We compared annual changes in selling price between all non-subsidized units sold from 1992 through 1996 in the 14 selected subdivisions and all sales in the counties in which they were located. Median annual price changes in dollars, whether upward or downward, were determined for four time periods: 1992-93, 1993-94, 1994-95, and 1995-96. As noted earlier, the fact that most subdivisions with subsidized units had a mix of detached homes and townhouses made price levels generally lower in such subdivisions than elsewhere in their counties, where single detached homes strongly predominate. Therefore price changes rather than absolute prices were compared.

We also show percent changes here rather than changes in dollars. This has been done to minimize the possibility of confusion resulting from differing overall price levels between the selected subdivisions and their counties. However, using dollar values would have made no difference in our conclusions. Readers wishing to confirm this are referred to the appendix tables, where both dollar and percent changes are shown along with actual median prices in dollars.

Overall, housing prices in both Montgomery and Fairfax Counties were up slightly in the first two of the four time periods, 1992-93 and 1993-94, down sharply in 1994-95, and down slightly in 1995-96.

Table 1 shows these changes, and also indicates whether or not the price trend for units sold in the subdivisions in this study was better than or not as good as that for the county as a whole. "Better" was defined as:

• if prices increased in both cases, those in the tested subdivision increased more;
• if prices decreased in both cases, those in the tested subdivision decreased less;
• if one increased and the other decreased, the prices in the tested subdivision were the ones that increased.

In Montgomery County, Maryland, price behavior in the eight subdivisions with subsidized units was better than in the county as a whole in two of the four time periods, 1994-95 and 1995-96. Prices increased for the tested subdivisions in both periods, while they decreased for the county as a whole. It was not as good in the other two, 1992-93 and 1993-94. In both of these periods, prices increased in both the tested subdivisions and the entire county, but rose to a lesser degree in the former.

In Fairfax County, Virginia, price behavior in the six tested subdivisions with subsidized units was also better in two time periods -- in this case, 1992-93 and 1994-95. In 1992-93, prices rose both in the tested subdivisions and in the county as a whole, but the price increase for the tested subdivisions was greater. In 1994-95, prices rose in the tested subdivisions while declining in the County as a whole.

In the other two time periods, the entire county did better than the six subdivisions. In 1993-94 prices rose in the county while dropping in the subdivisions, while in 1995-96 they declined in both cases but to a greater degree in the subdivisions.

Thus, in both Fairfax and Montgomery Counties there was no overall difference in the price behavior of market rate homes in the subdivisions with subsidized housing vs. the county as a whole.

Thus, in both Fairfax and Montgomery Counties there was no overall difference in the price behavior of market rate homes in the subdivisions with subsidized housing vs. the county as a whole.

For the same reason, we again show percent changes here rather than changes in dollars.This has been done to minimize the possibility of confusion resulting from differing overall price levels between the nearer and farther dwellings. Here, also as can be seen from the appendix tables.

In Fairfax County, VA, the units closer to the subsidized housing came out better in two of the four periods, 1993-94 and 1995-96. In 1993-94 prices declined for both categories, but the closer homes dropped by less. In 1995-96, the closer homes increased in price, while those at greater distances decreased. In the other two periods, the houses farther away did better. In 1992-93 both nearer and farther units rose in price, but the increase was greater for the latter. In 1994-95 the closer homes decreased in price while those more distant gained.

In Montgomery County, MD, the price change for the closer houses also was better in two time periods, 1992-93 and 1994-95. In 1992-93 prices declined in both the nearer and farther homes, but the decline was smaller for those closer to the subsidized housing. In 1994-95 prices rose in the closer units while those farther away dropped. In both the remaining two periods, 1993-94 and 1995-96, the closer homes did not do as well. Prices rose for those at both nearer and farther distances, but the increase was smaller for the nearer dwellings.

When data for the two counties were combined, once more the closer units came out better two out of four times, and not as well the other two. In 1992-93 both gained, but those closer gained more. In 1994-95 the closer homes gained while those farther away declined. On the other hand, in both 1993-94 and 1995-96 the closer units lost while the more distant ones gained.

In sum, while the homes farther from the subsidized housing sometimes came out better and sometimes not as well as those closer, there were an equal number of cases where the closer homes were ahead. Put still more succinctly, a tie.

