Performance shocks, turnaround strategies and corporate recovery: Evidence from Australia

Transcription

1 Performance shocks, turnaround strategies and corporate recovery: Evidence from Australia Alfred Yawson School of Banking and Finance, the University of New South Wales, Sydney, Australia Abstract Using a sample of Australian firms, we document improvements in operang performance following performance shocks. These improvements result in part from turnaround strategies adopted by management. Evidence suggests that changes made to leverage and operang expenses result in a negave contemporaneous effect on performance improvement. The adjustments made to working capital (revenue growth) result in a lagged negave (posive) impact on performance. Furthermore, whilst asset sales have a negave contemporaneous effect, layoffs, divestures and new CEOs have a lagged posive impact on performance improvement. The interacon of financial and restructuring strategies is found to result in an incremental impact on firm performance. Overall, there is evidence to suggest that financial and corporate restructuring strategies are efficient responses to performance shocks. JEL Classificaon: G32; G34 Keywords: Operang performance; Financial strategies; Corporate restructuring

2 1. Introducon A corporate turnaround may be defined as the recovery of a firm s financial performance following a performance decline. Turnaround strategies have been described in the business strategy literature as a master plan of acons necessary to reverse a declining business situaon (Barker and Duhaime, 1997). Although a host of business condions may reflect such an exigency, a substanal decline in financial performance is often considered a prime movaon for turnaround acons (eg., John et al., 1992; Ofek, 1993; Kang and Shivdasani, 1997; Denis and Kruse, 2000). A successful turnaround strategy results in a firm achieving a considerable improvement in performance. This paper focuses on two turnaround strategies (financial and corporate restructuring) that are ulised by firms when they experience performance shocks. The paper proceeds as follows. First, we analyse financial strategies adopted by management in response to performance shocks. Specifically, we examine changes to financial leverage, dividend payout rao, operang expenses, revenue growth and working capital, and relate these variables to changes in operang performance following a performance shock. Second, previous studies suggest a relaonship between corporate restructuring acvies and firm performance (Kang and Shivdasani, 1997; Denis and Kruse, 2000). This paper complements and extends prior studies by invesgang the impact of restructuring acvies on the corporate recovery process. In this regard, the current paper examines the use of asset sales, employee layoffs, new CEO appointments and sale of subsidiaries (divestures) as restructuring efforts aimed at achieving a turnaround. Third, there are suggesons in the turnaround literature to indicate that turnaround acons could have interacve effects on firm performance (Arogyaswamy et al., 1995). For example, the appointment of a new CEO coupled with an aggressive revenue growth strategy by a firm can jointly result in a stronger performance improvement than what could be achieved by either of these events implemented alone. As a result, this paper invesgates the impact of the interacon between financial and corporate restructuring acons on firm performance. Fourth, the implementaon of turnaround acons can take a long me, but management will usually concentrate acvity at the onset of the problem. 2

3 The mely implementaon of turnaround acons is therefore necessary to prevent further losses. What is less emphasised in the literature is the me limit by which a turnaround acon will be expected to impact firm performance. Consequently, this paper invesgates both the contemporaneous and lagged impact of turnaround acons on firm performance. This invesgaon will undoubtedly help firms in the planning and implementaon of turnaround acvies. Our movaon for pursuing these invesgaons is threefold. First, although anecdotal evidence suggests that management are concerned about the financial strategies that affect their performance, the impact of a financial strategy change on firm performance has not received the attenon it deserves in the academic literature. A thorough understanding of the effect of a financial strategy change on firm performance will enable firms adjust their financial strategies accordingly with the view to performance improvement. Second, a good exploraon of corporate restructuring strategies available to firms and their impact on future performance can serve as a guide to firms designing workable strategies to counteract performance shocks. Third, an understanding of the impact of the interacon between financial and corporate restructuring strategies on performance improvement can go a long way in assisng firms to design a combinaon of strategies to achieve a turnaround. Our sample consists of Australian firms that achieved a posive industry adjusted operang performance in one year followed by a substanal decline the following year over the period 1991 to Similar to the findings by Denis and Kruse (2000), our sample firms experience significant improvements in operang performance in each of the first three years following the performance shock. Consistent with predicons from the turnaround literature, there is evidence that adjustments made to financial leverage and operang expenses have a negave contemporaneous effect on performance improvement. Revenue growth has a one year lagged posive impact on firm performance. We attribute these findings to the efficiency gained through appropriate adjustments to financial policies. In addion, over 19% of the sample firms responded to performance shocks by employing corporate restructuring acvies. These 3

4 restructuring acvies, which are similar to those documented by Ofek (1993), Kang and Shivdasani (1997) and Denis and Kruse (2000), include employee layoffs, asset sales, divestures, new CEO appointments and friendly takeovers. Further analysis revealed that asset sales have a negave contemporaneous effect, whilst employee layoffs, divestures and CEO appointments have lagged posive effect on firm performance. Furthermore, the results suggest an interacve impact on firm performance when both financial and corporate restructuring strategies are used to deal with performance shocks. A number of studies have examined restructuring acvies pursued by poorly performing firms (eg., John et al., 1992; Ofek, 1993; John and Ofek, 1995; Kang and Shivdasani, 1997; Denis and Kruse, 2000). This study is closely related to Ofek (1993), Kang and Shivdasani (1997), and Denis and Kruse (2000). Ofek (1993) was primarily concerned with the effect of leverage on restructuring decisions by poorly performing firms. However, Ofek (1993) did not extend his invesgaon to the performance of firms following performance declines. Kang and Shivdasani (1997) examine corporate restructuring acvies following performance declines for a sample of Japanese and US firms. They find that both the Japanese and the US firms engage in various restructuring acvies following performance declines. The study by Denis and Kruse (2000) focuses on restructuring acvies by poorly performing firms during the acve takeover period ( ) and less acve period ( ). Both Kang and Shivdasani (1997) and Denis and Kruse (2000) conclude, from univariate results, that a firm s performance improves following a restructuring acvity. However, these studies do not provide any informaon on the funconal relaonship between a restructuring acvity and firm performance. The current paper differs in several important respects. First, whilst previous studies are primarily concerned with corporate restructuring acvies, this paper highlights the relave importance of both financial and corporate restructuring strategies in dealing with performance shocks. By modelling financial and corporate restructuring acvies together, the paper attempts to address the issue of which, if any, better results in performance improvement. Furthermore, previous studies do not emphasise the important issue of mescale required for turnaround 4

5 acons to impact firm performance. This informaon is important for managers in selecng appropriate turnaround acons that could result in a quick performance improvement whilst keeping the long term corporate strategy in mind. The current paper fills this gap. Another important disncon is that, this paper evaluates the interacon between financial and corporate restructuring strategies in dealing with performance shocks. The rest of the paper is organized as follows. Secon 2 sets out the theorecal background of the paper. Secon 3 describes the data and reports some descripve stascs. Secon 4 focuses on the impact of a financial strategy change on future firm performance. Secon 5 examines corporate restructuring acvies and future firm performance. Secon 6 examines the effecveness of both financial and corporate restructuring acvies in dealing with performance shocks. Secon 7 concludes the paper with a summary and discussions of the main results. 2. Theorecal background The literature has idenfied several generic turnaround strategies that could be employed to deal with performance declines. 1 This paper summarises these turnaround strategies in two parts; financial and corporate restructuring strategies Financial strategies As a first step to improving efficiency with the view of achieving a turnaround, management should review the financial strategies that affect the operaons of the firm (Pound, 1992; Ofek, 1993; Chowdhury and Lang, 1996). Financial strategies tend to provide a short term 1 These turnaround strategies include change of management, strong financial control, organizaonal change, product-market orientaon, improved markeng, and growth via acquisions, asset and cost reducons, investment, debt restructuring and other financial strategies. For detailed discussions of these turnaround strategies, see Slatter (1984). 5

6 soluon to performance problems. The literature suggests that a firm that has suffered a performance shock could recover by properly evaluang its cash generaon policies to ensure the availability of liquid resources to sustain operaons. Firms can increase their cash flows by increasing sales revenue, reducing dividend payments and controlling operang costs. 2 Sales revenue could be improved by raising selling prices, increasing cash discounts to customers and relaxing customer credit criteria. Pant (1987) shows that revenue generaon strategies account for most profit turnarounds. Also, a reducon in dividend payments will allow firms to conserve cash to sustain operaons (Grullon et al., 2002; Lie, 2004) whilst a decrease in operang costs can conserve cash, can also lead to an improved operang margin at the same me. Firms may also restructure their debt obligaons in response to performance shocks (Ofek, 1993). Debt restructuring could result in either an increase or a decrease in the proporon of debt in the capital structure. An increase in debt obligaon can improve liquidity, and also, provide incenves for management to improve performance (Jensen, 1989). In theory, it should be relavely difficult for a firm experiencing performance problems to raise addional loans. Empirical evidence suggests that such firms will attempt to rere some of their exisng loans and also convert part of it into equity to reduce interest payments and the likelihood of bankruptcy (Slatter, 1984). Furthermore, as a high level of debt can cause financial distress, a firm is likely to reduce its debt levels when it experiences a performance shock (Ofek 1993; Kahl, 2002). Prudent working capital management is also important in turnaround situaons. Firms are expected to reduce excess investments in working capital as part of the recovery process. A reducon in working capital could be achieved by reducing debtors by instung efficient debt collecon mechanisms, and also, reducing the amount ed up in inventories. Firms can also, though rarely, extend payments to creditors (Slatter, 1984). It is expected that prudent 2 Although theorecal models suggests that firms are reluctant to decrease dividends due to the negave signal it may send to the market (eg., Bhattacharya, 1979), firms may be compelled to do so in the presence of severe performance declines to improve liquidity (eg., Ofek, 1993; Gullon et al., 2002; Lie, 2004). 6

7 adjustments to financial policies following performance shocks should result in a performance improvement Corporate restructuring strategies Firms can also implement corporate restructuring strategies following performance shocks. This paper evaluates change of management, asset reducon (sale of assets and divestures) and employee layoffs as restructuring strategies necessary to migate a profit shortfall Change of management Turnaround situaons often require new CEOs (Slatter, 1984). A new CEO is required to provide a new sense of direcon, develop new financial and operang strategies and revitalise the firm. A change in CEO may occur even if the performance decline was brought about by condions beyond the control of the incumbent management. For example, if the enre industry is not performing well due to an industry specific shock, management should not be held responsible for poor performance (Morck et al., 1989). Even though CEOs may become scapegoats in those instances, their removal signal to the stakeholders that something posive is being done to improve performance. There is overwhelming empirical evidence to support this theory. Prior evidence suggests that board of directors demonstrate their responsiveness to poor performance by replacing poorly performing CEOs. For example, Warner et al., (1988), Weisbach (1988), Ofek (1993) and Denis and Kruse (2000) all show that the likelihood of top management changes is negavely related to firm performance. These empirical findings suggest that a replacement of a CEO should result in a performance improvement. In support of this view, Denis and Denis (1995) document an improvement in operang performance following dismissals of CEOs. Thus, new CEOs are expected to play important roles in the corporate recovery process Asset reducon 7

8 Empirical evidence suggests that managers who value firm size are more reluctant to reduce the assets under their command (Stulz, 1990). However, asset reducon becomes necessary when a firm suffers a performance shock (Lang et al., 1995; Kang and Shivdasani, 1997; Berger and Ofek, 1997; Denis and Kruse, 2000; Denis and Shome, 2004). The logic behind an asset reducon strategy is that, by disposing of redundant assets, a firm can operate the more useful assets. Furthermore, asset reducon could be used to improve cash flow. For a firm that has a huge debt overhang, cash realised from asset reducons could be used to reduce financial leverage (Kahl, 2002). Asset reducon could be accomplished in different ways, including closure of plants, the sale of assets and divestures (Ofek, 1993; Denis and Shome, 2004). In a smaller firm with no subsidiaries, asset reducon could be accomplished through disposing of nonperforming assets. Divestures are the preserve of larger firms that have substanal investment in different business segments and subsidiaries. In support of this view, John et al., (1992) and John and Ofek (1995) find divestures to be a dominant strategy for large firms coping with performance declines. It follows that firms that are able to dispose of their unwanted assets following performance shocks should achieve improvement Employee layoffs Employee layoffs have become a widespread turnaround strategy in recent years (eg., Inverson and Pullman, 2000; Chen et al., 2001). Layoffs can occur when a firm experiences a declining product demand, and also, when a new technology changes the producon process in a way that reduces the demand for labour. Although firms can provide several reasons for reducing their workforce, layoff decisions are usually made after a period of remarkable underperformance measured by a firm s accounng earnings and stock returns (Kang and Shivdasani, 1997; Chen et al., 2001). The successful implementaon of a layoff strategy will enable firms to cut down on labour costs and also increase labour producvity, especially when the layoff decision stems from 8

9 getng rid of unproducve workers. The variables used to proxy the above theories and their expected signs are briefly summarised in Table 1. Insert Table 1 about here 3. Data and descripve stascs The objecve of our sampling methodology is to idenfy firms that have experienced a substanal drop in operang performance. Operang performance is defined as the rao of earnings before interest, taxes, depreciaon and amorsaon (EBITDA) to total assets. Operang performance is preferred because share price incorporates the market expectaon of the value of any turnaround strategy that may be employed by firms following performance shocks (Morck et al., 1989; Denis and Kruse, 2000). This measure has consistently been applied in recent studies that examine firm performance in different situaons (eg., Kang and Shivdasani, 1997; Denis and Kruse, 2000; Lie, 2004; Kim et al., 2004). As a starng point, we idenfy the populaon of firms listed on the Australian Stock Exchange (ASX) for the period 1991 to Financial data for these firms are obtained from the financial informaon stored on the Fin Analysis database (Aspect Financial). To ensure the integrity of the data, we cross checked the data with the annual financial statements stored on Connect 4 database. Similar to Kang and Shivdasani (1997), Denis and Kruse (2000) and Kim et al. (2004), each firm s rao of EBITDA to total assets is adjusted by subtracng the median rao of EBITDA to total assets for all firms that fall into the same ASX industry sector descripon. This results in a firm specific measure of performance, which is within the control of management. The industry adjusted performance should also alleviate concerns about industry effect in operang performance that is likely to bias the analysis. Firms are included in the sample when their industry adjusted operang income is posive in one year but the rao becomes negave the following year. In other words, the sample includes firms that perform above the industry median in one year but performs below their industry median the following year. Thus, the sample consists of firms that underperformed their industry peers from 1992 to This 9

