Common wisdom says home sales increase when mortgage interest rates are low and fall when rates are high. Interest rates have been rising lately, but with the benchmark 30-year fixed rate mortgage (FRM) currently just north of 5%, rates are still fairly low – at least by historical standards. So why have home sales and refinances fallen so dramatically? The answer: Everything is relative. While 5% may be well below the 50-year average of 7.77%, as illustrated in Exhibit 1, it is also the highest rates have been in more than a decade. The last time rates were this high was in April 2010.
Exhibit 1: Historical Mortgage Rates When viewed over the last few years (Exhibit 2) we see a dramatically different picture. In August 2021, 30-year fixed rate mortgages averaged 2.84%, while in early August of 2022 they are about 5.43%, an increase of 91% in just one year. Many lucky borrowers were able to lock in low rates and, as discussed in prior blogs, the vast majority now hold mortgages with historically favorable low rates (under 4%). With rates rising again, many would-be borrowers are choosing to remain on the sidelines, holding on to their low-rate mortgages instead of refinancing or purchasing a new home.
Exhibit 2: Recent Mortgage Rates So what does this mean for the housing market? Does history provide any guidance? And how do historically low inventories of homes available for sale change the picture?
The Market for Mortgages
We know mortgage rates directly affect affordability via the impact on payments. As of mid-year 2022, approximately 30% of home buyers are paying with 100% cash as opposed to 31% for all of 2021. Stated another way, some 70% of buyers use a mortgage, but many buyers do not need one. Many older retired homeowners have no mortgages, which stabilizes some housing markets and makes them less vulnerable to mortgage rate swings. One implication of higher fixed rates is that we should expect a significant increase in adjustable-rate mortgage (ARM) borrowing as one way to offset the higher FRM mortgage rates. ARM borrowing has been near nil the last several years (under 3% of market share), but that share is likely to increase, given that ARM rates tend to run 100 to 150 basis points below FRMs.
Prior Analysis on the Impact of Mortgage Rates
Housing prices and sales volume tend to exhibit inertia or momentum, and recent trends foretell the near-term future. Stated another way, sales prices and sales volumes do not change instantly the way the stock market may react to new information. At the same time, fundamentals matter and interest rates are critical determinants of affordability and purchase decisions. For decades our research has tracked a direct and inverse relationship between mortgage rates and housing sales. That is, when mortgage rates have gone up, sales volumes declined and vice versa. We have also explored the impact of mortgage rates, combined with employment trends, on home prices. As expected, mortgage rates matter a great deal, as does employment growth, and historically, higher mortgage rates were correlated with slower or negative employment patterns. Positive employment trends are associated with stronger housing markets. Weaker employment trends foretell the opposite. This year is proving to be an exception to the pattern of higher mortgage rates with lower employment, as employment hiring through mid-2022 has remained strong, even in the face of higher interest rates and high inflation. Our prior research also suggests the markets most sensitive to mortgage rate changes are those near the coasts. In general, the least sensitive markets are in the middle of the country, where new supply can adjust more rapidly and affordability has been historically less of a critical issue. Looking at fundamentals as long term drivers of the housing market, we currently have an unusual combination of rapid increases in mortgage rates and strong employment gains, coupled with a lack of housing inventory.
A Look at the Last Forty Years of Housing Data
In Exhibit 3, we examine the last forty years of annualized percent changes in U.S. existing single-family sales versus annual changes in mortgage rates using quarterly data. Some relationships between mortgage rates, home sales and home prices can be seen with time series data which are quite compelling. This data encompasses real estate and economic cycles with different driving forces and characteristics. The fact that these relationships persisted throughout such a long period suggests they reflect true underlying fundamentals. In particular, the time series data shows that mortgage rates lead home sales by one quarter and home sales lead home prices by three quarters. With regard to home sales, we observe a negative correlation coefficient of .64 and the regression coefficient suggests that a relative 1% increase in the mortgage rate leads to a negative 8% change in sales volume on average. Astute housing analysts should also keep in mind that relationships between mortgage rates and sales volumes and between sales volumes and prices are likely to be non-linear. There will be a natural floor level of housing sales independent of interest rate related factors, including but not limited to sales driven by births, deaths, divorces, job moves and foreclosures.
Take this one step further and you can see that changes in sales volumes lead changes in prices. Exhibit 4a shows the annualized change in U.S. single family home prices (quarterly data) versus annual changes in the number of sales, with the strongest relationship found with sales volumes lagged three quarters. That is, mortgage rate changes lead sales volume changes by several months and sales volume changes lead price trends by several quarters. With mortgage rates currently about two percentage points above a year ago, these simple lagged relationships suggest that the median existing single-family price should be flat to slightly down on a year-over-year basis by the end of 2022. Similar relationships hold when analyzing individual markets as seen in Exhibits 4b and 4c for Chicago and New York, respectively.
Exhibit 4a: US Home Prices Versus Sales Volumes Lagged Three Quarters Exhibit 4b: Chicago CBSA Home Prices Versus Sales Volumes Lagged Three Quarters Exhibit 4c: New York Metro CBSA Home Prices Versus Sales Volumes Lagged Three Quarters Conclusions
Normally, we expect higher mortgage rates to have a significant impact on housing demand, resulting in far fewer sales and most markets softening with respect to prices. The pattern is fairly clear: mortgage rates increase, sales volumes decline and home prices soften with time lags of several months to several quarters as the market adjusts to new realities.
Mortgage rate increases in the past year are dampening demand, not only by affecting affordability – especially for first-time homebuyers – but also by impacting existing borrowers who do not wish to give up bargain mortgage rates by selling. These same, well-timed borrowers are therefore also unlikely become active buyers. We will see both demand and supply curtailed in the housing market for a few years, but it will not drop to doomsday levels and there are many non-financial drivers of the housing market.
Sales volumes are likely to fall off in the last half of 2022 and throughout 2023, even if employment figures remain strong. Historical relationships suggest sales could drop by 15% or more. The strong employment figures and the increased percentage of cash buyers (30%) or those who use ARMs will help mitigate what would otherwise be a temporary suspension of the real estate market. Other drivers of demand and supply, including job moves, deaths, divorces, child births and foreclosures will help keep sales volumes higher than would be the case if mortgage rates were the only factor driving sales, but a significant drop off in sales is possible, and this will in turn, soften the entire economy and slow down job growth.
Home prices will likely flatten or decline modestly, mitigated by historically low inventories. But we caution against over-reliance on short term median or average sales prices for a metro as there is normally a seasonal pattern in both sales and prices and we should not confuse the normal pattern with the long-term trend. Reports of lower sales prices based on broad geographic median price trends will likely be larger than the corresponding reality when controlling for seasonality, size and quality.
Markets with strong employment growth, low inventories and constrained supply will see uneven price trends within their metro markets. Lower price tiers will be more sensitive to the interest rate increases while higher price tiers will tend to remain more resilient against price dips. Average price trends will be somewhat overstating the underlying trends as the mix of homes that sell may be smaller on average, which is one way the market reacts to less affordability from higher mortgage rates. The rapid appreciation of the last few years drew some otherwise indifferent sellers into the market. These listings are likely to be withdrawn and, in other cases, listing prices will be adjusted downward. Whether anyone will classify these reactions as a housing bubble crashing is a function of how sensational they wish to make their headlines. Sales volumes will likely slow and prices will soften, but as of mid-2022 this appears to be a return to a more normalized market.