Informing the Vote: The Business Rescue vs Liquidation Decision

Background: Non-unitary insolvency systems require an initial choice between liquidation and rehabilitation. For those systems, the decision to further support rehabilitation is often reinforced by the estimate of a lesser return in liquidation. The creditors’ decision to either accept or reject the business rescue plan depends substantially on comparing the proposed liquidation value with the business rescue value that is mandated in the business rescue plan. Purpose: The purpose of this study is to compare the proposed liquidation value with the actual liquidation value of firms that had commenced business rescue proceedings but were subsequently liquidated. This study is exploratory; therefore, our aim is to use the findings of this study to derive future research opportunities for scholars to explore. Methodology: This research adopted a content analysis research approach by collecting documents such as business rescue plans and liquidation accounts, which are qualitative in nature, and describing these documents through quantitative data analysis. Findings: The study revealed that for the secured creditors’ sample, there was a significant difference between the proposed liquidation value and the actual liquidation value. For the unsecured creditors’ sample, this study found no significant difference between the proposed liquidation value and actual liquidation value. Limitations: The major limitation of this study was access to data; therefore, the huge decline Lastly, the findings of this study derived future research opportunities for scholars to explore.


Introduction
Undressed, a business rescue plan seeks to convince, by appealing to reason, the notion that through its approval the creditors will remain better off than in liquidation. This should not be done by misleading anyone or strong-arming them. Instead, it should influence and alter the way a creditor sees its investment in the distressed business. Often, though, the reason is overshadowed by a simplified comparison between the return that business rescue and liquidation will offer. A failure to appreciate the derogation of assets in either of these two procedures can significantly alter any promised return. When turnaround efforts fail to revive a business, the cost is often far more significant than expected, making the previously promised liquidation dividend regrettable, to say the least. We asked ourselves what the eventual liquidation return was for creditors not fortunate enough to achieve the plan's desired outcomes, and how far it was from the one proposed in the plan? In doing so, this study reveals a glimpse into a hidden cost within the business rescue process and how misleading the business rescue versus liquidation dividend comparison can be for decision-makers.
South Africa adopts a non-unitary insolvency system, requiring a firm to distinguish between business rescue and liquidation at the onset of proceedings. For a financially distressed firm, the choice between liquidation and business rescue is not always that clear and is often influenced by fear and hopefulness. Business rescue is usually initiated first in order to preserve the firm as a going concern, should it have a reasonable prospect of survival. However, more importantly for creditors, business rescue is preferred, so long as it offers a higher return on their claims than in liquidation. Fortunately for creditors, their role in the insolvency process allows them to levy their preference. Therefore, choosing the proper insolvency process is an economic decision and must be made timeously to prevent the erosion of the firm's assets. That being said, business rescue offers two major instances for the decisions to be contested by affected parties. The first is the commencement decision, which is made prior to initiating the formal proceedings. The decision can be made by the court through compulsory application by an affected party-but more often than not, over 90% is taken by directors through a voluntary application (Pretorius 2015). Ideally, one would prefer to select the best procedure at this point in order to prevent the unnecessary erosion of the firm's assets. However, while few to no administrative costs have been incurred at this point, limited information exists to make a clearly informed decision. Decision-making tools such as the Likelihood of Liquidation Framework can be used for the commence decision (Rosslyn-Smith and Pretorius 2018). Should business rescue proceedings be initiated, the business rescue practitioner will assume control of the firm's affairs and begin drafting a business rescue plan. The process of preparing and negotiating a plan is costly, as it generates direct administrative costs which include, but are not limited to, the basic remuneration of the business rescue practitioner, legal fees, professional and expert and advisory services, and contingency fees (De Abreu and Rosslyn-Smith 2021;Rosslyn-Smith, De Abreu, and Pretorius 2020).
The second occasion is at the second creditors' meeting, where creditors are expected to vote on the business rescue plan (Conradie and Lamprecht 2015). In this instance, however, creditors should be in a far more informed position regarding the turnaround potential of the firm (Mumford 2003). The expectation of the business rescue plan is that it should serve (Rosslyn-Smith and Pretorius 2015): • As a tool for feasibility declaration.
• As a medium of communication.
• As an enabler of transparency.
• To clarify the contractual obligations.
• To attract and secure post-commencement financing.
Falling short of the above expectations would compromise the creditors' decisionmaking abilities and possibly result in a less desirable outcome for all the affected parties. Furthermore, Section 150(2) of the Companies Act No. 71 of 2008 (hereafter referred to as the Act) prescribes the compulsory elements required to be presented in the business rescue plan. One of these is the probable dividend that would be received by creditors, in their specific classes, if the firm were to be placed in liquidation. The probable liquidation dividend must be stated as at the date of the commencement of the business rescue proceedings, and not at the date of publication of the business rescue plan-as was the case in Commissioner for the South African Revenue Service v Beginsel and Rennie NNO and Others (2013).
This study refers to this estimation as the proposed liquidation value (hereafter referred to as PLV). The rescue plan, in addition, is required to outline the probable dividend that a creditor would be expected to make if the rescue plan were to be implemented (hereafter referred to as the business rescue value [BRV]). While analysis of the rescue plan by the creditors should be comprehensive, it often boils down to whether the probable dividend creditors would receive in business rescue will exceed that of the PLV. As indicated by Le Roux and Duncan (2013), most creditors have little knowledge of the business rescue procedure and are often persuaded by this simple comparison. For all intents and purposes, it would be safe to assume that the creditors' vote is significantly influenced by this comparison, and that the business rescue plan is likely to succeed so long as the following is present:

