Corporate Restructuring and its Impact on Value-added, Productivity, Employment and Wages

What kind of influence does corporate restructuring have on a firm's real value-added, labor productivity, employment figure and wages?

It is said that the result of corporate restructuring in Japan was massive job losses and redundancies in the 1990s. Indeed, looking at the specific reasons for leaving one's job out of all of those unemployed, the number of people who left work involuntarily - for reasons attributable to the workplace or business - rose by almost five times from 320,000 in 1992 to 1.51 million in 2002 (according to the "Labor Force Survey" conducted by the Ministry of Public Management, Home Affairs, Posts and Telecommunications). Furthermore, 64% of those involuntarily dismissed from work were core household breadwinning males aged 25 to 64. Today's insecurities over household income and employment have been attributed to firms, which had hitherto placed emphasis on job security, changing their principles and rapidly restructuring.

However, what if these firms had not been restructured - would it have really stopped unemployment from increasing? If a corporation did its utmost to safeguard jobs, that action in itself would hike up product costs and discard competitive power, making the firm liable to suffer a management crisis and increasing the chances of bankruptcy. In actual fact, many companies, even if they manage to avoid bankruptcy, have been suffering from falling orders and forced to cut back on payrolls. In the end, exactly what sort of influence does restructuring have on a firm's competitiveness, the number of its employees and their salaries?

One uses the word "restructuring" in a general sense, but exactly what it encompasses will differ from firm to firm. We want to compare two types of firms - those that have made changes to organizational structure and those that have not - and investigate the effects that restructuring has on value-added and labor productivity. Furthermore, by comparing subsequent changes in the number of employees and the level of their wages in the respective firms, we would like to analyze the effect of organizational restructuring. Of course, depending on the scale and kind of restructuring conducted, the effects will differ, but by analyzing an average case, we hope to provide resources for conducting debate hereafter on realignment of corporate organizational structure.

The trouble with this sort of analysis, though, is the time that must pass once restructuring measures are carried out before the effects are realized. The period immediately after restructuring has taken place is seen by employees as all bad - redundancies have been made and wages decreased. However, in restructuring, if the firm's competitiveness has been increased, bankruptcy avoided, and large lay-offs have been prevented, then the final conclusion will be very different. Conversely, if no impacts are felt immediately following restructuring, once a certain amount of time has passed and those effects have filtered through, then it is quite possible there will be adverse effects on employment.

Owing to the fact that the effects of restructuring change over time, there is a strong chance that the results of analysis will be distorted if they are based on data from only one point in time. Accurate investigation of the long-term effects absolutely requires panel data derived from follow-up surveys on the same firm throughout a number of years. The quantitative analysis conducted in this paper uses corporate panel data that RIETI has developed based on the "Basic Survey of Japanese Business Structure and Activities" conducted hitherto by the Ministry of Economy, Trade and Industry (METI) (see note 1).

The "Basic Survey of Japanese Business Structure and Activities" was begun in 1992 and is being continued even today. Those businesses surveyed by it include the manufacturers, mining operations, wholesale retailers, and businesses in the services industry - all with 50 or more employees and capitalized at ¥30 million or more - that fall within the jurisdiction of METI. This survey is notable for two characteristics. Firstly, it differs from the conventional "Census of Manufacturers" and other surveys focused on individual establishments - factories, works, plants, etc - in that it is a survey that targets company-wide activities. If this were an analysis of production activities, looking at raw materials and production volume in individual establishments, would suffice. However, when analyzing the influence of economic activity tackled by an enterprise as a whole - the effects that organizational restructuring, expansion overseas, research and development have on profitability, competitiveness and employment - conventional establishment-based statistics alone will not do. Statistics targeting the firm are essential.

The second characteristic of this survey is that it has been designed so that each firm is assigned a permanent corporation code, allowing the same firm to be traced throughout many years. Using this data, we analyzed time-series changes in the effects of corporate restructuring by comparing the changes in performance, employment and salary of corporations that had carried out organizational restructuring with those that had not.

The very first round of the "Basic Survey of Japanese Business Structure," which was implemented in 1992, asked companies whether they had made any changes to their organizational structure in the three years prior to that. In conducting our analysis, we would like to borrow the same questionnaire.

In doing this, it was important to bear in mind if there were any other factors apart from organizational restructuring that could be affecting the firm's performance. For example, the performance of each firm differs according to the type of firm it is, and a firm's technological capability can even be affected by factors such as research and development. Consequently, we analyzed the effects of organizational restructuring with econometric techniques, specifically, by adding all such factors that could be thought of to explanatory variables and expressing the growth rate of each type of company in respondent or explained variables.

For a long time now, corporate performance has been analyzed by investigating what growth factors influence it. According to this kind of research, the bigger and older a firm is the lower the firm's chances of going bankrupt but so is the growth rate of surviving corporations (see note 2). Taking this previous analysis into account, the research conducted for this paper used an analysis technique that also considered the impact on corporate bankruptcy in analyzing organizational restructuring (hereinafter referred to simply as "restructuring") of companies and subsequent changes in corporate performance.

The model employed in this research is as follows:

The coefficients were estimated so as to simultaneously maximize the likelihood function for the two above equations. Since the influence exerted by restructuring on corporate bankruptcy was not found to be significant, it was decided as a result that the following analysis focus on debating the effects exerted by restructuring on a firm's growth rate, based on the second equation and with regard to corporations that stay in business.

