“Global Spillovers and Economic Cycles”

That was the title of an illuminating EABCN/CEPR/Banque de France/European University Institute conference I attended last week (organized by Philippe Bacchetta, Laurent Ferrara, Jean Imbs, and Massimiliano Marcellino). The program is here, and papers here.


The conference call described the focus:

In an increasingly integrated global economy, assessing the propagation of shocks is becoming of major interest to the international economic cycles analysis. It is indeed challenging to evaluate the impact of certain types of shocks on global business cycles and to disentangle the various transmission channels, such as trade flows, financial linkages or confidence effects. This conference will focus on empirical and theoretical contributions providing an assessment of various spillover effects at a global level and new perspectives on structural analysis, forecasting and economic policy design and assessment.

The conference papers provided a refreshing combination of advanced theory and econometrics and relevance to current policy concerns. All the papers were extremely interesting (I certainly learned a lot — there’s no way to keep up with all the developments, even when restricting oneself to international finance). I have to say it was also exhilerating to talk to people who believe in measurement, believe in the use of statistics, and who consider the possibility that policy could be welfare-enhancing (those following policy debates in the US know that not all people adhere to such views [1] [2]).


Lucrezia Reichlin (LBS) discussed the challenges of modeling of how central bank balance sheet measures affect output and other macrovariables. She highlighted the fact that using consolidated data were unlikely to be very useful in assessing what happens in the financial sector. Unfortunately, there’s no paper — although some of the material is available in published papers.


“International Spillovers: Real and Financial Channels,” by Ayhan Kose (International Monetary Fund), *Christopher Otrok (University of Missouri) and Eswar Prasad (Cornell University), is a must-read. In his presentation, Otrok elaborates on the decomposition of business cycles into global, regional and country specific factors, by further decomposing the global factor into spillover effects.


Institutional Investors Flows and the Geography of Contagion, by Damien Puy (European University Institute), used an incredibly detailed dataset of thousands of funds, to investigate the behavior of investors, by decomposing into global, regional and country-specific funds. One of his more notable findings is that the country-specific or idiosyncratic components of bond and equity flows is relatively small, suggesting that “push” “pull” factors are fairly unimportant [and “push” factors quite important] — mdc 6/5 . This was such a remarkable finding that I recommended that this finding be subjected to lots of robustness testing, in my discussion.


Perhaps the most provocative papers was The Puzzling Change in the International Transmission of U.S. Macroeconomic
Policy Shocks
, by Ethan Ilzetzki (London School of Economics) and Keyu Jin (London School of Economics). From the paper:

[T]his paper raises two challenges to existing knowledge on spillovers. First, we document a dramatic shift in how
U.S. monetary and fiscal policy shocks are transmitted to other economies. In an earlier period prior to 1984, a contractionary monetary shock in the U.S. lowers output of the 8 largest high-income economies in the world (ex-US), and appreciates the US dollar in both
nominal and real terms against these currencies. This accords with conventional views. In the sample post-1984, in which among the largest developing countries are also included, a contractionary monetary shock now raises output overseas and depreciates the U.S. dollar in real and nominal terms. Fiscal policy shocks have witnessed a similar metamorphosis: while unexpected fiscal expansions appreciated the dollar in both nominal and real terms in the earlier period, it depreciated the dollar in the later period. A natural question to ask is whether this qualitative shift in the nature of spillovers owes to changes in the foreign monetary policy reaction, and the answer is decidedly no: policy responses overseas are broadly
similar across these two sample periods.


The second challenge is that the responses to U.S. monetary policy shocks in the latter period have become somewhat of an empirical conundrum. Exchange rate and output effects in foreign economies are opposite of predictions from standard theories. While aptly capturing
many aspects of reality in earlier periods, these theories have a hard time describing recent patterns. There is reason to believe that an altogether di¤erent channel of transmission from the ones manating from existing models is required to understand even the basic effects on output and exchange rates. With a radical shift in the nature of transmission comes along a need for rethinking of the theoretical reference point on which policy prescriptions and welfare analysis are invariable based.

Ilzetzki and Jin attempt to rationalize these results by appealing to modifications to the theoretical framework (assumptions regarding pricing behavior, elasticities of substitution, etc.). Some of the commentators (myself included) conjectured that the use of bilateral models (e.g., US versus other countries) might be problematic, especially as other global factors might have become important in the latter period — including large purchases of government bonds by East Asian and oil-exporter central banks.


Other papers presented during poster sessions ranged in topic from elasticity optimism (Corbo and Osbat) [poster] and whether a Taylor rule is appropriate in a currency union characterized by nominal rigidities and financial frictions (Bhattarai, Lee, Park) (it isn’t).


Unfortunately, not all the papers are online, but over time more may be uploaded.

