Preliminary draft
Career Mobility in the Embedded Market
How Workers Find Jobs in the Japanese Financial Sector
Hiroshi Ono
Texas A&M University
March 2012
ABSTRACT
This project studies how macro-level global forces are shaping the behavior of firms and of individuals in the
Japanese financial sector, with particular focus on the increasing influence of foreign firms. These firms bring with
them employment practices that are more market-driven and less socially embedded compared to the Japanese status
quo. The co-existence of the foreign and the domestic in the Japanese labor market provides a fascinating test bed to
examine how local firms adapt to global pressures, and how workers navigate the changing institutional
environment. I employ mixed methods approach using in-depth interviews and two separate datasets of finance
professionals in Japan. My analysis reveals clearly demarcated patterns of career mobility. There is considerably
higher mobility in the foreign firms compared to their domestic counterparts. Mobility between the two sectors is
frequent, but it is predominantly a one-way transition from domestic to foreign. But I also find signs that mobility in
the foreign firms is constrained by local norms and conventions. The labor market does not operate in a vacuum
but is shaped by larger social forces.
* Direct correspondences to Hiroshi Ono, Texas A&M University, Department of Sociology, College Station, TX
77843-4351, USA.
INTRODUCTION
Scholars in organizations and human resources have long studied the question of why
some workers are more mobile than others. Research in this field has been profoundly
influenced by the sociological concept of embeddedness (Block 1990; Granovetter 1985), i.e. the
idea that organizations are embedded in their institutional environments. Their boundaries are
porous, and organizations are continuously shaped by what takes place outside of their borders
(Scott 2002). Workers in turn are embedded in the organizations that they work for, as well as
by the larger social and cultural framework that they are situated in. Workers are not
autonomous and they do not act in isolation. Their actions are determined and constrained by the
institutional context, by ongoing social relations, and by the behavior and expectations of others.
Complications arise when we compare turnover rates across cultures because we
confound different levels of analysis (Friedland and Alford 1991). For example, an observation
that turnover is lower in Japan than in the U.S. raises the crucial empirical question in
organizational analysis: Is it organizational culture or national culture that is driving the
differences in turnover? Can we conclude that Japanese workers are more “loyal” than are
American workers based on cross-country comparisons?
In this paper, I examine how organizational attributes and national culture affect career
mobility. I choose Japan as a test bed for empirical analysis. The Japanese labor market is de
facto characterized by low turnover and this makes it easier to observe deviations from this norm.
In this study, organizations are separated into two distinct types – the foreign and the domestic –
according to the foreign ownership structure of the firm.
My work is driven by the core hypothesis that loyalty is endogenous. I argue that the
high-commitment culture commonly associated with the Japanese workforce is not a unique
Japanese quality as suggested by some cultural scholars (e.g. Abegglen 1958), but an outcome of
the normative environment of the Japanese firm. I employ a unique research design which
clearly delineates the levels of embeddedness (Dobbin 1994) in organizational analysis. The
study contributes to our understanding of how organizational processes affect individual career
outcomes.
BACKGROUND
In Japan, the labor market functions under norms and institutions that value loyalty and
trust. There is an implicit understanding between workers and employers that their relationship
will be long-lasting, and that benefits over the long-run will override short-term interests (Ono &
Rebick 2003). Workers who breach this implicit contract and switch jobs are viewed as “social
defects,” and are penalized for their actions through lower pay.
Along with globalization, a new labor market has emerged in the country alongside its
Japanese counterpart. This is the market for foreign-owned firms. In recent years, we are
witnessing an increasing pattern of highly-skilled professionals who seek work in the foreign
firms in Japan.
The labor market for foreign firms represents the extreme opposite of the Japanese status
quo. It is deeply rooted in an Anglo-American home base, and operates more closely along
market principles dictated by quantity and price signals. There is more fluidity and less loyalty.
Workers switch jobs in response to higher pay. Employers dismiss workers when they are not
needed or when they underperform. In short, workers act like free agents, and firms prioritize
short-term profits in place of long-term relations.
In a country that features one of the lowest turnover rates in the world, the workers in the
foreign-owned firms stand out for their seemingly unrestrained pattern of interfirm mobility.
Some workers in the more traditional Japanese settings envy the greater independence and higher
pay, and yet others feel disdain for their money-motivated and short-sighted values.
Who are these free agents? What motivates them to move? What are the consequences
for defecting from the status quo? How do foreign firms attract workers in a labor market where
loyalty and long-term employment remain the norm? Comparing domestic versus foreign
highlights underlying differences between the market versus culture, formal versus informal, and
economic versus social.
MARKET OVERVIEW
The current study focuses on the finance sector in Japan. The Tokyo Stock Exchange is
the third largest exchange in the world by market value. Japan’s financial sector is a truly global
operation. The leading brokerage firms from around the world are represented there. Among the
workers in foreign firms, 68 percent are employed by the U.S. brokerage firms (JETRO 2004).
We therefore designate the U.S. firm as a close approximation and benchmark for the foreign
firm hereafter.
2
There are important substantive reasons for examining the finance sector. The financial
market epitomizes the advanced capitalist system. Financial products are exchanged, often
amongst anonymous buyers and sellers, over a market interface where transactions are conducted
instantaneously based primarily on prices and quantity. As Podolny (2005) explains: “Few if
any sectors contain markets that better conform to assumptions associated with the economic
model of perfect competition – specifically, the profit maximization of firms, the utility
maximization of individuals, and perfect information about the opportunities for exchange.”
Based on the market-dominated view of the financial sector, it is tempting to extrapolate
that the labor market for finance professionals also resembles a market with perfect information
and anonymous short-term trade. However, in the case of Japan, the labor market in finance is
actually well-sheltered from the pure market forces. The internal labor market is well-developed
in the finance sector, and accordingly, there is lower turnover than the private sector average. In
2006, over 60 percent of workers in finance were employed in large firms (greater than 1000
employees), and 43 percent were estimated to be covered by lifetime employment (Ministry of
Health, Labor and Welfare statistics); these numbers were18 percent and 21 percent, respectively,
for all industries in the private sector.1
The market for foreign firms in Japan
The human resource (HR) practices differ strikingly between domestic and foreign firms.
First, domestic firms conform to the internal labor market setup by recruiting larger numbers of
new graduate hires relative to mid-career hires.2 The new graduates enter the organization at the
bottom, and rise through the ranks after receiving extensive training from the firm. There is an
opposite pattern in the foreign firm, with greater numbers of mid-career hires compared to new
graduates. In fact, some foreign firms rely exclusively on midcareer recruiting of managers and
specialists (Robinson 2003). This hiring pattern suggests that the port of entry in the foreign
firms is open at all levels of the organization.
If the modus operandi of Japanese corporations is implicit and long-term, the counterpart
for the foreign firms is explicit and short-term. In place of a reward structure where wages are
determined by seniority and service to the firm, compensation in the foreign firm is tied more
closely to the workers’ performance. Foreign firms are less likely to support lifetime
employment, and more likely to use merit pay instead of seniority pay (JILPT 2005). They also
3
are more likely to use the annual salary system, where terms and conditions of pay are negotiated
on an annual basis. Wages are thus determined more by the worker’s marginal product, and less
by non-market attributes such as loyalty and commitment.
THE SETTING
We begin our theoretical groundwork by laying out the relationship between marketness
and employment duration. Following Block (1990), we define marketness as the extent to which
prices dictate economic transactions. Movement toward low marketness implies that nonprice
considerations take on greater importance in the transactions. In exchange theory (Blau 1964),
points D and F in Figure 1 indicate the ideal types representing the opposite positions of social
versus economic exchange. Economic exchange assumes that individuals are motivated
primarily by prices and quantity. Every transaction has a price, and individuals maximize utility
through cost-benefit calculations. Social influences are seen as something that disturbs
economic action (Block 1990). Under economic exchange, terms and conditions of employment
are documented explicitly. Job functions are clearly defined. Under high marketness, there is
little room for ambiguities. Workers who compete under high marketness know their market
price.
FIGURE 1 HERE
Social exchange is the status quo of the Japanese employment relationship (Murakami
and Rohlen 1992; Ono 2007). Under social exchange, workers may be motivated by prices, but
their actions are likewise influenced, and in many cases constrained by their more intimate
relationships with their employers. Lifetime employment – one of the main pillars of the
Japanese employment system – is not an explicit contract, but an implicit agreement formed
between workers and employers with the expectation that the relationship will be long-lasting
(Moriguchi and Ono 2006). Lifetime employment is thus a prime example of social exchange,
and occupies point D in Figure 1.
The intimacy of the employment relationship presumes closure, where favors are
exchanged on a daily basis. This closure fosters trust. Employers are then able to invest in the
workers, and the workers reciprocate by committing their utmost loyalty. The closure further
4
enables workers to build human capital and social capital within the firm. Much of this capital is
specific to the firm, and thus will be lost once the relationship is breached; it cannot be easily
codified or transmitted, and it cannot be quantified or priced. Workers who are employed in
these firms are not fully aware of their market worth.
Institutional complementarities
The Japanese employment system is best understood as a system of interdependent and
complementary institutions (Aoki 2001). The crucial parts that comprise the Japanese
employment system include extensive training, seniority based wages, internal promotion, job
rotation, and job security, among others. These parts are complementary to one another to the
extent that “the marginal returns from using one practice increase with the usage of another
practice” (Moriguchi and Ono 2006: 154). The cluster of these interdependent HR practices
constitutes a self-enforcing equilibrium, an outcome of which is a low-turnover model of
employment, characterized by high trust between workers and employers. If any of the parts is
removed or altered, the equilibrium is disturbed and the model becomes unstable.
