Slide 10 walsh and co authors estimated this equation

This preview shows page 5 - 7 out of 8 pages.

Slide 10 Walsh and co-authors estimated this equation in their analysis. The log of price is the dependent variable and it is a function of things we know are important for housing prices: housing characteristics (the vector H), and location and neighborhood characteristics (e.g., proximity to major roads, L). The equation also includes a year-specific variable to pick up any fluctuations in the housing market. The main independent variable we are intested in is clarity – or more precisely K- D, the inverse of clarity – which is highlighted in blue here. Notice that, first, it is in natural log form so the relationship between housing price and water quality is specified to log-log. This means we can interpret the coefficient, gamma, as an elasticity. It will be the percent change in housing price for 1% change in water
quality. Second, notice that log water quality and its gamma-coefficient are interacted with the distance dummy variable of Di. This is set to one as appropriate for the given buffer of “water front,” 0 to 500 meters, and 500 to 1000 meters to match the home. The result is three separate estimates of gamma, the water quality coefficient, one for each set of houses. Again, we’d generally expect the value of gamma to diminish (in absolute terms) as we get farther and farther away from the waterfront. Slide 11 Here is a representative set of water quality results from the regression. Please feel free to see the study for the full set of variables and how they related to housing sales price. And also recall that these are just three of the fourteen counties estimated in the study. A priori, we expect the coefficients (gammas) on water quality to be negative. This is because the measure used for water quality was K-D, the inverse of water clarity. A positive gamma suggests that housing prices decrease as gamma decreases and the water gets clearer. This is unexpected if people prefer clearer water, all else equal. In a large dataset like this, we are likely to see some of these unexpected outcomes as for Calvert county homes 500 to 1000 meters away from the water where home prices seem to rise with murkier water. This result is an outlier. Let’s look at Anne Arundel county shows a more typical result. Here we see the expected sign on our coefficient estimate and see that the effect is diminishing as distance from the water increases. In fact, in the 500 to 1000 meter buffer the effect is statistically insignificant. We interpret these results as elasticities. So a 10% increase in water quality would result in a 1.26% increase in value for Anne Arundel waterfront homes. Slide 12 Let look at how these results are applied to estimate the benefits of a change in water clarity. Walsh and co-authors calculated the benefits to the average home for a -0.1 change in measured K-D; or, loosely speaking, a 0.1 increase in measured clarity. The “0.1” is a change in the K-D index and was a number chosen by the authors for illustration, but which is similar to changes that are

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture