IRS Criminal Tax Statistics (4/3/17)

The IRS has issued its 2016 Data Book.  The pdf for the Databook is here.  I have extracted here the part under the heading "Collections, Penalties and Criminal Investigations."  In this extraction, there are two pages of graphics and information and then the next pages offer the following tables:

• Table 16. Delinquent Collection Activities, Fiscal Years 2015 and 2016
• Table 17. Civil Penalties Assessed and Abated, by Type of Tax and Type of Penalty, Fiscal Year 2016
• Table 18. Criminal Investigation Program, by Status or Disposition, Fiscal Year 2016

IRS CI also offers its Fiscal Year 2016 statistics here and its cumulative statistics for the years 2007 - 2016, here.

I have been extracting tax crimes data from both the IRS Annual Data Book and annual report of statistics since 2005 and offer it here.  I have been keeping in a spreadsheet a subset of these statistics since 2005 (note that I have two years not included in the IRS offering linked above).  The spreadsheet is here.  Here are the cumulative statistics for 2005-2016 and for 2012-2016:

FROM THE IRS DATA BOOK
LEGAL SOURCE STATISTICS
		2005-2016	2012-2016
1	Indictments	12,353	6,801
2	Convictions	10,850	6,093
3	Percentage Convicted (l. 2 / l. 1)	87.8%	89.6%
4	Sentenced	10,583	5,842
5	Incarcerated	8,396	4,642
6	Percentage Incarcerated (l. 4 / l. 5)	79.3%	79.5%
FROM THE IRS CI WEBSITE STATISTICS:
EXCLUDING FINANCIAL CRIMES
		2005-2016	2012-2016
1	Indictments	17,072	8,695
2	Sentenced	15,012	7,544
3	Percentage Sentenced (l. 2 / l. 1)	87.9%	86.8%

I have not tried to reconcile these statistics (I am sure it is in differences in the data in each set.)  I infer, however, that, for the mainstream tax crime, the conviction rate is far less than the general conviction rate of 95% touted by DOJ Tax.  Of course, some tax crimes are charged without an IRS investigation, but DOJ Tax would have to have one hell of a conviction rate on those crimes to move the overall conviction rate up from the convictions obtained from CI investigations -- again, if only mainstream tax crimes are considered.  And, in any event, at least from my practice, the tax crimes cases that I have been involved with have involved CI investigations except in two prominent instances.  For most tax practitioners doing this type of work, the cases will generally progress from an IRS CI investigation (which would cause the ultimate results to be in the CI statistics), although they may stop for some investigation by the grand  jury (in which case they would also be in the CI statistics).

I would appreciate hearing by comment or email from those having more knowledge of how DOJ Tax calculates its statistics and the differences between the IRS statistics and the DOJ Tax statistics.