>
(1) Table 1 provides a clear breakdown of the three acronyms that will be commonly used throughout the study. The probable dividend creditors would receive should the firm immediately liquidate. This estimated value is presented in the business rescue plan and accounted for by the business rescue practitioner. Actual liquidation value (ALV) The actual dividend creditors received once the firm was liquidated. This value is presented in the final liquidation, distribution and contribution account and is drawn up after the sale of assets by the liquidator. Business rescue value (BRV) The probable dividend creditors would receive should the business rescue plan be implemented. This value is presented in the business rescue plan and is estimated by the business rescue practitioner.

Source: Authors' own compilation
The information in the business rescue plan, therefore, should detail the risks associated with pursuing the plan and adjudicate against a liquidation scenario. If the liquidation scenario is misrepresented, decision-makers will be unable to determine the best course of action. It is imperative for all parties involved to understand how the liquidation scenario is impacted by turnaround efforts and reveal the hidden costs within the business rescue process. This study, therefore, aims to evaluate the business rescue versus liquidation dividend comparison, in order to better equip decision-makers. In addition, this study aims to use the findings of this study to derive future research opportunities for scholars to explore. To do so, this study compares the PLV to the ALV of firms that had commenced business rescue proceedings but were ultimately unsuccessful and subsequently liquidated. Creditors are the principal decision-makers of the plan and should be divided into two groups based on their propensity for risk (Levenstein 2015). Considering these segments of creditors, the objectives of this study are set as follows: • To investigate whether there is a discrepancy between the PLV and ALV for secured creditors. • To investigate whether there is a discrepancy between the PLV and ALV for unsecured creditors.
While the vote to adopt the rescue plan is a critical moment in the rescue process, the timing of the PLV often precedes it by several months (considering extensions), while the ALV may in practice take several years. Therefore, as shown in the timeline, the depreciation of assets may be influenced (figure 1):

Figure 1: Time-lapse between Business Rescue and Liquidation
Notes: BRP = business rescue plan As figure 1 illustrates, since a non-unitary approach is used in South Africa, there exists a time-lapse between the two insolvency procedures represented by the time-scale break. Should business rescue proceedings fail to deliver a higher return to creditors, the responsibility then lies with the directors to place the firm into liquidation (voluntary application) or the court (compulsory application) (South African Bank of Athens Ltd and Another v Zennies Fresh Fruit CC and Others, 2018). Both of these options have a time break as there is no specified time for either one to file for liquidation after business rescue proceedings have been terminated.

Literature Review
There are two types of financially distressed firms: those that should enter liquidation because they are economically inefficient, and those that should remain in operation because they are economically efficient (White 1994). The latter should initiate business rescue proceedings in order to have an opportunity to be saved.
When a financially distressed firm is faced with the decision to liquidate or enter business rescue, the goal should be to select the option that is most efficient and maximises the overall value for affected parties (Gilson 2012;Levenstein 2015;Rosen, Nicholson, and Rodgers 2011;Wihlborg 2002). A successful business rescue attempt is dependent on whether the firm is still economically feasible; ultimately, this decision is the responsibility of the practitioner (Le Roux and Duncan 2013).
Section 150(2) of the Act stipulates compulsory elements to be included in the business rescue plan. The above-mentioned section requires the business rescue plan to disclose, inter alia, the PLV, the BRV, as well as the preference ranking that will be applied to pay creditors, should the business rescue plan be adopted (Conradie and Lamprecht 2021;Loubser 2010). In order to determine the PLV and the BRV, the practitioner will need to determine both the liquidation value and the going concern value of the firm as well as the value of the firm's assets in both cases (Conradie and Lamprecht 2021). The creditors will then evaluate and compare the PLV and the BRV and vote in favour of or against the business rescue plan (Levenstein 2015;Rosen et al. 2011).
It could be argued that these are the most critical elements of the business rescue plan when creditors are faced with voting in favour of or against the business rescue plan. A creditor would not vote in favour of a business rescue plan if the proposed BRV is less than the PLV (Levenstein 2015;Loubser 2010). According to Bradstreet, Pretorius, and Mindlin (2015) and Le Roux and Duncan (2013), the majority of creditors (except for large creditors such as banks) involved in business rescue proceedings have little understanding of business rescue and what it encompasses. The fact that the business rescue plan is written in legal jargon further contributes to creditors' lack of understanding of the business rescue plan (Rosslyn-Smith and Pretorius 2015). It may be difficult to determine these values as there are various elements (that cannot be estimated with any degree of precision) that may ultimately have an impact on the accuracy of these values (Conradie and Lamprecht 2021;Loubser 2010). Thus, in order to enable creditors to make an informed decision, it is imperative that the information disclosed in the business rescue plan be unbiased, transparent and reliable (Pretorius and Rosslyn-Smith 2014). If these values are incorrect and the business rescue plan is accepted, this could result in inefficient outcomes (Levenstein 2015).

What if the Proposed Liquidation Value is wrong?
As central tenets, the valuations presented in the business rescue plan, such as the PLV and the BRV, play a critical decision-making role for creditors (Conradie and Lamprecht 2021). If the practitioner overestimates the going concern value of the firm, it will result in an inflated BRV. Therefore, creditors could vote in favour of the business rescue plan even though it might not be the most efficient outcome and the result thereof might be the erosion of assets, ultimately leading to a reduced return to all affected parties (Levenstein 2015). The converse situation, however, is equally detrimental. In the event that the practitioner undervalues the going concern value of the firm, this could result in the firm being liquidated even though the most efficient option would have been its rehabilitation.
If the business rescue proceedings are terminated and liquidation proceedings are commenced, discrepancies might occur where the PLV and the ALV values do not equate to the same amount. Several contributing factors can explain this outcome, such as the erosion of assets by the rescue attempt, the incorrect valuation of assets at the commencement, or the manifestation of the principal-agent dilemma. How these factors may influence the outcome is detailed below.

