NYC-WTC_000187725

NYC 9/11 Public Portal Document

35

where the choice probabilities are given by (8) and where is the share of dwellings that are offered for rental (or sale). As in CATLAS, this share is specified to be an increasing function of rent and given by a binary logit model. Hence,

(10) g for each (j,k) [l+expX(Fj*-djP]

where A. > 0 is the coefficient of rent and the djj^'s are fixed effects to be calibrated as proxies for the differential cost of renting versus keeping vacant housing units of type k in residential zone j. The equalities (9) for each (j,k) are held by the equilibrating action of rents. Solving these equations simultaneously for the JK rent vector of Rjk*’s determines equilibrium rents. To ensure equilibrium in the shopping market, the total number of shopping trips arriving at a shopping zone 1 must be accomodated by a correspondingly intensive use of shopping floor space. Hence, an increase in the demand for shopping in zone 1 should result in a higher shopping rent, rj^, per square foot of shopping space in that zone 1. Letting, Sy(ri^) be the rate of shopping space utilization (an increasing function of r^^) , demand and supply equilibrate as follows in each zone 1:

(11) ■5 ijkjnn V2

Where Fj^ is the quantity of floor space in zone 1. The summation is the total number of shopping trips arriving at shopping zone 1. By specifying the function S|(ry) as p > 0, the equilibrium shopping rents are obtained explicitly as:

i^ln; ijtan] (12) rl= ^ijkian y 1 ^1^1

Once equilibrium is obtained, we can calculate the aggregate consumer surplus of households using the technique discussed in Small and Rosen (1981). In NYREG's case, the consumer surplus will be approximate because the marginal utility of income, ju., is not constant. Calculated at the equilibrium, the weighted average marginal utility of income is as follows for workplace i:

CITY_21MC100_006662266

NYC-WTC_000187766