Hi guys!
My question is a bit long-winded, so please bare with me.
I'm doing research into M&A deals and post-acquisition innovative performance. I currently have two different kinds of data sets. The first is the data set on the deal level, showing target and acquiring firms, along with their patent assignee number, and the date that they were announced. (attachment 1)
The second is data on patents, shown per individual patent, and then some data I am interested in, such as assignee number, date, forward citations, classification code. (attachment 2)
My plan is basically as follows. I want to see how the knowledge overlaps, and non-overlaps, of the target and acquiring firms affects the post-deal performance of the acquiring firm.
To look at the post-deal innovative performance, I will choose a time frame of the total number of patents up to 1-3 years after the M&A deal. This is to account for the time it takes for the acquiring firm to absorb the knowledge and apply for the patent, while also considering the fact that the technology covered by a patent could lose most of its value within 5 years.
As to determine whether knowledge overlaps, I will consider the class of the patents of the target and acquiring firm. Those which are in the same class can be a sign of overlapped knowledge, while those in different classes are not overlapped. This can either be taken as a count variable.
Second, to see the quality of the overlapped and nonoverlapped knowledge components, I will consider the impact of each patent, which refers to the number of forward citations . Hence, the overlapped knowledge quality and nonoverlapped knowledge quality are calculated as the average impact of all the target firm’s patent in the overlapped and nonoverlapped patent classes.
Lastly, I will check for a moderating effect of M&A experience, simply including an interaction term with a dummy variable of experience (=1 if acquired in previous 5 years or so, =0 if not)
Hence, I want to work with firm-level data. (attachment 3 is a preview of how I want my data set to look like (including some control variables), for one particular acquiring firm) My question is then a guidance on which code to use/steps to take to get these variables. I was thinking a separate code to look at how an example target and acquiring firm would give me the variables. I was thinking to match the assignee code of the target and acquirer (or the name in some cases where the name was registered in PATSTAT) to get all the data per company. Then, from there, create the other variables by using commands looking at their respective class code of the patent.
Any help would be appreciated!
Kind regards,
Chris
My question is a bit long-winded, so please bare with me.
I'm doing research into M&A deals and post-acquisition innovative performance. I currently have two different kinds of data sets. The first is the data set on the deal level, showing target and acquiring firms, along with their patent assignee number, and the date that they were announced. (attachment 1)
The second is data on patents, shown per individual patent, and then some data I am interested in, such as assignee number, date, forward citations, classification code. (attachment 2)
My plan is basically as follows. I want to see how the knowledge overlaps, and non-overlaps, of the target and acquiring firms affects the post-deal performance of the acquiring firm.
To look at the post-deal innovative performance, I will choose a time frame of the total number of patents up to 1-3 years after the M&A deal. This is to account for the time it takes for the acquiring firm to absorb the knowledge and apply for the patent, while also considering the fact that the technology covered by a patent could lose most of its value within 5 years.
As to determine whether knowledge overlaps, I will consider the class of the patents of the target and acquiring firm. Those which are in the same class can be a sign of overlapped knowledge, while those in different classes are not overlapped. This can either be taken as a count variable.
Second, to see the quality of the overlapped and nonoverlapped knowledge components, I will consider the impact of each patent, which refers to the number of forward citations . Hence, the overlapped knowledge quality and nonoverlapped knowledge quality are calculated as the average impact of all the target firm’s patent in the overlapped and nonoverlapped patent classes.
Lastly, I will check for a moderating effect of M&A experience, simply including an interaction term with a dummy variable of experience (=1 if acquired in previous 5 years or so, =0 if not)
Hence, I want to work with firm-level data. (attachment 3 is a preview of how I want my data set to look like (including some control variables), for one particular acquiring firm) My question is then a guidance on which code to use/steps to take to get these variables. I was thinking a separate code to look at how an example target and acquiring firm would give me the variables. I was thinking to match the assignee code of the target and acquirer (or the name in some cases where the name was registered in PATSTAT) to get all the data per company. Then, from there, create the other variables by using commands looking at their respective class code of the patent.
Any help would be appreciated!
Kind regards,
Chris
Attached Files
Comment