mainly because its portfolio mixed private and government paper. Finally, usury laws did not prevent the rate of return on Bank stock or consols from rising above the legal limit, because stock prices were free to fluctuate in any range, and shares were freely transferable throughout the period. If the yields on Bank stock and the consol rates were the nearest things to our ideal long rates, what was the best index of the short-term rate of interest? The answer is provided by Thomas Mortimer, in his do-ityourself guide Every Man His Own Broker in 1761: “India Bonds are the most convenient and profitable security any person can be possessed of, who has any quantity of cash unemployed but which he knows not how soon he may have occasion for; the utility and advantage of these bonds is so well known to the Merchants, and other Traders of the City of London, that it is wholly unnecessary to enlarge upon it. There is as little tropble with an India Bond as with a Bank Note” (quoted in Presnell, 1956, p. 266).’ The only other serious source for short-term interest rates in the early and mid-18th century are Navy victualling bills, which were vouchers issued by Navy captains in payment for ship provisions and redeeme ’ East India bonds were India Company semiannual payment should have been indeed this instrument was

acceptable at par in exchange for goods at the time of East sales. The potential use of this instrument as a means of incorporated into the quoted market discount or premium, if used for the purchase of goods.

6

WEILLER

AND MIROWSKI

in order of isssue by the government. We do not believe they are adequately representative of short-term rates for a number of reasons. First, they were not a very liquid asset, and for that reason alone traded at a substantial discount. Because they were retired sequentially, they had no fixed term to maturity; in fact, the term tended to lengthen appreciably during times of war. They were not freely assignable in law (Dickson, 1967, p. 401). Due to their nature, they were issued in irregular amounts, and their aggregate volume fluctuated wildly, although in general their volume was smaller than that of the East India bonds (see Dickson, 1967, p. 403). These considerations resulted in Navy bills fluctuating much more drastically in price than East India bonds, and thus they were not generally representative of public or private interest rates. The effective quarterly rates of return on Bank stock and East India bonds are plotted in Fig. 1 and listed in Appendix II. The method of calculation is described in Appendix I. As for broad trends, the long rate (the dividend yield on Bank stock) falls from above 5% in the South Sea Bubble to around 3.7% in 1736, a point of relative stability until the threat of the Pretender in 1745, when it rises to nearly 4.5% for a short time. Thereafter, it falls to a never-to-be-regained low of 3.1% in 17.52, and rises continuously to a peak of nearly 4.8% in 1762. From the end of the Seven Year’s War the long rate falls to a trough of 3.3% in 17681769; it rises continually until the beginning of the American Revolution, Percent 6

Short

l718

l728

1738

1748

Ten+*

1778

Bank of England Dividend Price Yield. * East India Company Bond Rote. l

l

FIG.

1. Long-term

and short-term interest rates, 1718-1796.

1788

1796

INTEREST

RATES IN 18TH CENTURY

ENGLAND

7

when it experiences a very abrupt further rise to a level to somewhat above 5%, at which plateau it remains until 1785. From 1786 to 1792 it falls nearly continuously to a level of 3.3%, only to rise again at the end of the year as a credit crisis develops. The quarterly interest rates on consols and those of the East India bonds are plotted in Fig. 2. At the time of their introduction, consols sold at nearly par, thus yielding a 3% return in 1750. The rate fell to a low of 2.8% in 1752, subsequently rising to 4.4% in 1762. From the end of the Seven Year’s War through the American Revolution the consol rate remained stable in the 3.3 to 3.4% range. During and after the American Revolution the consol rate rose to a high of 5.5% in 1782, and stayed in the 4.5-5.3% range until 1785 when it began to fall until the 1790s. At this time an abrupt rise begins. As is to be expected (although we postpone detailed discussion of this issue until the next section), the short-term rate is preponderantly lower than the long rate, and its fluctuations are broadly correlated with the long rate. This does not imply that one can safely treat the complex of maturities as a single rate, however, because there are distinct periods when the “term structure” is downward sloping: the South Sea Bubble, 1726-1727, 1747-1749, 1761-1763, 1765, 1783-1785, and 1791-1792. When the short rate rises well above the long rate, one potential interpretation is that it may be seen as an indication of a liquidity crisis, i Percent 6 ,

Short

l750

Term

Rate*+

1760

* 3 percent cons01 yield. *+ East India Bond Rote.

FIG. 2. Consol and short-term interest rates, 1750-1796.

8

WJZILLER

AND

MIROWSKI

that the requirements for current borrowing far outweighed any estimates of the future course of interest rates. Using the quantitative indices, contrary to Hoppit (1986), we find that parts of Ashton’s narrative of 18th century “financial crises” (Ashton, 1959, pp. 114-132) are not corroborated. Ashton’s thesis that government borrowing in wartime is the major cause of credit stringency apparently holds for 1746-1750, 1761, and 1784-1785; but it is interesting that the other incidents seem to have been due to endogenous causes in the money market. To illustrate this further, for instance, consider the claim of Hoppit (1986) to have identified 18th century financial crises in Britain. One discovers that his preferred indicator of financial crises is the magnitude of deviations from a 61-quarter moving average of the number of bankruptcies reported in the London Gazette, supplemented by court records from the Public Record Office. Hoppit interprets a sharp rise in this index as indicative of a financial crisis. While this index has something to teach us, it cannot be considered an adequate indicator of financial crises for reasons explained in Mirowski (1985, pp. 217-219). First, the level of bankruptcies is generally procyclical, in the sense that they tend to rise in business expansions; and second, any smoothing and filtering procedure such as a moving average distorts the variance properties of the time series, shifting frequencies as well as peaks and troughs. The resultant time series has no coherent theoretical interpretation, and therefore any derived chronology of crises must also be called into question. We have proposed to tackle this problem by taking the Steuart/Thornton position seriously. Consequently, we have tried to come up with an appropriate index of liquidity for the 18th century money market and settled upon the term structure of interest rates. Using this index, one could construct a chronology of credit crises which diverges in some respects from that offered by Ashton and Hoppit (see Table 1). We view this research as the first step toward the clarification of the issue of the role of government borrowing in the precipitation of financial crises and liquidity problems begun in Heim and Mirowski (1987). This discussion suggests that market instability due to internal functioning can be found as early as the mid-18th century. In the next section, we employ some work by Shiller (Shiller, 1979, 1981) to frame this question of market instability in the form of some testable hypotheses. We propose to test the existence of a rational expectations (RE) theory of the term structure in the 18th century context. In most general terms, this asserts that the slope of the term structure has something to do with expectations about future interest rates; more specifically, this posits a view of the Iong-term rate (&) as a weighted average of past expected short-term rates (r,). The weights used are an exponentially declining

INTEREST

RATES IN 18TH CENTURY

ENGLAND

9

TABLE 1 Dates of Financial Crisis in 18th Century England Ashton 1720 1726 1733 1745 1761 1763 1772 1778 1783 1788 1792

Hoppit 1720 1726 1763 1772 1778 1793 1797

Term structure inversion 1720 1726 1747 1761 1765 1783 1791

Sources. Ashton-see Mirowski (1985, p. 212); Hoppit-see Hoppit (1986, p. 45); Term structure inversion-see Appendix II.

discount rate 4 = l/(1 + R), where R is the mean of the long-term The long-term rate is given by the relation

rate.

where D, is a liquidity or risk premium. The long-term rate is therefore determined by the discount rate (q), future expectations (as of time t) of a sequence of short rates (YJ, and a sequence of liquidity and/or risk premia (0). Clearly this expression indicates that short and long rates will move toward each other when they differ. Below we will describe the method used by Shiller (1979) to test (1). The first thing we notice is that given a constant liquidity premium (0) the series R, will vary no more than the series r,. This stems from the fact that the weights decline geometrically, hence imparting less weight to Y in past periods, and further because an average has a lower variance than its components. It is also true that the variance of an average of expectations is less than or equal to the variance of an average of actual values, which yields our first relation, V(r) > V(R).

@I

Hence, an implication of this theory is that the variance of the short rate exceeds that of the long rate. Thus [RI, should be a smoother series than [r],.

10

WEILLER

AND

MIROWSKI

Our next step is to compare the actual R, series with the one that would result if (1) were the process driving the long rate. We use a methodology proposed by Shiller (1979, pp. 1190-1199) to linearize (l), and obtain R, = qnWC+J which is a first-order this, and assuming,

+ (1 - q&t

rational

+ Dn)

expectations

O
difference equation.

(3) Solving

D = 1 - 4 n-1 c qkD’“-k’ 1 - qn k=O along with a terminal

condition Rt

=

gives us (1

-

4)

5 k=O

$b+k.

(4)

Equation (4) also assumes perfect foresight (i.e., E,[rJ = r,) and no liquidity or risk premium. We then calculate the following synthetic series recursively from a terminal value to obtain R; = qR;+, + (1 - q)r,.

(3

We call this new series (RF) the expost rational rate. In other words, given the above characterization (rational expectations, perfect foresight, D = 0, etc.), R: represents the rational expectations of the long rate R,. For purposes of comparison, we have plotted the expost rational rate and the actual long rate for Bank stock in Fig. 3A and for consols in Fig. 3B. Furthermore we know that R; = -W-t) f u,,

(6)

where ut is a white noise error. This gives us the variance of R* V(R*) = V(R) + V(u) + 2 C(u, R).

(7)

Since all available information is incorporated into the rational forecast of R(R*), orthogonality of information implies that cov(u,R) = 0. Further the V(u) must be positive, giving V(R*) > V(R). By combining

(8)

Eqs. (8) and (2) we obtain V(R) < V(R*) < V(r),

(9)

since [R”], is simply a weighted average of R, and r,. Equation (9) gives us our first test of the null hypothesis of rational expectations, perfect

INTEREST

RATES IN 18TH CENTURY

ENGLAND

11

Percent

Expost

l750

1760

Rational

Rot&*

l770

* Bank of @gland dividend price yield l * Weighted average of Bank rate and East India bond

1780 yield according

1790 to equation

I796

(5).

Percent 6 B

2

I

Consol

I,

I,,,,,,,,,,,,,,,,,,,,,,,,,,,,,L,,,,,,,,,,,,,LI

1750 * 3 percent +* Weighted

Rate*

1760 consol yield. average of conrol

1770 rate and East India bend

1780 yield according

1790 to equation

1796

(5).

FIG. 3. Long-term and expost rational interest rates. (A) Bank stock. (B) Consois.

12

WEILLER

AND MIROWSKI

foresight, and no liquidity premium.’ If the V(R*) lies outside that bounds given by (9) we can reject the null hypothesis that [RI, evolves according to (4), and is thus consistent with the rational expectations hypothesis. The next step of our analysis involves looking -at the holding period rate of return defined as H’“’

I

z2

p$;‘)

- PI”’ + c PC”, f



(10)

where C is the coupon payment, superscript n refers to the date to maturity, and Pt = [(C/R?)) + (RJ”))(l + Rj”‘)]. Linearizing this around RI”-‘) = Rj”) = R = C gives

H’“) ‘) qn= 40(1-_ P) t = RI”’1- -q&Y-; q”) * %I

(11)

Our theory of the term structure implies that under risk neutrality, the short rate (rJ should be equal to the one period holding rate (H,), giving us a second test of our hypothesis. The variance (or standard deviation) of H, written o(H), must be less than or equal to the standard deviation of P, written as u(r), times a factor of proportionality (a), where a is obtained by maximizing the standard deviation of H (i.e., taking the first derivative of the expression for H and setting it equal to zero), setting R, = R,-l. We can then use following relations to test rational expectations: c+(H) < 44 efl< where a = (1 - q’)-“’

&L

(12)

(13)

and Ft is a forecast of the short rate by fitting

rt on R,.

If (12) and (13) are satisfied, then the holding period return series is smoother than the short-rate series as predicted by the theory. Another way to look at this result is to realize that H, is simply a linear combination of R, and P,H., therefore if Pp’ = Pj”-I), V(H) < V(R), i.e., the series [Hj, is smoother. A major problem with (12) and (13) is that they do not allow for the possibility that the short rate, [t-Ii, may have an infinite variance (i.e., [rl, is nonstationary). This would be the case if [rlt followed a random walk. As a result Shiller looks at (rr - rt-l) and (H, - r,), and their corresponding standard deviations cr(Ar) and a(H - r). The term u(H - r) reflects the difference between the one-period rate that can be read off the term structure (HJ and the actual short rate (r,). Again we max’ The null hypothesis is actually a joint hypothesis of more than just these assumptions. We will designate this set of assumptions simply as “rational expectations.”

INTEREST

RATES

IN

18TH

CENTURY

ENGLAND

13

imize the standard deviation of (H, - r,); however, this time we make assumptions about the stochastic process which drives (r, - r,- J. Under the assumption that [r], is an AR(l) process (i.e., rr = p rt-, + F,) we obtain the bounds cr(H - r) < bw(Ar)

b = a/R.

