Corporate Governance Explanations effects on Firm Value

Hi all,

I'm quite new to statistics and need some guidance for my dissertation. The University is messing up with my supervisor resulting in me working on my own for the past month on this... What I now need is to decide on the research methodology. But I'm not quite sure what would be the best type of regression to apply, and what the major concerns would be. I'm hoping someone will be able to help me in the right direction so I can start looking for specific information on how to proceed.

Basically what I'm trying to do is measure the effect of a firm's explanation of their non-compliance to the Dutch code of corporate governance on the firm's value (using Tobin's Q ratio).

Some backgrounds information: The Dutch code is full of provisions with best practices that companies are to apply or explain why they don't.

I will categorize these explanations in categories such as: "it is a temporary deviation of the code", "laws in other countries the firm is active in do not allow for the application of this best practice", "too costly to apply", etc. (a total of 8 categories).

I would control for certain factors such as firm size and leverage ratio. and the sample size would be the top 75 traded companies based on the AEX, AMX and ASX indices.

Goal of the research is to measure whether investors care solely about the application rate of the code (i.e. 94% of the codes best practices have been applied) or care about the (type of) explanations of non-appliance as well. As a lot of explanations provided are of low quality that don't say much (i.e. "we do not apply this best practice as the board decided against it") I felt the need to distinguish between them.

The reason the code is voluntary is that it gives the freedom to firms to use other 'best practices' that are more suitable for that type of firm.

One issue I'm struggling with is how I would offset appliance vs explanation. (i.e. firm Y applies 90% and all 10% explanations are of category A, firm X applies 80% and all explanations are of category B). If X is now worth less than Y, we still don't know if it's due to the explanation or due to the level of application of the code. Of course realistically a company would offer various explanations that fall into different categories.

I'm sorry for the unorganized post but I'm just hoping on some guidance here to point me in the right direction so I can keep moving.

Feel free to ask more questions and raise any concerns you might have :)

Hi all,

I suppose my previous post was somewhat generic and unorganized, but I have advanced somewhat since then so hopefully someone can provide me with some insights.

In brief, I am running one dependent variable (a ratio of firm value) against 4 IV's:

1. Appliance rate of the code in %
2. Number of category 1 explanations given for non-appliance
3. Number of category 2 explanations given for non-appliance
4. Number of category 3 explanations given for non-appliance

IV 1 will be highly correlated with IV 2 - 4; 100% compliance means no explanations given and as the appliance rate gets lower, the number of explanations will rise.

To solve this I think I can use Principle Component Analysis. This can reduce the number of variables and address the multicollinearity. Basically I would want to get rid of the appliance rate, as that is the correlator; between the explanations themselves there should be no correlations.

Goal of the research is to find out whether the categories of explanation influence firm value, so I don't want to get rid of those variables.

So can I use PCA to 'integrate' the appliance rate into each of the explanation variables? This would reduce the correlation, although it wouldn't fully get rid of it as now the new composite factors for the explanations + appliance rate would be more correlated as they would rely on the 'base rate' of appliance rate.

Hoping someone can help me or at least tell me if I am barking up the right tree here :confused: