School District Income Verification Program

 

Year 11 Summary Report

   

State of New York

George E. Pataki, Governor

 School District Income Verification Liaison Committee

January 2005

 

Executive Summary

  

Since 1980, personal income wealth has been used in the calculation of state aid to public school districts.  From that time on, income wealth has been attributed to school districts based on the school district codes reported by residents on their personal income tax returns.  Concern over the accuracy of taxpayer reported school district codes resulted in the creation of a statewide system for the assignment of school district codes to taxpayer addresses.

 

First implemented in accordance with Chapter 50 of the Laws of New York State (1993), the statewide School District Income Verification (SDIV) Program has continued to demonstrate that a statewide computerized matching system is an appropriate and cost effective means of assigning school district codes to taxpayer address records.  The current system was first authorized by a 1994 amendment to Section 171, subdivision twenty-fifth, paragraph d of the Tax Law, and assigned school district codes to 1995 personal income tax returns for the purpose of aggregating income data to be used in determining state aid for education for the 1997-98 and 1998-99 school years.  School district codes assigned to 2002 tax returns will be used for determining state aid for the 2005-2006 school year as required by the Tax Law.  As also required by the Tax Law, a Memorandum of Agreement (MOA) governing the program was executed by the Executive Director of the Office of Real Property Services, the Commissioner of Taxation and Finance, and the Commissioner of Education and was approved by the Director of the Budget on October 17, 2003.

 

The system developed by the Office of Real Property Services (ORPS) under the direction of an interagency Liaison Committee was again able to correct or newly assign school codes to over 750,000 (from Appendix B) taxpayer addresses.  As in previous years, districts were provided an opportunity to participate in the review of several categories of address records attributed to their district.  This voluntary review was undertaken by 247 New York State school districts in Year 11 whose aid is dependent upon personal income wealth.  Any changes recommended by a participating school district as a result of the review were implemented if they were certified by the District Superintendent (DS) of the appropriate Board of Cooperative Educational Services (BOCES).  Of the 723,598 address records for which school codes were corrected or newly assigned, 11,232 resulted from this school district address review process.  The remaining 712,346 addresses had school codes corrected or newly assigned by one of the automated methods of the program.

 

A comparison of correction rates over the eleven-year period of the SDIV program is illustrated in Figure 1.  System corrections refer to instances in which the taxpayer address is attributed to a different school district than the first one reported by the taxpayer.  These corrections may be due to computerized assignment methods or to DS corrections described above. These corrections, when divided by the total number of address records processed, yield a “system correction” rate expressed as a percentage.  The system correction rate increased from 5.80 percent in Year 1 to 11.82 percent in Year 5, before dropping to 7.69 in Year 6.  However, after the relatively flat trend of Years 7 and 8, the system correction rate climbed back to 10.5 percent and 9.9 percent in Year 9. Based upon the system rates of the last years, it appears that these correction rates may have plateaued, and may in fact be dropping further in the future.

 

 

The District Superintendent correction rate considers DS corrections to a select group of "anomalous" addresses in the address review phase of the program only.  The DS correction rate is the ratio of DS corrections to the total number of addresses changeable by participating districts.  These address codes are either initially assigned by ORPS’ computerized methods and differ from taxpayer reported codes, are previous year corrections to the same address, or are those which can not be assigned by the previous two categories.

 

As the figure reveals, since Year 2, when the DS correction rate was 6.89 percent, there has been a steady but slowing reduction in DS error correction rates, which were down to 2.00 in Year 7.  Since then, there has been a generally upward trend in the D. S. correction rates.

The taxpayer address information provided by the Department of Taxation and Finance (DTF) was used solely for the purpose of assigning school district codes corresponding to the permanent residential address of taxpayers.  The information was divulged only to persons with the need to know in furtherance of this purpose.  All persons permitted to inspect the information were required first to sign DTF's "Agreement to Adhere to the Secrecy Provisions of the Tax Law and the Internal Revenue Code" (Form EN-202).  This extra measure provides further confidentiality (“secrecy”) assurance of taxpayer information.  As an aid to understanding the terminology of this report, a glossary appears in Appendix D.

Phase 1: School District Boundary Review/Correction

 

The SDIV program uses taxpayer supplied address information in concert with digital school district boundaries to assign school district residence codes to these addresses. The first phase of the program allowed districts to review and correct as necessary, their boundaries as shown on ORPS produced maps.  In October of 2002, maps were sent to districts for review.  Most districts (634), in the ORPS GIS had complete digital tax map (DTM) boundaries, that is, boundaries taken from approved digital or paper real property tax maps.  Others (34), have boundaries composed of both DTM and non-DTM portions.  While, still others (15 - excluding New York City districts), provide boundary data from only non-DTM sources.  To correct DTM boundary portions, districts sent ORPS paper or digital tax maps that portrayed the correct boundary portion(s). To correct non-DTM portions, districts manually drew approximate boundary corrections with the written consent of the affected district(s).  The following counties did not have complete DTM school boundaries in the ORPS GIS at the time maps were created for review:

 

·         Westchester (certain towns only)

·         New York

·         Queens                                                   

·         Richmond

·         Onondaga (one district)

·         Kings

·         Bronx

 

 

DTM boundary portions for districts in these counties will be incorporated into the ORPS GIS as they become available.

 

DTF Address Processing

 

DTF performed address “scrubbing” to correct and standardize address components of the 8,095,223 records that made up the tax year 2000 PIT data transmitted to ORPS for Year 11 processing.

 

In October 2003, DTF prepared address labels for New York State personal income tax forms for the over eight million taxpayers who reported their 2003 income.  In order to increase the accuracy of taxpayer school district code reporting, the codes printed on the mailing labels were updated using the corrections made in the Year 10 SDIV program.

 

Phase 2:  Assignment of School District Codes by ORPS

 

ORPS Address Scrubbing

As mentioned in the DTF Address Processing section above, DTF supplied ORPS with taxpayer addresses which had been scrubbed by DTF software.  ORPS then scrubbed the taxpayer addresses using a different software package in order to delineate address components for various stages of processing.

ORPS Assignable Address Records

As a result of the address scrubbing, 7,636,459 address records were identified as ORPS "assignable," i.e., records eligible for assignment of a final school district code unless changed during the school district address review.  A total of 469,052 address records were deemed "unassignable" or ineligible for address assignment.  These included two categories of unverifiable records and the previous year corrected records (see Appendix B).

The first unverifiable category consists of addresses to which mail is delivered but which are either outside of New York State or are those at which no taxpayer physically resides.  This group includes 445,072 out-of-state, post office box, general delivery, and business firms or non-residential addresses.  A school district code assigned based on the location of these addresses would not be reliable because persons using these addresses either live out of state or choose to have their mail delivered to these addresses but do not necessarily reside there.  These 445,072 address records, whose taxpayer reported school codes were not changed by ORPS processing, were excluded from the Phase 3 School District Address Review.

The second category of unverifiable address records consists of mailing addresses which correspond to actual taxpayer residence addresses, but which are not specific enough to accurately determine their geographic location by automatic processes. This group is made up of rural route and highway contract addresses.  Since it is possible for a rural route or highway contract to cross school district boundaries, any school district code assigned to these type of addresses by the current SDIV processing methodology on an automated basis would be unreliable.  However, since the SDIV liaison staff in these rural districts are often familiar with these addresses and their locations, the 23,066 rural route and highway contract addresses identified statewide were eligible for review and correction by school districts participating in the address review on the Unassigned lists.

As part of the address scrubbing process, an address record type code was assigned which provided information about the nature of the address. This was used to identify unverifiable address records.  The following table shows the breakdown of ORPS unassignable address records that were identified:

 

A total of 8,095,223 PIT address records were processed by the SDIV system as described above.  These records can be grouped into a number of categories that indicate their final disposition as a result of ORPS processing and school district review and correction.  Figure 2 illustrates these "address record categories."

 

Address records could be finally assigned by one of the four ORPS assignment methods (NYC, PIT/TAR, GIS, GIS/ZIP+4), or by superintendents of schools and certified by District Superintendents in the school district address review process (DS Corrections).  These corrections were either:  (1) originally assigned by one of the four assignment methods; (2), unassigned by the four methods and appeared on "Unassigned" lists; or (3), were previous year corrected records.

