Effects of diversification strategies on investment in railway business: The case of private railway companies in Japan

Research in Transportation Economics xxx (2016) 1e9

Contents lists available at ScienceDirect

Research in Transportation Economics journal homepage: www.elsevier.com/locate/retrec

Effects of diversification strategies on investment in railway business: The case of private railway companies in Japan Yeon-Jung Song*, Kenichi Shoji Graduate School of Business Administration, Kobe University, 2-1, Rokkodai, Nada, Kobe, 657-8501, Japan

a r t i c l e i n f o

a b s t r a c t

Article history: Received 31 October 2015 Received in revised form 30 June 2016 Accepted 20 July 2016 Available online xxx

Operating deficits and the increasing subsidies given to public transport are the main topics of discussion for both policymakers and scholars in the transportation sector. As a solution to these issues, the participation of private transport operators and creation of other sources of income than simply providing transportation services have been discussed in recent years. In Japan, private railway companies are the main providers of railway services, especially in urban areas. Although most such private railway companies provide adequate services and operate at a surplus, they receive no subsidies to cover their operating deficits and maintenance costs. Further, the most remarkable feature of private railway companies in Japan is their widely deployed diversification strategy, even into areas barely related to the transportation business, by using physical or tangible assets. This study investigates how the diversification strategies of private railway companies in Japan influence the railway business, focusing on the relationship between diversification and investment in transportation and the use of internal capital by diversified private railway companies. We review the research findings on investment and estimate quantitative models to examine the effect of diversification on transportation investment, based on the internal capital market theory discussed in the corporate finance literature. We find that investment in the railway business is influenced by diversification in terms of capital resource allocation. Internal capital is likely to be utilised for investment in not only the transportation business, but also diversified businesses. We conclude that this result is driven by the expected limitations of growth and profitability, as well as the financial support schemes available to transportation businesses. © 2016 Elsevier Ltd. All rights reserved.

JEL classification: D220 G310 H410 R420 Keywords: Private railway Urban transportation Diversification strategies Investment Regulations

1. Introduction One of the most popular topics in public transport is using private sector operations such as franchising and competitive tendering to overcome the low profitability and operational inefficiency of public transport (Hensher & Stanley, 2008; YvrandeBillon, 2006). In fact, passenger transportation services operated by the private sector were a general form of provision during the ‘railway boom’ that began in Europe, Japan, and North America at the turn of the 20th century and led to the establishment of many private railway companies (PRCs hereafter). However, while numerous such PRCs went bankrupt because of financial difficulties, PRCs in Japan survived by operating profit-making passenger railways in urban areas. Further, most PRCs in Japan are now also engaged in sectors such as real estate, retail, taxis and buses in

* Corresponding author. E-mail address: [email protected] (Y.-J. Song).

addition to their core railway business. While the success of PRCs is driven by densely populated urban areas and mass prepaid commuter passengers (Killeen, 1999), the difficulty in raising railway fares and demand decreases due to the decline in the productive age population are threatening the operation of PRCs. Nevertheless, as private enterprises, PRCs create their own business portfolios including making investment decisions and compete with other companies on certain routes based on market principles (Shoji, 2001). Indeed, the implementation of a diversification strategy is a unique feature of PRCs in Japan compared with those in other countries. Thus far, the diversification strategy of PRCs in Japan has been traditionally considered to be motivated and deployed in order to support the railway business (Saito, 1993; Shoji, 2001). However, the operating profit of diversified segments has been higher than those generated by railways and related businesses in some companies. Fig. 1 illustrates the change in the ratio of the operating profit of the transportation segments and diversified segments of major PRCs in Japan from 2000 to 2012. This figure shows that the ratio of the total operating profit of

http://dx.doi.org/10.1016/j.retrec.2016.07.022 0739-8859/© 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Song, Y.-J., & Shoji, K., Effects of diversification strategies on investment in railway business: The case of private railway companies in Japan, Research in Transportation Economics (2016), http://dx.doi.org/10.1016/j.retrec.2016.07.022

2

Y.-J. Song, K. Shoji / Research in Transportation Economics xxx (2016) 1e9

is needed to guarantee a certain level of investment in railways. Finally, we provide conclusions in Section 7.

2. Overview of Japanese PRCs in urban areas

Fig. 1. Change in the ratio of the operating profit of the transportation segments and diversified segments of major PRCs in Japan (annual means).

diversified segments tended to increase compared with that of transportation segments. Some PRCs even announce expansions to the scale and extent of their diversified businesses in their annual reports or mid- to long-term management strategies, thereby emphasising the importance of diversification. In other words, diversification may no longer be considered to support the railway business, but rather be an essential business strategy for the financial sustainability of PRCs.1 Although diversification strategies seem to play a key role in PRCs, it has not been investigated how they influence the railway business in terms of internal capital resource allocation for investment, focusing on the coexistence of railways, which are considered to be public utilities, and diversified businesses. Previous research has noted the diversification strategies of Japanese PRCs and investigated their effects, focusing on their contribution to the profitability of the railway business and to the entire company. However, most previous studies focus on the relation between diversification and performance at the firm level; finding research on how diversification strategies influence investment in the railway business for acquiring and upgrading assets is more challenging. Based on the foregoing, we aim to investigate whether the internal capital of PRCs is utilised for investment in the railway business to improve railway services through route expansion and large-scale maintenance, considering the diversified business portfolio of the firm, and whether it is utilised to invest or improve other diversified businesses if it is not reinvested in the railway business. For these purposes, we explore the investment decisions of Japanese PRCs from the perspective of internal capital markets (ICMs) as well as free cash flow and life-cycle theory, which are discussed in the corporate finance literature. The structure of the remainder of this paper is as follows. First, we describe the railway system in Japan, including Japanese PRCs in large urban areas and their business strategies. We then review the literature related to ICMs, which determine capital resource allocation by diversified firms. In Section 4, we quantify the effects of diversification on investment in railway companies by using data on Japanese PRCs. Our analysis includes only major PRCs operating in large urban areas. In Section 5, we refer to the interviews with three major Japanese PRCs to interpret the results. We next discuss whether the private operation of the railway business is desirable in terms of attracting investment and whether a regulatory scheme

1 On whether diversification strategies support the transportation business, especially the railway business, while diversification may be beneficial in terms of cost and future income, little quantitative research has thus far examined these opinions.

