Filtering Paper 230528.pdf

Using data on all buildings & their occupants from 1990-2017 in Sweden, the authors find that new buildings trigger moving chains which quickly benefit even the lowest income quartile. This is true for all building types and even for buildings which attract the highest-income residents. These moving chains improve the quality of buildings (measured by both age and floorspace), with poor households especially benefitting from larger units. Authors conclude: “Thus, it is not important to build homes directly affordable for low-income residents, they will reap the benefits of more housing space anyway through ripple effects.”

  1. Background/Motivation: Debates about whether new construction only benefits the rich, or whether it creates ‘moving chains’ which quickly free-up housing for poorer households. Rosenthal (2014) found that an in-mover to a 50y/o home will have income 60% lower than the first occupant.
  2. Model/Data: GeoSweden database covers the building address of every Swedish resident from 1990-2017. Didn’t thoroughly read the model, but they’re basically trying to map the filtering function through successive rounds. Round 0 = person who moves into the new building, Round 1 = person who moves in to that person’s old address, etc.
  3. Key Threats/Challenges: Key issue is they can only observe addresses at the building level, so they’re not mapping the exact person who moved into the exact unit, only the type of people moving into buildings that are vacated. Not a critical flaw. Likewise, addresses are only observed annually, so they can’t quite observe the immediate re-shuffling.
  4. Results: Individuals from the poorest quartile make up the majority of in-movers after the first round of moves. This rapid income decay occurs across all building types (detached homes, coop, rentals), although rental buildings filter faster than owned apartments, which filter faster than detached homes. “people with considerably less than the national mean income benefit from moving chains in later rounds no matter whether they access new homes directly” 30y/o buildings have a proportional mix of residents from different income groups. Mean income falls by 15% in 50 years, which is slow, but this largely reflects Sweden’s rent regulation and housing allowances for the poor. Municipalities with higher construction rates produce improvements in quality (floorspace and building age) for all income groups. Poor people get the most benefit in terms of more floorspace
  5. Discussion/Conclusions: “new homes of all types lead to more housing space for poor people compared to rich people. Thus, it is not important to build homes directly affordable for low-income residents, they will reap the benefits of more housing space anyway through ripple effects.” “building new homes, even expensive ones, leads to important trickle-down effects; thus, stimulating the supply of new homes is a viable approach to improving the housing situation of the poor”