Response to Data Cited from CT Post on Fairfield Home Values
At the RTM meeting on Monday, April 22, 2013, a member of the public cited an article from the April 12, 2013 edition of the Connecticut Post, as a basis for suggesting that the pressure on Fairfield’s home prices is perhaps not as great as some, including Fairfield Taxpayer, have feared, both in absolute terms and relative to other towns. We believe this conclusion is both intuitively and demonstrably wrong.
We believe that the numbers published by the CT Post – which implausibly showed Fairfield’s home prices down only 15% from their 2007 peak – are distorted and misleading, probably because they are based on median prices. Based on what we believe is better data from the University of Connecticut Business School that tracks mean (average) selling prices for “constant quality” single-family homes in most CT towns, Fairfield’s home prices are among the weakest over the same 2007-2012 period used by the CT Post, down almost 35% versus declines of only 21% and 24% for towns like Greenwich and Darien, where tax rates are substantially lower.[1]
Mean (Average) Constant Quality Single Family Home Sales by Town
|
Towns
|
|
2007 Value
|
|
2012 Value
|
|
Decline
|
| |
|
|
|
|
|
|
|
Greenwich
|
|
$1,186,600
|
|
$943,300
|
|
(20.5)%
|
|
Stamford
|
|
657,300
|
|
520,800
|
|
(20.8)
|
|
Guilford
|
|
430,200
|
|
328,000
|
|
(23.8)
|
|
Darien
|
|
1,102,900
|
|
833,500
|
|
(24.4)
|
|
Madison
|
|
493,500
|
|
372,500
|
|
(24.5)
|
|
Weston
|
|
767,500
|
|
573,800
|
|
(25.2)
|
|
Westport
|
|
1,314,100
|
|
978,400
|
|
(25.5)
|
|
Norwalk
|
|
541,300
|
|
402,500
|
|
(25.6)
|
|
Wilton
|
|
908,300
|
|
671,700
|
|
(26.0)
|
|
Shelton
|
|
454,000
|
|
332,800
|
|
(26.7)
|
|
New Canaan
|
|
1,357,700
|
|
969,500
|
|
(28.6)
|
|
Milford
|
|
353,700
|
|
246,800
|
|
(30.2)
|
|
FAIRFIELD
|
|
553,900
|
|
361,800
|
|
(34.7)
|
|
Bridgeport
|
|
247,000
|
|
158,300
|
|
(35.9)
|
|
Stratford
|
|
285,700
|
|
179,500
|
|
(37.2)
|
|
Trumbull
|
|
475,900
|
|
292,400
|
|
(38.5)
|
As it happens, Fairfield Taxpayer has a major research effort underway to quantify exactly what is happening to home prices in Fairfield both in absolute terms in each segment of the market (i.e., higher end, middle and lower end) and relative to surrounding towns. We will present our full analysis as soon as possible, but meanwhile wanted to respond to the CT Post data presented at the RTM meeting.
The CT Post data, which are provided in the next table, show Fairfield’s home prices down only 15% over the 2007-2012 period, which we all dearly wish were true.
Median Selling Price Data from Connecticut Post Article
| |
|
|
|
|
2007-12
|
|
Jan.-Feb.
