Prizes, Patents and the Search for
Longitude
M. Diane Burton
Tom Nicholas
Working Paper 16-046
Prizes, Patents and the Search for
Longitude
M. Diane Burton
Cornell University
Tom Nicholas
Harvard Business School
Working Paper 16-046
Copyright © 2015, 2016 by M. Diane Burton and Tom Nicholas
Working papers are in draft form. This working paper is distributed for purposes of comment and discussion only. It may
not be reproduced without permission of the copyright holder. Copies of working papers are available from the author.
Prizes, Patents and
the Search for Longitude
M. Diane Burton†
Cornell
Tom Nicholas*
Harvard Business School
March, 2016
Abstract
The 1714 Longitude Act created the Board of Longitude to administer a large
monetary prize and progress payments for the precise determination of a ship’s
longitude. It is frequently cited to justify the use of prize-based incentives over
patents. We use new data on marine chronometer inventors from 1714 to 1939 to
examine long run technological development. We show that while the initial prize
spurred entry by key inventors and progress payments helped to finance interim
R&D, patent usage was also high. We highlight a complex interplay between
prizes and patents over the life cycle of the industry.
Keywords: Prizes, Patents, Innovation
†
ILR School, Cornell University 170 Ives Hall Ithaca, NY 14853, USA. Email: burton@cornell.edu * Soldiers Field
Road, Boston, MA 02163, USA. Email: tnicholas@hbs.edu. We are very grateful to Steve Broadberry, Zorina Khan,
Naomi Lamoreaux, Ramana Nanda and Joachim Voth for comments and to Richard Dunn and Rory McEvoy from
the Royal Observatory Greenwich for sharing their expertise on chronometers and for invaluable help with data
sources. The Division of Research and Faculty Development at Harvard Business School provided funding.
I. INTRODUCTION
The race for global trade and maritime supremacy led the Spanish monarchy to offer a prize
for the discovery of longitude at sea during the late 1590s, to be followed in the early 1600s by
the Dutch Republic’s sequence of prizes for finding “the East and West” (Davids, 2011).1 In
response to a prominent navigation disaster and growing demands for a solution to the longitude
problem, a 1714 British Act of Parliament created an even larger award.2 Using a prize of up to
£20,000 (around £2.5 million today) a commission of adjudicating experts (known as the Board
of Longitude) and resources for inventors to engage in experimentation, the government aimed to
encourage knowledge accumulation in an area of high social value but relatively low private
investment. By the early 1770s following a long and acrimonious dispute with the Board of
Longitude, John Harrison (1693-1776) an English clockmaker, was awarded monetary values
approximately equivalent to the prize for his marine chronometer innovations.3
The longitude prize is frequently cited in the innovation literature as the most prominent
example of a non-patent based mechanism designed to spur technological development (Wright,
1983; Kremer, 1998; Kalil, 2006; Kremer and Williams, 2010; Williams, 2012; Brunt et al.,
2012; Moser and Nicholas, 2013; Murray et al., 2014). It has motivated a range of modern prize
competitions. For example, the recent X-Prize contests for space innovation, developments in
super-efficient vehicles or cost-effective gene sequencing revolve around a substantial monetary
award. The America COMPETES Act passed by Congress in 2010 provides Federal agencies
with the necessary authority to conduct prize competitions. In the spirit of the 1714 Act, the 2014
Longitude Prize supported by the British government consists of a £10 million prize fund in an
1
Phillip II initiated the prize offer in 1597 which was formalized by Phillip III in 1598 as a prize of 6,000 ducats
plus an annual annuity of 2,000 ducats and a gratuity of 1,000 more. In 1600 the States General of the United
Provinces of the Dutch Republic offered a prize of 5,000 guilders plus an annual annuity of 1,000 guilders. The
Dutch raised the prize amount to as much as 50,000 guilders by 1738, although there is some uncertainty in the
historical record concerning the exact amount of the prize award (Gould, 2013, p.12). Assuming this latter amount is
accurate, it would be equivalent to about £450,000 today.
2
In 1707 during the War of the Spanish Succession, Admiral Sir Cloudesley Shovell led a fleet of British naval
ships back to England from Gibraltar but the fleet lost its position in fog around the Isles of Scilly off the Cornish
peninsula of Britain, leading to the loss of four ships and almost 2,000 lives. Also, in July 1713 two mathematicians
William Whiston and Humphrey Ditton published an idea in The Guardian newspaper whereby ships in set
locations in sea trade channels would explode a mortar at a set height and time to allow location by other ships to be
determined by calculating the time elapsing between the explosion flash and the corresponding sound. They also
engineered a petition submitted to Parliament seeking to establish an award for a method to establish longitude.
3
It is important to note that the longitude prize was not actually awarded. The monetary amounts given to Harrison
represented only a tacit admission that the navigational problem had been solved.
2
effort to solve the problem of global antibiotic resistance. This is an area with substantial social
value that has not been addressed through patent-based incentives.
Several studies have analyzed the search for longitude in the eighteenth century from an
historical or a legal perspective (e.g., Sobel, 1996; Siegel, 2009). Harrison’s efforts are especially
well-documented (e.g., Andrewes, 1996). There is considerable debate about the Board of
Longitude’s decision to refuse payment to Harrison given that a key member, the Astronomer
Royal Nevil Maskelyne (1732-1811) favored “high-science” lunar observation methods relative
to Harrison’s “low-science” mechanical solution. On the other hand, we know very little about
the process of technological development that led to the navigation challenge being solved. No
empirical analysis has been conducted on the inventors who developed marine chronometers or
on the relationship between the prize and the patent system. Hence, the longitude prize is
commonly referenced in the innovation literature in the absence of hard evidence to support the
hypothesis that it is the exemplar case for the superiority of prize-based incentives.
Using a new dataset on chronometer inventors this paper examines the effect of the longitude
prize focusing on three main areas: the extent to which the Board of Longitude influenced entry
and patenting; the propensity of inventors to patent chronometer inventions; and relative changes
in the level of chronometer patenting over time. Results from each of these areas underscore a
complex interplay between the prize and the patent system over time. Though we cannot
pinpoint causality in our empirical analysis, we use a mix of qualitative and quantitative data to
provide suggestive evidence for our argument that the development of the marine chronometer
over the life cycle of the industry reflected the contribution of multiple inducement mechanisms.
In particular we pay close attention to the set of prize and patent-based mechanisms that jointly
created an environment which spurred entry and competition by key inventors, provided
financing for experimentation and interim R&D, and facilitated the diffusion of ideas.
The period we consider from 1714 to 1939 covers the duration of the prize competition under
the Board of Longitude, the rise of the industry after technical challenges were overcome as “the
British turned the marine chronometer into an object of industrial manufacture and commercial
use” (Landes, 1998, p. 171), and its interwar decline when radio signals displaced navigation by
chronometers. In other words, we are able to take a long run perspective on the process of
technological development. Data were assembled from Chronometer Makers of the World
(CMW) compiled by Tony Mercer (1919-2012), the grandson of the founder of Thomas Mercer
3
Ltd., a prominent English watch and chronometer manufacturer established in London in 1858.
CMW includes information on individuals, firms and governments active in the marine
chronometer industry. CMW entries were hand-matched against a data base containing all
British patents granted from 1714 to 1939 to determine the propensity to patent.
Our analysis begins by focusing on the Board of Longitude era. The monetary prize offered
by the British government spurred entry and led to intense competition by a small group of key
inventors, including John Harrison. But the Board could not easily manipulate the direction
(lunar versus mechanical solutions) or the timing of technological change. We show this using
the Board’s interim progress payments (totaling £52,535 or around £6.6 million today) paid out
to inventors in certain years for successful developments, which do not predict entry or
patenting. The arrival rate of new inventions was highly uncertain due to the level of
technological complexity and the time it took to design, test and manufacture an accurate marine
chronometer. We argue instead that the progress payments performed a more valuable function
by providing research investment for improvement and experimentation. Harrison’s first marine
chronometer (called H1) was completed in 1735 but his final H4 chronometer was not completed
until 1759. More generally, we show that the number of active chronometer makers and
craftsmen did not peak until the 1870s and 1880s, more than 150 years after the Longitude Act
had been passed. This highlights that key aspects in the development of the industry depended on
innovation incentives that extended beyond the impetus created by the 1714 prize.