Erosion of Assets
When an economically inefficient firm enters business rescue, in some cases, the rescue attempt simply delays the inevitable liquidation of the firm (White 1989). As a result, the creditors' dividends may be diminished due to the erosion of the firm's assets while rescue attempts are made (Levenstein 2015;Meeks and Meeks 2009). The erosion of a firm's assets may be caused by a number of factors.
One type of asset erosion that occurs is the depletion of cash during the rescue attempt. Cash may be depleted by the direct reorganisation costs such as the practitioner's fees and other professional fees such as those of lawyers (Branch 2002;Meeks and Meeks 2009). The longer the duration of business rescue proceedings, the higher the level of asset erosion, and the lower the dividend for creditors in liquidation (Meeks and Meeks, 2009). According to Branch (2002), these costs represent wealth transfers from creditors to these professionals.
During the rescue attempt, the practitioner may obtain additional financing by utilising any unencumbered assets as security for such finance (Leontsinis, Seymour-Cousens, and Ireland 2014). The practitioner may also seek cash in order to meet debt by selling off other unencumbered assets that are not critical for the firm's operations (Meeks and Meeks 2009). A practitioner who tries to save an inefficient firm may thus draw on the firm's assets in an attempt to rescue the economically inefficient firm (White 1994).
When the inefficient firm eventually ends up in liquidation, the creditors' dividend may be significantly decreased for the above reasons. As the Act allows the practitioner to utilise any unencumbered asset as security for PCF, unsecured creditors may receive a lower dividend than predicted in the PLV, because there may be no unencumbered assets left to satisfy these claims (Bradstreet 2015, 125). The practitioner's fee and any expenses associated with business rescue proceedings increase with the duration of business rescue proceedings and are ultimately paid from the assets once the firm is liquidated (De Abreu and Rosslyn-Smith 2021; Morrison 2007). Ultimately, creditors will bear the costs of the failed business rescue attempt and could possibly receive a lower dividend in the event of a subsequent liquidation. These costs would not exist if the firm went straight into liquidation (Le Roux and Duncan 2013; Loubser 2010).

Incorrect Valuation of Assets
According to Argenti (1976), incorrect valuation of assets is associated with failure and is, therefore, a symptom of financial distress. Determining the value of assets is a difficult task as it is a subjective process that requires assumptions to be made (Erasmus, Van Rooyen, and Oberholzer 2012). The PLV and the proposed business rescue values are dependent on the valuation of the firm's assets. Several factors may influence these values, for example, whether or not all creditors have proved their claims or if the costs involved in both liquidation and business rescue have been included in these estimates (Kim 2009;Loubser 2010).
The valuation of assets in financially distressed firms can be further complicated by data integrity issues. According to Argenti (1976), distressed firms have data integrity issues, as a common symptom of a distressed firm is accounting manipulations. The management of the firm is a key source of information in rescue proceedings and could provide information that is false, thereby influencing data integrity. This is a critical issue for the practitioner, as predicting the value of these firms could be based on distorted figures.

Principal-agent Dilemma
Pretorius (2014, 321) defines the principal-agency relationship as "a contract where one or more persons (principals) engage another person (agent) to perform some service on their behalf, which involves delegating some decision-making authority to the agent." The crux of the principal-agent predicament is the presumption that the principal and agent have conflicting interests and it is believed that the agent may not behave in the best interests of the principal (Parvinen and Tikkanen 2007;Pretorius 2014).
In the case of business rescue, the agent is the practitioner appointed by the distressed business or court, and the principals are the creditors (Pretorius 2014). Agency problems arise in business rescue as each party is affected by the outcome of the business rescue plan, consequently leading to conflicts between practitioners and creditors regarding remuneration and the desired outcome (Weiss and Wruck 1998). Although the practitioner (as the agent) makes all the decisions regarding the business rescue, he does not bear the financial risk in the event of the business rescue failing and the firm being liquidated. If liquidation ensues, the practitioners are most likely to withdraw their fees during the business rescue process. This could result in the practitioner, out of pure selfinterest, "attempting" to rescue a business that would be better off if liquidated immediately. There is also a possibility that the practitioner could manipulate information presented in the business rescue plan, such as the PLV and the BRV, in order to persuade the creditors to vote in favour of the plan.
Based on the above discussion, it is hypothesised that:

H1(alt):
There is a significant difference between the PLV and the ALV for secured creditors.

H2(alt):
There is a significant difference between the PLV and the ALV for unsecured creditors.
The methodological approach to testing the hypotheses will now be explained.

