Andrew Seltzer

Andrew Seltzer is a Senior Lecturer in Economics at University of London, Royal Holloway College. He holds a B.A. in economics from Colby College and an M.Sc. and Ph.D. in Economics from the University of Illinois at Champaign/Urbana. He wrote his dissertation under the guidance of Lee Alston with Pablo Spiller, Jeremy Atack, and Wallace Hendicks, and he was on the EHA dissertation panel in 1994. He is an avid cyclist and chess player and also enjoys fine Italian food, good bagels, and the occasional glass of wine or beer.

Seltzer's research has focussed primarily on the history of labor markets in the United States, Australia, and Great Britain. His research has been centered around two major projects. The first examines the passage and early functioning of the Fair Labor Standards Act (FLSA), which established the American minimum wage law. The second focuses on personnel practices in the Australian financial industry using the payroll records of several banks.

In "The Political Economy of the Fair Labor Standards Act" Seltzer looks at the origins of the American minimum wage law. The minimum wage provisions of the FLSA were an important deviation from earlier legislation in that they provided for a single minimum wage for all covered workers. Apart from the period between 1940 and 1946, when Industry Committees were given limited discretionary powers, the minimum rate could only be changed by an act of Congress. To this day Australia and much of Western Europe retain systems where minimum wages are determined by an arbitration process and contain differences across occupations, industries, age groups and regions. In the past decade Great Britain and New Zealand have abandoned this sort of arbitration and adapted American-style minimum wage legislation. To understand why in 1938 a proposed arbitration system was abandoned and why decades of congressional resistance to national minimum wage legislation were reversed, Seltzer examines the passage of the act. Much of the historical literature has portrayed the Congressional battle as a struggle between North and South, with the South siding with Republicans in opposition to the bill. Seltzer argues that this view is misleading as minimum wages had considerable popular support in the South, and regression analysis shows that the groups opposing the minimum wage (employers of black and low-wage labor) were the same in both regions. Furthermore, one of the original sponsors of the bill, Hugo Black, was from Alabama, and the pivotal event in generating legislative support for the bill was the Florida Senatorial election of Claude Pepper, who had made the FLSA a centerpoint of his campaign. Seltzer argues that voting on the FLSA was almost entirely determined by economic factors and legislator ideology, although the importance of ideology waned as the bill became an election issue. Finally, he argues that the South was successful in reducing the role of arbitration in the bill. In the final compromise, Industries Committees were given very limited discretionary powers, only being empowered to determine the timing of industry-specific implementation of the final $0.40 minimum rate. Southerners had feared that an unelected Wages and Hours Board would impose high minimum wages in order to reduce the competitive advantages of their industries.

In "The Effects of the Fair Labor Standards Act" Seltzer examines the consequences of the minimum wage on two of the largest low-wage southern industries. Both the seamless hosiery and lumber industries had high-wage firms located primarily in the North and low-wage firms located primarily in the South. Employment at seamless hosiery firms that paid an average wage of less than $0.325 per hour in 1938 (the minimum rate implemented in October 1939) decreased by about 17 percent between 1938 and 1940, whereas employment at firms paying over $0.325 per hour remained virtually unchanged over this period. Much of the decrease in employment at low-wage firms was due to automation of the knitting process. It is likely that there would have been greater job losses except for the War, which created an increase in product demand and resulted in the procurement of the machine tools necessary for further automation. The lumber industry epitomized regional differences in labor markets. Lumber production in the South tended to occur in small, atomized units where unskilled workers processed second or third-growth timber, whereas in the North and particularly the West highly skilled lumbermen processed virgin forest. In July 1940, fully 94.3 percent of Southern workers earned exactly the $0.30 minimum wage, whereas approximately 80 percent in the North earned over $0.475. Between 1939 and 1942 employment in the industry increased by over 9 percent nationwide due to War-related demand increases; however, the southern share of output of both hardwoods and softwoods decreased considerably. Finally it is argued that southern lumber managed to mitigate the effects of the FLSA by legal avoidance and illegal evasion of the minimum wage. The FLSA initially only covered firms that engaged in interstate commerce. In 1938, prior to the implementation of the minimum wage, approximately 60 percent of the 500 Texan sawmills participated in interstate commerce; by 1941 only about 4 percent continued to do so. Furthermore, the atomized nature of southern lumber made inspection difficult. Department of Labor statistics show that in 1946 lumber firms were inspected less frequently than those in other industries and were significantly more likely to be in violation of the law.

