Please note: if you don't agree with my methodology, that's fine with me. All of the following proves exactly nothing. I'm stating what I think the numbers mean, and how they could be used - what you do with them is entirely up to you. This research is to establish pricing guidelines, not definitive “if you pay more than $x you're an idiot” price fiats (if you'll pardon the pun).
What I have to work with are asking prices, NOT actual sale prices. I include a very brief synopsis of the ad for comparison purposes. Mileage in miles follows price where available. Currency conversions courtesy of Xenon Lab's Universal Currency Converter. All figures are as measured in the USA. Cars without prices but for which mileage was included are used for mileage calculations only. Close cover before striking.
1152 cars aren't a statistically valid sample. What little I
remember of conducting statistically valid surveys suggests that after 1,700 samples a trend is
definitively established, so I've got a ways to go yet. And it should be obvious to everyone
that individual cars and circumstances will dictate much higher or lower prices - this is
intended to be baseline information, not the official selling price for all Fiat Spiders. YMMV.
I have made every attempt to avoid duplicate entries, both of actual ads and in the data entry end of things (this is hand-crafted data, let me tell you), but as many as 10 percent of these cars are indeed one and the same. (I mean, 10 percent of the cars appear more than once in the Index, not that one car is represented 10 percent of the time … got that?) The careful reader will note that there are many cases where what appears to be the same car is offered at different prices. You're correct. The trend is that the asking price goes down - I assume until the phone finally rings.
These tell such a interesting story that I've left them in, inasmuch as each is a “discrete” offering of a given car at a given price. This does produce a leveling effect - a $2500 car that later is advertised at $1500 has the net effect of adding one car at $2000 to the overall Index, and adding one data point of a 40% discount (to a current overall rate of apx. 16%) to the thumbnail estimate of how much lower a seller may go. I believe this to be a desirable effect, inasmuch as it tends to keep the average prices higher (but not too high, as you'll see later). I don't intentionally include multiple occurrences of the same car at the same price, but it could happen. The law of large numbers will eventually drive out any overall adverse affect these few idiosyncratic entries may cause. Multiple car deals (buy one, get parts car with purchase) are counted as one car.
When there was any doubt that a particular car wasn't a 124S Spider, I left it out. Ads that said only Fiat convertible could be for an 850—ads for a 124 without additional hints (like, needs new top) could have been for a coupe. There actually were very few cases where any confusion was possible.
One correspondent from Australia questioned the validity of using mileage figures - how can I know if a 68,000 miles-as-advertised car is not actually a 168,000, or 268,000 mile car? Well, I don't. But either there is a stupendous amount of fraud in selling these cars, or most of the mileage figures reflect the Spider's usual role as an “occasional” car. If some of the numbers are suspect, the bulk are accurate, and the results will be close enough for Government work when the sample becomes big enough. In the meantime (in fact, always) use your own best judgment - butt-polished seats, worn pedal pads, and foot-weary carpets are not typical of 40,000 mile creampuffs, even if they are 18 years old.
The point of providing mileage estimates is to provide a reliable gauge against which claims of “low mileage” can be compared. A 1978 Fiat 124S Spider with 78,000 miles on it is not a “low mileage” car. It's actually about average.