Sunday, March 1, 2015

Housing Tax Policy, A Series: Part 16 - What a difference framing makes.

Here is a graph of the Price to Rent ratio for housing, from CalculatedRiskblog.com .  That really looks like irrational exuberance.  A case of national insanity, which has fallen back to reality.

What if we flipped this over, so that it is in terms of yield, like a bond?  Here is a graph of Rent to Price Ratio for housing.  (Here I am using annual data from the BEA for rent and the Federal Reserve for real estate values, which has similar behavior to the Case-Shiller National index in the first graph.)  I have also included two other real financial indicators - real GDP growth and real 20 year treasury bond rates.  I have also included a Net Rent to Price ratio, in order to provide a more accurate comparison between the bond returns and the net returns to homeownership.

No more insanity.  It was that easy.

Here's what happens when I try to create a Price/Rent version of 20 Year TIPS bonds and real GDP growth rates.  I added a 1% "premium" to the GDP growth rate and the treasury bond, before inverting them.

The treasury bond market and the residential real estate market are similar in size.  Could it be that long term TIPS should be yielding about 2%, nominal long term treasuries about 4%, and homes about 3%?  And, the dislocation in the housing market is keeping capital out of housing, and pushing it into treasuries, pushing home returns up (prices down) and treasury yields down?

This is a reason why, while I think a big inflationary boost would help get housing back in order, if we aren't going to get that, I am fairly sanguine about near term interest rate movements.  The demand (investment) end of the credit market would probably be pulling long term treasury rates up to 4% or so if we didn't have this supply (savings) glut.  And, rates would really have to jump before it would move home returns below the alternatives.

By the way, here is a graph of YOY changes in home prices, adjusted for rent inflation.  There is nothing unusual about the 2000s.  And, we can see from the graph above that Price to Net Rent ratios in the late 1970's were not substantially lower than they were in the 2000's.  Real interest rates and real GDP growth were also relatively low in the late 1970's.  There is no mystery to solve here.

I just don't understand why sophisticated, numerate observers are so insistent on finding a cause to the non-existent problem of the housing boom.  (I do think that homes have moved to permanently higher prices because of tax issues, which should be reversed, but this is simply another explanation for the price level.  There is still no mystery.)

Look at the first Fred chart above.  Net rent to price has been above real GDP growth almost permanently since the late 1980's.  Before that, real GDP regularly pushed well above implied returns to homeownership during expansions.  This suggests that home prices have been too low for the past 30 years, if anything.

5 comments:

  1. Kevin, as always a thought-provoking post. Towards the end of this post (http://bit.ly/1zazodK) on price-to-rent ratios, I look at % of peak value reached by state. It seems like, in a lot of states, we are approaching the peak "bubble" value. That seems to also suggest that housing prices are simply tying to poke up back up to where they should have been all along. A big factor in the 2007-09 collapse in housing prices seems to have been over-tight Federal Reserve policy (re: Scott Sumner). Would you agree with that assessment? If not, then I would follow up with why the collapse?

    ReplyDelete
    Replies
    1. Great visuals on that post!

      On your post, I would expect rent and price to move up together in cases where things like neighborhood amenities were the drivers of value. I would expect P/R to be more indicative of different tax rates or future expected changes (P/R would be higher in an "up and coming" neighborhood, because rents would be expected to rise there. Current rents wouldn't price in future expectations nearly as much as price would.)

      I would also say that home ownership is a real asset while the mortgage is a nominal debt, so in equilibrium, mortgage payments should be much higher than rent payments on a fully leveraged house. If you're putting 20% or less down on a house, and rent would cover the mortgage payment, you're pulling in mucho excess returns, because the house will be gaining unrealized capital gains over time, while the mortgage face value declines.

      I completely agree that tight policy caused the collapse. I also think tight policy made prices higher to begin with because tight policy made inflation premiums very low, and even probably caused long term real rates to be slightly lower than they would have been because of the danger of AD shocks at such low inflation rates. Both of these factors would lead to higher home prices. So, tight money led a little bit to the boom and was pretty much totally responsible for the bust, IMHO.

      But, to be clear, given the interest rates we had, the boom prices were reasonable. The bust prices however are simply out of equilibrium because of the decimated credit markets and frictions that keep institutional buyers from immediately picking up the slack. If recovery continues, P/R as high as they were before are justified. But, keep in mind, that's because treasury bonds are less than 3%. If P/Rs do recover, they may not be a good investment for everyone, just like 30 year treasuries aren't right for everyone.

      Delete
    2. I was a little sloppy in my response. Of course, you would need to factor in expenses and taxes on the rent to mortgage payment comparison. It is net rent that would be the important comparison.

      Delete
    3. Thank you for the excellent response. As a follow-up, you're saying that current (maybe higher) P/R is justified at current interest rates, right? So if interest rates go up, those ratios will have to adjust downwards (either rents increase because we're seeing more inflation / prices fall because the Fed has tightened too quickly). Assuming my logic is right, does the same hold true for stock prices (replace rent with earnings)?

      BTW, those visuals were created using Tableau Software (http://www.tableau.com/). If you get the chance, I highly recommend it. Also, I'm not a sales rep, just a huge fan.

      Delete
    4. That's right. But, home prices are way out of whack. Interest rates would need to rise a lot before they would be the constraining factor on home prices.

      It is technically true for stocks, but I think stock prices are much less rate sensitive than bond and home prices. This is because total required returns to the firm tend to remain fairly stable. Equity risk premiums are currently about 6% for the S&P 500. If long term rates go up 2%, I wouldn't be surprised to see equity risk premiums fall by 2%, so that discount rates on equity will be relatively unaffected. Earnings levels and growth rates are more important for stock valuations, I think. Also, PE ratios are somewhat high right now for stocks because firms are so deleveraged. All else equal, lower leverage should mean higher valuation ratios. This is slightly counterintuitive, but think of what the PE ratio would be on a firm that was completely equity funded and whose business plan was to invest all the capital in 10 year treasuries. It would have a very high PE ratio.

      Delete