curves from 5 years on were plotted on semilogarithmic graph paper, they would approximate straight lines, indicating that the decrease in the proportion of patents kept in force from year to year follows the simple exponential curve of decay.28

The next table presents estimates for two countries, Sweden and France. For Sweden, the number of fees paid during the 10 years 1930-39 was available,20 but not the base figures of the number of patents issued arranged by filing date, and these were estimated. The figures for France were calculated from a sampling taken in 1937. At that time the term of French patents was 15 years and the remaining 5 years in the table are merely extrapolations. The Swedish figures do not show the same regularity as those of the countries previously considered. The regularity of the French figures is not significant since this was assumed in making the estimates.

In the five countries which have been mentioned the term of the patent and the time for payment of the annual fees are counted from the filing date of the application. In a sixth country, the Netherlands, the timing of the payments is different and its figures cannot be directly compared with those given for these countries. In the Netherlands the term of the patent is 18 years from the date the patent is granted and the annual fees similarly date from the grant. There is a fee for the first year, which is due before the end of the month following the issue of the patent; if this first fee is not paid the patent, though granted, never does become effective and thus has no duration whatever. The following table gives the proportion of patents kept in force for the indicated years. Two sets of percent

30

See P. J. Federico, Taxation and Survival of Patents, 19 Jour. of the Pat. Off. Soc'y. 671-691 (1937), for the development of equations for these curves.

These were supplied by an official of the Swedish Patent Office.

The data is given in the published annual report for 1939, Verslagen en Mededeelingen van de Directie van Handel en Nijverheid van het Departement van Handel, Nijverheid en Scheepvaart, Jaargang 1940, Nummer 2; Verslag der Werkzaamheden van den Octrooiraad over het Jaar 1939.

ages are given, one (A) takes as the base the number of patents issued whether or not the initial fee was paid and the other (B) takes as the base the number of patents for which the first fee was paid.

Column A is based on the number of patents issued as 100 percent. Column B is based on the number paying the first fee as 100 percent. As seen from the data presented, the actual duration of most patents is less than the normal term for which they are issued. The data in the tables permit the approximation of the average duration of patents for the period preceding the war. These times are, for the first five countries: Great Britain, 6.8 years; Germany, 7.1 years; Switzerland, 5.3 years; Sweden, 6.9 years; France, 5.0 years. All of these periods are from the filing date of the application. The differences in the lengths of the normal terms do not affect these averages very much in view of the small proportion of patents still in force during the last few years of their normal terms.

If there were no renewal fees, the number of patents in force at a particular moment in one of these countries would be the sum of the patents issued on applications filed during the preceding 16, 17, 18, or 20 years, according to the length of the normal term, but with the renewal fees in existence, the number actually in force is much less. The ratio of the patents actually in force to the number of patents issued and whose normal terms had not expired cannot be calculated for the five countries from the data presented here since the period of pendency of the applications is not specified and is variable. This ratio would be from about one-fourth to about one-third and would be affected by the differences in the normal length of the terms.

For the Netherlands, where the term starts from the granting of the patent, the average duration of the patents granted calculated from the table (col. A) is 4.4 years. The ratio of the patents in force to those granted can also be calculated from the table and is 24

percent; that is, only 24 percent of the patents issued during the preceding 18 years are actually in force.

WAR AND POSTWAR PERIOD

The preceding section dealt with the period 1930-39. The decade 1940-49 cannot be treated for a number of reasons. The same statistics. are not available for all of the same countries during this period and what figures are available show many irregularities and inconsistencies. During the war years and after, many special laws were in existence which waived or deferred fees, reinstated lapsed patents, etc., and the statistics could not be considered normal or representative. Consequently only some partial figures beginning with 1948 are given, and these are on a different basis than those given for 1930–39.

The column for Germany represents the summation, reduced to percentages, of the fees paid up to the end of 1957 on patents granted on applications filed in 1948 and subsequent years.31

The column for Switzerland represents the summation, reduced to percentages, of the fees paid up to the end of 1956 on patents granted on applications filed in 1947 and the following years.32 Fees paid on patents resulting from applications dated before this period are not included.

The column for the Netherlands is based on the patents issued in 1948 to 1957 and the fees paid on these patents up to the end of 1957.33 In each instance the proportion of patents kept in force for a given year is much higher than the proportion for the same year shown in the prewar period. The level of payment of the fees during the 1930-39 period is higher than during a much earlier period 34 and the most recent period is higher still.35 The figures in the preceding table should be taken as indicative only as they are not homogeneous; the last line is based on 1 year only, the next to the last line on 2 years,

The number of fees paid during this period is given in the Blatt für Patent-, Muster-, und Zeichenwesen for March 1958.

The basic figures are in the annual index for 1957, see note 20, supra.

These figures are derived from unpublished annual reports, see note 19, supra.

34 Bee P. J. Federico, op. cit., supra, note 28, supra.

The British reports ceased publishing the base figures of patents arranged by filing date and there is no sound basis for making estimates; consequently, a table for the recent period similar to the others cannot be presented. However, the fees paid show a considerable increase over the 1930-39 period.

etc., and individual years are still somewhat too irregular for an average to be very meaningful.

EFFECT OF INCREASE IN FEES

The effect of an increase in fees on the proportion of patents on which fees are paid may be observed in connection with some of the countries in which the fees were increased in recent years.

As has been stated, the annual fees for the 7th to the 16th year were raised in England in 1955, coming into effect about the middle of the year. In 1954, under the previous scale of fees, the number of patents in force for which the 7th to the 16th year fees were due was 54,493 (these being the patents on which the 6th to the 15th year fees were paid in 1953); of this number the fees were paid on 48,088, or 88.2 percent. In 1955, when the new scale of fees was in force for part of the year, the 7th to the 16th year fees became due on 56,735 patents (the number for which the 6th to the 15th year fees were paid the preceding year) and were paid on 49,928, or 88.0 percent. In 1956, when the new scale applied to all patents, this proportion was 90.0 percent. In 1957 the proportion was 87.1 percent. These variations in the proportion paying the 7th to the 16th year fees do not appear very significant.

In Switzerland, the fees for the 8th to the 15th year were increased beginning January 1, 1956, the total amount of the increase being 36 percent. The proportion of the patents in force on which these fees were due in 1955 and on which these fees were paid was 86.0 percent. In 1956 this proportion, paying the higher scale of fees, was 86.1 percent.

In the Netherlands all the annual fees were increased about the middle of 1957, the total amount of the increase being 75 percent. In 1956 the fees were paid on 91.2 percent of the patents in force on which fees became due. In 1957, with part of them paying the higher rate, the proportion paying was 91.0 percent. Both years, however, are higher than preceding years.

For Germany, the patents granted on applications filed in 1948 and later have at most reached only the 10th year. Considering these early years, the proportion paying the 4th to the 10th year fees, of the patents on which these fees were due, was 87.0 percent in 1954, 88.0 percent in 1955, 86.6 percent in 1956, and 86.9 percent in 1957. The increase in fees occurred in 1955.

In general, the data available does not indicate that the increases in the amounts of the annual fees have affected the proportion of patents on which the fees are paid. The variations in the proportion paying the fees following an increase are much smaller than random variations occurring from year to year. It may be concluded that, within limits, the amount of the annual fees has little or no effect on the keeping of patents in force.

5. PROPOSALS FOR RENEWAL FEES IN THE UNITED STATES

The patent laws of the United States have never required the payment of periodical fees during the life of a patent, although the laws have contained two analogous features. As has been stated, design patents are granted for terms of 31⁄2, 7, or 14 years, with a different