DIVERSIFICATION EFFECTS: A REAL OPTIONS APPROACH

Transcription

1

2 DIVERSIFICATION EFFECTS: A REAL OPTIONS APPROACH A dissertation submitted to the Kent State University Graduate School of Management in partial fulfillment of the requirements for the degree of Doctor of Philosophy by Aiwu Zhao November 2008

3

4 Dissertation written by Aiwu Zhao B.S., Tsinghua University, 1993 M.A., Kent State University, 2003 Ph.D., Kent State University, 2008 Approved by Chair, Doctoral Dissertation Committee Members, Doctoral Dissertation Committee Accepted by Doctoral Director, Graduate School of Management Dean, Graduate School of Management ii

5

6 ACKNOWLEDGEMENTS So many people make a dissertation possible, and I am grateful to all those who have helped and supported me. I want to express my deep gratitude to my advisor, Dr. Mark Holder, for his insightful guidance and constant encouragement. Without him, completing of this dissertation would not have gone smoothly. I would also like to thank other members in my committee: Dr. Richard Kent, whom I respect for his vast knowledge and thoughtfulness, Dr. Emmanuel Dechenaux, whose objective evaluation on my work greatly benefited me, and Dr. Paul Dawson and Dr. James Boyd, whose detailed comments and solid scholarship helped me continuously. My special thanks also go to Dr. Michael Hu, Dr. Jayaram Muthuswamy, and Dr. Marvin Troutt for their help in my moment of need. I am also deeply thankful to my family. I am indebted to my dear husband. It is his love that has supported me this far. Finally, my heartfelt thanks should go to all those who have helped and encouraged me during this process. Without them, it would have been much more difficult. iii

7 TABLE OF CONTENTS Chapter 1: Introduction... 1 Chapter 2: Literature on Diversification as Value-Reducing Strategy Early Evidence and Inefficiency Explanation Inefficiency Explanation Other Explanations of Costs of Diversifications Agency Cost Explanation Counter Arguments to Value Destroying Explanations Chapter 3: Literature on Diversification as Rational Strategy Discount Reflects Pre-existing Characteristics Real Options Explanation Previous Examination on Value Changes around Diversification Chapter 4: Data, Methodology and Research Design Hypotheses Define Variables Dependent Variable Resource-Based View Related Explanatory Variables Real Options Theory Related Explanatory Variables Control Variables Empirical Methodology Data Chapter 5: Empirical Results Descriptive Statistics Value Changes around Diversification Regression Tests Chapter 6: Conclusion References iv

8 LIST OF TABLES Table 1 Definition of variables.. 39 Table 2 Number of observations 45 Table 3 Number of firms that add new segments ( ). 46 Table 4 Number of firms that reduce number of segments ( ).. 47 Table 5 Descriptive statistics of focused-to-diversified firms market-to-book ratios one year before diversification 49 Table 6 Industry level comparison of numbers of above and below industry median firms that diversified. 50 Table 7 Overall market level M/B ratios 51 Table 8 Industry level comparison of differences in M/B ratios between diversified and non-diversified firms. 52 Table 9 Descriptive statistics of firm level variables. 53 Table 10 Value changes of focused-to-diversified firms around diversification. 57 Table 11 Adjusted value changes of focused-to-diversified firms around diversification 59 Table 12 Comparison of value changes between above and below industry average focused-to-diversified firms. 60 Table 13 Value changes of focused-to-diversified firms around diversification after adjusting industry trends Table 14 Value changes of related vs. unrelated diversification activities.. 62 Table 15 Logit model on the possibility of having positive or negative value change in different diversification scenarios.. 64 Table 16 Differences in M/B ratios between focused-to-diversified firms and Table 17 industry average firms OLS regressions testing the resource-based approach vs. real options approach 70 Table 18 Explanatory power of the RBV and RO models Table 19 Regression analysis using dummy variables. 75 Table 20 Detecting multicollinearity by VIF Table 21 Regressions with reduced number of variables. 79 v

9 Chapter 1 Introduction Empirical studies have shown that diversified firms trade at a discount compared to stand-alone firms (Lang and Stulz, 1994; Berger and Ofek, 1995; Lins and Servaes, 1999; Shin and Stulz, 1998). This phenomenon is termed as the diversification discount. Early studies usually explain the phenomenon with the conclusion that diversification destroys value because diversified firms tend to show inefficiency in asset allocation, in management capability, or in other measures of operational efficiency. However, such an argument does not explain why firms diversify if diversification is ex ante inefficient. Recent studies try to find an answer to the puzzle and have shown that diversification discount does not necessarily reflect value-reducing effects related to declining operational efficiency. Gomes and Livdan (2004) suggest that diversification is a rational choice for firms with declining or no growth in existing business lines to explore new opportunities. They find that firms with lower values are more likely to diversify, so a cross-sectional discount related to diversified firms is not sufficient to show whether diversification activities would increase or decrease firm value, because the lower value might be mostly driven by pre-diversification characteristics of the firms. The commonly used measures in previous cross-sectional studies on diversification effects are market multipliers such as market-to-book (M/B) ratio, which is usually calculated as the sum of market value of equity and book value of debt divided by the book value of assets. The value measures are considered better than performance measures such as profitability ratios or market returns when investigating diversification effects, because they reflect future cash flow s impact by using capitalized value (Lang

