A LONGITUDINAL ANALYSIS OF THE SETTLEMENT PATTERNS, OCCUPATIONAL MOBILITY AND WEALTH ACCUMULATION OF EUROPEAN IMMIGRANTS TO THE U.S., 1840-1860

Joseph P. Ferrie Northwestern University

I. Introduction

The period of adjustment undergone by immigrants is a crucial part of the process of immigration. Though this process is interesting in its own right, as an indication of an economy's ability and willingness to absorb and embrace new members from vastly different backgrounds, it is important as well as a determinant of subsequent immigration: migrants whose expectations have not been met may discourage others from migrating. At the same time, the adjustment of immigrants to a new economy may postpone the benefit to the economy of the receipt of millions of working-age, often highly-skilled individuals.

For the U.S., the question of the extent of adjustment experienced by immigrants has received a great deal of attention. But few answers have been available for the early nineteenth century when the first great waves of immigration occurred. Though millions of European immigrants arrived in the years after the Napoleonic Wars and before the American Civil War, no data suitable for a large-scale analysis of the process of immigrant adjustment have been available for this period.

The decennial federal censuses of population -- the best source of economic and demographic information on America's population -- suffer from three shortcomings in addressing this issue: 1) the census did not ask immigrant respondents to state their year of arrival until 1890; 2) the census provides no sense of immigrants' circumstances in Europe or at arrival (e.g. their occupation); and 3) each census is a snap-shot of the population at a single point in time, while many questions of adjustment can best be addressed with data which are longitudinal.

To overcome these difficulties with the census and yet still have a body of data broad enough to allow generalizations to be made about the experiences of a large number of immigrants, I have produced a sample of more than two thousand male immigrants who were followed from their arrival in the U.S. (as documented in passenger ship arrival records) until they turned up in the manuscript schedules of the 1850 or 1860 federal census of population; more than five hundred immigrants in the sample appeared in all three sources, providing a record of their experiences which extends from their arrival for up to twenty years.

In this essay, I will use these data to explore three aspects of the process of immigrant adjustment: patterns of settlement, occupational mobility, and wealth accumulation. The questions I will address are: 1) how extensive were the changes in location and occupation made by immigrants, and how likely were immigrants to make such changes as their time in the U.S. increased?; and 2) what was the impact of those changes, as measured by their effect on wealth accumulation?

Few of the migrants studied here simply "uprooted" in the sense of the trauma, confusion, and paralysis described by Handlin -- a clear process of adjustment is evident beginning from the time of arrival. Many appear to have been "transplanted" instead --arriving with numerous handicaps (a lack of skills, capital, or knowledge of English), but acting after their arrival in ways consistent with attempting to reduce the impact of those deficiencies. The data suggest that immigrants continued to search for areas in which their greatest promise lay from arrival right through at least the end of the twenty years covered by the sample.

II. Patterns of Settlement

Where immigrants were located by the time of the 1850 and 1860 censuses is shown in Table 1. [Notes, references, tables, and figures will be provided at the session] Seventy-two percent of the immigrants in the sample in 1850 were located in the Northeast, and more than half were found in just three states -- New York, Pennsylvania, and Ohio. Smaller numbers were found in New England and the states closest to the frontier, while very few were found in the South.

Immigrants were clearly over-represented in cities: though more than 87% of the total U.S. population in 1850 lived in places of fewer than ten thousand people, only just over half of the immigrants in the sample were found in such places in 1850, and only 44% were found there in 1860. Part of these disparities might result from the fact that many cities were transportation hubs through which immigrants had to pass before they could reach the interior where most natives were located. Immigrants were indeed over-represented in New York City, Philadelphia, Cincinnati, and Buffalo, all of which served as gateways to the interior. But this does not account for the concentration of immigrants in smaller cities, places with fewer than 50,000 people.

Even the earliest arrivals in the sample were more concentrated in cities than were natives, which suggests that immigrants were attracted to urban centers for reasons other than the fact that they had to pass through them to get anywhere else. The attractions of cities included the high level of average wages, ready sources of employment for the unskilled and, in many cases, the presence of large numbers of one's countrymen -- an important source of both potential employers and potential customers.

Immigrants who arrived at the end of the 1840s were no less likely than those who arrived at the start of the 1840s to be found in 1850 in locations far removed from New York, their point of entry. At their arrival in New York, immigrants faced a variety of relatively inexpensive routes to the interior of the U.S., which many had arranged to employ even before their departure from Europe. The total fare from Europe to Buffalo, Cleveland, or Detroit was probably only one fourth greater than the $35 fare to New York city alone, with the total fare to Chicago and interior points such as Cincinnati, Pittsburgh, and Louisville only a third more than the fare to New York alone.

Though early and recent arrivals were found in many of the same locations in 1850, it was the recent arrivals who were more likely to relocate between 1850 and 1860. Table 2 shows that thirty percent of all immigrants in the sample linked from 1850 to 1860 changed region between 1850 and 1860, with net movement being from east to west: 15% moved to more western locations, while 9% moved from western locations back into the Northeast. Even among the earliest arrivals, movement between 1850 and 1860 was not negligible. Of the one hundred immigrants in the sample who arrived in 1840 and 1841, nineteen changed regions between 1850 and 1860 -- the process of adjustment clearly continuing into the second decade of these immigrants' lives in the U.S. But the magnitude of this change is dwarfed by change among the most recent arrivals: of the 105 arrivals in 1849 and 1850, 50 had changed locations between 1850 and 1860.

It thus appears that immigrants moved quickly from their point of entry to their 1850 location, with the vast majority (seventy percent) remaining in the same region over the period 1850 to 1860. But nearly thirty percent seem also to have found an 1850 location rapidly, but only over the course of several years found a location in which they were likely to remain. Changes in location between 1850 and 1860 could have resulted from several circumstances: receipt of information regarding superior locations, spending time at one location to accumulate the human or financial capital which would permit a move to a better location, or simply a change in fortune which permits or requires such a change.

The other important change in location which occurred in the sample between 1850 and 1860 was net migration from rural to urban locations. Nearly twice as many migrants went from rural to urban places as from urban to rural places. Those most likely to make either sort of move were the most recent arrivals. But the average age of those going from rural to urban places was more than two years less than that of those who remained in rural places, while urban-to-rural movers were no different in age from those who remained in urban areas.

Since younger workers had probably made less of an investment in their occupation by 1850, while they had a longer time over which to reap the net benefits of changing occupations, it is possible that those rural-to-urban moves represented changes in occupation as well, while urban-to-rural moves were immigrants merely passing through on their way to a rural occupation in which they expected to remain. In fact, of the 63 immigrants who made rural-to-urban moves, 40 (64%) changed occupation as well as location, while only 40% of the rest of the entire sample changed occupation. Among those making urban-to- rural moves, the same percentage changed occupation (43%) as among the rest of the sample. A more detailed consideration of occupational mobility follows.

III. Occupational Mobility

The distribution of 1850 and 1860 occupations in the sample shown in Table 3 is more concentrated in white collar and skilled blue collar occupations than the total U.S. population: 44 percent of the sample was in these occupations by 1850, while only 30 percent of the total U.S. population was so employed. Immigrants were correspondingly under-represented as farmers and laborers (56 percent versus 63 percent for the total U.S.). This difference between the sample distribution and the U.S. distribution widened between 1850 and 1860.

Most immigrants changed occupation at least once in the years after their arrival. Table 4 reports these findings. More than half changed occupation between arrival and either 1850 or 1860, with just over 40% changing between 1850 and 1860. For example, 9% of those in the sample described themselves as farmers at arrival and as laborers in the 1850 census. Of those moving from farmer to laborer by 1850, 63% had moved back up out of the laboring class by 1860. Overall, 16% of the immigrants in the sample arrived as farmers but had changed occupation by 1850. Of these, only 9% had returned to farming by 1860, while 67% had entered white collar, craft, or mining occupations.

Although there appeared to be considerable downward occupational mobility among these farmers at arrival who were found in occupations associated with lower wealth by 1850 or 1860, at least some of it may have been the result of a planned change of occupation, rather than an inability to re-attain a previous occupation. Farmers who were temporarily unable to obtain their own farms to work would have been most likely to show up in the census as farm laborers, rather than undertaking the investment necessary to convert to a white collar or craft occupation. At the same time, those who left farming for another occupation, both between arrival and 1850 and between 1850 and 1860, were more than five years younger on average than those who remained in farming.

That the decision to leave farming was made by just those immigrants who would have been most likely to make a decision to do so permanently (younger farmers), while those changes were not the sort we would expect to see made by those hoping to remain in farming suggests that leaving farming was for at least some the result of a decision to forsake permanently farming for another occupation rather than a temporary expedient before an eventual return to farming. The next section considers the impact of such changes on wealth.

