2 ms·
>What the authors are saying (I think) is that they tried the same tricks on the Panama data and got nothing useful out of it. The standard graph techniques don
by nobody9999 6y ago
>What the authors are saying (I think) is that they tried the same tricks on the Panama data and got nothing useful out of it. The standard graph techniques don’t explain anything useful about what’s going on.
The impression I got was not that they found nothing useful, but that they didn't find broadly interconnected networks within the data.
If that's the case, it's likely that the transactions in the data aren't tightly linked.
Which is useful, in that it tells us that those involved in such activities are not a cabal of evil elites, rather they're disparate groups with little in common except their desire to avoid taxes/hide assets.
That conclusion certainly implies that addressing these activities will be more difficult as they are decentralized and widespread, but it's certainly not useless IMHO.
- ColanR 6y ago> that they didn't find broadly interconnected networks within the data. That seems to agree with what the parent said. It sounds to me like if a broadly interconnected network was found, that would be a failure on the part of the people trying to facilitate the tax avoidance schemes - the closer to homogenous noise the graph looks, the harder it is to trace. Based on that, I'd say that there's no utility in trying to prove the negative, i.e., that larger networks do not exist. If the people hiding their assets actually did their job well, they would expect and hope that you (as an outside analyst) would come to exactly the conclusion that you have come to.
- newen 6y agoIt just means that they didn’t find useful network structures with the specific network analysis methods they used and the specific networks that they constructed using the data. It leaves room for tons of other network analysis and data analysis.