fftfk

I've decided to do a little analysis of 964 prices using data from the Rennlist 964 Price Sticky (https://rennlist.com/forums/964-foru...64-prices.html), 2015 Fleabay sales prices as documented in the 964 Price sticky, and prices from BAT. The data set includes 427 records starting in 1991 and going through 2018. It includes sales data including price, mileage, model year, body type, Manual/Tip, and data source. I excluded outliers such as speedsters, RS America, etc. as I felt these cars unfairly skewed the prices paid.


The raw data I used can be found here:

https://docs.google.com/spreadsheets...DVrpxV/pubhtml


I originally started this project to kick the hornet's nest and prove that while your average Porschophile would say the Tiptronic transmission is absolutely worthless actual sales data does not support this view. Generally, prices for manuals are about 14% higher than tiptronics. In addition, I felt the data lent itself to additional analysis so decided to go a little further with it.


Manual Vs Tiptronic Vs Body Type:


The average purchase price of a manual transmission was $29,593 while the average purchase price of a tiptronic was $25,921. In other words, the average tiptronic purchase price was 87.5% of the average manual purchase price. The average sales price of a cabriolet was $30,047 with a manual going for $31,451 and a tiptronic going for $25,641. The average sales price of a coupe was $28,713 with manual's going for $28,933 and tiptronics going for $27,207. The average sales price of a Targa was $23,155 with a manual going $23,979 and a tiptronic going for $18,625.



Trends


I excluded sales for 2007 and prior for the next analysis in order to better analyze trends with the air-cooled "bubble." As expected, the average price of model years 89-91 ($27,444) is less than the average price of model years 92-94 ($34,506). Manual transmissions are generally exhibited stronger prices than tiptronics although the past two years have shown increasing prices in tiptronics. It is possible the reported 2018 tiptronic price is an outlier or else it could be the continuation of a trend started in 2016. Finally, average prices of all 964 have increased pretty steadily from 2010 on. In 2010 the average sales price was $18,717 while the average price in 2017 was $51,155. An exponential smoothing model with a 90% prediction interval suggests prices will continue to increase. The predicted price in 2018 is $53,253 and in 2019 is $60,314.



Updated on 3/12/18 at 10:24

All - I have performed additional cleaning of the data based on feedback below. The cleaned file is linked here:

https://docs.google.com/spreadsheets...DVrpxV/pubhtml

Mileage Impact On Price

I used an ordinary least-squares algorithm to perform a regression analysis using mileage to predict price. While the results indicate that increasing mileage leads to lower prices the impact is not as great as expected. The resulting model is Price = 37,740 - .1275M + e where M is the mileage and e is the error term. The summary results:

Call:
## lm(formula = Purchase.Price ~ Mileage, data = porsche.clean)

##Residuals:
## Min 1Q Median 3Q Max
##-25496 -7919 -3970 5422 101187

##Coefficients:
## Estimate Std. Error t value Pr(> t )
##(Intercept) 3.774e+04 1.503e+03 25.101 < 2e-16 ***
## Mileage -1.275e-01 1.782e-02 -7.155 3.72e-12 ***
## ---
## Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

##Residual standard error: 12870 on 423 degrees of freedom
##Multiple R-squared: 0.108, Adjusted R-squared: 0.1058
##F-statistic: 51.19 on 1 and 423 DF, p-value: 3.72e-12

The intercept is 37,740 basically meaning the model predicts the price of a 964 with zero mileage is $37,740 +/- error term. The mileage coefficient is -.1275 meaning that each mile driven decreases the car's value $.1275. Both the intercept and skew were determined to be significant with small p-values. The adjusted R-squared is relatively small indicating that the mileage is responsible for about 11% of the car's value. I believe this is indicative of the common phrase that Porsches are meant to be driven and not sitting in a garage. The overall p-value is extremely small indicating that the model results are significant. In other words, while mileage is not the primary driver of a 964's price is a significant driver of the price. The red line in the following chart is the predicted value of a 964 based upon price while the o's are actually recorded data points.

Marine Blue

John McM

Railmaster.

Will the 964 eventually pass its main enemy the 993?

And what do You do for a living? Just curious!

GDOG8

fftfk

Marine Blue - While it is fully possible I made transcription mistakes in creating the data overall I think it is fairly accurate especially in that we are dealing with a decent sized data set's averages. A few errors shouldn't have that large of an impact. I agree that among similar quality cars the pricing should generally be Coupe, Targa, Cabriolet. Since the data doesn't reflect this the question must be asked why. My guess is that Coupe's are generally used more/more aggressively than cabriolets. For example, I have a cabriolet in Chicago and it generally only get's used during about 6 months of the year and I wouldn't consider using it for anything more than casual DE events. In fact, my cab only has about 44,000 miles. A coupe can be used year round, tracked, etc. I think the data is reflecting that there are more low quality coupes that get sold bringing the overall price averages down and that Cabriolet's tend to get used less and less harshly. The Targa pricing I think is just a reflection of the limited number of sales and the time of those sales (pre-bubble). As for C2/C4 delineation, I thought to myself about 3/4 of the way through the data that I should have included that data point!!!!! If you want to go back and add it in I would be happy to do further analysis.

John McM - Is there a source you are getting those prices from or is it just your guy feeling? I'm not denying them just wondering where they are from.I really wish you could look up car sales prices the way you can house prices. It would make the data much more accurate.

Railmaster - I'm a fan of the 993 too but only own a 964 so I hope it does! I am taking Master's level data analytics courses. I was going to attempt to do some regression analysis using R but then thought it would just be easier to feed the data into tableau.

Gdog8 - I agree that condition, history, and mileage are the driving forces but tip/no-tip definitely has a strong impact.

fftfk

Also...I hope the predictions are accurate!

John McM

Ralph3.

Did you do anything with mileage?

fftfk

I'll see about regressing price against mileage when I get home. I'll also look at adding C2/C4 info. I think I can do it without too much additional work.

Marine Blue

Genuinely appreciate the effort, thank you!

Ralph3.

I was playing with your data after dumping it into excel. You might look at a couple more outliers, line 18 looks like the mileage is off, line 91 because it is the only flatnose in the data.

Also the model year is wrong on line 411, should be 1990, I know because I am the idiot who purchased it.

cobalt

I applaud the work but the problem with the data is based strictly on rennlist, Ebay and Bat sales which is not the be all and end all. On average I find rennlst sales to be 25-30% shy and only about 70% of all sales. Of what I see sell privately most cars are far exceeding 993 sales especially for clean 92-94 C2 coupes which have brought some huge numbers well into 6 figures. The data doesn't lie but it doesn't tell the entire story either.

fftfk

Ralph 3 - I just double checked and the reported mileage is actually 51 miles (post #29). I agree that is most likely an error and took it out. I took out the flatnose and adjusted your model year although I have to say you can't be that big of an idiot...you bought a 964.

@cobalt - I agree the data doesn't tell the whole story. If you have a better/additional source I will include it.

cobalt

I wish I could but I am privy to private sales data that I could provide but can't corroborate without upsetting those involved. All I can say is there are considerable cars that sell outside our world here.