The Longitude Act did not prohibit inventors from acquiring patents. In fact, a primary
objective of the Board was to ensure the diffusion of knowledge across innovators, which was
difficult to achieve if inventors opted to protect their inventions through secrecy. Harrison did
not patent his inventions because he was concerned about reverse engineering and the Board
expended considerable efforts on encouraging him to disclose the details of his chronometers.
Another early entrant, John Arnold (1736-1799), used the patent system, which required
disclosure. We find a high propensity to patent among marine chronometer inventors relative to
benchmarks in the literature. Although patents can block innovation, an alternative interpretation
is that they provide an important basis for the market for inventions to function. Recent research
on the British patenting system indicates that patents were privately beneficial by helping
inventors to appropriate on the basis of their ideas while at the same time being socially
4
beneficial by promoting the disclosure of useful knowledge (Bottomley, 2014). In line with this
perspective we find a high demand for patenting despite the presence of the prize.
We also examine the determinants of patenting in the cross section of inventors during, and
after, the Board of Longitude era. We confirm econometrically that the prize did not preclude
patenting and show that the probability of using the patent system increased with the quality of
invention. Finally, we compare marine chronometer patenting to patenting in other technology
areas. Time varying estimates show that patent usage spiked during the 1820s, 1830s and 1840s
relative to patenting in a control group of scientific instrument patents, before attenuating later in
the life cycle of the industry. We link the relative spike in patenting to prize competitions
conducted by the British Admiralty at the Royal Observatory in Greenwich, London to promote
cumulative improvements to Harrison’s marine chronometer design. This evidence reinforces
our argument that the development of the industry from its nascent eighteenth century stage
revolved around a complex link between prizes and patents.
Overall our findings suggest that the development of the marine chronometer reflected an
environment in which multiple innovation incentives were available and adopted by inventors.
This heterogeneity is overlooked when the development of the marine chronometer is
contextualized as a simple contrast between prizes and patents. We maintain that the prize
provided the catalyst for skilled inventors to direct their efforts towards solving the longitude
problem because patents alone could not generate sufficient incentives for private investment. As
the prize provided a boost to innovation, inventors used patents to secure private returns.
Because the patent system mandated inventors to disclose, it simultaneously corrected for a
defect in the design of the longitude prize where a lack of disclosure created a barrier to
incremental invention. Patenting became even more important as the industry focused on
cumulative refinements to chronometers and the design of marketable products. A simple
account of award-driven innovation cannot explain our results. Rather, a more complex set of
prize and patent incentives influenced the behavior of inventors developing what Mokyr (2010,
p.42) describes as “one of the epochal innovations of the eighteenth century”.
The remainder of the paper is structured as follows. In the next section we outline in detail
the main aspects of the longitude prize and the functioning of the Board of Longitude. Section
three describes the dataset and section four presents the results. Section five concludes with an
assessment of the broader implications of our results and caveats to our analysis.
5
II. INSTITUTIONAL DETAILS AND ASPECTS OF THE LONGITUDE PRIZE
To determine location at sea it is necessary to know both latitude and longitude accurately.
However, while the measurement of latitude at the time of the Longitude Act involved using a
simple scientific instrument (a sextant) to observe the angle between the horizon and a known
celestial object such as the Sun or the North Star, finding longitude was far more complex.4
Although the theory was well-known (every 24 hours the earth turns through 360 degrees, so
each 15 degrees of longitude corresponds to a time-difference of one hour) without highly
accurate sea-based timekeeping it was impossible to compare the local time with the time at a
particular reference point. Longitude could potentially be identified by a variety of celestial
methods, such as eclipses of Jupiter’s satellites, but with considerable measurement error. For
example, in 1713 Sir Isaac Newton pointed out that the lunar approach of measuring the distance
between the moon and a fixed star would establish the location of a ship only to within two or
three degrees – about 138 to 207 miles at the equator. This level of error was not much better
than navigating by the rudimentary method of dead-reckoning (Gould, 2013).5
In his comments to the committee established by the British government to determine the
feasibility of a prize for longitude, Newton suggested that an accurate time keeper was the most
promising potential solution.6 Because he also suggested it was unlikely that a watch overcoming
formidable technological challenges (coping with variable temperature, pressure and humidity)
would ever be constructed, the government created a substantial award for a practicable solution.
The Longitude Act of 1714 authorized the following payments to be made to anyone regardless
of nationality: £10,000 for a method that could determine longitude within one degree, £15,000
if it was determined within two-thirds of one degree and £20,000 for positioning within half a
degree.7 Although it is not known how the government arrived at these “shadow prices” for
innovation, the Act created a premier targeted (i.e., ex ante) prize.
4
A sextant measures the angle between two objects. At night, with clear skies, a navigator could determine a ship’s
latitude by the angle between the horizon and the North Star. The North Star is directly overhead the North Pole at
an angle of 90 degrees north latitude whereas at the Equator it is at zero degrees latitude. During daytime the angle
of the Sun above the horizon at noon could be used along with nautical almanacs to correct for the Sun’s declination.
5
Dead-reckoning involves using calculations about direction and distance traveled to pinpoint location, but is
complicated by the effects of factors such as wind drift and tide.
6
Newton stated: “[F]or determining the Longitude at Sea, there have been several Projects, true in the Theory but
difficult to execute[.] One is by a Watch to keep time exactly[,] But by reason of the Motion of a Ship, the variation
of Heat and Cold, Wet and Dry, and the Difference of Gravity in different Latitudes, such a Watch hath not yet been
made” (Gould, 2013, p. 13).
7
At the equator each degree of longitude is approximately 69 miles converging to zero miles at the poles.
6
The longitude prize had three key design elements of importance to our analysis. First, the
Act created the Board of Longitude, a 23 person body of elites like the President of the Royal
Society and the Speaker of the House of Commons, who ultimately met about three or four times
a year, typically at the Admiralty Building in London. At the Board’s discretion awards could be
made or withheld according to the criteria stipulated in the Act. With majority agreement one
half of an award could be paid out, with the remainder held back until the method had been
scientifically assessed on a voyage from a British port to a port in the West Indies. Further Acts
were passed between 1714 and 1828 to promote additional areas in need of navigational
improvement, clarify the powers of the Board and the conditions of the longitude award.
Second, the Board of Longitude provided progress payments. The 1714 Act stipulated a sum
not exceeding £2,000 could be sanctioned by a quorum of five Board members for inventors
making significant breakthroughs or interim research and development, but larger amounts were
also paid out. For example, Leonhard Euler (1707-1783) received £300 in 1765 for his
contribution to Tobias Mayer’s (1723-1762) lunar tables. Mayer’s widow received £3,000 the
same year as a posthumous award for her husband’s work. Importantly, the amounts paid by the
Board were counted against the value of the 1714 prize. In Harrison’s case he received £23,050
in a sequence of installments over more than three decades, which included the amount provided
to him to construct his chronometers (Gould, 2013, p. 67).8 During its 114 year existence the
Board of Longitude spent about £157,000 on items such as salaries for administrators and
expeditions. About one-third can be attributed to progress payments and prizes (Howse, 1998).
Third, the prize did not substitute for patenting, but rather created an additional layer of
incentives on top of the patent system. Notwithstanding the concept of intellectual property was
still rudimentary at the time, a means of establishing proprietary rights over invention did exist in
Britain (MacLeod, 1988). The 1624 Statute of Monopolies provided the basis of British patent
law until the 1852 Patent Law Amendment Act was passed (Alpin and Davis, 2013). Although
the 1624 Statute did not require detailed disclosure for a 14 year patent term, by the 1720s
comprehensive written descriptions had become the norm and they were widely diffused.9
8
There is some disagreement about the precise amount. Howse (1998, pp. 407-8) documents that Harrison received
a total of £22,000 from the Board of Longitude: one payment of £250 in 1737, six payments of £500 each between
1741 and 1760, £1500 in 1762, £1,000 in 1764, £7,500 in 1765 and a final payment of £8,750 in 1773.