The cluster analogy does not apply to the foreign firms, as there are few institutional
complementarities. HR practices are less interdependent, and there is no kernel that holds them
together, thus resulting in a model of high turnover and low trust. In this system, loyalty does
not pay, and is in fact, irrational.
In Figure 1, the domestic firm is positioned at D, and the foreign firms at F. There are
several reasons why foreign firms occupy this position. First, the majority of workers in foreign
firms are employed by U.S. firms, as previously discussed. Since the U.S. has one of the highest
labor turnover in the world, and Japan the lowest (Ono 2010), the contrast between the domestic
versus the foreign constitutes the extremes. The second reason, which is related to the first,
concerns the distinction in corporate governance. Japanese management is characterized by the
stakeholder model, which prioritizes long-term gains over short-term profits, and looks after the
well-being of their workers and other stakeholders (Jacoby 2005). Anglo-American firms, on the
other hand, uphold the shareholder model of governance, and operate under different conceptions
of legitimacy. Their management style is driven more by maximizing shareholder value, and
does not uphold the long-term approach that is common in Japanese management. With respect
to lifetime employment, Ahmadjian and Robinson (2001) explain that the practice may be
5
deemed legitimate among Japanese investors, but it is less so among U.S. and European
investors who demand immediate attention to shareholder value. This view is consistent with the
survey results which show that foreign firms are less likely to support lifetime employment than
are domestic firms (as discussed earlier).
Third, by default, foreign firms that set up their business operations overseas face a
latecomer disadvantage arising from differences in local customs, language, and institutions
(Fukao and Ito 2003). These firms do not have the luxury of time. Overcoming the latecomer
disadvantage places these firms in high-pressure situations where they are expected to produce
immediate results, so that they can catch up to the local competition. These three factors
combined push the foreign firms away from D and toward F in Figure 1.
Organizations and their environments
A key insight from institutional analysis is that what takes place within organizations is
affected by what takes place outside of it. For example, Scott et al (1994) explains:
(T)he visible structures and routines that make up organizations are direct reflections and effects of rules and
structures built into (or institutionalized within) wider environments. Organizations reflect patterns or templates
established in a wider system. (p.2)
Organizations are “open systems” (Scott 2002) which continuously interact and respond to
outside forces. Organizations and their behaviors are best understood by considering the specific
institutional environment that they are situated in.
Figure 2 illustrates the relationship between organizations and their environments
intervened by national culture. Here the vertical dimension shows the marketness of the
organization, and the horizontal shows the marketness of the environment. Firm L represents the
U.S. firm in the U.S. which is a high marketness firm in a high marketness environment. On the
opposite end is firm M which is the status quo in Japan – a low-marketness firm in a low
marketness environment. Indeed, Jacoby (2005) coined the expression “the embedded
corporation” to describe the socially embedded nature of the Japanese firm. In the current
typology, the “embedded corporation” is embedded in a two dimensional schema.
Now consider the concrete example of an American investment bank such as Goldman
Sachs that enters the Japanese financial sector. This represents firm K, which is a high
6
marketness firm operating in a low-marketness environment. Their competitive environment
consists of the key Japanese (or domestic) brokerage firms such as Nomura Securities and Daiwa
Securities. Although these domestic firms may not have a full-blown lifetime employment
system in place, they still maintain an internal labor market setup with extensive job training and
considerable job security. Indeed, their organizational blueprint (Baron et al 2001) stands at
opposite ends in comparison to the U.S. brokerage firms. The Japanese financial sector thus
showcases a population of two vastly contrasting organizational forms – the highly market
driven foreign firm and the highly embedded domestic firm. It is in essence a market where the
economic meets the social.
A strategic decision for the American firm is whether to import their HR strategies with
them, or to adopt an organizational blueprint that better conforms to local labor market norms
and institutions. In other words, does firm K converge in the direction of firm M, or do they
maintain their core identity as firm L? This is a crucial decision that affects all HR operations,
including recruiting, training, compensation, retaining, and dismissal. Likewise, as the foreign
firms increase their presence in the local labor market, does firm M maintain their status quo or
do they converge towards firm K to remain competitive?
MODELS OF JOB TURNOVER
Internal labor market and human capital formation
Job turnover in Japan is remarkably low compared to the other industrialized economies
(Ono 2005). Much of this is due to the strong internal labor market setup that emphasizes job
training, job security, internal promotion, and fringe benefits (Doeringer and Piore 1971). From
the viewpoint of human capital theory, extensive on-the-job training is associated with low job
mobility, because workers accumulate skills that are specific to the firm. In fact, two of the top
reasons that workers do not change jobs in Japan are because they fear that their skills cannot be
transferred across firms, and because they will lose their returns to seniority (Ono and Rebick
2003). The longer a worker remains with the firm, the more embedded her skills become, and
the more costly it becomes to change firms. Maintaining mobility is synonymous with
avoiding specificity (Hirsch 1987).
If workers do not receive training, they do not acquire firm-specific skills. They are free
agents in the sense that they do not have any attachment to the firm beyond contractual
7
provisions. They have not invested in the firm, and the firm has not invested in them. Job
mobility is costless; there is no wage loss if they relocate to another firm. For free agents, there
is no time dimension. Their level of attachment is independent of the time they spend with the
firm. Free agents are better endowed with general human capital, or skills that are portable
across firm boundaries. Because it is easier to put a price on general skills than it is on firmspecific skills, free agents have a better sense of their market worth.
Features of the internal labor market depart from the spot-market conditions, and are
closer to the position of low marketness as depicted in Figure 1. The spot-market presupposes an
organizational setup where the ports of entry are open at all levels of the hierarchy. Workers can
be replaced by other workers as long as the incoming offer a more competitive price than the
incumbents. The internal labor market is less open because new recruits can only enter at the
bottom. By creating a market within a firm, the internal labor market shelters workers from the
brunt of business cycle fluctuations and the market volatility that accompanies it. Employers use
internal promotion to induce worker commitment, which in turn lowers overall turnover.
The movement from D to F implies a shift from firm-specific to general skills, and from
high-attachment to low-attachment. Workers that move from domestic to foreign firms
relinquish their job security, because they have left the internal labor market setup, and are no
longer insulated from external market forces. But they receive higher pay in return. This is
intuitively straightforward. The movement towards low job security involves higher risk for the
worker. Thus, difference in wages is essentially a risk premium. Alternatively, from the
perspective of workers in domestic firms, they pay an “insurance” to shield them against the risk
of low employment security. Put another way, it is inconceivable, if not irrational, that a worker
would voluntarily move to a position that carries with it higher risk but lower pay. The foreign
firm wage premium has been documented extensively in empirical studies (Ahmadjian and
Robinson 2001; Fukao and Ito 2003; Ono 2007; Ono and Odaki 2011), and we take this as given.
8
Social capital and job matching
The analysis of how workers are matched to jobs is one of the key contributions of
sociologists, and an area much neglected by the economic approach. Under the neoclassical
framework, employers have perfect information about who is looking for jobs, and how
productive they are; workers in turn know which firms are looking to fill job vacancies, and they
know their exact market worth. Job matching is costless because workers and employers are
matched to each other instantaneously in a perfectly operating market. In reality, there are
considerable transaction costs to job matching. Employers must invest time and resources
searching for the right candidate through a number of channels, e.g. posting advertisements,
hiring headhunters, spreading information by word of mouth, etc.
Social capital theory asserts that job matching can be facilitated by taking advantage of
the resources that are embedded in social relations. The quantity, quality and value of the social
capital vary by an individual’s location in a social network (Lin 2000). For free agents,
attracting outside options requires maintaining high visibility and cultivating social networks
(Hirsch 1987). To this end, free agents maintain extensive professional networks both within
and outside of the firm. These resources can be used strategically in securing the next job, and in
advancing their careers in general.
While social capital may facilitate career changes for some, it may inhibit mobility for
others. Like the concept of firm-specific human capital, a worker may accumulate social capital
that is specific to the firm, which would be lost if she changed jobs. Fear of losing personal
contacts is the other top reason that workers do not change jobs in Japan (Ono and Rebick 2003).
DATA AND METHODS
I use multiple methods and data in my empirical analysis. I conducted exploratory
interviews with finance professionals in Tokyo to examine preliminary hypotheses and to
generate new ones. These interviews provided an informative starting point from which to
discern the social dynamics of the labor market for finance professionals in Japan, and the key
determinants of their mobility (or immobility). During the time period from 2007 to 2009, I was
able to conduct a total of 28 interviews with finance professionals in Tokyo. All interviews were
conducted in Japanese. Respondents were selected mainly on the basis of snowball sampling.
9
The sample consists of individuals who switched from domestic to foreign firms, as well as those
who have never changed employers.
I conducted statistical analysis from two data sources. The first is individual-level data
from the “Survey of finance professionals” collected by the Nippon Life Insurance Research
Institute in June 1999 (hereafter NLI survey).3 The final sample size of 340 finance
professionals was selected from questionnaires that were distributed to 252 finance firms listed
in the Japan Company Handbook (Kaisha Shikiho, published by Toyo Keizai). The survey
collected a wide range of information relating to the respondent’s employment – e.g. description
of current work, career history, education and training – in addition to the basic demographic
information. 33 percent of the respondents are currently employed in foreign firms.
The second dataset, which I compiled and constructed, is an individual-level database
consisting of security analysts and their career histories (hereafter security analysts database).
Data were collected from analysts featured in the Nikkei Financial Daily. I employ a procedure
similar to Groysberg et al (2008) who collected demographic and job history data of security
analysts featured in Institutional Investor. I briefly describe the data collection process below.