IV. Wealth Accumulation

In his landmark study of 19th century wealth-holding patterns, Soltow (1975) found a concave relationship between age and wealth. A large literature in economics describes a variety of mechanisms which could account for such a relationship. But Soltow also discovered marked differences in wealth between the native and foreign born, as well as differences by country of origin among the foreign born. He suggested that these differences might have resulted from differences in immigrants' time of arrival, and in their ability in the years since arrival to overcomes barriers posed by language, prejudice, and low levels of skill at arrival. But, lacking any information on immigrants at their arrival, he was unable to assess the importance of these factors.

Tables 5 and 6 show the results of regressions on the natural logarithm of real wealth in 1850 and 1860 for all immigrants located in either 1850 or 1860. The explanatory variables include age and its square, fixed factors such as nationality (over which the immigrant had no control), variable factors such as location and occupation (which immigrants could alter), combinations of fixed and variable factors such as the impact of changes from occupation at arrival (which the immigrant could not control in 1850 or 1860) to occupation in 1850 (which he could control), and the independent effect of time since arrival.

The results display the expected concave age-wealth relationship: wealth rises with age at a diminishing rate in both years. But the age-wealth profile is both steeper and later- peaking in 1860. At age 30, for example, wealth is growing at 9.3% in 1860 but at only 7.2% in 1850; the 1860 profile achieves a maximum at age 59.2, while the 1850 profile peaks at 50.9 years. The Irish were 30% less wealthy than the British and Germans in 1850, but this difference was halved and reduced to statistical insignificance by 1860.

The results show much higher wealth for those in the Northwest states in both years and in the North Central states in 1860, and for those outside urban places. Those located in the Northwest, for example, were nearly two and a half times wealthier in 1860 than identical individuals located in the Northeast. Such differences probably resulted from the superior rate of return on investment in land in the west, as well as from a return to the riskiness of locating there. Again, though, there are pronounced differences between 1850 and 1860. The advantages of those in the Northwest and North Central states, as well as the disadvantage of those in urban places, are all greater in 1860 then they had been in 1850.

The results also display a clear hierarchy of wealth by occupation, much like that observed in other studies of nineteenth century wealth. Farmers are clearly the wealthiest: they own on average 20 times the wealth of laborers in 1850. Those in white collar occupations are one and a third times wealthier and craftsmen and miners are 91% wealthier than laborers in 1850. But again, the advantages of particular occupations are considerably greater in 1860 than they had been in 1860. Farmers were nearly 60 times wealthier than laborers by 1860, while the gaps between white collar workers and laborers and between craftsmen and laborers widened correspondingly.

The effect of occupational change since arrival shows the role of previous occupation (and presumably previous ability to accumulate wealth) on the level of wealth in 1850 and 1860. Craftsmen and miners in 1850 who were described as farmers in the passenger ship lists were nearly 50% wealthier than craftsmen and miners who were craftsmen and miners in the ship lists. Since farmers possessed the greatest wealth (both real and total) of any occupation in the 1850 and 1860 census, this suggests that immigrants who arrived as farmers may have had more wealth at arrival than those in other occupations, and that some of this advantage persisted even after a change in occupation. Laborers who had risen into the ranks of craftsmen and miners were 53% less wealthy than craftsmen and miners who had not changed occupation between the passenger ship lists and the 1850 census. This might be the result of the persistence -- even after occupational change -- of the disadvantage in wealth of those who arrived as laborers. By 1860, though, these effects were greatly reduced in statistical significance.

Finally, the return to duration after these other factors have been taken into account is both substantively and statistically significant in both years, though it smaller in 1860: wealth rose by 8.2% for every year since arrival in 1850, but by only 6.0% per year in 1860. It is possible that the return to duration in the U.S. is an illusion. If there was a secular decline over the 1840s in the wealth of all immigrants at their arrival -- a cohort effect -- then we might in considering a single cross-section of individuals mistakenly attribute the greater wealth of earlier arrivals to their greater time in the U.S., when in fact they were merely wealthier on average when they got here than later arrivals.

However, the fragmentary evidence which can be mustered on this point suggests that, if there was any cohort effect, it actually works to make the observed duration effect smaller than the true effect. The average wealth of German emigrants at their departure from Germany, as recorded by officials of their principalities, shows a slight rise over the 1840s, so the actual rate of wealth accumulation with duration in the U.S. for Germans at least is probably greater than the rate implied by Tables 5 and 6.

How can we account for the magnitude of this duration effect? One possibility which we can immediately discount is that the duration effect represents mainly the acquisition of language skills by non-English speaking immigrants. When regressions are estimated separately by country of origin, the Germans do indeed have a larger duration effect than the Irish, but the British have the greatest effect of all. The return to duration in the U.S. might also represent the acquisition of skills and knowledge other than language specific to the U.S. economy.

Another possibility is that the effect of duration in the U.S. is capturing general growth in the value of real estate or general economic growth over the 1840s and 1850s. Though there are no figures on the rate of growth of the value of real estate in the U.S. between 1840 and 1850, the corresponding figures for the next decade suggest that per capita real wealth grew at perhaps 5% per year between 1850 and 1860. Though the return to duration was higher than this average in some places -- as high as 8.3% per year between 1850 and 1860 in Chicago -- nearly two thirds of the immigrants in the sample were located in the Northeast where such high rates of return to duration in specific local economies were probably rare.

Much of the return to duration may simply represent the higher average level of per capita income in the U.S. relative to Europe. The fact that an immigrant who has been in the U.S. for only a year is 8% wealthier than an otherwise identical immigrant who has just arrived might result from the fact that the immigrant in the U.S. for a year has had the opportunity to earn for that year an income well above that earned in Europe by the more recent arrival in the same year.

Another result of duration in the U.S. is more subtle, though of perhaps greater magnitude: the increase in the returns to the locations and occupations into which immigrants moved after their arrival and the increase in the age at which wealth peaked. The increased advantage to those in particular locations and occupations is consistent with immigrants better matching themselves to places and jobs in the years after arrival. By 1860, a larger number of successful matches had been made than by 1850. The reduction in the wealth disadvantage of the Irish between 1850 and 1860, after accounting for all other observed characteristics, is also consistent with more successful matching: having come from Ireland with few skills and little capital may have been less of a handicap after a decade in which one could acquire additional skills or learn where best to put one's existing skills to work.

The later age peak in wealth in 1860 is consistent with immigrants over the course of their time in the U.S. economy being able to take better advantage of their comparative advantage in particular pursuits. For example, laborers relied more on simple physical strength than those with greater skill (craftsmen and those in white collar occupations) or those whose income was earned with inputs complementary to physical strength (farmers who used land, animals, and farm machinery). The peak of the age-earnings profile for laborers would have occurred earlier than for workers with greater skill or other inputs to employ, and the age-wealth profile of skilled workers or farmers would have occurred later as a result. More immigrants had been able to move into occupations which made use of more than physical strength by 1860, allowing them to postpone the age at which earnings and wealth peaked.

To see whether immigrants making changes in location or occupation (two of the most important choice variables in the regressions) are indeed responsible for shifting out the peak of the age-wealth profile, the regressions in Tables 5 and 6 were re-estimated, on the sample of immigrants located in both 1850 and 1860. The estimated age-wealth profile for the entire linked sample is shown in Figure 1. The peak rises from 48 years in 1850 to 62 years in 1860.

Figure 2 shows the result when immigrants who left the Northeast (most of whom relocated to the North Central and Northwest states) between 1850 and 1860 are excluded from the sample. The result is a much smaller shift in the peak of the age-wealth profile, moving from 47 years to 59 years, a shift of only 12 years rather than the 14 year shift when those who left the Northeast were included. Figure 3 does the same, but excludes those who entered farming between 1850 and 1860; the result is similar -- the peak shifts only 11 years. Finally, Figure 4 shows the result when those relocating from urban to rural places between 1850 and 1860 are excluded: the peak of the profile shifts only nine years.

The extent of adjustment resulting from such changes is probably a lower bound on the amount of adjustment actually experienced. The results in Figures 1 through 4 include the impact on wealth of changes in location or occupation, but exclude the impact of changes within locations or occupations -- finding a more favorable location within the Northeast or a better way to utilize one's existing skills, for example. The effect of such hidden adjustments is another component of the 6 to 8% return to duration.

V. Conclusions

Several of the same trends in the relationship between the age-wealth profile and duration in the economy uncovered here have been observed in the development of the Utah economy and the economies of several western states and counties. (Pope 1989 and Soltow 1975) In those cases where age-wealth profiles can be traced over a period of decades for an economy, an increasingly mature economy consistently displays a higher and later-peaking profile than had the same economy earlier in its development.

The explanation offered for those changes -- that in an increasingly mature economy, individuals may have had more opportunities to take advantage of their skills and thereby change the shape and peak of the age-earnings and age-wealth profiles -- might account for much of the difference between the profiles in 1850 and 1860. Immigrants who had spent a longer time in the U.S. (the 1860 sample) might be like individuals in an economy at a more mature stage of development, while recent arrivals (the 1850 sample), like individuals in a newly-formed frontier economy, had yet to sort themselves into the occupations and locations in which their skills could earn the greatest return.