 

Since New York City represents such a large portion of New York State taxpayers, it is not surprising that the NYC method captured the highest percentage of PIT address records (38.8 percent).  The PIT/TAR method accounted for about 35.3 percent of address assignments, followed by 17.7 percent assignments under the GIS method.    In addition:

 

    1)    The percentage of address records that could not be verified diminished slightly (dropping from 5.7

            last year to 5.5 percent this year).

 

    2)    The DS correction rate, always small, decreased from last years rate of 0.2 percent, to 0.1 percent; and

 

    3)    The GIS method of assignment garnered 17.7 percent of assignments, up slightly from the 17.2 percent

            in the prior year.

 

 

Figure 3 and Table 3 (taken together) illustrate the performance of the SDIV process in yet another way.  Here, the number of District Superintendent corrections to an initial ORPS assignment, are shown for each of the four methods of assignment, as well as for previous year corrected, rural route, highway contract, and other unassigned address records.  It is assumed that a large number of District Superintendent corrections to records initially assigned by a given method, indicate that the given method was less accurate than other methods with fewer District Superintendent corrections.  Figure 3 shows the absolute number of District Superintendent corrections by initial method of assignment or category.  Table 3 shows, for each method or address category, the number of District Superintendent corrections as a percentage of the total number of records that could possibly have been corrected.

 

An inspection of both Figure 3 and Table 3 reveals that the New York City method was the most accurate of the four methods of assignment: i.e., this method had the smallest correction rate of .26.  That is, only .26 percent of addresses eligible for correction under this method were in fact corrected.  The PIT/TAR method was also proven to have been accurate as well, with only .46 percent of addresses corrected.  By way of contrast, 6.24 percent of the ZIP+4 method addresses processed were corrected by district superintendents.  Although a high rate relative to the aforementioned methods, the ZIP+4 method actually declined from last year’s 15.51 percent figure.   Finally, the highest correction rate, reflecting inaccuracy of computerized assignment, is associated with unassigned addresses.  This year’s rate of 29.1 percent is more than five times the rate of last year at 5.54 percent.  It is not yet clear what explains this increase, which is such an outlier in its change over the prior year.  The other methods’ correction rates generally changed little.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 3
District Superintendent Correction Analysis
Method or Address Category	# Eligible  For D.S. Correction (A)	D.S. Corrections         (B)	D.S. % Correction Rate (B/A)
From Additions and Deletions Lists
NYC	5,608	15	0.26
PIT/TAR	196,672	913	0.46
GIS	146,367	6008	4.10
GIS/ZIP+4	4,630	289	6.24
From Unassigned Lists
Rural Rts. & Hway Contracts	3,255	98	3.01
Other Unassigned	13,430	3,909	29.10
Sum	369,962	11,232	3.00

Continuing Quality Improvements:

 

1)  Resolve school district coding anomalies on tax maps, assessment rolls, and the SDIV program's    digital boundary file:  A number of inconsistencies with regard to individual parcels still exist throughout the State.  In numerous situations, a parcel listed on the roll of one school district is wholly surrounded by parcels from another school district. The improper portrayal of certain types of properties on the various files and maps also needs to be resolved.  These include Native American Reservations, the federal installation at West Point, and Special Act (Institutional) School District properties.

 

2)  Obtain better reference address data: The advent of 911 emergency management systems has contributed to many countywide efforts to associate an address with each tax parcel.  These efforts will allow for computation of the precise geographic location of each tax parcel address record in the county, thereby facilitating school district code assignment for most taxpayers in that county.  Also, ORPS has formed a Real Property System Team to concentrate on improving the current computerized tax administration system, including identification of correct geographic locations for each tax parcel, as well as establishment of parcel address standards.

 

3)  Over time, many of the Previous Year Corrected address code changes have resulted in appropriately corrected mailing labels by taxpayers.  As a result, many of these previous year corrected addresses have proven to be verified – that is, the taxpayer reported address and the ORPS assigned district code are now one and the same the majority of the time.  In the interests of improved efficiency and, reduced correction workload by school district staff – such verified address locations will no longer be sent for review.^[1]

 

 

 

 

 

---

 

 

 

APPENDICES

Appendix A