9
The famous 1778 Liardet v. Johnson case established that the specification should include disclosures to allow
anyone skilled in the art to be able to replicate the invention.
7
Bottomley (2014, p. 28) describes patent specifications as being “diligently prepared, at least
from the 1770s” and reflecting “a uniquely reliable corpus of cutting edge technology.”
Records of the Board of Longitude reveal the interaction between prizes and patents. Thomas
Mudge (1715-1794), who patented the “lever” escapement in 1766, was awarded £3,000 by the
Board for design ingenuity and workmanship.10 Thomas Earnshaw (1749-1829) patented his
original invention of a “spring detent” escapement through a sponsor, while John Arnold
patented a variant of this technology in 1782 (Betts, 1996). Although the origins of the discovery
were contentious, both were awarded £3,000 by the Board of Longitude in 1805, the joint highest
amount for a mechanical solution, along with the sum paid out to Mudge.11
Generally, the Board was concerned with incentivizing inventors while also ensuring that
knowledge was diffused. In that sense allowing inventors to patent had benefits in terms of
opening technical ideas up to the public. This is perhaps one of the reasons why the Board
declined an offer made by Arnold in 1782 to “buy-out” his patents (Gould, 2013, pp. 113-114).
As Harrison opted for secrecy to protect his ideas, the Board was forced to take additional
measures to promote disclosure. Following a partially successful trail of Harrison’s H4 to
Jamaica and back in 1761-62, a 1763 Act was passed to encourage Harrison to reveal
information to make his chronometers fully replicable by British watchmakers.12 Because
placing this information in the public domain would disadvantage Harrison, the Act stipulated
that no other watchmaker could win the Longitude Prize within a four year time period.
In the decades after the 1763 Act was passed, refinements to Harrison’s H4 pocket watch
design focused on overcoming obstacles to accurate timekeeping at sea due to variable factors
such as heat, cold and lubrication. Manufacturers like Larcum Kendall (1721-1795), who was
tasked by the Board of Longitude with replicating Harrison’s design, aimed to produce at a price
that was practicable for widespread usage, while Mudge, Earnshaw and Arnold continued with
technical improvements to achieve greater accuracy. Earnshaw and Arnold in particular were
responsible for more cost-efficient production of pocket and box chronometers in what became
“the elite branch of the British watchmaking industry” (Davies, 1978, p. 524). Harrison’s
chronometer, or more specifically Kendall’s replica (called K1), was tested between 1771 and
10
An escapement stalls or creates mechanical movement in a chronometer according to a set rhythm.
John Roger Arnold received the award since his father died in 1799.
12
The 1763 Act was titled: “An Act for the Encouragement of John Harrison, to publish and make known his
Invention of a Machine or Watch, for the Discovery of the Longitude at Sea.”
11
8
1775 on James Cook’s second voyage of discovery and according to Cook’s assessment it had
“exceeded expectations” by the end of the expedition (O’Sullivan, 2008, p. 210).13 Logbooks of
the East India Company show the use of chronometers for navigation was systematized by the
1790s (Sobel, 1996, p. 162). By contrast, the Royal Navy was a late adopter as lunar distance
continued to be the preferred method. Less than 20 naval voyages per year used chronometers
before 1815, at a time when the Navy had almost 1,000 ships (Ishibashi, 2013, p. 56)
In 1774 Harrison successfully petitioned Parliament and received a final monetary award of
£8,750 to supplement the amounts he had already received from the Board of Longitude, yet this
represented only a tacit acknowledgement that the longitude problem had been solved. In fact,
the same year a new schedule of awards was established under an Act of Parliament that repealed
all preceding Acts. A prize of £5,000 was set for a method that could determine longitude within
one degree, £7,500 for within two-thirds of one degree and £10,000 for within half a degree.14
That is, the new legislation maintained exactly the same location parameters as the original 1714
Act, but with prize levels set at half the original amounts.
Several decades later, in 1818 the Board of Longitude was reorganized. More scientific
members from the Royal Society were added and new positions were created including a
Superintendent of Chronometers with the responsibility for administering trials at the Royal
Observatory in Greenwich. These took place as “premium trials” between 1823 and 1835 with
monetary awards for first, second and third place in an attempt to encourage cumulative
improvements in chronometer design and accuracy. The trials extended the stringent tests
according to scientific standards that the Board of Longitude had first initiated. Nevertheless, and
in spite of advances that it had made, critics of the Board argued it was institutionally inept. Most
notable among the critics was Member of Parliament John Croker, who was also the First
Secretary of the Admiralty, who argued in 1828 that the Board of Longitude “was of no use
whatever” (Howse, 1998, p. 405). The Board was dissolved by an Act of Parliament in 1828.
III. DATA CONSTRUCTION AND SUMMARY STATISTICS
Our empirical analysis is based on a dataset of chronometer inventors compiled from CMW.
Pioneers in the lunar approach to discovering longitude are not included in this source, though
13
H1 had been on a short sea trial to Lisbon.
The new Act also stipulated that if longitude was determined within given tolerances using solar and lunar
methods, the prize award would be £5,000. It also shifted the emphasis from longitude to navigation more generally.
14
9
we do observe progress payments made by the Board of Longitude to those working in this area,
which we do utilize in our empircs.15
Variables in CMW
CMW represents “a comprehensive list of makers and craftsmen... who worked in the
industry, their places of work and dates and serial numbers of their instruments.” Mercer
compiled the data using Thomas Mercer Ltd records and a series of supplementary sources
including extensive firm archives, records of the British Admiralty and the British Horological
Institute, and drawing also on a number of experts in marine history. His data include
government end users and public institutions like the British Ministry of Munitions established in
1915, which we exclude in our analysis. This exclusion, plus further data cleaning produced
1,336 observations from the original 1,701 entries in CMW. Table 1 provides descriptive
statistics for entry years during the Board of Longitude era (1714 to 1828), the post Board era
from 1829 to 1939 and the life cycle of the chronometer industry from 1714 to 1939.
Entry and exit are defined in CMW by “dates of occupation” – when activity in the industry
first started and then ceased. Although the list includes entry years up to 2006, we truncate the
data on the entry side to be between 1714 and 1939 because this is the period for which we have
patent data. Moreover, Mercer himself acknowledges a drop in data quality around 1939 as
records were lost during the Second World War. Entry and exit are provided in the majority of
cases but to maximize use of the data, we linearly interpolated when one of the dates was
missing. An active career was around 26 years for individual inventors. By comparison Khan
(2015) finds an average career length of 18 years for a pre-1820 cohort of famous British
inventors. For firms (just over one third of the data), an active period was about 44 years.
Figures 1 and 2 reveal the distributional characteristics of the data over the full non-truncated
range of observations in CMW. Prior to the longitude prize being established very few inventors
were engaged in this area of technical development. Pre-1714 entrants include Christiaan
Huygens (1629-1695), a famous Dutch scientist, mathematician and horologist, who attempted to
patent his discoveries in France. He built a marine chronometer in 1660 which used a pendulum
and a coiled spring to moderate mechanical movement. This timepiece, however, proved to be
15
Lunar methods were un-patentable but the prize had an important effect on knowledge diffusion through the
publication of ideas. Maskelyne argued that improvements in calculations by lunar distance should be made more
accessible by including them in a publication, which became the Nautical Almanac. It was first published in 1766
and widely diffused thereafter.
10
erratic in anything but calm seas. It was unable to correct for measurement errors caused by
temperature variations and gravity changes at different latitudes (Gould, 2013, pp. 27-30).
Even after the Longitude Act was passed entry into the industry was quite protracted. Figure
1 shows that peak activity for the individuals in the data occurred more than a century and a half
after 1714. The time scale associated with chronometers is quite distinct. In historical cases like
U.S. automobiles or tires, entry and exit dynamics played out over a much shorter interval of a
few decades with industry evolution typically characterized by a period of limited and then
general entry. In tire manufacturing, for example, Klepper and Simons (2001) document that a
peak in the distribution was reached in 1922 by which time 274 firms had entered. At this point
the industry was only about 25 years old. The kind of industry evolution shown in Figure 1
indicates that technological development in marine chronometers was a function of long run
dynamics that extended well beyond prize-based incentives created in 1714.