10 2 and Stulz, 1994). Besides being used as value measures, market multipliers have also been used widely as proxies for investment opportunities and growth potentials. For example, high market-to-book (M/B) ratio firms are considered growth firms, which are firms with high growth opportunities. One important piece of information reflected in market-to-book (M/B) ratio is the value of real options. Based on Myers (1977), who initiated the concept of growth options, market value has two components: growth options and present value of assets-in-place, which is usually proxied by the book value of the firm. Investors evaluate the market value of a firm by discounting all future possible income, so the growth potentials of a firm are incorporated in its market value even though these potentials have not been materialized as real assets investments (De Andrés-Alonso, Azofra-Palenzuela, and De La Fuente-Herrero, 2005). A growth firm will naturally turn into a mature firm with lower market multipliers after exploiting a series of growth opportunities and investing in real assets. Since multiple things can affect value measures such as market-to-book (M/B) ratio; as a result, even though decreased operational efficiency may lead to lower value measures, the converse is not necessarily true. A firm that operates its existing assets efficiently with a high return-on-assets ratio can still have relatively low market-to-book (M/B) ratio because of the lack of new growth opportunities. Previous diversification studies only focus on operational efficiency changes related to diversification activities and perhaps have not paid enough attention to the change in the future growth opportunities implied in value measures. Diversification activities, which are associated with real asset investments, are strategic decisions that can change the growth potentials of a firm and will create value impacts that are different

11 3 from those caused by changes in operational efficiency. For example, when a firm adds a new line of business that is similar to its existing business, it may extract excess value from its unique resources because of economies of scale. But heavy investment in closely related business sectors makes it harder to switch gears if the market of existing business deteriorates, because the firm now faces more transaction costs than before if it decides to change the direction of its business. While operational efficiency is the focus of previous diversification studies, our study will concentrate on the value impacts related to changes in future opportunities and try to differentiate between the two arguments. In our study, we formalize the two groups of arguments into their corresponding theoretical paradigms. Different predictions on the change of value measures are generated from these two groups of arguments. These contrasting predictions are the basis of the conceptual framework of this study. The argument that focuses on the value impact of operational efficiency follows the resource-based view of firm and transaction cost economics, which are the dominant theories on diversification in the strategy literature (Villalonga and McGahan, 2005). The resource-based view argues that the value of excess capabilities in resources such as superior production and managerial skills, patents, marketing abilities, and consumer goodwill can be enhanced through economies of scale. Under this framework, unrelated diversification processes will generate more transaction costs than related diversification processes, because it is usually more difficult to transfer the resources to a different segment than to a similar segment. Therefore, unrelated diversification will be less beneficial or even create negative impacts to firm value when compared to related diversification.

12 4 The argument that focuses on the value impact of growth potentials follows the real options theories in finance literature. The real options approach applies financial options theory to real assets investments. Financial options give the owner the right, but not the obligation, to buy or sell a security at a given price. Companies holding real options have the right, but not the obligation, to exploit different opportunities in the future (Rappaport and Mauboussin, 2001). Similar to financial options, the value of real options increases with the uncertainty level of the investment environment and with the time horizon to delay the investment, because uncertainty and longer waiting period make it more valuable to hold more alternative choices. Any exercise of potential opportunities will end the time value of the options. Under the real options framework, related diversification tends to reduce such market multipliers as market-to-book (M/B) ratio, because a related diversification is a process for a firm to materialize its potential growth opportunities. In related diversification, the prior portion of the firm s growth potentials reflected in market value will turn into assets-in-place, leading to a lower market multiplier ratio (Bernardo and Chowdhry, 2002). On the other hand, from the real options approach, unrelated diversification is expected to have a more positive value impact than related diversification, a prediction in opposition to that of the operational efficiency argument, because adding a segment with a business that is different from existing industry may add growth options to a new area. For firms with closely related segments, the value of growth options to new areas will be much lower, because it is more costly to enter than to expand.

