1. Clyde Barr Policy Analyst cbarr@mainehousing.org 207.624.5772
Continuum of Care Program: Solicitation of Comment on Continuum of Care Formula Docket No. FR5476N044
Clyde Barr
Policy Analyst

2. Current PPRN Calculation
Combination of
The formula used to award ESG program grant funds and
CDBG funds
Current CDBG “dual formula” system:
Larger of the two calculations assigned, less a pro rata reduction*
*ensures the total amount allocated is within the amount appropriated for funding
Emergency Solutions Grants (ESG)
Community Development Block Grants (CDBG)
Under current PPRN formula, after a .2 set aside for US Territories and insular areas
75% of total CoC allocation is distributed to ESG entitlement communities (generally large metro cities and urban concentrated)
Remaining 25% of CoC allocation is distributed to ESG non-entitlement communities
Portland only ESG entitled community in Maine (ESG Entitled Principle City)
Auburn: CDBG-MC (ESG Non-Entitled CDBG Principal or Metropolitan City)
Bangor: CDBG-MC (ESG Non-Entitled CDBG Principal or Metropolitan City)
Biddeford: CDBG-MC (ESG Non-Entitled CDBG Principal or Metropolitan City)
Lewiston: CDBG-MC (ESG Non-Entitled CDBG Principal or Metropolitan City)
Cumberland County: CDBG-UC (ESG Non-Entitled CDBG Urban County)

3. Previous Comments
Not representative of the number of individuals & families experiencing homelessness in their geographic area
CDBG formula was not appropriate for PPRN
Disliked the reliance on “urban blight” as reflected in the age of the housing stock
predominantly from western States, counties, and cities
Opposed to reductions for renewal projects

4. Developing New PPRN Formulas
HUD sought to maintain the basic structure
Data sources need to be:
Relevant
Accurate
Timely
Readily available
Chose not to incorporate PIT directly
Used average of two years
Relevant – to measuring homelessness
Accurate – correct in all details
Timely – of the moment
Readily available – attainable for all jurisdictions
WHY NO PIT
Not all CoCs use the same methodology
Not all CoCs conduct annual PIT count
Direct inclusion of PIT counts into an allocation formula may create perverse incentives against objective PIT count methodologies
HUD used an average of two years of PIT count data
Helped quantify the relevance of potential formula factors measuring homelessness
Insulated potential formulas from limitations of directly including PIT counts
Using factors correlated with PIT count, proposed formulas mitigate the risk of data fluctuations in PIT counts that may be less prevalent

5. Overview of Proposed Formulas
Goal: to approximate the actual homeless need in communities
Diminished reliance on annual PIT
Pre 1940s housing & Overcrowding not included

6. Overview of Proposed Formulas
Formula A
Formula B
Formula C
Formula D
10% * population
15% * poverty
25% * affordability gap
25% * rent-burdened ELI households
25% * rental units
25% * poverty
25% * population
50% * hybrid factor
Overcrowding and pre 1940s housing not included in new formulas
Affordability gap: measures the gap between the demand for and supply of rental units that are both available and affordable to ELI renter households
Rent burdened ELI households: ELI households that pay more than 30% of income for housing.
Rental units: renter occupied units
Hybrid factor: combination of Rent burdened ELI households & Rental units –
calculated by multiplying the number of rent-burdened ELI households by the ratio of: the jurisdiction’s percentage of renter occupied units divided by the national percentage of renter occupied units
All data from ACS & Comprehensive Affordability Strategy (CHAS)

7. 34% of CoCs funding could decrease between $307 - $25,955,807
Alt Formula “A”
COCNUM
COCNAME
PPRN15
Baseline Allocation
Alt Formula
$ Diff
PCT Diff
ME-500
Maine Balance of State CoC
$3,712,281
$4,157,086
$4,288,146
$575,865
15.51%
ME-502
Portland CoC
$1,222,088
$1,884,953
$624,340
-$597,748
-48.91%
ME-5**
One Maine CoC
$4,934,369
$6,042,039
$4,778,650
-$155,719
-3.16%
FORMULA WEIGHTS PEARSON’S CORRELATION
10% * population No significant correlation
15% * poverty
25% * affordability gap
25% * rent-burdened ELI Households .336
25% * rental units
Pearson’s correlation
a measure of the linear correlation between two variables X (population etc. . .) and Y (average 2 year PIT count), giving a value between +1 and −1 inclusive, where 1 is total positive correlation, 0 is no correlation, and −1 is a negative correlation
Calculated by HUD
Not based on 75/25 ESG Entitled/non entitled split – HUD is interested in removing this split
Baseline allocation calculated as: ((Total CoC PIT Count/National PIT Count)*Total Funding)
34% of CoCs funding could decrease between $307 - $25,955,807