The Nikkei Financial Daily is a Japanese daily newspaper specialized in reporting
financial news. Every week, the newspaper features a column by a security analyst, who
diagnoses a particular company’s financial standing and performance. The column also features
a brief profile of the analyst. A typical profile looks like the following:
Analyst X, Morgan Stanley
Male. Born 1973. Joined Daiwa Securities after graduating with B.A. in Economics from Hokkaido University in
1996. Employed at Morgan Stanley since 2004.
Using a customized template, I then coded this information for a total of 322 security analysts
from the Nikkei Financial Daily between January 2000 and December 2007. 58 percent of the
sample is currently employed in foreign firms. Provided that these analysts were selected to be
featured by the Nikkei Financial Daily, it does call into question the extent to which the current
sample represents the general population of security analysts. We do not have information about
the selection criteria, but presumably these analysts represent the higher end of the ability
distribution. We would therefore not be able to generalize our results beyond this pool of
10
relatively talented analysts. We proceed with care and caution in interpreting the results of our
empirical investigation.
ANALYSIS AND RESULTS
Worker profiles: Who works in the foreign firms?
We begin by profiling finance professionals employed by foreign firms. Table 1 shows
probit coefficients describing the determinants of being employed in foreign firms. Column (a)
shows the results using the NLI data. The basic variables included in this regression are sex,
education, age, and tenure. In addition, we include: (i) foreign university dummy variable
indicating whether the respondent graduated from a university outside of Japan; and (ii) English
certificate dummy variable indicating whether the respondent holds an English proficiency
certificate approved by the Japanese ministry. I also conducted a similar estimation for the
sample of security analysts to predict their entry into the foreign firm. These results are shown in
column (b).
And finally, to distinguish generalists from specialists, we include the variable, years of
specialization (in a specific area of finance) and its quadratic. Here, the areas of specialization
may include for example, financial product development, personal loans, international finance,
mergers and acquisitions, risk management, etc. Workers with longer years of specialization
correspond to specialists, and workers with shorter years of specialization correspond to
“generalists” (not to be confused with general human capital).
It is important to note that years of specialization is a separate measure from tenure. A
specialist may be highly specialized in a particular area of finance for a number of years, but she
may have changed firms in the interim, in which case her years of specialization would be
greater than her tenure. On the other hand, a generalist may have experienced a number of
different sections within the same firm, e.g. through job rotation, in which case her tenure would
be longer than her years of specialization.
The results show some interesting patterns (Table 1). In sum, the profile of finance
professionals in the foreign firms deviates from the status quo in several ways. The average
worker in Japanese firms fits all of the stereotypes of the so-called salaryman figure. He is male,
a graduate from a Japanese university, a generalist, and a job stayer. In contrast, the foreign firm
11
counterpart has a higher likelihood of being a female, a graduate from a foreign university, a
specialist, a job mover, and an English speaker.
TABLE 1 HERE
The issue of selection
A fundamental problem in the empirical analysis concerns the selection of workers into
foreign firms. Much of what we have discussed so far suggests that worker selection into foreign
firms hinges on some systematic process, thus violating the assumption of random selection. I
use the Heckman method (1979) to address the selection problem. We use the probit coefficients
obtained from Table 1 to construct the inverse Mills ratio (λ) which is then entered into the main
equations of interest to correct for selection bias. We use the variable, graduating from a
university outside of Japan, for the exclusion criteria. This variable is a significant predictor of
employment in foreign firm, but it has no effect on the outcomes of interest in our analyses.
Why do foreign firms have higher turnover?
Table 2a shows summary statistics for job changes. In the security analysts data, I also
provide categorical dummies to differentiate the types of domestic and foreign firms, according
to their ranking of security analysts as listed in the publication Institutional Investor for various
years. Top domestic firms include Nomura and Daiwa. Top foreign firms include Deutsche
Bank, Goldman Sachs, JP Morgan, Merrill Lynch, Morgan Stanley, Nikko Salomon Smith
Barney, and UBS Warburg.4
In both samples, workers in foreign firms have experienced a greater number of job
moves. On average, the number of job changes in foreign firms (versus domestic firms) is
greater by more than one in both datasets. In the security analysts data, we find that there is clear
differentiation among domestic firms, and among foreign firms. Specifically, the firms in
ascending order of job changes are: Top domestic < Other domestic < Top foreign < Other
foreign. The median number of job changes also follows this rank order.
Job mobility may be more common and tolerated in the foreign firms, but these workers
are still constrained by the broader institutional context and the normative environment of the
Japanese labor market. Too much mobility may signal distrust and defection even among the
12
foreign firms. The aversion towards frequent job changers still holds especially among the highstatus foreign firms. One industry insider explained:
At the end of the day, Goldman Sachs in Japan is Goldman Sachs Japan Company Limited. It operates like a
Japanese company, much more so than it does in New York or in London. So there is obviously a lot of influence
from Japanese culture.
Further, since the market for foreign firms is still relatively small, their HR departments
are tightly interconnected. Information about particular candidates travels across firms. Finance
professionals that have reputations for frequent job hopping are identified and given certain
labels, e.g. “gypsies” (Ikaros 2000), and then avoided by some potential employers.
Under the neo-classical framework, mobility is frictionless, and workers are free to move
in and out of firms as they please. But it says nothing about how the high mobility may affect
workers’ reputation in the labor market. In reality, reputations can travel across firm boundaries,
and a reputation for high mobility can impair an individual’s career trajectory.
At the top domestic firms, the median number of job changes is zero, i.e. the majority are
inbred and have never changed jobs. The percentage that has never changed jobs is in fact found
to be 87 percent (not shown here). Clearly there is no culture of high turnover in the top
domestic firms.
Table 2b shows Poisson regressions predicting the frequency of job changes as a function
of firm- and individual-level characteristics. In the Poisson model, the outcome μ is a positive
count. μ has a Poisson distribution with a conditional mean that depends on the set of covariates
x. The basic form of the model is:
μi = exp (xiβ)
(1)
In estimating the frequency of job changes, we account for the fact that different respondents
have different exposure times (t). This is done by multiplying both sides of equation (1) by t
(Long and Freese 2006). Since t = exp (ln t), equation (1) becomes:
μiti = exp (xiβ + ln ti)
(2)
13
In my estimations, we use the respondents’ age as the measure of exposure time.
The results show clearly that workers in foreign firms have higher turnover than do
workers in domestic firms. In the NLI data (column [a]), workers are predicted to have a
turnover rate that is 2.4 times greater (= e0.877) than their domestic counterparts. In Model (b),
we find that analysts who are currently employed in foreign firms have a significantly higher rate
of turnover than do analysts in domestic firms. More interestingly, we find that analysts who
started out in the top domestic firms (as their first employer) are significantly less likely to have
changed jobs, in comparison to analysts who started in other firms. This finding thus suggests
that the point of origin determines the frequency of mobility. Top domestic firms generally have
the most well-established internal labor market setup. They provide the most extensive training,
and pay the highest salaries (among the domestic firms). The cost of defecting – be it economic
or social – can be considerable.
In Model (c), we break down the current employer into four categories, designating the
top domestic firm as the reference (omitted) category. We confirm the same rank order of job
changes, with analysts in top domestic firms predicted to be the least frequent and analysts in
other foreign firms predicted to be the most frequent job changers.
TABLE 2 HERE
We have established that there is greater turnover in the foreign firm. We now
investigate why. Table 3 shows logit coefficients predicting differences in employment
characteristics between domestic and foreign firms using the NLI sample. Each row reports the
results of separate logistic regressions. In the interest of space, I present only the coefficient for
the foreign firm dummy. All regressions control for the vector X and the Heckman selection
term (λ), but these are suppressed from the output.
TABLE 3 HERE
Human capital development
One of the reasons that foreign firms provide less training is because they expect their
workers to be productive immediately. This leads to a key difference in the human capital
14
investment strategy. While workers in domestic firms invest in the firm and build firm-specific
skills, workers in foreign firms invest in themselves, to further specialize in their (general)
professional skills. I examine survey and interview results in light of these key differences.
First, workers in foreign firms have more years of specialization, and less tenure than do
their domestic counterparts (Table 1). On average, the mean years of specialization is greater
than tenure in foreign firms, but the reverse is true in domestic firms. Workers in foreign firms
may change jobs more frequently, but they are less likely to change their areas of specialization.
Second, workers in foreign firms are less likely to undergo job rotation as shown in Table
3. Job rotation is a common HR practice in the Japanese organization (Aoki 2001; Robinson
2003). Workers advance through the internal labor market in an upward spiral, by experiencing
a wide range of assignments in different sections. They may, for example, start out in
investment banking, then move to sales, then to HR or accounting. In so doing, they become
generalists because they develop knowledge and establish personnel networks across a variety of
sections within the firm. Job rotation is less common in the foreign firm. The emphasis is
instead for workers to become specialists, and to develop a deeper sense of professionalism in
their respective areas.
Third, if there is high turnover in the foreign firm, then the worker is also likely to be
working under a supervisor who came from another firm. This is an important but neglected
factor that is related to lower worker commitment. In Table 3, the coefficient for “I work for an
inbred boss” is negative. Here, an inbred boss refers to an internally promoted boss who has
never left the firm. The results thus imply that workers in foreign firms are more likely to be
working under a boss who has previous work experience with another employer.
Interview respondents were quick to point out that their bosses were highly influential in
building their careers. This is partly rooted in the fact that Japan is a highly hierarchical society
to begin with (Nakane 1970), so there is greater dependence between workers and their bosses.
Since the boss usually sets the example, and the subordinate follows, the culture of mobility can
be easily transmitted to the subordinates. One analyst explained: “It’s hard to pledge loyalty to
your employer when your boss himself is not loyal.” The tighter linkage with their bosses also
explains why many bosses invite their team members to change jobs with them. We discuss
these examples in the next section.
