Friday, October 16, 2015

State price indices! Get em while they're hot!


Martin Beraja, Erik Hurst, and Juan Ospina have constructed state price indices from retail scanner data! That's pretty badass. Nice job, guys.

I don't know what state-level price indices already exist, but these look pretty good, so if you need state-level price data for your research, you might want to mail these guys! It's only from 2006, of course, and it's biased because it only includes stuff that people buy in stores.

Anyway, Beraja et al. use their state-level data to help make a model of regional shocks. (I don't really believe the model that much, but that's not the point.) This kind of micro-flavored macro - "dissagregated" macro, if you will - seems to be getting more important. I've heard a number of other macro folks talk about making these sorts of models, and there seems to be a lot more focus on collecting region-, industry-, and firm-specific data. Related trends include models with heterogeneity, network models, and institutional models (for example, this one that I just saw presented).

That's cool. A lot of macro people seem to have woken up to the fact that their old model paradigms (NK, RBC, and the like) weren't really working, and new stuff is needed. And also, at the same time, woken up to the fact that the available data aren't very informative. The new micro-focused macro models, and new data sets, seem to be ways of trying to chip away at the problem. The new stuff hasn't taken over yet, but it's spreading.

Also, Beraja et al.'s data set demonstrates quite clearly how the IT revolution has dramatically increased data availability. I think everyone realizes that that's the big force changing econ right now.

8 comments:

  1. Anonymous5:12 PM

    Interesting and ripe for comparisons with other regions of the world. Europe has in a sense had disaggregated data for a long time so comparisons might unearth something new about the differences between the state markets of a politically unified region (USA) and the national markets of a politically fragmented region (Europe).

    ReplyDelete
  2. BLS has been releasing price indices at even the MSA level for a long time. And you can even get breakdowns by rent/food/etc.

    https://research.stlouisfed.org/fred2/graph/?g=1SFQ
    http://www.bls.gov/regions/west/news-release/consumerpriceindex_sanfrancisco.htm

    ReplyDelete
    Replies
    1. Hmm...that's cool! But I can't find it by state, only by MSA...

      Delete
    2. Anonymous11:38 PM

      This should help , assuming you're only interested in the state of Massachusetts , from 1910 to 1943 :

      https://research.stlouisfed.org/fred2/graph/?g=1SFQ

      Delete
    3. Anonymous11:42 PM

      Sorry , here's the right link :

      https://research.stlouisfed.org/fred2/graph/?g=2bQK

      Delete
  3. One of the Federal Reserve Blogs states that inflation in the US is in the west. I suppose they can't say that without some data from the west and some data from the other regions and then making a comparison. This Berjara data can't be the first source of such data. Is that not the regional Fed's job already. May be the Fed do it by regions and not individual states.
    https://www.stlouisfed.org/on-the-economy/2015/october/regional-cpi-inflation

    ReplyDelete
  4. Anonymous6:41 PM

    Noah,
    How come you economists dont switch to writing your stuff in code? Arent most of you younger, less computer illiterate kids armed with R or Python capacity? Wouldnt it be great if I could just open your repository and grab your data and code and play around too ?

    ReplyDelete
    Replies
    1. Yep, economists mostly use Python or Matlab, sometimes R. Not posting code is just a cultural quirk.

      Delete