The main operators of urban railways in Japan are usually classified into, JRs (Japan Railway), private railways, and public and quasi-public firms. Among these operators, PRCs play a large role as organisations run independently of the government. According to the Ministry of Land, Infrastructure, Transport and Tourism of Japan, there are 177 passenger railway operators in Japan (excluding cable car, ropeway, and non-passenger rail operators such as freight operators) and 142 of these firms were categorised as PRCs in 2012. Of these, 16 firms were ‘major’ companies, namely those that provided services in Japan's three metropolitan areas, which are characterised as having large demand for railway services (except the Nishi Nippon Railroad).

2.1. Private operators independent of the government and railway regulations Many of the companies established during the large-scale construction of railways in the late 19th century and the early 20th century in Japan, still exist today in the form of PRCs. On the contrary, the main operators of urban railways in the United States and Europe have changed from the private to the public sector because of financial difficulties or through the discontinuation of deficitridden railway lines. PRCs in Japan offer not only railway services, but also bus and taxi services based on the principle of selfsufficiency in order to generate a commercial return. In addition, most of Japan's 16 PRCs are listed on the Tokyo stock market, and these have formed group firms that have several subsidiary companies. They receive few operation subsidies from the government, although some financial support programmes are available such as loans with interest relief that is lower than that provided by banks or other capital markets and tax relief on funds for infrastructure investment (Shoji, 2001). Owing to public concerns, all railway operators including PRCs must adhere to regulations related to market entry (e.g. the permission-based system),2 fares (e.g. full-cost pricing with rateof-return and ceiling price and yardstick regulation), and safety standards according to the Railway Business Act and the Railway Operation Act. The Ministry of Land, Infrastructure and Transportation and Tourism formulates and enforces these regulations (Mizutani, 2014a).3 However, although several direct and indirect regulations exist, the government does not tend to intervene in planning new routes or operating railways, except for the highspeed railway (called Shinkansen).4 In addition, urban railway systems in large urban areas including Tokyo and Osaka are not operated or managed by the public sector, except for the municipal subway system.

2 Although railway operators had to acquire an operating license from the government before 2000, the entry regulation was changed to a permission after the abolition of the regulation controlling supply and demand. Exit regulation was also relieved after 2000, and operators can stop providing services by reporting it in advance (1 year before). For details, see Shoji (2001) and Mizutani (2014a). 3 See Mizutani (2014a) for details of entry, fare, and other regulations (p. 6e8). 4 The Ministry of Land, Infrastructure, Transport and Tourism has established the Council for Transport Policy (Koutsu Seisaku Shingikai). While this council suggests transportation policies and plans, it has no legal force and no compulsion, and the government hears the opinions from Council on railway policy. Although the government can offer proposals to railway operators on planning railway routes and improving railway systems, it cannot enforce operators to implement them.

Please cite this article in press as: Song, Y.-J., & Shoji, K., Effects of diversification strategies on investment in railway business: The case of private railway companies in Japan, Research in Transportation Economics (2016), http://dx.doi.org/10.1016/j.retrec.2016.07.022

Y.-J. Song, K. Shoji / Research in Transportation Economics xxx (2016) 1e9

3

2.2. Balance between public transportation operators and private enterprises

extent to which diversification strategies support the transportation business remains an important task for future research.

Although PRCs are regulated by such legislation, they are free to decide on new routes and infrastructural improvements based on their own profit motivations. Therefore, Japanese PRCs play two different roles: 1) as an operator providing public transportation services and 2) as a private enterprise pursuing profit maximisation. As operators of public transportation, PRCs should offer stable and safe services because urban railway is an indispensable part of daily life and they have to take the heavy responsibility for the railway accident. At the same time, they should try to provide better customer service to enhance profitability by maintaining and increasing ridership. On the other hand, in view of profitability, the transportation segments of diversified companies including bus and taxi modes as well as railways are profitable. The operating income of the transportation business can cover its own total costs and can generate sufficient additional capital and cash flow. Although cross-subsidies between transportation segments and other segments to make up for any losses are prohibited by law, as private firms, they have the right to allocate their own funds as they see fit. Thus, the headquarters can transfer the capital resources of the transportation segments to other segments. Thus, it is important for them to balance the profitability and sustainability of transportation services. For example, while they seek profitmaking businesses to ensure the profitability of the entire company, they also invest in transportation business to maintain service quality and safety. If the headquarters were concerned only about future profitability and the cost of capital motivated by profitseeking, they would push the transportation business aside when deciding on the priority order of investment because of its limited ability to generate future profitability.

3. Investment and ICMs

2.3. Diversification strategies Major PRCs in Japan have mostly diversified into real estate (e.g. selling and leasing houses and buildings) and distribution including retail (e.g. department stores, convenience stores, and supermarkets inside or around their terminal buildings and stations). Leisure and service businesses such as hotels operated locally and nationwide are also important. The maintenance of buildings and the manufacturing or maintenance of railway carriages are other main businesses, although the proportion of revenue of these is low compared with that of other businesses because of vertical integration, which reduces operating costs. Considering the limited profitability of the railway business because of fare regulations, it is regarded as natural and rational for PRCs, as private enterprises, to participate in new businesses that are free from regulation. Japanese PRCs have thus diversified 1) to create demand and/or increase ridership, 2) to develop other income sources, 3) to capture externalities resulting from the investment in the railway infrastructure, and 4) to exploit economies of scope and the strategic assets built by the railway business (Shoji, 2001), aiming to overcome limitations with regard to the profit and growth of railway business. Likewise, these diversified businesses have been considered to support the transportation business by reducing costs through vertical integration and absorbing the shock of decreasing future demand in the railway business. Regarding the effect of diversification strategies on the transportation business, only a few quantitative researches has estimated the relation between the number of passengers and the investment and income of diversified businesses. Killeen (1999) assumes that investment in diversified businesses might influence passenger use except commuters, but not to a statistically significant degree. Therefore, quantifying the

3.1. Diversified firms and ICMs The resource allocation in diversified firms is different from that in focused firms because the former consist of many subsidiaries, divisions, or business units that operate various businesses. ICMs allocate limited resources to the various divisions or segments inside the diversified company and the headquarters then have the authority to make allocation decision (Peyer, 2002; Stein, 1997). ICMs generated in diversified firms give the firm the opportunity to depend on the firm's cash flow by transferring capital between segments, thereby avoiding the frictions of external capital markets owing to the limited availability of funds (Hovakimian, 2011; Shin & Stulz, 1998). However, ICMs do not work well if the segment relies on its own cash flow to raise funds for its projects (Shin & Stulz, 1998). Peyer (2002) points out that establishing ICMs is a potential benefit of diversification in two main ways. The first is that internal resource allocation can be more efficient than allocation by external capital markets because of the relaxed constraints on managerial behaviour and use of less strict monitoring systems (Jensen, 1986). The second advantage is that ICMs affect transactions with external capital markets by reducing information asymmetry, because the headquarters can reallocate investment across divisions based on a more precise estimation of value-maximising investment needs compared with external investors (Shin & Stulz, 1998). If reallocating capital is based on the productivity of a segment (Khanna & Tice, 2001; Maksimovic & Phillips, 2002) and if the headquarters do not pursue the incentive, but rather engage in winner-picking (Stein, 1997), ICMs can facilitate efficient resource allocation. Stein (2003) also suggests two major benefits of ICMs, namely the ‘more-money’ effect and ‘smarter-money’ effect. The more-money effect is that diversified firms can acquire more total external financing than stand-alones, which is beneficial for the underinvestment problem. The smarter-money effect assumes that the CEO in ICMs knows about the prospects of the firm's divisions well and uses high-quality information to make value-enhancing reallocations across divisions if s/he behaves in the interests of external shareholders. 3.2. Life-cycle theory and free cash flow theory Free cash flow theory, developed by Jensen (1986), refers to the problem of agency costs and overinvestment. Before Jensen (1986), Mueller (1972) discussed the decreased marginal returns from investment in the core business as a consequence of the life-cycle effects (see also Markides, 1995; Thompson, 1999). Mueller (1972) argued that firms operating core businesses in mature industries that generate sufficient cash to be reinvested are likely to aim to maximise growth, rather than profit and stockholder welfare, and that internal capital makes investment funds available to managers. This therefore motivates managers to expand investment into different activities through conglomerate mergers. Jensen (1986) develops his theory by reformulating that of Mueller (1972). Free cash flow is defined as additional cash that can be used in other projects after financing all projects with positive net present values (NPVs). Free cash flow can be used for the firm's expansion and diversification as well as for dividends for share holders. Jensen (1986) also assumes that overinvestment results from the firm's weak governance and insists that free cash flow tends to be invested in unprofitable expansion owing to managerial

Please cite this article in press as: Song, Y.-J., & Shoji, K., Effects of diversification strategies on investment in railway business: The case of private railway companies in Japan, Research in Transportation Economics (2016), http://dx.doi.org/10.1016/j.retrec.2016.07.022

4

Y.-J. Song, K. Shoji / Research in Transportation Economics xxx (2016) 1e9

behaviour, which is characterised by the pursuit of rent. However, he also insists that this value-decreasing behaviour can be held back by the precommitment of cash flows generated by the core business to value-maximising projects. In addition to facilitating growth, corporate income fluctuations can further be attenuated by this investment expenditure (Amihud & Lev, 1981).5 3.3. Investment opportunity and cash flow A positive NPV is considered to be the most important factor of internal capital resources when deciding on capital allocations (Lamont, 1997; Shin & Stulz, 1998). However, other factors should also be considered such as the firm's strategy and future growth. Moreover, some investment decisions are based on endogenous factors such as the extent of the authority of the headquarters (Stein, 1997), organisational structure (decentralised or hierarchical; Stein, 2002), job histories of CEOs (Xuan, 2009), managerial incentives (Harris & Raviv, 1996; Scharfstein & Stein, 2000; Wulf, 2002), and aspirations for a division's future performance (Arrfelt, Wiseman, & Hult, 2013). However, neoclassical theory emphasises that the segment's investment tends to depend on the firm's perceived investment opportunities and its own cash flows and other segments' cash flows, rather than intra-firm factors (Maksimovic & Philips, 2008). In particular, the majority of previous research focuses on the effect of investment opportunities and liquidity, as measured by operating cash flows. A segment's investment opportunities influence firms' investment decisions markedly. Several corporate finance studies have focused on Tobin's q, which is the value of the investment opportunity divided by the cost of the required investment (Maksimovic & Philips, 2008). Previous studies of this topic suggest that segments of diversified firms do not tend to respond properly to investment opportunities, although opportunities measured by the firm's Tobin's q should be considered to decide the level of investment. This finding implies that diversified firms tend to underinvest in high opportunities, although they have a tendency to overinvest in low opportunities and even allocate too much to low-q segments and too little to high-q segments (Scharfstein, 1998; Shin & Stulz, 1998; Rajan, Henri, & Luigi, 2000; Billett & Mauer, 2003).6 On the contrary, as well as Tobin's q, a firm depends on cash flow to decide on investment. Much previous research suggests that a segment in diversified firms may rely both on its own cash flows and on those of other segments (Lamont, 1997; Shin & Stulz, 1998; Whited, 2001). A positive coefficient of the other segments implies transfer or subsidies among the segments (Shin & Stulz, 1998), with some previous studies investigating the effects of such transfers and subsidies (Rajan et al., 2000; Khanna & Tice, 2001; Billett & Mauer, 2003). Based on the discussion above, previous studies suggest the following investment equation, including investing opportunities and the cash flow of a segment and a firm (Lamont, 1997; Maksimovic & Philips, 2008; Shin & Stulz, 1998):

Ij ¼ b  qj þ d  CFj þ 4  CFj þ g  zj þ Dummyyear þ Dummyfirm where Ij is segment j's capital expenditure, qj is the Tobin's q of segment j, CFj is the cash flow of segment j, CFj is the cash flow of other segments except segment j, and zj represents the other explanatory variables.

5

See also Thompson (1999, p. 27). Stein (2003) finds that diversified firms engage in socialist cross-subsidisation in investment (p. 149). 6

