The Compustat Landscape

Post your research in political economy, share ideas and workshop your empirical work.

Moderator: sanha926

The Compustat Landscape

Postby blairfix » Mon Mar 20, 2017 11:18 am

I have been working on a paper where I use the Compustat database to model the hierarchical structure of firms. I will not discuss the details here, but will share a new method that I have developed for visualizing the model.

In the plot below, individual firms appear as pyramids. The size of the pyramid is proportional to the firm's employment. Vertical levels correspond to hierarchical levels of the firm. Color corresponds to individual pay (individuals are visible as tiny vertical bars, if you look closely).

Of course, individual pay is not available for Compustat firms. I have simulated it using Compustat's data for total labor expenses and CEO pay. The model assumes that pay increases exponentially by hierarchical level. Given this assumption, each modelled firm shown below has mean pay and a CEO-to-average-employee pay ratio that matches Compustat data.

The image is very high resolution that appears to be too big for the CasP server. I've provided a link below


https://drive.google.com/file/d/0B_AS9iet5yFMcVdwZEtYVWRVQVU/view?usp=sharing


Blair
blairfix
 
Posts: 88
Joined: Fri Dec 20, 2013 3:45 pm

Re: The Compustat Landscape

Postby rsalisbury » Mon Mar 20, 2017 6:19 pm

Is there any meaning to the XY positions of the pyramids? It seems a bit hard to compare when they're not lined up or ordered. Maybe you can order by firm size on one axis and income or hierarchical levels on another axis.

Also, as a software developer, I'm interested in what you used to generate this. I'm not very familiar with scientific programming tools. Is it something in Python or R?
rsalisbury
 
Posts: 6
Joined: Fri Aug 28, 2015 4:40 pm

Re: The Compustat Landscape

Postby blairfix » Mon Mar 20, 2017 6:59 pm

In this plot the x-y axis have no particular meaning. But that's actually a good idea to sort firms. Here's a new version where firms are sorted by employment.

https://drive.google.com/open?id=0B_AS9iet5yFMa1R1OXJvT1FDM00

This graphic is produced with R using the scatterplot3d package. The graphing is actually the easiest part. The hard part (at least for me) was generating the data underlying the plot.
blairfix
 
Posts: 88
Joined: Fri Dec 20, 2013 3:45 pm

Re: The Compustat Landscape

Postby blairfix » Mon Mar 20, 2017 7:00 pm

In this plot the x-y axis have no particular meaning. But that's actually a good idea to sort firms. Here's a new version where firms are sorted by employment.

https://drive.google.com/open?id=0B_AS9iet5yFMa1R1OXJvT1FDM00

This graphic is produced with R using the scatterplot3d package. The graphing is actually the easiest part. The hard part (at least for me) was generating the data underlying the plot.
blairfix
 
Posts: 88
Joined: Fri Dec 20, 2013 3:45 pm

Re: The Compustat Landscape

Postby Rob_Cayman » Tue Mar 21, 2017 3:28 pm

After a cursory glance at this thread i feel compelled to ask if you are familiar with Henry Mintzberg's Competing Values Framework with regards to organisational cultures?

Briefly, Mintzberg did research in the field of business management back in the 70's and 80's, and identified 4 main organisational cultures that are prevalent in most organisations. Clan, Adhocracy, Hierarchy, and Market.

There has been a lot of research since then, but in essence these cultures are still present today.

If the organisations that you have used for your analysis are predominantly of the Hierarchy culture, it would stand to reason that you would see strong hierarchical tendencies in the organisational structures of these firms.

However, if the firms you have analysed can also be viewed as one of the other 3 cultures, it may offer some explanatory power in the differences you observe in the various hierarchies through out your sample.
Rob_Cayman
 
Posts: 1
Joined: Fri Aug 28, 2015 4:53 pm

Re: The Compustat Landscape

Postby rsalisbury » Tue Mar 21, 2017 6:11 pm

Ah, now I understand what you meant by organizational culture, Rob.

I haven't really announced it anywhere here but I run a Capital as Power Facebook group, and Rob and I were discussing this there. I think what he might be getting at, or at least what I'm now interested in, is the difference in span of control of these different firms. So like Celadon Group and CBeyond appear to have the same number of hierarchical levels but different spans of control. Perhaps that has to do with organizational culture, or some characteristics of the social niches they occupy?
rsalisbury
 
Posts: 6
Joined: Fri Aug 28, 2015 4:40 pm

Re: The Compustat Landscape

Postby blairfix » Tue Mar 21, 2017 6:20 pm

The span of control is identical for all firms modelled here. The actual hierarchical structure of these firms is not known. I've assumed a span of control based on an in-depth analysis of existing firm case studies.

But don't read too much into what you see in the graphic. There are many assumptions underlying it.
blairfix
 
Posts: 88
Joined: Fri Dec 20, 2013 3:45 pm


Return to Research

Who is online

Users browsing this forum: No registered users and 0 guests

cron