From: Subject: WHAT SMALL ASSESSING OFFICES NEED FOR CAMA Date: Tue, 14 Oct 2003 13:54:53 -0400 MIME-Version: 1.0 Content-Type: text/html; charset="Windows-1252" Content-Transfer-Encoding: quoted-printable Content-Location: http://www.microsolvecama.com/BASICCAMA.HTM X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 WHAT SMALL ASSESSING OFFICES NEED FOR CAMA

CAMA : WHAT SMALL ASSESSING OFFICES NEED=20



Jan Schreiber


President, MicroSolve Corp.

A computer assisted mass appraisal (CAMA) system = means=20 different things to different people. Some stress the system=92s ability = to value=20 properties using a cost approach. Others emphasize data management tools = like=20 import/export utilities or routines for printing standard reports. Still = others=20 look for analytic tools to ensure equity, such as sales ratio statistics = or=20 graphic presentation formats. The best CAMA systems combine all of these = features and in addition are designed to accommodate a wide variety of = database=20 designs, valuation methods, and report types.


In this article I survey the principal features = that any good=20 CAMA system should have, and I give some suggestions for making them = efficient=20 and capable of achieving appraisal equity.


Database

A database is a repository for all the information = in the=20 assessor=92s office about parcels. It is organized by ownership units, = meaning=20 that each record holds information describing the contiguous land under = one=20 ownership at any one location, together with all improvements on that = land.=20 Nowadays databases are usually relational. A relational database = is=20 really a set of tables, each pertaining to a major feature of property = (for=20 example, land, outbuildings, dwelling, and all the appurtenances of a = dwelling,=20 such as porches, attached garages, heat, floor, etc.) These tables are = linked=20 together by an identifier called the parcel ID. This type of structure = makes it=20 possible to have as many types of land, as many outbuildings, or as many = porches=20 as needed, because records can always be added to the land, outbuilding, = or=20 porch tables. The database is not limited by an arbitrary number of = fields=20 established when it was created.


Databases of this sort are equipped with automatic = indexes.=20 Once an index is made, it is updated whenever data is changed; the user = does not=20 have to update it explicitly.


The database allows any user to add and delete = records as=20 necessary. It also includes a "filter" capability, which means the user = can set=20 selection rules that will display or process only records meeting = certain=20 criteria. For example, a user can set a filter so that the database will = display=20 or process only records of properties that were built after 1965 or = properties=20 that were sold during 2001. More complex filter settings might stipulate = that=20 all properties should consist of ranch-style houses, of good quality or = better,=20 in a certain neighborhood. A good database program will allow users to = display=20 complete records, or to show data in tabular form, where each row is a = different=20 record and each column is a feature (or factor) of that record.


More sophisticated databases will also include = sketches of=20 the building footprint, which can be drawn in the process of data entry. = They=20 will allow users to attach digitized photos to a record, so that one or = more=20 views of a property can be displayed along with the numeric = data.


Ideally, users will be able to modify factors in = the database=20 as needed. If they need a new field, they can add it. If they need to = modify an=20 existing one (for example, to accept multiple decimal places), they can = do that=20 as well.


Finally, the database should provide a means of = exporting=20 data in a format that can be used elsewhere: perhaps in an Excel = spreadsheet, or=20 perhaps by another database, which may need to import a file in ASCII=20 format.


Cost Approach

Even though everyone recognizes that the market is = the best=20 determiner of value, most assessors rely on the cost approach as their = primary=20 valuation tool. The reason is that market analysis requires a minimum = number of=20 valid sales per year of each major property type =96 something not = always easy to=20 achieve in small jurisdictions =96 whereas the cost approach simply = associates a=20 dollar figure, stored in a table, with each building component, and = thereby=20 compiles a total property cost.


The major supplier of property cost figures in the = United=20 States is Marshall & Swift, but there are other sources as well, = including=20 some states, which have developed their own costing algorithms and = associated=20 rates. A functional CAMA system should include the Marshall & Swift = cost=20 approach, and a versatile one should make it possible to use other types = of cost=20 approach at the user=92s option. In either case it is important that = users be able=20 to update cost rates when they need to, or change local multipliers,=20 depreciation rates, and the like. In the best systems, any aspect of the = cost=20 approach, including the algorithms used and the printed report layout, = can be=20 modified to suit user preferences.


