Port St. Lucie Housing Market Project - Research Paper Example

Lecturer’s and Number Submitted Port St. Lucie Housing Market Project Introduction Over the years, Florida has felt the effects of the national mortgage troubles and housing slowdown just like other places in the country. The ever presence of constructions has changed, home sales dropped as the job losses in the housing related industries have worsened. Meanwhile, the residents of Florida have been hit by some rise in property taxes, especially after being hit by the natural calamities like hurricanes in 2004 and 2005, coming in the increased insurance rates. In the analysis of the issues affecting the people living in this fast-growing Central Florida city, it was found that among all the issues affecting the state, none was as important to most voters as the real estate crisis. Almost all said that slumping property values and rising taxes and insurance costs were at the top of their concerns. Florida was hit significantly hard by the crash in the housing market. It is the belief of some Florida politicians that real estate bubble was caused by loan officers pressuring real estate appraisers to over-value homes. They cite testimonials from real estate appraisers being pressured by loan officers to hit certain price targets on their appraisals or they would not be hired for future appraisals. In order to reform the real estate market in Florida, the state is considering legislation to require a purely quantitative approach to real estate appraisals. That is, they would like to construct a mathematical ‘formula’ that would take the characteristics of the property and provide an estimate of the value of the property. Specifically, the formula would provide a price estimate and a likely range around the estimate (i.e. a 95% confidence interval). As a consulting firm, given the responsibility of testing the use of this approach as, this paper purpose to construct a formula that would predict sale prices of properties in the county of Port St. Lucie. Port St. Lucie County is located on the east coast of Florida just north of West Palm Beach. Most of the residential properties in the county are located within a few miles of the Atlantic Ocean. Methodology The research design used in this study was descriptive survey whereby the study aimed at collecting information from respondents on their attitudes and opinions in relation to the sale prices of properties in the country of Port St. Lucie located at the east coast of Florida. By assuming that as the case study the study utilizes both qualitative and quantitative research techniques into analyze the topic of study using statistical techniques and tools like Stata to construct variables, test hypotheses, estimate models, perform inference, and test for the robustness of the models. Using raw datasets (one that may contain errors) on residential property sales, clean the dataset, construct analytical variables, test hypotheses – bivariate and multivariate, this study checked for the robustness of the models, in explain the results. The study used the over 24,000 records from the Multiple Listing Service (MLS) from 2001 to 2005 to determine this mathematical formula. The data contains all listings of single-family homes, condominiums, and townhouses that were sold, cancelled, expired, or are pending sale in Port St. Lucie County. This data was used to construct analytical variables, the analytical variables being used to construct linear regression models to determine residential property values, and to check the robustness of the models for the report. Analysis Descriptive statistics The dependent variable that the study aimed at explaining was the Sold price of the houses (use for market value of house). The variations of the sale prices of properties were found to be caused by a number of factors. These factors include number of bedrooms, number of full baths (i.e. sink, toilets, bath/shower), number of half baths (sink and toilet only), garage spaces, living square feet, and the total square feet. All these were expected to explain the dependent variable, the market value of houses in Florida. The bivariate relationship between the variables Reg sold price bedrooms full baths, half baths, garage spaces, living sqft , totals qft Source | SS df MS Number of obs = 19610 -------------+------------------------------ F (6, 19603) = 1035.95 Model | 1.2808e+14 6 2.1347e+13 Prob > F = 0.0000 Residual | 4.0394e+14 19603 2.0606e+10 R-squared = 0.2407 -------------+------------------------------ Adj R-squared = 0.2405 Total | 5.3202e+14 19609 2.7131e+10 Root MSE = 1.4e+05 According to the Stata output above, the R-Squared and the Adjusted R-Squared are 0.2407 and 0.2405 respectively suggesting that 24.05% of the market value of houses in Florida (dependent variable) are explained by the number of bedrooms, number of full baths, half baths, garage spaces, living square ft, and total square ft (independent variables). With the number of the 19610 observed cases in the data, 24.05% is a little weak under the circumstance, but used for statistical references. In other output where collinearity (where a multiple regression have a non-zero correlation), the following result was obtained. . reg sold price bedrooms fullbaths, halfbaths, garage spaces, livingsqft , totalsqft CSQFT SSQFT TSQFT POOL note: CSQFT omitted because of collinearity Source | SS df MS Number of obs = 19610 -------------+------------------------------ F( 9, 19600) = 731.57 Model | 1.3378e+14 9 1.4864e+13 Prob > F = 0.0000 Residual | 3.9824e+14 19600 2.0318e+10 R-squared = 0.2515 -------------+------------------------------ Adj R-squared = 0.2511 Total | 5.3202e+14 19609 2.7131e+10 Root MSE = 1.4e+05 In this case, from the 19610 observations, the R-Squared, and the Adjusted R-Squared are 0.2515 and 0.2511 respectively, implying that 25.11% of the dependent variable is explained by the dependent variables in the study. Regression model The model below was used to analyze this relationship: Y= α+ βi X1+ βi X2+ βi X3+ βi X4+ ε Whereby Y = sold price α= Constant term βi = Beta Coefficient X1= bedrooms X2= full baths X3= half baths X4= garage spaces X5= living sqft X6= total sqft ε= Error term Consider the output below: ------------------------------------------------------------------------------ Sold price | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- bedrooms | -9226.705 1737.654 -5.31 0.000 -12632.65 -5820.756 fullbaths | 58076.74 2979.906 19.49 0.000 52235.87 63917.61 halfbaths | 53419.46 3415.05 15.64 0.000 46725.67 60113.24 garagespaces | 9734.17 1371.886 7.10 0.000 7045.158 12423.18 livingsqft | 95.57525 2.438534 39.19 0.000 90.79551 100.355 totalsqft | .5390799 .2572997 2.10 0.036 .0347506 1.043409 _cons | -91374.27 5394.029 -16.94 0.000 -101947 -80801.52 ------------------------------------------------------------------------------ At the 95% confidence interval, the coefficient table above show the relationship between the sold price and the factors affecting the sold price of houses in Florida. This implies that the regression model becomes as stated below: Sold Price in Florida= -91374.27 -9226.705 bedroom+ 58076.74 full baths +53419.46 halfbaths+9734.17 garage spaces +95.57525 living square ft + 0.5390799 total square ft This implies that all other factors held constant, the number of full baths contribute the highest as far as sold prices of houses in Florida is concerned. Meaning that, a 1% increase in the number of full baths increases the sold price of houses in Florida by 58076.74 times. It also statistically shows that a 1% increase in half baths contribute to an increase of 53419.46 in the Sold price while a 1% increase in garage spaces contribute to 9734.17 of the sold price. 1% increases in living square ft and total square ft contribute increments of 95.57525 and 0.5390799 of the Sold Price of houses in Florida respectively. However, the number of bedrooms negatively contributes to the sold prices, since 1%, increase in the number of bedrooms decreases the Sold Price by 9226.705. The constant value of the Sold Price at any case as stipulated above is -91374.27. Conclusion While examining all the relevant factors amounting to the significant crash on Florida housing market, it is also important to look at the importance of the economic relevance of the factors affecting housing in every state. The fact is that the number of bedrooms, full baths, half baths, garage spaces, Total Square, and the living squares all have varied impacts to the market value of the houses in Florida. The concerned parties and the relevant policy makers in Florida, including politicians, should adopt the above model (formula) for pure quantitative approach to real estate appraisals for their supposed reforms. As it is the belief of some Florida politicians that real estate bubble was caused by loan officers pressuring real estate appraisers to over-value homes, they should then have a proper statistical evidence to table their concerns. Appendix DATA CLEANING Dummy variable for relevant zipcodes gen zipcodePSTL=. replace zipcodePSTL==1 if zipcode==34952 | zipcode==34953 | zipcode==34983 | zipcode==34984 | zipcode==34985 | zipcode==34986 | zipcode==34987 | zipcode==34988 replace zipcodePSTL=1 if zipcode==34952 | zipcode==34953 | zipcode==34983 | zipcode==34984 | zipcode==34985 | zipcode==34986 | zipcode==34987 | zipcode==34988 Dummy Variable for views tab view, gen(dview) Data cleaning. Making a new dummy variable for city where =1 is all variations of Pt St Lucie gen DCITY=. replace DCITY=1 if city=="PORT SAINT LUCIE" | city=="PORT ST LUCI" | city=="PORT ST LUCIE" | city=="PTSTLUCE" | city=="SAINT LUCIE" | city=="SAINT LUCIE WEST" | city=="STLUCIEW" Data cleaning. Making a new variable for pool that can be used in the regression gen POOL=. replace POOL=0 if pool=="N" replace POOL=1 if pool=="Y" Data cleaning making a dummy variable for if sqft is lower than 650 gen SQFTDUMMY=. replace SQFTDUMMY=0 if totalsqft=1000 Dummy variable for if living sqft is larger than total sqft gen SQFTREAL=. replace SQFTREAL=0 if totalsqftlivingsqft GENERATING NEW VARIABLES FOR VIEW generate golf = regexm(view, "A") generate ocean = regexm(view, "B") generate intercoastal = regexm(view, "C") generate river = regexm(view, "D") generate lake = regexm(view, "G") INTERACTIONS Creating interaction terms for type of housing and living sqft Creating dummy variables for the types of housing tab type, gen(DTYPE) Generating interaction between Condo and sqft gen CSQFT= DTYPE1*livingsqft Generating interaction between Single family home and sqft gen SSQFT= DTYPE2*livingsqft Generating interaction term between Townhome and sqft gen TSQFT= DTYPE3*livingsqft Generating interaction terms for views/sqft generate oceanlivingsqft = ocean*livingsqft generate golflivingsqft = golf*livingsqft generate intercoastallivingsqft =intercoastal*livingsqft generate riverlivingsqft =river*livingsqft generate lakelivingsqft =lake*livingsqft END INTERACTIONS REGRESSIONS Regression with the basic variables. All vars are statistically significant reg soldprice bedrooms fullbaths halfbaths garagespaces livingsqft totalsqft Regression with type/sqft interaction and fixed pool var. All vars are stat sig reg soldprice bedrooms fullbaths halfbaths garagespaces livingsqft totalsqft CSQFT SSQFT TSQFT POOL Read More