Indicator Development

The project explored a variety of data items intuitively related to property abandonment. Guided by data availability and the advice of our user community, five primary indicators comprise the composite summary used in our early warning system: property tax lien, housing code notice of violation, environmental ticket, structure fire, and sale or foreclosure. Indicator Math describes how the primary indicators below were summarized into a composite score.

Property Tax Lien (also Proxy for Utility Shutoff)

Tax sale advertisement may be the leading early warning indicator for fiscal abandonment in Providence for two reasons. First, the delinquency thresholds for inclusion ($1000, two quarters late) are low. Second, tax sale advertisement serves as a strong proxy for utility payment delinquency. At a meeting of Housing Network RI, several community partners concurred that tax sale advertisement would encompass most properties that were subject to water and sewer liens. While initially considered as factors in the abandonment equation, data related to forced shutoffs of electric and gas service are problematic because of the prevalence of oil as a home heating fuel, distinguishing whether an owner or renter is the responsible party, and the unresponsiveness of utility companies to our inquiries.

Providence now holds a tax sale once per year in late summer. Three weeks before the auction is held, the Tax Collector advertises affected properties in the local newspaper. In the 2004 sale, 1800 properties were advertised and 330 went to sale. Over half of the sold liens had been redeemed by the original property owner the following May, with three months remaining in the 365 day redemption protection period. After one year, lien redemption can become considerably more costly and difficult. The number of advertised tax liens and the number of sold tax liens in 2004 were less than half what has been witnessed in recent years. Of the 1500 properties identified as unutilized in our survey, 25% had been advertised for tax sale in the past five years and 13% had been sold at tax sale during that time span.

Housing Code Notice of Violation

Code enforcement activity is an essential indicator in an abandonment early warning system but availability of detailed, timely, accurate, and complete data is currently very limited in Providence. Housing code enforcement (here known as “Minimum Housing”) is not presently using the centralized Govern Land Management System. Instead they use an antiquated Access-based system to generate correspondence and keep track of most enforcement activity on paper within a property’s physical manila folder.

The most reliable indicator we could glean from this system is issuance of a “Notice of Violation” (NOV) which lists specific violations. The system can not easily indicate current violation abatement status so we are forced to simply use issuance of NOVs to indicate some problem with the property.

Environmental Ticket

The Department of Public Works issues tickets for violations related to improper storage of trash, dumping, and failure to abate rat infestation. Data are available for all violations but for this system it seemed appropriate to include tickets of $250 or more issued for severe refuse/dumping (greater than 1 cubic yard). Ticket data as recorded in the Municipal Court database include only address and not plat-lot necessitating conversion using a lot-to-address lookup table.

Because of recent database changes, only fifteen months of comparable data are currently available. Several more years of data need to accumulate before a statistical relationship to abandonment can be established.

Structure Fires

Structure fire data from the Providence Fire Department is a potentially valuable indicator. Some properties will be repaired quickly while others may be substantially destroyed and create a long-term liability for the community. A strong indicator would be if a property had several fires within a short period of time.

Sales and Foreclosures

While not necessarily a trouble indicator per se, sales and foreclosures are key data items in the system. Both can mean a new start, status quo, or the beginning of decline. In any case the date of sale relative to other key indicator dates is important and needs to be interpreted at the individual property level as well as within the automated early warning indicator count.

Several other data items were considered but for various reasons were not suitable as primary indicators of impending abandonment:

Municipal Liens (other than property tax liens)

Boarding and demolition liens issued by the Department of Inspections and Standards (DIS) are exceptionally rare in the present real estate market, as are environmental liens issued by the Department of Public Works (DPW). Both DIS and DPW have a backlog of unpaid liens; however, a review of these properties revealed these outstanding liens – in the absence of a formal collection mechanism and often overlooked at the time of deed transfer – are not reliably indicative of current trouble.