In this research, based on the above investment equation, we focus on the extent to which the cash flow of the transportation segment and of other segments influences investment sensitivity in the transportation segment, rather than that of investment opportunity. In particular, we focus on the transfer of free cash flow, instead of operating cash flow, to investigate whether the firm's additional cash after operating activities can be invested in the transportation segment. Section 4 explains the variables used in the estimation equation and analysis strategies more in detail. 4. Empirical analysis 4.1. Sample and data description The sample includes 15 major PRCs,7 which are diversified and which report segment information in their consolidated financial statements, and three JR companies (JR East, JR Central, JR West), which were privatised in 1986 and listed on the stock markets for the 1999e2012 fiscal years. We obtain data on all sample firms during 1999e2012 since we use time-lagged data on decision timing, but our regressions cover 2000 to 2012. As a result, 228 observations are included in the sample. The data we use are extracted from the NIKKEI Financial quest database, which contains data on the consolidated financial statements published by each company. 4.2. Variables To investigate how ICMs affect investment in the transportation segment, we use the investment sensitivity of the transportation segment as the dependent variable, which is the ratio of the capital expenditure of the transportation segment to the firm's total assets in the previous year, following Shin and Stulz (1998). Investment sensitivity is usually defined as the ratio of capital expenditure at t to the total assets of the firm at t  1 (Shin & Stulz, 1998; Whited, 2001).8 The free cash flows of the transportation segment and other segments at t are also normalised by the firm's assets at t  1. We further include five independent variables including the share of transportation segment sales at t which controls for the effect of segment size (Scharfstein, 1998). Table 1 presents the descriptive statistics of the variables used in the regressions. 4.2.1. Cash flow As we presented in investment equation mentioned above, cash flow is a proxy of ICMs, and many studies estimating investment equations define cash flow as the sum of operating income and depreciation (Ahn, Denis, & Denis, 2006; Duchin, 2010; Lamont, 1997; Shin & Stulz, 1998; Whited, 2001). On the contrary, cash flow is also defined as the segment's after-tax cash flow (Billett & Mauer, 2003), net cash from operations minus dividends (Peyer, 2002), and free cash flow (Faulkender, Flannery, Hankins, & Smith, 2012). The definition of cash flows influences the conclusion drawn on the efficiency of ICMs. Indeed, Shin and Stulz (1998) regress a segment's capital expenditure on an imputed segment's Tobin's q, the segment's own cash flow, and the cash flow of the firm's other segments and suggest that a positive coefficient estimate on the other segments' cash flow implies cross-subsidised capital expenditure between segments. However, Billett and Mauer (2003) support the efficiency of ICMs. In this research, as

7 We do not consider Tokyo Metro as diversified PRC because it did not diversify and publish segment-level data during the study period. 8 According to Whited (2001), the variable of total assets is widely used in the existing literature as a scaling variable. In addition, if we assume that investment decisions are made at the firm level, total assets are appropriate for scale variables.

Please cite this article in press as: Song, Y.-J., & Shoji, K., Effects of diversification strategies on investment in railway business: The case of private railway companies in Japan, Research in Transportation Economics (2016), http://dx.doi.org/10.1016/j.retrec.2016.07.022

Y.-J. Song, K. Shoji / Research in Transportation Economics xxx (2016) 1e9

5

Table 1 Descriptive statistics.a Variable

Obs

Mean

Std. Dev.

Min

Max

Investment sensitivity (t) Free cash flow of the transportation segment (t) Free cash flow of the other segments (t) Return on assets (ROA) of the transportation segment (%) (t-1) Herfindahl index (t-1) Transportation segment's sales of overall sales (t)

228 228 228 228 228 228

0.0295 0.0020 0.0034 4.0971 0.6899 0.4106

0.0171 0.0135 0.0119 1.4043 0.1036 0.1848

0.0039 0.0486 0.0458 0.96 0.245 0.1329

0.0914 0.0483 0.0312 8.44 0.828 0.9020

a We calculated the correlation coefficients between all the variables and found no evidence of multicollinearity owing to high correlation. As well as investment sensitivity, the free cash flow of the transportation segment and other segments were normalised by total assets.

mentioned above, we focus on the additional cash after used in current or positive NPV projects. Moreover, following Jensen (1986), we adopt the free cash flows of the transportation segment and aggregated free cash flows of the other segments. Free cash flows are defined as operating income before depreciation less tax expense less capital expenditure, in line with Billett and Mauer (2003) and Faulkender et al. (2012). We expect cash flow to lead the headquarters in diversified firms to invest in growing businesses instead of the passengerdecreasing transportation segment, which has matured and thus has restricted profitability. In terms of the extra cash holdings of the firm, the free cash flow of other segments also affects investment in the transportation segment because of the limited capital resources. Hence, we use the free cash flow of the transportation segment and that of the other segments at t of investment. 4.2.2. Investment opportunities (investment profitability) As a proxy of investment opportunities, Tobin's q has been widely used in the literature. Whited (2001), however, states that Tobin's q is an inadequate variable with which to control for investment opportunities. At the firm-level, Tobin's q is generally calculated by using disclosed financial data, as the ratio of the total market value of the firm (the sum of total stock market value and total liabilities) to total asset value, while it is not observed directly at the segment level because the segments of diversified firms are not traded. Therefore, many studies use the median of the Tobin's qs of the focused firms, particularly based on a segment's threedigit industry SIC code. However, Whited (2001) insists that a segment's Tobin's q tends to lie below the industry median when a diversification discount exists. Much of the empirical literature also suggests that the investment opportunities of same-industry focused firms are not representative of those of the segments of diversified firms (Berger & Ofek, 1995; Billett & Mauer, 2003; Lang & Stulz, 1994; Villalonga, 2004). Moreover, most empirical studies use data on US companies, which classify segments by using SIC codes. However, in Japan, segment classification is not based on SIC code or on the standardised rule, but rather on the decision of a firm's manager. This distinction implies that matching the segments of diversified and focused firms is difficult because various businesses are mixed in one segment in this case, as is shown in the case of PRCs. Considering the difficulties and limitations mentioned above, Lang, Ofek, and Stulz (1996) use a segment's ROA, which is the ratio of operating income to segment assets, as a proxy of investment opportunities. In this vein, we also use the segment's ROA at t  1 as a variable of investment opportunities. 4.2.3. Extent of diversification Considering that diversified businesses in PRCs have become a large part of the firm, we assume that expanding diversification strategies may influence the investment policy of PRCs. In the literature, the extent of diversification is one factor influencing investment decisions (Maksimovic & Phillips, 2002; Shin & Stulz,

1998). The number of segments (Shin & Stulz, 1998) and Herfindahl index (Khanna & Tice, 2001; Maksimovic & Phillips, 2002; Rajan et al., 2000) are used as variables representing the extent of diversification. Maksimovic and Phillips (2002) show that investment sensitivity increases with a firm's Herfindahl index. We thus use the Herfindahl index at t  1 as a variable to define the extent of diversification, which computes a firm's dispersion across the segment in which it operates by summing the squared shares of total firm sales for each segment.

