Table 19.-United States-Mexican Travel Account, 1938-39 1

1 Preliminary for both years. Based upon estimated volume of travel and average international expenditures.

2 Rounding of data prevents exact addition.

3 It is estimated that goods purchased in Mexico and brought back to the United States, largely by motorists visiting border towns, comprised between at least $25,000,000 and $35,000,000 of this total (a larger share of total expenditures than travelers to Canada allot to purchases).

It will be noted that the slight increase in expenditures from 1938 to 1939 was confined to outlays by residents of the United States, travel expenditures by Mexicans remaining at the same sum as in the previous year. The predominance in total expenditures of those by persons who travel merely to border towns is particularly noteworthy. Although spending at low rates which range down to nothing, the large number of such persons, as shown in table 20, accounts for their important influence upon the United States-Mexican travel account. The question of whether expenditures by border visitors have a different significance from expenditures for travel to interior points is discussed at the end of this section. Rounded estimates of the number of travelers in the various broad categories upon which the total expenditures shown in table 19 were based appear in table 20.

Table 20.-Number of Trips to Mexico by Residents of the United States and to the United States by Residents of Mexico, 1938–39 1

1 Preliminary for both years. Based chiefly upon break-down of statistics of the United States Customs Bureau through utilizing observations secured from border officials.

2 Roughly estimated by inflating data from Mexican sources to allow for travel at ports probably not covered by the figures originally reported.

Rounding of data prevents exact addition;

GENERAL METHODOLOGY

The method described below for estimating travel expenditures between the United States and Mexico has been applied only to the years 1938 and 1939. Therefore, estimates previously made for other years now lack comparability with those for 1938 and 1939.

Earlier rough data will not be revised, however, until more satisfactory means can be found of using present methods to obtain estimates of total expenditure data for years prior to 1938.

The problems involved in reckoning the United States-Mexican travel account are generally those encountered in making all estimates of travel expenditures. The specific approach is similar to, although differing in detail from, that for the United States-Canadian travel account, which likewise involves the treatment of travel across a land border. The general method thus consists of applying average expenditures by various types of travelers to the number of travelers in the groups to which the expenditure averages pertain. In following this method in the present instance, it was obviously necessary to have data showing the volume of travel to Mexico by residents of the United States and of travel to the United States by residents of Mexico. Average expenditures for this travel by residents of each country were likewise needed. In developing a method of estimate, knowledge of the nature of the volume-of-travel data to which expenditure averages were to apply was prerequisite to the construction of the expenditure averages themselves. On the other hand, the groupings of volume-of-travel data needed to be so set up as to secure maximum homogeneity of individual outlays within each group.

Other factors determining the grouping of the expenditure averages were the form and classification of the basic volume-of-travel data available and the ease or difficulty of obtaining estimates of average expenditure to apply to alternative groupings of these basic data.

The general considerations just mentioned indicate problems which were involved in the utilization of data on volume of travel and in the construction of estimates of average expenditures. Below are stated (1) the process by which volume-of-travel data, classified by various groupings and subgroupings, were obtained to meet present requirements; (2) general details concerning the derivation of expenditure averages; and (3) a detailed description of the derivation of total expenditures according to type of transportation employed by the travelers in each grouping. The manner of deriving a break-down of total expenditure estimates into certain classifications within the individual type-of-transportation groupings is also described under the last heading.

Data on Volume of Travel.

The Bureau of Customs of the United States Treasury Department, as well as the Immigration and Naturalization Service of the United States Department of Labor, compiles data regarding the total number of travelers crossing the southern land border of the United States. The customs officers of the Mexican Government arrive at a count of the number of "tourist" cars entering and leaving Mexico, and the Mexican immigration_officers derive a count of "tourists." Of these data, those of the United States Bureau of Customs seemed most generally adaptable to the special requirements of computing travel expenditures. The Mexican data cited comprise a count of "tourist" cars at the chief port of entry only, and of "tourists" at various and perhaps all ports. The latter statistics exclude most travel which does not extend beyond the Mexican border ports, and for that reason are useful for estimating only one element (and as yet a relatively unimportant element) in the United States-Mexican travel account.

The series of the Immigration and Naturalization Service includes both "tourists" and other types of travelers. Compiled for the special purposes of immigration work, it segregates the travelers into citizens and aliens, but does not subdivide these categories into classifications which would be useful in reckoning travel expenditures. Moreover, the fact that this series is compiled on a fiscalyear basis rather than on a calendar-year basis, as far as travel at the land border is concerned, makes it less adaptable to the purpose in question.