Comparable Properties

Also known as "comparables sales," "sales = comparison," or=20 simply "comparables," this is probably the oldest appraisal tool around. = Using=20 it, an appraiser can specify several factors that are taken to be the = most=20 essential features of a property (land area, building = area,=20 building quality, number of bedrooms, and year = built might=20 be good examples). The user assigns a weight to each characteristic, = indicating=20 its importance in determining the property=92s value. Starting with a = subject=20 property chosen by the user, the computer scans through all the = properties in=20 the database, comparing each of them with the subject. Those in which = the=20 weighted differences are smallest are considered most similar to the = subject.=20 The five "most comparable" properties are listed in a report.


If each of these properties has recently sold (the = program=20 can be instructed to choose only sold parcels), the sale price is = adjusted by a=20 value representing any difference between this property and the subject. = For=20 example, if the subject parcel had three bedrooms and one of the = comparable=20 parcels had two bedrooms and sold for $150,000, the program would adjust = the=20 selling price to reflect what it would be if the property had an = additional=20 bedroom. The appraiser who set up the program might have instructed it = to treat=20 the value of a bedroom as, say, $10,000. So the adjusted sale price = would now be=20 $160,000. In this way, each of the comparable parcels is given a value = derived=20 from its sale price, and their average can be taken as a good = approximation of=20 the value of the subject parcel in the current market.


The comparable properties routine is usually seen = as a=20 defensive tool for the assessor =96 that is, one that he can bring out = if a=20 taxpayer comes in to protest his appraisal, saying, "My neighbor=92s = house was=20 appraised a lot lower. Why are you picking on me?" This tool allows the = assessor=20 to show what the truly comparable properties are in a given = neighborhood, and=20 what their value in the market appears to be. It is a cumbersome tool to = use for=20 mass appraisal, however. Modern statistical procedures like multiple = regression=20 work much more efficiently.


Sales Ratio Studies

The aim of the assessor=92s office is to provide an = equitable=20 basis for levying property taxes. "Equitable" means that parcels of = similar=20 size, quality, and construction in a given location are valued = similarly, much=20 as the real estate market values them through frequent arm=92s-length=20 transactions. For that reason, the market is considered a standard in=20 determining equity. The standard mathematical measure of equity is the = sales=20 ratio, or the ratio between the appraised value of a property and the = price at=20 which it sold. Those two numbers should be very close, and on the = average the=20 ratio should be close to 1. A sales ratio study is simply the = calculation of=20 this ratio for a series of sold properties, together with the = examination of=20 some statistical measures indicating how much deviation from the ideal = ratio=20 there is in the population studied.


Here are the principal measures:


A good CAMA system should be able to provide these = measures=20 readily, and for any subgroup of parcels the user chooses to define. In=20 addition, the program should be capable of displaying graphic indicators = such as=20 histograms (bar graphs) and scatter diagrams that show salient features = of the=20 data in a way that can be quickly grasped visually.


Beyond the Basics

Good CAMA systems do much more than I have = described. They=20 should be able to run multiple regression or other statistical processes = and=20 apply the derived coefficients to a range of parcels in the database. = They=20 should permit the user to develop various income approaches, including = direct=20 capitalization and discounted cash flow, for the valuation of commercial = properties. They should allow the creation and use of separate sales = files, or=20 of files representing subgroups of parcels that users may want to take = into the=20 field on a laptop for on-site updating. Further, they should permit = integration=20 with geographic information systems and with administrative and = accounting=20 software. CAMA vendors can often facilitate such integration.


It is important to be sure any software you = purchase will=20 operate efficiently, in a user-friendly, graphical user interface. At = least as=20 important is the support you get from the software provider. You should = expect=20 same-day service whenever you have a question, and quick turnaround when = you=20 need modifications in your program. Software is a great timesaver, but = only when=20 it doesn=92t bog you down in delays that smart, responsive human beings = can=20 prevent.