A firm's growth rate - it - is the growth rate of a firm, represented by i, over a number of years, represented by t. Here, growth rate is measured in terms of employment and value-added. In addition, we also analyzed the influence of restructuring on corporate performance measured in the rate of increase in growth rate of productivity (value added per worker) and growth of wages. The scale of the firm is determined using the number of employees and the age of the firm is derived by computing the year of the survey minus the year of establishment. The restructuring dummy variable is set as 1 if a firm i has been restructured, or 0 if it has not. The number of years elapsed - t - is a variable that expresses the number of years that have elapsed since restructuring. Finally, the interaction term (restructuring * number of years elapsed) is a variable that multiplies the restructuring dummy variable with the number of years elapsed, and the coefficient η3 represents how the performance of companies that have undergone restructuring differs from those that have not. In comparison to η3, the coefficient η1 of the restructuring dummy variable can be explained as showing the level of difference between the performances of companies immediately after restructuring and those without any restructuring. In addition to the variables introduced above, this research also used the following variables for analysis: the ratio of research and development cost to sales, the degree of dependence on specific customers, industry dummy variables.

The results of the research are presented in table 1. For each variable, the upper figure is the coefficient that was estimated, while the lower figure in parentheses shows the t-value. This t-value expresses the statistical significance of each variable and if it exceeds approximately 1.96, then the hypothesis that the estimated coefficient is zero is rejected with the probability 95%.

Let us see if a firm's performance - in other words, its growth rate of real value-added and rate of increase in labor productivity - is affected by restructuring.. The results can be seen in the first and second columns of table 1. First of all, in the case of the growth rate of value-added, the coefficients for the restructuring dummy variable and the number of years elapsed are positive, but the interaction term between them is negative. However, the t-value is below 1.96 for each variable. This means that the estimated coefficients measured here cannot be significant and it can be concluded that there is no great difference in the growth rate of value-added between a company that has undergone restructuring and one that has not. So, how about the rate of increase in labor productivity? The restructuring dummy variable is positive, and the t-value is also high, from which can be concluded that labor productivity increases sharply immediately after restructuring. However, the coefficient of interaction between the restructuring dummy variable and the number of years elapsed is negative and thus it shows that the effects of restructuring in this area lose strength as time passes.

So, is the improvement in corporate performance stemming from the restructuring shown here in any way linked to the expansion of employment?

In the third column, the impact that restructuring has on employment growth rate is analyzed, using the employment growth rate as an index for the firm's growth, and since all of the values of t are high, it suggests that a significant effect is being exerted. Focusing on each individual coefficient, it can be seen that the restructuring dummy variable has a negative t-valuehis means that employment decreases immediately after restructuring has taken place. On the other hand, the interaction term between the restructuring dummy variable and the number of years elapsed is positive. This shows that the employment growth rate of a corporation undergoes a relatively high increase once restructuring has taken place, compared to firms that have not been restructured. Graph 1 plots the path of employment growth rate based on this coefficient.

As can be understood from graph 1, companies that have been restructured see a substantial fall in employment in the period immediately after the restructuring has taken place compared to those that have not. The rate of change remains negative and the number of employees continues to decrease. But approximately four years after restructuring has taken place, the rate of decline falls below that of a firm that has not been restructured. In other words, the trend that can be understood from all this is that, on average, although the number of employees at a firm that has been restructured will fall sharply immediately after the restructuring, the rate of reduction will slow down as time passes, and from the fourth year onwards following restructuring employment at the firm will be more safeguarded than at a firm that has not been restructured. The fact that a difference in the employment growth rate appears between restructured firms and nonrestructured firms can, as has just been seen, be considered as one of the factors behind the discrepancy between the two types of company in the recovery of performance.

So, it has been confirmed that while restructuring will on average expand value-added per employee, on the other hand it also has the effect of raising productivity by reducing employment for several years subsequent to restructuring. However, in the end, does this kind of strengthening of the firm's competitive power, from the perspective of the remaining employees, bring economic benefits by raising wages? Since the total amount of cash earnings was recorded by the "Basic Survey of Japanese Business Structure and Activities," by dividing this figure among the total number of employees the wage per person can be calculated, which can then be used as an explained variable to analyze the effects of restructuring. The results are shown in the fourth column of table 1. Because the restructuring dummy variable and the variable produced by multiplying the restructuring dummy variable with the number of years elapsed do not hold any significance, and therefore these coefficients are not statistically significant, the conclusion can be drawn that reorganization does not have any effect on wages. In short, while the restructuring of a company is connected to improvement in productivity, the link does not go as far as raising wages to reflect this.

In this paper, the influence exerted on the employment, labor productivity, and wages of a firm following its restructuring has been investigated using corporate panel data. Up until now, corporate restructuring has on many occasions been accused of being a great contributing factor to the decrease in employment and the increase in unemployment. However, what has been understood based on the results of this analysis is that while there will be a temporary decrease in employment following the restructuring of a corporation, in the long-term it will not decline as sharply as that of a firm which has not been restructured. However, employees do not benefit from the fruits of restructuring in terms of wages.

Hereafter, in order to uncover what sort of effect corporate restructuring which has continued since the 90s has had on the labor market, analysis that delves even deeper than this research is needed. For example, there are many different kinds of restructuring, and research is called for that examines what type of restructuring takes place in what kind of situation, and if it is at all connected to the improvement in a firm's performance and expansion of employment. Furthermore, it is essential to investigate what differences exist in the effects of restructuring depending on the length of time and scale with which it is carried out, and conduct this analysis over an extended period of time.

>> Original text in Japanese