Non-Utilization Tax

Nonutilization tax is applied only to structures, at the exclusion of unused vacant land. Although used in Providence, this tax – a fee equal to 10% of the property tax and disqualification for tax exemptions – appears to be missing the local adoption ordinance for compliance with state enabling legislation. Perhaps for this reason, the nonutilization tax has been applied sparingly, with only 152 properties fined in 2003. Presence of the nonutilization tax is noted on the Detail Page of the interface but is not a factor in the incident summary.

Tax Class = Vacant

Common attributes of the property from the Assessor’s database can provide the earliest clues to abandonment. A “vacant” tax class (land use) designation is a first flag – keep in mind that “vacant” in this context refers to parcels without a structure. Our survey recorded 880 Vacant Residential parcels that appeared unutilized, meaning not in use as a sideyard or as orderly side parking. This is over one in every five such parcels citywide. Likewise, 10% of all Vacant Other parcels (admittedly, a catch-all classification) registered as unutilized in our driving survey. Other abandonment trends by land use are interesting but less than helpful. According to our survey, multifamily properties are more than twice as likely as single family properties to fall into disuse but neither type of property has more than 1% of its stock surveyed as unutilized. One item of concern was that over 50 city-owned properties appeared unutilized.

Proximity to Other Unutilized or Abandoned Properties

Spatial correlation – i.e. liened properties near liened properties, environmental tickets occurring in clusters, a rash of mortgagee sales in a neighborhood – is unmistakable. While there are anecdotes and some evidence that the actions, or inaction, of one land holder will affect the decisions of surrounding land holders, much of the spatial clustering could be attributed to economics: abandonment factors will be concentrated where low incomes are concentrated. We try to address this by summarizing incidents within census block groups (see Warning Sign Summary).

Location

Neighborhood can also be a useful filter, as seven neighborhoods had at least 100 unutilized parcels and two, West End and Lower South Providence, had over 200. Along with Lower South Providence, Olneyville and Upper South Providence were the three neighborhoods with at least 10% of all real estate deemed to be unutilized.

Assessed Value

Unutilized properties tend to be assessed below the citywide median for their land use classification. However, this tendency is not evident when compared to the same type of properties in the same neighborhood. At the time of project outset, the tax revaluation for year end 2003 was nearing completion. We supposed that properties with a drastic assessment hike might be prone to tax delinquency, but an adjustment to the homestead exemption more than offset the higher assessment in most cases.

Owner Characteristics

The current database captures two types of owner characteristics through exemptions: owner-occupancy and elderly ownership, both derived from property tax exemption status. Owner-occupancy is shown in the Detail Page of the search application and both characteristics are displayed in the mapping component.

Owner-occupancy is mapped at the property level but elderly ownership is aggregated to block group to preserve confidentiality. We discussed the possibility of identifying “problem owners” but did not pursue it for legal and practical reasons. One major barrier is the practice of purchasing property using different company names.

Warning Sign Summary or “Indicator Math”

A Warning Sign Summary is dynamically created for each property as well as for the surrounding block group, which we term the immediate vicinity. This provides context without having to launch the mapping interface and also tries to get at the indirect effect of warning signs on nearby properties.

One of the main pieces of feedback we received is not to provide a composite indicator score at the expense of detailed data. Partners believed that a composite indicator score was useful, but still wanted to be able to search on and view property-level detail for each indicator. Taking this into account, we devised a non-weighted indicator score based on simply adding up Tax Lien Advertisements & Sales, Housing Code Violation Notices, Public Works Environmental Tickets, and Structure Fires (Foreclosures are also included for Immediate Vicinity analysis).

We did not weight any of the factors differently, feeling that a sophisticated static weighting system is difficult to justify. Partners emphasized that the system had to be easy to understand and that the composite indicator was merely a way to narrow their research.

Sales and foreclosure data are used to optionally “reset” the indicator count, assuming that change of ownership constitutes a new start. Admittedly this may not always be the case, so users are given the option to tally indicators since last ownership change or regardless of change. Each warning sign also has a time threshold so that after a certain period of time incidents are no longer considered active. See table below:

* Only included in immediate vicinity analysis.