Term Paper – 2015 ECO 5520/5930/6520/Honors
It’s early, but … Several have asked. Deliverable is due Monday, March 30. Paper is due Wednesday, April 22.
3 Housing – Why is it Different? Why? –Housing is heterogeneous –Housing is immobile –Housing is durable –Housing is expensive –Moving costs are high –Neighborhood comes with housing … and it matters!
4 Heterogeneous? Dwellings differ in: –house size (sq. feet) –lot size (sq. feet) –configuration –quality People seem to value these qualities differently.
5 Immobile? It is where it is. Where you buy it, you get: –Accessibility (to good and bad things) –Package of local public services –Environmental quality Further –You can’t (really) “move” houses –You can’t rebundle them (use half of two different houses at the same time).
6 Price: The Hedonic Approach Hedonic approach looks at house as a bundle of components. Analogy: Suppose that when you went to the grocery store, all you could buy were “filled” shopping carts (food, soaps, etc.), and each one had a price. You know what’s in them, but you can’t take things out or put things in.
7 Price: The Hedonic Approach How do you figure out what the individual components are worth? A> If you had a large sample of carts, and each had different amounts of goods in them, then you could come up with the value of the individual components.
8 Example for Hedonic Prices Suppose that sq. feet of living space was ALL that mattered in the price of house. You collect data on lots of houses. Sq. feet Price
9 Example for Hedonic Prices What does this suggest? –A> Bigger houses have more value. Let’s draw a line. Sq. feet Price ? ?
10 Example for Hedonic Prices Line has a form: Price = a + b*size Sq. feet Price What does a mean? What does b mean? a slope = b
11 Example for Hedonic Prices Says that for each additional sq. ft., house price is $b more. Sq. feet Price Although it is hard to think of, we could draw this diagram in n dimensions! a slope = b b is the hedonic price of house size.
12 n dimensions? Let’s look at a house with 2000 sq.ft., 5 rooms for $75,000 Price Sq. feet Let’s look at a house with 3000 sq.ft., 6 rooms for $100, Rooms Line has a form: Price = a + b*size + c*rooms
Your Task Use this analysis to say something about taxes and what they buy. We have over 125,000 housing transactions for Ohio for They have –House variables –Neighborhood variables –Tax variables
Hypotheses Let’s look at the value of a house. All else equal, what do you think will be the impact of better services? All else equal, what do you think will be the impact of higher taxes?
Here is an example !!! Dependent variable is log of the transactions price When we write Log P = b 0 + b 1 X 1 + b 2 X 2 + e The coefficient of X is the percentage change in P brought about by a one unit change in X.
Here is an example Parameter Estimates VariableDFParameter Estimate Standard Error t ValuePr > |t| Intercept lotsize Lot size in square feet bedrooms Number of bedrooms brick 1 if brick; 0 otherwise fullbath How many full bathrooms agehouse Age of house in years buildingsqft Building square feet effmills_sd Effective tax rate in mills – What’s a mill? cprate_sd Percent in college prep track Cincinn 1 if Cincinnati metro; 0 otherwise Cleveland 1 if Cleveland metro; 0 otherwise Columbus 1 if Columbus metro; 0 otherwise Dayton 1 if Dayton metro; 0 otherwise Toledo 1 if Toledo metro; 0 otherwise Ytown 1 if Youngstown metro; 0 otherwise One mill is equivalent to one- tenth of a cent or $0.001.
Here is an example Parameter Estimates VariableDFParameter Estimate Standard Error t ValuePr > |t| Intercept <.0001 lotsize E E <.0001 bedrooms <.0001 brick <.0001 fullbath <.0001 agehouse <.0001 buildingsqft <.0001 effmills_sd <.0001 cprate_sd <.0001 Cincinn <.0001 Cleveland <.0001 Columbus <.0001 Dayton <.0001 Toledo <.0001 Ytown <.0001
Here is an example Parameter Estimates VariableDFParameter Estimate Standard Error t ValuePr > |t| Intercept <.0001 lotsize E E <.0001 bedrooms <.0001 brick <.0001 fullbath <.0001 agehouse <.0001 buildingsqft <.0001 effmills_sd <.0001 cprate_sd <.0001 Cincinn <.0001 Cleveland <.0001 Columbus <.0001 Dayton <.0001 Toledo <.0001 Ytown <.0001
How well does it fit? Number of Observations Read Number of Observations Used Number of Observations with Missing Values 7224 Analysis of Variance SourceDF Sum of Squares Mean Square F ValuePr > F Model <.0001 Error Corrected Total Root MSE Dependent Mean R-Square Adj R-Sq Coeff Var
But … I’m sure you can do better than that.
Topic: Data-based Finance Analysis I am creating a file for your use. You will create hypotheses and test them.
Topic: Data-based Health Analysis Student will form null hypotheses. Test them using appropriate data analysis. Write up findings in a format to be provided by professor.
Specifics ECO 5520 – No less than 10 pages; no less than 5 references from scholarly journals. ECO 5930/6520 – No less than 15 pages; no less than 10 references from scholarly journals. –Wikipedia is not a scholarly journal. –Fisher is not a scholarly journal.