Seltzer's other major project uses the payroll records of Australian banks and contract theory to examine personnel practices in the industry. The primary data set is the career histories of employees of the Union Bank of Australia (UBA), one of the largest trading banks at the turn of the century. The records contain date of birth, details of previous experience, date and reason for exit, and the wage, position, and branch location over employees' entire careers at the UBA. The data set records annual wages, position, and location for 1,767 employees who either were present in November 1887, entered prior to April 1900, or entered between 1917 and 1927 and were still present in 1932. In addition separate samples have been gathered from the bank of South Australia, The Bank of Australasia, the Bank of Adelaide, and the Queensland National Bank.

In "Personnel Practices at the Union Bank" Seltzer examines the structure of internal labor markets using primarily the 1888-1900 entry cohort data. The paper begins by examining the internal hierarchy of the bank. The paper then examines compensation. It is shown that pay was tied to a few simple, and easily observable rules. The bank maintained a 7-year salary scale for junior employees, and rarely deviated from the scale. After the UBA reduced the salary scale in 1888, it only adjusted the pay of new recruits and not those already in employment. Even after seven years tenure, by which time employees were no longer covered by the scale, salaries seem to have been determined by a few simple, impersonal rules. Seltzer shows that the primary determinates of salary were prior experience, position, branch characteristics, and particularly tenure. The influence of ability on wages is more difficult to measure, as the records do not contain a direct proxy for ability. The paper attempts to separate out the influence of ability on wages using two types of fixed-effect models. In the first model, he examines the annual increment of log real wages given to newly promoted employees using tenure as a control variable and the increments in prior and subsequent years as an employee-specific indicator. Consistent with a rules-based wage policy, it is shown that there was a strong tenure-based component of the increment and that employees received a one-off pay rise at the time of promotion in addition to their normal tenure-based increase. In the second model he examines the natural log of real wages using the pooled career data. Again the results are consistent with a rules-based wages policy as a few position, tenure, location, and outside experience variables explain over 80 percent of variation in wages. Tenure is the most important explanatory variable. The bank operated a policy of deferred compensation, with log real wages increasing at an increasing rate after about 30 years tenure. Using a fixed effects specification Seltzer shows that about half the effect of position in the salary regression is due to the fact that promotions were systematically given to higher ability employees. Finally, he examines wages policy over the business cycle. The evidence suggests that nominal wages, and to a lesser extent real wages, were shielded from the external economy and that cyclical factors add little explanatory power to the wage regressions.

In "Salaries and Career Opportunities" Seltzer further examines the influence of tenure in banking personnel practices using the data from the UBA and four other banks. The theoretical literature has provided three reasons why one would expect tenure-based personnel practices in Australian banking. First, workers with sufficient secondary education to enter banking employment were scarce in Australia. Secondly, most of an employee's training was done on the job, and thus turnover was very costly. Finally, banks were extremely vulnerable to theft and other forms of employee opportunism. Seniority-based personnel practices are likely to reduce opportunism because dishonest employees risk the loss of deferred compensation, such as future wage increments and their pension. The paper outlines the role of seniority in several personnel practices. In promotion decisions seniority played a limited role. Contemporary accounts indicate that competent and experienced workers should not be by-passed for promotion; however, experience alone was not sufficient. The evidence indicates that there was a fairly standard promotion ladder and that accounting positions were a "feeder path" to the level of manager. There existed "fast-tracks" to the level of manager; however, even the most talented junior employee needed considerable time to work their way up. The paper then examines wages. The UBA offered employees a generous pension from at least as early as 1877, and cross-sectional regressions show considerable deferred compensation in the 1880s and 1890s. Beginning in 1921 the relationship between seniority and wages became enshrined in law as government tribunals set industry-wide wages covering the first 13-18 years of tenure. Time series evidence shows that the late-career real wage profiles of 1917-27 entrants were also steeper than those of earlier entrants. The paper then considers the effectiveness of these seniority based policies by examining turnover. Turnover at the UBA was exceptionally low by contemporary or even modern standards, with all available evidence showing less than 10 percent annual turnover prior to 1900. Employees on high wage or at higher-level positions were less likely to resign or be dismissed, employees accruing rapid pension growth were less likely to retire, and more senior employees were less likely to leave for reasons other than retirement. Finally, the paper shows that other banks used very similar personnel practices to those of the UBA.

"Controlling and Motivating the Workforce" and "The Nature of Bank Work" examine the monitoring system that was an essential complement to the internal labor market. In addition to the possibility of theft or fraud, the bank had to ensure that its employees were acting in the bank's best interests. For example, without monitoring a manager could make loans on insufficient collateral or make inherently risky loans to his friends. Much of the book-keeping procedures were designed to minimize the bank's exposure to fraud. For example, the bank maintained rules that separated handling money and recording transactions, employees had to balance the ledgers before they were allowed to go home at the end of the day, and over half of all employees at the larger branches spent at least some of their time checking books that had been prepared by other employees. Employees were frequently moved between branches in order to make it more difficult for employees to form collusive arrangements to defraud the bank with other employees or with customers. Another consequence of frequent moves was that an employee's successor would carefully check the books upon commencing the assignment. These rules were enforced by an inspection staff that frequently made unannounced visits to the branches, and had the authority to examine all of the books. Even behavior outside the workplace was regulated, and employees were prohibited from engaging in activities that might create the temptation to steal such as gambling, posting bail, holding certain shares, accumulating debt, or even getting married on a low salary. Transgressions of any of the banks rules were swiftly and severely punished. Despite these practices and the extensive internal labor markets outlined in the papers cited above, the careers of approximately 7 percent of 1888-1900 entrants ended in dismissal for dishonest behavior.

"Was the 1931 Wage Cut Successful?" examines the impact of a national incomes policy designed to counter the depression. The Great Depression hit Australia worse than any nation except the United States and Germany. Australia's economy was dependent on primary exports and consequently the collapse in commodity prices and world trade had a flow-on effect throughout the economy. In order to counter the effects of this supply shock, the Commonwealth Court of Conciliation and Arbitration decided in 1931 to cut the real value of the Basic Wage by 10 percent. The banking industry was covered under a separate award; however, this too was cut by 10 percent in nominal terms in February 1931. The paper examines the effect of the Award wage cut on the wages of the 1917-27 entrants. It is shown that the Award cut was not passed on to workers. Controlling for tenure, real wages actually increased by 2.65 percent in 1931, a magnitude slightly larger than the increase in 1928, 1929, or 1930. In 1930 approximately 25 percent of sample employees earned more than the Award rate; however, in 1931 this increased to 69 percent. Finally, the paper speculates on why the UBA refused to pass the award wage cut on to its workers. It is argued that not passing on the wage cut was a form of shielding workers from the external labor market similar to that analyzed in "Personnel Practices". It is further argued that this sort of implicit contract may have been fairly prevalent in the 1930s, by which time at least half the Australian workforce was employed in firms that were sufficiently large to support internal labor markets.

"A Comparison between Unit and Branch Banking" examines the consequences of the institution of branch banking on the Australian financial system and the wider Australian economy. Much of the literature on branch banking considers its effects on the diversification of risk and the exposure of banking systems to large-scale crises. However, there exists an older literature that argues that branch banking facilitates functional specialization across banking offices. The key to this argument is that a unit bank must both collect deposits and issue loans, whereas a branch can specialize in only one, provided that other branches of the bank can make up the difference. The paper shows that this was particularly important in rural areas, which often supported large-scale capital intensive primary industries, but could not raise sufficient capital though local deposits to support these industries. Using data from the Annual Reports of the UBA's branches it is shown that urban branches tended to specialize in deposit collection and rural branches tended to specialize in issuing loans. Furthermore, it is shown that even under conservative estimates of the fixed costs of running a branch, the majority of rural branches would not have been viable had the loans they could issue been limited by the deposits they were able to raise.

"The Impact of Career Interruptions" uses the sample of 1917-27 entrants to evaluate the effects of volunteering for the Australian Imperial Forces during the Second World War on employees' subsequent careers. Approximately 38 percent of those remaining at the UBA in 1939 enlisted in the AIF, taking an average leave of 3 years and 7 months. The paper begins by comparing the careers of volunteers and non-volunteers prior to 1943. It is shown that after controlling for tenure, age, and branch location there were no systematic differences in salary between the two groups. Even in the year prior to taking military leave, the increment received by volunteers is almost exactly the same as the increment received by non-volunteers. Likewise prior to 1943 volunteers were as likely to be in position and branch assignments that indicated strong future promotion prospects as were non-volunteers. The paper then compares career outcomes after 1946. Surprisingly, regressions that control for personal characteristics but not position show no differences between volunteers and non-volunteers. Even more surprisingly, adding position dummies shows that volunteers ceteris paribus received higher wages than non-volunteers. Because volunteers averaged 2 fewer years from hiring to exit and further shortened their career by 3.6 years of military leave, this evidence is interpreted as positive discrimination by the bank. This is then compared to the numerous studies of career interruptions as a result of child-birth which invariably show that time out of the labor force has a large adverse affect on future prospects.

One paper that deviates from the theme of historical labor markets is "Neighbourhood and Family Effects", which uses a survey of students in their final year of secondary education to analyze the determinates of Melbourne youths' decision on whether to proceed to tertiary education. Considerable research has been devoted to the role of families in youths' decisions; however, far less has been devoted to neighborhoods. If neighborhoods play a large role, for example through peer groups, networks, or demonstration effects, then rising regional inequality may be creating poverty traps. The analysis shows that neighborhoods indeed have a large and significant effect on youths' educational aspirations, and that there exists a strong peer group effect. Even after controlling for family background, the average wealth of the neighborhood has a strong positive effect on the whether a student intends to pursue tertiary education. Finally, it is argued that policies that promote equality and integration may have intergenerational effects through their influence on educational choice.

Seltzer's future research on banking personnel records will include the English banking industry. At present he is gathering data from the 1890-1940 wage books of Williams Deacon's Bank (WDB). At the turn of the 20th century the WDB and UBA were similarly sized branch banks. However, there are two important differences between the two banks. First, Australian labor markets, particularly for workers with a secondary education, were much thinner than English labor markets. Secondly, the WDB branches were geographically highly concentrated in the Manchester region, whereas the UBA branches were spread thinly over Australia and New Zealand. This difference in geographic concentration meant that inspection and other forms of monitoring were much easier and less costly for the WDB than for the UBA. In "Salaries" he speculates that tenure-based personnel practices were largely driven by search costs and agency problems, and one test of this hypothesis is whether these practices were less prevalent at the WDB. Preliminary evidence comparing the 1887-1899 UBA cross-sections to the 1890-1901 WDB cross-sections indicates that deferred compensation was extremely important at the UBA and virtually non-existent at the WDB.

A second paper will look at the use of performance-based pay and promotions in the banking industry. Both the UBA and WDB retained records of the annual deposits, advances, and profits of the individual branches. These data can be linked to the payroll records of employees in order to create an indicator of performance. This research will examine whether branch performance was used to make immediate salary adjustments, used over time in promotion decisions, or largely ignored in personnel practices. Another issue that this research will address is the difference between managerial and non-managerial compensation. It is likely that branch results were more important for managerial compensation, as the manager had more direct control over bringing in and retaining business than other branch employees.