8. 35% of CoCs funding could decrease between $646 - $25,410,952
Alt Formula “B”
COCNUM
COCNAME
PPRN15
Baseline Allocation
Alt Formula
$ Diff
PCT Diff
ME-500
Maine Balance of State CoC
$3,712,281
$4,157,086
$4,205,686
$493,405
13.29%
ME-502
Portland CoC
$1,222,088
$1,884,953
$631,500
-$590,588
-48.33%
ME-5**
One Maine CoC
$4,934,369
$6,042,039
$4,703,350
-$231,019
-4.68%
FORMULA WEIGHTS PEARSON’S CORRELATION
25% * poverty
25% * affordability gap
25% * rent-burdened ELI Households .336
25% * rental units
35% of CoCs funding could decrease between $646 - $25,410,952

9. 37% of CoCs funding could decrease between $406 - $25,701,362
Alt Formula “C”
COCNUM
COCNAME
PPRN15
Baseline Allocation
Alt Formula
$ Diff
PCT Diff
ME-500
Maine Balance of State CoC
$3,712,281
$4,157,086
$4,254,856
$542,575
14.62%
ME-502
Portland CoC
$1,222,088
$1,884,953
$612,211
-$609,877
-49.90%
ME-5**
One Maine CoC
$4,934,369
$6,042,039
$4,536,819
-$397,550
-8.06%
FORMULA WEIGHTS PEARSON’S CORRELATION
25% * population No significant correlation
25% * poverty
50% * hybrid
Hybrid is a combination factor of rental units & rent-burdened ELI Households
37% of CoCs funding could decrease between $406 - $25,701,362

10. 37% of CoCs funding could decrease between $8,279 - $22,436,373
Alt Formula “D”
COCNUM
COCNAME
PPRN15
Baseline Allocation
Alt Formula
$ Diff
PCT Diff
ME-500
Maine Balance of State CoC
$3,712,281
$4,157,086
$3,795,378
$83,097
2.24%
ME-502
Portland CoC
$1,222,088
$1,884,953
$701,758
-$520,330
-42.58%
ME-5**
One Maine CoC
$4,934,369
$6,042,039
$4,166,388
-$767,981
-15.56%
FORMULA WEIGHTS PEARSON’S CORRELATION
25% * poverty
25% * affordability gap
50% * hybrid factor
37% of CoCs funding could decrease between $8,279 - $22,436,373

11. A Deeper Look. . . 100 Largest PIT 100 Smallest PIT
38% - 44% of CoCs see decreased funding
Mean PIT: 4,075
Median PIT: 2,165
100 Smallest PIT
23% - 26% of CoCs see decreased funding
Mean PIT: 189
Median PIT: 185
Top 100 PIT Max Top 100 PIT Min
NY-600 New York City CoC 75,323 SC-503 Myrtle Beach, Sumter City & County CoC 1,319
Formula A -18% Formula A 20%
Formula B -17% Formula B 27%
Formula C -11% Formula C 24%
Formula D -0.3% Formula D 14%
Bottom 100 PIT Max Bottom 100 PIT Min
IL-501 Rockford/Winnebago, Boone Counties CoC 327 MD-510 Garrett County CoC 9
Formula A -14% Formula A 70%
Formula B -13% Formula B 69%
Formula C -16% Formula C 51%
Formula D -14% Formula D 75%

12. Big Picture of “Alt” Formulas
% Negative Funding by Census Regions
Region
Alt “A”
Alt “B”
Alt “C”
Alt “D”
National
34%
35%
37%
New England
54%
50%
Mid-Atlantic
65%
70%
67%
74%
Midwest ENC
42%
44%
45%
Midwest WNC
39%
South Atlantic
22%
25%
28%
East South
13%
West South
15%
18%
Mountain
11%
17%
Pacific
24%
29%
Region 1: Northeast
Division 1: New England – CT, ME, MA, NH, RI, & VT
Division 2: Mid Atlantic – NJ, NY, & PA
Region 2: Midwest
Division 3: East North Central – IL, IN, MI, OH, & WI
Division 4: West North Central – IA, KS, MN, MO, NE, ND, SD
Region 3: South
Division 5: South Atlantic – DE, FL, GA, MD, NC, SC VA, DC, & WV
Division 6: East South Central – AL, KY, MS, TN
Division 7: West South Central – AR, LA, OK, TX
Region 4: West
Division 8: Mountain – AZ, CO, ID, MT, NV, NM, UT, & WY
Division 9: Pacific – AK, CA, HI, OR, & WA
*Puerto Rico & other US territories are not part of any census region or census division

