Leveraging Open Source Data: a Panama Case Study
Leveraging open source data can be a critical complement to a thorough investigation in Latin America. As an example, on a recent engagement, we understood from numerous contacts that the subject, a consumer goods company, had particularly close political ties to the previous government of Ricardo Martinelli (2009 to 2014) and had found itself unable to obtain the same sort of contracts from the current government of Juan Carlos Varela (2014 to 2019).
Using online data from the General Department of Public Contracting, we found that the total amount of the company’s public contracts had declined 74% between Martinelli’s term and Varela’s term to date. In fact, the company had obtained an average of 78 contracts per year during Martinelli’s term versus about 20 per year in Varela’s term to date. In effect, this data provided some level of objective verification of both source comments and of the company’s political connections.
It’s clear that public tenders in Latin America are often an area of potential anti-corruption concerns. As we have discussed previously, it often makes sense to include some analysis of tender data as part of a due diligence in the region. Looking at this data over time may reveal interesting trends in the volume, type, and dollar amount of contracts awarded that can complement the efforts of a larger investigation and provide insights into the target of the investigation.
Bearing this in mind, we sought to carry out an illustrative analysis of an actual company in Panama, which we will call “Company A,” using a combination of public and private databases that index public tender data from Panama. Company A is a large commodity distributor reportedly owned by a powerful family with political ties to the current government and by a high-ranking current government official.
The first and most notable trend is the sudden change in the volume of contracts received by Company A. From 2009 to 2013, during the bulk of Martinelli's term (Varela took office on 1 July 2014), the company received about 16 contracts. In contrast, nearly three years into the Varela government, the company has received over 300 contracts. The trend in contracts per year is illustrated in the below chart.
Using a third-party database that extracts online data from the General Department of Public Contracting, we proceeded to analyze the trend over time in total dollar amount of contracts awarded per year during the Varela government. Not surprisingly, we found a corresponding trend of increasing total dollar amounts per year awarded to Company A as illustrated in the below chart.
Between 2014, when the current government began, and 2015 there was a 2,223% increase in the total dollar amount of contracts awarded. From 2015 to 2016 there was a 1,187% increase. 2017 is on track to being a record year with dollar totals for the first half of the year nearly equaling the total for all of 2016. In other words, there is an unmistakable trend toward increased contracts during the current government, which would appear to confirm Company A’s political alliances.
Having noted an overall trend, we can proceed to an analysis of the numbers sourced from the third-party database. With just over 300 records between 2009 and the first half of 2017 and entries as low as two digits, a Benford Test would not be all that interesting or useful. However, basic data analysis techniques can yield interesting insights once one has cleaned up the data and copied it to Excel.
For example, given Company A’s position as a commodities distributor, one may want to isolate round dollar payments using the MOD function. Not surprisingly, we find that the bulk of these payments have occurred during the current government, when the majority of Company A’s contracts were awarded; they include three 2016 contracts for $1,000.00 each from a single government Ministry.
One might also decide to use a string function to isolate payments that end in say, “0.99” on the theory that contracts with these amounts might also warrant further scrutiny or might be aimed staying below a certain limit. We find there are there are a total of four such payments, which, interestingly, include a payment for $999.99.
The general idea is that this analysis can complement a broader investigation of Company A. In the case of the $1,000.00 contracts, one might want to look further at whether these are direct contracts and whether anyone at Company A has connections to someone at the Ministry in question. In the case of the $999.99 contract one might want to verify potential relationships between Company A and the government agency involved as well as reviewing the contract documents/specifications and applicable legal limits on contract amounts, if any.
Caveats and Conclusions
On their own, the above trend and data analysis do not constitute an indication of any improper conduct. However, it’s clear that looking at the volume, type, and dollar amount of contracts awarded to the target of, say, a diligence investigation, can yield potentially interesting insights and complement information gathered in the broader investigative effort.
Any analysis of contract amounts should be done carefully. Data must be properly formatted and time invested in thinking about what sorts of analyses make sense given the company’s profile and the size of the data set. For instance, while round dollar amounts might be considered unusual for a company selling products, they would not necessarily be unusual for a company selling consulting services.
It’s also worth noting that data quality and ease of collection can vary significantly across jurisdictions in Latin America. Even Panama’s public contracting website can prove difficult for extracting certain data. When using third-party sources, one should always be careful to keep in mind that data may contain errors or omissions.
All of the information gathered should be viewed in the context of the results of the broader investigation.