All the good in this article is to the credit of the Holy Spirit and all the bad is entirely my own doing. For a while now I have been blessed with accurate forecasts of drug seizures. Any economists interested in working on Drug Data Science should email me at nicholsezekiel@gmail.com as I have been fervently seeking some help with drug research. This article is on methamphetamine seizures with potency greater than 85%. The model uses a vector autoregression to forecast methamphetamine seizures from DEA STRIDE (System to Retrieve Drug Evidence) data. The more I have worked on drug modeling the more I have found that macroeconomic variables have a profound effect on drug seizures. I am limited in my market data to using law enforcement data and as it has been stated in the literature this data is not a random sample of drug market data but rather a measure of cases brought by the DEA. As to the differences between this data and market data, only the Lord knows.

​

The best model consists of the log difference of the case count of methamphetamine seizures of a kilo or more of 85% potency or greater equal to the log difference of the case count of small quantity seizures of 4 grams or less, the log difference of total net grams of small quantity seizures of 4 grams or less and the log difference of the M1 currency supply. The out of sample forecast for Q42014 to Q32015 with the actual numbers. You can see it is extremely accurate with a MAPE of 7.5.  This shows that law enforcement seizures of drug trafficking can be accurately forecast.

Quarter    Prediction    ActualKiloOrMoreIceMethamphetamine    Percentage Error
2014Q4    95.71656                             97                                          1.323134021
2015Q1    83.78989                             103                                       18.65059223
2015Q2    106.1235                              112                                         5.246875
2015Q3    108.5175                              114                                         4.809210526
           Average Out of Sample Mean Percentage Error                    7.507452945

​

Here are my do files and DEA STRIDE Data acquired by FOIA request to the DEA.

Meth Predictors Do file Stata 13

High Quality Meth Seizure VAR Do file Stata 13

DEA STRIDE DATA Stata 13

​

Blessings, Peace and Grace to you all!

​