5. Results To investigate the investment policy of ICMs in the transportation segment to confirm the robustness of the analysis, we first estimate the baseline equation and model including the diversification index for the total sample. Then, we categorise the sample into two groups by operating region, namely the Kanto region surrounding the Tokyo metropolitan area (TMA) and the Kansai regions including Osaka and Nagoya. Finally, we estimate the same models for both groups. Shoji (2001) suggests that the business strategies and types of diversification strategies differ between the Kanto and Kansai regions and Mizutani (2014b) also finds a considerable difference between TMA and other regions except TMA with regard to the effect of yardstick regulation. In addition to strategic type, we expect categorising the sample into two groups to offer implications on relations among railway demand, diversification, and investment in the railway business. This is especially so considering the different demand levels of the two regions: railway demand is gradually increasing in the Kanto region because of the population influx into TMA, whereas demand is decreasing in the Kansai region. In addition, if ICMs benefit diversified PRCs, especially the transportation segment, the transportation segment of diversified PRCs may be less likely to depend on its own free cash flow compared with focused firms (e.g. Tokyo Metro) since ICMs can subsidise it. Following Shin and Stulz (1998), we assume that Tokyo Metro is comparable focused firm to diversified major PRCs. Then, we estimate the investment equation without the free cash flow of other segments and add a dummy variable of Tokyo Metro as well as interaction term of the dummy variable and the firm's free cash flow. We estimate the regressions by using unbalanced panel data. All coefficients are GLS estimates of the random effects model9 and

9 The difference between fixed and random effects is whether unobserved variables are correlated with the regressors. As Greene (2012, p. 387) puts it:Again, the crucial distinction between fixed and random effects is whether unobserved individual effect embodies elements that are correlated with the regressors in the model, not whether these effects are stochastic or not.Although fixed effects investigate the impacts of time-variant variables by removing time-invariant individual characteristics of each entity (e.g. operating area of the PRC), random effects allow timeinvariant differences across entities to play a role as explanatory variables (TorresReyna, 2007). Therefore, we estimate panel data models with random effects and find that the null hypothesis of orthogonality was not rejected by the Hausman test.

Please cite this article in press as: Song, Y.-J., & Shoji, K., Effects of diversification strategies on investment in railway business: The case of private railway companies in Japan, Research in Transportation Economics (2016), http://dx.doi.org/10.1016/j.retrec.2016.07.022

6

Y.-J. Song, K. Shoji / Research in Transportation Economics xxx (2016) 1e9

Table 2 Variable estimation of the investment equation for diversified PRCs: GLS estimates.d Investment sensitivity Total Free cash flow of the transportation segment Free cash flow of other segments Segment's ROA

Kanto c

Kansai c

Total c

0.8850 (19.78) 0.0364 (1.49) 0.0011 (1.57)

0.8777 (18.92) 0.0520a (1.74) 0.0006 (0.93)

0.9000 (8.00) 0.0473a (1.74) 0.0017 (1.16)

0.0284c (4.29) 0.0113c (2.65) 228 (18) 0.8987

0.0273c (3.62) 0.0156c (3.46) 117 (9) 0.9534

0.0303 (2.21) 0.0069 (0.99) 111 (9) 0.8112

Herfindahl index Sales share of the transportation segment Constant Observations (Groups) R2 a b c d

Kanto c

0.8797 (19.60) 0.0501b (2.31) 0.0011 (1.58) 0.0373 (1.32) 0.0415b (2.47) 0.0188 (0.71) 228 (18) 0.9098

Kansai c

0.8781 (18.87) 0.0510a (1.90) 0.0006 (0.91) 0.0022 (0.10) 0.0277b (2.02) 0.0167 (0.78) 117 (9) 0.9533

0.9070c (10.77) 0.0671b (2.12) 0.0010 (1.05) 0.0860b (2.55) 0.0690b (2.34) 0.0647a (1.86) 111 (9) 0.8779

Significant at the 10% level. Significant at the 5% level. Significant at the 1% level. The variables using price such as capital expenditure and cash flows are deflated by using a GDP deflator.

calendar-year dummies for each year, and the heteroscedasticity robust t-statistics are reported in parentheses (the coefficients of calendar-year dummies have been omitted because of space restrictions). Table 2 presents the estimates of the baseline investment equation and the model including the diversification index, which implies the extent of diversification. In the estimation of the total sample, the coefficients of the free cash flow of the transportation segment are statistically significant in all models and the absolute value of the coefficients of the transportation segment's free cash flow is much larger than that of the free cash flow of other segments. However, the signs of the coefficients of the free cash flow of the transportation segment and other segments are negative. This fact implies that an increase in the free cash flow of the transportation segment does not seem to be connected with additional investment in this segment. Further, segment ROA, which is a proxy for investment opportunities, is not significant in all equations, and this is a noticeable difference to that for non-regulated firms. As discussed above, investment opportunities affect the decisions of ICMs and are positively related to a segment's investment. This finding implies that investment in the transportation segment is unlikely to depend on the profit decision completely, in contrast to a non-regulated firm's segments. In the model including diversification index, its coefficient is

only positive and statistically significant in the Kansai subsample. This finding suggests that deploying a diversification strategy positively influences transportation investment under in specific circumstances when controlling cash flows, investment opportunities and other factors. In addition, an increase in the extent of diversification has about twice the impact of an increase in the extent of diversification, although its coefficient is not statistically significant. Although the passenger KM in the PRCs operating in urban areas of Japan is decreasing, the Kansai region shows a steeper decline than the Kanto region in recent years. Thus, expanding the extent of the diversification strategy seems to positively affect investment in the transportation segment if PRCs experience a sharp decrease in the number of railway passengers. Table 3 presents the results of the estimation of the investment equation for diversified PRCs and single-segment firm (Tokyo Metro). The results show that Tokyo Metro tends to invest in the transportation business more than diversified PRCs because the coefficient of dummy variable of Tokyo Metro is positive. In addition, the interaction term of the dummy variable of Tokyo Metro and free cash flow has a positive coefficient. This finding implies that a focused firm is more sensitive to the free cash flows than diversified PRCs; however, generated additional cash tends not to be used for investment in the transportation segment.

Table 3 Variable estimations of the investment equations for diversified PRCs and Tokyo Metro: GLS estimates. Dependent variable: Investment sensitivity

Coefficients

Segment's ROA

0.0010 (1.51) 0.8888c (20.76) 0.0550c (5.20) 0.4257a (1.73) 0.0246c (8.65) 236 (19) 0.8970

Free cash flow of the transportation segment Tokyo Metro dummy Tokyo Metro dummy  Free cash flow of the transportation segment Constant Observations (Groups) R2 a b c

Significant at the 10% level. Significant at the 5% level. Significant at the 1% level.