The reason for this timing was technological complexity. Although Harrison’s dominant
design was largely perfected by the 1770s, it took decades of incremental advance to make the
chronometer fully accurate, cost effective and commercially practicable. Chronometers consisted
of numerous complex parts including a movement, drawn by a weight or spring and an
escapement, all of which required significant breakthroughs to be made. Following the growth of
the industry during the Board of Longitude era and beyond, the interwar “shakeout” and decline
illustrated in Figure 2 were reasonably rapid. This was driven by a fundamental technology shift
away from navigation using marine chronometers and towards radio time signals.
The longitude prize was a “public reward” that was open to inventors of any nationality.
CMWs coverage is global, including chronometer makers like Pierre Le Roy (1717-1785) and
Ferdinand Berthoud (1727-1807), two prominent innovators and rivals residing in Paris. Outside
of Britain, 23 countries are represented, with the largest concentrations in France, Switzerland
the U.S. and Germany. Figure 3 shows the location of the makers and craftsmen within Europe
broadly defined. Britain was a significant location accounting for at least 70 percent of the
entries with clustering in London, the county of Lancashire, and in cities along the south and
northeast coast. These clusters reflected a division of labor in watch making. In 1800 70,000
workers were employed in the industry, mostly in Clerkenwell, London where final assembly
took place (New Scientist, 1957). Specialist craftsmen in Lancashire produced tools and partly
finished watch movements for the London market. Harrison’s early experiments culminating in
11
H1 were conducted in Barton-upon-Humber, in North Lincolnshire, but fewer than 3 percent of
CMW entries were located within 50 miles of there. Harrison moved to London in 1730.
Chronometers varied by the quality of craftsmanship. We were able to parse the data based
on a quality indicator of the individual or firm listed. According to Mercer (1991, p. ix) “a
marine chronometer was of no value whatsoever until it had passed the exacting tests laid down
by the British Admiralty.” In 1805 the Admiralty instigated a formal plan requiring that
chronometers could only be sold to the public when they had been officially tested. Moreover,
the best chronometer makers were permitted to use the prestigious title of “Maker to the
Admiralty” which was used in catalogues and advertisements to aid appropriability (Ishibashi,
2013). During the Board of Longitude era, 35 percent of CMW entries supplied the Admiralty
with chronometers falling slightly to 27 percent in the post Board of Longitude era.
Board of Longitude Awards and other Prize Competitions
We matched the names in CMW to individuals who received a Board of Longitude award
using the list in Howse (1998). Although the longitude prize was never technically awarded,
interim disbursements made by the Board can be thought of as progress payments or awards for
exceptional developments. 3 percent of the CMW entries for the period 1714 to 1828 received a
Board of Longitude award, with the largest cumulative amount being paid to Harrison.
Finally, we coded a dummy variable for whether an inventor received an award at other prize
competitions. The Board of Longitude was not the only source of public approbation for
innovation. Prize competitions were commonplace during the eighteenth and nineteenth
centuries, with a multitude of European, American and Asian industrial exhibitions. Mercer
includes biographical details on entries which we augmented with our own literature searches.
Table 1 shows 7 to 9 percent of CMW cohorts received a prize for chronometer innovation. For
example, Jean Adrien Philippe (1815-1894), a French horologist who co-founded Patek Philippe
& Co., received a gold medal at the 1844 French Industrial Exposition.
Matching to Patent Data
We identified patent usage by hand matching CMW names to a comprehensive dataset of
British patents, including those granted in Scotland, which had its own separate registration
system for patents prior to 1853.16 We matched by the title of the invention and by name in order
16
Ireland also had its own separate registration system prior to 1853, but most of the records no longer survive.
12
to get as close a correspondence as possible between the invention and inventor. Because the cost
of obtaining a patent was high (by the middle of the nineteenth century a patent could cost £120
in England and as much as £350 in Scotland and Ireland) liquidity constrained inventors had the
option to patent through a sponsor or an agent, who would sometimes have their own name on
the patent.17 Although we made every effort to correct for such instances, our method will lead to
a downward biased estimate of the propensity to patent because of this practice.
We find that 15.9 percent of inventors patented their technical discoveries during the Board
of Longitude era (Table 1) rising to 18.9 percent if we exclude all foreign domiciled inventors. In
the post-Board of Longitude era for entry cohorts after 1828, the figures are 18.1 percent and
20.3 percent respectively. Either way, these shares are much larger than the 9.6 percent patenting
rate observed for all inventors exhibiting scientific instruments at the Crystal Palace Exhibition
in 1851 (Moser, 2012, p.55), a category that would include chronometers.
Furthermore, the propensity to patent was even higher among the best inventors. 63 percent
of inventors who received a Board of Longitude progress payment patented. This can be
compared to 15.8 percent of award winning inventors exhibiting scientific instruments at Crystal
Palace who patented. At a time when it was common for inventors not to patent, chronometer
inventors appear to have disproportionately protected their technical developments using formal
measures. Clearly, the longitude prize did not preclude patenting and in the rest of the paper we
explore this aspect of the competition and its implications in more detail.
IV. EMPIRICAL STRATEGY AND RESULTS
The Effect of the Board of Longitude on Entry and Patenting
We first examine entry and patenting in relation to incentives for innovation created by the
Board of Longitude. A key requirement of a prize competition is that it stimulates competitive
entry. It is reasonable to assume that the 1714 longitude prize met this objective. Harrison was
motivated by the prospect of winning and the most prominent horologists of the seventeenth and
eighteenth centuries were incentivized to discover a solution to the longitude problem. Moreover,
the contest was not restricted to experts, so it was associated with a specific type of entry
dynamics. Its design was analogous to the modern crowd sourcing tournaments examined by
17
Thomas Earnshaw could not afford to patent given that fees were so high. He received financial assistance from a
watchmaker, Thomas Wright, who used his own name to patent Earnshaw’s escapement invention. Even by the
middle of the nineteenth century patent fees were about four times larger than per capita income (Khan, 2005, p. 31).
13
Boudreau et al., (2011). They show that if the problem to be solved is particularly difficult,
opening up to a broad array of entrants can elicit the production of high-value long-tail ideas
beyond what could be achieved by restricting the tournament to domain experts alone.
Widespread entry meant that the longitude prize attracted practically-oriented inventors like
Harrison, not just experts in celestial methods. Qualitative accounts indicate that the
announcement of the longitude prize encouraged entry and competition.
An important additional issue is the extent to which prize administrators can influence the
timing of entry through the certification of particular areas of innovation (Kalil, 2006). Given the
richness of the data available from the Board of Longitude era, we can implement a test. We use
the monetary amounts paid out by the Board to inventors for successful experimentation as a
proxy for “certification” of particular approaches. The first payment of £250 was made to
Harrison in 1737 for his H1 and H2 chronometers. We consider awards in aggregate and we also
distinguish between awards for “mechanical” and “lunar” methods. While it is often argued that
the Board was against mechanical solutions, Figure 4 shows that by around the 1770s it was
vested in this area of development. If these awards created incentives, we might expect to see a
response in the data through an effect on entry rates and patenting.
To gain a better understanding of how the progress payments functioned, we exploit time
series variation in awards to determine if these predict changes in entrants and chronometer
patents. We use variants of the following specification where t indexes time and AWARD is the
logarithm of monetary awards. We allow awards to influence our outcome measures with a one
period lag, though all of our results are robust to including longer lag structures as well. We also
use a one period lag of the outcome variable as an additional control, which helps to mitigate
omitted variable bias under the assumption that unobserved factors that would influence entry
(such as common trends) will be correlated with the prior year outcome.
OUTCOMEt   1 AWARDt  1   2OUTCOMEt  1  t
Results in Table 2A are from negative binomial regressions given the count distribution of
the outcome measures. Baseline estimates suggests a positive relationship between awards in the
prior year and entry in the current year. Column 1 shows a statistically significant relationship
between progress payments and entrants, implying that a 10 percent increase in awards at time t1 is associated with a 1.1 percent increase in entry at time t. The coefficient on awards in column
14
2 is also statistically significant when controlling for lagged entry. That is, the Board of
Longitude progress payments have explanatory power even in the presence of natural dynamics
where the previous level of entry activity is a strong positive predictor of current entry.
However, when we disaggregate the award payments into “mechanical” and “lunar”
components the regressions produce counter intuitive results. We would expect entrants into the
chronometer industry to respond positively to changes in mechanical awards if there was any
Board of Longitude certification effect, but the relationship is actually negative in columns 3 and
4. By contrast, in columns 5 and 6 we find positive and statistically significant coefficients on
variables measuring lunar awards. When conditioning on both award types in the same
specification in columns 7 and controlling for dynamics in column 8, the pattern in the
coefficients remains the same. In column 9 we take the additional step of restricting the entry
count to only CMW entries domiciled in Britain and include covariates controlling for British
real GDP and population using the data provided by Broadberry et al., (2015). Yet we still find
no systematic evidence of a link between entry and the Board of Longitude payments. Figure 5A
shows that there is no observable relationship present in the raw data.
Neither do we detect any economically meaningful associations when we use chronometer
patents as an outcome variable in Table 2B. We find no evidence of a statistically significant
effect of awards on patents in columns 1 or 2. And when using mechanical and lunar awards as
predictors (columns 3 to 8) the coefficients are also imprecisely estimated. In a similar way to
the entrant specifications, we find the most important determinant of current period patents in
these regressions to be prior period patents. In fact it is only the autoregressive terms in the
specifications that have any statistical power. Column 9 uses real GDP and population controls
and the coefficients remain imprecisely estimated. Even a parsimonious model does not generate
a statistically significant coefficient on mechanical awards. Figure 5B plots the raw relationship
between patents and mechanical awards, again showing no evidence of a systematic pattern.
There are two potential explanations for these null results. First, the level of technological
complexity associated with chronometer innovation meant the arrival rate of entry and patents
was highly uncertain. This is consistent with the protracted pattern of entry shown in Figure 1,
where over a century and a half elapsed from 1714 to the industry reaching its peak. Entry and
patenting may have been subject to some degree of path dependency set in motion by the
announcement of the 1714 prize, but the Board of Longitude did not determine the direction of
15
innovation or the timing of entry and patents beyond this point in time. Second, we are
generating null results because the confidence intervals are wide and not because of a precisely
estimated zero effect. The spread in the confidence interval is plausibly driven by idiosyncrasies
in individual behavior. Some inventors may have been strongly impacted by these progress
payments, whereas for others the entry and patenting decision was probably more stochastic. We
suspect that the results are at least partly a reflection of this underlying heterogeneity.
This is not to say that the progress payments were irrelevant. To the contrary, they performed
a central role by providing financing for intermediate research and development efforts and for
experimentation. As Macleod (1988, p. 193) points out, “without the grants made him by the
Board… it is most unlikely that Harrison could have perfected his chronometer.” The effect was
not confined to Harrison. The Board awarded the horologist Thomas Mudge £500 in 1777 for
further experiments associated with precision watchmaking. In that sense the payouts functioned
more like modern government procurement contracts with progress payments, or “advance
market commitments” where R&D funding is often allocated in step-by-step increments (Kremer
and Williams, 2010). These types of stimulus may be more likely to affect the intensive margin
of technological development. This would help to explain why our regression specifications
reveal no effect of progress awards on the extensive margin of entry and patenting.
Inventor-level Correlations with Patenting
Next, we turn to the cross-sectional analysis of patenting at the inventor-level for the Board
of Longitude and the post-Board eras. For each of the CNW entries we constructed an indicator
for patent usage. Although we do not exploit the timing of patenting with respect to awards given
the limited number of observations we have, Figure 6 shows that pre- and post-award patenting
were both possible. While around two-thirds of patenting occurred roughly contemporaneously
with an award, patenting either side can also be observed in the data. For example, Arnold first
patented in 1775, four years after he received a Board of Longitude award, whereas Earnshaw
patented in 1783, 17 years prior to his first award. These observations indicate that there was not
a simple sequential relationship between prizes and patents. Indeed, insofar as the progress
payment awards were made by the Board for intermediate innovation and R&D, it may have
been optimal for inventors to patent ideas at these different stages of the technology cycle.
Beyond this basic descriptive evidence, we can enhance our understanding of the factors that
were correlated with the discrete decision to patent. We use variants of the probit specification
16
below where i indexes inventors. We measure awards using both the logarithm of monetary
values received and an indicator variable coded 1 for an award and 0 otherwise. PRIZE is an
indicator for awards received in other prize competitions, which provides us with a benchmark
for comparing the propensity to patent for longitude award winners. BL is coded 1 for inventors
entering the chronometer industry in the post-Board of Longitude era (i.e., after 1828) to control
for any changes in the propensity to patent by specific cohorts of inventors. We use a vector of
control indicators (X) to identify the geography of inventors (Britain is coded 1), firms versus
individual inventors and whether or not an inventor was a “Maker to the Admiralty.”
Pr(Patent  1)   1 AWARDi   2 PRIZEi   3 BLi  X i  i
Table 3 presents estimates of marginal probability effects. The use of patents by chronometer
inventors during the Board of longitude era was increasing in the monetary value awarded. The
estimates in column 1 show that a doubling in the award value is associated with an increase in
the probability of patenting of 4.7 percent. The economic magnitude of this estimate remains
approximately the same when controlling for other covariates in column 4. A discrete change
had a much larger effect. Column 2 shows that wining an award is associated with an increase in
the probability of patenting of 49 percent, which is robust to including controls in column 5.
A sense of the economic magnitude of the discrete change estimates can be gained by a
comparison to the coefficients for prize winners at other competitions. The longitude award
effect in columns 2 and 4 is much larger than the other prize winner effect in columns 3 and 5 –
it is about 4 times larger in column 7 when the effects are jointly estimated. Furthermore, in
column 7 the longitude award coefficient is statistically larger than the prize coefficient under a
Wald test (χ²=4.08, p=0.044). A plausible explanation is that the Board of Longitude effectively
screened for quality. Evidence from industrial exhibitions shows that high quality inventors tend
to win prestigious prizes and have a higher probability of patenting (Moser, 2012).
Perhaps surprisingly we do not find a statistically significant effect on patenting for inventors
located in Britain, under the assumption that it was easier to manage complex patent filing
procedures domestically as opposed to from abroad (Khan, 2005). On the other hand, given that
patent agents could facilitate transactions from overseas and London was the main center of the
watch making industry during this time period, even foreign inventors would have had incentives
to protect their intellectual property rights against expropriation in this geographic area.
17
We do not detect any differences between individuals and firms in terms of patenting
probabilities, but we do find that inventors designated as a “Maker to the Admiralty” were more
likely to patent. During the late eighteenth and early nineteenth centuries the Admiralty’s
demand for chronometers increased and it put in place an institutional framework (including
using the Royal Observatory at Greenwich as a store of government-owned time pieces) to
ensure the supply of marine chronometers to Navy captains (Ishibashi, 2013). To the extent that
being a “Maker to the Admiralty” reflected a coveted standard of excellence, the estimates
suggest that patents were positively associated with the quality of technological discovery.
Finally, we find limited evidence of a change in the probability of patenting conditional on
entry cohort. In column 8 we do not find any statistically significant interactions on variables
measuring inventors entering the chronometer industry in the post-Board era. Although, the main
effect of the Board of Longitude variable across the specifications is positive and sometimes
statistically significant, it is always economically small. This is consistent with the statistics
shown in Table 1 where the share of inventors patenting only marginally increases over time.