(14)

If alternatively, we assume [(r, - r,-J] follows a white noise process, then [RI, is a random walk (i.e., R, = r, + EJ, thus implying E(R, rJ = 0. We obtain a(H

- r) < ccr(Ar)

c = l/R.

(1%

The test statistics in (14) and (15) allow us to look at nonstationary series, thus negating one often cited criticism of this methodology. Flavin (1983) raises a number of points concerning the biases found in these types of tests. Most prominently cited is the potential for the sample moments of a distribution to differ from its population momeuts. To this end, Flavin derives a bias correction for c(r), given as k in Eq. (16) (Flavin, 1983, p. 946): [l - (1 + p) + 2p(l - pT)]*” (1 + p)T (1 - p)‘T’ ’ where p is the autoregressive parameter on the short rate process and T is the number of observations. Our empirical work .results adjuste for this criticism are presented along with those using the standard methodology. Our next task is to test covariance restrictions. A truly efficient market implies that there should be no way to make a profit through arbitrage using currently available information, Therefore, we look at the covariance between the one-step-ahead forecast error (R”, - R, f II) = Z,, and either R,, or r,. Theory indicates that if the market is efficient E[Z,, R,] = E[Z,, r,] = 0.

To test this we can run a regression of (R*, - R,) on Ii,, R,-l, r,, rrmI, etc. Since this regression would likely suffer from serious serial correlation, we use a generalized least squares regression of (H, - rJ on R, R,- 1, . . . rt-- r,- I, . . . , or more simply on R,, . . . , r, and perform a joint F-test. If the market is efficient we should accept a null hypothesis of all coefficients equal to zero, indicating the absence of covariation as stated above. This indicates that all available information was used in the formation of these asset prices, lending credence to the hypothesis of rational expectations. A final test along similar lines looks at both the covariance restriction and the critical implication of rational expectations theories of the term structure, that short rates and long rates converge toward each other

14

WEILLER

AND MIROWSKI

when they differ. Put in another way, this is the tendency for long-term rates to move in such a way as to equalize holding period returns. These tests are performed by regressing (H, - VJ on R,, and (H, - rJ on (B, rJ, with the null hypothesis indicated above (p = 0). If instead we regress (&+r), which is a linear combination of (H, - rJ, on (R, - r,), the null hypothesis would be p = ((1 - q)/q). We use only the most recent observations of R, and r, because according to rational expectations these prices should be a sufficient statistic for all information available up to time t. Finding a coefficient significantly different from zero would indicate the role of outside events in the formation of asset prices. Furthermore, a negative coefficient would indicate that the long rate moved so as to widen the divergence between the holding period rates and the short rate, providing us with a strong rejection of this aspect of the rational expectations theory. This would point to R, rising temporarily relative to r, with a shock, and subsequently falling to a normal level.3 IV In this section we implement the tests described in Section III by first looking at the full sample and then at three subperiods. The first subsample runs from the beginning of our sample through the first appearance of consols (1718 Q4-1750 Q3), the second subdivision looks at the period prior to the American Revolution (1750 Q4-1775 Q4), and the last considers the period from the American War through the end of our sample, (1776 41-1796 Q4). We will present results in two parts, looking initially at variance bounds tests then at tests on the covariance restrictions imposed by the theory. IVa Table 2 contains all information pertinent and (13). We look initially at Eq. (9)

to tests of Eqs. (9), (12),

V(R) < V(R*) < V(r) which is violated in all samples tested. The strongest violation occurs in the last subsample, where [RI, and the [rlr series have a similar standard deviation while the expost rational rate ([R*]J is about one-third as volatile. Therefore, our first tests indicate a rejection of the null hypothesis of rational expectations. The next two tests are given by Eqs. (12) and (13), respectively,

3 Shiiler (1979, p. 1212). This iast comment as well as much of the above derivations are due to Shiller.

INTEREST

RATES

Tests

Bank stock 1718:1-1796:IV 1718:I-1X0:111 175O:IV-1775:IV 1776:I-1796:IV Consols 175O:IV-1796:IV 175Q:IV-1775:IV 1776:1-1796:IV

Note.

The

expression

IN

18TH

TABLE Assuming

CENTURY

ENGLAND

15

2 Stationarity

44 aa

dR*) ao(?)

cm

Eq.(12)

Eq. (9)

Eq. (13)

0.0520 0.2769 0.0484 0.2888 0.0305 0.2557 0.0562 0.1993

0.0502 0.2436 0.0431 0.2288 0.0232 0.2557 0.0216 0.1979

0.765 Reject 0.0809 Reject 0.0680 Reject 0.0571 Reject

Reject Reject Reject Reject Reject Reject Reject Reject

0.0688 0.2708 0.0294 0.2706 0.0617 0.1993

0.0507 0.2671 0.0235 0.2702 0.0281 NA

0.0760 Reject 0.0680 Reject 0.0571 Reject

Reject Reject Reject Reject Reject Reject

for a is given

in Eq. (13);

q

is given

in Table

0.4511 0.4859 0.4059 0.4958

0.5595 0.4660 0.6133

3.

They test the standard deviation of the linearized holding period yields (H,) against those of the short rate (rJ and the predicted short rate (PJ, respectively. As described in Section III, the factors of proportionality are obtained by maximizing the standard deviation of H,. The data strongly violated these bounds for all samples considered for both versions of the tests. In Table 3, we present an alternative set of variance bounds tests on Eqs. (12) and (13) suggested by Flavin. The correction involves adjustment of the standard deviations of Y and r by a factor l/k, where k is given by Eq. (16). Using this correction, we note that the null hypothesis of rational expectations is rejected less than 50% of the time. While we are unable to compute k in the final period of our study due to a value of p equal to or exceeding one, the rational expectations hypothesis is accepted for both consols and Bank stock in the period extending from 1750 to 1776, and rejected in the earlier period (1718-1750) for Bank stock. The Flavin critique clearly has strong, yet mixed, results on our conclusions. Since computation of k requires knowledge of the serial correlation parameter, p, Flavin’s critique also forces us to look at the stationarity of these interest rate processes. Without exception, every p falls within a 95% confidence interval of one, forcing us to question the robustness of tests such as those in Eqs. (9), (12), and (13). In Table 4, we give

16

WEILLER

Test

Using

Eqs.

by Flavin

au(f)

Eq.

