|
HUD
has never, to my knowledge, published any kind of statistical study on
REAC scoring trends or the accuracy and reliability of the scoring
results of the REAC inspection.
Raw scoring data has been published for many years now, in the
form of scoring data for each property that is inspected, but nobody to
my knowledge is processing that information into meaningful statistics
for public consumption. This
raw scoring data can be downloaded from two different web sites that
present the data in two different formats.
One of these presents a set of data for multi-family housing and
another set for public housing, each of which is huge but simple list of
inspection results, inspection by inspection.
Another is a large spreadsheet that combines into one record per
property of the last three consecutive REAC scores for each of over
20,000 multi-family (non-public-housing) properties. I
have found the latter of these datasets, the one presenting the last
three consecutive scores for each non-public-housing property, to be the
most useful for extracting information that is, in my personal opinion,
useful for assessing the success of the REAC inspection in achieving its
objective. Let me state here,
that the underlying objective of the REAC inspection is to identify
which properties are in need of HUD’s intervention due to poor physical
condition which indicates unsuitable habitation standards and/or
significant mismanagement or of the property. I
entered into my analysis of REAC scores with a set of assumptions or
possible biases relating to my own personal observations and persuasive
anecdotal evidence that I have been able to verify to my own
satisfaction by visiting many properties where the following
propositions seem to have been reality: |
||
|
Propositions: 1 -
REAC inspections do not consistently produce reliable, objective, useful
results 2 –
REAC inspections often mis-identify properties that are in significant
decline as being in acceptable or even such good condition that
oversight can be relaxed 3 –
REAC inspections often mis-identify properties in very good condition as
being in such poor condition that they require referral for remedial
action 4 –
REAC scoring can vary so drastically between two inspection dates that
the “swing” or “churn” of the scores calls the accuracy of the
inspection into question With
these propositions in mind, I designed a set of questions that I
endeavored to answer by a simple straight-forward analysis of the raw
data. I then reduced the
dataset to include only properties for which three inspections results
were available, and for which the most recent inspection was no earlier
than 2012. I then applied
spreadsheet functions to extract the answers to the questions below. |
||
| Note: These are not multiple choice questions - the answer shown are all true | ||
|
Question 1: What
percentage of properties are identified as being in such good condition
that HUD may relax their oversight by granting a one year or two year
exemption from the REAC inspection?
In other words: How many properties score 80 or higher, or 90 or
higher, respectively?
Answers: A -
95.5% of the 22,808 properties have scored 80 or above at some time in
their last three inspections, earning a “one-year skip.” B -
78.6% of the 22,808 properties have scored 90 or above at some time in
their last three inspections, earning a “two-year skip.” C –
76.6% of all inspection results for the 68,424 inspections (three
inspections of 22,808 properties) were scores of 80 or higher. D –
49.0% of all inspection results for the 68,424 inspections (three
inspections of 22,808 properties) were scores of 90 or higher. E -
51.6% of the 22,808 properties have always scored 80 or better. F –
19.4% of the 22,808 properties have always scored 90 or better. |
||
|
Question 2: What
percentage of properties fail the REAC inspection or score so badly that
REAC must refer the property to a HUD Field Office, Program HUB, or to
the Departmental Enforcement Center (DEC) for sanction, supervised
remedial action, or a penalty of some kind?
Answers: A –
10.8% of all 22,808 properties have failed an inspection with a score
below 60 at some time. B –
0.7% of all 22,808 properties (164) have failed an inspection with a
score below 30 at some time. C –
4.3% of all inspection results for the 68,424 inspections (three
inspections of 22,808 properties) were scores of less than 60. D –
0.3% of all inspection results for the 68,424 inspections (three
inspections of 22,808 properties) were scores of less than 30. E –
0.3% of the 22,808 properties have always scored less than 60. F –
0.09% of the 22,808 properties (2) have always scored less than 60. |
||
|
Question 3: What
percentage of properties experience a significant scoring variation over
the course of 2 or 3 inspections?
Answers: A –
25.6% of all 22,808 properties have experienced a scoring variation of
20 points or more from one inspection to the next. B –
10.0% of all 22,808 properties have experienced a scoring variation of
30 points or more from one inspection to the next. D –
1.2% of all 22,808 properties have experienced a scoring variation of 50
points or more from one inspection to the next. E –
8.2% of all 22,808 properties have had both a score below 60 and a score
above 79 at some time. F –
5.0% of all 22,808 properties have had both a score below 60 and a score
above 89 at some time. G –
0.42% of all 22,808 properties have had both a score below 30 and a
score above 79 at some time. H –
0.22% of all 22,808 properties have had both a score below 30 and a
score above 89 at some time. |
||
|
Question 4: What
are the average and median scores, and the frequencies of significant
ranges of scores? The
average REAC score across 68,424 inspections of 22,808 properties was
85.7. The
median REAC score across 68,424 inspections of 22,808 properties was 89. The
mode (most frequent score) was 99. |
||