C.  DIFFERENCE IN PRICE BEHAVIOR BETWEEN HOUSES ADJACENT TO SUBSIDIZED HOUSING AND ALL

We looked at changes in median sales prices for homes adjacent to subsidized units vs. all other dwellings in the same subdivision. Table 3 shows the results.

In Montgomery County, MD, the price trends for the adjacent units were better than for all other sales in the subdivision in two time periods, 1993-94 and 1994-95. In both instances, the adjacent houses gained while all others as a group lost. The trends for the adjacent dwellings were less favorable in the remaining two periods. In 1992-93 both lost but the adjacent units lost more. In 1995-96 the adjacent homes lost while the rest gained.

In Fairfax County, VA, there were also two time periods in which the price changes were better for the homes adjacent to subsidized units: 1994-95 and 1995-96. In the former, both gained but the adjacent dwellings gained more. In the latter, both lost but the adjacent houses lost less than the remainder. In the other two periods, the adjacent homes did less well. In 1992-93 both gained but the adjacent units rose less. In 1993-94 both lost but the adjacent homes lost more.

For both counties combined, the results were similar. In 1993-94 and 1994-95, both the adjacent dwellings and the remaining units sold rose in price. However, those adjacent to the subsidized housing rose more. On the other hand, in 1992-93 both gained, but the adjacent units gained less. And in 1995-96 both lost, but those adjacent lost more.

We also looked at price changes within individual subdivisions as compared to the zip codes in which they were located. Table 4 shows the results for Fairfax County, VA and Table 5 shows them for Montgomery County, MD. In Fairfax County, VA, six subdivisions were analyzed, while in Montgomery County, MD there were eight, for a total of 14 subdivisions in both counties combined. Descriptions of the 14 communities are in Appendix 4.

1. Subdivisions in Fairfax County, VA (Table 4)

In Fairfax County, VA, four of the six subdivisions had data for all four time periods. They were Burwell (The Park), Dunn Loring, Village Park, also known as Chatham Towne (La Cross), Summerhill (Sheffield), and Village Park. In the other two, Waters Edge (Green Duck) and Misty Woods, there were enough resales to calculate a change reliably in only one of the four time periods.

In Waters Edge (Green Duck) that time period was 1995-96, and in that period it did not do as well as its zip code; median sales prices in the subdivision declined while those for the zip code as a whole increased. In Misty Woods, the period for which adequate data were available was 1992-93. In that period, prices dropped in both the subdivision and the zip code, but by more in the subdivision.

Among the remaining four subdivisions, all had enough sales in each of the four time periods to compute a price change. In one of these, the subdivision excelled in all four time periods. In two more, there were ties -- the subdivision excelling in two periods, the zip code in the remaining two. In the fourth subdivision, the zip code was ahead in three time periods and the subdivision in only one. Taking all four time periods for all four of these subdivisions together, a total of 16 cases, the subdivision was ahead in nine and the zip code in seven.

In Burwell (The Park), the first of the four subdivisions with complete data for four time periods, prices in the subdivision behaved better in all four. In two, 1992-93 and 1995-96, prices in the subdivision rose while those in the zip code declined. In a third, 1993-94, they rose in both, but the subdivision gained more than the zip code. In the fourth, 1994-95, there was no change in median sales price for Burwell but a decrease for its zip code.

Summerhill (Sheffield), the third, gained in two and lost in two. The subdivision did better in 1992-93 and 1995-96. In the former, it gained while the zip code lost; and in the latter, both lost but Summerhill lost less. The two periods in which Summerhill did not do as well as its zip code were 1993-94 and 1994-95. In the first of these, both lost but Summerhill lost more; and in the second, the subdivision lost while the zip code gained.

Village Park (La Cross), the fourth subdivision, did better than its zip code in only one year, 1994-95, when it gained while the zip code lost. In 1992-93, Village Park gained but the zip code gained more. In both 1993-94 and 1995-96, Village Park lost while the zip code gained.

Taking all six Fairfax County subdivisions together, there were a total of 18 time periods where enough data on price changes were available for analysis. In nine of these the subdivision was ahead, while in the other nine the zip code did better. Another tie.

2. Subdivisions in Montgomery County, MD (Table 5)

All of the eight subdivisions in Montgomery County, MD had adequate numbers of sales to permit computation of a price change for all four time periods. In none of these eight did the price change for either the subdivision or the zip code excel in all four periods.

