I think the notion of "as is" value of site improvements is silly and baseless. If every residential property has a driveway, sidewalk, and water and sewer source, there are not any sound ways to gauge the "as is" value, even when new. Even if one has all those components but the sidewalk, can you really parse the value of a missing sidewalk out of the mix? I have spent hours cleaning the data to assess whether or not there is a measurable difference between a property with a well and one with a cistern, to no avail. Same with septic vs public sewer. Rarely, a difference might seem apparent, but it is just as likely to be due to another difference (I have one model here where the difference between public and private utilities can be shown, but the public utilities are on the valley floor and the private utilities are in the steep, adjoining hills...where people fight the roads every winter hauling water home every night after work or pay a premium to those whose business is to delver water, who fight the same battle for a higher price...is the difference in value due to location, or due to the difference in utilities?). I expect the results would be almost identical if one went to the trouble of compiling data that included the sizes and ages of every component. But, around here, a well could be 30 feet deep and cost $2,500 or 1,200 feet deep and cost in excess of $100,000, and a difference in value can't be measured reliably. The entire approach is diminished by forcing the inclusion of a baseless guess.