DSS Case Study A knowledge-based Decision Support System

DSS Case Study: A knowledge-based Decision Support System for enterprise mergers and acquisitions Faculty of Computer Science and Engineering, HCMC University of Technology October 2011 1

Outline n n n 1. Introduction 2. Mergers and acquisition method and process 3. The architecture of the knowledge-based DSS. 4. Conclusions References 2

Introduction n n The merger and acquisition offers incredible market opportunities. The main objectives of M&A are to capitalize on gaps in the market, increase market share, improve customer service, and eliminate local and global competition, not simply to cut operation costs. M&A have been applied in a wide range of industries such as banking, communications, airline, manufacturing, and service industries. Because of the complexity and importance of M&A, decision support systems are frequently used as tools to support decision-making. 3

Mergers and acquisition method and process n n Merger means acquiring control of a target company through stock purchases or exchange. Mergers include the acquisition of the essential assets of a company by another company or the acquisition of shares in a company by another company. Acquisition is the combining of two or more enterprises into one through the purchasing of a target company’s assets or stocks. Mergers can be divided into two types: absorptive merge and creative merge. Absorptive merger means the target company will be eliminated. Creative merger means that both companies including the target company and the buying company will be erased and then a new company will be established. 4

Ten actions by the acquiring firm to assure success in M&A. n n n n n 1. Set the company’s long-term objectives 2. Select the means to achieve the objectives 3. Make sure that the proposed acquisitions can achieve the objectives. 4. Establish candidate search procedures 5. Develop strategies and criteria for analyzing potential acquisitions 6. Prepare negotiations with the target firm. 7. Prepare actual consummation of the acquisitions. 8. Plan the post-acquisition changes and integration. 9. Allocate budgets and responsibilities. 10. Exercise leadership and control. 5

Figure 1. The process of mergers and acquisitions 6

Five steps of M&A process In practice, the M&A process can be broken down into five steps: q q q 1. Perform a pre-acquisition review 2. Search and screen the target company 3. Investigate and value the target 4. Acquire the target through negotiation 5. Perform post-merger integration 7

Step 1 - Pre-acquisition review n n In this step, a pre-acquisition review is performed to assess the company’s own situation and decide whether a M&A strategy should be adopted. If the company finds that it’s difficult in the future to maintain its core competencies, market share, return on capital, etc. , the a M&A program may be necessary. If a company fails to protect its valuation, it may find itself the target of a merger. The main task in a pre-acquisition review process is to determine if the desired growth rate of the target company can be achieved in the future. q If not, an M&A program should establish a set of criteria whereby the company can grow through acquisition. 8

Step 2: Search and Screen targets n n The second step is to search for possible M&A candidates Target companies must fulfill a set of criteria so that the fitness of the target company is good to the acquirers. The search and screening step process should be performed in-house by the acquiring company. This step consists of the following activities: q q q Develop a growth strategy defining the role of M&A. Set criteria for candidate screening, evaluation and selection Identify, collection information about and assess potential candidates. Determine which candidate offers the best fit for a deal. Develop an action plan for executing the deal. 9

Step 3: Investigate and value the target n n The step is to conduct more detailed analysis of the target company. This will require review of q q Operations Strategies Financials Other aspects 10

Step 4 – Acquire through negotiation n n After the target company is selected, the process of negotiating an M&A begins. In this step, several key questions should be considered: q q n How much resistance will come from the target company? What are the M&A benefits for the target company? What will be the bidding strategy? How much does the acquirer offer in the first round of bidding? The most common M&A method used to acquire a target company is for both companies to reach agreement about the M&A. 11

Step 5: Post-merger integration n n In this step, the integration of two M&A partners is achieved. The post-merger integration step is the most difficult phase. This step requires extensive planning throughout the entire organization. The integration process can take places at three levels: q q q Full: all functional departments Moderate: some key departments Minimal: some selected personnel 12

The architecture of the knowledgebased decision support system (KDSS) n The KDSS for mergers and acquisitions is composed of 6 components: q q q n A database A case base A rule base A model base An inference engine A user interface. It provides users with a friendly and effective interface to communicate with the system and find alternative actions for decision-making. 13

