Private Firm Valuation Using Multiples: Can Artificial Intelligence Algorithms Learn Better Peer Groups?

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

‘In the aspect of IPOs [10]–[12], and of leveraged buyouts [13], [14], authors conclude that selecting peers based on the industry criterion is superior to using pure total market multiples’ – specific clarifications are needed. ‘(i) Transaction (acquisition) characteristics are related to synergy filters (e.g. [27]), control filters (e.g. [28]), diversification filters (e.g. [29]), regional filters (e.g. [30]), and a filter indicating the method of payment (e.g. [31]). (ii) Market characteristics are related to market condition/activity filters lending support to the managerial herding hypothesis [32]; examining the relation between the premium paid in acquisitions and deal size [29]; examining valuation differences in boom and crash market periods relative to stable periods of IPOs [33]; comparing the effectiveness of various industry classification codes e.g., [34]). (iii) Company characteristics are related to profitability (ROA and ROS e.g. [35]; examining the performance of various industry groupings [36]; examining the performance of industry-related as compared to cross-sectoral multiples [25]; elaborating upon accuracy and drivers’ evidence of multiples [37]; examining sell-side analyst’s choice on peers [38]; examining optimal peer selection formulae [39]; examining CEO compensation e.g. [40]), risk (size of the target and acquiror firm e.g. [41]); business risk e.g. [34]; legal form of the target firm [42]; size of the acquired stake [32]; size ratio target firm to acquirer firm e.g. [29]), growth (e.g. [43]), and further criteria (type of accounting [44]).’ – also insufficient specific info here for each source. The same here: ‘(e.g. sequential or batch training algorithms [58]), Growing SOM [59], LVQ [56], etc.)’. ‘By adding a regional filter [18], adding filters for profitability and intangible assets [19], adding a growth filter [20], and adding some combination of profitability [21], [22], growth, risk, and further financial ratios filters similarly find an improvement’ – develop on each aspect. ‘Finally, find the isolated use of a company size/risk, growth or debt filter to be inferior to the industry filter was found [20], [25], [26]’ – poorly constructed and indicate the particular results of each source. ‘4. Research Methodology and Results’ – these should be distinct sections and of course their contents should be redistributed. ‘The setup for the SOM algorithm for clustering similar input instances is outlined in Error! Reference source not found.3’ – check this. ‘5. Discussion’ – missing comparisons with other research results. The reference list is not properly edited and most of the cited sources are extremely old.

Comments on the Quality of English Language

‘Finally, find the isolated use of a company size/risk, growth or debt filter to be inferior to the industry filter was found [20], [25], [26]’ – poorly constructed.

Author Response

Dear Reviewer,

We express our gratitude for dedicating your time to reviewing our manuscript and providing valuable feedback. Enclosed herewith, you will find our responses.

Best regards

 

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The following aspects need to be improved:

1. It is necessary to clarify why the proposed model is better than mathematical and statistical models in the economic context (see line 57). What other models have been studied in the literature, the context and benefits of machine learning are not clear.  It is also not understood in what type of research the previous studies have been developed and what is the motivation of the research. Why is linear regression chosen? The research is not well understood.

2. The literature review is not clear about the problem and adds the results that have been determined in the studies, but it does not explain what is being investigated. It mixes strategic, economic and financial aspects, but since the hypotheses or research questions are missing, it is difficult to understand the contributions and limitations that previous studies have had and what aspects will be contributed in the present research. 

3. In 3.1 it mixes many aspects but it is not clear why the characteristics of the data are determined and why the sample is selected. What does it contribute to the research, what are the criteria for data analysis, are there more optimal ways of analysis?

4. In 3.2 Why is it studied with ML?why is linear regression chosen and not another one? are the variables dependent or independent? why are they stochastic? 

5. In 3.2 Why is it studied with ML? why is linear regression chosen and not another one? are the variables dependent or independent? why are they stochastic? why is Neighbourhood Component Analysis suitable for the analysis? there are many methods but it is not clear if they are necessary to solve the problem.

6.There is a lack of coherence between the introduction, literature review, development and methodology. 

7. Challenges could be added to the discussion.

8. The research is not understood, it needs to be revised and better structured.

9. Improve conclusions and summary. 

Comments on the Quality of English Language

Moderate editing of English language required

Author Response

Dear Reviewer,

We express our gratitude for dedicating your time to reviewing our manuscript and providing valuable feedback. Enclosed herewith, you will find our responses.

Best regards

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Please see the attachment

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Abstract should be re-written to state what you did in the manuscript and methodology and significance for the private firm valuation; Missing key Feldman's work on private firm valuations.

Author Response

Dear Reviewer,

We express our gratitude for dedicating your time to reviewing our manuscript and providing valuable feedback. Enclosed herewith, you will find our responses.

Best regards

Author Response File: Author Response.docx

Reviewer 4 Report

Comments and Suggestions for Authors

This article recommends using machine learning over older methods.

As a main drawback I can't see the dataset anywhere, so I just rely on the words of the authors without evidence. They haven't uploaded the dataset to a public repository, eg GitHub. (Maybe I didn't notice it). Both the code and the dataset are imperative to be seen online so that no one can dispute the results.

The discussion should perhaps be expanded by emphasizing the results.

Author Response

Dear Reviewer,

We express our gratitude for dedicating your time to reviewing our manuscript and providing valuable feedback. Enclosed herewith, you will find our responses.

Best regards

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript can be published, with two caveats: 1. sections cannot have only one subsection (there are at least two such instances). 2. the reference list is poorly edited.

Author Response

Dear Reviewer, 

thank you for taking your time for reviewing our manuscript. Our comments are in the file below.

Best regards

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The specific contribution of machine learning to this research remains unclear, as statistical models could be employed to achieve the same results.

It would be beneficial to include a discussion of the advantages and disadvantages of using machine learning versus mathematical formulae. This section would be enhanced by a discussion of the relevant literature.

Comments on the Quality of English Language

Minor editing of English language required

Author Response

Dear Reviewer, 

thank you for taking your time for reviewing our manuscript. Our comments are in the file below.

Best regards

Author Response File: Author Response.pdf

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

The text has undergone a significant improvement, and the structure and the new knowledge it generates are now clearly understood. 

Comments on the Quality of English Language

Minor editing of English language required