15
Fourth, workers in foreign firms are less likely to receive training from the firm. Table 3
shows the responses to the question: Which of the following has helped you improve your
expertise? Workers in foreign firms are less likely to respond that they learned from their bosses,
and more likely to respond that they achieved this through self-improvement. Table 3 also
shows that workers in foreign firms seek better education and training from their firms. These
responses thus reveal the relatively independent work environment of the foreign firm, where
workers are themselves responsible for investments in their human capital. The conditions differ
greatly from the Japanese employment setup characterized by substantial training, and hands-on
coaching from their supervisors.
In the finance profession, it is common market knowledge that domestic firms provide
more training than do foreign firms. Many finance professionals actually take advantage of this
notable distinction. Unsurprisingly, domestic firms become “feeders” for the foreign firms. One
investment banker gave the following advice for young professionals considering a career in the
foreign firm:
It is unwise for graduates to choose a foreign firm as their first employer. They should start in a domestic firm first,
get as much training as they can, then consider the next move.
Nomura in particular is well-known for their extensive training. Many finance
professionals start their careers at Nomura, then migrate to the foreign firms. This pattern has
become so well-established in the profession that the group is known as the Nomura alumni, and
they are highly sought after in the finance profession.5 The Nikkei Financial Daily provides an
annual ranking of the top security analysts. In most of the years surveyed (between 1994 and
2002), Nomura analysts and their alumni claimed more than 50 percent of the top positions
(according to author’s own estimations). The majority of the top ranking analysts are therefore
trained at Nomura.
In sum, the survey and interview results highlight key differences in human capital
development between domestic and foreign firms. Barriers to mobility in the foreign firms are
lower. The market is dictated by the supply and demand for general human capital. Lower
demand for firm-specific skills lowers their barriers to entry and lowers their attachment in the
foreign firms, which further enhances their mobility.
16
Autonomy, independence and self-worth
The next battery of questions concerns the link between performance and pay. Workers
in foreign firms are more likely to respond that: (i) their performance is directly linked to
compensation and promotion; (ii) the firm recognizes the importance of professional skills; and
(iii) they are highly evaluated for their expertise. These results all indicate a tighter linkage
between productivity and pay. Conversely, it can be inferred that the link between performance
and pay is more ambiguous in the domestic firms. This ambiguity stems partly from the
difficulty of placing a value on firm-specific skills, and the tendency to downplay individual
achievement in place of group performance.
A consistent pattern among the professionals that I interviewed was that they demanded
clear recognition for their achievement and performance. Many moved to foreign firms because
they were frustrated with the lack of recognition for individual achievement with their previous
(domestic) employers.
For these reasons, some contend that working in domestic firms is like working under
socialism (AERA 2005). The firm can exert excessive pressure on their workers to conform to
the prevailing norm towards consensus, harmony, and group identity. Those that do not conform
to these conventional views naturally lose their attachment over time. Many migrate to foreign
firms to experience market capitalism in its most naked form (Suenaga 1999). One trader who
moved from Mitsubishi to Goldman Sachs drew a direct analogy to the growing number of
Japanese professional baseball players who are now playing in the Major League in the U.S.
When I was with Mitsubishi, I was reasonably happy, but I always wanted to be in the big leagues. To me,
Mitsubishi is like the Yomiuri (Tokyo) Giants, and Goldman Sachs is like the New York Yankees. Who would turn
down an offer to play in the Major League?
If workers in foreign firms value independence and autonomy, then we would expect that
they are less likely to work in teams. The results, shown in the last row of Table 3, indicate that
teamwork is equally valued in the foreign firm as it is in the domestic firm. I shall follow up on
this in the next section.
17
The perception of time
Another consistent opinion in the interviews concerns the different perceptions of time.
Domestic firms are widely recognized for their long-term vision. Domestic firms provide
extensive training, with the expectation that the employment relationship will be long-lasting. In
the gift exchange rhetoric (Akerlof 1982), firms invest in their workers, with the expectation that
this gift will be repaid later on in their careers. But some workers may not appreciate the gift,
because it pushes back the timing of working and being rewarded in accordance to their marginal
product.6 Further, given the higher frequency of job rotation, workers may have to endure time
served in various sections of the firm until they are able to work in the job that they truly desire.
Clearly, the workers in the foreign firms have less employment security than do their
domestic counterparts, but this is of course an outcome of their rational decisions. As Baron et al
(2001) explain, “Employees who consider joining a firm that has a history of espousing a
particular model are likely to have a good sense of what they will encounter” (p.973).
Interview respondents who were employed by foreign firms repeatedly remarked that
they did not have the patience to deal with the long-term orientation of the Japanese firm, and
that they wanted to put their skills to use immediately. One respondent remarked that working in
Japanese firms is like working under hypnosis. Another respondent made the analogy to the
rabbit and the tortoise, to highlight the relatively slower pace of work in the Japanese firm.
Workers do not have the sense of urgency because they are in a long-term employment
relationship, and they know that they cannot be dismissed. Baron (1988)’s imagery of Newton’s
first law of motion may be appropriate here: “employees remain in a state of rest unless
compelled to change that state by a stronger force impressed upon them” (p.494).
In sum, the cluster of HR practices in place within the domestic firms is held intact
through institutional complementarities. Extensive training and job rotation within the firms
leads to the accumulation of firm-specific skills which results in an implicit, long-term
employment. Because it is difficult to place a value on firm-specific skills, the link between
individual performance and pay is ambiguous. These conditions do not hold true in the foreign
firms, because the HR practices there are less interdependent. Since they are better endowed in
human capital that is marketable elsewhere, they are fully aware of their market worth.
Accordingly, there is a clear link between their productivity and pay.
18
Analysis of movers
We now examine career mobility among the movers in the NLI sample. Here, movers
are defined as those who have previous work history with a different employer. Of the full NLI
sample of 340, there were 101 movers. The current analysis is restricted to this subsample.
TABLE 4 HERE
Table 4 shows responses to a number of questions from the NLI survey regarding career
changes. The outcome variable here is a yes/no binary response. I ran logistic regressions to
predict the “yes” outcome for each survey question. Table 4 shows only the coefficient for the
foreign firm dummy. All regressions control for the vector X and the Heckman selection term
(λ), but these are suppressed from the output.
We first confirm that workers who are currently employed in foreign firms are more
likely to have moved there from a foreign firm. We also confirm that workers who moved to
foreign firms resulted in higher salary.
We now examine the reasons for moving. Recognition for individual performance is
highly valued among workers in foreign firms. This finding is largely consistent with the earlier
claim, that workers migrate to foreign firms in pursuit of greater individual recognition.
Higher pay is a less important motive for moving to foreign firms. In fact, the response
rate is statistically insignificant between workers in domestic and foreign firms. This finding
does not necessarily contradict the one above. The two findings suggest that workers on average
end up with higher salary as an outcome of moving to foreign firms, but higher pay is not
necessarily the main motive for their move. The trader from Goldman Sachs explained that his
main motivation for moving was greater recognition, and his desire to test his skills in the big
leagues.
Goldman Sachs offered a higher compensation package than the one at Mitsubishi. But this was not the reason for
my move. I moved because they were a foreign firm. If Nomura had offered a higher salary, for example, I would
not have moved. (emphasis mine)
19
Job matching
We next examine how workers are matched to employers. We are particularly interested
in examining here the role of information in searching for jobs. Table 4 results do not show a
definitive pattern in the job search method. Workers in foreign firms were less likely to use
newspaper ads, suggesting that they found employers through other informal means. But they
were more likely to find jobs via headhunters. The other results – via acquaintance and via direct
invitation – are both insignificant.
Headhunters play a big role in job search for foreign firms (Ikaros 2002; Robinson 2003),
a point which was mentioned frequently in the interviews. Some respondents were approached
by headhunters, while others approached headhunters themselves to find their next jobs. And
still others made it a point to meet periodically with headhunters to learn about their market
worth, and also to update the headhunters with their latest achievements. Headhunters have
enormous databases of information, but that information can get old, according to one
investment banker in my interviews. He felt that it was in his interest to update his information
with the headhunters.
I learned during my interviews that there exist headhunters that specialize exclusively in
placement into foreign firms. This market niche suggests that the demand for midcareer
placement is considerably larger among the foreign firms. The internal labor market structure of
domestic firms is not accommodating to midcareer professionals. Domestic firms, in turn, are
not looking to hire midcareer professionals.
Team mobility and social capital building
The one recruitment channel that is not mentioned in the survey, but was raised
consistently in the interviews is the role of the former boss. Without doubt, bosses are
instrumental players in finding jobs and more generally, in advancing their careers. Because
finance professionals often work in teams, it is common practice for them to migrate in teams.
These teams then build up team-specific capital (Groysberg et al 2008) – be it human capital or
social capital – which loses value if broken up into smaller parts. The importance of maintaining
a good reputation with former bosses was emphasized repeatedly in my interviews.
One famous story in the finance sector is the case of the mergers and acquisition (M&A)
team at Merrill Lynch. The section head of the M&A team spun off and launched his own
20
company specializing in the M&A business. When he quit, he took his entire team with him,
thus leaving the M&A section at Merrill Lynch vacant and scrambling to fill up these vacancies.
Social capital can facilitate mobility in some situations, and constrain it in others. A
veteran researcher in a large domestic brokerage firm explained how he built his “empire” there.
Over time, I drew upon my resources and built my own “empire.” And once you build your empire, you are locked
in. If I were to move, I think about how many phone calls I have to make internally. I really cannot put a value on
this network I have built in-house. The headhunters used to call me frequently when I was younger, but they
stopped calling because they realized that I cannot be bought.