13 5 This dissertation proceeds as follows. In the next section, we review the literature that links the diversification discount to operational inefficiency. In Section III, we move to studies that question the inefficiency argument and outline the theoretical framework of the real options argument. In Section IV, we present the methodology applied in our study, with empirical evidence reported in Section V. Section VI concludes the paper.

14 6 Chapter 2 Literature on Diversification as Value-Reducing Strategy There is no unified definition of diversification in previous empirical studies. Some studies examine diversification effects by focusing on merger and acquisition activities (Morck, Shleifer, and Vishny, 1990; Hubbard and Palia, 1999; Matsusaka, 1993). There are also studies using the number of segments reported by the firms to decide the companies diversification levels (Lang and Stulz, 1994; Berger and Ofek, 1995; Lins and Servaes, 1999; Shin and Stulz, 1998). A business segment is considered an enterprise component that provides distinguishable product or service and has distinguishable production process. The quantitative thresholds set up by the Financial Accounting Standards Board (FASB) for segment identification are 10% of assets, revenues, or profit/loss of the overall enterprise. Studies based on reported segment data usually do not take into consideration how the new segment is created. We will review both groups of literature on diversification effects. We define that a company has a related diversification when it reports a new segment or acquires a new business line whose first two digits of the Standard Industrial Code (SIC) are the same as those of its existing segments (Holder et al. 1998). Unrelated diversification means the first two digits of the SIC code of the new segment are different from those of its existing segments. Firm expansion only refers to increases in investment within existing segments and no new segment or business line is added.

15 7 2.1 Early Evidence and Inefficiency Explanation The conventional rationale for diversification is to benefit from internal capital markets and synergy. The role of internal capital markets is considered especially important for firms operating in emerging and lesser-developed countries, where external financial markets are not very efficient (Khanna and Palepu, 2000), because the internal market generated within a diversified company can mitigate the inefficiency of external market by allocating more resources internally (Weston, 1970). In developed countries where specialized market exchange mechanisms are available, resources will be allocated more efficiently through external markets. As a result, the benefit of internal capital markets becomes a less important reason for diversification with the development of financial markets. But diversified firms still enjoy lower financing cost than focused firms. Especially, diversified firms can borrow at much lower costs compared to focused firms (Deng et al., 2007; Reeb et al., 2001; Penas and Unal, 2004), because besides being usually bigger in size than stand-alone firms, diversified firms may also use the multiple segments to generate a coinsurance effect to lenders and increase debt capacity (Lewellen, 1971). But lower financing cost is not the underlying force that drives firm value; much of the research regarding the impact of diversification on firm value focuses more on the benefits of synergy from diversification. Both business strategy and modern corporate literature has developed extensive theoretical paradigms in this area. The dominant theory on diversification in strategy literature is the resource-based view (Villalonga and McGahan, 2005; Wernerfelt, 1984; Barney, 1991). The resource-based view argues that the value of excess capabilities in resources such as superior production and managerial

16 8 skills, patents, marketing abilities, and consumer goodwill can be enhanced through economies of scale. A similar idea in the modern corporate model is the concept of organizational capabilities (Chandler, 1990). The corporate organization view posits that organizational capabilities such as marketing, distribution, and development skills of top and middle management are transferable across products and industries and are a source of company value (Matsusaka 2001). However, empirical evidence from early studies indicates that benefits from diversification can be hard to achieve. Diversified firms are usually found to be traded at a discount compared to stand-alone single segment firms. Some empirical finance studies draw this conclusion by using Tobin s q as the value measure. Tobin s q, by definition, is the present value of future cash flows divided by the replacement cost of tangible assets. Tobin s q theoretically measures the contribution of a firm s intangible assets, such as organizational and reputational capital, monopolistic rents, and investment opportunities to firm value. Intangible assets can have both positive and negative values. For example, managerial entrenchment is viewed as a negative value intangible asset. Lang and Stulz (1994) introduce Tobin s q in diversification studies in finance literature to investigate whether a firm s degree of diversification will affect its market valuation. Previous diversification studies usually use performance measures, such as profitability ratios or market returns. Lang and Stulz (1994) indicate that results based on accounting or stock market performance are affected by the chosen sample period and are also difficult to interpret in terms of poor performance. When using accounting measures, it is difficult to decide the appropriate tradeoff between current and future performance. The choice to use one-year or five-year performance measures is arbitrary. When using stock price as a

