does contain gold coin import and export figures for 1873-80 and there is a disparity between Bordo's estimated figures and those of the Abstract. There have been a number of criticisms of international gold movement statistics that have produced serious doubts concerning their reliability, but Goodhart has reviewed some of the data sources and is optimistic about the usefulness of official gold trade statistics (see Appendix 1). However, Bordo's interpolation between the benchmarks, using the total net change of mintings and exports, resulted in figures lying on a steeper trend and higher than the next benchmark figure. This was partly because this method did not allow for the of coin or the unrecorded carriage of coin by migrants. Bordo did not discuss this specifically but reduced his estimates by multiplying the figures interpolated between the benchmarks, using the relationship that the benchmark bore to the estimated figure. The results do not allow for changes in the Bank of England's holdings of gold coin and in fact relate to the total amount of gold in the United Kingdom. For the years after 1880 Bordo used Sheppard in unadjusted form. 2. CALCULATION OF A GOLD COIN SERIES, 1868-1914 Deciding that previous estimates could be improved on, we have calculated series for gold coin, silver coin, bronze coin, total coin, and coin in circulation, using only those benchmark estimates that might be regarded as reliable. We have also examined our interpolated figures in relation to those contemporary benchmark figures which we can determine to have been overestimates or underestimates. Jevons provided a for 1868 which is the only single-year estimate capable of being reworked to form a suitable starting point for our series of total gold coin [12, pp. 438-47]. The Bank of England and the Royal Mint provide figures from 1905 for coin in circulation and for the total volume of silver and bronze coin respectively; we have accepted these for the years 1905-14, and they form a convenient end to the series in the period 1868-1914 (see Table 2). We have attempted to correct Jevons's and to provide a gold coin series by interpolating between his estimate and the official figures from 1905 onwards, using data for the net issue of coin from the mint and net recorded exports of gold coin, together with our estimates of net unrecorded coin exports and an allowance for the of coin. The starting point is our evaluation of Jevons's figure, taking into account deficiencies that he acknowledged and criticism he received from contemporaries. This allows his results to be reworked to form a starting point for our series. The starting point for this, however, is to understand how Jevons produced his estimate. Jevons This content downloaded from 207.46.13.76 on Thu, 25 Aug 2016 06:26:20 UTC All use subject to http://about.jstor.org/terms 28 : MONEY, CREDIT, AND BANKING sent questionnaires to a large number of banks and obtained an analysis by date of minting of 90,074 sovereigns and 75,036 half-sovereigns and the amount of gold coin held in the Bank of England when the sample was analyzed in March 1868. He also had figures for sovereigns registered for export in 1865, 1866, and 1867. The next step was the calculation, that of 14.6m sovereigns minted in 1863-64, 0.6m were in the Bank of England, and 1.75m had been exported by 1868. This export was determined on the assumption that the proportion of exported sovereigns dated 1863-64 to all sovereigns exported in 1865-67 was the same as the proportion of sovereigns dated 1863-64 to those in his sample (i.e., in circulation). These adjustments left 12.25m sovereigns dated 1863-64 in circulation in the United Kingdom in 1868. Because coins dated 1863-64 formed 18.671 percent of his sample, he obtained a total circulation for sovereigns in March 1868 of £65.6m. He finally decided on £64.5m so that when £3.5m of sovereigns held in the Bank of England were added to £12m of half-sovereigns similarly determined, a round figure of £80m (instead of £81.1m) was obtained for the total gold coinage of the United Kingdom. However, Jevons had to admit that this method did not allow for the extensive melting of new coin by bullion dealers and capitalists, and consequently this was an upper limit for gold coin [12, pp. 443-44]. Further, Jevons did not allow for, or acknowledge, unrecorded movements of coin by migrants (Table 1). John B. Martin of Martin's Bank in London criticized Jevons's estimate because the sample contained 48,000 sovereigns held by the Bank of Scotland, which he pointed out were chiefly supplied by the Bank of England as a reserve for note issue [14, pp. 219-20]. Because the quality or newness of these coins was higher than the