This is a notable example of firm-specific social capital. Clearly, his reason for not moving is
social and not economic; he does not move because the social costs outweigh the benefits.
In sum, the findings from the analysis of movers underscore the important role of
information in the job matching process. Workers in foreign firms are sufficiently informed
about potential job opportunities, and they are well informed about their own market worth.
These workers maintain close ties with their supervisors, and communicate regularly with
headhunters. These investments in information and social capital improve the quality of the job
match, and facilitate their career mobility.
Career mobility among security analysts
In our final analysis, we examine career mobility among security analysts. We are
primarily interested here in examining the extent to which mobility is barrier free. Empirically,
this corresponds to the condition of random mobility; there is no dependence between the origin
and destination.
In Table 5, column (a) records the number of job moves, which ranges from zero for nonmovers up to five for those who experienced six employers. Column (b) records the number of
employers, ranging from one to six. Column (c) records the number of persons corresponding to
each category, and column (d) records the total number of moves, which is the number of moves
multiplied by the number of persons.
The most frequently observed pattern is the group of non-movers in the domestic firm.
Among the total sample of 322 analysts, 91 (or 28 percent) fall in this category. In contrast, only
9 analysts (or 3 percent) are in the category of non-movers in the foreign firm. Only 40 analysts
(or 12 percent) started out in foreign firms. This is consistent with earlier discussion, that foreign
21
firms are generally the destination and not the origin of careers. We begin to see a pattern where
the domestic firms are positioned as “feeders” for the foreign firms. As suggested by interviews,
these analysts may be taking advantage of the extensive training offered by domestic firms, then
moving on to the foreign firms where they can be rewarded with higher compensation or greater
personal recognition.
I highlight the transition from foreign (F) to domestic firms (D) using boxes in Table 5.
The results again clearly show the un-likelihood of this move. Only 20 analysts moved from
foreign to domestic firms. This is just 6 percent of the total number of analysts, or 5 percent of
the total number of possible moves shown here. With respect to first jobs, 88 percent started off
in domestic firms. With respect to current jobs, this ratio firms drops to 42 percent.
TABLE 5 HERE
Table 6 presents the mobility table that illustrates how security analysts move from their
origins to their destinations. The left-hand column shows the origins, i.e. their first employers.
The remaining columns are the destinations of their first job move. The firms listed in Table 6
are the top performing financial institutions that employ the highest ranking security analysts, as
tabulated and ranked by Institutional Investor, and other publications in finance.
TABLE 6 HERE
The numbers in the cells indicate job-moves. If a person moves from Nomura to
Goldman Sachs, then this is recorded as a job move. If the person does not move at all, then this
is recorded in the diagonal. In this way, I recorded all analysts and their career mobility patterns.
The results are striking. There is a clear association between the origin and destination, thus
rejecting the condition of barrier free mobility.
Domestic firm is the origin, foreign firm is the destination. The domestic firms feed
analysts into the foreign firms. This can be read directly from Table 6. We observe a
conspicuous pattern where the total outflow is greater than the total inflow among the domestic
firms, and a reverse pattern among the foreign firms. For example, there were 48 analysts who
22
originated at Daiwa and 24 who remain there. On the other hand, 3 analysts started at Goldman
Sachs, but 15 ended up there.
Mobility and immobility at the top. The top two domestic firms have the highest share of
immobility than do other firms. Among those who start in these firms, about half remain as
recorded in the last column of the table labeled “% stayers.” For the movers from these top
domestic firms, the destinations are the top foreign firms. For the analysts in the other (lowerstatus) domestic firms, the majority move on to the lower-status foreign firms.
Point of origin determines not only how frequently these analysts move between jobs
(Table 2b), but also where they move. Analysts from the top domestic firms move to top foreign
firms. Of the 33 analysts who left Nomura, 67 percent were placements into the high-status
foreign firms. At Daiwa, this ratio was 71 percent. In contrast, among the other domestic firms,
placement into the high-status foreign firms was only 33 percent.
Moving into the top two domestic firms is virtually impossible. Mobility into Nomura
and Daiwa from foreign firms was zero. Nine analysts moved into the two top domestic firms,
but these moves originated from other domestic firms. This finding confirms again the rigid
internal labor market setup of the top Japanese firms, and the barrier to entry for midcareer
professionals. An analyst from Nomura Securities explained as follows:
We are gradually adapting to the competitive environment in the finance sector, but we are still a Japanese company.
This means that we train our analysts in-house, and we still try to fill vacancies internally.
Interestingly, there is no mobility between the top two domestic firms; no one moves
from Nomura to Daiwa, or the other way around. One possible explanation is because these are
the top two domestic competitors, so mobility between competitors may be limited. But more
relevant to our discussion is the position that changing employers under the Japanese
employment system is a deviant act and a breach of the implicit contract. Quitting a Japanese
employer has enormous consequences for subsequent career trajectories, much more so than
quitting a foreign one. If one were to move, s/he will do so with the maximum benefit. Hence,
as one analyst explained, “it would not make sense to move to another Japanese firm which
offers the same pay scale and level of compensation.” This is another reason why most analysts
that quit the top domestic firms end up in the top foreign firms. It also explains why only one
analyst from Nomura and one from Daiwa moved to the other (lower-status) domestic firms.
23
Team mobility. In Table 6, we observe that mobility from the top domestic firms to the
top foreign firms is not random. For example, at Nomura, 8 analysts moved to Goldman Sachs;
at Daiwa, 7 moved to Deutsche Bank. While there is a strong Nomura-Goldman Sachs
connection, there is no migration from Nomura to Deutsche Bank.
This herd behavior could be related to team migration or some type of an institutional
linkage between firms, although the numbers are too small to make statistical inferences.
Working in teams is standard practice among security analysts, be it domestic or foreign. When
one moves, s/he may take their team members with them, especially if s/he happens to be a team
leader. One analyst at Deutsche Bank explained:
At one point there was a massive influx of analysts that came from Daiwa Securities. Much of this move was
initiated by upper-level managers who were formerly at Daiwa, and they pulled their teams into Deutsche. In some
cases, an entire section can be formed of Daiwa analysts.
An underwriter at Morgan Stanley explained:
In the 1990s, there was a strong presence of Nomura Securities at Morgan Stanley, because we had a leader from
Nomura. We poached a lot of specialists from Nomura, and fired our own. These days, we have a strong presence
from Daiwa Securities.
Heterogeneity: The positive association between workforce heterogeneity and turnover is
one of the key claims in the organizational demography literature (e.g. Sorensen 2000). If
outflow is greater than inflow in the domestic firms, and the opposite is true in the foreign firms,
then we would expect greater heterogeneity in the foreign firms.
In Table 6, the last two rows report the percentage of analysts that are inbred, and the
index of differentiation in each firm. I note that the numbers reported here are approximations.
Since I only tabulate the first job change, the destination is the second employer which is not
necessarily the current employer (I re-estimated the numbers using a table that records the
current employer instead of the second employer, and achieved similar results). Also, the sample
of analysts recorded here is not comprehensive, but a sub-sample that was featured in the Nikkei
Financial Daily. The reported numbers should be interpreted in the context of these
simplifications.
For each firm, the index of differentiation is estimated by the following equation (Gibbs
and Poston 1975):
24
C
Index of differentiation = 1 
N
i 1
2
i
 C

  Ni 
 i 1 
2
(3)
where N is the number of analysts who moved to the current employer from firm i. The index is
bound by the absolute minimum of zero, and a relative maximum, 1 – 1/C where C is the number
of previous firms represented in the current employer. The condition of zero heterogeneity, or
perfect homogeneity is achieved when C is equal to one, or when all analysts in the firm come
from the same origin.
We first observe that the percentage of inbred analysts is over 85 percent in the top
domestic firms. This is a remarkably higher percentage compared to 5 percent in the top foreign
firms, and 2 percent in the other foreign firms. These numbers confirm that the culture of
inbreeding – where analysts are trained and raised internally – remains the norm in the top
domestic firms, and a culture of poaching is pervasive in the foreign firms.
Accordingly, we find that the index of differentiation is low – 0.23 and 0.27 – in the top
domestic firms, and high in the top foreign firms, where it ranges from 0.69 to 0.97. The
heterogeneous composition of the analysts at foreign firms can be confirmed by eyeballing the
numbers on Table 6. At Morgan Stanley for example, there are analysts who moved there from
Nomura, Daiwa, and JP Morgan, as well as other domestic and other foreign firms. This
demographic heterogeneity is both the cause and the outcome of high turnover in the foreign
firms. The low percentage of inbred workers means that there is a large influx of workers with
previous experience. The large influx results in a more diverse and heterogeneous workforce,
which in turn leads to higher turnover.
Finally, I estimate mobility probabilities between origins and destinations among the
security analysts. Table 7(a) shows the 5 × 5 mobility table. This is a collapsed version of Table
6 where I group the top foreign firms into one category. I compare five different specifications
of models. Four of these are standard in the literature – independence, quasi-independence,
symmetry, and quasi-symmetry. In the fifth model, I construct my own design matrix as shown
in Table 7(b), which I call the asymmetry model. Our analysis thus far has found strong
25
evidence that the mobility from domestic to foreign goes only in one direction. The asymmetry
model is a direct test of this one-way transition. I hypothesize that the mobility between the
origin and destination is not symmetric across the diagonals. The designated reference category
is cell number 1 which corresponds to immobility in the top domestic firms, i.e. this is the group
of analysts in the top domestic firms that did not move. I note one caveat here. In the three
diagonal cells numbered 5, we are unable to distinguish between those who moved and those
who did not. For example, in the category of top foreign firms, cell 5 is confounded by the
immobility of analysts who remained in the top foreign firms, as well as the transition of analysts
that moved from one top foreign firm to another top foreign firm. The pattern of immobility is
rare among the foreign firms and in the other domestic firms (as shown in Table 6), but we
cannot rule out the possibility that the parameter estimates in these three diagonal cells may be
confounded by this heterogeneous population.
