Banking on Decent Jobs

New technologies came slowly to the banking industry, but change is now coming faster and faster. The question is, who will benefit?

Consider the check, a banking mainstay introduced shortly after the Civil War. For more than a century, checks were processed by hand. When a check was received, a clerk, sometimes wearing a green eyeshade, wrote down the amount, whether for $1 or $1 million. The check, bundled with others, was sent, often in a cardboard box, to another clerk. This clerk added up all the checks in the batch and verified that the sum equaled that of the deposit slips. Other clerks then posted balances and sent each check for collection to the bank on which it was drawn. Basically the same processes used to handle J.P. Morgan's checks in the 19th century were used to handle David Rockefeller's in the mid-20th century.

One highly visible innovation appeared in the mid-1980s, the ubiquitous automated-teller machine. But such innovation lagged in bank back offices. No more. Today, information technologies are dramatically altering back offices, in America and abroad.

An earlier breakthrough came to back offices in the 1950s, when Bank of America introduced magnetic ink character recognition. Machine-readable numbers began appearing on checks -- the number of a check and of the bank account as well as the bank's routing number. Reader-sorter machines were developed, and checks could be handled more quickly and by fewer workers. Productivity increased, costs were reduced.

But this soon became old stuff -- the machinery of the punch-card era. Now new computer technologies, including optical character-recognition technology and check imaging, have arrived.

To examine the impact of these new systems and technologies on banking and its workforce, three academicians -- David H. Autor, an economist at the Massachusetts Institute of Technology, Frank Levy, a professor of urban economics at mit, and Richard J. Murnane, a professor of education and society at Harvard University -- obtained access to Cabot Bank, a pseudonym for one of America's 20 largest banks. Cabot is fairly typical, with extensive retail and commercial banking operations. It moved aggressively to acquire other banks. It now has branches in numerous states and many countries, and its retail operations have more than doubled in the last decade.

In the mid-1990s, Cabot, like many other banks and as in many other industries, began adopting new computer-based technologies. The Cabot study focuses on two check-processing departments: deposit processing, which handles the 2.8 million checks that come to the bank each day, and exceptions processing, which handles the 3 percent (84,000 checks) that require more detailed handling. Checks sent to the exceptions department include those that need verification of signatures, ones returned for insufficient funds, stop-payment orders and checks greater than $2,000.

The two departments were operating under an old-fashioned organization system, with extensive paperwork and routinized jobs often requiring little skill and offering numbing repetition. Most were filled by workers with only high-school educations. These jobs were perfect targets for new computer systems.

In the processing department, the key job was proof-machine operator, which required a series of routine steps. Workers took staples and paper clips off checks, organized the checks so that they all faced the same way, keyed in the amounts, added up the total and made sure it equaled the total on the deposit slip, attempted to reconcile mistakes and sent finished checks to a machine that sorted them by account numbers.

The checks that required additional handling were forwarded to the exceptions department. Jobs in this division were narrowly defined. Workers in one group verified signatures; another group handled stop-payment orders; a third handled checks to be returned for insufficient funds. Here, too, work was organized in an archaic manner.

For example, if a check had a questionable signature, someone in the exceptions department (the employees were usually women) searched for an authorized signature card and compared the signature to the one on the check. If a discrepancy was found, the worker filled out a paper form and sent the questioned check and the form to a superior for action. The paperwork could go through three or four levels before someone finally decided whether the check should be paid.

Each worker spent a significant part of the day searching in boxes for correct checks or moving checks to other workers. The jobs were frustrating, and turnover was substantial, about 30 percent a year. The bank accepted this turnover because, with such minimal job skills, workers could be easily replaced.

Workers in the exceptions department knew only one task and had little knowledge of other processes. "We were in a situation where people checked their brains at the door," a manager told the authors. Delays and complaints were common.

Then, in the 1990s, circumstances forced Cabot and other banks to address how the work was organized and to consider investments in new technologies. The number of checks to be processed was increasing rapidly, in part because of acquisitions. Federal banking regulations limited the number of days a bank could hold checks before customers could receive their money and required paper checks to be returned quickly to the banks on which they were drawn before the deposit bank could obtain its money. Cabot also wanted to decrease the handling time for checks to minimize the bank's costs.

In 1994, Cabot seized on the new imaging technology. With check imaging, a high-speed camera takes a picture of the front and back of checks. People in several departments can examine checks at the same time, and checks no longer have to be walked from floor to floor, relieving a time-consuming and costly bottleneck. A second new technology scans and captures most checks and deposit slips as they are processed. These two technologies eliminated the most labor-intensive tasks in the departments.

As with much new technology, however, the Cabot plan brought not just solutions but questions, too. If machines were taking over part of a job, how would the remainder of the job be structured? Could current workers be retrained to use the new technologies? How would costs and productivity gains be balanced? And who would benefit from the savings in costs and increased productivity? This last is an age-old question when new technologies are added, going back two centuries ago to textile mills, iron and steel mills, and railroads.

The Cabot managers took two differing approaches. In check processing, the bank adapted the old tradition of narrowly defined tasks to new technologies. In the other, exceptions, management used the new technologies to reorganize work to demand greater problem-solving skills. To Autor, Levy and Murnane, this is an example of the discretion managers have in choosing how new technologies affect work. It suggests, the authors say, that new technologies, properly installed, can bring better-paying, more interesting work.

Here is what happened at Cabot: In the processing division, following the advice of the image-processing equipment vendor, the traditional job of proof-machine operator was split. A low-level job, much like the old one, was created for a worker to remove paper clips and staples from checks and put the checks in the reader-sorter machine. That job paid $9.51 an hour in 1998, less than the previous similar jobs in that division, which had paid $10.03 (in 1998 dollars).

