committed in the three towns from 1993 to 1998. In fact, the abrupt jump in number of crimes from the year 1992 to 1993 is mainly due to the addition of Part II crimes described above. Thus, crime estimates as given by the Department of Public Safety for these years are biased downwards. In fact, the total number of 'out of casino' crimes in Ledyard declined from 535 in 1993 to 364 in 1998. 'In casino' crimes also show a decline from 1,212 in 1994 to 989 in 1998 with 60 percent of them being larceny. Table 4 summarizes these numbers.

Table 4. Total Crimes in the Town of Ledyard

Data Source: Division of State Police, Crimes & Data Analysis Unit, Department of Public Safety. Note: Prior to 1993 we have only index crime data for Ledyard. From 1993 onwards crimes are separated into 'in casino' and 'out of casino' crimes.

Aggregating over these three Towns, we conclude that 'out of casino' crimes have increased only marginally. Moreover, the MPTN contributes regulatory fees to the Connecticut State Police and Liquor Control Division, which accounted for $4.4 million in the fiscal year 1999. The State Police prosecute crimes on the reservation.

Property Value Analysis

This part of the study analyzes the impact of the Foxwoods Resort Casino and related operations of the Mashantucket Pequot Tribal Nation on surrounding residential property values. Specifically, we analyze the annual growth rate of residential property values (proxied by the property sale price) in Ledyard, North Stonington, and Preston, and, compare this trend with the annual growth rate of residential property in the Hartford LMA.

When performing a study on residential properties, one needs to guarantee that the properties' characteristics do not change during the study period. There are different ways to arrive at a consistent calculation. One is the conventional method in the appraisal profession, that is, extract properties of constant quality that are sold more than once during the study period, and calculate the sales price change for these properties. An alternative method is to conduct an econometric analysis on the property sales prices controlling for the characteristics of the properties, and then use the estimates from the model to calculate the predicted property value growth rate for a specific menu of characteristics. Due to a limitation of the available data on property characteristics in the three towns (see the next section), we choose to use the first method, that is, matching properties that have been sold more than once.

In order to separate the Casino's impact on housing prices from the general trend in the housing market from 1981 to 1999, we separate the study period into two periods, 1981-1989 and 1990-1999. The reason we use 1990 as the break point instead of 1992, the year when the Casino opened its doors, is to take into consideration people's expectation of the Casino's future impact on housing price trends.

The impact of the Foxwoods Resort Casino on the adjacent three towns' property values can be found by comparing the housing price trend in these three towns with the trend in a broader or different geographical area. Due to the irregular behavior of

housing prices in the southwest area of Connecticut, for example, Fairfield County, we use the Hartford LMA as the basis for comparison.

The housing sales price data is extracted from two sources. We obtained housing sales data for 1990-1999, for the three towns from the home sales website of Dow Jones & Company, Inc. Because this data does not provide the characteristics of the transacted properties, we were unable to perform the econometric analysis described above. Instead, we used the matching property method to conduct the analysis. The Center for Real Estate and Urban Studies at the University of Connecticut provided the housing sales data for 1981 through 1989.' The annual growth rate of the constant quality house price for the Hartford LMA was obtained from the Center for Real Estate as well.8

We obtained 683 sales records by matching properties that were sold more than once during 1981-1989 in the three towns. The mean annual growth rate of these house prices is 11.42%, compared to a 9.03% annual growth rate in the Hartford LMA. Note that these growth rates as well as the other growth rates used in this report are based on nominal prices. That is, the sales prices at each date are not adjusted for inflation. Therefore, part of the price increase is due to inflation rather than increased property value. Our conclusion is unaffected despite these nominal growth rates. For the second period, 1990-1999, we obtained 251 matched sales in the three towns adjacent to Foxwoods Resort Casino. These properties' sales price growth rate averages 0.57% annually, compared to a -1.16% annual growth rate for the Hartford LMA during the same time period. Table 5 summarizes the descriptive statistics for the housing sales trend for the three towns and the Hartford LMA during the two periods.

*These price indices are constructed through an econometric analysis by controlling the housing characteristics, and track the value of the standardized house over time. They are not the actual housing sales prices, but the prediction of the sales price if the standardized house is on sale at a certain time. For detailed methodology, see Clapp and Giaccotto (1994).

Table 5 demonstrates that during the first period, 1980-1989, the three towns
enjoyed a much larger increase in property value relative to the Hartford LMA. This is
partly due to the attractiveness of the (near) waterfront properties in these three towns. It
can also be attributed partly to the increased demand for housing as people migrated to
this area to fill the high paid jobs in the defense industry. Table 5 also shows that there is
a sharp decline in the housing price growth rate for both geographical regions. The
average annual housing price growth rate for Hartford LMA plunges from 9.03% in
1980-1989 to -1.16% in 1990-1999. This sharp decline in housing price trend during the
1990s is attributed to the recession in the early 1990s. For the three-town area, the hit
was even harder. About the same time the housing market plunged due to a larger cycle,
the defense industry began contracting its facilities in the New London region. If it were
not for the Foxwoods Resort Casino that started its operation in 1992 and immediately
pumped thousands of new jobs into the surrounding area, the decline in housing prices in
this area would have been more dramatic than in the Hartford LMA. Although there is
also a decline in the growth rate in housing prices in the three towns adjacent to the
Casino relative to the 1980s, the decline is much less dramatic than in the Hartford LMA.
As a result of Foxwoods Resort Casino and related operations, property value growth
rates in the three adjacent towns were actually positive compared to the Hartford LMA.
Instead of losing value over time, the properties in the three adjacent towns have slowly
increased their value. Statistical tests show that the median annual growth rate of
housing prices in the three towns is significantly higher than in the Hartford LMA in both
periods (see Table 6). This result is consistent with the findings in another study on the
Foxwoods Resort Casino in 1993 by Arthur Wright and Associates.

We conclude that the development of the Foxwoods Resort Casino and other MPTN operations in New London County dampened the recession in employment and housing prices in the early 1990s and contributed substantially to the economic rebound of the region through the decade. This included a positive return to housing investment.