13. 37% of CoCs funding could decrease between $8,531 - $18,982,831
Custom Calculation
COCNUM
COCNAME
PPRN15
Baseline Allocation
Alt Formula
$ Diff
PCT Diff
ME-500
Maine Balance of State CoC
$3,712,281
$4,157,086
$5,194,189
$1,481,908
39.92%
ME-502
Portland CoC
$1,222,088
$1,884,953
$841,561
-$380,527
-31.14%
ME-5**
One Maine CoC
$4,934,369
$6,042,039
$5,705,002
$770,633
15.62%
FORMULA WEIGHTS PEARSON’S CORRELATION
15% * poverty
15% * pre-1940 housing
20% * affordability gap
50% * hybrid
Inclusion of pre-1940s housing
Overcrowding (.277) has similar result – might be multicollinearity
Multicollinearity
is a phenomenon in which two or more predictor variables in a multiple regression model are highly correlated, meaning that one can be linearly predicted from the others with a substantial degree of accuracy.
Problem with pre-1940s housing
Decreases over time – especially if it is a static measurement
A dynamic measurement would be “housing stock 75+ years old”
37% of CoCs funding could decrease between $8,531 - $18,982,831

14. % Difference Maine Alt Formulas
COCNUM
PPRN15
Alt A
Alt B
Alt C
Alt D
Custom
ME-500
$3,712,281
15.51%
13.29%
14.62%
2.24%
39.92%
ME-502
$1,222,088
-48.91%
-48.33%
-49.90%
-42.58%
-31.14%
ME-5**
$4,934,369
-3.16%
-4.68%
-8.06%
-15.56%
15.62%
Including pre-1940s housing has a positive impact in the northeast and a negative impact in the west and south Atlantic
Have not found any formula that PCoC has the potential for gaining funding over previous year

15. “Alt” Formulas vs. Custom
% Negative Funding by Census Regions
Region
Alt “A”
Alt “B”
Alt “C”
Alt “D”
Custom
National
34%
35%
37%
New England
54%
50%
39%
Mid-Atlantic
65%
70%
67%
74%
Midwest ENC
42%
44%
45%
43%
Midwest WNC
South Atlantic
22%
25%
28%
40%
East South
13%
West South
15%
18%
24%
Mountain
11%
17%
Pacific
29%
Nationally % of CoCs losing funding is consistant

16. HUD wants to know our thoughts:
The four proposed formula options
Factors and corresponding weights
Dual or multi-formula system
On language that would prevent a CoC from losing more than a certain portion of their PPRN
Other comments on CoC formulas
Including should they keep 75/25 split or adjust it

17. The four proposed formula options
Thoughts on. . .
The four proposed formula options
More than 1/3 of CoCs could see funding decrease with all proposed funding formulas
New England & Mid-Atlantic states are disproportionately affected

18. Factors and corresponding weights
Thoughts on. . .
Factors and corresponding weights
Including pre-1940s housing increases funding in the
Models with the hybrid factor have the most CoCs that could lose funding
There is not a one size fit all

19. DUAL OR MULTI-FORMULA SYSTEM
Thoughts on. . .
DUAL OR MULTI-FORMULA SYSTEM
Support a dual or multi-formula system
Regional differences should be accounted in funding system
Pre-1940s housing
Variable’s impact will decrease with time

20. Thoughts on. . .
language that would prevent a CoC from losing more than a certain portion of their PPRN
YES – HUD should adopt language that would prevent a CoC from losing more than X% of its funding

21. Thoughts on. . . Other comments on CoC formulas
ESG 75/25 split should be discontinued
Other comments on CoC formulas
ESG 75/25 split should be discontinued
75/25 ESG Entitled/Non-entitled Split
Alt A
Alt B
Alt C
Alt D
MCoC
-35%
-36%
-38%
PCoC
-32%
-28%
75/25 ESG Entitled/Non-entitled Split
Alt A
Alt B
Alt C
Alt D
MCoC
-35%
-36%
-38%
PCoC
-32%
-28%
PCoC has best funding within these formulas
However 45% - 51% of CoCs nationally see funding decline
HUD has indicated interest in removing this split
Open to other ways to target densely populated urban areas
NYC/LA/Houston see funding increase
Chicago/DC/Philly see finding decrease

22. Other comments on CoC formulas
Thoughts on. . .
Other comments on CoC formulas
Declining importance of annual PIT a good idea
Hypothesis, including more robust count data from new HUD Performance Metric 3 (number of homeless persons) could:
Smooth out funding discrepancies
Provide a truer picture of need

23. Questions?