With the advance of the digital economy, an increasing number of companies are facing changing business characteristics. As the uncertainty and complexity surrounding technological trends and the global situation grow, risks are increasing for business strategies that depend entirely on traditional business models. As a result, companies in infrastructure and smokestack industries have also started to invest in new businesses which were previously unfamiliar to them. Additionally, as the scope of activity for new business models represented by terms such as "platform" and "ecosystem" expands, companies are eagerly engaging in open innovation and business partnerships in order to become platform operators in their business domains.
Changes in business characteristics alter the optimal personnel management system. For example, let us define two simple different business categories which I call "Guardian" and "Star" businesses. Figure 1 shows the distribution of corporate profit for the two. Corporate profit and its frequency are represented by the horizontal and vertical axes, respectively. Let us define businesses whose upside potential is limited, but downside risk is significant as "Guardian businesses." Typical examples include infrastructure companies, including power generation, communication and transportation companies. While guardian-type companies can rarely earn profits much higher than expected, they could suffer huge losses in the event of an accident or failure. On the other hand, for "Star"-type companies, downside risks are limited, while their corporate value may rise much higher than expected if their business proves successful. Platform companies as represented by GAFA (Google, Amazon, Facebook, Apple) and other companies that have the potential to create new markets belong to the "Star" type.
The personnel management systems that are optimally suited to these two types of businesses are entirely different. As shown in Table 1, whereas a flat organization that enables prompt decision-making is desirable for star type businesses ("Stars"), a multi-layered organization that reduces the risk of making mistakes is necessary for guardian type businesses ("Guardians"). For Stars, a bonus system that encourages risk-taking—namely a system under which employees receive bonuses in accordance with their contributions to profits but under which their pay is not affected by losses—is desirable, because risk-taking raises the option value of the business. For Guardians, the best pay system is a penalty system which punishes mistakes by significantly reducing pay in order to discourage risk-taking but which does not reward success with a pay increase. The two types of businesses also need individuals of different talents. Stars need workers who have high growth potential but may lack track records and workers with diverse backgrounds (let us call them "risky workers"), while Guardians need workers with such qualities as stability and reliability ("safe workers") and shun diversity, which is considered to hamper smooth coordination.
Table 1: Business Category and the Optimal Personnel Management System
Optimal personnel management system
Tendency to Overvalue Safe Workers
The advance of digitization has expanded the possibilities for new businesses that use information technology (IT) and artificial intelligence (AI) in a variety of industries, and this may be increasing the share of Stars in the economy as a whole. Even so, it appears that many companies have largely maintained the same hiring policies and practices and still refrain from hiring workers with diverse and uncertain profiles.
I have many years of experience teaching at Japanese and U.S. business schools, and I have realized that MBA students' attitudes toward hiring differ significantly between Japan and the United States. For example, when given the hypothetical choice of hiring Job Applicant A who has a 100% chance of generating financial value equivalent to 10 million yen (safe worker) or Job Applicant B who has an even chance of generating value equivalent to 20 million or zero yen (risky worker), around 80% of U.S. MBA students would choose Applicant B, the riskier choice. However, around 80% of Japanese students would choose Applicant A, a safer choice. What is the reason for this difference?
Let us look at Table 2 in order to systematically examine the difference. The decision-making matrix contains two right decision possibilities and two wrong decision possibilities. In the case of deciding whether or not to hire a job applicant, the applicant may be either able or unable to make positive contributions to the prospective employer's business performance, and the prospective employer cannot know for sure which of the two possibilities is right. In this case, the right decision is (1) to hire the applicant if he/she is able to make positive contributions and (2) not to hire the applicant if he/she is unable to do so. The wrong decision is (1) to hire the applicant if he/she is unable to make positive contributions (Type 1 error, or false positive error) and (2) not to hire the applicant if he/she is able to do so (Type 2 error, or false negative error).
Table 2: Two Wrong Decision Possibilities Regarding Hiring
Applicant A contributes to the company
Applicant A does not contribute to the company
Type 1 error False positive error
Type 2 error False negative error
Japanese companies prefer safe workers because they want to minimize the risk of making a false positive error. Underlying this tendency is the lifetime employment arrangement which is pervasive in Japan and which makes it difficult to fire workers once they have been employed even if the worker makes no contribution to the company. In short, the cost of making a false positive error is high. In contrast, in the United States, workers who do not contribute to the company can be fired more easily, so there is a stronger tendency to place more emphasis on applicants' potential than on qualities like stability. In this case, the cost of making a false positive error is low. However, the above is not the only reason why Japanese companies prefer stability. Unlike U.S. companies, where front-office managers typically have the authority in hiring decisions, Japanese companies give the human resources department centralized authority over hiring. When front-office managers have found hired workers to be useless, they blame the human resources department for making the wrong hiring decision. To avoid the blame, the human resources department makes conservative hiring decisions.
However, amid the growing perception that there is a shortage of capable young workers due to the low birth rate and the value of "risky workers" is increasing due to changing business characteristics, the cost of making a false negative error is steadily rising in Japan.
