|
Every once in awhile we all dig around trying to figure out where we justified our current methods of rounding for HUD. Version 202C (the next version of the TRACS interface) includes Rounding guidelines for all financial data plus voucher adjustments to be implemented with 202C. We coded based on these specifications because there has been nothing definitive prior. The 202C spec Appendix H also mentions that in the transition to 202C, all CAs should accept adjustments within $1 of the number they compute until all vendors are on board with the new rounding. Some history: 202B (the currently implemented TRACS interface) originally included these guidelines, then was split into 202B and 202C, 202C is now delayed. Below is a link to get to the 202C spec yourself. http://www.hud.gov/offices/hsg/mfh/trx/trxdocs.cfm See Also: |
There are two main reasons why Bostonpost should continue to round voucher adjustments the way we do:
The 202C guidelines were defined with consistency in mind and are the only guidelines with specific instructions available.
Contract Administrators are repeatedly instructed by HUD to accept discrepancies within $1; TRACS itself has always been implemented to accept data within $1.00.
The most frequent point of contention is how to handle adjustments for a mid-month Gross Rent Change: In the past, convention (and intuition) has been to take the new contract rent minus old contract rent, and then prorate that delta change as the adjustment to subsidy. This results in every tenant having the same increase. The new recommended algorithm results in adjustment amounts varying by $1 between different tenants, depending on the prorating of their assistance amount.
The algorithm is:
Resulting Subsidy billed for the month = first part of month prorated at old amount + second part of month prorated at new amount.
HAP Adjustment = Old Subsidy amount for whole month - Resulting Subsidy billed for the month.
The philosophy behind this is: One algorithm works for all scenarios - for example, a Move-in Correction of a mid-month Move-in. Here you could take the new rent - the old rent and prorate the difference over the days of occupancy. However, philosophically you would want the ending result of a correction to be the same as if you had the correct information at the time you originally performed the move-in. The new algorithm will successfully compute the adjustment as if the move-in was originally calculated with the new rent.
-Janet Wittenberg, BostonPost Development