TABLE 7 HERE
I estimate mobility probabilities by fitting generalized linear models where the link is the
natural logarithmic function, and outcome y takes on a Poisson distribution:
ln {E(y)} = xβ
y ~ Poisson
(4)
Table 7(c) shows the model comparisons. The asymmetry model, which assumes an
asymmetric pattern of mobility between domestic and foreign firms, achieves the best fit as
evaluated by the BIC statistics. The significant improvement over the independence model
suggests that the condition of no association between origin and destination (corresponding to
null hypothesis condition [1]) is rejected. I proceed to present results of the asymmetry model in
Table 7(d).
All parameters in Table 7(d) are negative, thus indicating that all patterns of mobility
shown here are less likely to occur compared to the reference group of immobility between the
top domestic firms. We again confirm the overwhelming pattern of one-way transition from the
domestic to the foreign. The results clearly show that the transition into the domestic firms is
significantly less likely to occur than is the transition into the foreign firms, regardless of origin.
26
The coefficients reported under the top domestic firms and other domestic firms are significantly
more negative than the ones reported under top foreign and other foreign.7 Regardless of one’s
origin, analysts who change jobs in midcareer are more likely to end up in foreign firms. These
findings are consistent with the commonly observed cycle of job hopping in the foreign firms.
Moreover, we confirm that entry into the top domestic firms from any firm is highly
restricted. Among the four columns, the coefficients for top domestic are the most negative.
Analysts who started their careers elsewhere are unlikely to find midcareer opportunities in the
top domestic firms.
If analysts in the top domestic firms move, they are least likely to move to other domestic
firms. As discussed earlier, there is no incentive for these analysts to move to another domestic
firm, because any domestic firm will be inferior to their current employer. If they move at all,
analysts from the top domestic firms will choose foreign firms as their destinations because of
the reasons aforementioned – higher salaries, better recognition of their worth, etc. And if they
have the choice, they will choose the top foreign firms over other foreign firms because the
benefits generally are better there. This is in fact what we find. For the analysts from the top
domestic firms, the likelihood of mobility is lowest going into other domestic firms, and highest
for the top foreign firms.
SUMMARY

There is greater turnover in the foreign firm. Workers in foreign firms change employers
more frequently, and they are more likely to be hired under short-term contracts. The
higher turnover is related to human capital development and skill formation. Foreign
firms do not offer training, so the workers do not acquire firm-specific skills. These
workers invest mainly in general skills, which improves their outside options, and
subsequently weakens their attachment to the firm.

Mobility is not entirely uninhibited and barrier-free. Job mobility in the Japanese
financial sector is not flexible in all directions, but “sticky.” Workers only move in one
direction, from the domestic to the foreign, and not the other way around. Instead of
random mobility between origin and destination, there is a systematic pattern where the
destinations are determined by their origins.
27

The top domestic firms serve as feeders for the foreign firms. Workers in the domestic
firms receive extensive training, then migrate to the foreign firms where they can receive
higher compensation. Lastly, instead of an atomized pattern of individual mobility, we
find evidence of team mobility, dictated by informal contacts between workers and their
former bosses.

Finance professionals move in teams. They do so because they have built team-specific
capital that can be transferred with the team across firm boundaries, but that is difficult to
rebuild if the team is broken up.

Workers are not anonymous and identity is important. Workers want to maintain a
positive reputation with their bosses because the bosses will pull them into the valuable
jobs. Workers avoid the reputation for job-hopping, because excessive mobility signals
defection even among the foreign firms.

Workers in foreign firms are better informed about their outside options than are their
domestic counterparts. They know their market worth. They cultivate professional
networks, maintain contacts within and outside of their organization, and actively
exchange information. They use headhunters strategically, to keep themselves informed
of outside options, and to periodically reassess their market value.

Foreign firms offer higher salaries on average, but this is not the key driver of mobility.
The main motivation for moving is the pursuit of greater personal recognition. Many
workers who migrate to foreign firms do so because they do not conform well to the
group-oriented approach of the Japanese organization. They seek positions in foreign
firms because they value recognition for individual achievement over the long-term
orientation of the domestic firm.

For some, the cost-benefit calculation is more social, and economic motivation is less
important. They build and invest in firm-specific social capital. The social costs of
moving outweigh the economic benefits from moving.

Workers in foreign firms receive compensation that is more closely tied in to their own
performance and contribution. It is also easier to place a value on general skills than it is
on firm-specific skills.
28
Discussion
The current study has shown that whether one is loyal to the employer or not depends less
on individual attributes, and more on the institutional context that she is situated in. The findings
weaken cultural perceptions of the highly-committed Japanese worker by demonstrating that
loyalty is endogenous. Japanese workers in foreign firms are highly mobile and they show little
signs of loyalty.
At the same time, we uncover evidence which suggests that foreign firms adapt to local
labor market conventions. The Japanese financial sector operates under different notions of
organizational legitimacy; the highly atomized and market-driven Anglo-American approach can
be implanted but only up to a point. Referring back to Figure 2, firm M and firm K are distinct
organizational forms by virtue of their differing positions of marketness. However, firm K may
be a highly market-oriented firm in Japan, but it is still lower on the marketness scale compared
to firm L. For example, workers are free to move but too much mobility could send a bad signal
in a market where the status quo is still low mobility. Career mobility in the embedded market is
thus constrained by social conventions.
Over time, the Japanese employment system is destined to move from social to the
economic mode of exchange (Murakami and Rohlen 1992). Along with globalization, the
foreign firms are increasing their market presence in the Japanese economy. As Westney (2001)
explains, these global pressures are pushing Japan’s business practices towards a systematic
model for change: The move to U.S.-style shareholder capitalism. Recent reactions by domestic
firms illustrate best the pressure to normalize and to converge to global competition. Domestic
firms lost their best and brightest to the foreign firms during the 1990s. For the first time,
Nomura’s uncontested top position in the analyst rankings started to slip. The following excerpt
from Institutional Investor describes the brain drain problem, and how Nomura is fighting back:
To some, Nomura’s slip reflects the same problem that equity researchers are grappling with at many of the
Japanese companies they follow: a resistance to change. The country’s largest brokerage firm remained true to
Japan’s seniority-based system of compensation for most of the past decade -- and paid a price. Scores of young,
talented analysts defected to U.S. and European banks, where the pay structure is weighted more toward merit than
tenure… It wasn’t until February of last year that Nomura revamped its research compensation system to be more
competitive, a move that has met with some success. The firm didn’t lose a single analyst to a competitor last year
and even managed to rehire a former analyst. (Shimizu 2001)
29
Throughout my interviews, I learned about specific actions that the domestic firms are taking to
counter the exodus of their best human capital, e.g. introducing separate career tracks for
specialists, and overhauling the compensation system to better match that of the foreign firms.
These actions, especially calling back their former people who defected to the foreign firms,
would have been unthinkable under the traditional Japanese employment system.
From the viewpoint of institutional complementarities, these market corrections will
undermine the overall interdependence of HR practices, and may eventually destabilize the
Japanese employment system. But such actions may be inevitable in the face of globalization
and increased market competition from the foreign firms.
At the same time, there are signs that some foreign firms are moving towards an internal
labor market setup. In Figure 2, the domestic firms (M) are moving in the direction of higher
market orientation, and foreign firms (K) are moving towards non-market orientation. For
example, foreign firms that have a longer history of conducting business in Japan are now
systematically recruiting new university graduates and intensifying their internal training (Ikaros
2002). These recent developments in both domestic and foreign firms suggest a pattern of
convergence (Jacoby 2005). The co-existence of domestic and foreign firms, and the expanding
presence of the latter are fascinating areas of research, and I reserve this for my future work.
Notes
1
Lifetime employment figures are estimated from Ministry of Health, Labor and Welfare statistics using method
described by Ono (2010). It measures the number of workers in the age group 50 to 54 who have never changed
employers divided by the total number of workers in that age group.
2
For example, JILPT (2005) reports that the percentage of midcareer hires in domestic firms was 60 percent in
domestic firms, compared to 82 percent in foreign firms.
3
The NLI data was provided by the Social Science Japan Data Archive, Center for Social Research and Data
Archives, Institute of Social Science, The University of Tokyo.
4
I experimented with several variations of these classifications, and achieved similar results.
5
Source: Nihon Keizai Shimbun. 2005. “Nomura soken, jiyu ou DNA” (Nomura Research Institute: Opportunities
and freedom). September 19, 2005.
6
This is the case of deferred compensation. If workers accept the training, they will be paid lower than their
marginal product, and they will have to wait until the training period is over to recoup their losses.
7
The one exception here is the transition from top foreign to other domestic (cell 11), and the transition from top
foreign to other foreign (cell 12). The coefficient in cell 11 is more negative than the one in cell 12, but this
difference is not statistically significant.
30
References
Abegglen, James C. 1958. The Japanese Factory: Aspects of Its Social Organization. Glencoe: Free Press.
AERA. 2005. Gaishikei he yokoso. (Welcome to the Foreign Firm). Tokyo. Asahi Shimbunsha.
Akerlof, George A. 1982. “Labor Contracts as Partial Gift Exchange.” Quarterly Journal of Economics 97:543-569.
Aoki, Masahiko. 2001. Toward a Comparative Institutional Analysis. Cambridge: MIT Press.
Baker, Wayne E. 1984. “The Social Structure of a National Securities Market.” American Journal of Sociology
89:775-811.
