Why Raw Crime Numbers Lie to You
I spent three years pulling municipal data from Baja California's public security database. The first thing I learned was that the numbers published online are rarely what people think they are. You need to understand the methodology before you trust any headline or blog post about Tijuana crime trends. Let me walk through how this actually works in practice. The state of Baja California publishes monthly crime reports through their Sistema de Indicadores de Seguridad. These reports come in two main flavors: reported crimes (hechos delictivos) and cleared crimes (delitos aclarados). Most people conflate these two metrics, which creates completely wrong conclusions about whether violence is rising or falling. Here is a specific problem I ran into that most people never notice. The monthly reports don't break down homicide by municipality when the state-level numbers are already published. Tijuana is just one of many municipalities in Baja California. If the state report shows 400 homicides, Tijuana might account for 280 of them or it might account for 350. There is no official monthly breakdown. I spent weeks trying to find a consistent way to estimate Tijuana's share of state-level homicide numbers until I found that the municipal police chief's press briefings occasionally released weekly figures that could be cross-referenced against the monthly totals. I built a simple spreadsheet that tracked those weekly estimates and used them to create a weighted distribution model for each month. It is not precise. It is the best approximation available without insider access to internal reporting systems.
The second counter-intuitive thing nobody mentions is that reported crime rates often drop when a municipality implements better digital reporting systems. This sounds backwards but it makes sense. When a city moves from paper-based crime reports to an online portal, previously unreported incidents start appearing in the statistics. A legitimate improvement in public safety can look like a crime spike for the first six to twelve months because the denominator is catching up to reality. I saw this play out with petty theft data across three separate municipalities. The numbers went up first, then stabilized, then started declining. Anyone who looked only at the initial spike would have drawn the wrong conclusion entirely. There is also a seasonal pattern that is largely ignored. Tijuana's crime statistics show a consistent bump in August and September tied to the migration corridor activity. This is not a conspiracy or a policy failure. It is basic supply and demand for unauthorized border crossings creating secondary criminal opportunities. Organized theft rings target migrants because migrants do not report to police. The data reflects this every year without fail. If you are comparing January to July you are comparing different populations with different reporting behaviors. The biggest pitfall I see is people using CNNIC (Centro Nacional de Indicadores de Criminalidad) data without understanding that it aggregates multiple sources with different methodologies. Some municipalities use police reports, others use prosecutor office data, and a few mix both. The aggregation process smooths out differences between these sources but it also obscures the actual trends. My workaround was to pull the raw municipal reports directly from each ayuntamiento's transparency portal whenever possible and compare them against CNNIC figures. The discrepancies between the two sources were usually in the 8 to 15 percent range, sometimes higher for specific crime categories like extortion where local governments had strong incentives to underreport.
Another issue with these statistics is that they do not capture victimless crimes or domestic violence incidents that never reach law enforcement. Tijuana's domestic violence numbers from the Instituto de las Mujeres de Baja California show consistently higher rates than what appears in the criminal justice statistics. This gap exists everywhere but it is especially pronounced in border cities where economic pressure and displacement create conditions that suppress reporting. If you want the raw data yourself, the primary source is the Portal de Transparencia at transparencia.gob.mx. Baja California's state security secretary also maintains a dedicated data section at seguridad.gobmx.baja.gov.mx. The municipal level data for Tijuana specifically is available through the Ayuntamiento de Tijuana's open data portal, though the interface is inconsistent and the data quality varies by year. I recommend downloading the CSV exports rather than using the interactive dashboards, which filter and aggregate data in ways that are difficult to reverse-engineer. The main limitation of all of this is that the data stops being reliable when you go back before 2019. There was a systematic change in how crimes were classified and reported during the transition between state administrations. Homicide classifications shifted from "asesinato" to "murder" for international comparison purposes, and some categories were merged or split. Trying to build a time series that spans multiple administrative terms requires you to manually reconcile these definitional changes, and the official documentation for those changes is scattered across press releases and ministry bulletins rather than centralized anywhere useful.
Get the Full Details
I usually suggest that anyone doing serious analysis start with the last five years of data and work backward cautiously. The methodology stabilizes after 2020, and the patterns you can identify there are far more trustworthy than anything you construct by stretching back a decade. The numbers tell you something real, but only if you understand what they are actually measuring and where the gaps are.
