9.2The Effect of Information Technology on Income Inequality

As discussed in the previous section, one demand-side explanation of increasing income inequality proposes that globalization and the extraordinary increase in high technology investment has shifted demand from low-skilled to high-skilled workers. For example, computers may be seen as a factor increasing income inequality, as reported by The Economist (1999):

‘Information technology replaces the unskilled; less demand means lower wages. At the same time, computers complement the skills of more sophisticated types – the “knowledge workers” who represent (...) the future of work. This complementarity raises individuals’ productivity and thereby increases their earning power. The prosperous get more so, the unskilled get dumped. ’

Levy and Murnane (1996) studied the effects of computerization on the demand for skilled and unskilled labor. They found that the computer revolution seems to have been responsible for a large part of the decline in the demand for unskilled labor during the 1980’s. Levy and Murnane (1996) studied the custodian unit of a bank intensively and noticed that computers generate two opposite effects:

‘By changing skill requirements, computerization increases the optimal ratio of skilled to unskilled labor per unit of output. By improving labor productivity, computerization nonetheless reduces the quantity of skilled labor per unit of output.’

Computerization has, however, not eliminated the need for knowledge that underlies the routine work. That is why training must also be accompanied by hiring of new graduates. Levy and Murnane found that more than one out of four trained workers left the bank after the training period, generating extra costs for the bank. Those costs are associated with the computerization of the firm, and might constitute another reason for the failure to see the important productivity gains expected from IT. The trained workers who have left the bank are most likely to sell their new skills to another company, which will benefit indirectly from the bank’s spending in training program. The negative effect on the bank and the positive effect on the new hiring firm should even out at the aggregate level, but are part of the reason why some firms might succeed better than others with IT. This fact might generate the need for government intervention, which will adjust the skills of the labor force to the needs of the new “computerized” firm. However, private markets might have ways of reducing turnover of trained workers. For instance, firms may avoid these extra costs by paying their workers a wage equal to their marginal revenue product while they are being trained (trainees get usually paid less than trained workers). Hence, trainees would have incentives to stay with the training firm.

As discussed in Chapter 3, David (1990) offered an explanation of the productivity paradox through historical consideration, pointing out the “diffusion lags” that have accompanied most of the great scientific discoveries of the twentieth century. The reason for this delay in productivity results comes in part from the labor force, which is not ready to use new technologies right away. Workers have to gain the necessary skills, and producers of new technologies have to improve their interface with humans.

Autor, Katz and Krueger (1998, hereafter AKK) studied the effects of computers on the labor market. They looked at the changes in the relative supplies and wages of workers by education from 1940 to 1996. They note that the literature relates the importance of technology to wages because of a skill-biasing effect and a demand shift to more educated workers. However, AKK argued that in order to evaluate the skill-biased technological change, it is necessary to (1) use a framework combining shifts in both the relative demand and relative supply of skills, (2) consider a longer time frame, (3) look at the relationships among observable technology indicators and skill upgrading over this long time period. They found that the utilization of more-skilled workers is greater in the most computer-intensive industries, but could not conclude whether a causal interpretation of this relationship is appropriate or not.

Krueger (1993) used data from the Current Population Surveys (CPS) to determine, on one hand, if workers using computers at work earn more than similar workers without computer skills, and, on the other hand, if the premium associated with computer skills can account for the increasing wage inequalities in the 1980s. According to Krueger (1993):

‘The new computer technology may be a complement or a substitute for skilled labor. In the former case the computer revolution is likely to lead to an expansion in earnings differentials based on skill, and in the latter case it is likely to lead to compression in skill-based differentials.’

Krueger estimated wages equations by OLS, and found that the wage premium to workers using a computer at work was 10 to 15 percent between 1984 and 1989. Krueger also estimated a wage equation with and without a dummy variable controlling for computer use and found that “nearly 40% of the increase in the return to schooling can be attributed to the expansion in computer use.” He concludes that these results suggest that the computer revolution has certainly contributed to changes in the wage structure of the 1980s.

However, Krueger’s results must be taken with caution. A study from Dinardo and Pischke (1997) has stressed the importance of causality and cast some doubt on the interpretation of results concerning the wage differential associated with computer use. Indeed, the authors have measured a “large differential for on-the-job use of calculators, telephones, pens or pencils, or for those who work while sitting down.” Obviously, these characteristics do not have a real effect on wages. Thus, this study was intended to be a warning for careful interpretation of results.

Artus and Lefeuvre (1998) argued that between groups inequality rose mainly in the 1980s, and remained relatively stable in this last decade, which was “the most computerized.” However, in the 1990s, there appeared a new type of inequality: wealth inequality. They reported that over the 1989-1998 period, the wealth of households who earn income greater than $100,000 increased by 18%, and by only 10% for others. Similarly, the wealth of college-educated workers has increased by 46% and 20% only for others between 1995 and 1998. Still, Artus and Lefeuvre noted that this type of inequality may not be attributable to the “new economy” only. Indeed, the wealth of the richest households may be simply due to the growth of the stock market, which constitutes an important part of this wealth. In the new economy, workers’ capacity to react, innovate and adapt to new challenges becomes more important than their education level. The next section looks at income inequality in a spatial dimension, through a survey of some literature on regional income inequality.