Please cite this article in press as: Song, Y.-J., & Shoji, K., Effects of diversification strategies on investment in railway business: The case of private railway companies in Japan, Research in Transportation Economics (2016), http://dx.doi.org/10.1016/j.retrec.2016.07.022

Y.-J. Song, K. Shoji / Research in Transportation Economics xxx (2016) 1e9

6. Discussion We interpret our estimation results of the investment equation of PRCs and Tokyo Metro in Tables 1 and 2. These tables indicate that the investment policy of these companies differs from that of the firms operating in non-regulated industries. In particular, we confirm that it is hard to expect the free cash flow generated in other segments to be transferred to the transportation segment; indeed, even the free cash flow of the transportation segment is unlikely to be invested in transportation. Regarding a focused firm, this tendency is even noticeable from the estimation results. In addition, expanding the extent of diversification strategies can positively influence transportation investment when the number of passengers changes rapidly, although we cannot ensure that the increase of investment results from internal cash. In addition to free cash flow, investment opportunities are thus not a critical factor influencing investment in the transportation segment. First, the operating profit of the railway business was positive throughout the analysis period, allowing it to generate the free cash necessary for investment, although some diversified segments showed a negative operating profit. While some PRCs have negative free cash flow in the transportation business in some fiscal years, the average free cash flow of the transportation segment is positive, suggesting that it can generate the cash flow available for investment. Such investment is not only for maintaining the current state, but also for expansion and large-scale improvement. In addition, in the estimation results including Tokyo Metro, free cash flow is unlikely to be used for the transportation segment, even in Tokyo Metro. Hence, if someone expects PRCs to benefit from diversification by building ICMs and utilising internal capital for transportation investment when a positive free cash flow is generated, this result is contrary to the expectation that the internal funds of diversified firms may contribute to transportation investment. Evidence from previous research also supports this result. Thompson (1999) examines the growth in the non-core activities (deploying businesses except utilities) of newly privatised UK utilities and finds that they tend to focus on expansion through diversification given the generous regulation and limited opportunities for growth of the core business (utility business). In addition, the estimation results suggest that the free cash flow of other segments has little impact on investment in the transportation segment. Thus, the estimation results in Tables 2 and 3 imply that ICMs hardly contribute to transportation investment by using free cash flow. The presented findings are inconsistent with previous research regarding the influence of ROA, which is a proxy of investment opportunities. Although the coefficient of the ROA of the transportation segment is positive, it is not statistically significant and its impact is slight. This finding implies that investment opportunities emphasising profitability, an influential factor affecting investment decisions, are not critical to deciding investment in the transportation segment compared with other factors. We assume that the slight impact and insignificance of ROA may come from the social responsibility placed on the railway business (i.e. duty to achieve safe operations). PRCs play the role as provider of public transportation service by maintaining safety of railway operation and consider social losses from accident and operation trouble, while the government also encourages PRCs to enhance safety improve user environment by supporting some of the cost.10

10

The government partially supports projects such as ensuring earthquake resistance, setting a barrier-free environment in railway stations, and improving railway station facilities. For details, see the report of The Association of Japanese Private Railways (2015).

7

As the reason free cash flow does not contribute to investment in the transportation segment, the limited profitability and growth of the railway business are also considered. As mentioned in Section 1, all railway operators including major PRCs face fare regulation. The full cost principle has been applied to the fares of JR companies. In particular, the rate-of-return regulation applied to major PRCs and subways (Mizutani, 2014a, 2014b).11 Fares based on rate-of-return regulation may induce the firm to overinvest. However, a ceiling price for the railway fare of each operator is set based on its cost structure and this is approved by the government (Mizutani, 2014a). Despite concerns about the over-assessment of cost, yardstick regulation is applied and the government evaluates the performance of operators by using common measures such as average cost. Quantitative evidence supports the effectiveness of yardstick regulation (Mizutani, Kozumi, & Matsushima, 2009). Finally, on some routes (e.g. from Kobe to Osaka in the Kansai region), the competition between JR company and major PRCs make it hard to increase fares up to the ceiling price. In fact, while railway fares have increased recently following the increase in the consumer tax rate, PRCs have not revised their fares drastically because of the degree of market competition. Consequently, the present fare system does not fully consider future investment in the railway business, as Shoji (2001) points out.12 Therefore, limited profitability owing to fare regulation may be an obstacle that prevents PRCs from investing in the transportation business using free cash flow. The operating profit of the transportation segment in major PRCs is sufficient to generate free cash flow owing to stable demand from large commuter and high population density in urban areas. Moreover, Lamont (1997) shows that investment in non-core segments depends on the cash of core businesses in the case of oil companies. Thompson (1999) also suggests that a firm whose core businesses are stable cash generators, but market-matured and regulated by price regulations, has few real opportunities to grow its core businesses and tends to invest in other businesses that are expected to achieve growth in the case of newly privatised UK utilities. As presented in Fig. 2, it is hard to expect future growth in the railway business because of the declining population. In particular, the decrease in the productive age population is related to a decline in railway ridership; thus, PRCs cannot ensure stable operation because of the reduction in railway business revenue. In this context, major PRCs are expanding their business portfolios into various areas beyond their traditional diversified businesses, such as housing development projects in foreign countries of Tokyu, to explore new income sources and thereby enhance the firm's financial sustainability. However, compared with the transportation segment, the free cash flow of other segments is negative on average. Thus, firms may rely on funds from the transportation segment to expand diversified businesses. Although the transportation business is mature and can generate additional cash, free cash flow may thus be utilised to expand diversified businesses. Finally, we pay attention to the differences in the operating environments between the Kanto region and the Kansai region. First, productive age population, which generates stable demand for railway companies by commuting, in the Kansai region is shrinking at a greater rate than that in the Kanto region (see Fig. 2). Although the population is decreasing in both regions, the Kanto region shows a smaller drop than the Kansai region. Thus, potential demand for the

11 Except JR companies, major PRCs, and subways, the full cost is calculated as the sum of operating cost, depreciation, interest expense, and reasonable profit for minor PRCs and local railways. 12 For a discussion of the effect of fare regulation on investment incentives and management improvement, see Shoji (2001).

Please cite this article in press as: Song, Y.-J., & Shoji, K., Effects of diversification strategies on investment in railway business: The case of private railway companies in Japan, Research in Transportation Economics (2016), http://dx.doi.org/10.1016/j.retrec.2016.07.022

8

Y.-J. Song, K. Shoji / Research in Transportation Economics xxx (2016) 1e9

Fig. 2. Trend of the productive age population (2005 ¼ 100). Source: (Actual value) National census, Japanese Statistics Bureau and Statistics Centre (2010) (Estimated value) Regional Future Estimated Population in Japan, National Institute of Population and Social Security Research (2013).

railway business is comparatively more stable in the Kanto region. Second, in the Kansai region, three companies (JR, Hankyu, and Hanshin) are competing on the same route from Kobe to Osaka, while there are hardly any competing routes in the Kanto region. As well as potential demand, the existence of competitors on the same route also influences the relationship between diversification and investment in the railway business. When we look at the estimation results shown in Table 2, the coefficient of the diversification index of Kansai subsample is only significant and positive. This finding implies that the extent of diversification strategies influences investment in the transportation business positively, controlling cash flows, investment opportunities, and other factors only in the Kansai region. We infer from the above that the relatively poor prospect for future demand and competitive situation may drive PRCs in the Kansai region to strengthen the railway business by using diversification strategies, even if they already deploy diversification strategies widely. Further, expanding the extent of diversification strategies has a positive impact on the railway business through the ‘more money’ effect (e.g. acquiring more external funds than single segment firms through ICMs) if PRCs in the Kansai region diversified into the businesses enabling to make the railway business more attractive. In terms of diversified businesses, as shown in the estimation results of Table 2, although the free cash flow of the transportation segment and other segments is unlikely to be invested in the railway business because of the use of diversification strategies, PRCs in the Kansai region may diversify to make the railway more attractive, thereby capturing demand. Further, the free cash flows from the transportation segment may be deployed to support these diversified segments. This view is supported by Shoji (2001), who suggests that the proportion of dominated and related types according to Rumelt's categorisation in the Kansai region is higher than that in the Kanto region. Since the Herfindahl index can only capture the effect of increased sales compared with the main business, we cannot confirm whether the direction of the diversification strategies supports the transportation business. In addition, because we do not consider the effect of ICMs on the external capital market, whether the ICMs of the PRCs in the Kansai region acquire more external funds by diversification is also in question. Further research is needed on these issues.

work in major PRCs and whether internal funds generated from transportation segments and other segments can be invested in transportation segments. By estimating panel data models allowing random effects, we demonstrate that free cash flow is unlikely to be invested in the transportation segment, even in the case of a focused firm, although it tends to be invested in the transportation segment more than in diversified PRCs. We also show that diversification has a positive impact on investment in the railway business given the decrease in future demand and competitive market. However, we cannot confirm how the diversification strategies used support investment in the transportation segment and how ICMs work through expanding diversification, except utilising internal capital. These issues are still questionable. One practical implication of our research is that it is difficult for diversified PRCs to utilise internal capital for investment to improve and expand railways. In the case of Japanese PRCs, when the railway business is making a profit and generating cash for investment, this cash tends to be allocated to expand diversification. However, regarding capital investment in railways, Japanese PRCs raise funds from operating income and are financed by private banks or public bank (Shoji, 2008). We interviewed the managers of three major PRCs in the Kansai region and can confirm that most PRCs tend to rely on funds from the Development Bank of Japan13 (e.g. low interest loans) or financial support programmes that operate jointly with the public sector (e.g. the establishment of the quasi-public sector) when they plan to construct or improve infrastructure at a large scale. Japanese PRCs may not utilise internal capital because of the existence of such support schemes, although this is minimal considering that capital investment in railways is supported by funds from subsidies and taxes in Europe and North America. However, considering the low promise of future growth in the railway business, support schemes offered by the public sector, although not large, may serve as insurance to guarantee a certain level of quality in the provision of railway services. Therefore, a political framework (e.g. regulation) that encourages active investment in the railway business using the firm's own cash is also needed given that diversification strategies result from a firm's strategic motivation to improve performance and maximise future growth. Diversifying properly can therefore contribute to creating

7. Conclusions This study examines how diversification influences investment in the railway business, focusing on the capital allocation performed by the firm's headquarters in ICMs and using data on Japanese PRCs. We aim to investigate how ICMs built by diversification

13 The Development Bank of Japan provides long-term loans with low interest to PRCs up to 50% of the cost of construction; however, it only supports specified projects related to disaster prevention, increasing railway capacity, and improving user convenience.

Please cite this article in press as: Song, Y.-J., & Shoji, K., Effects of diversification strategies on investment in railway business: The case of private railway companies in Japan, Research in Transportation Economics (2016), http://dx.doi.org/10.1016/j.retrec.2016.07.022

Y.-J. Song, K. Shoji / Research in Transportation Economics xxx (2016) 1e9

abundant capital resources inside a firm. However, the allocation decision is made by the headquarters, which consider future profitability and growth first. Although private operators have the flexibility and freedom to design service and business strategies, owing to the expected decreasing railway demand, enhancing the railway infrastructure may be overwhelmed by other segments expected to gain more profit in the future. Hence, restricting the diversification of private operators is not a desirable choice. The aggravated environment surrounding passenger railway services, such as competition with individual transportation and decreases in population size, burden railway operators; thus, they need strategies to guarantee their financial sustainability, and diversification may play a role in this regard. Diversification may be an incentive for operators because it opens up the possibility of creating various income sources and seeking further growth opportunities. This means that the firm may overcome the limited profitability of railway services owing to fare regulation by diversifying. Therefore, designing a financial support programme for improving infrastructure and formulating a regulatory scheme to monitor the weak governance systems of private operators, which is related to wasteful investment (Jensen, 1986; Markides, 1995; Mueller, 1972; Thompson, 1999), are needed to minimise the influence on railways. We confirm that the internal capital of PRCs is mainly used to diversify businesses, rather than railway businesses; however, we cannot investigate whether diversification is deployed desirably for future growth and financial sustainability, nor verify the extent of the governance systems of PRCs. Studying the forms of external capital that affect the investment decisions of the railway business such as loans, subsidies, and stocks and examining the support programmes to mitigate the tax rate or interest are also recommended to clarify the relationship between diversification and investment further, when focusing on the ICMs of a firm. These questions are left to future research. Acknowledgments We are grateful for the comments by two anonymous referees and participants at the Thredbo 14 Workshop 7 ‘Market initiative: regulatory design, implementation and performance’. The responsibility for any remaining errors in the authors. References Ahn, S., Denis, D. J., & Denis, D. K. (2006). Leverage and investment in diversified firms. Journal of Financial Economics, 79(2), 317e337. Amihud, Y., & Lev, B. (1981). Risk reduction as a motive for conglomerate mergers. Bell Journal of Economics, 12(2), 605e617. Arrfelt, M., Wiseman, R. M., & Hult, G. T. M. (2013). Looking backward instead of forward: Aspiration-driven influences on the efficiency of the capital allocation process. Academy of Management Journal, 56(4), 1081e1103. Berger, P. G., & Ofek, E. (1995). Diversification's effect on firm value. Journal of Financial Economics, 37(1), 39e65. Billett, M. T., & Mauer, D. C. (2003). Cross-subsidies, external financing, constraints, and the contribution of the internal capital market to firm value. The Review of Financial Studies, 16(4), 1167e1201. Duchin, R. (2010). Cash holdings and corporate diversification. The Journal of Finance, 65(3), 955e992. Faulkender, M., Flannery, M. J., Hankins, K. W., & Smith, J. M. (2012). Cash flows and leverage adjustments. Journal of Financial Economics, 103(3), 632e646. Greene, W. H. (2012). Econometric analysis (7th ed.). Essex, England: Pearson Education Limited. Harris, M., & Raviv, A. (1996). The capital budgeting process: Incentives and information. The Journal of Finance, 51(4), 1139e1174. Hensher, D. A., & Stanley, J. (2008). Transacting under a performance-based contract: The role of negotiation and competitive tendering. Transportation Research Part A: Policy and Practice, 42, 1143e1151.