Baron, James N. 1988. “The Employment Relation as a Social Relation.” Journal of the Japanese and International
Economies 2: 492-525.
Baron, James N., and Michael Hannan. 1994. “The Impact of Economics on Contemporary Sociology.” Journal of
Economic Literature 32:1111-1146.
Baron, James N., Michael T. Hannan, and M. Diane Burton. 2001. “Labor Pains: Change in Organizational Models
and Employee Turnover in Young, High-Tech Firms.” American Journal of Sociology 106: 960-1012.
Becker, Gary S. 1993. Human Capital. Chicago: University of Chicago Press.
Blau, Peter M. 1968. “Social Exchange.” Pp. 452-458 in International Encyclopedia of the Social Sciences, Vol.7.
Edited by David L. Sills. New York: Macmillan and Free Press.
Block, Fred. 1990. Postindustrial Possibilities: A Critique of Economic Discourse. Berkeley: University of California
Press.
Burk, James. 1988. Values in the Marketplace: The American Stock Market under Federal Securities Law. New York:
Walter de Gruyter.
Egami, Go. 2007. Shikkaku shain (Failed workers). Tokyo: Shincho bunko.
Friedland, Roger and Robert R. Alford. 1991. “Bringing society back in: Symbols, practices, and institutional
contradictions.” Pp.232-263. In The New Institutionalism in Organizational Analysis. Edited by Walter W.
Powell and Paul J. DiMaggio. Chicago: University of Chicago Press.
Fukao, Kyoji, and Keiko Ito. 2003. “Foreign Direct Investment and Services Trade: The Case of Japan.” Pp. 429-480.
In Trade in Services in the Asia-Pacific Region. Edited by Takatoshi Ito and Anne O. Krueger. Chicago:
University of Chicago Press.
Gibbs, Jack P., and Dudley L. Poston, Jr. 1975. “The Division of Labor: Conceptualization and Related Measures.”
Social Forces 53: 468-476.
Gilpin, Robert. 2001. Global Political Economy: Understanding the International Economic Order. Princeton:
Princeton University Press.
Granovetter, Mark S. 1985. “Economic Action and Social Structure: The Problem of Embeddedness.” American
Journal of Sociology 91:481-510.
Groysberg, Boris, Linda-Eling Lee and Ashish Nanda. 2008. “Can They Take It With Them? The Portability of Star
Knowledge Workers’ Performance.” Management Science 54: 1213-1230.
Heckman, James J. 1979. “Sample Selection Bias as a Specification Error.” Econometrica 47:153–61.
Hirsch, Paul M. 1987. Pack Your Own Parachute: How to Survive Mergers, Takeovers, and Other Corporate
Disasters. Reading: Addison-Wesley Publishing.
Hirsch, Paul M., Stuart Michaels, and Ray Friedman. 1987. “‘Dirty Hands’ Versus ‘Clean Models’: Is Sociology in
Danger of Being Seduced by Economics?” Theory and Society 16:317-336.
Ikaros. 2000. Gaishikei kigyo shushoku guide book 2000 (Employment guide book for foreign firms 2000). Tokyo:
Ikaros Publication.
_______. 2002. Gaishikei kigyo shushoku guide book 2002 (Employment guide book for foreign firms 2002). Tokyo:
Ikaros Publication.
Jacoby, Sanford M. 2005. The Embedded Corporation: Corporate Governance and Employment Relations in Japan
and the United States. Princeton: Princeton University Press.
Japan External Trade Organization (JETRO). 2004. Gaishi kei kigyou koyou chosa (Employment survey of foreignowned firms). Tokyo: JETRO.
Tainichi chokusetsu toshi ni kansuru gaishikei kigyou ishiki chousa (Attitude Survey of Foreign Direct investment in
Japan). Tokyo: JETRO.
Japan Institute for Labour Policy and Training (JILPT). 2005. Gaishikei kigyo no roushi kankei tou jittai chousa
kekka houkokusho (Results from the survey of labor and management relations of foreign-owned firms). Tokyo:
JILPT.
31
Lazear, Edward P. and Kathryn L. Shaw. 2007. “Personnel Economics: The Economist's View of Human Resources.”
Journal of Economic Perspectives 21: 91-114.
Lin, Nan. 2000. “Inequality in Social Capital.” Contemporary Sociology 29: 785-795.
Long, J. Scott and Jeremy Freese. 2006. Regression Models for Categorical Dependent Variables Using Stata (2 nd
Edition). College Station: Stata Press.
Mincer, Jacob. 1962. “On-the-Job Training: Costs, Returns, and Some Implications.” Journal of Political Economy
70: 50-79.
Ministry of Economy, Trade and Industry (METI). 2006. Gaishikei kigyo doko chosa (Survey of Trends in Business
Activities of Foreign Affiliates). Tokyo: METI.
Moriguchi, Chiaki, and Hiroshi Ono. 2006. “Japanese Lifetime Employment: A Century’s Perspective.” Pp. 152-176.
In Institutional Change in Japan. Edited by Magnus Blomström and Sumner La Croix. London: Routledge.
Murakami, Yasusuke, and Thomas P. Rohlen. 1992. “Social-Exchange Aspects of the Japanese Political Economy:
Culture, Efficiency, and Change.” Pp. 63-105 in The Political Economy of Japan, Volume 3: Cultural and Social
Dynamics. Edited by Shumpei Kumon and Henry Rosovsky. Stanford: Stanford University Press.
Nakane, Chie. 1970. Japanese Society. Berkeley: University of California Press.
Ono, Hiroshi. 2007. “Careers in Foreign-Owned Firms in Japan.” American Sociological Review 72: 267-290.
_______. 2010. “Lifetime Employment in Japan: Concepts and Measurements.” Journal of the Japanese and
International Economies 24(1): 1-27.
Ono, Hiroshi, and Kazuhiko Odaki. 2011. “Foreign ownership, human capital, and the structure of wages in Japan.”
The International Journal of Human Resource Management 22: 3036-50.
Ono, Hiroshi, and Marcus E. Rebick. 2003. “Constraints on the Level and Efficient Use of Labor.” Pp. 225-257. In
Structural Impediments to Japan's Economic Growth. Edited by Magnus Blomstrom, Jennifer Corbett, Fumio
Hayashi, and Anil Kashyap. Chicago: NBER and University of Chicago Press.
Pfeffer, Jeffrey. 1983. “Organizational demography.” Research in Organizational Behavior 5: 299-357.
Podolny, Joel M. 2005. Status Signals: A Sociological Study of Market Competition. Princeton: Princeton University
Press.
Robinson, Patricia A. 2003. “The embeddedness of Japanese HRM practices: The case of recruiting.” Human
Resource Management Review 13: 439-465.
Scott, W. Richard. 2002. Organizations: Rational, natural, and open systems (5th edition). Englewood Cliffs: Prentice
Hall.
Scott, W. Richard, John W. Meyer and Associates. 1994. “Developments in institutional theory.” In Institutional
Environments and Organizations: Structural Complexity and Individualism. Edited by W. Richard Scott, John W.
Meyer and Associates. Thousand Oaks: Sage Publications.
Shimizu, Kazuhiko. 2001. “The 2001 All-Japan Research Team.” Institutional Investor 35: 83-86.
Sorensen, Jesper B. 2000. “The Longitudinal Effects of Group Tenure Composition on Turnover.” American
Sociological Review 65: 298-310.
Suenaga, Toru. 1999. Make Money! (in Japanese). Tokyo: Bungei shunju.
Toyo Keizai. 2006. Gaishikei kigyo souran (Foreign affiliated companies in Japan: A comprehensive directory).
Tokyo: Toyokeizai shinposha.
Westney, D. Eleanor. 2001. “Japanese Enterprise Faces the Twenty-First Century.” Pp. 105-143. In The Twenty-First
Century Firm: Changing Economic Organization in International Perspective. Edited by Paul DiMaggio.
Princeton: Princeton University Press.
32
Table 1: Determinants of entry into foreign firms (probit coefficients)
(a)
(b)
NLI sample
Coef
Schooling completed
High school
University
Graduate school/ MBA
Foreign university
Female
Age
Tenure
English skills
Years of specialization
Years of specialization
squared
Constant
Security analysts
S.E.
(reference)
-0.482
(0.310)
1.011
***
(0.347)
0.617
0.052
-0.110
0.432
0.165
**
***
***
**
***
(0.214)
(0.020)
(0.019)
(0.186)
(0.048)
-0.006
***
(0.003)
-2.522
(0.758)
Coef
S.E.
(reference)
-0.762
(0.514)
1.788 ***
(0.531)
0.989
0.042
-0.152
-0.535
**
**
***
(0.430)
(0.017)
(0.018)
(0.636)
Log likelihood
-149.0
-135.8
Pseudo R2
0.26
0.35
* p<.10, ** p<.05, *** p<.01. Standard errors in parentheses.
Security analysts data also includes survey year dummies to control for yearspecific fixed effects.
33
Table 2a: Summary statistics for the number of job changes
Mean
S.D.
Median
NLI data
Domestic
Foreign
0.21
1.36
0.63
1.33
0
1
Security analysts
Domestic
Foreign
0.63
1.91
1.00
1.09
0
2
Top domestic
Other domestic
Top foreign
Other foreign
0.19
1.33
1.59
2.28
0.55
1.15
1.08
1.21
0
1
1
2
Table 2b: Poisson regression coefficients predicting frequency of job changes
NLI data
Security analysts
(a)
(b)
Coef
S.E.
Coef
S.E.
Coef
Current employer
Foreign firm
0.877
***
(0.316)
0.417
***
S.E.