The data assembled by the Bureau of Customs are fortunately compiled according to type of transportation used by the traveler; but the lack of subdivisions into categories of United States residents and Mexican residents, and into length-of-stay groups, constituted a serious obstacle to the utilization of this series. At the request of the Finance Division, however, United States customs officers at the chief ports of entry submitted estimates of the percentages of the trips by each important type of transportation made by residents of the United States, on the one hand, and by residents of Mexico, on the other. The officers also furnished percentages for application to the number of travelers in various of these subclassifications which permitted a division of the latter into lowerexpenditure and higher-expenditure groupings.

The crude classification thus made into roughly homogenous spending groups, however, did not provide for all types of transportation the direct means for a separation of travel destined to the interior of Mexico or of the United States from travel destined merely to border towns. Below, in connection with the discussion of specific types of travel, are described the methods used for estimating, separately from border travelers, the number of travelers to the interior, whose rate of expenditure is far higher than that of travelers who visit border towns only.

Average Expenditures.

On behalf of the Finance Division, some 12,000 questionnaire cards were distributed during 1939 by the Immigration and Naturalization Service of the United States Department of Labor among travelers returning to the United States from a trip to the interior of Mexico. The returns provided inductive data for estimating average expenditures on trips to the interior of Mexico by residents of the United States. Automobile travelers, chiefly, received these cards upon return from Mexico. Some were filled out by travelers returning from the interior by other modes of transportation, and some by automobile travelers and others returning from trips to border towns. Estimates of average expenditures were based, whenever possible, upon actual averages obtained. In instances where

this was not feasible because of lack of data, other indicators, such as judgments by consular officers, were employed. Below, under the headings of the various types of transportation employed, are described the further details of derivation and application of the averages used.

Improved knowledge about the volume of travel of various types, which was obtained through customs officers early in 1940, as described above, should make it possible to obtain more satisfactory questionnaire averages for some of the categories involved. Average expenditures as estimated for 1939 were used in calculating total expenditures for 1938 as well as for 1939, since equally satisfactory questionnaire data were not available for the construction of separate averages for the former year.

Expenditures by Automobile Travelers.

Data on volume of travel were obtained for automobile travelers in the various classifications by the general methods set forth above. Some special adaptations of the methods there described were employed for greater accuracy. The classifications referred to were obtained for automobiles according to whether registered in the United States or in Mexico, rather than for travelers in automobiles, and average expenditures per car were applied directly to the number of automobiles in the various categories in deriving total expenditures by automobile travelers. The Finance Division's estimates of the number of automobile travelers were subsequently obtained by multiplying the number of automobiles in various categories by the respective average number of persons per car. The latter averages were based on questionnaire data rather than on a count by border officials. The customs officers' estimates of the percentages of United States-registered (and likewise of Mexican-registered) cars crossing the border which were registered in border towns and in towns away from the border, respectively, made a rough stratification possible in estimating the expenditures of automobile travelers. An average of $7.50 for the group of United States cars registered in border towns and proceeding only to an opposite border town was obtained without regard to the locality where they crossed the border. In the case of returns applying to such United States cars registered in other than border towns, a sufficient number was received to make possible an additional stratification according to some of the particular border localities visited. To the data on travel at other localities for which returns were insufficient, a uniform rate derived from all the questionnaires received was applied.

To apply these rates of expenditure per automobile to the total number of automobiles visiting Mexico from the United States without making adjustment for such of these cars as penetrated to the interior of Mexico would result, of course, in an understatement of expenditures in Mexico by motorists from the United States. Means were therefore found of estimating the total number of cars which went to the interior and of applying to this number average expenditures representing typical outlays in Mexico by their occupants. The number of these cars was then subtracted from the total number of crossings of United States cars to Mexico, to leave, as remainder, the number of crossings by United States cars to Mexican border towns. The corresponding subtractive adjustment in expenditures by United States motorists was made at that rate per car which represented a weighted average of expenditures by the occupants of United States cars registered in other than border towns. This procedure is supported by evidence that but a fraction of motorists to the interior of Mexico reside in border towns of the United States.

