Brexit, State of the Nation's Housing, and home sales: the week in charts.
26 June 2016
IT WAS A BUSY WEEK FOR ECONOMIC AND HOUSING DATA. Existing and new home sales came out, the Joint Center for Housing Studies of Harvard University released their annual State of the Nation’s Housing, and the U.K. voted to leave the European Union (the Brexit). We’ll recap these data and events through charts I’ve created and shared throughout the week.
In this post, I’ll share each of the charts with some commentary, and then below, I’ll include the R code I used to generate the charts.
Home sales catching up to the fastest pace in a decade
Both existing and new home sales for May 2016 came out this month, with Existing home sales released on Wednesday and new home sales on Thursday. Home sales have been picking up steam throughout 2016. Total home sales are likely to reach the highest annual total since 2006. In order to do that, sales have to outpace 2007.
Existing home sales post best monthly sales since 2007
Existing home sales for May 2016 were at the highest level since 2007 and seem to be getting traction. The graph below compares monthly existing home sales (not seasonally adjusted) across years. This year’s pace is running below 2007, but the gap is narrow. In 2007 home sales slowed in the summer and fall, so if this year’s pace maintains its momentum it’s likely this year will be the best year for existing home sales since 2006.
New home sales drop, but estimates uncertain
New home sales had an extremely strong initial reading for April, but the May release revised down the estimates substantially. In May 2016, new home sales were up 8.7 percent year-over-year. The estimates for new home sales are subject to a lot of sampling uncertainty, and Census provides estimates of that uncertainty. In the chart below I plot not only the level of new home sales, but also the 90% confidence interval constructed using data from Census.
Total home sales behind 2007, but looking to gain
Through May total home sales (new + existing) are running below 2007 pace, but not by much. The chart below compares monthly total home sales (NSA) since 2007. In 2007 home sales slowed considerably in the seconed half of the year, leaving room for 2016 to catch up if we can maintain a pace similar to last year.
Ahead of Brexit, interest rates already falling.
On Friday after the Brexit leave vote was announced, financial markets roiled, with stocks tumbling and Treasury yields falling. But even before the vote was announced, long-term interest rates had been falling.
Will mortgage rates fall even more next week? With 10-year Treasury yields down almost 19 basis points on Friday, we should expect to see mortgage rates trend lower in the next week. But even if the declines in Treasury yields stick next week, in times of volatility some of the declines in Treasury yields do not flow into mortgage rates, instead spreads widen.
The chart above shows that in general, the 30-year fixed mortgage rate follows the 10-year Treasury pretty closely. However, in 2012 there were several points below the line. Those weeks correspond to the spring of 2012 when 10-year Treasury yields fell more rapidly than mortgage rates. This was partially due to increased volatility and risk aversion around other European macroeconomic risks.
The chart below shows the time series for the 10-year Treasury and the 30-year mortgage rate, with the spread shaded.
The chart below plots the spread, with the 2011-2016 average (about 1.7 percentage points) spreads denoted by the dotted lines. Periods when the spreads were above average (2011-2012) and again in 2016 were periods with increased macroeconomic risks.
While risks may be elevated, it might be helpful to compare to 2008 when spreads were considerably wider:
This week the Joint Center for Housing Studies (JCHS) of Harvard University released their annual State of the Nation’s Housing. The report is full of useful information about housing markets in the United States.
One area of concern for analysts of U.S. housing markets is the rising cost burden of housing. In the report the JCHS computes costs burdens using Census data. Cost-burdened households are those that pay more than 30 percent of income for housing, while severely burdened households spend more than 50 percent. The chart below compares the percentage of cost-burdened households by metro area to the U.S. average broken out by household income.
Code to make selected charts
In this section I’ll provide code to make a few of the charts I posed above.
First we’ll load the required libraries.
Existing home sales chart
Next, we’ll make a static version of the bar chart:
Animated existing home sales chart
Make the animation with the following code:
New home sales chart
Animated new home sales chart
The animated gif for new sales is pretty straightforward to construct. We won’t be using tweenr. Instead, we’ll just loop through each row in the data frame, plotting all the points before that row.