10 process results in an inial sample of 415 observaons. Since we are interested in turnaround strategies following performance shocks, we attempt to avoid including firms that may be financially distressed prior to the year of performance shock (Ofek, 1993; Kang and Shivdasani, 1997). Such firms might have already put some turnaround acons in place, making it difficult to determine their immediate and subsequent impact on firm performance. Consequently, we require that the rao of interest expense to operang income in the year prior to the performance decline be less than one (Kang and Shivdasani, 1997). This criterion resulted in the eliminaon of 16 firms, leaving a final sample of 399 firms. Even though not a prior condion, none of the firms entered the sample twice Annual distribuon of sample firms Table 2 presents the annual distribuon of the sample firms in the year of the performance shock. The total number of firms idenfied for each year ranges from 13 (3.26% of sample) in 1992 to 68 (17.04% of sample) in The mean and median number of firms across the sample period is 44, indicang that performance shocks do not cluster in any parcular year. The mean annual performance, however, differs across years. The mean performance over the sample period is -0.18, but individual years exhibit deviaons from this value. For example, the average performance for 1992, 1997 and 1999 is lower than the sample average, whilst the remaining period recorded higher annual averages. The median firm performance is fairly distributed across the sample period with the excepon of 1992 value, which seems to be a funcon of the sample size. Insert Table 2 about here 3.2. Analysis of firm performance The operang performance of the sample firms in the base year (year -1) and the year of the performance shock (year 0) are reported in Table 3. The table provides industry adjusted 10

11 operang performance (panel A) and changes in performance (panel B) from year -1 to year +3. Barber and Lyon (1996) demonstrate that nonparametric tests are more powerful than parametric test in studies of operang performance. Moreover, mean operang performance values are more likely to be influenced by outliers. Hence while both mean and median values are reported, the subsequent discussions focus mainly on median values. As reported in Table 3, the median industry adjusted EBITDA as a proporon of total assets in year -1 is The median industry adjusted operang performance declined to in year 0. The median change in performance from year -1 to year 0 is Stated differently, the sample firms suffered a median performance decline of 231%. This change in performance is stascally significant at the 1% level. Pourciau (1993) provides evidence that newly appointed CEOs manage earnings in order to report lower income in the early years of their tenure. We invesgate whether the appointment of new CEOs can explain the reported performance declines of the sample firms. Out of the 399 firms in the sample, 4 firms appointed new CEOs in year -1. However the performance declines of these firms are not stascally different from those of the remaining firms. Thus, the sample firms were performing well but suffered a huge industry adjusted performance shock necessary to movate turnaround acons. The stascs in panel B of Table 3 show that, consistent with Denis and Kruse (2000), the sample firms achieved significant improvements in operang performance from year 0 through to year +3. The median change in performance from year 0 to year +1 was The median firm consistently achieved performance improvement in years +2 and +3, with median industry adjusted values of and 0.026, respecvely. The improvement in the median firm's performance is stascally significant at the 1% level in each of the three horizons. 3 3 It is possible that the improvement in operang performance reflects mes series properes of accounng earnings (Penman, 1991; Fama and French, 1995). We follow the procedure in recent research in dealing 11

12 Insert Table 3 about here 3.3. Survivorship bias Naturally, we are able to report operang performance for only those firms that survived the sample period. Thus, we are unable to trace the future performance of firms that exit ASX by way of acquision or bankruptcy. This is likely to introduce survivorship bias in the analysis. As indicated in Table 4, the sample firms decreased from year 0 to year +3. By the end of year +3, the sample firms had reduced from 399 to 367, a reducon of 8.02%. Of the 32 firms that could not survive the sample period, 30 (93.75%) of them were acquired in friendly deals whilst 2 (6.25%) of them could not meet the data requirements. It is possible that firms that survived as independent firms performed better in the year of the performance shock than those that could not survive, making the performance improvement reported in secon 3.2 a suspect. To assess the extent of survivorship bias in these results, we compare the performance of survivors and non survivors in year 0. Although not a complete measure, poorer performance of non survivors relave to the survivors in year 0 would indicate a potenal upward bias in the reported changes in performance. The results presented in Table 4 indicate no significant difference in the median performance of the survivors and non survivors. Hence, there is no evidence to conclude that survivorship bias gives rise to the improvement in performance reported in secon 3.2. Insert Table 4 about here 4. Financial strategies following performance shocks The theory on corporate turnaround suggests that the performance improvements achieved by the sample firms following the performance shock may result in part from the with this problem by including a change in performance in the previous year in the esmated models (eg., Aboody et al., 1999; Lie, 2004). 12

13 financial strategies adopted by management. We invesgate this view by first analysing the changes made to financial strategies in response to performance shocks. It is expected that there will be significant differences in financial strategies before and after the performance shock. Table 5 reports the mean and median changes in financial strategies from year -1 to year 0 and from year 0 to year +1. The change in dividend payout rao from year -1 to year 0 is whilst the change from year 0 to year +1 is The mean difference between these changes is stascally significant at the 5% level. Thus consistent with the findings by Grullon et al., (2002) and Lie (2004), an average firm reduces its dividend payout rao when it experiences a performance shock. This seems to suggest that firms become financially constrained when they encounter performance difficules and they attempt to conserve liquid resources by cutng down on dividend payments. Moreover, even though the average change in revenue growth from year 0 to year +1 is negave, it represents a significant improvement over the growth achieved from year -1 to year 0. A median firm, however, achieves a posive growth in revenue from year 0 to year +1. This is significantly different from the median growth achieved from year -1 to year 0, suggesng that firms pursue aggressive growth strategies when they encounter performance problems. Furthermore, a median firm is able to reduce its operang expenses from year 0 to year +1 and the median change compared with the change from year -1 to year 0 is stascally significant at the 1% level. The reducon in operang expenses is necessary to improve operang margins. As hypothesised, sample firms reduce their investments in working capital from year 0 to year +1 as a way of reducing costs associated higher investment in current assets. Finally, firms reduce their financial leverage after a performance shock even though the difference is not stascally significant. The descripve stascs provide evidence that firms experience significant changes in their financial strategies following performance shocks. Insert Table 5 about here 4.1. Impact of financial strategies on firm performance 13

14 Indeed, if a change in a financial strategy reflects an appropriate managerial response to a performance shock, we shall expect such change to be significantly related to performance improvement. We conduct this invesgaon by esmang the following cross seconal equaon. EBITDA t+ τ, i + β GRO (1) 4 = α + β DIV + β WC β LEV + β EBITDA 6 2 t 1, i + β EXP 3 + β SIZE 7 + ε We esmate equaon 1 separately for changes in industry adjusted operang performance from year t to year t + τ, where τ = years +1, +2 and +3 and measures the change from year 0 to year +1 (see table 1 for definion of variables). The log of total assets (SIZE) in the prior year controls for a possible size effect in the corporate recovery process. The change in the rao of EBITDA to total asset in prior year controls for the me series properes in accounng earnings that can affect future operang performance (Penman, 1991; Fama and French, 1995). Although in theory, past earnings should impact current earnings, we are unsure of the funconal form of this relaonship. Following Aboody et al., (1999) and Lie (2004), we assume that future performance is linearly related to past performance. To determine the parameter esmates of the models, we first attempt to idenfy the correlaon between the independent variables (financial and corporate restructuring) used in the models. The correlaon coefficients are reported in Table 6. The highest correlaon coefficient of 0.36 is between divestures and firm size. This is to be expected because large firms are more likely to divest when they encounter performance shocks (John et al., 1992; John and Ofek, 1995). Although there are correlaons between a few other explanatory variables, their coefficients are quite low so mulcollinearity should not pose a problem. Insert Table 6 about here 14

15 Table 7 presents OLS regression results from equaon 1. 4 The results provide evidence that improvements in operang performance in year +1 are significantly and negavely associated with changes to financial leverage and operang expense (t=2.45 and 2.19). These results demonstrate that the ability of management to reduce leverage and operang expense results in an immediate improvement in firm performance. There is weak evidence that financial leverage has a one year lagged impact on performance improvement. One explanaon to this finding is that a reducon of leverage following performance shocks potenally releases funds from interest payments into other producve areas, which potenally improve operang performance. Growth in revenue, dividend payout rao and working capital do not have any contemporaneous effect on firm performance. Consistent with expectaons, the financial improvement achieved in year +2 is significantly posively related to revenue growth (t=2.32). Thus, as predicted, sample firms are able to pursue revenue growth strategies to improve performance but their impact is felt only in the subsequent years. Surprisingly, none of the financial strategies explains the performance improvement in year +3, supporng the view that financial strategies are short term efficiency measures aimed at a quick turnaround (Chowdhury and Lang, 1996). The results in Table 7 also show that, the change in operang performance in year t is significantly and negavely associated with performance improvement in year +1 (t=2.08). It is significantly posively associated with changes in performance in years +2 and +3 (t=12.71 and 6.11), confirming the theory that prior accounng earnings can influence current performance (Penman, 1991; Fama and French, 1995). Although financial strategies play a role in the corporate recovery process, most of the performance improvement, especially in years +2 and +3, are explained by prior operang performance. 4 The reported stascs reflect the winsorisaon of the observaons that lie ±3 standard deviaon from the mean. The reported t-stascs are based on White (1980) robust standard errors. 15

16 Insert Table 7 about here 5. Corporate restructuring acvies following performance shocks This secon describes the appointment of new CEOs, asset sales, divestures and employee layoff acvies following performance shocks. The list of divestures is compiled from Thomson Financials Securies Data Collecon Planum database. A divesture is defined as a sale of subsidiary by the parent to a third party, which could include investor group comprising the management of the divested subsidiary. In order to eliminate very small divestures that are likely to introduce noise in the models, we require the value of the transacon to be at least $US10 million. This value is much lower than US$100 million and US50 million cut off points used by Mulherin and Boone, (2000) for the U.S and Powell and Yawson (2004) for the UK market, respecvely. However, the smaller size of the Australian market warrants the use of a much lower value. Comparavely, the minimum transacon value used in this paper is much higher than the minimum value of A$0.5 million imposed by da Silva Rosa et al., (2004) in their study of the market for takeover advisers in Australia. The informaon on the appointment of new CEOs and employee layoff announcements are obtained from Signal G records through the Securies Industry Research Centre of the Asia-Pacific (SIRCA). 5 Asset sale is defined as sale of plant, property and equipment with a value of at least 5% of the total book value of assets. Table 8 summarises the major restructuring acvies pursued by the sample firms. Consistent with US evidence, the most common corporate restructuring acon following performance shocks is asset reducon (10.8% of the sample). Within this group, 12 firms (3.01% of sample) divested subsidiaries whilst 31 firms (7.8% of sample) disposed of assets. 6 All the 5 Signal G is the data feed provided by the ASX to communicate corporate announcements to brokers and investors. 6 In idenfying divestures, layoffs and CEO appointments, mulple events for a given firm are consolidated. For example, if a firm divested two or more mes in the same year, only one observaon is recorded. This approach reduces the number of acvies, but it is unlikely to bias the results. 16

17 subsidiaries and the assets were sold for cash, indicang the importance managers put on cash inflow in the corporate turnaround process. Also, twenty two firms (5.5 % of sample) reported a change in CEO. In comparison, Ofek (1993) reports 21% CEO replacement for US firms whilst Kang and Shivdasani (1997) document 14% for Japanese firms. Thus, the replacement of CEOs following performance shocks in Australia is lower than those reported for the US and Japanese firms. It should be emphasised that the studies used for comparison are rather old. To the extent that the speed at which CEOs are replaced following performance shocks has changed in the US and Japan over the past decade, the comparison will not be valid. Also, in the pre-performance shock year, none of the sample firms announced employee layoffs. However, 8 firms (2.0% of sample) announced employee layoffs between year 0 and year +1. There were 10 firms (2.5% of sample) that engaged in at least two different restructuring acvies from year 0 to year Insert Table 8 about here To appreciate the extent of a performance shock necessary to movate a restructuring acvity, the sample firms are paroned into quarles condioned on their industry adjusted rao of EBITDA to total assets in the year of the performance shock. Overall, 30.3% of restructuring acvies occurred in the 1 st quarle whilst 27.6% occurred in the 4 th quarle. There are however, some differences in the frequency of individual restructuring events across quarles. Over 8% of divestures occurred in the 1 st quarle whilst 66.7% occurred in the 4 th quarle. Although asset sales are common across quarles, they are more pronounced in the 4 th quarle. The appointment of new CEOs mostly occurs in the 1 st quarle. Whilst 37.5% of employee layoffs occur in the 1 st 7 Although our research design does not allow us to directly test the impact of corporate control, it is important to realise that the external takeover market plays an important role in restructuring Australian firms that have experienced performance shocks. To provide informaon on this issue, we compiled a list of takeovers from the SDC Planum database. In all, 13 firms (3.3% of sample) were acquired in friendly deals between year 0 and 1. 17

18 quarle, none of this event occurs in the 4 th quarle. Despite these differences, our Pearson s χ 2 test suggests that there is no difference in the frequency of restructuring acvies among firms in the 1 st and 4 th quarles of performance declines, with the sole excepon of divestures. This suggests that corporate restructuring acvies are pursued by poorly performing firms irrespecve of the magnitude of the performance shock. This seems to indicate that self selecng a sample from a list of poorly performing firms, which has characterised most previous studies, to evaluate their restructuring acvies is unlikely to add any value to the analysis Determinants of corporate restructuring acvies following performance shocks In view of the fact that over 80% of the sample firms did not engage in any restructuring acvity from year 0 to year +1, we are inclined to invesgate the determinants of restructuring likelihood following performance shocks. To pursue this issue further, we esmate a mulnomial logit model, tesng the associaon between the likelihood of a restructuring choice and a set of financial variables. The mulnomial logit model specifies the probability P ij that firm i will select outcome j following a performance shock (be a non-restructuring firm if j=0; be a takeover target if j=1; layoff employees if J=2; divest if J=3; appoint a new CEO if J=4, and sell assets if j=5). Hence, the dependent variable takes the values 0, 1, 2, 3, 4 and 5. X ij, is a vector of explanatory variables and β is a vector of unknown parameters to be esmated. The vector of variables include: return on assets, dividend rao, financial leverage, operang expenses, revenue growth and working capital. These variables measure the industry adjusted change in financial condions from year -1 to year 0. We also include firm size because large firms are more likely to restructure following performance shocks. The model is specified as follows: P ij = exp( β ' j X i ) 1+ exp( β ' j X i ) (2) 18

19 In order to idenfy the parameters of the model, the normalisaon β 0 = 0 is imposed (Maddala, 1983). The maximum likelihood technique is used to esmate the model s parameters. The esmaon procedure yields five sets of coefficients, represenng each of the restructuring choices relave to the non restructuring firms. Insert Table 9 about here The results in column 5 of Table 9 indicate that the decision to sell assets following performance shocks is negavely related to return on investments. This result is consistent with the findings by Denis and Shome (2004) who document significant negave relaonship between prior operang performance and the likelihood to downsize. Furthermore, the results in column 1 and 3 of Table 9 indicate that the likelihood to divest or be acquired following a performance shock is posively related to firm size. The coefficients indicate that divestures are more sensive to firm size than takeovers. This result is consistent with prior evidence that suggest that large firms coping with performance declines are more likely to divest in order to focus operaon on core business areas (eg. Lang et al., 1995; Berger and Ofek, 1999). Moreover, there is evidence to indicate that large firms experiencing low revenue growth are more likely to replace their CEOs following a performance shock. Note that there is no relevant variable that explains the decision to layoff employees following performance declines. The general implicaon of these findings is that, although firms respond to performance shocks with a variety of restructuring acvies, the likelihood of a restructuring event cannot be easily determined by the change in financial condions in the prior year. Perhaps ownership and strategic factors can better explain choice of a restructuring acvity following performance shocks. This is an avenue for future research Impact of corporate restructuring acvies on firm performance 19