17 9 performance measure, firm performance needs to be adjusted for risk. Otherwise, firms can be considered to perform better in terms of stock return just because they have higher risk. Lang and Stulz (1994) suggest that using Tobin s q avoids these problems because Tobin s q uses the capitalized value, which reflects future cash flow s impact. In addition, Tobin s q does not require risk adjustments for firm comparison. In empirical studies, researchers have developed different methods to compute Tobin s q. Lang and Stulz (1994) calculate Tobin s q as the sum of market value of common stocks, the book value of debts and preferred stocks divided by the sum of the book value of assets other than plant, equipment, and inventories and an estimated replacement cost of plant, equipment, and inventories. However, the accuracy of this calculation method has been questioned. Berger and Ofek (1995) argue that calculating Tobin s q needs to estimate replacement value based on an assumption about the rates of depreciation and inflation, which have large variations among different industries. But industrial level adjustment for replacement costs in each segment is impossible from the available data. Therefore, Berger and Ofek (1995) use market multipliers, such as market-to-book, market-to-sale, and market-to-ebit (earning before interest and taxes) in their study. Despite the differences in value measures, empirical studies usually show similar results (Lang and Stulz, 1994; Berger and Ofek, 1995; Lins and Servaes, 1999; Shin and Stulz, 1998). Both Lang and Stulz (1994) and Berger and Ofek (1995) find that diversified firms trade at a discount of about 15% compared to stand-alone firms. Most of the following studies on diversification effects follow Berger and Ofek s (1995) method and use market multipliers as the value measures.

18 Inefficiency Explanation Based on the above observation, different explanations of the costs of diversification have been studied. The negative impacts from diversification usually are focused on cross-subsidization to unprofitable lines and higher management costs due to information asymmetry. Berger and Ofek (1995) attribute the value loss to cross-subsidization from better-performing segments to poor segments. They find that diversified firms tend to over-invest in segments with poor investment opportunities. They argue that the existence of multiple business lines makes it possible for the firm to allocate too many internal resources to under-performing sectors or even a negative equity sector, reflecting an inefficient use of resources. These under-performing segments should likely have been closed to let their resources flow to more efficient segments. Shin and Stulz (1998) find similar evidence showing internal capital markets may bring more value loss than value increase. Because managers in poorly performing segments have strong incentive to maintain their positions and to lobby the top management; the opportunity cost to these managers of taking away from productive work to engage in lobbying is lower (Scharfstein and Stein, 2000). Thus the relative capital expenditure of a segment will not solely depend on whether it has better investment opportunities compared to other segments. Rajan, Servaes, and Zingales (2000) share the same view that the relative value of a segment will not determine the allocation of resources, and they point out that the easiness of channeling capital resources within the internal capital market will aggravate the agency cost problem. They create an influence cost model based on Coase s (1937) theory that decisions within a hierarchical organization are based on power

19 11 considerations rather than relative prices. They posit that the greater the diversity of resources and opportunities among segments, the more difficult it is to make resource allocation decision only based on segment value or performance. There will be more chances for poor performing segments to get resources within a diversified environment, because more factors other than performance are available to help under-performing segments increase their lobbying power, whereas performance evaluation will play a more important role when segments operate in similar business areas. Harris et al. (1982) stress that when only division managers know the productivity of the resources in their divisions, the information asymmetry problem will give rise to inefficient resource allocation. Following this mechanism, Rajan et al. (2000) indicate that firms diversified into less similar multiple business lines will be more likely to experience over-investment or undertake negative net present value (NPV) projects than firms with similar multiple business lines. Evidence on increased share value with reduction in diversification level has also been used to indicate that diversification destroys value. Daley, Mehrotra, and Sivakumar (1997) observe that spin-offs involving segments in different industries add more value than those occurring within the same industry. Desai and Jain (1999) investigate threeyear abnormal returns after spin-offs and find that the abnormal returns of spin-offs that increase firm focus are 47% higher than those that do not increase firm focus. John and Ofek (1995) also observe improved operating performance following asset sales that increase firm focus.

20 Other Explanations of Costs of Diversifications Though Lang and Stulz (1994) establish a convention in finance literature to employ Tobin s q as a measure of diversification effects, their study is not the first one to use Tobin s q to evaluate diversification effects. Montgomery and Wernerfelt (1988) in economics literature also apply Tobin s q in their diversification study. In their research, Montgomery and Wernerfelt focus on the decrease in the value of marginal economic rents during the diversification process. The economic or Ricardian rents that can be extracted from the firm s inimitable factors are similar to the transferable resources defined in the resources-based view and the organizational capabilities in modern corporate model. Diversification is a process for a firm to exploit the excess capacity of these unique factors. The concept of rent-extraction through diversification takes both the internal market in finance literature and resource-based view in the strategy literature into consideration. On one hand, Ricardian rents only accrue to the owners of unique factors. This argument is similar to the idea of internal markets in that both are referring to resources that cannot be obtained from external markets efficiently. On the other hand, because the uniqueness of the resources will decrease when the company enters into sectors farther away from its original business, Montgomery and Wernerfelt posit that the competitive advantage of the unique resources will decrease with increase in diversification level and there is a negative relationship between the extent of diversification and firm value. Such negative relationship is consistent with the conclusion based on the resource-based view. As a result, the company will realize lower marginal rent and lower value with increases in diversification level.

