Random Snippets

Apologies for thin posting.  Lots in the air, even more than usual, and this blog sits low on the priority list.

A few brief thoughts below.  Maybe more on these later.

Macro Observations

Stocks and homes look like an incredibly good buy.  Price to earnings ratios for stocks and price to rent ratios for homes are around historic averages.  But prices are a lot better than average because the opportunity cost, reflected in interest rates, are at historic lows.  Put it all together an stocks and houses haven't looked so cheap in at least 60 years.

Housing and broader economic recovery.  Prices are finally rising nationwide, as are housing starts.  We appear to have a shortage of homes for an economy at full employment.  Thus, we should expect building to resume to at least historic norms soon.  That should bring back a lot of construction jobs.  Add a modest multiplier, consumer confidence with higher home equity and portfolio values, could bring about a virtuous cycle that could have our economy humming in another year or two.  This is the basic story Bill McBride's charts show, and he's been right on just about everything this past business cycle.  It pays to pay attention to the data.

ARMs safer than fixed-rate mortgages.  Conventional wisdom, past and present, has always been than a 30-year fixed-rate mortgage is safer than a variable rate mortgage.  That may be true depending on your definition of safe.  An ARM may seem risky given rates could spike and your mortgage payments could go up.  But if you've taken finance and digested the basic lessons of portfolio theory, you know that you need to think about how risks covary with other investments.  Then consider that the only way mortgage rates could spike is if we have real recovery in the economy (no, Greece is not a  likely possibility for us [1, 2]).  With real recovery, your wages, savings and home value are likely to go up with your mortgage payment and it should be easy to manage.  Alternatively, rates could fall in the event the economy stagnates (see Japan).  If this happens you could be stuck in your 30 year mortgage because your home value falls and you cannot refinance at the lower rate.  I don't mean to give financial advice here---you have your own expectations---but I'm pretty sure an ARM is the less risky thing to do.

Political Stuff

Nate Silver: Was awesome to behold this election.  Sam Wang deserves a lot of credit too.  Statistics wonks like me wonder why Nate Silver and Sam Wang had similar predictions but very different confidence intervals for their predictions.  In my own experience, I find it very easy to underestimate standard errors and very hard to overestimate them.  This, plus Nate Silver's occasional quips about the kinds of subtle error correlations he accounts for (e.g., parenthetical note in second paragraph below the map here), make me lean in his direction over Sam Wang.  On the other hand, Sam Wang is a lot more transparent about his model. He evens posts his Matlab code.  Hmmmmm...

Vacuum in the middle: With Tea Partiers displacing moderate Republicans and the Blue Dog Democrats retiring, we seem to have a vacuum in the middle.  This seems to confound basic models of political economy.  There would seem to be a big incentive for a small coalition of Republicans amass extraordinary power by thumbing their nose at Gover Norquist and taking a big step toward the center. After all, this elections proved that the power of money is limited and slapping backs with Obama can pay handsome rewards (Chris Christy's spike in approval).  It's hard to see Republicans maintaining their tight coalition.

Academic pursuits

Fixed effects gone bad:  It's fashionable in empirical economics to run regressions with giant data sets and a zillion fixed effects to control for unobserved factors or isolate a particular "source of identification." This is easy to do.  But if it's done in a thoughtless manner, bad things can happen.  This is an over-arching theme of our comment on Deschenes and Greenstone coming out in in the American Economic Review next month (finally).

Supply and Demand:  My paper with Wolfram Schlenker wherein we identify global supply and demand elasticities for food commodities using weather and yield surprises was recently accepted as a full-lenth article in AER.  There are some substantive revisions since the NBER working paper, but no major new insights.  This is probably my best academic contribution to date.

Personal stuff

Michelle and I have just adopted a son from South Korea. It's the best Holiday gift ever--we are in love already.  No pictures because it really won't be official until the home study is done.  The other big news item is that I've taken leave from NC State and have accepted a position at the University of Hawaii at Manoa.

Aloha and Happy Holiday.


  1. Regarding things Nate Silver, on the third hand, wasn't the truth, more or less, revealed after the election?

    David Simas tells us that "the race was "fixed" in the battleground states at a 3-to-4 point margin" (http://www.huffingtonpost.com/2012/11/21/obama-campaign-polls-2012_n_2171242.html?utm_hp_ref=politics).

    Then Jim Messina reveals that the Obama campaign's models (that generated the aforementioned "nowcast/snapshot") did very well indeed. From http://www.buzzfeed.com/rubycramer/messina-obama-won-on-the-small-stuff-4xvn we learn:

    "Their models ultimately predicted Florida results within 0.2%, and Ohio within 0.4%. The only state they got wrong, noted Messina, was Colorado, "where we got one more point than we thought we would."

    So, how might these facts be used to adjudicate the "dispute" between Nate and Sam? Well, first, the average of Sam's meta-margin over the campaign was indeed in the 3 to 4% range. It was just that the poll median was still too noisy. And, second, at a "fixed" 3 to 4% advantage in the swing states, Sam's standard errors could be adjusted upward significantly, still painting the picture that Romney basically had no real chance never! Nate Silver actually having played the role of Romney shill? What irony!

  2. Schtevie, I see your point. But it's hard to evaluate the standard error based on a single election. The problem, as I see it, is that systematic components of the error that span multiple states happen to land pretty close to zero. Nate assumes that systematic error component has a larger standard deviation than Sam does. But just because the draw we got was close to zero, that doesn't give us much information about the actual SD--maybe Sam just got lucky. Now, Sam seems to claim that we know this SD is fairly small from past elections. I (and Nate?) would counter that it might be larger as things change from one election to the next, like response rates, households without land lines, demographics of the likely voters, etc.

    I might have a stronger opinion if I spent some time studying Sam's code. That would be a really fun thing to do, but it's not something I really have time to do right now...

    1. Michael,

      Not sure if you have looked at Sam Wang's "worst case" calculation (http://election.princeton.edu/2008/08/28/technical-note-correlated-change-among-states-revisited/) from the 2008 cycle, what suggests that the covariation issue is of no great importance (at least in the relevant range of electoral competition as was recently witnessed).

      And of related interest is Messina's interview at Politico (http://www.politico.com/multimedia/video/2012/11/jim-messina-at-politicos-playbook-breakfast-event.html). He talks about the shortfalls of typical polling practices (and to some degree about how the Obama team overcame them).

      The point then being, for this particular election, the SDs were in fact very small: the Obama campaign identified its electorate, cultivated it, and to a large degree "controlled" it. Anyway, it's definitely worth a listen.

  3. Interesting thoughts on home prices, and I agree with you. Except for one thing that I've been trying to wrap my own mind around. What happens to home prices when interest rates rise?

    Obviously an ARM ticks upward and your mortgage payment gets larger. But what about the overall price of the asset, the home itself? You can think about a home as a type of bond (with a negative coupon in the form of taxes maintenance expenses and so forth). Thinking of homes this way makes the "interest rates are low so homes are a good deal" argument seem pretty thin. What do you think?

    Dan @ Casual Kitchen

  4. Daniel: it depends on whether the interest rate rise is expected or unexpected, and how much is inflation and how much of it is real. Way back in the late 70s and early 80s, we had huge interest rates and rapidly rising home values.But most of the rise in interest rates was inflation, and home prices rose even faster than inflation did. Someone from one of the big universities--I think it was James Poterba--wrote a paper on this, attributing the excess rise to tax incentives or something. Anyway, a rise in nominal rates means faster nominal home price growth. A rise real rates should depress home values all else the same. But the only reason real rates would rise is if the real economy grows much faster, and faster growth almost surely means higher home values all else the same.
    On net, I think higher interest rates almost surely means higher home values, at least for the US for the foreseeable future.


Post a Comment

Popular posts from this blog

Nonlinear Temperature Effects Indicate Severe Damages to U.S. Crop Yields Under Climate Change

Commodity Prices and the Fed