9

Hovakimian, G. (2011). Financial constraints and investment efficiency: Internal capital allocation across the business cycle. Journal of financial Intermediation, 20(2), 264e283. Jensen, M. (1986). Agency costs of free cash flows, corporate finance and takeovers. American Economic Review, 76, 323e329. Khanna, N., & Tice, S. (2001). The bright side of internal capital markets. The Journal of Finance, 56(4), 1489e1528. Killeen, J. B. (1999). Business diversification and integration strategy of urban private railways in Japan: An economic analysis of strategy, market performance and policy implications. Doctor of Commerce Thesis. Graduate School of Business Administration: Kobe University. Lamont, O. (1997). Cash flow and investment: Evidence from internal capital markets. The Journal of Finance, 52(1), 83e109. Lang, L., Ofek, E., & Stulz, R. M. (1996). Leverage, investment, and firm growth. Journal of Financial Economics, 40(1), 3e29. Lang, L., & Stulz, R. M. (1994). Tobin's Q, corporate diversification, and firm performance. Journal of political economy, 102(6), 1248e1280. Maksimovic, V., & Philips, G. (2008). Conglomerate firms and internal capital markets. In B. E. Eckbo (Ed.), Handbook of corporate finance: Empirical corporate finance (pp. 423e479). North-Holland: North-Holland. Maksimovic, V., & Phillips, G. (2002). Do conglomerate firms allocate resources inefficiently across industries? Theory and evidence. The Journal of Finance, 57(2), 721e767. Markides, C. (1995). Diversification, restructuring and economic performance. Strategic Management Journal, 16(2), 101e118. Mizutani, F. (2014a). Looking beyond Europe with a special focus on Japan. Discussion paper series, 2014-24. Graduate School of Business Administration. Mizutani, J. (2014b). Yardstick regulation in Japan's railway industry. Transport Policy Studies' Review, 17(2), 20e27 (in Japanese). Mizutani, F., Kozumi, H., & Matsushima, N. (2009). Does yardstick regulation really work? Empirical evidence from Japan's rail industry. Journal of Regulatory Economics, 36(3), 308e323. Mueller, D. (1972). A life cycle theory of the firm. The Journal of Industrial Economics, 20(3), 199e219. Peyer, U. (2002). Internal and external capital markets. s.l.:Available at: SSRN 309746. Rajan, R., Henri, S., & Luigi, Z. (2000). The cost of diversity: The diversification discount and inefficient investment. The Journal of Finance, 55(1), 35e80. Saito, T. (1993). Private railways: Development of Japanese style railway management (Shitetsu sangyo: Nihongata tetsudo keiei no tenkai). Kyoto: Koyo Shobo (in Japanese). Scharfstein, D. S. (1998). The Dark side of internal capital markets Ⅱ: Evidence from diversified conglomerates, Working Paper 6352. National Bureau of Economic Research (NBER). available at: http://www.nber.org/papers/w6352. Scharfstein, D. S., & Stein, J. C. (2000). The dark side of internal capital markets: Divisional rent-seeking and inefficient investment. The Journal of Finance, 55(6), 2537e2564. Shin, H. H., & Stulz, R. M. (1998). Are internal capital markets efficient? The Quarterly Journal of Economics, 113(2), 531e552. Shoji, K. (2001). The policy of urban public transportation (Toshi Koukyou Koutsu Seisaku) (in Japanese). Tokyo: Chikura Shobo. Shoji, K. (2008). A note on urban public transport policy in Japan: Financing and funding structure of urban rail. Journal of Economics and Business Administration, 198(2), 21e38 (in Japanese). Stein, J. C. (1997). Internal capital markets and the competition for corporate resources. The Journal of Finance, 52(1), 111e133. Stein, J. (2002). Information production and capital allocation: Decentralized versus hierarchical firms. The Journal of Finance, 57(5), 1891e1921. Stein, J. C. (2003). Chapter 2-Agency, information and corporate investment. In G. M. Constantinides, M. Harris, & R. M. Stulz (Eds.), Handbook of the economics of finance (pp. 111e165). North Holland: Elsevier. The Association of Japanese Private Railways. (2015). The databook of the major private railways, Japan. The Association of Japanese Private Railways. Thompson, S. (1999). Increasingly marginal utilities: Diversification and free cash flow in newly privatized UK utilities. Review of Industrial Organization, 15(1), 25e42. Torres-Reyna, O. (2007). Panel data analysis fixed and random effects using stata (v. 4.2), s.l.: [online] available from: https://www.princeton.edu/~otorres/Panel101. pdf (Accessed 18 May 2016). Villalonga, B. (2004). Diversification discount or premium? New evidence from the business information tracking series. Journal of Finance, 59(2), 479e506. Whited, T. M. (2001). Is it inefficient investment that causes the diversification discount? The Journal of Finance, 56(5), 1667e1691. Wulf, J. (2002). Internal capital markets and firm-level compensation incentives for division managers. Journal of Labor Economics, 20(S2), 219e262. Xuan, Y. (2009). Empire-building or bridge-building? Evidence from new CEOs' internal capital allocation decisions. Review of Financial Studies, 22(12), 4919e4948. Yvrande-Billon, A. (2006). The attribution process of delegation contracts in the French urban public transport sector: Why competitive tendering is a myth. Annals of Public and Cooperative Economics, 77(4), 453e487.

Please cite this article in press as: Song, Y.-J., & Shoji, K., Effects of diversification strategies on investment in railway business: The case of private railway companies in Japan, Research in Transportation Economics (2016), http://dx.doi.org/10.1016/j.retrec.2016.07.022