44

k

1718:1-175O:III 175O:IV-1775:IV 1776:1-1796:IV Consols 175O:IV-1796:IV 175O:IV-1775:IV 1776:1-1796:IV

Note. The expression

MIROWSKI

TABLE 3 (12) and (13) as Adjusted

da

Bank stock 1718:1-17%:IV

AND

(1983)

(12)

n

9

0.4511 0.5880 0.4859 0.6145 0.4059 0.2288 0.4958 NA

0.4709 316 0.4699 131 1.1178 101 NA 84

0.3815 0.961 0.3723 0.960 1.1178 0.964 NA 0.958

Accept 0.996 Reject 0.988 Accept 0.998 NA 1.005

0.5595 NA 0.4660 0.2288 0.6133 NA

NA 185 1.1829 101 NA 84

NA 0.963 1.1829 0.968 NA 0.958

NA 1.002 Accept 0.998 NA 1.005

for

k is given

in Eq.

Eq. (13)

P

Reject Reject Accept NA

NA Accept NA

(16).

results of variance bounds test, suggested by Shiller, that do not require the data to be stationary. In testing Eq. (14), cr(H - r) < bo(Ar), we obtain b by assuming that Ar is an AR(l) process. Equation (15), cr(H - r) < ccr(Ar), uses the assumption that [Y], is a random walk to obtain c. In 9 of 10 cases, the inequalities in (14) and (15) are satisfied by the data. Satisfaction of those inequalities indicates that there is no evidence to contradict the existence of efficient markets and RE, based only in knowledge of (Ar), and the

Tests

Not

TABLE Assuming

4 Stationarity

dH-4

bdr)

c&9

Eq.

Bank stock 1718:1-1796:IV 1718:1-175O:III 175O:IV-1775:IV 1776:1-1796:IV

0.4479 0.4831 0.4071 0.4885

3.194 4.213 2.384 1.780

.8820 1.180 .634 .510

Accept Accept Accept Accept

Accept Reject Accept Accept

Consols 175O:IV-1796:IV 175O:IV-1775:IV 1776:1-1796:IV

0.5543 0.4647 0.6049

2.282 2.808 1.797

.605 .705 ,515

Accept Accept Accept

Accept Accept Reject

Note. The expressions

for

b and c are given

in Eqs.

(14) and (15),

(14)

respectively.

Eq. (15)

INTEREST

RATES IN 18TH CENTURY

ENGLAND

17

above arguments. According to Shiller (1979, p. 1205), “In simple terms, it is conceivable that even if a(Ar) is very small, W(Y)may be large, if Y is expected to drift far above its historical range in the future.” This, in fact, occurs in the later part of the 18th century when consol rates rise from 3 to 5% and above. IVb

Our next step is to analyze the covariance between the one-step-ahead forecast error (R*, - R, + D), and information available at the time and before (i.e., R,, r,; s < t). In doing this we used a correction for serial correlation suggested by Shiller (1979, p. 1202). We ran regressions of (H, - rJ on Rt, Rt-l, Rt-2, rf, rf-,, Y,-~and found the null hypothesis of all coefficients except the constant term equal to zero to be rejected at a minimum of the 10% level in all samples. This result could cast further doubt on our hypothesis of rational expectations by violation of the orthogonality of information assumption.4 These tests are not presented in the paper, but are available from the authors. Our final test of this type considers the relation between the spread of the holding period rate of return and the short rate, or a linear combination of these, and the long rate (or the spread between short and long). Table 5 gives the results of regressions of (H, - ut) on R,. Both these results and those in Table 6 are corrected for first-order serial correlation using the AR(l) routine in TSP, and test the critical rational expectations assumption of equalization of one-period holding rates and short rates, as well as the covariance restrictions discussed above. In this case we again find evidence contrary to rational expectations. 6;tests similar to those described above give us a rejection of the nu hypothesis of p = 0 at the 5% level for all periods and assets except 171&I-1750:IV for Bank stock. Two problems exist with these results: First, according to Shiller (1979, p. 1211) these estimates may be biased upward due to simultaneity bias arising from the fact that R, appears on both sides of the equation Second, there is the fact of the small size of the R-squares. Since a rejection of p = 0 implies that exogenous events and unused, but available current information were important in the formation of asset prices, a low R-square may indicate that the information contained in R, is not very useful in helping to predict H, - Y*. Table 6 gives results of regressions of (H, - YJ, which is a linear combination of (H, - rJ, on (R, - Y,). We test the hypothesis of fi = 0, or that errors are orthogonal to the currently available information in the term structure. We accept the null hypothesis at the 5% level for 4 As pointed out above, rational expectations posits that all errors in estimates are independent of information .availabIe at that time. This results from the full exploitation of all information available to agents at any given time.

* Significant ** Significant

1776:L1796:IV

175O:IV-1775:IV

Consols 175O:IV-1796:IV

1776:1-1796:IV

175O:IV-1775:IV

1718:L1750:111

Bank stock 1718:L1796:IV

at 10% level. at 5% level.

-0.38** (-2.01) -0.124** (-2.63) -0.088** (-2.15)

-0.080** (-3.46) -0.090** (-2.15) -0.184** (-3.34) -0.137** (-2.68)

1.98** (2.14)

G.64)

0.986** (2.01) 3.68**

1.30** (2.34) 1 s2 (1.51) 4.33** (2.94) 2.44** (2.12)

TABLE

5

0.093 (-1.25) -0.044 (-0.44) -0.084 (-0.75)

0.20s** (3.75) 0.124 (1.41) 0.326** (3.45) 0.379** (3.67)

(r,)

