## Project-II (Serial confidence, Overconfidence and CEO’s Empire Buildings Case Solution

**Estimate the following regression model with OLS**

Dependent Variable: NB_ACQUISITIONS | ||||

Method: Least Squares | ||||

Date: 03/13/17 Time: 16:15 | ||||

Sample: 1 712 | ||||

Included observations: 633 | ||||

Variable | Coefficient | Std. Error | t-Statistic | Prob. |

GROWTH | 3.83E-05 | 3.07E-05 | 1.244547 | 0.2138 |

INSIDER | -0.010362 | 0.003165 | -3.273848 | 0.0011 |

AVERAGE_RELATIVE_SIZE | 0.037109 | 0.011306 | 3.282326 | 0.0011 |

TE_TA | 0.135573 | 0.061296 | 2.211784 | 0.0273 |

SIZE | -0.545269 | 0.159014 | -3.429061 | 0.0006 |

FIXED_ASSETS | -0.160198 | 0.341992 | -0.468426 | 0.6396 |

FINANCE | 0.288161 | 0.328596 | 0.876946 | 0.3809 |

C | 10.07546 | 1.498498 | 6.723702 | 0.0000 |

R-squared | 0.092650 | Mean dependent var | 5.483412 | |

Adjusted R-squared | 0.082488 | S.D. dependent var | 2.894800 | |

S.E. of regression | 2.772838 | Akaike info criterion | 4.890178 | |

Sum squared resid | 4805.395 | Schwarz criterion | 4.946424 | |

Log likelihood | -1539.741 | Hannan-Quinn criter. | 4.912020 | |

F-statistic | 9.117002 | Durbin-Watson stat | 0.129146 | |

Prob(F-statistic) | 0.000000 | |||

**Evaluate the explanatory power of the model**

According to the linear regression analysis conducted above, the probability calculated for the explanatory powers of the model. It can be evaluated that, the fixed assets, finance and growth were the variable with the highest probability values in the entire model. Which meant that, these variables would have the highest probability of generating values in the model, with respect to the dependent variable (NB acquisition). Whereas, the insider and average relative size variable amounted towards the same probability values. Which mean, they had an equal chance of generating values in the model, and the size variable had the least chance of generating value in the model.

**Examine the overall significance of the model**

It can be determined, after examining the significance of the entire model that, the significance values generated under the linear regression model amounts to at 9.11. Which was significantly higher, than the benchmark set at 0.05. Which meant that, there was sufficient evidence present to prove the null hypothesis and therefore, it should be accepted. In which, the null hypothesis was formed on the basis that, the NB acquisition would be affected by all the other variable present in the dataset. Hence, it was estimated that, the null hypothesis was true.

**Verify the different ideas suggested by the literature regarding the determinants of multiple acquisitions**

It can be determined, through the linear regression analysis conducted above that, the different idea suggested to impact all other variables on the NB acquisition realized by the bidders. As far as, the growth variable is concerned.It can be determined that, sales growth over the yearscould have a significant impact on the NB acquisition, attributed to its high probability values generated through the linear regression model. Furthermore, the bidder ratios of total equity to total assets also have a significant impact on the NB acquisition, attributed to its high probability value generated. Moreover, the fixed asset to total assets determinant of the bidders could also have a significant impact on the NB acquisition. As all three of these variables significantly impacted the NB acquisition. However, it was estimated that, other variables presents in the dataset did not necessarily impacted the NB acquisition, as much as the variable mentioned above.Which could be attributed to their low probability value generated in the linear regression model.

**What is the main weakness of the use of OLS to estimate this model? Why?**

The main weakness of the use of OLS pertains is its inability to effectively consider, relevant other factors such as market risks and market volatility. Which could significantly contribute towards increasing the overall NB acquisition realized by the bidders. Whereas, the market volatility and risk increases, it decreases the bidders appeal to invest in the market, as they believe that the probability of incurring more losses was higher in a risky and volatile market.

**Use the Poisson regression methodology to estimate the model presented and evaluate its explanatory power and its overall significance.**

Dependent Variable: NB_ACQUISITIONS | ||||

Method: ML/QML - Poisson Count (Quadratic hill climbing) | ||||

Date: 03/13/17 Time: 17:08 | ||||

Sample: 1 712 | ||||

Included observations: 633 | ||||

Convergence achieved after 5 iterations | ||||

Covariance matrix computed using second derivatives | ||||

Variable | Coefficient | Std. Error | z-Statistic | Prob. |

GROWTH | 8.38E-06 | 5.85E-06 | 1.431420 | 0.1523 |

INSIDER | -0.001402 | 0.000395 | -3.548549 | 0.0004 |

AVERAGE_RELATIVE_SIZE | 0.005868 | 0.001553 | 3.777745 | 0.0002 |

TE_TA | 0.019627 | 0.007992 | 2.455775 | 0.0141 |

SIZE | -0.106739 | 0.024543 | -4.349116 | 0.0000 |

FIXED_ASSETS | -0.034911 | 0.052225 | -0.668462 | 0.5038 |

FINANCE | 0.051771 | 0.050958 | 1.015948 | 0.3097 |

C | 2.611692 | 0.229703 | 11.36988 | 0.0000 |

R-squared | 0.088069 | Mean dependent var | 5.483412 | |

Adjusted R-squared | 0.077855 | S.D. dependent var | 2.894800 | |

S.E. of regression | 2.779829 | Akaike info criterion | 4.971510 | |

Sum squared resid | 4829.655 | Schwarz criterion | 5.027756 | |

Log likelihood | -1565.483 | Hannan-Quinn criter. | 4.993353 | |

Restr. log likelihood | -1607.073 | LR statistic | 83.18041 | |

Avg. log likelihood | -2.473117 | Prob(LR statistic) | 0.000000 | |

According to the Poisson regression methodology, it can be determine that, the explanatory variables growth, fixed assets and finance had the highest probability values generated in the model. Which meant that, these variables were more likely to generate value in the entire model, compared to other variables present in the dataset. However, other variable generated significantly lessor probability values, compared to the variables mentioned above. On the other hand, the significance value generated under the model, was higher than the benchmark set at 0.05 at 83.18. Which meant that, the null hypothesis selected was true, attributed to the significant amount of evidence present, and hence it should be accepted...................................................

This is just a sample partial case solution. Please place the order on the website to order your own originally done case solution.