|
|
2012-13
|
|
Towns
|
2007 Value
|
|
2012 Value
|
|
Decline
|
|
2013 Value
|
|
Change
|
| |
|
|
|
|
|
|
|
|
|
|
Darien
|
$1,330,000
|
|
$1,200,000
|
|
9.8%
|
|
$950,000
|
|
(20.8)%
|
|
Guilford
|
425.000
|
|
380,000
|
|
10.6%
|
|
375,000
|
|
(1.3)%
|
|
Fairfield
|
615,000
|
|
521,500
|
|
15.2%
|
|
415,000
|
|
(20.4)%
|
|
Wilton
|
891,000
|
|
735,000
|
|
17.5%
|
|
631.250
|
|
(14.1)%
|
|
Weston
|
935,000
|
|
755,000
|
|
19.3%
|
|
793,000
|
|
5.0%
|
|
Stamford
|
695,000
|
|
550,000
|
|
20.9%
|
|
494,000
|
|
(10.2)%
|
|
Milford
|
344,000
|
|
270,000
|
|
21.5%
|
|
244,500
|
|
(9.4)%
|
|
Ridgefield
|
777,500
|
|
607,500
|
|
21.9%
|
|
541,500
|
|
(10.9)%
|
|
Norwalk
|
525,000
|
|
410,000
|
|
21.9%
|
|
361,000
|
|
(12.0)%
|
|
Shelton
|
380,000
|
|
294,900
|
|
22.4%
|
|
290,000
|
|
(1.7)%
|
|
Trumbull
|
460,000
|
|
350,000
|
|
23.9%
|
|
288,000
|
|
(17.7)%
|
|
New Canaan
|
1,685,000
|
|
1,280,000
|
|
24.0%
|
|
950,000
|
|
(25.8)%
|
|
Westport
|
1,315,500
|
|
1,000,000
|
|
24.0%
|
|
1,238,250
|
|
23.8%
|
|
Greenwich
|
2,030,000
|
|
1,520,000
|
|
25.1%
|
|
1,800,000
|
|
(18.4)%
|
|
Easton
|
745,000
|
|
555,500
|
|
25.4%
|
|
468,475
|
|
(15.7)%
|
|
Madison
|
539,500
|
|
401,875
|
|
25.5%
|
|
309,900
|
|
(22.9)%
|
|
Stratford
|
285,000
|
|
208,900
|
|
26.7%
|
|
197,500
|
|
(5.5)%
|
|
Bridgeport
|
248,000
|
|
130,000
|
|
47.6%
|
|
149,000
|
|
14.6%
|
The data for this table came from an organization called The Warren Group,[2] and anyone can obtain seven days of free, limited access to their database at no cost. When we checked the CT Post numbers against their source, we found a number of errors that change the results primarily for Guilford (which was down only 10.6%, not 27.3%), Norwalk (down 21.9%, not 26.7%), Westport (down 24.0%, not 34.0%), and Madison (down 25.5%, not 27.7%).
However, more important than these clerical errors is the fact that these data are based on median selling prices.
Why Median Prices Can be Distorted and Misleading
The fundamental problem with using median data to analyze changes in home values is that: (a) the trend in median home values over time (a.k.a., a “time series”) can be significantly distorted by shifts the mix of the homes being sold; and (b) comparisons of changes in median home values between two or more towns can be significantly distorted unless the towns have similar numbers of homes in each price segment.
If the mix of the homes that are sold changes significantly in terms of high, medium, and low segments: (a) the median selling price can change significantly, even if all the homes were sold at exactly the same price for which they sold five years earlier; and (b) the median selling price can remain unchanged even if home prices rise or fall significantly. A pretty good explanation from an independent source of how median prices can be misleading is as follows:
“We have observed before that most market areas contain a number of price ranges. Think of them as layers, something like geological strata. Consider a hypothetical example, greatly more simplified than any real market. There are three levels to this imaginary market: Starter Homes that currently range from $150,000 - $200,000; Mid-range Homes that are $300,000 - $500,000; and Luxury Homes that run from $800,000 - $1 million. Now, if in one year there were 20 sales of Starter Homes, 5 sales of Mid-range Homes, and 2 sales of Luxury Homes, the median would be somewhere in the Starter Homes range of $150,000 - $200,000. Suppose that in the next year, the Mid-range Homes and the Luxury Homes each experienced price declines in the neighborhood of 30%. (It happens.) Now, their respective price ranges would be $210,000 - $350,000 and $560,000 - $700,000. Suppose that, along with those price declines, sales of those higher priced properties picked up. Let's say that in the next year there were 5 sales of Starter Homes, 15 sales of Mid-range Homes, and 7 sales of Luxury Homes. The median then would be in the middle range of $210,000 - $350,000. It would be higher than the year before, even though it was reflective of sales of properties whose values had declined.”[3]
How Can We Explain the Results from the CT Post Data?