(H, - rr) = a + PR, + E,

0.041

0.056

0.016

0.042

0.069

0.011

0.015

4.59*

6.95**

4.04**

4.66**

8.51**

2.46

5.81**

F

1.91

2.00

1.95

1.90

1.92

1.95

1.93

DW

INTEREST

RATES

IN

18TH

CENTURY

ENGLAND

20

WEILLER

AND

MIROWSKI

the Bank for all samples except 1776:1-1796:IV, which we accept at the 10% level. For consols, we accept our null hypothesis at the 10% level for all but the subsample from 1776:1 to 1796:IV. However, it is important to notice the negative signs on the coefficients, as this provides further evidence, albeit weak, against the rational expectations theory of the terms structure. These results are again subject to the caveats mentioned above. Our empirical tests seem to provide us with a mixed assessment of a rational expectations theory of the term structure. We accepted our null hypothesis only when the tests suggested by Shiller were adjusted using Flavin’s methodology, and when we look at I. On the other hand, regression tests provided us with further limited refutation of the RE theory of the term structure. V As the reader of the previous sections will realize, the question whether 18th century British interest rates exhibited “normal” behavior from the viewpoint of 20th century neoclassical economic theory is a very complex and involved issue. The Shiller variance bounds tests and their amendments leave us with an equivocal answer: from some points of view and in some respects the behavior of 18th century term structure of interest rates may conform to the “rational expectations” characterization, and in some respects it does not. As Shiller and McCulloch (1987, pp. 4546) admit, “The volatility tests do not allow us to tell whether there is too much volatility in long rates or just nonstationarity in short rates.” In the case of 18th century London, nonstationarity of short rates seems to be the prime suspect, particularly with regard to the upward trend in all interest rates from midcentury to 1796, as one can observe in Fig. 2. One major lesson to be drawn from this exercise is the extent of the difficulties surrounding the portrayal of interest rates as purely mechanistic phenomena, responding in a normal manner to fixed determinants. One cannot simply assume that in all circumstances long rates were simply the weighted average of expected values of future short rates. Another thing we do not find is a term structure of interest rates which might be safely characterized by a single market rate, as for instance implied in (Williamson, 1984). The results in Tables 5 and 6 should be regarded as a first tentative attempt to dig deeper into the exact structure of “inefficient” behavior from the viewpoint of neoclassical theory. We suggest that interest rates should not only be portrayed as responding but also as responding to events to so-called “market fundamentals,” external to that theory (Mirowski, 1987). While most economic historians probably do not consider themselves partisans of the rational expectations school of economic theory, they probably do share a sense that economists have a fair understanding of

INTEREST

RATES

IN

18TH

CENTURY

ENGLAND

21

the functions that markets perform, that markets have been an organizing force throughout western history. This view provides a framework, for example, in Ashton’s (1959, 1955) classic works on the economy of 18th century England. In those books, Ashton applies early 20th century economic theory to the 38th century without much in the way of explicit rationale or justification. Stagnation is explained as the hampering of market forces by external agencies, most concretely by the English government. Interest rates played an important role in Ashton’s story, because they were thought to channel investment to appropriate areas. The government’s prosecution of wars through deficit finance channels finance to inappropriate uses, thus driving up interest rates and curtailing economic growth. Usury laws prevent the free transmission of market information, and also hamper growth. This analysis is repeated in Williamson ( 1984). This conventional view needs to be reconsidered at many levels. First, Ashton (1959, p. 104) needed to assert that the ensemble of interest rates moves in tandem so that he could discuss the empirical effects of “the” interest rate on the economy. As we have seen, this assumption cannot withstand empirical scrutiny. Second, Ashton assumes that there is an agreed-upon theory of the impact of interest rate moveinents upon the economy. This was not true in the 18th century, as we have shown in our initial discussion of the Thornton/Steuart and Smith/Hume schools; it may come as some surprise that the situation is little better at present (Heim and Mirowski, 1987). The reader need merely to compare Malinvaud (1972, Chap. lo), Dodds and Ford (1983), Leijonhufvud (I981), Evans (1985), and Garegnani (1972) to see that there is still much dissension over the function of an interest rate in a market setting. The empirical implications of the rational expectations conceptions of interest are formulated in a format amenable to statistical tests, which allowed us to deploy the tests reported in this paper. Unfortunately, the rational expectations hypothesis did not manage to illuminate the function of inWest rates either in the 18th century, or in Shiller’s work in the 20th century. This indeterminacy in the theoretical role of interest rates brings us to our third criticism of the conventional view. If there is as yet no consensus as to whether interest rates are uniquely procyclical or anticyclical, this raises the question whether the government activity in the market for debt was intrusively detrimental to economic grow&. If there is no unique specification of the role of debt markets in this period, then there is no solid case that they were internaIly sufficient in generating economic growth nor is there any reason to believe governmknt borrowing activities were uniquely an impediment. In essence, this dichotomy .echoes the contrast between Smith’s belief that ihe interest rates merely reflected the rates of profit which were the outcome of inde-

22

WEILLER

AND

MIROWSKI

pendent market forces, and Thornton’s conception of the debt market as an institution to be nurtured by government control. We bring some evidence to this dispute, showing that if an inverted term structure of interest rates is viewed as a symptom of a credit crisis, then there are a number of such instances in the 18th century which cannot be readily traced to wars or other obvious meddling by the government, contrary to Hoppit (1986). This raises the possibility of endogenous market instability. Of course, such instability can only be defined in relation to a theory of the role and function of market interest rates. The problem of constructing a coherent theory of financial crises in the spirit of Steuart and Thornton is a tremendously complicated issue, and one to which justice cannot be done in this context. (An extended discussion of this issue may be found in Mirowski (1985)). However, a few brief observations are in order. First, the present research should caution that the notion of a simple and transparent historical narrative of financial crises is a will-o-the-wisp, primarily because any interpretation of the primary historical materials is so very theory laden. Making the theoretical presuppositions a little more obvious-as we have attempted to do here by resorting to a theory based on rational expectations-often reveals that controversy over the historical record is really controversy over the meaning and interpretation of the theory. In summary, this paper provides historians with an improved and unified database, and raises some questions about relationships between theory and economic history. Sometimes it happens that economic theory helps to organize historical inquiry; but other times, historical inquiry can raise important questions about economic theory. APPENDIX

1

Data Sources All data for this paper were obtained from microfilm copies of The Course of the Exchange, the semiweekly report of the London Stock Exchange. Dividends are listed in the quarter in which they were paid but are averaged over two periods and multiplied by 4 to get a continuous series at an annual rate. This gives us the share yield defined as R, = 4*(Div,

+ Div,+,)/2)

where Div, is the dividend paid in quarter Bank of England stock defined as PBOE,

= 5 WPBOET

/ PBOEt, t,

(~‘4

and PBOET is the price of

/ 13,

W)

t=1

where WPBOE,

is the weekly price of Bank of England

stock. Short

INTEREST

RATES

IN

18TH

CENTURY

rates were also obtained from the “Course” The short rate is defined as Y, = [(Coupon/lOO)

23

ENGLAND

and are end of period quotes.

