14
© Property Data Solutions Over Priced Listings PriceFinder Research 7/1/09 Kent Lardner www.pricefinder.com.au

Over Priced Listings

Embed Size (px)

Citation preview

© Property Data Solutions

Over Priced Listings PriceFinder Research 7/1/09

Kent Lardner

www.pricefinder.com.au

© Property Data Solutions

Do over priced real estate listings usually lead to unsold properties?

To answer this question, PriceFinder Research conducted an analysis of properties yet to be sold and still listed on the market from 50 days (up to a maximum of 150 days). A total of 73 property listings were collected from the PriceFinder database from 13 suburbs in both NSW and QLD.

The PriceFinder system was then used to estimate the price for each listing. This new estimated price was then used to calculate how much each listing was over or under priced.

Our results confirmed that in nearly 50% of cases, the unsold properties had been listed at prices greater than 10% above the PriceFinder estimate.

Introduction

© Property Data Solutions

Methods Used

The variables measured and analysed included:

• the list price itself (to see if higher priced properties take longer to sell)

• the suburb and the total amount of listings for that location

• median price for that suburb

• the PriceFinder estimate

• the percentage over/underpriced and the total days on the market

The data was then prepared using Excel and analysed using the statistical packages of GRETL and MINITAB.

© Property Data Solutions

PriceFinder Estimation Method

PriceFinder allows a user to compare the quality of the comparable sale to the subject property.

The system has already adjusted for size differences, all the user needs to do is compare is the overall quality and street appeal of each property to the subject.

If a comparable is not suitable, click ‘Remove and replace’.

© Property Data Solutions

Results

The listings tested had a median of 72 days unsold. The median percentage over-priced according to the PriceFinder estimate was 10%.

Regression tests indicated a positive relationship between the number of days unsold and the total price error. As a general guide, for every 1% over priced a property would remain unsold for an extra 5 days. For example, a house over-priced by 10% would remain unsold for an extra 50 days based on this model.

Almost half of the properties re-valued fell within the 10% error range. In practice this is generally accepted as the norm given the heterogeneous nature of housing.

© Property Data Solutions

Conclusions and Further Discussion

When a property is over priced or stays listed on the market for a long period, interest wanes. Potential buyers will often become suspicious and ignore the listing, or move on to another property that is more realistically priced.

Whilst the total amount of listings on the market and median price has some level of impact on how quickly a property will sell, setting the right listing price is the most important factor.

Another key consideration is how much a market may change over an extended period of time. A property listed for 50 days or longer may experience a number of changes in the market as a result of new listings and recent sales during this period. Similar properties that sell within the period at prices below the current list price will have a significant impact on any subsequent valuation of the listed home. Likewise, a significant jump in listings will also place downward pressure on the price.

PriceFinder Research will continue it’s study on the relationship between listing price and days on the market. Further studies will include analysis of all listings (from 1 day to 200 days) and include total stock levels and ‘market absorption’ rates.

© Property Data Solutions

Appendix

1. Graph 1 - Total Days Unsold

2. Graph 2 – List Price

3. Graph 3 – Estimated Price Error

4. Least Absolute Deviation – Test Results

5. Histogram for Price Error (%)

6. Data Used

© Property Data Solutions

Graph 1 - Total Days Unsold

130110907050

95% Confidence Interval for Mu

908070

95% Confidence Interval for Median

Variable: Days

69.682

21.586

78.390

Maximum3rd QuartileMedian1st QuartileMinimum

NKurtosisSkewnessVarianceStDevMean

P-Value:A-Squared:

82.635

29.992

90.103

141.000 99.000 72.000 65.000 50.000

73-2.7E-010.980886

630.02225.100284.2466

0.0003.974

95% Confidence Interval for Median

95% Confidence Interval for Sigma

95% Confidence Interval for Mu

Anderson-Darling Normality Test

Descriptive Statistics

The graph indicates that the median average for total days unsold from our sample is 72. Please note that we have selected properties that are still listed after 50 days, so this average does not reflect a total market average.

Another key observation is the normality test. When the P-Value is <0.05 the data is not considered normally distributed, limiting some further model uses.

© Property Data Solutions

Graph 2 – List Price

2200000180000014000001000000600000200000

95% Confidence Interval for Mu

720000620000520000420000

95% Confidence Interval for Median

Variable: List

430000

328744

527632

Maximum3rd QuartileMedian1st QuartileMinimum

NKurtosisSkewnessVarianceStDevMean

P-Value:A-Squared:

590000

456770

706012

2300000 732000 500000 377500 189000

735.057132.07889

1.46E+11382267616822

0.0004.363

95% Confidence Interval for Median

95% Confidence Interval for Sigma

95% Confidence Interval for Mu

Anderson-Darling Normality Test

Descriptive Statistics

The graph indicates that the median average for list prices from our sample is $500,000. The data is highly skewed with the highest priced property at $2.3M and lowest at $189k.

Another key observation is the normality test. When the P-Value is <0.05 the data is not considered normally distributed, limiting some further model uses.

© Property Data Solutions

Graph 3 – Estimated Price Error

0.350.250.150.05-0.05

95% Confidence Interval for Mu

0.1350.1250.1150.1050.0950.0850.075

95% Confidence Interval for Median

Variable: Percent

0.076824

0.078897

0.090239

Maximum3rd QuartileMedian1st QuartileMinimum

NKurtosisSkewnessVarianceStDevMean

P-Value:A-Squared:

0.120000

0.109623

0.133049

0.3900000.1650000.1000000.040000-7.0E-02

730.2214510.7320978.42E-030.0917430.111644

0.0111.002

95% Confidence Interval for Median

95% Confidence Interval for Sigma

95% Confidence Interval for Mu

Anderson-Darling Normality Test

Descriptive Statistics

The graph indicates that the median average for our price error is 10%.

Another key observation is the normality test. When the P-Value is <0.05 the data is not considered normally distributed, limiting some further model uses.

© Property Data Solutions

Least Absolute Deviation – Test Results

LAD estimates using the 73 observations 1-73Dependent variable: Days

VARIABLE COEFFICIENT STDERROR T STAT P-VALUE

Percent 500.000 51.4366 9.721 <0.00001 ***

Mean of dependent variable = 84.2466 Standard deviation of dep. var. = 25.1002 Sum of absolute residuals = 3567 Sum of squared residuals = 259293

Using multiple independent variables including median price, total listings and percentage over/under priced, all were found to be statistically significant.

However the coefficient values for both total listings and median price was very low and hence removed from the final model in this instance.

As the sample was skewed (only 50 to 150 days listed) the regression model has been included as a general guide only.

LAD estimates using the 73 observations 1-73Dependent variable: Days

VARIABLE COEFFICIENT STDERROR T STAT P-VALUE

Percent 130.102 36.8794 3.528 0.00074 *** Median 4.00400E-05 1.06563E-05 3.757 0.00035 *** Listings 0.631788 0.0829639 7.615 <0.00001 ***

Mean of dependent variable = 84.2466 Standard deviation of dep. var. = 25.1002 Sum of absolute residuals = 1858.38 Sum of squared residuals = 79142.5

© Property Data Solutions

0.40.30.20.10.0-0.1

30

20

10

0

Percent

Fre

qu

en

cyHistogram of Percent

Summary Statistics for Discrete Variables

Percent Count CumCnt Percent CumPct

-0.07 1 1 1.37 1.37

-0.03 1 2 1.37 2.74

0.00 7 9 9.59 12.33

0.02 1 10 1.37 13.70

0.03 5 15 6.85 20.55

0.04 4 19 5.48 26.03

0.05 4 23 5.48 31.51

0.06 3 26 4.11 35.62

0.07 2 28 2.74 38.36

0.08 4 32 5.48 43.84

0.09 3 35 4.11 47.95

0.10 4 39 5.48 53.42

0.11 3 42 4.11 57.53

0.12 4 46 5.48 63.01

0.13 3 49 4.11 67.12

0.14 3 52 4.11 71.23

0.15 2 54 2.74 73.97

0.16 1 55 1.37 75.34

0.17 1 56 1.37 76.71

0.19 5 61 6.85 83.56

0.20 2 63 2.74 86.30

0.22 2 65 2.74 89.04

0.23 1 66 1.37 90.41

0.25 1 67 1.37 91.78

0.28 1 68 1.37 93.15

0.30 2 70 2.74 95.89

0.31 2 72 2.74 98.63

0.39 1 73 1.37 100.00

N= 73

The chart and data table display the counts and cumulated percentage for price errors. For example, 47.95% of the sample are 9% and lower.

Histogram for Price Error (%)