A third paper will look at discrimination in the industry by comparing the career experiences of different ethnic groups. Because the records contain the names of each individual it is possible to identify the ethnic origins of their father's family. Australian social historians have outlined discrimination against Irish and German migrants prior to Federation. This paper will test three different models of discrimination: employer discrimination, which implies lower wages being paid to minority groups; employee discrimination, which implies occupational segregation in the workplace; and customer discrimination, which implies that minorities would be less likely to be positions such as teller and manager. It is also hypothesized that the extent of discrimination may have differed considerably across colonies and later states.

A fourth paper will look at personnel practices over the business cycle. "Personnel Practices" shows that there was considerable wage shielding and "Salaries" shows that turnover varied relatively little over the business cycle. A document prepared by the auditors of the bank's pension raises another interesting possibility. The bank appears to have discouraged early retirements in the depression of the 1890s because its pension funds were performing poorly. These results suggest that, if widespread, these sorts of personnel practices may have increased the amplitude of the business cycle. In addition to examining the national business cycle, it is possible to identify localized business cycles resulting from commodity price swings. The Annual reports of the branches identify the major commodities produced by the bank's customers. The paper will examine the sensitivity of branch openings and closings, salaries, promotions, job rotation, and exits to local and national cyclical swings in the economy.

A fifth project addresses the effect of mergers on personnel practices. It is often argued that one of the primary purposes of mergers is that they provide a mechanism to renegotiate various contracts, including those with workers. The banking industry experienced numerous mergers during the time period studied. The banking data set is unique in that it contains numerous career records of employees who began working in a bank that was later absorbed into another bank. The mergers covered in this data set include: Manchester & Salford and WDB in 1890, Sheffield & Rotherham and WDB in 1907, the Bank of South Australia into the UBA in 1892, and the formation of the ANZ in 1951 with the merger of the UBA and the Bank of Australasia. It is hypothesized that labor contracts may have been restructured after the UBA absorption of the BSA, which was financial insolvent at the time of the merger. However, it is less likely that there was radical restructuring in the creation of the ANZ or the two WDB mergers because the merging firms were all financially sound.

Bibliography

"Salaries and Career Opportunities in the Banking Industry: Evidence from the Personnel Records of he Union Bank of Australia" (co-author Kenneth Simons), Explorations in Economic History, vol. 38 (April 2001), pp. 195-224.

"Controlling and Motivating the Workforce: Evidence from the Banking Industry in the Late 19th and Early 20th Centuries", Australian Economic History Review, vol. 40, no 3 (November 2000), pp. 219-38.

"Personnel Practices at the Union Bank of Australia: Panel Evidence from the 1887-1900 Entry Cohorts" (co-author David Merrett), Journal of Labor Economics, vol. 18, no. 4 (October 2000), pp. 573-613.

"The Nature of Bank Work and Worker Monitoring: A Study of the Union Bank of Australia in the 1920s" (co-author David Merrett), Business History, vol. 42, no. 3 (July 2000), pp. 133-152, also published in David Merrett, editor, Business Institutions and Behavior in Australia, Frank Cass, London & Portland, OR., 2000, pp. 133-152.

"Neighbourhood and Family Effects in Educational Progress" (co-author Ben Jensen), Australian Economic Review, vol. 33, no. 1 (March 2000), pp. 17-31.

"The Effects of the Fair Labor Standards Act of 1938 on the Southern Seamless Hosiery and Lumber Industries", Journal of Economic History, vol. 57, no. 2 (June 1997), pp. 396-415.

"The Political Economy of the Fair Labor Standards Act of 1938", Journal of Political Economy, vol. 103, no. 6, (December 1995), pp. 1302-1344.

"The Impact of Career Interruptions on Subsequent Career Prospects: Evidence from the Personnel Records of World War II Volunteers", unpublished manuscript.

"Was the 1931 Wage Cut Successful? Evidence from the Personnel Records of the Union Bank of Australia", Melbourne Working Paper in Economic History 20, 1997.

"A Comparison between Unit and Branch Banking: Australian Evidence on Portfolio Diversification and Branch Specialization, 1860-1930", unpublished manuscript.