20 We further invesgate the extent to which corporate restructuring events contribute to performance improvements. We relate restructuring acvies from year 0 to year +1 to the changes in firm performance from year 0 to year +3. Since we have no prior evidence on the funconal form, we assume based on the empirical literature that restructuring events will be linearly related to performance improvement (eg., Kang and Shivdasani, 1997; Denis and Kruse, 2000). Hence, we esmate the following cross seconal equaon; EBITDA t+ τ, i + λ EBITDA 5 = λ + λ LAY t 1, i 0 + λ SIZE λ DIV + ε 2 + λ CEO 3 + λ SALE 4 (3) Insert Table 10 about here where τ = years +1, +2 and +3 and captures the restructuring acvies from year 0 to year +1. Again, we esmate this regression for each of the three horizons. We also control for firm size and change in performance in the previous period. As presented in Table 10, the only restructuring acvity that results in a performance improvement in year +1 is asset sales (t=2.13). Asset sales have a negave contemporaneous impact on firm performance. This finding suggests that an ad hoc decision to eliminate assets following a performance shock could result in operang losses. Furthermore, the results show a posive and significant associaon between employee layoffs and performance improvement in year +2 (t=2.06). Thus, employee layoffs result in the firm achieving efficiency in operaons through improved producvity of labour and a reducon in labour cost. Furthermore, divestures have a posive impact on firm performance in year +3 (t=2.32). These results seem to suggest that employee layoffs and divestures are 20

21 strategic decisions that are taken with a long term corporate performance in view. Contrary to expectaons, the appointment of new CEOs has no significant impact on firm performance Addional analysis and robustness checks 6.1 Impact of financial and corporate restructuring acvies on firm performance As pointed out in secons 4 and 5, both financial and corporate restructuring acvies play significant roles in dealing with performance shocks. Consequently, we model both turnaround acvies together to assess their complementary impact on firm performance. This analysis also serves as a robustness check for the results reported in the previous secons. To invesgate this, we esmate the following cross seconal OLS regression where τ = years +1, +2 and +3 and measures financial and corporate restructuring acvies from year 0 to year +1. EBITDA + β WC 93 t+ τ, i = 03 β EBITDA β LAY 6 t 1, i β YR 0 + β DIVEST + β SIZE 11 + β DIV ε + β LEV + β CEO β EXP + β GRO + β SALE (4) Insert Table 11 about here 8 One cricism of this approach is that it ignores restructuring acvies that take place after year +1. To overcome this problem, restructuring acvies from year 0 to the year in which performance is measured are taken into account. In results not reported, we find a significant posive associaon between employee layoffs and performance improvement in year +2 but divestures become insignificant in year +3. Assets sales become negave and significant for both years +2 and

22 Since the Australian economy experienced a posive growth throughout the sample period, an improvement in firm performance could be attributed to the general growth of the economy and not necessarily to the turnaround acons pursued by management. We control for this effect by including yearly dummies in the models. The variable YR is an indicave variable that equals one if a performance improvement is recorded in year Y and zero otherwise. Table 11 presents OLS results from equaon 3 for each of the three horizons. Consistent with the previous results, financial leverage, operang expenses and asset sales are negavely and significantly associated with performance improvement in year +1. The results suggest that firms that are able to reduce their financial leverage probably from cash realised from asset reducon strategies and the conversion of debt into equity instruments experience performance improvement. The results for year +2 are essenally the same in signs and significance as those reported in Tables 8 and 10. Two major differences should, however, be noted. First, the results show that CEOs have a posive impact on firm performance in year +2, whereas working capital becomes negave and significant (t=1.72 and 1.71). Thus consistent with theory, new CEOs improve performance but there is a me lag for their impact, though marginal, to be felt. Similar to the previous models, divestures are posively and significantly related to performance improvement in year Overall, the results suggest that financial and corporate restructuring strategies play complementary roles in dealing with performance shocks Interacon effects 9 It is possible that the performance improvements achieved by the sample firms are influenced by the manipulaon of the asset structure by management. For example, management may use a substanal amount of off balance sheet assets, which can potenally decrease the asset base, resulng in the improvement in the rao of EBITDA to total assets. As a further robustness check, we remove firm size and replace it with change in size, but the results remain unchanged. 22

