Friday, June 22, 2018

Gimme shelter update: housing purchase affordability


 - by New Deal democrat

Let's update one measure of housing prices: comparing of the purchasing power of buyers vs. the price a typical house.  There are at least 3 indexes.

First, here is the N.A.R.'s "Housing affordability index," updated through May. This compares their estimate of median household income vs. the median price of an existing home:



Note that the data isn't seasonally adjusted, so there is an annual cycle. Even so, it is clear that affordability is at its lowest in 10 years. But on the other hand, the index is nowhere near its lows of 2005 or 2006.

Second, the private firm ATTOM Data Solutions publishes an index comparing annualized wages to median house prices:



This, like the N.A.R.'s Index, just made a 10 year high, or low, depending on how you look at the data, but again, is nowhere near the level of u affordability from the peak of the housing bubble.

Finally, here is the comparison, performed by Political Calculations, of monthly median household income, as measured by Rentier Research (which recently restarted this feature), compared with the median price of a *new* home:



New houses, as opposed to existing houses, are at new all time highs, even as compared with the peak of the housing bubble.

The sharp difference in affordability between new and existing houses is probably the main reason why existing houses most recently have sold after an average of only 27 days on the market, and months' inventory of existing houses is at all time lows.

Finally, just keep in mind that this measure does not tell us what the monthly mortgage payment for the median house will be. Because interest rates, even with recent increases, are still 1%-2% lower than they were during the 2000s housing boom, the monthly mortgage payment is nowhere near an extreme at this point.

Thursday, June 21, 2018

The state of the housing cycle


 - by New Deal democrat

The housing market generally proceeds according to an established cyclical rhythm.

My post describing the current status of that cycle as measured by interest rates, sales, prices, and inventory can be found at XE.com.

Wednesday, June 20, 2018

Further comments about the state of the housing market


 - by New Deal democrat

Here are some additional salient comments about the housing market right now.

1. Existing home sales are completely stagnant

Not only did existing home sales fall for the month, not only are they down slightly YoY, but they are now less than 2% higher than where they were 3 years ago:



In May 2015, houses sold at a 5.35 million annualized rate. In today's repot they were reported as selling at a 5.43 million annualized rate.

This, of course, is at complete variance to the very positive trend of new home sales over the last three years.

2. Not only sales, but inventories of new vs. existing houses have completely diverged

There was a great article this morning on housing over at Barry Ritholtz's The Big Picture. Be sure to read it all, but here are a couple of important graphs from the piece.

Inventories of existing houses are continuing to shrink, while inventories of new homes have been slowly rising for the last couple of years:



3, Why the divergence?

The K.I.S.S. explanation is that *prices* of new and existing homes have been moving roughly in lockstep, up about 6% YoY by nearly any measure, for the last 5 years. Pricewise, they have established a stable equilibrium. Thus, the other measures tend to gravitate back to whatever is necessary for that equilibrium  point.

An interesting point that the authors of the article at The Big Picture make is that the price differential between new and existing houses has expanded from 10% to 30% since 2006:



Meanwhile, the size of new houses has increased by 10%. That obviously explains some of the difference, but there is roughly another 20% that isn't so explained.

4. Foreign buyers?

The one possible explanation that isn't covered in The Big Picture piece is whether wealthy foreign (mainly Chinese) buyers are driving up the prices of (the relatively luxurious) new houses. I haven't seen any good articles about this for about three years. At that time it looked like foreign buying might be declining, but maybe it has picked up again since than.

If the profits in new homes are at the luxury end for wealthy foreign buyers with whom in the aggregate domestic buyers can't compete, that would drive domestic buyers to the "substitute goods" of existing homes. But that same dynamic would mean that *owners* (and potential sellers) of existing homes also, relatively speaking, would be unable to afford selling and moving up to the relatively luxurious new homes. Thus existing home inventory would be capped.

Foreign buying is an excellent fit for the missing part of the puzzle. I just don't know if it is factually true.

5. In the aggregate, there's not enough evidence that the housing market is at its peak.

Purchase mortgage applications -- including both new and existing homes -- have been very positive for the last few years. They made a new expansion high about 6 weeks ago. They have backed off since:


(h/t Ed Yardeni's blog)

but they are still running, on average this year, roughly 3% higher than their level of last year.