21 13 There are similar arguments on the negative relationship between the extent of diversification and firm value in strategy literature as well. Many studies indicate that whether diversification increases or decreases firm value depends on the relative size of synergies as well as the implementing costs. Some strategy studies find that related diversification shows superior performance, in terms of return-on-assets (ROA) or returnon-sales (ROS), over unrelated diversification (Bettis, 1981; Markides and Williamson, 1994, 1996), because in a more diversified environment, the synergy implementing process also becomes more difficult. There is international diversification evidence supporting the resource-based view and rent-extraction model as well. Morck and Yeung (1991) show that firm value, measured by Tobin s q, will increase with its multinational scale only when the firm has high level of spending in research and development (R&D) and in advertising, indicating that unique resources such as these intangible assets are necessary for diversification to increase firm value. Morck and Yeung call the argument internalization theory, meaning a firm can increase its value by internalizing markets for such intangible assets as superior production, marketing, or managerial skills, patents, and consumer goodwill. This idea is similar to economies of scale in the resource-based view. 2.2 Agency Cost Explanation Early empirical evidence that has identified the costs of diversification considers the diversification discount a result of costs outweighing benefits and concludes that diversified firms are less efficient than focused firms. However, the inefficiency argument falls short of explaining why firms still choose to diversify if diversification is

22 14 ex ante inefficient. Villalonga (2004a) indicates that from 1990 to 1996, the number of firms that diversified is almost the same as the number of firms becoming more focused. Some studies apply agency cost theory and suggest that private benefits to managers, such as empire building, are the driving motive for diversification (Baurnol, 1967). Shleifer and Vishny (1989) also indicate that diversification can help managers extract more compensation as the assets amount under their management increases. Increase in assets also reduces the probability for the management to be replaced because it is more difficult to acquire a big firm than a small firm (Shleifer and Vishny, 1989; Amihud and Lev, 1981). However, the entrepreneurship-related motivation does not necessarily lead to management behaviors that will destroy the firm value. Stein (1997) indicates that if top managers are motivated by entrepreneurship, they should make management decisions that will increase the efficiency and the value of firm, because the value of their entrepreneurship is consistent with the value of the enterprise. It is hard to justify that doing an inefficient job in asset allocation will be helpful for building entrepreneurship. As illustrated in previous literature, inefficient asset allocation is the result of the agency problem related to the managers in poorly performed divisions rather than to the top management. A higher level of managerial ownership should encourage the top managers to make efforts to reduce the agency costs generated by division managers. Scharfstein (1998) finds the inefficient asset allocation problem is reduced when managerial ownership level is higher. Palia (1999) finds that the diversification discount is reduced when the diversified firm has better corporate governance structures such as stronger management pay-performance sensitivity and smaller board size. In addition, Yang

23 15 (2006) finds that firms with weaker corporate governance are not more likely to diversify than firms with stronger governance, and they do not lose more value around diversification either. Therefore, empirical evidence suggests that agency costs may explain the value decreases in some diversification activities, but not the overall diversification discount phenomenon. 2.3 Counter Arguments to Value Destroying Explanations Besides the difficulty of agency arguments in justifying the motivation for diversification, the conventional explanation of the value destroying impact of diversification has also been challenged by some recent studies that question the validity of diversification discount and its related explanation. A serious challenge comes from the flaws in research method. Whited (2001) points out that there are measurement errors in using Tobin s q to proxy investment opportunities. Whited argues that previous studies calculate Tobin s q as the market value of the firm divided by its replacement value and use it to proxy investment opportunities. But the macroeconomic literature (Lucas and Prescott, 1971; Hayashi, 1982) indicates that the proper measurement for investment opportunities in intertemporal models should be marginal q, an unobservable quantity that reflects the expected present value of marginal product of capital, and the observable Tobin s q, which is an average value measure, can deviate from the unobservable marginal q considerably. When examining whether funds have been transferred across divisions efficiently, investment opportunities related to each division need to be identified accurately. The argument questioning the validity of using Tobin s q as the investment opportunity proxy

24 16 undermines the theoretical foundation that diversification destroys value because of inefficiency in asset allocation. After estimating a measurement-error and correcting the proxy problem, Whited (2001) observes no evidence of inefficient allocation in internal capital market. Another group of studies questioning the conventional wisdom on value destroying effect of diversification casts doubt on the comparability between single segment firms and segments within a conglomerate. Villalonga (2004b) documents that the average size of segments in diversified firms is larger than that of focused firms in the same industry. Following Lang and Stulz (1994), most of the conventional diversification studies usually calculate the excess value of diversified firms as the difference between the Tobin s q of a diversified firm and the weighted average q of its segments as if the q of each segment were the average q of the focused firms in its industry. But because it is widely documented that asset size and M/B ratio are inversely related, the q ratios assigned to the segments of a conglomerate are upwardly biased. As a result, the documented cross-sectional diversification discount is likely exaggerated. Maksimovic and Phillips (2002), in their profit-maximizing neoclassical model, predict that conglomerate firms perform differently from single-segment firms when they allocate resources across segments over the business cycle and when they respond to industry shocks. Therefore, research methods comparing conglomerates with focused firms are theoretically flawed.

25 17 Chapter 3 Literature on Diversification as Rational Strategy 3.1 Discount Reflects Pre-existing Characteristics Though recent evidence is not able to overturn the value discount phenomenon associated with diversification activities, the embedded shortcoming in the matching method used in previous studies challenges the view that the discount is caused by inefficient management of diversified firms. As more studies are carried out to fine-tune the investigation on the difference between diversified firms and single segment firms, some researchers provide rationalized explanations on diversification and suggest that the discount is not due to differences in efficiency but rather to other factors. Some studies (Maksimovic and Phillips, 2002; Campa and Kedia, 2002) indicate that the cross-sectional diversification discount can be generated even based upon the profit-maximization diversification model. The discount cannot be used as evidence to prove that diversification destroys value. Firms choosing to diversify are systematically different from focused firms prior to diversification. Many studies, including Lang and Stulz (1994), identify that diversified firms are poor performers and tend to trade at below industry average even prior to diversification. Campa and Kedia (2002) indicate that part of the diversification discount is the result of the endogeneity of the diversification decision. They show that firms choose to diversify as a means to move away from low growth industries. Diversifying firms have higher value than exiting or disappearing firms in their industry, but lower value than firms remaining focused in the industry. They suggest that there are systematic patterns in the relative value of diversifying firms because of the pattern in diversification strategies,

26 18 especially since exiting firms tend to be low value firms. As a result, the industry median Tobin s q will boost up, creating a cross-sectional value discount for diversifying firms. Such discount is due to changes in industry composition rather than changes in intrinsic value. Campa and Kedia (2002) control the pre-existing firm and industry factors in their tests. The results do not show significant diversification discount. This group of arguments indicates that the documented diversification discount reflects the pre-existing characteristics of firms that choose to diversify rather than the direct impact of diversification. Moreover, Gomes and Livdan (2004) suggest that the direct impact of diversification should be positive because it is an optimal choice for low or no growth firms to explore new growth opportunities. The diversification discount reflects that firms that choose to diversify are usually low or no growth businesses with low M/B ratios. Gomes and Livdan s argument is consistent with Hyland and Diltz s (2002) finding that conglomerate firms tend to be poor performers that adopt a diversification strategy in an effort to acquire growth opportunities. Following this rationale, it is necessary to revisit some of the empirical findings identified in early evidence. For example, Berger and Ofek (1995) find that diversified firms tend to over-invest in low Tobin s q industries, and they attribute this to operational inefficiency. Scharfstein and Stein (2000) find that diversified firms misallocate capital by over-investing in divisions with bad investment prospects and under-investing in divisions with good investment prospects. But cross-sectional evidence cannot reveal whether the segment with bad investment prospects is a long-existing or relatively newer business line. Without scrutinizing the investment trend of the firm, it would be too early to conclude that diversification causes inefficient asset allocation.

27 Real Options Explanation Gomes and Livdan (2004) propose that diversification is expected to enhance rather than destroy value for low growth firms through exploring new growth opportunities. The focus of their study is to explain the cross-sectionally identified diversification discount rather than to test the direct value impact of diversification activities. Moreover, Gomes and Livdan s (2004) explanation is mainly applied to the scenario when a low growth firm chooses to diversify into unrelated business. Bernardo and Chowdhry (2002), on the other hand, suggest that diversification may lead to lower M/B ratio for good performers as a result of materializing growth potentials into real investment. Our study is closer to Bernardo and Chowdhry s (2002) approach in that we focus on investigating the change of M/B ratio before and after the diversification. In addition, our explanation covers diversification scenarios for both low and high growth firms. Bernardo and Chowdhry (2002), in their theoretical model, argue that multisegment firms have lower market value because they have exhausted the options to expand, and the documented diversification discount does not necessarily relate to operational inefficiency. Our study expands the application of real options theory to diversification effects investigation. It has opposite predictions on the change in direction of firm values from what have been suggested by the resource-based view and rent-extraction arguments. Real options theories have important implication on diversification study, because it is a convention in finance literature to proxy the growth options using M/B ratio, which happens to be the major value measure for diversification effect as well.

28 20 Using M/B ratio to proxy growth opportunities, however, is not free of problems. For example, M/B ratio has been used to measure multiple factors such as corporate performance, intangibles, and market power. But an empirical comparison among different growth proxies carried out by Adam and Goyal (2003) shows that M/B assets ratio, calculated as the market value of debt and equity divided by the book value of assets, is still the best variable to proxy for investment opportunities. M/B assets ratio has the highest level of information content on investment opportunities, and it is least affected by other factors when compared to M/B equity ratio, calculated as the market value of equity divided by the book value of equity, P/E (price/earning) ratio, and firm s capital expenditures. Despite the inherent connection between growth options and diversification study through the shared evaluation variable, M/B ratio, not many studies, except for Bernardo and Chowdhry (2002), have been found linking these two fields together. In this section, we review the related literature in real options theory, which will help generate new perspectives on the diversification effect. Real options theory addresses three major characteristics of investment decisions. First, the uncertainty embedded in every investment project means that there is no 100- percent-sure investment decision in reality. There is always the possibility that the firm may regret in the future after investing. Second, once the investment materializes, it is hard to reverse it because the majority of the invested assets cannot be recovered and will become sunk costs if the investment is terminated. As a consequence, the third characteristic of investment decisions is that delaying an investment decision gives the firm an opportunity to learn more about the uncertainty, to wait for an optimal time for

29 21 investment, and to reduce the chance of regret (Dixit and Pindyck, 1998). Thus the feasibility of delaying investment is valuable to a firm. To link the real options idea to firm value, Dixit and Pindyck (1998) indicate that expandability of operations gives rise to a call option and to invest is to exercise the option. The value of the call option decreases with increases in expansion costs. So even though firms have the choice to invest in any industry, they usually obtain more expandability or call option value from their existing business areas, because it is more costly to enter a new area than to expand in existing business. On the other hand, Abel, Dixit, Eberly, and Pindyck (1996) illustrate theoretically that a firm making the investment partially or totally reversible acquires a put option. Both the call and put options have holding value or time premium because the uncertainty in the future may generate an adverse circumstance under which the firm may regret its decision of having exercised the option early. Based on real options theory, De Andrés-Alonso et al. (2005) indicate that M/B ratio can be illustrated as: M B = AIP + ( RO), AIP in which AIP stands for assets-in-place, usually measured by the book value of the assets; RO stands for real options, which are the different valuable alternative strategies available to a firm. Real options theory suggests that the more valuable choices the firm holds, the higher the M/B ratio will be. The irreversibility feature of real investments provides new interpretations on some of the phenomena documented previously as evidence for inefficient allocation of resources. For example, because of the irreversibility of real investment, firms facing low

30 22 or no growth will not exit one market easily. When the market situation changes and there comes the chance to gain from the low growth segment, it will be much less costly to expand an existing segment than to start a new segment. The identified overinvestment in low M/B industries of diversified firms might be a result of irreversibility of real investment, rather than that of an active investment choice. Alvarez (1999) shows theoretically that at the optimal exit threshold, operating revenues may be well below costs, meaning that production can be optimal even when net cash flows are negative because the value of future productive potentials is higher than current loss. Following this rationale, a diversified corporate structure reduces the exit cost for low growth business segments. Diversified firms can voluntarily close low growth business segments at an optimal time, whereas focused firms will be forced to go through bankruptcy process when they exit the market. The high costs related to bankruptcy process and the limitation in choosing an optimal time to exit for focused firms indicates that low growth firms are expected to gain when they choose to diversify into unrelated businesses. In addition, a more constrained financing market for single segment firms as identified by previous literature (Deng et al., 2007; Reeb et al., 2001; Penas and Unal, 2004) may lead to the case that only better performing single segment firms can survive, and the average M/B ratios for focused firms will be upward biased. The built-in differences between diversified and focused firms illustrated above are in accordance with the literature (Whited, 2001; Villalonga, 2004b) that questions the comparability between focused firms and conglomerates. Examining the intertemporal change rather than the cross-sectional difference becomes a necessity to investigate the value impact of diversification.

31 23 Though most of the real options discussions so far focus on investing or disinvesting decisions related to isolated single projects, the value impact of real options can be extended easily to cover interrelated multiple-project cases such as diversification. The real options view suggests that firms with good potential in existing businesses will have high value in call options because of the expandability. At the same time, these firms have minimal or no put options because of the high sunk cost of disinvesting. Under the diversification investment scenario, when a firm with great future in existing business expands into related new business segments, it exercises a call option and should expect a decrease in M/B ratio. AIP a + COa + PO Assuming the M/B ratio after diversification is AIP a a, in which CO stands for call option, PO stands for put option, and the sub-a stands for after diversification. The assets-in-place after the diversification (AIP a ) is usually larger than the assets-in-place before the diversification because diversification is an activity in need of real asset investment. The call option (CO a ) value decreases after the firm exercises the investment. Since the business of the old and new segments has a high level of correlation, related diversification does not provide the firm more flexibility to exit. Thus the value impact of put options (PO a ) is minimal. As a result, the overall value, in terms of M/B ratio, decreases after the diversification under this scenario. When a firm with good potential in existing business diversifies into an unrelated business, it also exercises a call option based on its existing resource, because some of the operational capabilities are transferable across different industries. But the creation of expandability in a new business gives rise to a new call option related to the new segment. At the same time, the value of put options increases. Since some common

32 24 resources are now shared by multiple business lines, the sunk cost related to disinvesting one segment decreases. The M/B ratio after the diversification for this scenario can be illustrated as AIP + CO + CO + PO a ao AIP a an a, in which CO ao stands for the call option related to the old business line after the diversification; business line after the diversification; and CO an stands for the call option related to the new PO a stands for the put option generated after the diversification. The call option value related to original resources decrease, while call option value related to new business lines increases, and the put option value increases as the sunk costs to terminate either business line decreases. The overall change in M/B could thus be either positive or negative. For a firm with not much potential in existing business, it usually has little competitive advantage and low or no expandability in current segment, so the rise of expandability in new business and the increase in put options in either business lines will affect firm value positively and outweigh the impact of exercising the prior negligible call option. The overall change in M/B for this case should be positive. When a firm with not much potential in existing business diversifies by adding a closely related segment, it will not gain much from either the creation of call option or put option. The firm is not able to increase its call option value because the industry of the new segment is closely related to its old business, in which it lacks competitive advantage. The put option value is also limited because the high correlation between the old and new segments will not provide much flexibility for the firm to exit either business line. On the other hand, since the value of real options prior to diversification is minimal,

33 25 the firm has not much to lose from the diversification process either. The overall change in M/B ratio for this case is unclear. Though the real options idea is believed to depict a more realistic investment decision process than the conventional NPV method, the complexity in measuring the real options value has hindered its application in practice. Most of the real options academic studies are on the theoretical conjecture level. One of the areas with relatively more empirical examination focuses on growth options. The concept of growth options is parallel to the idea of growth opportunity. The quantified proxy for growth options is launched as early as 1977 when Myers defines that the market value of a firm has two components, the current value of assets-in-place and the value of growth options. M/B assets ratio, or the closely related Tobin s q, has been widely used as the proxy for growth opportunities in corporate investment literature. The growth options idea follows the same rationale as call options, which arise from expandability, but the valuation of growth options has taken more external factors into consideration. For example, increased market uncertainty not only means bigger risk that may lead the firm to delay investment, but it also encourages early investment because higher uncertainty means more opportunity, and early investment and market entry is associated with greater privilege to expand in the future (Kulatilaka and Perotti, 1998).

34 Previous Examination on Value Changes around Diversification The difficulty of cross-sectional studies on diversification effects in generating a conclusive result calls for an examination on value changes around diversification, which is the approach this study follows. Previous effort in this direction has not been very successful and has generated mixed results. The evidence based on stock price reactions to diversification is not consistent. For example, Morck et al. (1990) find that investors respond negatively to unrelated acquisitions during the 1980s but not during the 1970s. John and Ofek (1995) show that in late 1980s, the market reacts positively when diversified firms decide to increase focus and sell asset. On the other hand, positive response to diversification has also been identified. Hubbard and Palia (1999) show that announcements of diversifying acquisitions during the 1960s and 1970s conglomerate merger wave give rise to positive returns. For the same period, Matsusaka (1993) also observes similar phenomena that acquirers in unrelated purchases realize positive abnormal returns upon the announcement of a merger, whereas those in related acquisitions realize negative abnormal returns. Besides the problem of mixed results, event studies using stock price reaction is not an ideal method to investigate diversification effect. The market response can be distorted by the price level paid and does not necessarily reflect the diversification impact itself. Loughran and Vijh (1997) show that five-year post-acquisition stock price performance can be either positive or negative depending on the type of acquisition and the method of payment applied during the transaction. Firms employing stock mergers