How I Track Political Corruption In Latin America
Most people think tracking corruption means reading the news. It doesn't. The actual work is usually done in government procurement databases, campaign finance filings, and court records that are intentionally scattered across dozens of portals in different countries, many of them in languages and formats that were never designed for external scrutiny. I've spent years pulling together ownership records and spending data across the region. What follows is a practical overview of how the process works, what actually moves the needle, and where it consistently breaks down.
A Field Guide to Political Corruption In Latin America
Corruption in Latin America operates through several overlapping channels, and they rarely appear in the same report. The most common mechanisms are public procurement manipulation, campaign finance opacity, illicit enrichment through shell companies, and regulatory capture where oversight agencies are staffed by people with direct ties to the industries they should be monitoring. The procurement side is where the bulk of documented activity lives. A typical pattern involves a ministry issuing a tender with specifications so narrow that only one preselected company can meet them, combined with a evaluation panel stacked with people who have no formal conflict of interest on paper but do on the ground. Campaign finance is another major vector. Many countries passed laws requiring disclosure, but the disclosures often arrive months after the election, filed in formats that are not easily searchable or machine-readable, and with thresholds high enough that small bundles of money from a single beneficial owner can be split across multiple donors to fall under reporting limits.
Then there is the regulatory capture piece, which is harder to pin down but often more damaging over time. An antitrust agency hires its senior staff from the very industries it regulates. A central bank governor rotates out into a board seat at a commercial bank that had been under investigation. None of this is illegal by itself, but the pattern is consistent and it shows up across the region regardless of which party is in power.
Get the Full Details
What Most People Get Wrong About This Work
Beginners tend to look for smoking guns, a signed document naming a politician as the direct beneficiary of a contract. Those exist, but they are rare and usually buried in trial records rather than public databases. The more useful signal is structural: repeated contract awards to the same companies, overlapping directorship networks between bidders and regulators, or procurement cycles where winning bids cluster around election dates. Another common mistake is assuming that a country without active corruption scandals is free of it. Quiet periods in anticorruption enforcement usually mean one of two things: either the institutions are working independently, or the enforcement apparatus has been repurposed to target only political opponents while allies operate openly. The difference shows up in who gets prosecuted, not in whether prosecutions happen.
How I Actually Do the Research
The first step is mapping the institutional landscape of the country you are studying. You need to know which agency handles procurement, which one oversees campaign finance, which tax authority holds beneficial ownership records, and which courts keep the relevant case files. This is not always obvious from a government website. For procurement, I start with the national procurement portal and then cross-reference against the public registry of companies, looking for matching contractor names. Many countries publish award data, but the data quality varies wildly. In some cases the award values are missing entirely, in others the supplier tax ID is redacted, and in still others the data is available but only in PDF format with no search function. On campaign finance, the useful documents are the annual and pre-election filings. The trick is catching the filings before the statutory deadline expires, because once a cycle closes, agencies often archive older reports and make them harder to access. I set calendar reminders for each country's filing dates and pull the raw data immediately when it appears.
For ownership links, I use the corporate registry to build a basic directorship and shareholding network, then layer in court records, customs data, and any available audit reports. The goal is not to prove a single chain of custody but to identify whether the network itself has suspicious density, with the same intermediaries appearing across multiple unrelated contracts or contributions.
A Specific Case That Almost Broke My Method
During a project tracing infrastructure contracts in a mid-sized country, I kept hitting a wall with a particular contracting authority. The procurement portal listed winning bidders and values, but every contract above a certain threshold was marked as "confidential" under a vague public secrecy statute that had been amended multiple times. The data simply was not there. What I found instead was a workaround that I now use as a standard fallback: the country's supreme audit institution publishes annual audit reports, and those reports often reference specific contracts and discrepancies even when the procurement portal does not disclose them. The audit reports are usually written in less polished language, but they contain the contract numbers, the awarded amounts, and sometimes the names of the evaluating officials. I cross-referenced those audit findings with the corporate registry and was able to reconstruct a significant portion of the missing procurement chain without ever accessing the primary portal. This is not a guarantee. Some audit institutions are routinely ignored or defunded, and some countries do not publish their reports at all. But where the audit institution has even partial independence, its archives can be more useful than the procurement system itself.
Tools That Actually Help
DataHub for Brazil and similar transparency portals in other countries are useful, but they are not complete. They cover federal-level data in some cases and leave state and municipal spending largely opaque. I combine them with the OECD's anti-corruption monitoring databases, the World Bank's procurement records where available, and regional bodies like the Inter-American Development Bank, which disclose funded projects separately from national systems. For beneficial ownership, the Open Ownership database and national registries are the starting point, but the coverage is uneven. Several countries have laws on the books that require disclosure of ultimate beneficial owners, but enforcement is inconsistent and many entities still operate through layers of nominees with no public link to the real owners. Network visualization tools help when you have enough data to plot directorships and shareholdings, but they are only as good as the underlying records. Garbage in, garbage out is the rule here, not an exception.
Where This Approach Fails Completely
It fails when the state itself is the primary vector for corruption, because the relevant databases become instruments of opacity rather than transparency. Procurement portals are maintained by the executive branch, and when the executive branch benefits from keeping information hidden, those portals will reflect that preference through selective disclosure, slow publication, and frequent format changes that break automated harvesting tools. It also fails when corruption is informal and cash-based. Petty bribery at the municipal level, expedite payments at customs, and informal kickbacks on small contracts rarely surface in any public record. The work in those spaces depends on survey data, whistleblower testimony, or investigative journalism, not on database queries. A third failure point is political turnover. When a new administration comes in after a corruption scandal, it often deletes or restructures the transparency portals to make it harder to trace the previous government's deals. This is not always illegal, but it is deliberate, and it has happened more than once in the region. The workaround is to preserve snapshots of any portal you are monitoring, ideally through web archiving or local backups, because the public version may not survive the next election cycle intact.
What This Means in Practice
If you are doing this work for an organization, your output should focus on patterns, not single incidents. A single suspicious contract is an anecdote. A cluster of contracts awarded to companies with shared directors, evaluated by the same panel, during a period of weakened audit oversight, is a finding. If you are doing it independently, budget more time for data cleanup than for analysis. Expect to spend the first few weeks simply making disparate sources talk to each other, not because the data is bad, but because the standards differ so much between agencies that manual reconciliation becomes necessary. And if you find yourself stuck on a country where the official portals are effectively locked, look for the audit institution, the electoral tribunal, and any court proceedings related to the contracts you are investigating. The real records are often sitting in places nobody expects to find them, waiting for someone to ask the right question in the right archive.