Estimates of the number of cars registered in the United States which penetrated to the interior of Mexico were based upon the count made by Mexican Customs officers of "tourist" automobiles entering Mexico at Laredo on permits for stays of 4 days or longer. Evidence from questionnaires seems to corroborate unofficial impressions that whether or not a permit is issued for an automobile depends, in practice, upon whether the car proceeds more than 10 miles into Mexico.

Since the Mexican count of "tourist" automobiles from the United States is described as applying only to those cars entering Mexico through Laredo, estimates of the number of similar cars entering at other ports were requisite in order that computed expenditures by United States motorists traveling to the interior of Mexico should not omit a considerable amount of the sums so spent. Therefore, tentative estimates of the volume of such travel at other ports had to be made, somewhat arbitrarily, to include not only cars entering on permits at ports where permits are similarly issued for cars of this nature, but also cars of this type entering at ports where permits appear not to be issued as they are at Laredo.

Of cars traveling to the interior of Mexico, however, those included in the figure for Laredo account for two-thirds of estimated expenditures by United States motorists traveling to interior points in Mexico.

The method of ascertaining the total number of United States-registered "tourist" cars to the interior of Mexico was as follows: The first step consisted of making an estimate of the maximum number of cars of this nature crossing at the various ports other than Laredo, as well as at Laredo. Such an estimate was made by comparing the number of questionnaires received from this type of traveler at the various ports with the number received from automobile travelers who merely visited border towns after crossing at those ports. Then, since the actual number of United States-registered permit-type cars crossing at Laredo was known, the ratios for Laredo only of the actual number of cars reported to the estimated maximum were computed separately for 1938-39. Finally, these ratios were applied to the estimated maximum number of permit-type cars of United States registry crossing during the respective years at (a) Hidalgo and (b) all other ports.

These roundabout methods of obtaining desired data are made necessary by the absence of statistics adapted to balance-of-payments requirements. Increas ing interest in this field may eventually give rise to the collection of travel data more suitable to the special purposes involved.

Average expenditures by travelers to the interior of Mexico in United Statesregistered cars were obtained by the Finance Division from questionnaires distributed by immigration officers to automobile travelers as they returned to the United States at various ports. The questionnaires were compiled separately for each border port from which they were received. Questionnaires from motorists of this type returning via Laredo and Hidalgo (whose expenditures comprise over 75 percent of estimated expenditures by United States motorists to the interior by the way of all ports) amounted to over 90 percent of such returns. Laredo averages were applied to Laredo travel, Hidalgo averages to Hidalgo travel. Averages obtained for other ports were applied to the estimated numbers of permit-type cars returning at each port for which returns were received; and for ports from which no permit-type returns were received, an average based on questionnaires covering travel at ports other than Laredo was used.

Methods similar to those used for estimating the volume of all automobile travel across the United States border by residents of the United States and for subdividing it into various classifications were employed in obtaining comparable data on travel to this side of the border in Mexican-registered cars. The treatment of United States customs data afforded a break-down of total crossings by Mexican cars in the manner previously described. The number of Mexican cars which traveled to the interior of the United States was also estimated and was included as a segment of the total of Mexican cars in the same way that the number of United States cars estimated as traveling to the interior of Mexico was included as a segment of all United States cars crossing the border.

The number of Mexican-registered cars crossing to the interior of the United States was derived, however, by applying to the number of United States permittype cars counted or estimated as crossing to Mexico at the various ports a ratio taken to represent the relationship of the number of Mexican cars in this class to the number of United States cars in this class. This ratio was derived from figures reported as pertaining to an earlier period and is subject to adjustment if actual data for more recent years can be obtained on this point. The results obtained by this procedure seemed preferable to a purely arbitrary estimate. Expenditure averages were applied to the estimated number of Mexican cars which traveled merely to border points by steps for the most part identical with those by which average expenditures had been applied to the corresponding category of United States-registered cars. Average international expenditures by Mexican motorists crossing only to border towns of the United States were computed on a different basis from average international expenditures by United States motorists crossing to Mexican border towns. It was decided, after surveying various estimates of expenditures by travelers of the respective nationalities, attributed chiefly to United States consular officers, to place average expenditures by Mexican residents visiting United States border towns at half the average expenditures by United States residents of the same type visiting Mexican border towns.

Average expenditures per car for Mexican motorists proceeding to the interior of the United States, however, were taken as equal to expenditures per car for United States motorists visiting the interior of Mexico. Because of the lack of definite data on the subject, and because it seemed as reasonable as any alterna