Friday, December 23, 2016

Macro Musings Podcast: Xmas Economics

My latest Macro Musings podcast is a special holiday edition on the economics of Christmas.Two special guests joined me all the way from Germany to discuss this topic. My first guest was Anna Goeddeke, a professor of economics at ESB Business School in Reutlingen, Germany. My second guest was Laura Birg, a postdoctoral researcher at the Center for European, Governance, and Economic Development Research, University of Göttingen 

Together they coauthored an article in Economic Inquiry titled “Christmas Economics—a Sleigh Ride” that surveys and summarizes the economics literature on Christmas. It is a great read for this time of the year and was the basis of our conversation. We touched on a number of interesting topics like the seasonal business cycle, the deadweight loss of Christmas, and charitable giving during the holidays.

The seasonal business cycle discussion was particularly fascinating for me. There is a literature that starts with Barksy and Miron (1989) (ungated version) that shows most of the variation in aggregate economic measures like GDP comes from seasonal fluctuations. Yet most macroeconomists, myself included, typically start our analysis with seasonally-adjusted data. Here is a Barky and Miron summarizing their findings on GDP for 1948:Q2-1985:Q5:
The standard deviation of the deterministic seasonal component in the log growth rate of real GNP is estimated to he 5.06%, while that of the deviations from trend is estimated to be 2.87%. Deterministic seasonal fluctuations account for more than 85% of the fluctuations in the rate of growth of real output and more than 55% of the (percentage) deviations from trend. Business cycle fluctuations and/or stochastic seasonal fluctuations represent a relatively small percentage of the fluctuations in real output. Plots of the log level of real output (Figure 1) and the log growth rate of real output (Figure 2) make this point even more clearly. The seasonal fluctuations in output are so large and regular that the timing of the peak or trough quarter for any year is rarely affected by the phase of the business cycle in which that year happens to fall. 
Unfortunately, the BEA no longer makes available non-seasonally adjusted GDP data. However, we can look at other times series to see how large seasonal swings can be relative to recessions. For example, below is retail sales:

What makes this really interesting is that these wide swings in economic activity are not matched by similarly-sized swings in the price level. Most of the seasonal boom is in real activity. Put differently, there is an exogenous demand shock every fourth quarter where prices remain relatively sticky so real activity surges. This is a microcosm of demand-side theories of the business cycle. It seems, then, that more could be learned about broader business cycle theory from studying GDP and other time series in their raw non-seasonally adjusted form. That will have to wait, however, until the BEA starts releasing the data.

This was a fascinating conversation throughout. You can listen to the podcast on Soundcloud, iTunes, or your favorite podcast app. You can also listen via the embedded player above. And remember to subscribe since more shows are coming.


  1. Terrific post. It confirms that money is not neutral.

  2. "Yet most macroeconomists, myself included, typically start our analysis with seasonally-adjusted data."

    One should always eschew the seasonally mal-adjusted data. Monetary policy objectives should be formulated in terms of desired rates-of-change, roc's, in monetary flows, M*Vt (our means-of-payment money times its velocity of circulation), relative to roc's in R-gDp.
    Roc's in N-gDp (though "raw materials, intermediate goods and labor costs, which comprise the bulk of business spending are not treated in N-gDp"), can serve as a proxy figure for roc's in all economic transactions, P*T, in Yale Professor Irving Fisher's truistic "equation of exchange". Roc's in R-gDp have to be used, of course, as a policy standard.

    - Michel de Nostredame

  3. "What makes this really interesting is that these wide swings in economic activity are not matched by similarly-sized swings in the price level."

    What it does is in-validate Scott Sumner's N-gDp targeting.

  4. What it does is validate Yale Professor Irving Fisher's "equation of exchange" (the distributed lag effect for real-output)

    Yale Professor Irving Fisher's transaction's concept of money velocity, or the "equation of exchange", is an algebraic way of stating a truism; that the product of the unit prices, and quantities of goods and services exchanged P*T, is equal, for the same time period, to the product of the volume, and transactions velocity of money M*Vt.

    The "transactions" velocity (a statistical stepchild), is the rate of speed at which money is being spent, i.e., real money balances actually exchanging counter-parties. E.g., a dollar bill which turns over 5 times can do the same "work" as one five dollar bill that turns over only once.

    It is self-evident from the equation that an increase in the volume, and/or velocity of money, will cause a rise in unit prices, if the volume of transactions increases less, and vice versa. Of course, this is just the "unified thread" of algebra, estranged from "general field theory" of macro-economic modeling, where the chorus is: "All analysis is a model" - Ken Arrow.

    In contradistinction, the mainstream Keynesian-economic metric variant, income velocity - Vi, a contrived figure, is calculated: by dividing N-gDp for a given period by the average volume of the money stock (M1, M2, & MZM), for the same period (viz., make believe). A decline in the income velocity of money (like during the Great-Recession), i, is supposed to suggest that the Fed initiate an expansive, or less contractive, monetary policy.

    This signal could be right - by sheer accident. I.e., the historical trend of Vt vs. Vi, at various intervals, moved in absolutely divergent paths - giving the income velocity economists false signposts:

    See: Sept. 30 1979, correspondence between R. Alton Gilbert (Ph.D., Senior V.P., FRB-STL), and Leland J. Pritchard, then Professor Emeritus, KU (Ph.D., "Chicago School", Economics, 1933, MS Statistics, Syracuse).

    Remember that in 1978 all economist's forecasts for inflation were drastically wrong. To put that into perspective (Business Week): there were 27 price forecasts by individuals & 9 by econometric models for the year 1978. The lowest (Gary Schilling, White Weld), the highest, (Freund, NY, Stock Exch) & (Sprinkel, Harris Trust & Sav.).
    The range CPI, 4.9 - 6.5 percent. For the Econometric models, low (Wharton, U. of Penn) 5.7%; high, 6.6% U. of Ga.).

    For 1978 inflation based upon the CPI figure was 9.018%. Roc's in M*Vt projected 9.0%

    The importance of Vt in formulating - or appraising monetary policy, derives from the obvious fact that it is not the volume of money which determines prices & inflation rates, but rather the volume of monetary flows, M*Vt, relative to the volume of goods & services offered in exchange. And the importance of Vt is demonstrated by the historical fact that it has fluctuated 2.5 times as widely as the primary money stock over a corresponding 50 year period.

    The impact of the expiration of the FDIC’s unlimited transaction deposit insurance on the economy lead to a May 2013 “taper tantrum”. But no such taper tantrum was forthcoming when QE3 actually ended. Why? Because savings were forced out of the payment’s system when the FDIC’s insurance coverage was reduced (the theory that was responsible for my stock market prediction for a “market zinger”, when previously during the GR, no such impetus was forthcoming).