Without access to the raw data to determine exactly how many sales took place in each meaningful segment of the market, we can’t definitively explain the results of the Warren Group data.
However, we know that two primary sources of housing mix distortion in the 2007-2012 period were: (a) foreclosure sales at distress prices, which have increased the number of lower-end home sales in some towns; and (b) changes in the tax laws, which stimulated the sale of primarily higher-end homes in 2012 in anticipation of a higher capital gains tax rate (20% up from 15%) and the imposition of a 3.8% Medicare tax on investment gains, both of which became effective on January 1, 2013.
Foreclosure sales matter because when there are a lot of them, as there have been in a town like Bridgeport, they pull down the median sales price. Thus, we should not be surprised that Bridgeport showed the greatest decline in median selling price between 2007 and 2012 in the CT Post data. For the same reason, we should not be surprised if, when most foreclosure sales have been completed, that Bridgeport shows one of the strongest increases in median home price.
Anticipated tax increases would have had the greatest impact on median selling prices in towns that had a significant number of higher-end homes that have been owned long enough that they are still worth significantly more than the owner paid, despite the substantial decline in housing values from the 2007 peak. This may have happened in Fairfield in 2012, and if so, it could have shifted the mix of sales toward the higher end and thus shifted the median price higher than it would otherwise have been. If this theory is correct, then Fairfield’s median price should show a greater decline in 2013 than otherwise. Indeed, the results from the Warren Group for the first two months show the median price in Fairfield down 20.4% from 2012. Other towns with similarly sharp declines include: New Canaan (25.8%), Madison (22.9%), Darien (20.8%), Greenwich (18.4%) and Trumbull (17.7%). Since these could be statistical flukes (particularly because so few sales close in the first two months of the year – only 35 in Fairfield versus 552 in all of 2012, according to Warren), we will continue to monitor the 2013 data to see if these declines persist, and will update this analysis.
As noted above, comparisons of changes in median home values between two or more towns can also be significantly distorted unless the towns have a similar number of homes in each price segment. For example, if one town has a more concentrated distribution of housing values around the center of its range (e.g., a bell shaped curve), then its median selling price will change less in response to any change in the mix of housing sales than another town with a perfectly flat distribution of housing values.[4] In other words, if there is a large enough homogenous "core" at the center of the housing stock, variations in the number of "transactions" at either the high end or the low end will not have as much influence on the median value, which will tend to remain within that relatively narrow core range. However, this would not prevent the median value from reflecting the full extent of any rise or fall in home prices.
We will also be tracking the number of transactions, which last year, as you can see in the table below, were still 30% below their 2007 level in Fairfield, versus only 3%-12% lower in Darien, Greenwich, New Canaan, Westport and Wilton.