Figure 2. The knowledge based DSS for mergers and acquisitions 14

The architecture of KDSS n n n The database provides all the information needed in the model base and rule base for calculating, analyzing and reasoning. The rule base stores all the necessary rules for supporting reasoning when the inference engine performs forwarding chaining. The model base employs critical statistical models to analyze, forecast, or evaluate business value. In KDSS, there are two models: q q q The discounted free cash flow Economic profit models. and adopt scenario analysis to perform business valuation. 15

n n The case base helps the user deal with a new problem while he/she faces a similar case. The inference engine uses rules in the rule base and data in the database to select suitable rules that it can use to infer new knowledge or models to predict or analyze the problem. 16

The database n n n The database stores a collection of end user data that consists of raw facts of interest to the end user and meta data which are the data about data. The meta data provide a description of the data characteristics and relationships. To manage data, answer ad hoc queries, gain better access, and reduce data inconsistency, a DBMS is required. This approach uses Microsoft Access to create a database and an ODBC driver to access the data in the database. The database contains all financial data used in the models, cases, or rules. 17

n n The database consists of 3 tables: Password, Financial. Data, and Model. Parameters The Password table is to verify user identification. The Financial. Data table provides diverse information about a company’s finance. The attributes of Financial. Data are Current_Liabilities, Net_Operating_Capital, Current_Assets, Net_Property_Plant_Equipment, Other_Operating_Assets, Short_term_Investment, Goodwill_amortization, Non_Operating_Investment, market_price, stockholder_equity, debt_ration, etc. The Model. Parameter is used to provide a wide range of variables for model analysis and forecasting. 18

The knowledge base n n n The knowledge base contains domain knowledge useful for problem solving. In order to use experiences effectively, knowledge extraction and collection must be continually performed. The knowledge stored in the knowledge base is in the form of rules or cases, which the inference engine in the KDSS can quickly and accurately infer and generate suggestions and actions. 19

The rule base n n n There is a rule base in the knowledge base. The rules in the rule base are represented as a set of rules. Each rule specifies a relation, directive, strategy, and recommendation. It is in the form of: if <antecedent clauses> then <consequent clauses> If the antecedent clauses are true, then the consequent clauses are true. 20

Some rules: n n n Rule: 1 If Growth_Rate 30% of a similar industry then Type = “Optimistic” Rule: 2 If 5% Growth_Rate 29% of a similar industry then Type = “Normal” Rule: 3 If -5% Growth_Rate 4% of a similar industry then Type = “Conservative” Rule: 4 If After_Merger_Business_Value 130%* Before_Merger_Business_Value then Strong_Suggestion_to_Buy Rule: 5 If 110%*Before_Merger_Business_Value After_Merger_Business_Value <130%* Before_Merger_Business_Value then Suggestion_to_Buy 21

n n n n n Rule: 6 If 100%*Before_Merger_Business_Value After_Merger_Business_Value <110%* Before_Merger_Business_Value then Suggestion_to_Wait Rule: 7 If Merger_Type = “Electronic Industry” then show Case 1 Rule: 8 If Merger_Type = “Chemical Industry” then show Case 2 Rule: 9 If Choice = “Discounted Free Cash Flow” then execute Model 1 Rule: 10 If Choice = “Economic Profit” then execute Model 2 22

The case base n n n Beside a rule base, the KDSS also uses a case base in the knowledge base. The case base records all past cases of mergers and acquisitions. For these cases, the basic financial data, procedures, policies, legal issues, and other related problems are evaluated, organized and saved in the case base. There are four pairs of successful merger and acquisition company, called X 1 Y 1, X 2 Y 2, X 3 Y 3 and X 4 Y 4, from Taiwan’s electronic and chemical industries. The four main cases in the current case base named CASE 01, CASE 02, CASE 03 and CASE 04. Each case links basic data, related regulations and taxes, procedures, expert audits, and agreements, all organized in a hierarchical format. 23

Figure 3. The hierarchical structure of a case representation 24

Case base reasoning n n Upon receiving a case, the KDSS first checks the case base to determine if a similar case exists. If so, the old case is used as guidance to solve the case. If not, the case is regarded as a new case, the whole process is performed. When the case is finished, it will be evaluated and added in the case base if it exhibits good performance 25

The model base n n n A model base in a DSS provides at least one management science model for analysis, evaluation, or forecasting. The result produced by the model will be used by the inference engine in the DSS to infer and suggest a feasible solution based on the rules in the rule bases. There a number of models can be used in evaluating business values: q q n Discounted cash flow (DCF) model Economic profit model Adjusted present value (APV) model Equity DCF model The first two models are most popular. Here, DCF model and economic profit model are used. 26

The discounted cast flow model Company’s Value = Present Value of Cash Flow During Explicit Forcast Period + Present Value of Cash Flow After Explicit Forcast Period When the business enters the mature stage, we can assume that g = 0 and the Eq. (1) can be simplified as follows: The definition of various variables in the above formulas are given as follows: (1) NOPLAT: Net operating Profit Less Adjusted Tax, (2) NOPLAT = EBIT(1 – Tax Rate) (3) (2) EBIT: Earnings Before Interest and Taxes 27

(3) ROIC: Return of Invested Capital ROIC = NOPLAT/Invested Capital (4) IC: Invested Capital IC = Net Operating Capital (Current Assets – Current Liabilities) + Net Property, Plant, and Equipment + Other Operating Assets + Short-term Investment + Goodwill Amortization+ Non-operating Investment. When the value of ROIC is high, it means the creative value on the investment is high. (5) g: Growth Rate in NOPLAT, g = ROIC Net Investment Rate, where Net Investment = Net Invested Rate NOPLAT (6) FCF: Free Cash Flow FCF = NOPLAT + Depreciation – Net Investment 28

(7) WACC: Weight Average Capital Cost, where WACC = Wd. Kd(1 -T) + Wp. Kd + Ws. Ks. Wd: Debt Ratio = Liabilities/(Liabilities + Preferred Stock Equity + Common Stock Equity). Wp: Preferred Stock Ratio = Preferred Stock Equity/(-Liabilities + Preferred Stock Equity + Common Stock Equity). Ws: Common Stock Ratio = Common Stock Equity/(-Liabilities + Preferred Stock Equity + Common Stock Equity). Kd = Cost of Debt = Debt i. Kp: Cost of Preferred Stock = (if the target company does not have this item, then ignore this item). Ks : Cost of Common Stock 29

Economic profit model Company’s value = Invested Capital + Present Value of Projected Economic Profit = Invested Capital The economic profit equals the spread between the return on invested capital and the cost of capital times the amount of invested capital. The model can be defined as follows: Economic Profit: = Invested Capital (ROIC – WACC) = NOPLAT – Capital Charge = NOPLAT – (Invested Capital WACC) 30

The reasoning mechanism n n n There are two kinds of reasoning. One is forward chaining, and the other is backward chaining. Backward chaining is goal-driven reasoning. In backward chaining, the inference engine first has a goal (hypothesis) and attempts to find evidence to prove it. The inference engine starts by searching for rules that might produce the desired solution and achieve the goal in the THEN (action) parts. Forward chaining is data-driven reasoning. Forward chaining does not start from a hypothesis, but with some confirmed findings. 31

Conclusions n n The KDSS system provides not only formal merger and acquisition procedures, relevant law and regulations, but also actions and feasible suggestions. By using the system, managers can easily deal with M&A decision-making problems via the Internet. The system consists of a database, case-base, rule-base and model-base. For assessing a business’s value, a discounted cash flow model and economic profit model are provided in the model base. 32

References n n 1. W. Wen, W. K. Wang, T. H. Wang, “A hybrid knowledge based decision support system for enterprise mergers and acquisitions”, Expert Systems with Applications, 28, 569 -582 (2005). 2. K. Pal, O. Palmer, “A decision support system for business acquisitions”, Decision Support Systems, 27, 411 -429 (2000). 33