Posted by Bill Dorotinsky
As donors and governments spend enormous sums of money on automating financial management information systems in countries around the world, a great deal of attention is paid to the hardware and computer software. Debates over off-the-shelf or customized software, getting the right hardware, and setting up the right procurement arrangements, would fill volumes. But some recent cases illus rate the importance of paying attention to the 'soft' systems that surround the automated system -- the human systems that both serve and are served by the FMIS, the rules and controls embedded in and surrounding the systems.
On February 17, 2007, the Ugandan Daily Monitor broke a story ("Exposed: money racket in government") on an alleged scam being run from the Ministry of Finance and Ministry of Public Works, where fictitious suppliers where created in the computerized database, and payments were made to these suppliers for fictitious goods and services delivered. The article describes some of the means used to skirt what controls did exist, and but also illustrated that there were other controls missing.
A continent away, in Washington, D.C., a similar story unfolded in the District of Columbia tax office, where sham corporations were recorded in the automated information system, and tax refunds approved for these corporations. In some cases, the properties did not even exist, indicating absence of automated data-matching with property records or tax records to help control for such abuses. (See , for example, the November 14, 2007, Washington Post article, "D.C. Tax Scam Could Total $32 million".)
Of course, collusion among key officials can overcome the best controls, and no system, automated or manual, can completely prevent abuse. But closer attention to the soft systems surrounding the automated systems can help reduce the risk of collusion and corruption. For example, annual internal or external audit review of control systems and system operation, or examination of refund patterns for anomalies, can help identify weaknesses before they mushroom into large losses for taxpayers.
Such soft-systems include the internal controls in government agencies and treasuries to identify irregularities and seek supporting documentation. For Uganda, some of the internal controls were working. The scam came to light when the internal audit office of the ministry of public works sought verification of previous payments and services rendered.
For D.C., the internal controls clearly broke down. The alleged masterminds operated with little or no managerial oversight, and internal audit did not detect the irregularities. The scam apparently came to light when an alert bank employee questioned a tax refund payment for a company that did not appear to exist, and alerted officials. The D.C. case suggests the importance of more transparency and enlisting vendors in helping police the system.
More broadly, when automated systems are designed and installed, more attention to these procedures and systems would be valuable. Too often, as suggested above, the hardware and computer software are the focus of attention. But it is the human dimension, and the rules and procedures put in place, that will ultimately determine if the automation is a good investment.