Rafael Alvarez, Tax Fraud, New York 2020

Bronx Tax Preparer Pleads Guilty to Filing Tens of Thousands of False Tax Returns, Causing $145 Million in Fraudulent Tax Loss

A high-volume tax preparation company located in the Bronx, New York, was at the center of a decade-long, $145 million tax fraud scheme orchestrated by Rafael Alvarez, also known as “the Magician.”

Edward Y. Kim, the Acting United States Attorney for the Southern District of New York, announced today the guilty plea of Rafael Alvarez, a/k/a “the Magician,” to a two-count Superseding Information charging Alvarez with one count of conspiracy to defraud the U.S. and steal government funds and one count of aiding and assisting in the preparation of a false and fraudulent U.S. individual income tax return.

As part of today’s guilty plea, Alvarez agreed to pay the Internal Revenue Service (“IRS”) $145 million in restitution and forfeit over $11.84 million in fraudulent proceeds he received from his criminal conduct.

Alvarez, 61, of Cortland Manor, New York, pled guilty today before U.S. District Judge J. Paul Oetken to one count of conspiracy to defraud the U.S. and steal government funds, which carries a maximum sentence of five years in prison, and one count of aiding and assisting in the preparation of a false and fraudulent U.S. individual income tax return, which carries a maximum sentence of three years in prison.

Alvarez is scheduled to be sentenced by Judge Oetken on April 11, 2025.

Alvarez became known as “the Magician” by his customers for his supposed ability to make their tax burden disappear. But, as today’s guilty plea shows, there was no magic to what Alvarez was doing – he was committing a serious federal crime by falsifying tens of thousands of tax returns and, in the process, depriving the IRS of $145 million in tax revenue.

Today’s guilty plea, in one of the largest ever tax frauds by a return preparer, should serve as an important reminder to tax professionals that this Office will vigorously investigate and prosecute tax offenses.