Changes in the Chronometer Patents Relative to other Technology Areas
The last component of our analysis examines the changing role of patents over the life cycle
of the chronometer industry. Although we have shown that the propensity to patent was high
among inventors engaged in this technology area, this does not mean that the level of patenting
was also high. The propensity to patent is an intensive measure whereas we are also interested in
changes along the extensive margin. Figure 7 illustrates a striking pattern. A small number of
chronometer patents (an average of 0.13 per year) were granted during the Board of Longitude
era, but the use of patents increased substantially to an average of 2.30 patents per year in the
period from the Board’s dissolution in 1828 to 1939. Towards the end of the time period the
number of patents granted began to fall. One explanation is that patents became more important
in the post-Board era due to commercialization efforts and cumulative improvements to the
chronometer’s design. As the industry faced relative decline, patents became less important.
Of course, the changes in patents shown in Figure 7 may also be driven by contemporaneous
demand shocks or patent policy reforms. An average of 46.5 British patents were granted per
year during the Board of Longitude era compared to an average of over 8,100 per year between
1828 and 1939. In an attempt to explore drivers of the trends in Figure 7 in a more robust
framework, we construct a panel to compare changes in chronometer patents to patents in two
18
technology areas. First, we benchmark chronometer patents against all British patents and second
(because variation in demand conditions or the rate of technological development may
fundamentally differ across industries) we construct a more closely comparable technology area
consisting of only scientific instrument patents. In the absence of an official concordance of
individual patents to technology categories for our time period, we follow the procedure outlined
in Brunt et al., (2012) and use keywords to identify these inventions from patent titles.18
We use negative binomial difference-in-differences specifications to measure relative
changes in the level of patents. This allows us to account for different baseline rates over time
and to control for the general upward trend in patenting. Due to the length of the timespans being
considered in Figure 7, we do not use a single post-period setup but rather allow for flexible
time-varying estimates. Given that we observe a small number of yearly patents in the data for
some of the time period, we collapsed the observations into a dataset of decade-level means.19
In the specification below j indexes technology categories and t indexes time. CHR is an
indicator coded 1 for chronometer patents and 0 for patents in the comparison category and TD
represents time period decade dummies. The technology area fixed effect ϕj captures any
systematic time-invariant differences in patenting across technology areas, while the time fixed
effect γt controls for unobservable factors such as variation in the cost of accessing a patent that
may have caused changes in patenting across all technology areas.
PATENTSjt   1CHRj  TDt  j  t  jt
Figure 8 plots point estimates and 95 percent confidence intervals for δ1 from the 1800s to
the 1930s relative to a baseline from the 1710s to the 1790s. Compared to all British patents in
Figure 7A we find no evidence of a change in chronometer patents until the 1870s when the
coefficients turn strongly negative. Relative to scientific instrument patents (Figure 7B) we also
observe negative effects from the 1890s, but the estimates are positive and precisely estimated
18
We derived list keywords from publications such as Gerard L’Estrange Turner’s, Scientific Instruments 15001900. These are: accelerometer, ammeter, amperage, anemometer, armillary, astrolabe, barometer, caliper,
calorimeter, catheter, chartometer, circumferentor, clinometer, dental, dipleidoscope, dynamometer, eidograph,
electrometer, electroscope, ellipsometer, graphometer, gravimeter, hodometer, hygrometer, inclinometer,
interferometer, kaleidoscope, lithotomy, magnetograph, magnetometer, manometer, micrometer, microscope,
nocturnal, ohmmeter, opisometer, optical, oscilloscope, pantograph, pedometer, planetarium, planimeter, reflector,
refractor, sand glass, scale, seismometer, sextant, spectrogram, spectrometer, spectrometer, stereoscope, sundial,
telescope, theodolite, thermocouple, thermometer, voltmeter, zograscope.
19
Convergence failures meant that we could not consistently estimate the models with annual data.
19
for the 1820s to the 1840s. Coefficients across these decades imply approximately a seven-fold
relative increase in patents. We do not detect any change in chronometer patents compared to
scientific instrument patents based on the estimated coefficients for the 1800s and the 1810s.
Because the patenting trends are close for these decades, this provides another baseline
comparison. A Wald test shows that coefficients from the 1820s to the 1840s are jointly
statistically different from the coefficients for 1800s and the 1810s (χ²=115.14, p=0.000).
Interestingly this spike in patenting coincides with intense competition associated with the
refinement of chronometers. Specifically, the Admiralty conducted trials at Greenwich to
improve the quality of chronometers being supplied to the Navy. The Board of Longitude had
initially sent Harrison’s chronometers there to be subjected to scientific scrutiny, and the
Admiralty’s “premium trials” from 1823 to 1835 reflected a continuation of this effort. Although
the trials did not solve the problem that chronometers tended to lose time during extreme
temperature fluctuations, the best chronometers achieved high standards of precision. For
example, according to the formula that was used to measure accuracy, a+2b, a chronometer
developed by a respected London-based maker Robert Molyneux (d. 1876) was awarded a £200
prize at the “premium trials” in 1832 (a+2b=12.3 comprising of 6.7 for a and 2.8 for b).20
Molyneux patented in 1840. We find a significant boost to patents during the premium trial era,
implying that these incremental prizes may have intersected with the patent system to provide a
mixture of incentives to promote further technological development.
Equally, by around the 1840s chronometers were sufficiently well made and durable that they
rarely needed to be replaced. Although the cost of accessing a British patent fell substantially
with the modifications to patent law over time, patents became increasingly redundant in an
industry that focused more on manufacturing cost efficiency. Kendall’s replica of Harrison’s
chronometer cost £450 to construct in 1769 (approximately £55,000 today). By 1840 a
chronometer could be produced for £40 (Davis, 2006, p. 515). As the chronometer industry
reached technological maturity, patenting inevitably began to attenuate. Private returns, imitation
and expropriation risk would have diminished as the technology approached obsolescence.
20
Daily measurements were taken against a synchronized clock with a being the difference in seconds between the
greatest and the least weekly sums of the daily rates and b being the greatest difference between two consecutive
weekly sums of the daily rate rates. The metric had a scientific and a practical rationale: a would reflect week-toweek precision, a key factor on short voyages and b would reflect long-term precision, a key factor on distant
voyages. A smaller value of a+2b was better. Measurement occurred over 29 weeks under variable simulated
temperature conditions. See further, Gould (2013, pp. 257, 267-272).
20
V. CONCLUSION
The longitude prize is the archetypal example of a government sponsored attempt to promote
technological development. It is oft-cited in the literature on the economics of innovation and by
policy makers considering the design, or administration of intellectual property rights regimes.
Typical accounts assert that the longitude prize led to a major technological breakthrough and so
it is frequently used in policy discussions to suggest that prizes can successfully substitute for
patents. Yet, such strong normative inferences are not justified by the available evidence. There
are still major gaps in our understanding of how the longitude prize actually functioned, and also
in the development of the marine chronometer industry more generally.
We have combined qualitative and quantitative information with a view to providing a more
comprehensive analysis of how innovation incentives influenced the way the navigational
challenge was solved. Our results suggest that the search for longitude did not reflect a contrast
between prize and patent-based incentives; rather the prize did not preclude patenting and
patenting was not precluded by the prize. While patents had not created sufficient incentives to
generate private investment in this area of navigational science, the initial 1714 prize led to entry
by a group of inventors many of whom still relied on patents. Patenting was advantageous from a
government policy perspective because it led to the disclosure of proprietary ideas.
Furthermore, the industry developed for reasons beyond the initial spur to entry created by
the longitude prize. Although our empirical analysis shows that the Board was unable to
precisely control the direction of innovation over time, its progress payments for experimentation
facilitated research and development and incremental invention. Notably, most of those awarded
payouts also used patents. As the industry progressed during industrial maturity and a dominant
design was established, premium award trials were used by the Admiralty to support cumulative
innovation, which also coincided with patent usage. Our analysis underscores the presence of a
multifaceted system of prize and patent-based innovation incentives affecting the behavior of
inventors over the life cycle of the marine chronometer industry.
