GY314 Economics of Housing Markets

Readings Notes

Harmon and Gill (2002)

W1: Married military men living in bases where length of stay is unknown.

Result: Expected stay goes up by 1 year= 3 percentage point increase in homeownership, ith transaction costs (TC) estimated to be roughly 3% to 4% of total income

Hilber and Liu (2007)

W1: Binary logit estimates of housing tenure

Result: Black households have less HO→ intergenerational externalities (parental wealth) and location

Hilber (2005)

W1/ W2: Locational risk and argues that HO is a risky financial investment created by neighborhood externalities like junk, littler, noise and crime. Externalities measured as perceptions of residents through interviews. Controls for endogeneity and unobservable factors by using a fixed effect regression.

Result: HO will be must lower where exposed to such externalities vs if there were positive externalities. As such people do respond negatively to negative externalities (particularly in cities)

Topel and Rosen (1988)

W2: Investigates investment in single-family housing in the U.S. using a real hedonic price index (Pt) to control for quality changes. They model housing as a durable asset where current investment depends on expectations of future prices and internal adjustment costs.

Result: Housing supply is much more elastic in the long run than the short run. Their estimates show a short-run elasticity of approx. 1.0, while the long-run elasticity is significantly higher at approx. 3.0, proving that builders adjust production levels slowly over time.

Di Pasquete and Wheaton (1994)

W2: Analyzes the long-run supply of new housing and the total housing stock. They critique previous models by accounting for the fact that as construction increases, land costs and construction costs also rise, which eventually dampens the supply response.

Result: They find lower elasticities than earlier studies: new housing supply elasticity is 1.0–1.2, while total housing stock elasticity is 1.2–1.4. This suggests that because of rising input costs (land), the long-run supply is less responsive than previously thought by Topel and Rosen.

Mayer-Somerville (2000)

W2: Critiques traditional models by arguing that housing construction is a flow (change in stock) that should be explained by changes in prices rather than price levels. They utilize a "flow vs. stock" framework to show that once prices stabilize, construction returns to steady-state levels regardless of how high those prices are.

Result: They find an extremely low long-run supply elasticity of approximately 0.08. While new building starts respond significantly to price shocks in the short term, the overall long-run elasticity of the housing stock remains very small due to the constraints of land availability and urban growth boundaries.

Hilber and Mayer (2004)

W2: Examines how physical land scarcity affects supply responsiveness across different locations→ utilize a sensible policy instrument (changes in local demand patterns) to isolate how the availability of undeveloped land dictates how much new housing can be built when prices go up.

Result: Locations with less undeveloped land exhibit a much more inelastic housing supply. In supply-constrained markets, demand shocks translate primarily into higher property prices rather than an increase in housing volume, whereas "unconstrained" areas see more construction and smaller price jumps.

Ermisch et al (1996)

W3: Analyzes UK housing demand across six regions, emphasizing how household formation and tenure choice respond to costs and income. They avoid 2SLS (Two-Stage Least Squares), arguing that house prices are exogenous to individual households in a regional market.

Result: They find a price elasticity of -0.4 and income elasticity of 0.5, confirming that UK housing demand is highly inelastic and slow to react to market shifts

Rosenthal et al (1991)

W3: Use American housing survey from 1981 looking at veteran loans which have no down payments→ constraint on the quantity

Result: Demand elasticity of 0.92 , however based on credit constrained HH if drops between -0.38 and -0.52 → liquidity constraints make demand much less sensitive to price changes

Shiller (2006)

W3: Historical prices in Amsterdam, Norway and US: similar patterns emerge when shown together

Gohl et al (2022)

W3: Plays down irrational exuberance: Short run dominated by momentum effects yet long run shows mean reversion (German setting)→ indicates that prices eventually correct back towards fundamental values.

Capozza et al (2004)

W3: idea of irrational exuberance: US setting: serial correlation stronger in booming markets→ consistent with euphoria → momentum driven and price increase today is predictive of price increases tomorrow

Stein (1995)

W3: Liquidity constraints: Income shock effects the ability of first time buyers to afford down-payments→ dramatic effect on overall housing market→ want to finance trade up and sell high

Result: Findings show that positive income shocks cause demand and prices to rise, facilitating trade-ups. Conversely, when prices are low, owners are "locked in" by their mortgages and must wait longer to sell, leading to a strong positive correlation between house prices and the total volume of transactions.