(0.105)
Top domestic firm
Other domestic firm
Top foreign firm
Other foreign firm
First employer
Top domestic firm
Schooling completed
University
Graduate school/ MBA
Female
Foreign national
Cohort 30
Cohort 40
Cohort 50
Lambda (λ)
Constant
(c)
(reference)
0.841 *** (0.328)
0.962 *** (0.322)
1.160 *** (0.331)
-0.104
-0.026
-0.210
0.843
1.102
0.049
-1.531
-1.019
(0.241)
(0.295)
(0.210)
***
***
***
(0.267)
(0.331)
(0.619)
(0.307)
(0.764)
-0.316
***
(0.085)
-0.281
***
(0.082)
-0.431
-0.369
-0.203
0.586
0.585
0.532
-1.421
-1.114
***
**
(0.129)
(0.164)
(0.189)
(0.289)
(0.293)
(0.317)
(0.129)
(0.315)
-0.368
-0.328
-0.215
0.586
0.544
0.517
-1.307
-1.801
***
**
(0.128)
(0.162)
(0.176)
(0.271)
(0.279)
(0.297)
(0.143)
(0.425)
**
**
*
***
***
**
*
*
***
***
-337.01
Log likelihood
-217.10
-340.19
Pseudo R2
0.37
0.26
0.27
* p<.10, ** p<.05, *** p<.01. Standard errors in parentheses.
NLI data also controls for years of specialization.
Security analysts data also includes survey year dummies to control for year-specific fixed effects.
34
Table 3: Logit coefficients predicting differences in employment characteristics between domestic and foreign firms
Foreign firm
Log likelihood
Pseudo R2
coefficient
Prospect of job rotation
Boss is inbred (internally promoted)
(0.414)
(0.328)
-158.8
-139.1
0.26
0.34
Q: Which of the following has helped you improve your expertise?
Boss supervision on the job
-0.628 **
Self-improvement
0.625 **
(0.300)
(0.310)
-336.3
-270.3
0.03
0.03
Q: What areas do you seek improvement in your firm?
0.850
Education and training
Match between compensation and performance
-0.596
***
*
(0.330)
(0.317)
-183.9
-202.7
0.06
0.05
My performance is directly linked to compensation
My performance is directly linked to promotion
1.098
0.634
***
**
(0.282)
(0.302)
-358.4
-332.0
0.05
0.05
0.624
**
(0.285)
-369.2
0.04
0.796
**
(0.313)
-285.8
0.08
(0.339)
-178.6
0.03
My firm recognizes the importance of
professionals and professional skills
I am evaluated highly for my expertise
I work in teams
-1.730
-1.861
-0.036
***
***
Short-term contract
1.261 **
(0.616)
-51.5
0.34
Annual salary negotiation
2.440 *** (0.399)
-117.4
0.37
* p<.10, ** p<.05, *** p<.01. Standard errors in parentheses.
Shown are coefficients for the foreign firm dummy for separate ordered logit regressions. All regressions control
for sex, education, age, tenure, years of specialization, and lambda (λ).
35
Table 4: Employment characteristics of movers (logit coefficients)
Foreign firm
Log likelihood
coefficient
Pseudo R2
My previous employer was a foreign firm.
My salary is now higher as a result of the move.
2.251
1.874
**
***
(0.785)
(0.667)
-40.7
-56.5
0.35
0.16
Reasons for moving:
I moved because there was a well-established
system of performance evaluation in place at
the current firm.
I moved because of higher pay
2.039
**
(0.908)
-31.5
0.22
(0.687)
-43.2
0.11
-47.5
-42.1
-48.1
-37.2
0.18
0.14
0.04
0.13
-0.484
Job search channel:
I found the job through a newspaper ad.
-1.674 *** (0.645)
(0.629)
I found the job through a headhunter.
1.045 *
(0.668)
I was recommended by an acquaintance.
0.240
(0.830)
I was directly invited by the current employer.
0.575
* p<.10, ** p<.05, *** p<.01. Standard errors in parentheses.
Shown are coefficients for the foreign firm dummy for separate ordered logit regressions.
for sex, education, age, tenure, years of specialization, and lambda (λ).
All regressions control
36
Table 5: Mobility table of security analysts
Number
(b) Number of employers
Number of
of moves
persons
1
2
3
4
5
6
(a)
(c)
0
D
91
0
F
9
1
D
D
27
1
D
F
57
1
1
F
D
1
F
F
13
2
D
D
D
9
2
D
D
F
8
2
D
1
F
D
2
D
F
F
33
2
D
0
F
D
2
F
1
F
D
2
F
0
F
D
2
F
F
F
9
3
D
D
D
D
1
3
D
D
D
F
0
3
D
D
0
F
D
3
D
D
F
F
4
3
D
D
1
F
D
3
D
F
2
F
D
3
D
F
7
F
D
3
D
F
F
F
29
3
D
D
0
F
D
3
D
F
0
F
D
3
0
F
D
F
D
3
F
F
0
F
D
3
F
D
0
F
D
3
F
F
0
F
D
3
F
F
2
F
D
3
F
F
F
F
3
4
D
D
D
F
F
1
4
D
D
F
F
F
1
4
D
D
F
1
F
D
4
D
F
F
1
F
D
4
D
F
F
1
F
D
4
D
F
F
F
F
4
4
F
F
F
1
F
D
:
(other possible mobility patterns suppressed)
:
5
D
F
F
F
1
F
D
5
D
D
F
F
F
F
1
5
D
F
F
F
F
F
1
5
F
F
F
F
F
F
1
:
(other possible mobility patterns suppressed)
:
Total
322
D: Domestic firm, F: Foreign firm
Total number
of moves
(d) = (a) × (c)
0
0
27
57
1
13
18
16
2
66
0
2
0
18
3
0
0
12
3
6
21
87
0
0
0
0
0
0
6
9
4
4
4
4
4
16
4
:
5
5
5
5
:
427
37
Table 6: Mobility of security analysts
Destination
Origin
NM
DW
NSSB
GS
DB
JPM
ML
MS
UBS
Other
D
Other
F
Total
D: Nomura (NM)
D: Daiwa (DW)
D/F: Nikko/NSSB (NSSB)
F: Goldman Sachs (GS)
F: Deutsche Bank (DB)
F: JP Morgan (JPM)
F: Merrill Lynch (ML)
F: Morgan Stanley (MS)
F: UBS Warburg (UBS)
D: Other domestic
F: Other foreign
35
0
0
0
0
0
0
0
0
5
0
0
24
0
0
0
0
0
0
0
4
0
3
2
19
1
0
1
0
0
0
7
6
8
0
0
1
0
0
0
0
0
4
2
0
7
3
0
0
0
0
0
1
5
0
2
1
1
0
0
1
0
0
0
3
0
4
2
0
0
0
0
0
0
0
5
0
4
4
0
0
0
1
0
1
0
6
2
1
1
0
0
0
0
0
1
1
8
1
1
1
5
0
0
0
0
0
0
38
2
10
6
9
1
0
0
0
0
0
48
13
68
48
37
3
0
3
0
2
2
133
26
Total inflow
40
28
39
15
16
8
11
18
13
47
87
322
% Inbred
87.5
85.7
48.7
29.8
2.3
30.4
Index of differentiation
0.23
0.27
0.73
0.99
0.95
0.23
4.9
0.69
0.77
0.89
0.83
0.88
0.97
%
Stayers
51.5
50.0
51.4
40.0
10.5
7.7
D: Domestic firm, F: Foreign firm
38
Table 7: Mobility tables
(a) Counts
Origin
Nomura
Nomura
Daiwa
Top Foreign
Other Domestic
Other Foreign
Daiwa
35
0
0
5
0
0
24
0
4
0
Destination
Top F
22
17
31
38
11
Other D
Other F
1
1
5
38
2
10
6
11
48
13
(b) Design matrix – Asymmetry model
Origin
Nomura
Nomura
Daiwa
Top Foreign
Other Domestic
Other Foreign
Daiwa
1
6
10
6
15
6
1
10
6
15
Destination
Top F
7
7
5
13
16
Other D
Other F
8
8
11
5
17
9
9
12
14
5
(c) Model comparisons
Model Specification
Independence
Quasi-independence
Symmetry
Quasi-symmetry
Asymmetry
G2
206.34
16.35
174.04
7.26
0.39
df
16
11
13
6
7
BIC
154.84
-19.06
132.19
-12.06
-22.15
(d) Estimated parameters
Origin
Nomura
Daiwa
Nomura
-
-4.094
(1.008)
Daiwa
-4.094
(1.008)
***
-
Destination
Top F
-0.405 **
***
(0.204)
Other D
-2.996
(0.592)
***
Other F
-1.261
(0.275)
***
-0.405
(0.204)
**
-2.996
(0.592)
***
-1.261
(0.275)
***
Top Foreign
-6.196 ***
(1.475)
-2.053
(1.069)
*
-3.798
(1.165)
***
-3.061
(1.123)
***
Other Domestic
-4.094 ***
(1.008)
-2.053
(1.069)
*
-2.053
(1.069)
*
-1.822
(1.066)
*
-5.349 ***
-2.213
(1.489)
(1.142)
* p<.10, ** p<.05, *** p<.01. Standard errors in parentheses.
*
-3.740
(1.272)
***
-2.053
(1.069)
*
Other Foreign
39
Long
D
Employment
duration
F
Short
Low
High
Marketness
Figure 1: Foreign versus domestic firms
40
High
Marketness
of the
organization
Low
K
L
U.S. (foreign)
firm in Japan
U.S. (domestic)
firm in the U.S.
M
N
Japanese (domestic)
firm in Japan
Japanese (foreign)
firm in the U.S.
Low
High
Marketness of the environment
Figure 2: Foreign versus domestic firms
41