What we cannot do in this paper is to establish causal relationships or make hard inferences
concerning welfare. Over the duration of the Board of Longitude, the government did not insist
on a transfer of patent rights, nor did it engage in a buyout of inventors who did patent influential
inventions. The literature on the theory of optimal innovation incentives has long debated the
type of incentives that maximize social welfare (e.g., Wright, 1983; Kremer, 1998; Shavell and
21
Ypersele, 2001; Chari et al., 2012; Spulber, 2014). Even the most optimistic proponents of prizes
from a theory perspective note the potential for downside risk. While Polanvyi (1944) argued
that government sponsored prizes would create a welfare-efficient solution to incentivizing
innovation, he also indicated the potential for “corruption and arbitrary oppression”. Khan (2014)
finds imperfect administration in several cases of historical prize competitions.
Our analysis should be interpreted more as providing suggestive observational evidence that
the longitude prize and patents had jointly beneficial attributes in this historical context. The
intersection of prize and patent-based incentives helped to create an inducement effect to spur
entry, financed interim R&D and promoted disclosure, thereby helping to facilitate a major
technological breakthrough in an area of high social value. It is also important to clarify the
circumstances surrounding the use of the longitude prize given its preeminence in policy
discussions where the development of welfare beneficial innovations is often thought to be
constrained under market mechanisms and patents.
22
References
Aplin, Tanya and Jennifer Davis. 2013. Intellectual Property Law. Oxford, Oxford University
Press.
Betts, Jonathan. 1996 “Arnold & Earnshaw: The Practicable Solution” in William J. H.
Andrewes, ed. The Quest for Longitude: The Proceedings of the Longitude Symposium
Cambridge, Mass.: Harvard University Press., pp. 311-330.
Bottomley, Sean. 2014. The British Patent System during the Industrial Revolution 1700-1852.
Cambridge, Cambridge University Press.
Boudreau, Kevin J., Nicola Lacetera, and Karim R. Lakhani. 2011. “Incentives and Problem
Uncertainty in Innovation Contests: An Empirical Analysis.” Management Science 57(50:
843-863.
Broadberry, Stephen, Bruce Campbell, Alexander Klein, Mark Overton and Bas van Leeuwen.
2015. British Economic Growth, 1270-1870. Cambridge, Cambridge University Press.
Brunt, Liam, Josh Lerner and Tom Nicholas. 2012. “Inducement Prizes and Innovation”, Journal
of Industrial Economics 60(4): 657-696.
Chari, V.V., Michael Golosov, and Aleh Tsyvinski. 2012. “Prizes and Patents: Using Market
Signals to Provide Incentives for Innovations.” 147(2): 781-801.
Davids, Karel, 2009. “Dutch and Spanish Global Networks of Knowledge in the Early Modern
Period: Structures, Changes and Limitations,” VU University Amsterdam Working Paper.
Davies, Alun C. 1978. “The Life and Death of a Scientific Instrument: The Marine Chronometer,
1770-1920,” Annals of Science 35(5): 509-525.
Gould, Rupert T. 2013. The Marine Chronometer, Its History and Development. Woodbridge,
Suffolk, England.
Howse, Derek. 1998. “Britain’s Board of Longitude: the Finances, 1714-1828,” The Mariner's
Mirror 84(4): 400-417.
Kalil, Thomas. 2006. “Prizes for Technological Innovation.” Brookings Institution Working
Paper no.8.
Khan, B. Zorina. 2005. The Democratization of Invention: Patents and Copyrights in American
Economic Development, 1790-1920. Cambridge, Mass.: Cambridge University Press.
Khan, B. Zorina. 2014. “Inventing in the Shadow of the Patent System: Evidence from 19thCentury Patents and Prizes for Technological Innovations” NBER Working Paper.
Khan, B. Zorina. 2015. “Knowledge, Human Capital and Economic Development: Evidence
from the British Industrial Revolution, 1750-1930” NBER Working Paper.
Klepper, Steven and Kenneth L. Simons, 2000. “The Making of an Oligopoly: Firm Survival and
Technological Change in the Evolution of the U.S. Tire Industry,” Journal of Political
Economy 108(4): 728-760.
Kremer, Michael. 1998. “Patent Buyouts: A Mechanism for Encouraging Innovation,” Quarterly
Journal of Economics 113(4): 1137-1167.
Kremer, Michael and Heidi Williams. 2010. “Incentivizing Innovation: Adding to the Tool Kit”,
Innovation Policy and the Economy 10: 1-17.
Landes, David S.1983. Revolution in Time. Cambridge, Mass.: Harvard University Press.
MacLeod, Christine. 1988. Inventing the Industrial Revolution. The English Patent System,
1660-1800 Cambridge. Cambridge University Press.
Mercer, Tony. 1991. Chronometer Makers of the World. London, N.A.G. Press.
23
Mokyr, Joel. 2010. “The Contribution of Economic History to the Study of Innovation and
Technical Change: 1750-1914,” in Bronwyn H. Hall, and Nathan Rosenberg, Handbook of
the Economics of Innovation, Volume 1. North Holland, Elsevier, pp., 11-50.
Moser, Petra. 2012. “Innovation without Patents: Evidence from World’s Fairs,” Journal of Law
and Economics, 55(1): 43-74.
Moser, Petra and Tom Nicholas. 2013. “Prizes, Publicity and Patents: Non-monetary Awards as
a Mechanism to Encourage Innovation,” Journal of Industrial Economics 6(3): 763-788.
Murray, Fiona, Scott Stern, Georgina Campbell and Alan MacCormack, 2012. “Grand
Innovation Prizes: A Theoretical, Normative and Empirical Evaluation.” Research Policy,
41(10): 1779-1792.
Nicholas, Tom. 2013. “Hybrid Innovation in Meiji Japan,” International Economic Review
54(2): 575-600.
Officer, Lawrence H. and Samuel H. Williamson. 2014. “Six Ways to Compute the Relative
Value of a U.K. Dollar Amount, 1270 to present,” MeasuringWorth, 2014
Polanvyi, Michael. 1944. “Patent Reform,” Review of Economic Studies 11(2): 61-76.
Shavell, Steven and Tanguy van Ypersele. 2001. “Rewards versus Intellectual Property Rights,”
Journal of Law and Economics 44(4): 525-547.
Siegel, Jonathan R. 2009. “Law and Longitude,” Tulane Law Review 84(1): 1-19.
Sobel, Dava. 1996. Longitude: The True Story of a Lone Genius Who Solved the Greatest
Scientific Problem of His Time. New York: Penguin Press.
Spulber, Daniel. 2014. “Prices versus Prizes: Patents, Public Policy, and the Market for
Inventions,” Northwestern University Working Paper.
Wright, Brian. 1983. “The Economics of Invention Incentives: Patents, Prizes, and Research
Contracts,” American Economic Review 73(4): 691-707.
Williams, Heidi. 2012. “Innovation Inducement Prizes: Connecting Research to Policy,” Journal
of Policy Analysis and Management 31(3): 752-776.
24
Table 1. Descriptive Statistics
Board of Post Board Full Time
Longitude of Longitude
Period
1714-1828 1829-1939 1714-1939
Observations
Entry Year
Exit Year
Patented
Longitude Award (progress payment)
Prize Winner (other competitions)
Britain
Firm
Maker to Admiralty
271
1802
[23.9]
1841
[33.7]
15.9%
3.0%
6.3%
74.0%
36.5%
35.1%
1,065
1875
[23.9]
1906
[29.1]
18.1%
1,336
1860
[39.6]
1893
[39.7]
17.7%
7.1%
69.7%
35.0%
26.9%
7.0%
70.6%
35.3%
28.6%
Notes: Means and standard deviations in parentheses. Data compiled from information in
Mercer (1991). Statistics on the percentage who patented are constructed by hand
matching observations with a dataset on British patents described in Bunt et al., (2012).
