Review of Fed Study Using Survey Data to Understand Housing Dynamics
Sein is always interested in new research that helps shed light on the complex dynamics of the housing market. A recently updated research note by Andreas Fuster and Basit Zafar of the Federal Reserve Bank of New York (FRBNY) provides important insights in understanding the impact of monetary policy on the dynamics of housing demand in relation to mortgage rates and down payment requirements. Specifically, using the change in willingness to pay (WTP), Fuster and Zafar conclude:
- A decrease in down payment requirements or an increase in non-household wealth has large affects on WTP;
- A decrease of 2-percentage points in rates changes WTP on average by 5%.
While these finding are important, the differences the authors found between homeowners and renters and the implied effects from household financial situation is, we believe, even more important from a policy perspective. More specifically, we use the findings from the research to forecast the following:
Without a significant change in credit availability and down payment requirements, we expect origination to decline in the coming months as all-cash sales continue to decline. Further, rate hikes expected by the Fed in the coming months should sharply curtail refinance activity.
The survey data used by Fuster and Zafar is from a special online survey that is part of the FRBNY’s Survey of Consumer Expectations (SCE). The survey sample comprised 1,420 household heads with a response rate of 85.3% (1,211). In addition to questions on each respondent’s expectations for housing prices, and the likelihood of them moving and purchasing a home, the researchers also asked information on debt levels, liquid savings, credit score range, zip code and household income. The final sample size was further reduced by:
- Dropped 28 respondents who did not answer all questions,
- Dropped 70 respondents younger than 21 or older than 70,
- Dropped 59 respondents who spent less than 3-minutes or more than 30-minutes answering the questions,
Final sample size used in the analysis was 962 for which the authors report the following key statistics:
- 698 are homeowners and 264 are renters.
- Even geographic distribution with approximately 23% of respondents from each of the Northeast, Midwest and West, and 30% from the South.
- Average age of respondents was 47.7
- Average household income of $181,000
- Average liquid savings of $368,000
- Average non-housing debt of $54,000
When compared to the overall population we find the following:
- Sample homeownership rate of 72.5% is 9% more than the 64% homeownership rate for the US as of January 2015
- Geographic distribution of the sample closely reflects the geographic distribution as of 2014
- The average income of the sample population is significantly larger than the $52,000 median income of the US, however, if inspecting the average income by quintile, we notice that the sample average is very close to the average income of the top quintile.
- Average non-housing debt of the sample is close to the average US household consumer debt profile of $48,000
- Finally, the average liquid assets of households included in the survey is significantly higher than the approximately $50,000 for the US according to the Pew Charitable Trust.
In short, the survey population is significantly skewed towards the top 2% of the US population in terms of income and household liquidity. This point not withstanding, Fuster and Zafar highlight some very important heterogeneity in the effects of rates and financing options within the sample survey.
All respondents are assumed to move to a similar locality and asked their WTP under four scenarios:
- I: All respondents are told that should they decide to purchase a home in their new locality they would be required to make a down payment of 20%, with half the sample assigned a rate of 4.5% and the remaining half a rate of 6.5% on a 30-year mortgage.
- II: Each respondent keeps the assigned rate from scenario 1 but is allowed to specify a down payment amount, subject to a minimum down payment of 5%.
- III: Respondent rates are switched to study and facilitate in-sample sensitivity of WTP to down payment and rates
- IV: Each respondent is told that they have inherited $100,000 in cash to study the effect of non-housing wealth on WTP.
Effects of lowering down payment
The authors report that the effect of a change in down payment requirement strongly depends on the household’s financial situation. Fuster and Zafar also argue that macroprudential measures such as LTV caps, predominantly affect the lower end of the housing market. Specifically, when comparing WTP between scenarios I and II the authors find:
- For the entire sample population, 11% lower their WTP, 46% leave their WTP unchanged and 43% increase their WTP.
- Renters 56% of renters increased their WTP
- The average increase in WTP is 10-15%, but for renters the increase it is around 40%
Effects of changing mortgage rate
- A 2% decrease in rate increases WTP by 3-4% and an increase lowers WTP by 5-7%.
Effects of Wealth Shock
- Average WTP increases by $15,000
- Down payments increased on average by $35,000
- Increase in WTP was larger for renters than for homeowners.
While it has long been understood that financing conditions affect housing demand, the survey method used by Fuster and Zafar provides some quantitative measure for further analysis. We believe the effects of down payment requirements is probably more pronounced than the analysis would suggest given the income skew in the sample population. Of particular interest to us was the disparity in the effects between renters and owners.
For the mortgage market as a whole the study seems to suggest that we should not expect any significant increase in origination volume through the first half of 2016 because down payment requirements remain high. Further, as the percentage of all-cash sales continues to decline, we should expect originations to decline in the coming months and quarters.