There were five cases where the two were tied: the subdivision itself excelled in two time periods and the zip code in the other two. There were two cases where the subdivision was ahead 3-1, and one in which the zip code led by the same margin. Looking at all eight subdivisions for all four time periods, 32 cases in all, the subdivision was ahead in 17 (53 percent) and the zip code in 15 (47 percent). This was the only comparison we made in which we found anything but a tie, and here the difference was small and not sufficient to conclude the presence of assisted units positively affected adjacent sales prices, hence, we do not think it contradicts the conclusion that there is essentially no difference.

Taking the eight subdivisions one by one:

Fallsberry did better than its zip code in one time period, 1994-95, when both lost but Fallsberry lost less. As for the remainder, there were two periods when both lost but Fallsberry lost more, 1992-93 and 1995-96. There was one, 1993-94, in which both gained but the zip code gained more.

Flower Hill (Oakwood) bested its zip code in three of the four periods. In the first of these, 1993-94, both gained but Flower Hill gained more. In the second, 1994-95, Flower Hill gained while the zip code lost. In the third, 1995-1996, there was no change in price for Flower Hill but the zip code lost. In the only period when it did not do as well as its zip code, 1992-93, both lost but Flower Hill lost more.

Hallowell did better in two time periods and its zip code in the other two. In 1992-93 both lost but Hallowell lost less. In 1994-95 both gained and Hallowell gained more. In both of the losing periods, 1993-94 and 1995-96, the zip code gained while Hallowell lost.

Oak Spring did better than its zip code two out of four times. In both 1992-93 and 1994-95, it gained while the zip code lost. In 1995-96, however, both lost but Oak Spring lost more. And in 1993-94, the zip code gained while Oak Spring lost.

Saddlecreek, like Flower Hill, beat out its zip code three to one. In both 1992-93 and 1995-96, it gained while the zip code lost. In 1993-94, both lost but Saddlecreek lost less. But in 1994-95 the zip code was ahead, although both gained.

Stonebridge was ahead in two of the four time periods, and not as good as its zip code in the other two. In both 1993-94 and 1995-96, Stonebridge gained while the zip code lost. In 1992-93 and 1994-95, however, the zip code gained while Stonebridge lost.

Stratford Knolls also excelled in two time periods while yielding to its zip code in two others. In 1995-96 it gained while its zip code lost, while in 1994-95 both lost but Stratford Knolls lost less. On the other hand, in both 1992-93 and 1993-94 Stratford Knolls lost while the zip code in which it is located gained.

Timberlawn, (Pinehaven and Misthaven Terrace), like Stratford Knolls, bested its zip code in 1995-96 by gaining while the zip code lost. Again, both subdivision and zip code lost in 1994-95, but the subdivision beat the zip code by losing less. Timberlawn, again like Stratford Knolls, was behind its zip code in 1993-1994 by losing while the zip code gained. But the similarity ends there. In 1992-93 Timberlawn and its zip code both rose but the subdivision rose less. Stratford Knolls, in the same year, lost while its zip code gained.

When we examined data comparing overall price changes by time period for subdivisions containing subsidized dwellings with the entire counties in which they were located, we found no difference. The same thing proved true when we compared price trends for market rate homes located within 500 feet of an subsidized unit with those at greater distances, and also when we looked at the trends for market rate homes adjacent to subsidized units vs all the rest of the unsubsidized units in the subdivision.

When we examined the price trends for individual subdivisions with their zip codes, however, we found that while 2-2 ties occurred most commonly, trends in the majority of subdivisions nonetheless favored one or the other. When we added up all the occurrences, the difference disappeared in Fairfax County, and there was again a tie. But in Montgomery County (and hence in both counties combined), the subdivisions ended up slightly ahead. Why did this occur, and does it affect the overall conclusion that there is no difference?

We hypothesize that the smaller numbers of sales in the individual subdivision allowed random price variations to exert more influence. There were 1,012 sales in all, but when broken down by subdivisions and time periods there were an average of slightly more than 20 sales for each of the 50 combinations (four time periods for each of 12 subdivisions, plus one time period for each of the remaining two). With this few cases, it was almost inevitable that random price fluctuations would come into play in individual instances. Nonetheless, taking the two counties together, there were 26 cases where the subdivision came out ahead vs 24 where the zip code excelled -- 52 percent vs 48 percent, an insignificant difference.