Table 2. The Effect of Longitude Awards on Entry and Patents
A. Entry
Longitude Awardt-1
(1)
(2)
0.106**
[0.048]
0.085*
[0.048]
Longitude Awardt-1 (Mechanical)
(3)
(4)
-0.029
[0.046]
-0.026
[0.046]
Longitude Awardt-1 (Lunar)
Entrants t-1
0.154***
[0.059]
(5)
(6)
(7)
(8)
(9)
-0.029
[0.051]
0.140**
[0.056]
-0.028
[0.048]
0.116**
[0.059]
0.149**
[0.060]
-0.037
[0.043]
0.092
[0.058]
-0.035
[0.029]
-0.001
[0.017]
0.027
[0.027]
0.140**
[0.058]
0.115*
[0.060]
0.151**
[0.059]
114
-227.35
114
-220.84
114
-227.18
114
-220.66
114
-159.31
(5)
(6)
(7)
(8)
(9)
0.074
[0.071]
0.098
[0.078]
0.101
[0.076]
0.057
[0.076]
0.346**
[0.157]
0.095
[0.075]
0.088
[0.085]
0.172
[0.193]
0.038
[0.030]
-0.042
[0.051]
114
-80.62
114
-79.70
114
-72.05
0.163**
[0.064]
GDPt-1
Populationt-1
Observations
Log pseudolikelihood
114
-228.40
114
-221.61
114
-230.88
114
-223.26
B. Patents
Longitude Awardt-1
(1)
(2)
0.112
[0.078]
0.102
[0.080]
Longitude Awardt-1 (Mechanical)
(3)
(4)
0.101
[0.073]
0.001
[0.078]
Longitude Awardt-1 (Lunar)
0.317**
[0.147]
Patents t-1
0.120
[0.080]
0.093
[0.079]
0.282**
[0.135]
114
-80.97
114
-80.34
0.394**
[0.177]
GDPt-1
Populationt-1
Observations
Log pseudolikelihood
114
-80.86
114
-80.01
114
-81.17
114
-79.87
Notes: The dependent variable in the top panel is an annual count of entrants into the chronometer industry, with
entry years determined by Mercer (1991). Columns 1 to 8 include all entrants, while column 9 includes only entrants
located in Britain. Indexes of real GDP and population in Britain are from Broadberry et. al., (2015). The dependent
variable in bottom panel is an annual count of chronometer patents. Award amounts reflect progress payments made
by the Board of Longitude and are compiled from Howes (1998). These were then categorized into mechanical (i.e.,
chronometer-based) and lunar awards and then converted to constant 1800 values using the retail price index of
Officer and Williamson (2014). Award amounts are specified in logarithms with one added to rescale zero values.
Coefficients are from negative binomial specifications with robust standard errors. Significance at the *** 1 percent
** 5 percent and * 10 percent levels.
Table 3. The Determinants of Patenting
(1)
(2)
(3)
Longitude Award (£ amount) 0.047**
[0.020]
Longitude Award (indicator)
(4)
(5)
(6)
(7)
(8)
0.046**
[0.019]
0.488***
[0.166]
Prize Winner
0.484***
[0.166]
0.141***
[0.050]
0.493***
[0.164]
0.118** 0.121**
[0.049] [0.049]
0.021
[0.109]
Great Britain
-0.013
[0.028]
-0.013
[0.028]
0.000
[0.028]
-0.004
[0.028]
0.066
[0.059]
Firm
0.020
[0.025]
0.019
[0.025]
0.014
[0.025]
0.017
[0.025]
0.050
[0.057]
Maker to the Admiralty
Post-Board Era (BL)
0.095*** 0.095*** 0.083*** 0.081*** 0.043
[0.030] [0.030] [0.030] [0.030] [0.062]
0.044
[0.027]
0.046*
[0.027]
0.027
[0.027]
0.051* 0.052**
[0.027] [0.027]
0.035
[0.027]
0.050*
[0.027]
0.084
[0.052]
BL×Prize Winner
0.113
[0.141]
BL×Great Britain
-0.085
[0.074]
BL×Firm
-0.039
[0.058]
BL×Maker to the Admiralty
0.044
[0.072]
Observations
Log pseudolikelihood
1,163
1,163
1,163
1,163
1,163
1,163
1,163
1,163
-564.25 -563.44 -562.99 -557.60 -556.81 -557.64 -553.26 -556.11
Notes: The dependent variable is an indicator coded 1 for patent usage and 0 otherwise. Award amounts are
specified in logarithms with one added to rescale zero values. The full dataset includes 1,336 observations,
but data are missing for some of the variables included. Hence, we use 1,163 observations for which we have
complete data across the specifications. Coefficients are marginal effects from probit specifications with
robust standard errors. Significance at the *** 1 percent ** 5 percent and * 10 percent levels.
Observations
200
300
400
500
Figure 1. The Number of Active Chronometer Makers and Craftsmen
Post Board of
Longitude Era,
1829-1939
0
100
Board of
Longitude Era,
1714-1828
1500
1600
1700
1800
1900
2000
Notes: See the text.
300
Figure 2. The Distribution of Entry and Exit
Post Board of
Longitude Era,
1829-1939
0
Observations
100
200
Board of
Longitude Era,
1714-1828
1500
1600
1700
Entry
Notes: See the text.
1800
1900
Exit
2000
Figure 3. Geographic Location of Observations
!
!
!
!
!
!
!
! !!
!
! !! !!!
!
!
!
! !!
!
!
!
!
!
!
!!!
!
! !
!
!
!!
!
! !
!
!
! ! !!
!
!
!
!
!
!
!
!
! !! ! ! !
!
!
! !
!!
!
!!
!!
!
!
!
! !
!
!
!!! ! !
!! !!!!
!
!
!
!!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!! !
! !!
! !
!
!
!
!
! !!!
!
!
!
!!
!! !!!
! !
!
!!!
!
!
!
! !
!
!
!
!
!!
!!!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!!
!
!
!
!! ! !
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!!
!
!
!
!!
!
!!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!
!!
Notes: Constructed using geocoding of the addresses given in Mercer (1991).
0
10000
British Pounds
20000
30000
40000
Figure 4. Cumulative Board of Longitude Award Payments
1720
1740
1760
Mechanical
1780
1800
1820
Lunar
Notes: Award amounts reflect progress payments made by the Board of Longitude and are
compiled from Howes (1998). These were categorized into mechanical (i.e., chronometer-based)
and lunar awards and then converted to constant 1800 values using the retail price index of Officer
and Williamson (2014).
Figure 5. Mechanical Area Longitude Awards, Entry and Patents
0
0
5
Number of Entrants
10
15
2000
4000
Mechanical Awards
20
25
6000
A. Entry
1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830
0
0
1
Number of Patents
2
2000
4000
Mechanical Awards
3
4
6000
B. Patents
1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830
Notes: Award amounts in solid red circles reflect mechanical (i.e., chronometer-based) progress
payments made by the Board of Longitude compiled from Howes (1998) converted to constant
1800 values using the retail price index of Officer and Williamson (2014). Bars reflect an annual
count of entry in the top figure and an annual count of patents in the bottom figure.
0
20
Percent
40
60
Figure 6. The Timing of Inventor Patents and Awards
-20
-10
0
10
Patent Lag Relative to Award Date
20
Notes: The timing of awards reflects the year a progress payment was made by the Board of Longitude
from Howes (1998). The timing of patents reflects when a patent was issued. Negative values on the x-axis
indicate patenting prior to an award and positive values reflect post-award patenting.
Patents
10 12 14 16 18 20
Figure 7. Chronometer Patents 1717 to 1939
Post Board of Longitude Era, 1829-1939
0
2
4
6
8
Board of Longitude Era, 1714-1828
1700
1750
1800
Moving Average
1850
1900
1950
Raw Data
Notes: Raw plot of chronometer patents per year overlaid using a 10 year moving average of the series.
Figure 8. Time Varying Effects of Changes in Chronometer Patenting
-6
Relative to Base Period = 1710s to 1790s
-5
-4
-3
-2
-1
0
1
A. Relative to All British Patents
1800 1810 1820 1830 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930
Decade
-6
Relative to Base Period = 1710s to 1790s
-4
-2
0
2
4
B. Relative to Scientific Instrument Patents
1800 1810 1820 1830 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930
Decade
Notes: Coefficients derived from a negative binomial difference-in-differences specification where the indicator for
chronometer patents was interacted with a decade indicator to obtain time varying estimates. The Board of
Longitude was dissolved by an Act of Parliament in the 1820s (1828).