Believe those who are seeking the truth. Doubt those who find it. Andre Gide

Wednesday, April 10, 2013

Poor Germany

Well, here's an eyebrow raiser: Germans Among Poorest in Europe: ECB Study

The paper is available here: The Eurosystem Household Finance and Consumption Survey. The cross country comparison of net wealth can be found in Table 4.1 on page 76.

Median net wealth in Germany for 2010 was 51K eur. Compare this to median wealth in Greece (101K), Italy (173K) and Spain (182K).

This just doesn't sound right to me, but I haven't gone through the report in detail. Evidently, the differences are driven primarily by real estate wealth. Thankfully (?), Germany escaped the housing price "bubble" that afflicted many European countries; see figure below.


Moreover, as I noted here, the German growth experience over the past 20 years has been nothing to write home about.

Germany: low growth, no asset price bubbles, low wealth, but...stable. Das ist gut?



April 16, 2013: Update here from VOX, who emphasize that the wealth distribution in Germany distorts the picture presented here. 

7 comments:

  1. These numbers are way too different, as you suggest. Must be some measurement or institutional difference driving this. I'd start with K/Y, then figure out why W/Y is so different. I'd love to know the answer, but doubt this is it.

    ReplyDelete
    Replies
    1. Dave, yes, maybe I should have put a "?" in the title of this post. You must have some smart grad student eager to take up this task? :)

      Delete
  2. Hi Dr. Andolfatto,

    It's nice to finally comment on your blog for the first time. I hope that you are doing well in St. Louis, and I also hope that the economics students at SFU still get to benefit from your teaching from time to time. I'm looking forward to the release of the 3rd edition of your textbook; I've enjoyed reading the parts of the 2nd edition that I can understand so far.

    In the plots that you showed above, is it possible to show some measure of uncertainty around those lines - perhaps a confidence band? That would be necessary to determine whether the differences between the different countries' real house prices are significant or not.

    I am not well versed in time series yet (a deficiency that I aim to correct), and I am not very familiar with reading econometric or macroeconomic data, so perhaps there are conventions that I don't know about. I was, however, quite pleased to see standard errors given for many of the estimates in The Eurosystem Household Finance and Consumption Survey.

    Thanks for educating a statistician interested in economics about data in your line of work.

    ReplyDelete
    Replies
    1. Hi Eric. Well, I'm no statistician, but I feel pretty confident in my (shared) belief that house price dynamics in Germany were quite different from those countries that experienced a house price boom-bust cycle!

      Delete
    2. I trust that this is true, Dr. Andolfatto. In general, though, if economists want to determine whether or not two curves (house prices, inflation, unemployment rate, etc.) measured over time differ, is it customary to include confidence bands to ascertain such differences?

      I recognize that this issue of confidence bands and significance may deserve examination for economic data in general, and not just for this particular post. (I didn't mean to single out this plot.) Nonetheless, perhaps further discussion or conversation is needed.

      I recently attended a seminar by a health journalist who was perplexed by the lack of confidence intervals for all types of estimates in health care when I raised this point; she was even more baffled considering how they are commonly given for political polls ("...19 times out of 20"). A productive Twitter conversation ensued with several others in the field acknowledging this lack of uncertainty in health journalism.

      I appreciate your and anybody else's thoughts on this while I dip into economics with a statistician's eye.

      Delete
  3. Thanks for the interesting post.

    I think that using medians here can be misleading.

    As shown in table 4.2, wealth is very unevenly split among homeowners and non-homeowners. That results in median wealth values extremely sensitive to the share of a nation households that are homeowners.
    Using averages instead brings Germany to 195,1 k€ and Italy to 275,1 k€. Still questionable but sounds more like the real thing...

    ReplyDelete
    Replies
    1. Ciao Gianni,

      I wasn't sure which to highlight, but as you suggest, it's still surprising.

      Delete