© Property Data Solutions

Data Used

STATE Suburb Address 1 List PF Percent Days Median ListingsQLD Albany Ck 18 Atkinson 600000 570000 0.05 80 475000 70QLD Albany Ck 4 Bauple 500000 476000 0.05 120 475000 70QLD Albany Ck 19 Silvereye 530000 495000 0.07 64 475000 70QLD Albany Ck 9 Iverness 750000 670000 0.12 66 475000 70QLD Albany Ck 3 Boronia 750000 670000 0.12 86 475000 70QLD Albany Ck 17 Windemere 430000 380000 0.13 138 475000 70NSW Belrose 19 Corbett 715000 715000 0.00 139 795000 32NSW Belrose 44 Perentie 900000 885000 0.02 98 795000 32NSW Belrose 41 Windrush 745000 725000 0.03 121 795000 32NSW Belrose 69 Ralston 665000 647000 0.03 133 795000 32NSW Belrose 35 Contentin 895000 860000 0.04 101 795000 32NSW Belrose 2 St Andrews 1050000 915000 0.15 85 795000 32NSW Belrose 61 Pringle 1000000 770000 0.30 64 795000 32NSW Camden 93 Fitzwilliam 390000 370000 0.05 62 356500 24NSW Camden 40 lerida 390000 345000 0.13 81 356500 24NSW Camden 43 menangle 895000 745000 0.20 141 356500 24NSW Camden 49 old hume 419000 320000 0.31 68 356500 24QLD Carindale 12 Turnbul 775000 740000 0.05 76 627000 74QLD Carindale 11 Cornouste 900000 820000 0.10 50 627000 74QLD Carindale 5 Mindup 670000 550000 0.22 70 627000 74NSW Coffs 20 Bray 285000 285000 0.00 64 350000 132NSW Coffs 69 Boultwood 300000 300000 0.00 64 350000 132NSW Coffs 165 Combine 275000 275000 0.00 65 350000 132NSW Coffs 19 Hull 339000 320000 0.06 65 350000 132NSW Coffs 3 Orman 330000 302000 0.09 135 350000 132NSW Coffs 24 Roselands 505000 460000 0.10 72 350000 132NSW Coffs 11 Michelle 375000 340000 0.10 127 350000 132NSW Coffs 6 Namoi 385000 324000 0.19 129 350000 132NSW Grafton 31 Cranworth 240000 226000 0.06 63 252000 60NSW Grafton 28 Smith 215000 195000 0.10 63 252000 60NSW Grafton 152 Turf 189000 170000 0.11 69 252000 60NSW Grafton 205 Hoof 280000 246000 0.14 98 252000 60NSW Grafton 108 Hoof 340000 296000 0.15 70 252000 60NSW Grafton 61 Riverdale 479000 388000 0.23 114 252000 60NSW Grafton 30 Oliver 315000 240000 0.31 54 252000 60NSW Kellyville 36 Janamba 575000 575000 0.00 70 555000 89NSW Kellyville 50 Fraser 489000 489000 0.00 100 555000 89

NSW Kellyville 5 Clovelly 590000 575000 0.03 62 555000 89NSW Kellyville 9 Indigo 719000 636000 0.13 66 555000 89NSW Kellyville 14 Hutchinson 610000 525000 0.16 60 555000 89NSW Murwillumbah 22 George 345000 345000 0.00 78 363000 93NSW Murwillumbah 9 Saddle 449000 430000 0.04 77 363000 93NSW Murwillumbah 24 prince 460000 432000 0.06 98 363000 93NSW Murwillumbah 23 Myrtle 425000 395000 0.08 141 363000 93NSW Murwillumbah 5 Tombonda 355000 325000 0.09 85 363000 93NSW Murwillumbah 18 MCPherson 420000 354000 0.19 80 363000 93NSW Murwillumbah 7 Golden Links 590000 471000 0.25 93 363000 93NSW Newtown 63 Ambermark 649000 626000 0.04 72 650000 30NSW Newtown 226 Edgeware 640000 560000 0.14 71 650000 30NSW Newtown 52 Simmons 600000 501000 0.20 71 650000 30NSW Parramatta 38 Crimea 435000 390000 0.12 100 430000 27NSW Parramatta 76 Hassall 439000 370000 0.19 85 430000 27NSW Parramatta 40 Crimea 550000 396000 0.39 100 430000 27NSW Pymble 46 Bannockburn 1275000 1180000 0.08 84 1233000 55NSW Pymble 20 Greenway 1750000 1570000 0.11 120 1233000 55NSW Pymble 20 Narrelle 1129000 1011000 0.12 64 1233000 55NSW Pymble 14 Anatoc 1750000 1540000 0.14 71 1233000 55NSW Pymble 38 Station 849000 715000 0.19 64 1233000 55NSW Pymble 74A Pymble Rd 2300000 1800000 0.28 69 1233000 55NSW Randwick 2 Chtham 1460000 1500000 -0.03 58 1120000 27NSW Randwick 44A Clovelly 1250000 1200000 0.04 66 1120000 27NSW Seven Hills 13 Zermatt 380000 370000 0.03 57 360000 45NSW Seven Hills 19 veronica 479000 447000 0.07 62 360000 45NSW Seven Hills 15 Veronica 570000 530000 0.08 82 360000 45NSW Seven Hills 2a Johnstone 300000 271000 0.11 68 360000 45NSW Seven Hills 51 Abott 300000 252000 0.19 67 360000 45NSW Seven Hills 11 ATHABASKA 365000 300000 0.22 137 360000 45NSW Seven Hills 428 Seven Hills Rd 335000 258000 0.30 72 360000 45NSW Toongabbie 377 Wentworth 430000 460000 -0.07 99 386000 72NSW Toongabbie 131 Rawsch 540000 524000 0.03 99 386000 72NSW Toongabbie 43 Fitzwilliam 390000 360000 0.08 63 386000 72NSW Toongabbie 49 Bungaree 430000 395000 0.09 76 386000 72NSW Toongabbie 4 Illoca 585000 500000 0.17 68 386000 72

The data table below has used listings from December 2008. PF indicates the PriceFinder estimate. Percent shows the difference between the List price and PriceFinder. Days is the total days listed. Median is the overall suburb median priced for Q4/2008 for houses. Listings is the total amount of listings for that suburb (from 1 to n days).

© Property Data Solutions