May 25, 2010 – Graziano Testimony

Appendix A

Portfolio estimates are critical to evaluating the proposed program

The utility of any portfolio-recapitalization proposal depends entirely on whether it works for the large inventory of properties, and that is ultimately a factual and quantitative exercise.  So, as HUD and the Congress consider TRA or some other form of property-based rental assistance (PBRA), we all need the best projections we can obtain as to the consequences of both a pilot and a universal program.  

Projecting TRA onto the public housing inventory, using sample properties

As part of this effort, to respond promptly yet quantitatively to HUD’s proposal, on short notice CLPHA convened a working group from among its members, and engaged a nationally recognized affordable housing expert to assist the working group and CLPHA in quantifying the impact.  We asked working group participants, who include several of the nation’s largest housing authorities, to identify properties they considered representative. 

For each selected property, the participants provided current operating data using a standard data collection instrument.  Participants are also providing their own estimates of their properties’ capital backlog, a concept that has to encompass more than a typical physical needs assessment and include the non-revenue components on public housing properties, such as community facilities and site infrastructure funded by the property, instead of being funded by the municipality as is the case for private affordable properties. 

The properties were self-selected, and the data was self-reported and is unaudited, so the results are not necessarily reflective of the entire portfolio.  Nevertheless, our survey sample encompassed roughly 19,000 apartments in fifteen housing authorities, and we asked them to pick typical properties.  We here report our findings in the interests of furthering the discussion.  

Basic assumptions in our analysis

The purpose of TRA is to standardize HUD programs and level incentives across those programs, while preserving public housing as a national resource.  In our projections, we have made the following assumptions that reflect those principles:

1. Rents at market, meaning 100% of FMR.  All properties are assumed to cancel their ACC's (which provide them with operating subsidy and modernization funds) and replace the ACC with a Section 8 contract at local, market, which we assume is 100% of FMR.  (We will also do sensitivity analysis using alternative rent assumptions.)
2. Properties retain their 'Other Income'¸ which is outside the ACC.
3. Assistance is portable, so that financial vacancy stabilizes at 5%.
4. No change in use, tenancy, or  income levels.  The properties will continue to operate as public housing, serving the poorest of the poor.
5. A one-time 10% increase in operating expenses, even if there is rehab, to account for marketing and competitiveness.  This is conservative but appropriate in light of the unknowns associated with a conversion.
6. All existing social programs continue.  Implied by keeping operating expenses unchanged.
7. New financing available on FHA-insured market terms, which are presumed to be 5.5%, 35 years, 117% debt service coverage.
8. Baseline capital backlog of $40,000 per apartment, which we think represents a decent starting point for national averages.  CLPHA is doing additional research to improve the accuracy of this estimate, which is obviously critical.
9. Annual new replacement reserve funding of $350 per apartment per year, a relatively low figure based on the presumption that the new financing will deal with the capital backlog, returning the property into sound and market-competitive condition prospectively.
10. Transaction costs of 3% of the new loan.
11. No continuing dividend limitations or restrictions on refinancing, so that post-TRA public housing authorities are placed in an equal position with their affordable and market competitors.

Estimated impact of TRA, as a pilot and as a permanent program

Assuming that the subset we have studied does in fact reflect the inventory as a whole, and using the baseline assumptions listed above, we project the consequences to HUD and to the inventory as follows.

New rents will be roughly $4,200 per apartment per year higher than current.  At 100% of FMR, the new rents will $350 monthly higher than the resources public housing now receives.  If we take this figure as reflective of the under-funding of public housing, and capitalize it at the assumed borrowing rate, it translates into $55,000 per apartment of value housing authorities have been deprived, which if multiplied across the entire 1,300,000 apartment inventory, represents $71.5 billion in financeable value – rehab plus equity housing authorities could use in furtherance of their mission.

The inventory divides into three groups: Viable, Sub-viable, and Social Assets.  Properties are Viable if, at market rents, they can generate new debt sufficient to cover at least the baseline capital backlog (projected at $40,000 per apartment).  Using that figure, and based on our portfolio sample, we find a portfolio distribution roughly as follows:

50-60% Viable.  These properties can support at least $40,000 per apartment of rehab.

30-40% Sub-viable.  These properties can support some rehab, but not enough.

  5-15% Social Assets.  These properties have negative Net Operating Income, and hence will need exception rents (see below).

Social asset properties will need rents above 100% of FMR, and project-based rents.  A property is a 'social asset' if it is both serving the cause of quality affordable housing, yet has negative Net Operating Income if rented at market.  These properties are not necessarily badly managed, and in fact most are well-managed; rather, they operate under handicaps (e.g. security services, social programs) the market competition does not.  Experience in HUD's mark-to-market program a decade ago revealed that these tend to cluster in two types:

• Rural high-rise elderly¸ where the competition is unprofessional walkups, and where the public housing property is built to a higher standard, including community facilities, and operated to enhance the elderly residents' quality of life.
• Urban family developments in difficult neighborhoods, where the property is maintained better, and provides better security, than its conventional competition.

Social-asset properties also tend to be concentrated in heartland America, where foreclosures and abandonment have weakened rents in the local submarket.

No capital subsidy can make a social asset property viable; only a budget-based exception rent, property-based in perpetuity, can assure their financial health.  These exception rents were an important feature in HUD's mark-to-market initiative and should be incorporated into TRA.

A $290 million pilot will fund 60-65,000 apartments.  HUD's initial proposal is for $350 million in funding, of which $50 million is for resident choice and $10 million is for technical assistance, leaving $290 million available for increased subsidy.  (We presume that this is intended to be an evergreen annual subsidy increase, since if it were a one-time payment it would be woefully inadequate to induce owners to participate.)  At a threshold of $40,000 per apartment, the pilot will fund 60-65,000 apartments nationwide.

Even this number of apartments participating may be optimistic.  Early-adopters in a voluntary pilot will be those properties that have the most potential to raise their rents, and to use the proceeds for substantial renovations. 

Based on an estimated conversion of 65,000 units with average rehabilitation of $80,000 per unit, the $290,000,000 initial TRA fund could lead to $5.2 billion of renovations a multiple of 18 times.