Kahneman and Tversky (1991)

W3: Loss aversion/ prospect theory credit constraints evidence consistent with loss aversion → idea that property owners are loss averse and not willing to sell in downturn

Result: Volume falls, time on market rises and prices then fall more

Genovose and Mayer (2001)

W3: Loss aversion: Add to K&T (1991) by saying loss aversion matters a lot

Result: liquidity constraints matter but much less (LTV>0.8)

Rosen (1974)

W4: Hedonic price model: provides economic model of hedonic (utility maximizing) equilibrium in markets where goods are not explicitly trades. Suppose we have a housing structure s, neighbourhood n, and consumption c:

U(s,n,c) and BC: m=c+p(s,n)

Ridker and Henning (1967)

W4: Air pollution: Examines the effect of air pollution on house prices using 167 census tracts in St. Louis (>60% single family units). They used median house prices as the dependent variable and sulfation levels as the primary proxy for pollution.

Result: While most results were significant, "above average schools" was insignificant and "time to CBD" returned a positive coefficient (it should be negative). This signaled a transfer of bias; because the location/school variables were poorly captured, the pollution coefficient likely absorbed that error, making the final estimate unreliable

Black (1999)

W4: One of the earliest hedonic analysis that was done very well→ matches 22,000+ single family residents School quality jumps at border, everything else smooth. Addresses OVB

Result: Without boundary fixed effects, the price premium is overstated (0.035). Using a narrow 0.15-mile band, the effect is refined to 0.016; specifically, a 5% increase in test scores leads to a 2.1% increase in house prices (approx. $3,948 at the time), proving parents pay significantly for school inputs.

Silva et al (2009)

W4: adds to Black (1999) in UK: uses data from primary schools between 2003-2006→ uses differences across LEA boundary to identify MWP, dependent variable is (log) house prices and use different aspects of school quality as independent variable.

Result: Once you control for boundary fixed effects: super stable at 3.70 vs 2.75→ same 2-3% evaluation of schools, so WTP is 3% more for 1 std deviation of school distrobution

Tiebout (1956)

W5: Proposes a theoretical model where local government expenditure is "market-like." He argues that individuals "vote with their feet" by moving to communities that offer their preferred package of public goods (e.g., schools, parks) and tax rates, assuming perfect mobility and full information.

Result: Under specific assumptions, this sorting process leads to an efficient equilibrium where local public goods provision matches resident preferences. Unlike national public goods, which suffer from "preference revelation" issues, the Tiebout mechanism ensures people end up in jurisdictions that reflect their willingness to pay.

Aaronson (1998)

W5: Investigates neighborhood effects by looking at how neighborhood poverty and high school dropout rates impact a child's probability of graduating. To address the "selection bias" (families choosing neighborhoods based on unobserved traits), he utilizes a sibling fixed-effects model, comparing siblings who lived in different neighborhoods during their teens.

Result: Finds a poverty coefficient of -0.129 and a dropout coefficient of -0.045. However, the study's reliability is debated because it relies on the assumption that family moves are exogenous; if a family moves because of a "shaky" home life, the neighborhood change is confounded by internal family instability.

Case and Katz (1991)

W5: Analyzes a 1989 survey of low-income youths in Boston to distinguish between neighborhood effects (peers) and "predicted location" (family background). They use a model to see if the behaviors of neighbors (crime, drug use, schooling) directly influence a youth’s own outcomes, effectively testing for social contagion.

Result: They found strong evidence of endogenous effects: a youth's probability of engaging in a behavior increases significantly if their neighbors do the same. For example, a 10% increase in neighborhood delinquency was associated with a roughly 5% increase in an individual's likelihood of being delinquent, even after controlling for family background.

Crane (1991)

W5: Investigates the existence of non-linear neighborhood effects, specifically testing the "epidemic model" of social problems. He looks at how the percentage of high-status workers (professional/managerial) in a neighborhood correlates with teenager dropout rates and teenage childbearing.

Result: Finds a sharp "tipping point" or threshold effect: in the very worst neighborhoods (bottom 5%), dropout rates jump dramatically. This proves the relationship is non-linear; moving from a "bad" to a "mediocre" neighborhood has a much larger impact on success than moving from "good" to "great."

Hilber and Mayer (2004) HO

W7: Land scarcity determines if homeowners support schools. In undeveloped districts, an increase in elderly residents drops school spending by -8.2% because they see no home-value gain. In highly developed districts, the same elderly demographic increases spending by +1.6% to boost their property’s resale value

SCCBS (2000)

W7: HO status affects social capital investment→ HO at 120.4 p.a vs renters 101.4 p.a

Green and White (1996)

W7: HO provides a good environment for upbringing of children

Galster (1983)

W7: HO treat housing stock more carefully→ better lived environment

Hilber and Robert-Nicoud (2007)

W7: HO affects planning regulation and restrictiveness of land use regulations

Hilber and Mayer (2004) W7

W7: Model on local public school as durable investment- households with children get benefit of schools vs no benefit, tax (Tau)→ median voters decision: community invests if the PV net of benefits is positive considering direct benefits, house price capitalization and duration of property. Renters do not benefit. Data from 11,000 US districts.