/ (1 + (Premium/lOO))l,

643)

where Coupon is the yearly coupon rate on the East India bond, and Premium is the premium or discount with respect to par ( = 100 pounds). The benchmark (R) for the discount factors (4) were calculated as simple averages over the appropriate samples. APPENDIX II Bank of England Dividend Yield, East India Bond Rate, and 3% Consol Yield Bank dividend yield

Date 1710 1710 1710 1710

Ql 42 43 44

6.62 5.70 6.23 6.80

1711 1711 1711 1711

Ql 42 43 44

6.77 6.81 6.76 7.30

1712 1712 1712 1712

QI Q2 Q3 44

7.23 7.12 6.99 6.92

1713 1713 1713 1713

Ql Q2 43 44

6.56 6.45 6.25 6.45

1714 1714 1714 1714

Ql Q2 Q3 Q4

6.58 6.64 6.26 5.85

1715 1715 1715 1715

Ql Q2 Q3 Q4

5.67 6.12 6.16 6.48

1716 1716 1716 1716

Ql Q2 Q3 Q4

6.25 6.05 5.82 5.78

1717 QI 1717 Q2

5.87 5.81

East India bonds

3% Consols

Bank dividend yield

Date

East India bonds

1717 Q3 1717 Q4

5.34 5.44

1718 1718 1718 1718

Ql Q2 Q3 44

5.09 5.37 5.34 5.53

4.91 0.99 3.98 3.99

1719 1719 1719 1719

Ql Q2 43 Q4

5.19 4.77 4.77 4.85

3.94 4.93 4.91 4.93

1720 1720 1720 1720

Ql Q2 Q3 44

4.62 3.76 3.08 3.89

4.99 4.90 4.94 5.04

1721 1721 1721 1721

Ql Q2 Q3 Q4

4.31 4.60 4.50 4.85

4.88 5.05 5.00 5.01

1722 1722 1722 1722

QI Q2 43 Q4

4.98 5.23 5.15 5.23

5.01 4.99 4.99 4.96

1723 1723 1723 1723

Ql Q2 Q3 Q4

5.13 5.13 4.99 4.98

4.94 4.92 4.91 4.91

1724 1724 1724 1724

Ql Q2 Q3 Q4

4.66 4.68 4.57 4.60

4.90 4.91 4.91 4.91

3% Consols

24

WEILLER

APPENDIX

Date

,Bank dividend yield

East India bonds

3% Consols

AND

MIROWSKI

II-continued

Date

Bank dividend yield

East India bonds

1725 1725 1725 1725

Ql Q2 43 44

4.54 4.51 4.42 4.55

3.91 3.91 3.92 3.94

1736 1736 1736 1736

Ql 42 43 Q4

3.70 3.70 3.66 3.69

3.30 3.28 3.28 3.30

1726 1726 1726 1726

Ql 42 43 44

4.78 4.86 4.72 4.97

4.96 4.95 4.95 4.97

1737 1737 1737 1737

Ql Q2 Q3 44

3.68 3.76 3.80 3.84

3.32 2.80 2.81 2.81

1727 1727 1727 1727

Ql 42 43 44

4.87 4.67 4.52 4.22

4.95 4.83 4.85 4.84

1738 1738 1738 1738

Ql 42 43 44

3.89 3.90 3.85 3.84

2.81 2.81 2.81 2.81

1728 1728 1728 1728

Ql Q2 43 Q4

4.03 4.04 4.07 4.12

3.83 3.82 3.82 3.87

1739 1739 1739 1739

Ql Q2 43 44

3.83 3.89 3.99 4.00

2.81 2.91 2.91 2.87

1729 1729 1729 1729

Ql Q2 Q3 Q4

4.07 4.05 3.97 4.35

3.85 3.82 3.82 3.81

1740 1740 1740 1740

Ql 42 43 44

3.95 3.88 3.86 3.97

2.87 2.85 2.91 2.89

1730 1730 1730 1730

Ql Q2 Q3 Q4

4.29 3.88 3.85 4.19

3.81 3.79 3.78 3.78

1741 1741 1741 1741

Ql 42 43 44

3.91 3.87 3.90 3.97

2.88 2.87 2.88 2.89

1731 1731 1731 1731

Ql Q2 43 44

4.14 3.76 3.72 4.07

3.78 3.77 3.77 3.74

1742 1742 1742 1742

Ql Q2 Q3 Q4

4.02 3.93 3.87 3.87

2.88 2.88 2.88 2.87

1732 1732 1732 1732

Ql 42 43 Q4

4.01 3.71 3.64 3.67

3.78 3.75 3.86 3.80

1743 1743 1743 1743

Ql Q2 43 Q4

3.79 3.74 3.72 j.74

2.87 2.87 2.86 2.86

1733 1733 1733 1733

Ql Q2 Q3 Q4

3.65 3.67 3.72 4.12

3.32 3.32 3.38 3.46

1744 Ql 1744 42 1744 Q3 1744.Q4

3.74 3.85 3.75 3.78

2.97 2.93 2.92 2.93

1734 1734 1734 1734

Ql Q2 Q3 44

4.12 4.20 3.99 4.00

3.47 3.41 3.37 3.37

1745 1745 1745 1745

Ql Q2 Q3 Q4

3.78 3.75 3.81 4.08

2.96 2.95 3.03 3.11

1735 1735 1735 1735

Ql Q2 Q3 Q4

3.93 3.99 3.94 3.82

3.34 3.37 3.33 3.31

1746 1746 1746 1746

Ql Q2 Q3 Q4

4.49 4.45 4.11 3.86

3.01 3.93 3.86 3.91

3% Consols

INTEREST

RATES

IN

APPENDIX

Date

Bank dividend yield

East India bonds

3% Consols

18TH

CENTURY

25

ENGLAND

II-continued

Date

Bank dividend yield

East India bonds

3% Cons&

1747 1747 1147 1747

Ql Q2 43 Q4

3.90 3.94 3,99 4.13

3.92 3.96 3.95 3.96

17.58 1758 1758 17.58

Ql 42 43 44

3.77 3.71 3.76 3.82

2.91 2.92 2.96 2.97

3.76 3.79 3.82 3.85

1748 1748 1748 1748

Ql 42 43 Q4

4.22 4.02 3.92 3.96

3.98 3.95 3.96 3.96

1759 1759 1759 1759

Ql 42 Q3 Q4

3.86 3.97 4.04 3.95

3.01 4.01 4.00 4.00

3.89 3.92 3.92 3.91

1749 1749 1749 1749

Ql Q2 Q3 Q4

3.89 3.75 3.62 3.67

3.92 3.85 3.85 3.92

1760 1760 1760 1760

Ql 42 Q3 Q4

4.05 4.08 4.05 4.12

4.00 3.99 3.97 4.01

3.91 3.91 3.90 3.90

1750 1750 1750 1750

Ql 42 Q3 44

3.75 3.75 3.71 3.71

3.96 2.84 2.85 2.87

3.16

1761 1761 1761 1761

Ql Q2 Q3 Q4

4.27 3.92 3.99 4.30

3.96 4.88 4.95 4.99

3.90 3.90 3.86 3.82

1751 1751 1751 1751

Ql Q2 Q3 44

3.66 3.62 3.56 3.55

2.87 2.85 2.85 2.84

3.17 3.18 3.20 3.21

1762 1762 1762 1762

Ql Q2 Q3 Q4

4.80 4.65 4.29 4.06

4.95 4.95 4.87 4.84

3.77 3.73 3.68 3.63

1752 1752 1752 1152

Ql Q2 43 44

3.48 3.43 3.40 3.12

2.83 2.81 2.79 2.79

3.22 3.24 3.26 3.21

1763 1763 1763 1763

Ql Q2 Q3 Q4

3.62 3.59 3.78 4.01

4.80 4.81 4.96 3.99

3.59 3.54 3.49 3.44

17.53 1753 1753 1753

Ql Q2 43 Q4

3.14 3.19 3.27 3.28

2.80 2.79 2.80 2.82

3.29 3.31 3.33 3.35

1764 1764 1764 1764

Q1 Q2 Q3 44

3.92 4.32 4.42 4.09

4.00 3.96 3.91 3.90

3.42 3.39 3.37 3.35

1754 1754 1754 1754

Ql Q2 Q3 Q4

3.35 3.36 3.36 3.40

2.84 2.83 2.82 2.85

3.37 3.39 3.41 3.44

1765 1765 1765 1765

Ql Q2 Q3 Q4

3.91 3.89 3.74 3.64

3.86 4.82 3.90 3.94

3.33 3.31 3.25 3.23

1755 1755 17.55 1755

Ql Q2 Q3 Q4

3.44 3.52 3.63 3.69

2.94 2.92 2.94 2.95

3.46 3.48 3.50 3.52

1766 1766 1766 1766

Ql Q2 Q3 Q4

3.70 3.70 3.65 3.63

3.95 2.96 2.96 2.97

3.20 3.17 3.18 3.19

1756 17.56 1756 1756

Ql Q2 Q3 Q4

3.73 3.80 3.85 3.73

2.96 2.95 2.92 2.95

3.54 3.57 3.59 3.61

1767 1767 1767 1767

Ql Q2 Q3 44

3.49 3.85 3.14 3.48

2.98 2.99 2.99 2.99

3.2 3.21 3.21 3.22

1757 17,57 1757 17.57

Ql Q2 Q3 Q4

3.85 3.80 3.76 3.74

2.93 2.92 2.92 2.93

3.64 3.67 3.70 3.72

1768 1768 1768 1768

Ql Q2 43 Q4

3.38 3.28 3.29 3.41

2.98 2.97 2.97 2.95

3.23 3.24 3.25 3.26

26

WEILLER

APPENDIX

Date

AND

MIROWSKI

II-continued

Bank dividend yield

East India bonds

3% Consols

Date

Bank dividend yield

East India bonds

3% Consols

1769 1769 1769 1769

Ql 42 43 44

3.36 3.32 3.29 3.47

2.94 2.95 2.95 2.95

3.27 3.29 3.30 3.31

1780 1780 1780 1780

Ql 42 43 44

4.87 4.90 4.78 4.90

3.97 3.95 3.98 3.98

4.33 4.34 4.36 4.37

1770 1770 1770 1770

Ql 42 43 44

3.58 3.59 3.65 4.02

2.94 2.94 2.98 2.96

3.33 3.34 3.36 3.37

1781 1781 1781 1781

Ql Q2 Q3 Q4

5.06 5.34 5.28 5.37

3.99 3.98 3.99 3.98

4.39 4.40 4.42 4.43

1771 1771 1771 1771

Ql 42 Q3 Q4

3.81 3.60 3.53 3.71

2.94 2.92 2.93 2.93

3.39 3.40 3.42 3.44

1782 1782 1782 1782

Ql Q2 Q3 Q4

5.38 5.25 5.24 5.14

3.98 3.99 4.00 4.03

4.45 4.47 4.49 4.51

1772 1772 1772 1772

Ql Q2 03 Q4

3.62 3.64 3.69 3.79

2.93 2.96 2.97 2.99

3.46 3.48 3.50 3.52

1783 1783 1783 1783

Ql Q2 Q3 Q4

4.72 4.49 4.74 5.08

4.00 5.01 5.07 5.12

4.53 4.55 4.53 4.51

1773 1773 1773 1773

Ql Q2 Q3 Q4

3.86 3.90 3.86 3.87

2.99 3.00 2.98 2.97

3.54 3.56 3.58 3.61

1784 1784 1784 1784

Ql Q2 Q3 Q4

5.26 5.16 5.18 5.26

5.05 5.05 4.99 5.02

4.48 4.46 4.44 4.42

1774 1774 1774 1774

Ql Q2 Q3 44

3.93 3.91 3.84 3.83

2.96 2.93 2.91 2.92

3.63 3.66 3.68 3.71

1785 1785 1785 1785

Ql Q2 Q3 44

5.22 -5.15 4.96 4.54

5:Ol 4.99 4.9s 4.94

4.40 4.37 4.35 4.33

1775 1775 1775 1775

Ql 42 43 44

3.82 3.85 3.86 3.85

2.91 2.92 2.91 2.90

3.74 3.77 3.81 3.84

1786 1786 1786 1786

Ql 42 43 44

4.28 4.27 3.94 4.01

4.89 4.83 4.79 4.82

4.30 4.28 4.26 4.24

1776 1776 1776 1776

Ql Q2 Q3 Q4

3.87 3.94 3.95 3.99

2.94 2.95 2.93 2.95

3.88 3.91 3.95 3.99

1787 1787 1787 1787

Ql 42 43 Q4

3.94 3.88 3.99 4.59

4.91 3.88 3.87 3.85

4.22 . 4.19 4.20 4.22

1777 1777 1777 1777

Ql 42 Q3 44

4.01 4.07 4.18 4.21

2.98 3.00 3.00 3.00

4.03 4.07 4.11 4.15

1788 1788 1788 1788

Ql Q2 Q3 Q4

4.36 4.03 4.00 4.04

3.84 3.87 3.86 3.85

4.23 4.24 4.26 4.27

1778 1778 1778 1778

Ql 42 Q3 Q4

4.67 5.11 4.96 4.86

3.01 3.96 3.91 3.91

4.20 4.24 4.25 4.26

1789 1789 1789 1789

Ql Q2 Q3 Q4

4.17 3.99 3.79 3.78

3.86 3.84 3.80 3.80

4.29 4.31 4.32 4.34

1779 1779 1779 1779

Ql Q2 Q3 Q4

5.03 4.88 5.01 4.96

3.96 3.98 3.96 3.95

4.28 4.29 4.30 4.32

1790 1790 1790 1790

Ql Q2 Q3 44

3.75 3.94 3.97 3.83

3.79 3.92 3.84 3.81

4.37 4.39 4.40 4.43

INTEREST

RATES IN 18TH CENTURY

APPENDIX

1791 1791 1791 1791 1792 1792 1792 1792 1793 1793 1793 1793 1794 1794

Ql Q2 43 Q4 Ql 42 43 44 Ql Q2 43 Q4 Ql Q2

Bank dividend yield

East India bonds

3% Consols

3.71. 3.80 3.60 3.53 3.34 3.33 3.42 3.61 4.12 4.10 4.00 4.06 4.36 4.24

3.92 3.83 3.79 3.83 3.79 3.80 3.80 3.98 4.00 4.02 3.98 3.96 3.97 3.99

4.45 4.47 4.49 4.52 4.55 4.58 4.61 4.64 4.66 4.69 4.71 4.74 4.77 4.80

ENGLAND

27

b-continued

Date 1794 Q3 1794 Q4 1795 Ql 1795 Q2 1795 Q3 1795 44 1796 Ql 1796 Q2 1796 Q3 1796 Q4 1797 Ql 1797 42 1797 Q3 1797 Q4