Multi-stage Interview System Is Aggravating the Problem
Even if companies plan to hire workers with diverse profiles or workers with higher potential for innovation, it is difficult to implement hiring according to plan in many cases. For example, a certain company used paper screening to select applicant students who were considered to have a high level of creativity, but unfortunately, none of the applicants survived the interview stage. One of the reasons why it is difficult for companies to hire workers according to plan is the use of the particular multi-stage interview system used in Japan. In order to narrow down the long list of applicants, companies use this system, under which only applicants who have successfully passed several interviews are hired. The problem is that typically, only two interviewers at most—one interviewer in many cases—handle each applicant, so personal biases of the interviewers are a significant factor.
As is widely known, there is a psychological tendency for humans to be biased or lack objectivity in evaluating other people. People tend to place high value on the skills that they themselves possess while undervaluing the skills in which they are lacking. As a result, they tend to have high regard for people who are similar to themselves. Let us assume that a company interviews an applicant with unique characteristics. In Japan, workers who stand out tend to be the source of personality conflicts that distract the workers from the main mission of the office. The multi-stage interview system is highly likely to eliminate applicants with unique profiles and favor "reliable" applicants who are viewed favorably by many interviewers but typical of the company.
To change this status quo, it is essential to re-organize the interview stage into a process whereby a team of interviewers screen applicants based on a structured interview approach with increased involvement of front-office personnel. To reduce false negative errors, it is important to conduct evaluations from as many perspectives as possible, to increase the amount of comparable information available through structured interviews, to ensure that the hiring policy is understood and followed by the whole hiring team and to decentralize the decision-making power so as to incorporate the input from the front-office operation.
Problem with the Use of Aptitude Tests to Narrow Down Pools of Applicants
Another problem is the method of using aptitude tests. As it has become possible for applicants to take aptitude tests online at low cost, many companies use these tests as a means of narrowing down the pool of applicants. Types of measures used by recruiting companies include a cognitive ability score to evaluate basic intelligence, a comprehensive score based on applicants' predicted achievement as measured by the test vendor, and a stress resilience indicator score. However, what happens when many companies use ready-made aptitude tests in order to narrow down the pool of applicants? Applicants are neatly divided into two groups—people who consistently succeed in advancing to the interview stage and people who are consistently eliminated before the interview stage. This presents a new problem—I would call "new statistical discrimination." Applicants whose predicted achievement levels are low are not allowed to move on to the interview stage.
It might still be a necessary evil if aptitude tests were effective predictors of future performance. It may not be the case, however. In the data science workshop for HR managers the author organizes, we had the participants analyze which applicants actually received job offers (including those who decline the offer) at the end of the interview process. They calculated how much of the difference in the probability of receiving a job offer between successful and unsuccessful applicants including those that failed in the initial test screening can be explained by the differences in test scores on the aptitude tests. According to the calculation, only 1-2% can be explained by the results of aptitude tests in most cases, with the highest rate at 5%. Not only is the accuracy of aptitude tests not very high in predicting which candidates will receive final job offers, but also online tests pose other risks, such as the possibility that applicants may ask other people to take the test on their behalf or that they may not answer questions truthfully.
There are other problems involved in the use of ready-made aptitude tests prepared by vendors. When the abovementioned workshop asked the participating companies to examine which applicants declined a job offer after they had passed the whole screening process, it was found that applicants who achieved high scores on aptitude tests are highly likely to decline a job offer. If aptitude tests correctly identify applicants' potential job competencies, this may be considered to be a natural consequence given that competent people receive job offers from many companies. However, on the premise that the correlation between the results of aptitude tests and the probability of receiving a job offer is not necessarily strong, as is indicated above, we may assume the following: applicants who achieve high scores on aptitude tests are invited to interviews by many companies, and as a result, even though their probability of receiving a job offer may be similar to the probability for other applicants who are invited to interviews by much fewer companies with lower test scores, they are likely to receive job offers from more companies. From the position of recruiting companies, offering interviews to only the highest scoring applicants may undermine their hiring efficiency by allowing many applicants who are likely to ultimately decline a job offer to advance to the interview stage.
The presence of these problems does not necessarily mean that companies should not use aptitude tests. In the case of large companies which have to handle a huge pool of applicants, it is inevitable to narrow down the pool somewhat through paper-screening. For companies, the important thing to do is to develop their own standards for evaluating applicants instead of only using ready-made indicators prepared by vendors. By doing that, companies can clarify the kind of skills and qualities they need from workers. In many cases, a company needs various types of leaders and high achievers. Therefore, companies should design multiple measures based on aptitude test results and other information on the application forms suited to identifying appropriate applicants and invite applicants who have scored high on some of those indicators to interviews. In that way, companies can avoid, to a certain degree, excessive competition for the same small pool of applicants.
As has been made clear above, many companies appear to be adopting flawed recruiting methods while complaining about a shortage of workers. Companies that select applicants with suitable qualities through an appropriate process and which spare no effort in communicating their own strong points will be able to meet their human resource needs. It is essential to take necessary steps, including ensuring that the hiring policy is understood and followed by the recruiting team, making selection based on a team approach and through structured questions, and developing unique measures for initial screening based on aptitude tests and other information on the application form. Companies should also make efforts to utilize AI in order to make paper screening more effective.