How the Estonian Tax and Customs Board Evaluated a Tax Fraud Detection System Based on Secure Multi-party Computation

被引：36

作者：

Bogdanov, Dan ^{[1
]}

Joemets, Marko ^{[1
]}

Siim, Sander ^{[1
,2
]}

Vaht, Meril ^{[1
]}

机构：

[1] Cybernet, Tallinn, Estonia

[2] Univ Tartu, Inst Comp Sci, EE-50090 Tartu, Estonia

来源：

Financial Cryptography and Data Security (FC 2015) | 2015年 / 8975卷

关键词：

Tax fraud detection; Risk analysis; Secure multi-party computation; Case study;

D O I：

10.1007/978-3-662-47854-7_14

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

The Estonian Tax and Customs Board (MTA) has identified that Estonia is losing over 220 million euros a year due to avoidance of value-added tax (VAT). The parliament proposed legislation that makes companies declare their purchase and sales invoices for automated risk analysis and fraud detection. The law was vetoed by the Estonian President on the grounds of confidentiality breach and unnecessary burden to companies. In this paper, we report on our collaboration with MTA to build a tax fraud detection system prototype that uses secure multi-party computation (SMC) to remove the companies' concerns over confidentiality. We estimate that the prototype could process a month of Estonian VAT data in ten days running on 20 000 euros worth of hardware.

引用

页码：227 / 234

页数：8