Bank dividend yield

East India bonds

3% Consols

4.29 4.46 4.56 4.41 4.11 4.11 4.02 4.25 4.66 4.80 5.07 5.75 5.42 NA

3.97 3.99 4.01 4.02 3.99 4.02 4.01 3.97 5.04 5.04 5.01 5.05 5.02 5.03

4.83 4.87 4.90 4.93 4.97 5.01 5.04 5.08 5.12 5.13 NA NA NA NA

REFERENCES Anderson, B., and Cottrell, P. (1974), Money and Banking in England. England: David & Charles. Ashton, T. S. (1955), An Economic History of England: The 18th Century. London: Methuen. Ashton, T. S. (1959), Economic Fluctuations in England, 1700-1800. London/New York: Oxford Univ. Press. Barre, R. .I. (1987), “Government Borrowing, Interest Rates, Prices, and Budget Deficits in the U.K., 1701-1918.” Journal of Monetav Economics, 20, 221-247. Carter, A. C. (1975), Getting, Spending and Investing in Early Modern Times. Assem: Van Gorgum. Chaudhuri, K. (1978), The Trading World of Asia. London/New York: Cambridge Univ. Press. Clapham, J. (1944), The Bunk of England. London/New York: Cambridge Univ. Press. Two Volumes. Cope, S. R. (1978), “The Stock Exchange Revisited.” Economica 45, l-22. Dickson, P. G. M. (1%7), The Financial Revolution in England. New York: St. Martin’s. Dodds, J., and Ford, J. (1983), Expectations, Uncertainty, and the Term Structure of Interest Rates. New York: Barnes & Noble. Evans, P. (198.5), “Do Large Deficits Produce High Interest Rates?” American Economic Review 75, 68-87. Flavin, M. A. (19$3), “Excess Volatility in the Financial Markets: A Reassessment of the Empirieal Evidence.” Journal of Political Economy 91(6), 929-956. Garegnani, P. (1972), “On a Change in the Notion of Equilibrium.” In M. Brown, K. Sato, and P. Zarembka (Eds.), Essays in Modern Capital Theory. Amsterdam: Nolrh Holland. Pp. 25-45. Grossman, H., and Shiller, R. (1981), “The Determinants of Variability of Stock Market Prices.” American Economic Review 71(2), pp. 222-227. Hargreaves, E. L. (1930), The National Debt. London: Arnold.

28

WEILLER

AND MIROWSKI

Heim, C., and Mirowski, P. (1987), “Interest Rates and Crowding Out in the British Industrial Revolution.” Journal of Economic History 47, 117-139. Hicks, J. R. (1967), Critical Essays on Monetary Theory. London/New York: Oxford Univ. Press. Hoppit, J. (1986), “Financial Crises in 18th Century England.” Economic History Review 39, 39-58. Hume, D. (1970) In E. Rotwein (Ed.), Hume on Economics. Madison: Univ. of Wisconsin Press. Kindleberger, C., and Laffargue, J. (1982) Financial Crises. Cambridge: Cambridge Univ. Press. Leijonhufvud, A. (1981), “The Wicksell Connection”, in his Information and Coordination. New York: Oxford Univ. Press. Malinvaud, E. (1972), Lectures on Mfcroeconomic Theory. Amsterdam: North-Holland. Mirowski, P. (1981), “The Rise (and Retreat) of a Market: English Joint Stocks in the 18th Century.” Journal of Economic History 41, 559-577. Mirowski, P. (1985), The Birth of the Business Cycle. New York: Garland. Mirowski, P. (1987), “What Do Markets Do ?” Explorations in Economic History 24, 107129. Mortimer, Thomas (1761) Every Man His Own Broker. London. Pressnell, L. (1956), Country Banking in the Industrial Revolution. London/New York: Oxford Univ. Press. Pressnell, L. (1960), “The Rate of Interest in the Eighteenth Century” In L. Pressnell (Ed.), Studies in the Industrial Revolution. London/New York: Oxford Univ. Press (Athlone). Sen, S. R. (1957), The Economics of Sir James Steuart. Cambridge, MA: Harvard Univ. Press. Shiller, R. J. (1979), “The Volatility of Long Term Interest Rates and Expectations Models of the Term Structure.” Journal of Political Economy 87(6), 1190-1219. Shiller, R. J. (1981), “The Use of Volatility Measures in Assessing Market Efficiency.” Journal of Finance 36(2), 291-304. Shiller, R. J., and McCulloch, J. H. (1987), The Term Structure of Interest Rates, NBER Working Paper 2341. Smith, Adam (1976), An Inquiry into the Nature and Causes of the Wealth of Nations. Chicago: Univ. of Chicago Press. Thornton, Henry (1802), An Inquiry into the Nature and Effects of the Paper Credit of Great Britain. London: Hatchard. Williamson, J. (1984), “Why Was British Growth So Slow During the Industrial Revolution?” Journal of Economic History 44, 687-712.