Methodology
This research adopted a content analysis research approach by collecting documents, such as business rescue plans and liquidation accounts, which are qualitative in nature, and describing these documents through quantitative data analysis (Mouton 2001).
The target population for this study consisted of firms that had commenced business rescue proceedings but were ultimately unsuccessful and liquidated between the period 9 of 1 May 2011 and 30 March 2016 in South Africa. The units of analysis for the research objective were 13 firms that were in business rescue but failed and were then liquidated.
This study used a non-probability sampling method as there was no access to an accurate and updated sampling frame from which to draw a probability sample (Cooper and Schindler 2014). A combination of two non-probability sampling methods, namely purposive and snowball sampling, was used. Purposive sampling was used to select liquidators, practitioners and the High Courts who were in possession of the documents of firms who had previously been under business rescue but were subsequently liquidated (Saunders, Lewis, and Thornhill 2009). Snowball sampling was used by asking liquidators and/or practitioners for referrals to other practitioners, as well as liquidators who had also dealt with similar cases (Saunders et al. 2009). However, as access to the business rescue plans and liquidation accounts was highly dependent on the practitioners, liquidators and the High Courts, the sample of failed business rescues was ultimately a convenience sample.
This study aimed to achieve a minimum sample size of 30 or more failed business rescue cases. The researchers contacted 79 different possible participants, which included practitioners, liquidators, large creditors such as banks and several Masters of the High Court. The response rate was 25%, as 20 participants agreed to take part in the study; this number was attributed mainly to the sensitive nature of the topic. The researchers received a total of 66 liquidation accounts and 50 business rescue plans. However, there were only 27 paired cases where the researchers had both the liquidation account and business rescue plan for a particular firm. Two cases were excluded due to the data integrity of the business rescue plan. One case was later excluded as the liquidation account was insufficient, preventing the extraction of data from the account. A further eight cases were discarded as there were no claims made by creditors in liquidation; therefore, these cases could not be compared with the business rescue plans. Finally, three more cases had to be excluded as the liquidation was still in process and there were still several assets to be realised. This resulted in a total of 13 usable failed business rescue cases.

Data Analysis and Measures
In order to calculate the PLV and ALV to be tested in the hypotheses, the business rescue plans and liquidation accounts were analysed in order to facilitate the sensemaking of the documents (Cooper and Schindler 2014). The liquidation account has a standardised format as prescribed by the Master of the High Court. However, despite Section 150 of the Act stipulating compulsory elements to be included in the business rescue plan, there were significant inconsistencies in the format of the business rescue plans obtained. Therefore, each of the practitioners reported the information in different formats.
Data conversion was necessary as the data contained in the liquidation account and business rescue plan were not presented in a format suitable for achieving the stated research objective (Zikmund et al. 2010). Therefore, data were extracted from these documents and entered into a database with various categories. The data were categorised into the following categories: the PLV for both secured and unsecured creditors and the ALV for both secured and unsecured creditors. The total value available for distribution after liquidation costs and post-commencement finance claims was also recorded for each of the firms from the liquidation accounts. The researchers were unable to extract the total value available for distribution in the PLV in the business rescue plans, as most of the plans did not contain enough details on the PLV analysis.

Proposed Liquidation Value
The practitioner produces the PLVs, which are generally presented in the section "liquidation scenario" as named in most cases. An estimated value is given of how much is predicted to be available for distribution for each class of creditors. This estimated value can be expressed in terms of "cents in the rand" or as a percentage. In this study, the PLV was calculated and recorded separately for secured and unsecured creditors and was calculated for each firm. The PLV was calculated as follows:

Actual Liquidation Value
The ALV was obtained from the distribution account section in the liquidation account. The distribution and contribution accounts provide details on the total amount each creditor claimed as well as received. In the study, the ALV was calculated and recorded separately for secured and unsecured creditors and was calculated for each firm. It must be noted that the researchers did not include post-commencement finance claims, as these claims would not have been made had the firms gone straight into liquidation. The ALV was, therefore, calculated as follows:

Ethical Considerations
Participants' informed consent to take part in the study was obtained and each participant was required to read and sign the informed consent form before providing the researchers with any data. The consent forms ensured that participants were provided with information regarding the purpose of the study, who the researcher was, and how the data would be used. Participants were ensured of the right of confidentiality and anonymity to prevent any harm against the participants or their firm.

