Home Price Expectations Flashcards
(17 cards)
why study home price expectations
- High stakes, most important asset to hh
- Behavioral relevance, decisions should be based on future market conditions
- Aggregate implications, spillover effects to financial and labor markets
Home prices exhibit
Short term momentum, long term mean revision
Properties of Home price expectations:
1.) Extrapolative, closely co-move with recent price changes
2.) Substantial Disagreement in cross-section
Cross section variation > time series variation
Which Informations are driving home price expectations
Information and experiences close to us
- close in time
- close geographically
- close socially
Study, Close in Time
Information Provision Experiment:
Measure peoples prior beliefs about past and future home price changes, 1&5 years (To get a perception gap).
Then inform participants about:
T1: past 1 year
T5: past 5 year
Control: No info
past home price changes and then re-elicit local home price expectations
What is key is the interaction term between the information provision T1perception gap and T5perception gap.
Study, Close in Time, Results
Home price expectations are extrapolative in both short and long term. Contrary to textbook logic, expectations across 5 years do not exhibit mean revision tendencies.
Study: Close Geographically
To what extent do people rely on local price information to form beliefs about aggregate (nationwide) home prices?
- Monthly datas on 1 year expectations
- Match respondents to past changes in their local zip code
- Relate expectations about aggregate to local experiences
Study: Close geographically, Result
Yes, local prices affect expectations about aggregate.
Study: Close Socially, Setup
Are expected home price changes propagated through social networks?
- Facebook data on friendship networks
- Ask Indiv. if investing in property in their zip code is a good/neutral/bad investment
- Calculate friends expectations by share of friends in San Francisco and delta Homeprices
- regress good investment response on constructed friendHPE
Study, Close socially, results
Local home price expectations of individuals increases with the share of firends they have in regions with recent home price growth.
Home price expectations and homeownership decisions, Study Setup
Does HPE, positively affect prob of buying a home among renters, negatively affect prob of selling among owners?
Hard to link correlations between decisions and expectations due to:
- omitted variable biases
- reverse causality
- classical measurement error in expectations
Use “Friends recently experienced local home price growth” as Instrument as they are likely uncorrelated with confounding factors
Home price expectations and homeownership decisions, Study Result
- Incrases prob of becoming a homeowner as a renter
- Increases prob of staying homeowner (not selling) among previous owners
Information Provision experiment, Expectation Shifter
Erklär ich dir später
Home Price expectations and current consumption decision. Setup
Field Experiment:
Respondents randomly receive high or low forecast about HP growth
- Forecast significantly shifts respondents own HP expectations
- Link survey data to scanner data from NIELSEN family to track consumption
Home Price expectations and current consumption decision. Result
No effect for the general population:
Renters:
- Significant decrease in consumption as they expect rents to increase. Or, future home purchase becomes more costly
-> both explain reduction of consumption
Owners:
- “Paper gains” if they do not plan on moving
- Higher proceeds from selling but also higher cost of buying new home.
But this only looks at private agents. There are no explanations for commercial participants in the real estate market. Hinted at the renters decision to save money as they expect rents to increase.
Internal Validity of Expectation Shifters
1.) Friends
+ geografically distant friends expectations probably orthogonal to confounding factors
+ within zip code comparison good control for local confounding factors
- Even distant friends might be socioeconomically similar
- Social influence might be endogenous
-> Moderately strong internal validity
2.) Information
+ Random assignment ensures confounding factors can be neglected
+ measure clean belief updating
+ includes incentivized behavior
- might react to survey framing or perceived credibility
- poses threat to experimental demand effects
-> High internal validity
External Validity Expectation Shifters
1.) Friends
+ Captures naturally occurring channel, social learning
- only uses Facebook data (representativeness issue)
- salience of friends expectations might limit generalizability
-> High external validity among digital-social populations, possibly lower for older cohorts
2.) Information
+ Mimics possible real world phenomenon with news updating
- might not be how people are exposed to information
- survey based, hypothetical investing decision
-> Moderate to good external validity