Yesterday I wrote of a poor report for single family housing permits, While I think the preponderance of evidence is that they are on the cusp of at least trending sideways, I should clarify that there have been several times during the last 5 years when permits have pulled back about 4% from the peak. The difference is that the pullback only lasted 2 months and then permits rebounded. This time the decline has lasted 3 months so far, and while we don't know what the future holds, the fundamentals of interest rates and prices have not improved.

Tuesday, June 19, 2018

Preponderance of evidence from poor housing permits points to slowdown in GDP


 -  by New Deal democrat

The preponderance of evidence, based on this morning's report on housing permits and starts, is that increased interest rates and continuing increased prices are beginning to take a bite out of the market. 

First of all, let's take a look at single family permits -- the most reliable, least volatile of all the measures -- (red, left scale) and total permits (blue, right scale):



Both declined this month, but more importantly, both made 7 month lows. Outside of the expiration of the housing stimulus way back in 2010, this is the first time that single family permits have made this significant a decline -- off about 4% -- during this expansion.

At the same time, since the peaks for single family homes were only in February, and overall in March, not enough time has passed to be confident that this was the peak.

Further, declines of 4% or more took place several times in the 1990s and 2000s without signaling the top of the market, as in 1994-95, 1996, and 2004:



Meanwhile, housing starts made a new high:



This includes on a three month rolling basis, which cuts down on volatility.

But permits lead starts, and in the last eight months there has been an increased backlog of housing permits which have not yet translated into starts. This month that number declined slightly to its lowest since December:



So it is not a surprise that starts have continued to rise even though permits have declined in the last several months.

At the same time, it's worth noting that, even though the economy didn't roll over, in two of the last 3 times -- 1994 and 2004 -- where there was a similar decline in permits outside of recessions, real GDP did slow down:




Should single family permits fail to make a new high for at least one more month, and should the decline be over 5% from peak, I will switch their rating from positive to neutral. And, while the evidence is by no means conclusive, I would say the preponderance of the evidence is that housing is slowing down, and that will have an effect on the economy over the next 6 - 12 months.

Monday, June 18, 2018

Prime age employment participation and wages: not so clear arelationship


 - by New Deal democrat

In the last couple of months variations of the same graph which is supposed  to "solve" the wage conundrum have been going around. I saw another version this weekend:




Easy to see, there is what looks like a nice, nearly linear relationship between the prime age (25-54) employment to population ratio (left scale) with wages as measured by the employment cost index (ECI)(bottom scale).

While I see some merit to the approach, I don't think the graph actually means what its purveyors think it does.

First, let me reproduce it with FRED data, which I am able to do through 2001:


Same pretty linear relationship, with maybe a slight bend at area of 79% participation and 2.5% wage growth.

Now let me do the same thing, except this time, while I am continuing to use the prime age E/P ratio, let me substitute the (inverted) U6 underemployment rate:


Same nice relationship, with a little bit bigger kink at the same point in the graph.

Now let me do the same thing again, this time using the U3 unemployment rate:


Once again, same nice relationship, with more of a kink at that same point in the graph.

In other words, in the source graph that allegedly "solves" the problem, it isn't the employment/population ratio that is doing the heavy lifting at all!  Rather, it is the substitution of the Employment Cost Index for average hourly wages that is responsible for the smoother relationship. That's because, unlike other measures of wages, the ECI has shown increasing wage growth in the last several years:



When we use average hourly wages as the measure instead of the Employment Cost Index, the nice relationship disappears:



And of course, since unlike the Employment Cost Index, average hourly wages go back all the way to the 1960s, we can see what the prime age e/p ratio looks like in previous periods.  Here it is for the 1980s and early 1990s:



Hmmmm, I don't think that is the same nice relationship whereby both increase and decrease together, do you?

When we confine the prime age E/P ratio to just men, to deal with the effect of the secular entry into the workforce by women, here's what we get:



Nope, that is not really helpful proving the case either.

Let me emphasize the I do think there is some merit to the approach. But at least part of the problem is that the labor market can only absorb so many new entrants at any given time, and when too many new entrants saturate the job market (as during the time that tens of millions of women entered the market during the 1970s and 1980s), wages will be depressed.

It is thus probably not a coincidence that the three periods of time during this expansion that have seen the highest rate of new participants in the last 20 years are the exact same periods when wage growth stagnated:


In short, it looks like the nice neat relationship between prime age employment participation and wage growth disappears when we use other measures of wages, and also when we apply it to prior periods.