3 thoughts on ““Global Spillovers and Economic Cycles”

  1. Jeffrey J. Brown

    Here is a composite summary of recent trends in US oil & gas production in the global oil export market.
    The implications of the extrapolation of the combined Brazil +Iraq ECI (Export Capacity Index) ratio shocked even me, to-wit, that they would collectively approach zero net oil exports in only 10 years.
    How many times have we read about the bonanza of oil riches that Brazil + Iraq represent?
    Recent Trends in US Oil & Gas Production in the Global Oil Market
    ExxonMobil put the annual decline rate from existing wellbores in the 4%/year to 6%/year range a few years ago. For the sake of argument, let’s assume that the decline rate from existing US oil wells was about 5%/year in 2008, when the US hit the post-hurricane low production point of 5.0 mbpd (Crude + Condensate, EIA).
    Let’s assume that US Crude + Condensate production averages 7.5 mbpd in 2013, and let’s make (in my opinion a conservative) assumption that the decline rate from existing US wellbores is about 10%/year this year, as an increasing percentage of US production comes from high decline rate shale/tight plays.
    At a 10%/year decline rate, in order to simply maintain a production rate of 7.5 mbpd out to 2023, the US oil industry would have to replace the productive equivalent of every single oil field in the United States of America–everything from Thunder Horse in the Gulf of Mexico, to the Eagle Ford Play, to the Permian Basin, to the Bakken Play to the North Slope of Alaska.
    Or let me put it this way, at 5%/year decline rate, in 2008 the US lost 250,000 bpd per year due to declining production. At a 10%/year decline rate and a production rate of 7.5 mbpd, we would lose 750,000 bpd this year due to declining production.
    In other words, a 50% increase in net production + an increase in the decline rate from 5%/year to 10%/year would lead to a 200% increase in annual volume of oil lost to declining production from existing wellbores.
    Assuming an annual loss about 750,000 bpd from existing wellbores, the gross increase in production would have to exceed 750,000 bpd in order to show a net increase in production. For example, let’s assume that we average 7.5 mbpd in 2013, and let’s assume that we lose 750,000 bpd (0.75 mbpd) from existing wellbores this year. In order to show a net increase of 0.25 mbpd from 2013 to 2014 (from 7.5 to 7.75 mbpd), the industry would have to show a gross increase in production of one mbpd, which would be the production from new wells in 2014 that were not producing in 2013.
    If you add it all up, assuming a 10%/year decline rate from existing wellbores, in order to show a significant net increase in production, the US oil industry would have to put on line, over a 10 year period, the productive equivalent Saudi Arabia’s 2005 crude + condensate production of 9.6 mbpd, which is also the US production peak that we saw in 1970.
    This is why Peaks Happen, and it’s why production declines are inevitable. On the upslope, new oil wells can offset the declines from existing wellbores, but with time, new oil wells can no longer offset the increasing volume of oil lost to production declines And of course the overall decline rate from existing US gas wells is almost certainly even higher than for oil wells.
    We are currently averaging about 66 BCF/day in dry natural gas (NG) production in the US (EIA). If we assume a 20%/year decline rate per year from existing NG wellbores, the industry would have to put online the productive equivalent of current US dry NG production over the next five years, in order to maintain a production rate of 66 BCF/day.
    So, based on a 10%/year decline rate for oil wells and a 20%/year decline rate for gas wells, in order to maintain a crude oil production rate of about 7.5 mbpd and a NG production rate of 66 BCF/day, in round numbers the industry would have to add the productive oil equivalent of one new Bakken play every year and the productive gas equivalent of 2.3 Barnett Shale plays–every single year, year after year.
    Globally, the dominant trend we are seeing is a post-2005 decline in Global Net Exports of oil (GNE), as the developing countries, led by China, so far at least have consumed an increasing share of a post-2005 declining volume of GNE. Of course, this means that developed net oil importing countries like the US have to make do with a declining share of a declining volume of GNE. And the US is still dependent on imports for the majority of crude oil process in US refineries. The post-2005 decline in GNE, combined with increasing demand from developing countries were the primary factors that contributed to global (Brent) crude oil prices more than quadrupling from $25 in 2002 to $112 in 2012.
    An extrapolation of the 2005 to 2011 rate of decline in the ratio of GNE to Chindia’s (China + India’s) Net Imports (CNI) suggested that some time around 2030, China & India would theoretically consume 100% of GNE. Recent released EIA data confirm the trend, at least through 2012, as the GNE/ECI ratio fell from 9.5 in 2005 to 5.0 on 2012.
    For more information on global oil exports, you can search for: ASPO + Export Capacity Index.
    For a concrete example of the Export Capacity Index (ECI) concept works, consider two countries that are widely considered to be critically important sources of future crude oil production: Brazil and Iraq.
    If we extrapolate the 2008 to 2012 rate of decline in Brazil + Iraq’s combined ECI ratio (the ratio of liquids production* to consumption), they would collectively approach zero net oil exports in about 10 years. Note that their combined net exports fell from 2.0 mbpd (million barrels per day) in 2008 to 1.8 mbpd in 2012 (EIA).
    Or in other words, the increase in Iraq’s net oil exports from 2008 to 2012 could not even offset the decline in net exports from Brazil (as Brazil slipped into net importer status, even if we count biofuels as “oil” production).
    *EIA data, production = total petroleum liquids + other liquids (mostly biofuels in the other liquids category)

  2. Johannes

    Hi Menzie, hope you had a nice time in Paris, sitting at the Champs-Élysées in a coffeehouse.
    Yes indeed, the “push” factor is more important than “pull”, hope my investment will work out properly.

  3. Bruce Carman

    Sum of annual change rates of non-farm employment and wages less productivity vs. the annual change rates of real final sales and real imports:
    http://research.stlouisfed.org/fredgraph.png?g=jgg
    The administration will be required to “manage” the deflator, inventories, deflator, and import prices to avoid reporting an annualized real GDP contraction hereafter.

Comments are closed.