Under the new system, the computer did the middle step, reading (instead of keying in) the amount of the checks that had clear printing or handwriting. For checks that were not clear, two new jobs were added: that of a keyer, who tried to decipher the amounts on unclear checks, and that of an image balancer, who balanced final deposits. The first job paid $10 an hour plus incentives, the second $11.

The bank retrained proof-machine operators for the new, higher-level jobs. Over the first 10 years under the new workplace structure, increases in efficiency cut the number of jobs in the division from 67 per million checks to 53 per million. The jobs were more narrowly defined than ever. About a third were less skilled than before, illustrating a point made by experts that technology often creates jobs with reduced skills. On the other hand, the keyer and image-balancing jobs required higher skills than the jobs they replaced.

The exceptions department took a different approach, with "fewer people doing more work in more interesting jobs," as the bank vice president in charge of the reorganization put it. Here, managers involved the workers in deciding how to rearrange their jobs to use the new technology. Managers met with workers in groups and asked them what irritated them about their work. One reply was that jobs were too narrowly constructed. For example, rather than have one worker deal with stop-check orders, one with overdrafts and one with other exceptions, why not cross-train workers so that a single employee would be responsible for all exceptions on a customer's account, regardless of the category? This would expand jobs, not reduce them, and make them more varied and interesting.

This made sense to managers, and the bank made the change before the new check-imaging technology was in place and after workers had received 80 hours of training. Productivity jumped immediately. By the end of 1995, 530 workers were handling 65,000 checks a day, compared with 650 before. Once the check-processing technology was in place, the number of employees needed for the work dropped to 470.

Overall, a greater proportion of workers received better-paying, more diverse jobs than received lesser-paying, narrower jobs. These jobs often demanded new problem-solving skills. Some workers felt overwhelmed by the new demands of their work, even after training. The bank told the authors that new hires in the exceptions division tended to have at least some college experience, which means that the changes have led to a decrease in the already short list of jobs available to high-school graduates. The technology also adds substantial job insecurity, as jobs are now easier to move to another part of the country -- or to another country. "If key punching can be centralized in Atlanta, it can be centralized in Bangalore," Murnane said.

There are other lessons to be learned from Cabot's experience. Managers indeed have discretion in how work is reorganized and information technologies are installed. But, according to Levy, the nature of the work that is done will, in most cases, determine how computer technologies are installed and what tasks will be eliminated. If tasks can be done faster and more cheaply by computers than by workers, he says, the work will be transferred to computers.

In the processing department, most tasks were simple, routine and conducive to being computerized. Computers in large part were thus substituted for workers. It's not at all clear that managers could have chosen to upgrade work quality and skill requirements in the processing department.

In the exceptions department, tasks were more complex. "Computers could not read what was on the mind of the people writing the checks -- that's nothing that could be programmed," Murnane says. In this department, Murnane observed, computers could complement the tasks of workers but not perform or eliminate them. Murnane also said that management in each department was equally humane, but that the processing department (or downstairs department), rather than the exceptions department (or upstairs department), offered more opportunities for substitution of computer tasks for human ones.

Each author made the point that bosses will be bosses. That is, in installing computer technologies, as in shrinking or expanding businesses, hiring or firing workers, setting sales and wages, executives will do what saves the most money or promises the largest return.

If one executive uses computers to cut costs regardless of the effect on workers, and another, perhaps more interested in workers' fates, attempts to use computer technology in a manner that is less destructive to workers and their jobs, the second executive, Murnane notes, is not going to fare very well in a competitive business environment if costs are not reduced as much as in the first situation.

"The high-road strategy is only going to last if it not only leads to improved jobs but also an improved bottom line or at least improved productivity," Murnane said. "It's not a question of 'good cop, bad cop.' If the strategy that leads to better jobs does not lead to improved productivity, that strategy is not going to survive."

The decision of how to reorganize a workplace and bring in new technologies belongs essentially to management, especially when there is no union. Cabot is unorganized, and only about 2 percent of the U.S. financial-services industry belongs to unions. Still, it is clear to researchers that managers have more discretion than they sometimes appreciate in reorganizing workplaces and as new technologies are introduced.

Eileen Appelbaum, director of Rutgers University's Center for Women and Work and a co-editor of "Low-Wage America," said the banking study shows that managers in the nation's new economy must "sit down with workers and discuss work processes before new technologies are installed." Cabot, she says, achieved a "majority of productivity gains by rethinking the work process" before new computers systems were put into use.

Appelbaum also notes that in times of full employment, higher-skilled workers are often difficult to find. The banking study, she says, shows that the "existing high-school-trained workforce can be trained" to use new computer systems. In most cases, she added, "Computer skills are not the barrier." What requires more attention, once new computer systems are installed and workers trained to use them, is training workers in problem solving, as Cabot had to do in the exceptions department.

Overall, the Cabot study authors say, employers should seek out employees' ideas on how their jobs should be organized. Employers should also understand, Appelbaum says, that workers, including those with only high-school educations, can be trained and retrained. Many observers of the American workplace have made the same points for years. One of them was William "Big Bill" Haywood, the Industrial Workers of the World organizer of a century ago who said, "The manager's brains are under the workmen's cap."

Change can benefit workers as well as employers and customers. If our society has the imagination to invent new technologies, it can be imaginative enough to deploy them in ways that create good jobs. Outcomes depend less on technological imperatives than on the ancient question of who decides.