Logit regression of rehousing

This topic discusses the estimation of the rehousing behaviour and relocation decisions of households in layer 3 of the Ruimtescanner XL

The general framework will be the maximum likelihood estimation, or more specifically the logit regression of household \( i \) moving or not moving to location \( j \)

This study focuses on preferred residential density, and direct neighbourhood characteristics and their interaction with household characteristics as a further refinement to Tigris that models LMS-zonal preferences. It is likely that we'll take such zonal preference per household type (or household \(i\)) as exogenous, which incorporates preferences such as inter-zonal accessibility.

Relevant characteristics related to household \(i\)
 * agegroup of oldest household member ( <65, >=65)
 * number of persons (1, 2, 3+)
 * number of working (0, 1+)
 * household income category (1, 2, 3).
 * original location before relocation (related to RIN i.e. Buurt-id)

Relevant characteristics related to j
 * j=0 is the don't move choice: is this j the don't move choice?
 * overall zonal attractiveness of j (as determined by Pearl).
 * price of j (NVM only, or also WOZ?).
 * waiting time

Location factors, relevant for both the origin of i and j, connected through RIN.
 * structure factors:
 * type: appt, terras-housing, (semi-) detachted, bungalow.
 * density indicator: (see conceptual model: buildingarea / groundarea, floorarea / appt., appt. / building)
 * environment factors:
 * neighbourhood characteristics (typically within 1km range). i.e.:
 * average WOZ < 300m.
 * parks > 1ha,
 * horeca,
 * supermarkt/levenmiddelen,
 * distance to 3rd closest primary school

Relevant characteristics related to the combination of i and j.
 * distance \( i \to j \) as an indicator for transition costs.
 * difference of location factor values between origin \( i \)
 * relative attractiveness of the Tigris-zone for actor \( i \).
 * relative attractiveness of any of the \( j \) characteristics for any of the \( i \) types.

= links =
 * Logit regression.
 * Ruimtescanner XL