23 An important consideraon in implemenng turnaround strategies is that a turnaround acon pertaining to one strategy may impact other types (Chowdhury and Lang, 1996) which could potenally lead to an incremental effect on firm performance. This hypothesis is grounded in the synergy effect theory, where the implementaon of two strategies may result in a better performance improvement. For example, if employee layoffs contribute to higher employee producvity, it will reflect in revenue growth which could impact firm performance. Also, a new CEO may pursue an aggressive revenue generang strategies, both of which could impact firm performance. As hypothesised in table 1, corporate restructuring acvies and financial strategies (revenue growth and financial leverage) are predicted to have a posive impact on firm performance. Hence, we expect the interacon of revenue growth and leverage with the restructuring acvies to provide a posive incremental effect on firm performance. 10 Specifically, we replicate equaon 3, permitng the coefficients on corporate restructuring variables to vary with revenue growth and leverage. The regression results are reported in Panel B of table 11. Consistent with predicons, the incremental coefficient on new CEO interacted with revenue growth is posive for all three horizons but significant in years +1 and +3 only (t=3.86 and 1.65). This result suggests that firms that appoint new CEOs and are able to increase their revenue have greater chance of achieving a recovery. Surprisinly, layoffs interacted with revenue growth result in a negave incremental impact on firm performance in year +1 (t=2.15). This finding is counterintuive. However, layoff interacted with revenue growth result in a posive incremental effect on firm performance in year +2 (t=1.76). Furthermore, divestures interacted with revenue growth results in a significant posive incremental effect on firm performance in year +2 (t=1.86). Thus firms that are able to streamline their operaons by eliminang misfit subsidiaries and are also able to pursue revenue growth strategies at the same me have a greater chance of recovery. Asset sales interacted with 10 We do not provide the interacve effects for all variables because there are no clear expectaons when a corporate restructuring event and a financial strategy result in an opposing effect on firm performance. 23

24 revenue growth results in an incremental negave impact on firm performance. Thus, the interacon of financial and corporate restructuring acvies can result in an incremental effect on firm performance. 7. Summary and concluding remarks This paper provides evidence that both financial and corporate restructuring strategies play important roles in dealing with performance shocks. Regarding financial strategies, the results show a negave contemporaneous effect of financial leverage and operang expenses on performance improvements. Growth in revenue has a one year posive lagged impact on firm performance. These results reflect short term efficiency gained through prudent adjustments to financial strategies (Chowdhury and Lang, 1996). Second, prior evidence suggests that corporate restructuring events are common among firms that have experienced performance declines (Ofek, 1993; Kang and Shivdasani, 1997; Denis and Kruse, 2000). Consistent with prior studies, we find asset sales, divestures, new CEO appointments, employee layoffs and friendly takeovers to be popular restructuring strategies pursued by Australian firms that have experienced performance shocks. Whilst the determinant of a restructuring choice following performance shocks cannot easily be determined by financial variables, there is evidence that lower return on investment compared with non restructuring firms can determine the likelihood of asset sales. This is similar to the findings by Denis and Shome (2004). Also, firm size is posively related to the likelihood of divesture and takeovers whilst CEO change is significantly related to lower revenue growth and posive firm size. Further analysis indicates that asset sales have a contemporaneous negave impact on firm performance. However, we find employee layoffs and new CEO appointments to be posively associated with performance improvement in year +2. Divestures are found to be posively associated with performance improvements in year +3. These findings suggest that employee layoffs, divestures and new CEOs have long term impact on firm performance. Furthermore, there is evidence that the interacon between corporate restructuring events and financial strategies provides an incremental effect on performance improvement. More interesngly, we find the appointment of new CEOs interacted with revenue growth to have a 24

25 posive contemporaneous effect on firm performance, whilst layoffs and divestures interacted with revenue growth have a lagged impact on firm performance. Our analysis offer several suggesons for corporate execuves responsible for designing effecve turnaround strategies to reverse a performance decline. First and foremost, it is important for management to make serious attempt to reduce financial leverage and operang expenses when their firms encounter performance shocks in order to achieve a quick turnaround. Furthermore, management should consider disposing of misfit subsidiaries in order to concentrate on core business areas to achieve organisaonal efficiency. The board can also appoint a new CEO who can competently facilitate the transion from the performance shock to corporate recovery. These specific acons are consistent with the proposals in the turnaround literature, and are essenal for firms to recover from performance shocks. It is worth nong that performance shocks are reversible through proper applicaon of financial and corporate restructuring strategies. However, management should be aware that turnaround acons could have either contemporaneous or lagged impact on operang performance. Consequently, we recommend management and the board to have construcve debates to enable them design effecve turnaround strategies with the view to achieving a recovery. References Aboody, D., Barth, M., and Kasznik, R., Revaluaon of fixed assets and future firm performance: Evidence from the UK. Journal of Accounng and Economics 26, Arogyaswamy, K., Barker, V. L., and Yasai-Ardekani, M Firm turnarounds: An interacve two stage model. Journal of management Studies, 32, Barker III, V. L., and Duhaime, I. M., Strategic change in the turnaround process: theory and empirical evidence. Strategic Management Journal 18, Berger, P., Ofek, E., Causes and effects of corporate refocusing programs. Review of Financial Studies 12,