vashon assessor’s database presentation

							Analysis of King County Assessor’s Database

Bob Powell, Feb. 8, 2007
bobp@dogpatch.com

Using the Assessor’s database and information from the island’s public water systems, I
analyzed metrics of Vashon's 7979 tax parcels by present and potential usage, created a
picture of Vashon’s present state of development, and predicted Vashon’s ultimate “build
out” under “unlimited water” and “today’s water” scenarios.

Again, this is predicting Vashon’s ultimate build-out – when every square inch of land
usable under present growth management, zoning, health and other requirements has
been built to its limit. This predicts the ultimate target – not a growth rate.

I consider zoning type and density for each parcel, hypothetical water and septic/sewer
availability, present usage and other factors.

My intent is to make a reasonable estimate – this requires many assumptions, some easy
to defend, some unscientific. I present the assumptions for debate, and at least, am
confident that the calculations accurately reflect the assumptions. Different assumptions
about water availability and what constitutes a buildable parcel can be plugged in to
observe the result.

The main report is in two parts: the statistics section, which is all hard data from the
Assessor’s office, and the predictions, which are computed from hard data but using
necessarily vague assumptions. The assumptions are questionable, but the computations
based on them are tangible.

This is a work in progress. Many parcel aspects are still treated in the abstract – the
present model doesn’t look at terrain, soil conditions, match up parcels with water district
boundaries, or attempt to discern the proximity or suitability of private water systems.

My original intent was to create a fully automated tool to take a fresh download of the
Assessor’s data, feed it through a program, and pop out these two reports from live data –
updated as often as you like. That would be a great way to refine the predictions – load
new data and see how new development is apportioned across the various parcel sizes,
types and conditions.

Alas, the raw data is so awkward and inconsistent that automating this is far more work
than it could possibly be worth. Clearly, the Assessor’s data serves a key role in the
King County economy – but my density calculations use data elements that apparently
aren’t used in any KC reports or calculations, so the Assessor has no reason to maintain
those elements in a suitably clean form.

The prediction report also uses data from Vashon’s public water purveyors. In Jan. and
Feb. 2007 I spoke with the managers of all the water systems mentioned in the report – as



Property Analysis Summary                                        printed 11/18/2009 page 1
well as Bill Lasby at the KC DOH (rural water systems), and the Vashon Sewer District
manager.

I also look at residential building permits in the DDES database, some of which are
referenced in this document, but not the statistics or predictions reports.

General observations:

As a rule, growth in Vashon Town is limited by water, and growth outside Town is
limited by zoning and the GMA. Actually not limited by the GMA, because small
parcels are grandfathered and need only meet techical and health requirements.

This is exactly backwards from the explicit intent of the KC plan, which is to encourage
small-footprint housing co-located with services, and discourage new development in the
more rural areas.

I’d like for this study to encourage the Vashon community to question the present
situation, and perhaps work towards a situation that enables more ecologically sound
development. Today, it’s … not economic to build a small-footprint home –
development cost for one unit on an RA parcel is too high to justify it.

DDES permits – “replacement” structures:

I looked at the 85 most recently opened active permits for new single family residential
structures – this covers mid 2005 through the present. There’s no particular significance
to the start date, I just stopped at 85.

Of the 85, only 17 are for structures in the “R” zones, i.e. Vashon Town. Of the 17, 16
are for the Vashon Household Roseballen project. (The project is for 19 homes but I see
only 16 active permits.)

The one project that isn’t Roseballen is for a new dwelling on a parcel that already has a
structure, that will be kept as an ADU.

Of the other 68 active permits – all 68 are in the RA zones, i.e. outside Town.

Of the 68, 14 are for “replacement” dwellings (replacing mobile homes, beach cabins or
older homes) and 54 are for new dwellings on previously vacant land.

Growth Management Act:

The GMA and revised KC plan put the emergency brakes on rural subdivision – in
general, only parcels over 10ac (RA5 zoning) or 20ac (RA10) can be subdivided at all,
and any substantial subdivisions require clustered development.




Property Analysis Summary                                       printed 11/18/2009 page 2
(Most small RA parcels on Vashon are zoned RA2.5, but RA2.5 no longer means 2.5 ac
… RA2.5 is now explicitly treated as RA5 or 5ac/DU for subdivision and other legal
purposes.)

That said, the plan grandfathers all existing substandard / non-conforming parcels. A
single dwelling unit can be built on any existing parcel in the RA zones, irrespective of
size, so long as the parcel meets basic technical and health requirements, i.e. water
availability, and either sewer availability or septic feasibility. On Vashon, this represents
over 80% of existing vacant residential parcels.

The Vashon Sewer District does not expect usage to exceed its capacity. The KC plan
says to use sewers in the “R” zones and not to use sewers in the “RA” zones. I make
assumptions that sewer connections are available to all parcels in the “R” zones, and not
in the RA zone. For an “R” parcel to be buildable it needs only water – there is no size
minimum. In general, multiple units can be built, up to the limit indicated by the lot size
divided by the zoned density, e.g. 2 units on a half-acre lot zoned R4.

For an RA parcel, the main question is whether it can be built or not – not how many
units; it’s either one or none. (About 175 of Vashon’s ~8000 parcels are eligible for
subdivision.)