Demographic Profile of Sample Firms
The sample consisted of 11 private companies, one closed corporation and one nonprofit company. The majority of the firms were located in the Gauteng province (n = 9), while two were located in the Western Cape province and one was located in Mpumalanga. The company size was determined by the company's public interest score. The sample consisted of three small firms, two medium firms and two large firms. The researchers were unable to determine the size of six of the firms as the business rescue plans did not provide the firms' public interest score or sufficient information to calculate the public interest score. A wide range of firms was represented in the sample, including publishing; transportation; wholesale trade; information technology; construction and related services; maintenance and repair services; catering and accommodation; motion picture and television services; and finally, community, social and personal services. This demographic profile of the firms indicates that even though the sample size was small, the study represented a variety of firm profiles.

Descriptive Statistics of Variables
Although the sample size initially consisted of 13 firms, the sample size for secured creditors was only 11. The reduced sample size is due to the fact that two of the firms in the sample did not have secured creditors. The descriptive statistics indicate a substantial difference (

Hypotheses Tests
Hypotheses H1 and H2 focused on whether significant differences exist between the PLV and ALV for secured and unsecured creditors, respectively. The null and alternative hypotheses of H1 and H2 are stated below:

H1(null):
There is no difference between the PLV and the ALV for secured creditors.

H1(alt):
There is a significant difference between PLV and the ALV for secured creditors.

H2(null):
There is no difference between the PLV and the ALV for unsecured creditors.

H2(alt):
There is a significant difference between the PLV and the ALV for unsecured creditors.
The above-mentioned hypotheses are all two-tailed (non-directional) hypotheses and were tested at a 95% level of significance (i.e., α = 0.05). As all the variables mentioned above were measured at a ratio level of measurement, the paired-samples t-test is the most suitable parametric test (Pallant 2007). If the assumptions of the paired-samples ttest cannot be met, a non-parametric substitute, known as the Wilcoxon signed-rank test, can be used (Pallant 2007).
Pallant (2007) indicates that there are two assumptions that have to be met in order to use the paired-samples t-test. Firstly, both the variables being compared should be measured at a ratio level of measurement. Secondly, the difference score between the two variables being compared should be normally distributed (Pallant 2007). The researchers assessed the assumption of normality through the Kolmogorov-Smirnov test, in conjunction with a visual perusal of histograms and normal probability plots (Field 2009). The tests displayed considerable deviations from normality in the difference scores of the variables being compared for both secured and unsecured creditors.
The non-parametric test, the Wilcoxon signed-rank test, was used to test H1 and H2 as the data violated the assumption of normality. The Wilcoxon signed-rank test works in a similar way to the paired-samples t-test. However, the Wilcoxon signed-rank test assigns a rank to the difference score of the two variables in order to be compared (Field 2009;Pallant 2007).

Hypothesis 1
The first hypothesis test focused on whether a significant difference exists for secured creditors. Table 3 contains the results of the Wilcoxon signed-rank test conducted for hypothesis 1. Notes: a. ALV < PLV, b. ALV > PLV, c. ALV = PLV, d. Based on positive ranks. Table 3 shows that there is a statistically significant two-tailed difference between the PLV and ALV for secured creditors, z = -2.19, p = 0.028, which was within the significance level of 0.05; therefore, H1(alt.) was accepted. According to Field (2009), because the z-score is based on positive ranks, the majority of ranks were negative. Therefore, we can conclude that the ALV is significantly lower than the PLV. It is clear that eight secured creditors had a lower ALV score compared to the PLV. In one of the instances, the inverse was true. There were two tied rank scores where the ALV and PLV were equal.

Hypothesis 2
The second hypothesis test focused on whether a difference exists for unsecured creditors. Table 4 contains the results of the Wilcoxon singed-rank test conducted for hypothesis 2. Notes: a. ALV < PLV, b. ALV > PLV, c. ALV = PLV, d. Based on positive ranks. Table 4 indicates, in contradiction to what was expected, that there is no statistically significant two-tailed difference between the PLV and ALV for unsecured creditors, z = -0.533, p = 0.594. Therefore, H2(null) cannot be rejected in support of H2(alt.), as the pvalue of 0.594 was higher than the significance level of 0.05. According to Field (2009), because the z-score is based on positive ranks, the majority of ranks must have been negative. Therefore, we can conclude that the ALV is lower than the PLV; however, this difference is not significant. The table shows that for six of the unsecured creditors, their ALV score was lower than the PLV score and there were only three instances where the ALV score was higher than the PLV score. There were two tied rank scores where the ALV and PLV were equal.

Additional analysis
Further investigation was conducted in order to determine possible contributions to the results above. The total amount claimed in liquidation by all creditors, including postcommencement finance creditors for the original 13 firms, was R 270 937 277.92. The total amount that was available for distribution to all creditors, once the administration costs of liquidation were deducted, was R 78 159 454.66. This indicates that only 29% of all creditors' claims were paid out in liquidation. Figure 2 displays the portion of the distribution each creditor group received.  Figure 2 shows that the total secured creditors only received 46% of the amount available for distribution and unsecured creditors only received 23%. The remaining amount left for distribution, which is a total of 31%, was distributed to postcommencement finance creditors. Interestingly, the post-commencement finance claims included practitioner fees, post-commencement finance salaries and postcommencement finance supply creditors, most of whose claims were treated as administrative expenses and were paid out of both the secured creditors' dividend and the free residue before any creditors were paid. This may shed some light on the reluctance of secured creditors (mainly banks) to support rescue, as the practitioners' fees amounted to a total of 19% and the remaining post-commencement finance creditors' claims amounted to 12% of the total 31%.  The sample size for secured creditors consisted of 11 firms, as two firms in the sample, C7 and C12, did not have any secured creditors. The figure shows that the ALV was the same as the PLV in only two cases (C6 and C8). Further, there was only one case where the ALV was higher than the PLV (C10). This may be due to deliberate manipulation or the potential incompetence of the business rescue practitioner. For the remaining eight cases, the PLV was higher than the ALV. In 10 out of the 11 cases, the BRV exceeded the PLV, where in nine out of the 11 cases, the BRV was 100 cents in the rand.  The sample size consisted of 11 firms, as two firms, C2 and C4, did not have any unsecured claims in the subsequent liquidation. The figure shows that in two cases, ALV and PLV were equal (C1 and C3). Further, there were three cases (C5, C10 and C11) where the ALV was higher than the PLV and one case (C10) where the ALV was higher than the BRV. Once again, this may be due to deliberate manipulation or the potential incompetence of the business rescue practitioner. In five cases (C6, C7, C8, C9 and C12), the PLV was higher than the ALV. Lastly, there was one case (C13) where the unsecured creditors had to make a contribution of 0.42 cents in the rand.
One may assume that because of the high BRV, creditors voted in favour of business rescue. However, the BRV may distort creditors' decision-making, as it does not account for the risk associated with business rescue, which is evident in these failed business rescue cases. As a result, creditors voted in favour of business rescue, even though it may not have been the most efficient outcome, resulting in the erosion of assets that ultimately led to reduced returns to all affected parties in the subsequent liquidation.

Discussion
The intention of the study was to determine whether there existed a discrepancy between the PLV and ALV, and then to direct future research towards understanding why such differences occurred. Initially spurred by speculation about deliberate manipulation and potential incompetence, this research shed some light on alternative contributors to the variations in PLV and ALV.
For the secured creditors' sample, the results indicate that there is a significant difference between the PLV and the ALV. These results further revealed the PLV was significantly higher than the ALV, indicating that secured creditors received a lower dividend than what was predicted in the business rescue plan. While there may exist numerous reasons for this, the following possibilities may be the cause. A possible explanation for the significant difference could, for example, be the incorrect valuation of the firm's assets, which resulted in an incorrect PLV. This is supported by Erasmus et al. (2012), who state that determining the value of assets is a difficult task, as it is a subjective process that requires assumptions to be made.
Alternatively, secured assets may have suffered damage or neglect during the time between calculating the PLV and when the ALV materialised (Krishnan and Moyer 1994;Rosslyn-Smith and Pretorius 2018). Lastly, the impact of business rescue administrative costs may well have reduced the ALV well below its potential, should liquidation have been chosen rather than adopting the rescue plan. This is evident in the findings of this study, as secured creditors received 46% of the amount available for distribution and unsecured creditors only received 23%. Should any of these possibilities be significantly responsible for the discrepancy, then secured creditors would be unable to decide the best option. Therefore, we propose the following research questions to guide future investigations: • Is the PLV greater than the ALV for secured creditors because of poor asset valuations? • Is the PLV greater than the ALV for secured creditors because of the deterioration of the firm's secured assets during business rescue? • Is the PLV greater than the ALV for secured creditors because of the added costs of the business rescue procedure?
For the unsecured creditors' sample, this study found no significant difference between the PLV and ALV. The results are surprising, given the fact that there was a significant drop-off of creditors from business rescue to liquidation in the sample; therefore, one could assume that the remaining unsecured creditors would have received a higher dividend in the subsequent liquidation compared to that when all unsecured creditors in rescue also submitted claims in liquidation.
Further analysis was conducted to explain the outcomes of the results for unsecured creditors. Analysis revealed that 31% of the available distribution in liquidation after the administration expenses were deducted was paid to post-commencement finance creditors. These post-commencement finance claims were paid out of the secured creditors' dividend as well as from the free residue before any creditors were paid. These findings provide a possible explanation for the significant difference between the PLV and ALV for secured creditors. Although there was no significant difference between the PLV and ALV for unsecured creditors, had these post-commencement finance claims not existed and had the firm gone directly into liquidation, the unsecured creditors could have received a larger dividend (Le Roux and Duncan 2013;Loubser 2010). This leads us to propose the following research questions: • Is the PLV not significantly different from the ALV for unsecured creditors because of poor asset valuation? • Is the PLV not significantly different from the ALV for unsecured creditors because of the deterioration of the firm's free assets during business rescue? • Is the PLV not significantly different from the ALV for unsecured creditors because of the added costs of the business rescue procedure? • Is the PLV not significantly different from the ALV for unsecured creditors because there are fewer creditor claims in liquidation? • Is the PLV not significantly different from the ALV for unsecured creditors because of the added post-commencement finance claims?
The above-mentioned findings suggest that the erosion of assets may be a cause of the results of the study's findings across both secured and unsecured assets. The erosion of assets may be the result of a number of factors. Although the researchers could not determine all the possible factors, it is clear that the secured and unsecured creditors' dividend in liquidation was affected by the costs of the business rescue. These findings are also congruent with the arguments of Branch (2002) and Meeks and Meeks (2009), who state that reorganisation costs may significantly decrease a creditors' dividend.
An unexpected finding from the study emerged from the data collection process. Information about the firm's assets and affairs in the rescue plan did not correlate in the majority of cases with the documentation regarding the firm's liquidation. This may suggest that in cases where firms commenced business rescue proceedings but were ultimately unsuccessful and liquidated, information regarding the firm's assets may not be transferred effectively. This may suggest a potentially serious flaw in South Africa's insolvency system. This has motivated our last research question, which may be of interest to insolvency practitioners: • Are the discrepancies between the PLV and the ALV for secured and unsecured creditors the result of data loss between the two insolvency processes?
Limitations A limitation of this study worth mentioning was the access to data that resulted in a decline in the sample size compared to what was expected. Seventy-nine participants were contacted over two months, with many resisting the request to share the information. The majority of the participants argued that the information contained in the business rescue plans was not public information or that retrieving the documents would be too time-consuming. Furthermore, most of the liquidation accounts requested from the High Courts were not finalised. This does reveal the need for more transparency in these records for research purposes.

Contributions and Value of the Study
This study offers several contributions. To the researchers' knowledge, this is the first study comparing the proposed liquidation value and actual liquidation value in order to determine whether differences exist. The study tackles a novel perspective on one of the influences that can affect the vote. This is of particular importance to creditors, who often have little knowledge of business rescue proceedings and may find the results of this study to be helpful. Legislators are also encouraged to consider the findings of this study compared with future amendments of the Companies Act. Lastly, the findings of this study derived future research opportunities for scholars to explore.