|
Constant Quality Single Family Home Prices by Town in Connecticut (Uconn School of Business)
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Number of Transactions
|
|
2007 Sales Prices
|
|
2012 Sales Prices
|
2007-2012
|
| |
2007
|
2012
|
Change
|
|
Low
|
Medium
|
High
|
Average
|
|
Low
|
Medium
|
High
|
Average
|
Change
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Bridgeport
|
137
|
349
|
155%
|
|
209.7
|
235.3
|
295.9
|
247.0
|
|
127.2
|
163.2
|
184.5
|
158.3
|
(35.9)%
|
|
Darien
|
306
|
270
|
(12)%
|
|
841.1
|
1,116.6
|
1,351.0
|
1,102.9
|
|
613.3
|
840.5
|
1,046.7
|
833.5
|
(24.4)%
|
|
Easton
|
na
|
na
|
na
|
|
na
|
na
|
na
|
na
|
|
na
|
na
|
na
|
na
|
na
|
|
Fairfield
|
744
|
521
|
(30)%
|
|
441.5
|
550.5
|
669.6
|
553.9
|
|
278.0
|
385.2
|
422.2
|
361.8
|
(34.7)%
|
|
Greenwich
|
428
|
402
|
(6)%
|
|
870.4
|
1,120.5
|
1,568.9
|
1,186.6
|
|
654.5
|
895.9
|
1,279.5
|
943.3
|
(20.5)%
|
|
Guilford
|
262
|
207
|
(21)%
|
|
262.0
|
370.6
|
658.0
|
430.2
|
|
211.8
|
284.4
|
487.8
|
328.0
|
(23.8)%
|
|
Madison
|
222
|
202
|
(9)%
|
|
282.9
|
437.9
|
759.8
|
493.5
|
|
225.7
|
336.5
|
555.4
|
372.5
|
(24.5)%
|
|
Milford
|
531
|
388
|
(27)%
|
|
289.3
|
341.2
|
430.5
|
353.7
|
|
191.6
|
238.4
|
310.3
|
246.8
|
(30.2)%
|
|
New Canaan
|
233
|
225
|
(3)%
|
|
1,029.1
|
1,436.9
|
1,607.1
|
1,357.7
|
|
728.4
|
1,035.2
|
1,144.9
|
969.5
|
(28.6)%
|
|
Norwalk
|
621
|
423
|
(32)%
|
|
326.9
|
488.9
|
808.2
|
541.3
|
|
240.4
|
364.4
|
602.9
|
402.5
|
(25.6)%
|
|
Ridgefield
|
na
|
na
|
na
|
|
na
|
na
|
na
|
na
|
|
na
|
na
|
na
|
na
|
na
|
|
Shelton
|
293
|
221
|
(25)%
|
|
330.3
|
441.8
|
590.0
|
454.0
|
|
249.3
|
330.5
|
418.5
|
332.8
|
(26.7)%
|
|
Stamford
|
674
|
460
|
(32)%
|
|
511.4
|
649.7
|
810.8
|
657.3
|
|
385.5
|
509.5
|
667.3
|
520.8
|
(20.8)%
|
|
Stratford
|
475
|
402
|
(15)%
|
|
189.1
|
263.3
|
404.8
|
285.7
|
|
121.1
|
167.1
|
250.3
|
179.5
|
(37.2)%
|
|
Trumbull
|
268
|
325
|
21%
|
|
388.1
|
457.7
|
582.0
|
475.9
|
|
225.1
|
279.1
|
372.9
|
292.4
|
(38.5)%
|
|
Weston
|
154
|
121
|
(21)%
|
|
568.2
|
747.0
|
987.3
|
767.5
|
|
432.6
|
553.0
|
735.7
|
573.8
|
(25.2)%
|
|
Westport
|
365
|
322
|
(12)%
|
|
691.6
|
1,220.9
|
2,029.7
|
1,314.1
|
|
491.1
|
899.2
|
1,544.8
|
978.4
|
(25.5)%
|
|
Wilton
|
204
|
197
|
(3)%
|
|
534.2
|
851.7
|
1,339.0
|
908.3
|
|
393.6
|
581.7
|
1,039.8
|
671.7
|
(26.0)%
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Source: http://www.business.uconn.edu/cms/p1175
|
|
|
|
|
|
|
|
|
|
[1] We would like to thank Neal Fink for referring us to the UConn Business School database.
[2] http://www.thewarrengroup.com/
[3] http://realtytimes.com/rtpages/20100302_medianprice.htm
[4] Think of one town in which there is a house, and just one house, at each and every price point in $100,000 increments from $100,000 to $2 million. Any change in the mix of sales in this town is going to move the median sales price up or down significantly. Now think of another town with exactly the same number of houses, but with 50% of them at a single price point, say $500,000. In Town #2, changes in the sales mix are not likely to change the median sales price from $500,000.