For an RA parcel to be buildable, I use the figure of 1/3 ac as the average minimum to
accommodate a septic system and reserve drainfield area, using an off-site water source
(public water system or shared well). I use 2.5 ac as the minimum for a septic system
and new private well.

Many parcels can get water from a neighboring private well (existing or creating a new
“Group B” private water system), or groups of vacant parcels can combine to share a well
within coordinated easements and covenants. But small parcels tend to border small
parcels, large near large, and typical parcels too small to accommodate a well are most
likely to border parcels that either can’t accommodate a new well, or have an existing
well that doesn’t meet current health standards and thus can’t be used as a new source.

Many vacant parcels are near the shoreline or other terrain, where sensitive areas
preclude any new use of the parcel at all, even with water available. Overall, I assume
that half of “small” (1/3 to 2.5ac) are usable, 2/3 of parcels 2.5 to 10ac are usable, and all
parcels over 10ac are usable.

The text in the “Prediction” report goes into more detail.

Data sources:

In late Jan / early Feb 2007, I spoke with the managers of WD19, Heights Water,
Dockton/Maury Mutual/Westside et. al. (Doug Dolstad) and Burton Water Co., and
obtained the number of existing paid unused connections, and number of additional new
connections available for purchase, to the extent of the purveyor’s ability to provide it.



Property Analysis Summary                                          printed 11/18/2009 page 3
In the case of WD19, I matched up the list of “zero use” accounts either with individual
parcels or at least with the parcel zoning (in Vashon Town in the “R” zone, or outside in
the “RA” zone). WD19 does not track whether a customer has a structure attached to a
connection – a customer either has an account or not. Customers who own a paid water
share for a vacant lot have an account like anyone else, they just get billed for 0 cf/month
usage.

In the case of the WD19 waiting list, I matched up the entries near the top of the list with
specific individual parcels, to the extent possible. See the “Prediction” report for more
detail.

Implications for other growth projections:

Accurate growth prediction affects many facets of Vashon economic life: ferry capacity,
school district capital budgeting, emergency services, sewer, road and other infrastructure
improvements, et. al.

The Washington State Ferry System uses a higher growth rate for Vashon’s population
and Vashon to Seattle commuting demand, and implicitly a higher population target than
Vashon will reach. The figures used by WSF should be questioned, and more realistic
figures used for budgeting capital needs and expenditures.

The 2000 Census reports 4867 “housing units” and 4193 “households”. (What’s a
“household”?) The 2005 Assessor data shows 5147 “dwelling units” – 280 more in 3
years, but both are suspect. But are there really nearly 1000 unoccupied units? How
many Vashon units are second homes? (How many unofficial units don’t appear in the
Assessor’s data or the census? What is Vashon’s actual population?)

Notes on the process – Assessor’s database:

This effort involved downloading several separate databases from the Assessor’s web
site, totalling about 2 gigabytes uncompressed, having hundreds of thousands of
individual records spread across many tables. There is a great deal of noise in the data,
making it extremely challenging even to identify just the Vashon parcels, and compute
an aggregate total of the land area.

Based on all the inconsistencies in the data, it’s unlikely that anyone else has ever run
this kind of lay-of-the-land analysis using the Assessor’s raw data. If they had -- the
quality of the raw data would of necessity reflect a higher standard, to enable such
computation.

The database is structured as separate tables and records for parcels, structures, tax
accounts, valuations for each calendar year, condomium complexes, individual
condominum units and many additional detail records. I extracted the Vashon data and
then “flattened” the data to create a single aggregate record with one record per parcel



Property Analysis Summary                                        printed 11/18/2009 page 4
that includes the most recent tax valuation, taxpayer, number of dwelling units, zoning,
usage, lot size and other figures. This is maintained as a single Excel spreadsheet. I then
separated the parcel records into ten sub-sheets to facilitate easier computation based on
different groups of zoning and usage.

Some problems preventing easy use of the Assessor’s raw data:

I don’t question the accuracy of the Assessor’s data for revenue purposes – although
anecdotally there have been problems in the past, I found no evidence to suggest present
problems. The database exists for revenue purposes, not geograpphic purposes

1. Many large parcels appear in the database coded as buildable land, but in reality are
tidelands with no indication that they shouldn’t be treated as usable land. The clue is an
implausibly low tax valuation. I have manually checked and removed these records.

2. The standard Assessor’s report for Vashon is “Area 100” plus “Area 47”, but this fails
to retrieve any of Vashon’s rcw-titled Forest Land parcels, which are coded differently
in the database. (Several dozen large parcels treated as residential for zoning purposes
don’t appear as part of Vashon.) I found and added these records.

3. Forest Land parcels that do appear in the report have no SqFtLot value. I computed
approximate sizes for these parcels using the parcel boundaries.

4. Many parcels have “999999” or other special codes stored in the “SqFtLot” field
which is the sole database representation for parcel size. This suggests that no “official”
calculations using the Assessor’s data consider parcel size. (Size calculations blindly
using this data would compute incorrect totals by treating many properties as 999999 sq-
ft in size.) I corrected the values to be meaningful for the calculations.

5. The database is always in a state of flux, as routine work is performed, and parcels are
subdivided or combined on a daily basis. Many parcel records have no corresponding
records in other tables. Some parcels have tax assessment records for the present year
and prior years, others only for prior years.




Property Analysis Summary                                        printed 11/18/2009 page 5