2017 Arizona Auction Review - higher estimate lots

The prior post reviewed lower estimate lots at the 2017 Arizona collector car auctions, which built on the post reviewing the performance of the auction companies Bonhams, Gooding & Company, and RM Sotheby’s. This post looks at the cars in the higher estimate group from those three auctions. A car is assigned to the higher estimate group if the mid-estimate (the average of the pre-auction low estimate and high estimate) was at least $250,000. To further help with the evaluation of the strength of the bidding and the accuracy of the pre-auction estimates, the lots are classified by the condition of the car. I did the classification and used the definitions in the Hagerty Price Guide - with an added condition (#5) for non-running basket-case cars. In this higher estimate group, there were no condition #5 category cars and only one condition #4 category car. Condition #3 category cars numbered 37 examples across the three auctions - while #2 category cars and #1 category cars totaled 62 and 10, respectively. Only cars that were sold are included and all prices are before fees.

The scatter plot below shows results for the three auction companies for the higher estimate group. Each auction company and condition group can be plotted separately, and by moving the cursor over each point, the lot, auction, mid-estimate, and sale price can be seen. A single lot can be highlighted after selecting one of the three auctions in the drop-down menu, and then typing the lot number into the text box. Regression equations and trend lines can be plotted too by checking the box. Selecting ‘All’ auctions will perform the regression on the condition group(s) for all three auctions combined (where there is sufficient data), while the other menu options do not combine results for the regressions. The prices are before buyer’s fees.

The only condition #4 category car was a single-family owner 1955 Mercedes-Benz 300 SL Gullwing. Offered by Gooding & Company, lot 18, had a pre-auction mid-estimate of $1,000,000 and sold for $1,325,000. While the condition of the car may have made it hard to accurately value before the auction, the single-family ownership provenance may have encouraged strong bidding too.

The three auction companies had more cars in the condition #3 category group, which resulted in the following regression equation:

$price = $7860 + 0.76 x $mid-estimate

The 0.76 regression coefficient (the beta) suggests that buyers paid 76 percent (before fees) of the pre-auction mid-estimate plus $7,860 (the alpha) for cars in this condition group. The r² of the regression was a high 0.99, which means that the mid-estimate explains 99 percent of the variance in the price. This high r² number may be due to three cars sold at the Bonhams auction. Lot 24 was the 1963 Jaguar E-type Lightweight Competition, which had a pre-auction mid-estimate of $8,250,000 and sold for $6,700,000 (all prices are before fees). 

Lot 44 was the 1952 Ferrari 340 America Spider Competizione s/n 0196A by Vignale, which had a pre-auction mid-estimate of $8,250,000 and sold for $5,800,000. 

Lot 30 was the 1931 Alfa Romeo 6C 1750 5th Series Supercharged Gran Sport Spider by Zagato, which had a pre-auction mid-estimate of $3,100,000 and sold for $2,550,000. 

These three points, and especially lots 24 and 44, make for a clean linear regression. One car that stands out from this regression was auctioned by RM Sotheby’s is lot 232, the 1924 Isotta Fraschini Tipo 8A Landaulet Sala & Riva, which had a pre-auction mid-estimate of $725,000 and sold for $395,000. 

While Isotta Fraschinis used to be popular cars in the U.S. back when they were new, their popularity may now be more limited, which makes pre-auction estimates more challenging and or limits buyer enthusiasm.

The condition #2 category cars were the most numerous of the higher estimate group with 62 cars sold across the three auctions. The regression equation for the category is the following:

$price = -$6224 + 0.84 x $mid-estimate

The 0.84 regression coefficient (the beta) suggests that buyers paid 84 percent (before fees) of the pre-auction mid-estimate, and as such were comfortable paying close to the pre-auction mid-estimate for a car in this condition #2 category. In contrast to the regression of condition #3 category cars, the intercept (the alpha) is a negative $6224, which means that after taking into consideration the pre-auction mid-estimate, cars in the condition #2 category were generally discounted by this amount. The r² of the regression was a slightly lower 0.95, which means that the mid-estimate explains 95 percent of the variance in the price. One car that seems to have played a big part in the regression was auctioned by Bonhams. Lot 54 was the 1928 Mercedes-Benz Type S 26/120/180 Supercharged Sports Tourer Erdmann & Rossi, which had a pre-auction mid-estimate of $5,500,000 and sold for $4,375,000 (all prices are before fees). 

Two lots that seemed to do well for the condition #2 category were sold by RM Sotheby’s, which were lots 121 and 263. Lot 121 was the 2003 Aston Martin DB AR1 by Zagato, which had a pre-auction mid-estimate of $312,500 and sold for $345,000. 

Lot 263 was the 1969 Ferrari 365 GTS s/n 12489 by Pininfarina, which had a pre-auction mid-estimate of $3,200,000 and sold for $3,275,000. 

These cars were either conservatively valued before the auctions and or saw strong bidding during the auctions. On the other side of the results for the condition #2 category of cars, lot 48 at Bonhams and lot 28 at Gooding & Company seemed to be optimistically valued before the auctions and or saw limited bidding during the auctions. Lot 48 at Bonhams was the 1966 Ferrari 275 GTB Scaglietti Pininfarina s/n 08973, which had a pre-auction mid-estimate of $3,000,000 and sold for $1,575,000. 

This result is interesting considering the two Ferrari 275 GTBs that Gooding & Company failed to sell. Lot 46, the 1967 Ferrari 275 GTB/4 s/n 10563 (estimate $3,000,000 to $3,400,000) stopped at a high bid of $2,400,000, and lot 146, the 1966 Ferrari 275 GTB Long Nose Alloy s/n 08117 (estimate $2,900,000 to $3,200,000) stopped at a high bid of $2,500,000. 

Lot 28 at Gooding & Company was the 1936 Bentley 4 1/4 Litre Cabriolet Antem, which had a pre-auction mid-estimate of $600,000 and sold for $320,000. 

Considering that RM Sotheby’s failed to sell lot 153, another condition #2 category car - a 1936 Bentley 3 1/2 Litre Sedanca Coupe by Windovers 

 

with pre-auction mid-estimate of $362,500 but only a high bid of $270,000 – perhaps the markets for Ferrari 275 GTBs and 1930s Bentleys are changing.   

There were fewer best-in-the-world condition #1 category cars in the higher estimate group than in the lower estimate group, at 10 and 15, respectively. With few observations available, the r² was 1.0, meaning that the mid-estimate of the cars in this category explained 100 percent of the variance in the sale price. With so few cars and apparently little disagreement about the value of those cars, the errors for lots in the condition #1 category regression tend to offset each other. The regression equation for the category is the following:

$price = $20 731 + 0.76 x $mid-estimate

In this condition category, the regression coefficient (the beta) of 0.76 means that buyers paid 76 percent of the mid-estimate plus an additional $20731 (the alpha) all before fees. Compared to the regression of condition #1 category cars in the lower estimate group, which had a beta of 0.62 and an alpha of $12963, the values of alpha and beta are greater in the higher estimate group. It suggests that the pre-auction mid-estimates are viewed as a bit more accurate in the higher estimate group and that bidders are willing to pay a higher premium for cars in this condition and estimate group. Two lots that stand out in this condition #1 category group were sold by RM Sotheby’s and were lots 231 and 258. Lot 231 was the 1961 Ferrari 400 Superamerica SWB Coupe Aerodinamico s/n 2841SA by Pininfarina, which had a pre-auction mid-estimate of $3,550,000 and sold for $2,800,000 (all prices are before fees). 

Lot 258 was the 1939 Mercedes-Benz 540K Special Roadster by Sindelfingen, which had a pre-auction mid-estimate of $7,900,000 and sold for $6,000,000. 

 

In terms of scooping up a good deal and or an example of a car with an optimistic pre-auction estimate, RM Sotheby’s lot 134, a 1939 Bugatti T57 Cabriolet by Letourneur et Marchand had a pre-auction mid-estimate of $1,375,000 and sold for $925,000.

Disclaimer

JohnWileyCars provides no warranty about the content or accuracy of the content published here. JohnWileyCars shall not be liable for any loss of profit or any other damages resulting from the use of the information published here.

Arizona Auctions Review

2017 Arizona Auctions Review

With the 2017 Arizona collector car auction cycle now complete, here is a look at how the results of the high-end auction companies (Bonhams, Gooding & Company, and RM Sotheby’s) compared. While all three had sales rates between 80 and 90 percent and each sold over $30 million of cars (all prices referenced in this post are before fees), differences in the results appear when making a distinction in the cars that were offered. Unlike a classification by condition as with the post on RM Sotheby’s Duemila Ruote auction, here the distinction is a price threshold of $250,000. Cars with a pre-auction mid-estimate (the average of the low and high estimate) that was less than $250,000 go in one group, and those with a pre-auction mid-estimate above that threshold go in the other group. Using this distinction, the first auction company performance difference is that one of the auction company’s pre-auction mid-estimates are much better at predicting the sale price of cars in the lower estimate group, but their mid-estimates are the worst at predicting the sale price of cars in the higher estimate group. Similarly, one auction company had pre-auction mid-estimates that were the best at predicting sale prices for cars in the higher estimate group, but their mid-estimates were the worst for cars in the lower estimate group. For all three auction companies, their pre-auction mid-estimates were either consistently too high or too low, depending on the price group. Given the price group variations among auction companies, these findings could be helpful in determining which is the best company to buy or sell a high or low priced car.

The scatter plot below shows results for the three auction companies. Each price group can be plotted separately, and by moving the cursor over each point, the lot, mid-estimate, and sale price can be seen. The prices are before buyer’s fees and prices for unsold lots are zero.

The company with pre-auction mid-estimates that were the most accurate in predicting the sale price of cars in the lower estimate group was Gooding & Company. Their sale rate for the group was 100 percent (compared to 89 percent for Bonhams and 99 percent for RM Sotheby’s), which helped raise the correlation coefficient of their mid-estimates to sale prices to 0.72 (compared to 0.50 for Bonhams and 0.61 for RM Sotheby’s). Their R² (the amount of variance in the sale price explained by the pre-auction mid-estimate) was also the highest at 0.74 (compared to 0.31 for Bonhams and 0.58 for RM Sotheby’s).

Gooding & Company’s accuracy with pre-auction estimates in the lower estimate group is despite the sale of lot 62, the 1958 Porsche 356 A Super Speedster, which sold for $310,000 (estimate range $200,000 to $275,000). This car was a basket-case, and as was observed in the results of the Duemila Ruote auction, these cars are difficult to value accurately.

Unfortunately, Gooding & Company seemed to do poorly with the sale of cars in the higher estimate group as shown by their low R² of 0.32 for this group (compared to 0.91 for Bonhams and 0.86 for RM Sotheby’s). While Gooding & Company’s sale rate was not the worst at 67 percent (compared to 66 percent for Bonhams and 80 percent for RM Sotheby’s), instead it is probably because Gooding & Company’s unsold lots had the highest total value relative to the total value of the lots they did sell in that higher estimate group. Adding up the mid-estimates of their unsold lots totals $17,775,000, which is 62 percent of $28,865,000, which is the sum of the mid-estimates of the lots they did sell. At Bonhams, adding up mid-estimates of unsold lots is only $8,387,500, which is just 23 percent of $36,587,500, which is the sum of the mid-estimates of sold lots. RM Sotheby’s has a similar percentage of unsold value to sold value at 21 percent, which is the $10,612,500 sum of mid-estimates for unsold lots to the $51,470,000 sum of mid-estimates for sold lots.

Gooding & Company seemed to suffer most in the higher estimate group due to a couple of lots that failed to sell. Lot 46, the 1967 Ferrari 275 GTB/4 s/n 10563 (estimate $3,000,000 to $3,400,000) stopped at a high bid of $2,400,000, and lot 146, the 1966 Ferrari 275 GTB Long Nose Alloy s/n 08117 (estimate $2,900,000 to $3,200,000) stopped at a high bid of $2,500,000. Had these two cars sold, along with lot 34, the 1936 Squire 1 ½ Litre Tourer Ranalah (estimate $1,500,000 to $2,000,000) that stopped at the high bid of $1,050,000 - Gooding & Company’s results in the higher estimate group would have improved.

Bonhams had pre-auction mid-estimates that were the best predictors for the final sale price for cars in the higher estimate group, but that was partly because they offered the fewest number of lots in that price group with just 29 (compared to 61 at Gooding & Company and 64 at RM Sotheby’s). The correlation coefficient for Bonhams was 0.76 (compared to 0.44 for Gooding & Company and 0.78 for RM Sotheby’s). The R² of their pre-auction mid-estimates to sale prices was 0.91 (compared to 0.32 for Gooding & Company and 0.86 for RM Sotheby’s).

Bonhams results in the higher estimate group were helped by two sales, which were lot 44, the 1952 Ferrari 340 America Spider Competizione by Vignale s/n 0196A, which sold for $5,800,000 (estimate range $7,500,000 to $9,000,000), and lot 24, the ex-Bob Jane 1963 Jaguar E-type Lightweight Competition s/n S850667 engine no. V682558P, which sold for $6,700,000 (estimate range $7,500,000 to $9,000,000).

In a similar reversal to the predictive power of Gooding & Company’s pre-auction mid-estimates with respect to sale prices, Bonhams had the worst performance when it came to cars in the lower estimate group. It is likely because their sale rate for the lower estimate group cars was the lowest at 89 percent (compared to 100 percent at Gooding & Company and 99 percent at RM Sotheby’s). The correlation coefficient for Bonhams in the lower estimate group of pre-auction mid-estimate to sale price was just 0.50 (compared to 0.72 for Gooding & Company and 0.61 for RM Sotheby’s). The R² value of variation explained also was the lowest at 0.31 for Bonhams (compared to 0.74 for Gooding & Company and 0.58 for RM Sotheby’s).

One final observation of the results of the auction companies was the direction and magnitude of the difference between the pre-auction mid-estimate and the sale price. In the lower estimate group, all three auction companies had regression intercepts that were greater than zero. The implication is that (after multiplying the mid-estimate by a factor in the range of 0.50 to 0.72) the estimates for the lower estimate group were consistently less than the sale price by $13,441 to $13,904 – except for Gooding & Company which had estimates that were just $3,492 too low. The opposite was true in the higher estimate group, where excluding the poor results from Gooding & Company, the intercept was -$211,893 for Bonhams and -$93,699 for RM Sotheby’s (Gooding & Company had an intercept of +$54,261). The implication being that both Bonhams and RM Sotheby’s had pre-auction estimates for cars in the higher estimate group that (after multiplying by a factor of 0.76 to 0.78) overestimated the sale price by between $93,699 and $211,893. Given the coefficients for both price groups were between 0.50 and 0.78, but the resulting intercepts were positive for the lower estimate group and negative for the higher estimate group, it seems buyers were more aggressive in bidding for cars in the lower estimate group.

Disclaimer

JohnWileyCars provides no warranty about the content or accuracy of the content published here. JohnWileyCars shall not be liable for any loss of profit or any other damages resulting from the use of the information published here.