Result: 3 predictions which are: likelihood of school investment increases with greater capitalization, more HO= more school spending and elderly HO may even support if it boosts house prices.

Hilber (2007) [Social Capital]

W7: Social capital= neighbors co-operating create a local “club good”→ HO means they invest more → effect is stronger when land is scarce and no effect on non neighborhood social capital

Oswald (1996)

W7: Oswald hypothesis= HO reduces mobility→ leads to higher unemployment→ accept local jobs more easily and have lower reservation wages→ net effect is they find jobs faster+ housing increases entreupreneship as they use the house as collateral

Aksoy et al (2022)

W8: WFH survey: 1.5 average days WFH vs 0.7 before the pandemic→ desired 1.7 per week WTP= 5% for 2-3 days WFH

Card (1990)

W10: 7% increase in labour force in Miami as a result of Mariel boatlift in 1980 (125,000 Cubans arrived overnight)→ yet no effect on minimum wages and employment outcomes of other workers relative to comparison cities.

Results: Miami’s labor market absorbed the influx remarkably well, suggesting that the local economy adjusted through industry expansion or that the increased demand from new residents offset the supply effect. Labour demand adjusts

Pischke and Velling (1997)

W10: Investigates the impact of immigration on local labor markets in Germany during the late 1980s. Unlike the US-based "Mariel Boatlift" study, this research examines a period of high immigration into a European labor market characterized by more rigid institutions and different welfare structures.

Result: Finds small and often insignificant falls in the employment-to-population ratio. They conclude there are no systematic negative effects on the employment outcomes of native workers, suggesting that even in a more regulated labor market like Germany's, local economies are capable of absorbing immigrant inflows without significant displacement

Borjas (2003)

W10: effect of immigration on national labour market- main idea is one should consider skills and experience in labour markets to classify workers→ national skill-cell approach

Result: A 10% supply shock of immigrants reduces weekly earnings in that catergory by 4% + reduces annual earnings by 6.4%

Dustman et al (2005)

W11: Labour force surveys in UK→ analyses regional markets in 17 regions→ study both employment effects and wage effects.

Result: They find no significant overall effect on native employment or aggregate wages. However, when disaggregating by education, they find some significant negative effects for the intermediate education group, though these effects are not large.

Manacorda et al (2012)

W11: They find that immigrants and natives are imperfect substitutes—even within the same skill cell. As a result, a 10% increase in the immigrant share has almost no effect on native wages but widens the native-migrant wage gap by about 2%. Essentially, new immigrants primarily compete with existing immigrants, driving down their wages rather than those of the native-born population.

Kerr and Lincoln (2010)

W11: Suggestive that immigration brings valuable skills such as innovation and entreuprenship → in 2000, 12% of US were migrants and yet…

Result: 26% nobel prize winners (10yrs period), 25% of high tech companies over $1m, 24% of patents etc

Many in highly educated fields→ innovation clusters in these areas→ knowledge transfer→ amplified if they self-select and migrate (Roy Model)→ positively affect natives

Saxenian (1999)

W11: Studies SV and effect of asian migrants→ direct and indirect→ 11,000+ high tech founders + 100 in depth interviews

Result: 25% foreign born, 30%+ of high tech workers, accounted for $16.8bn sales + 58,000 jobs, more persuasive over time (13%→ 29%)

Hunt and Gauthier-Loiselle (2010)

W11: Use panel data covering US states over period 1940-2000→ 1% increase in highly educated immigrant means a 9-18% increase in patenting per capita

Issue is that share of migrants is endogenous→ skilled workers tend to migrant to states already experiencing positive technological shocks

Kerr and Lincon (2010) …

W11: Examines the impact of the H-1B visa program on US innovation and labor markets. The study focuses on the "supply side" by exploiting fluctuations in the H-1B cap (which varied between 65,000 and 195,000 during 1995–2008). They targeted firms and cities with high "H-1B dependency," noting that the majority of these visas go to Scientists and Engineers (S&E), with roughly 50% arriving from China and India.

Result: A 10% increase in the H-1B population leads to a 2–4% increase in the growth of S&E patenting for each standard deviation increase in a city's H-1B dependency. Crucially, they found little evidence of "crowding out" native workers; instead, a 10% increase in H-1B visas led to a 0.5% increase in total S&E employment, suggesting that the influx of foreign talent expands the overall technical workforce

Gagliardi (2013)

W11: Investigates skilled immigration’s link to UK innovation across 103 TTWAs (2001–2007) using Fixed Effects and Instrumental Variables to control for settlement bias.

Result: Confirms highly skilled migrants significantly boost regional technological innovation, proving foreign talent drives UK knowledge economies similarly to the US

Nathan (2015)

W11: Focuses on patenting behavoiur and asks if it is affected by ethnic minority investors and diversity→ results also parallel the US showing ethnic investors boost innovative activities