Opinion

Felix Salmon

Chart of the day: GDP growth and volatility

By Felix Salmon
June 8, 2011

Alan Taylor has one of those op-eds today which is crying out for a chart. He’s comparing DMs, developed markets, to EMs, emerging markets*:

A central macroeconomic indicator, gross domestic product growth and its volatility, speaks to the reversal of fortune. Circa 1970, EMs exhibited high mean and high variance relative to DMs; but after 1980 this risk-reward combination evaporated as EMs suffered a lost decade. But from the 1990s EM growth picked up and volatility moderated; DM growth slowed and, after the crisis, volatility spiked.

What does this look like in graphical form? Here’s the chart, which is taken from a recent Morgan Stanley research report:

emdm.tiff

As with any chart where time isn’t on the x-axis, this is very interesting and also takes a little time to fully comprehend. Basically, if you look at the DM line — that’s the dark blue one, representing developed economies — you see average growth rates declining steadily all the way from 1980 through the present day. That was something we accepted during the Great Moderation, kidding ourselves that even though we weren’t growing as fast as the emerging world, at least we were growing with very low volatility.

Then the crisis hit, and GDP volatility hit levels unprecedented in recent economic history, at least in the developed world, with standard deviations moving up towards 3 percentage points.

Meanwhile, the emerging world has a much higher mean growth rate — around 9%, compared to the developed world’s 3.5% or so — which has been rising steadily in recent years even as volatility in that growth rate has declined towards developed-world levels. Emerging-market GDP volatility isn’t yet lower than developed-world GDP volatility, but it’s close. And given the tiny difference in volatility, the enormous gap in absolute size is all the more striking.

This chart, in a nutshell, explains why American companies are all falling over themselves to make inroads in China, and other emerging markets. They’ve had strong growth for a while, but now they’re pretty stable, too. Here’s Taylor, again:

What would have been your reaction, circa 2006, to someone peddling the following forecast? There would soon erupt the worst, synchronised global financial crisis in 80 years, or possibly ever. World trade would start to collapse as fast as in 1929-31. Many developed markets (DMs) would experience deep recessions, the costliest banking crisis ever and would stagger toward fiscal calamity. But – guess what! – emerging markets (EMs), after a brief panic, would sail on unscathed. They would have no significant crises: currencies, banks and fiscal positions would retain stability. A two-track recovery would take EM growth on to a trajectory away from a troubling slump in the DM world.

That world might not have seemed likely five years ago. But it’s the world we’re living in today. And it’s the base case scenario for the foreseeable future, too.

*Update: In response to inboulder, in the comments, here’s the list of the countries used, from page 11 of the PDF:

List of EM countries and ISO codes: Taiwan (TWN), India (IND), Indonesia (IDN), Korea (KOR), Malaysia (MYS), China (CHN), Singapore (SGP), Hong Kong (HKG), Thailand (THA), Brazil (BRA), Mexico (MEX), Peru (PER), Colombia (COL), Argentina (ARG), Venezuela (VEN), Chile (CHL), Russia (RUS), Poland (POL), Czech Republic (CZE), Hungary (HUN), Romania (ROU), Ukraine (UKR), Turkey (TUR), Israel (ISR), SA (ZAF).

List of DM countries and ISO codes: United States of America (USA), Germany (DEU), France (FRA), Italy (ITA), Spain (ESP), Japan (JPN), United Kingdom (GBR), Canada (CAN), Sweden (SWE), Australia (AUS), New Zealand (NZL), Austria (AUT), Belgium (BEL), Denmark (DNK), Finland (FIN), Greece (GRC), Iceland (ISL), Ireland (IRL), Luxemburg (LUX), Netherlands (NLD), Norway (NOR), Portugal (PRT), Switzerland (CHE).

Comments
7 comments so far | RSS Comments RSS

I’m curious what they mean by “GDP volatility” here (the MS report doesn’t state their methodology as far as I can tell). I’m guessing it’s something like the variance of a recent horizon of GDP growth, in which case any sudden shift in growth rates is likely to keep GDP volatility high for a few of years just as a matter of methodology. On the DM side, I wouldn’t read a ton into that; it only tells us what we already know, which is that we (DMs) had a nasty shock to our GDP growth rate.

Posted by absinthe | Report as abusive
 

John Joseph Wallis (University of Maryland) has written a brilliant paper on that matter. his conclusion is that high growth in good times is not as important for high living standards as we usually tend to think. How often a country is hit by economic downturns is what matters most, as well as how severe these crises are. Wallis: “Rich countries are not rich because they grow faster when they grow, but because they have fewer episodes of negative growth and shrink more slowly in times of crisis.”

I blogged about that paper last year: http://olafstorbeck.com/2010/05/17/shrin k-baby-shrink/

Posted by OlafStorbeck | Report as abusive
 

I couldn’t find the methodology for determining DMs and EMs, or which countries were included even. This seems vital to the Morgan analysis, doesn’t anyone have this?

Posted by inboulder | Report as abusive
 

I’m with inboulder in questioning the definition of volatility. Stdev is a lousy volatility metric, as you can see if you just ask yourself which of these two sequences seems more volatile:

+2, +2, +2, +2, -2, -2, -2, -2

+2, -2, +2, -2, +2, -2, +2, -2

Stdev thinks they’re identical. You really need a metric that takes account of self-correlation, the degree to which the recent past can be used to project the future.

Obviously the past can never be used to COMPLETELY predict the future, but if you can at least use it to make an educated guess, that means you have low volatility.

Posted by Auros | Report as abusive
 

I’m still curious about the formula of GDP Volatility. Is the formula lagged variable from spesific period? If we count the GDP volatility of country in period, does it mean we just have one value or variate value of years?

Posted by Vinland | Report as abusive
 

I’m still curious about the formula of GDP Volatility. Is the formula lagged variable from spesific period? If we count the GDP volatility of country in period, does it mean we just have one value or variate value of years?

Posted by Vinland | Report as abusive
 

I’m still curious about the formula of GDP Volatility. Is the formula lagged variable from spesific period? If we count the GDP volatility of country in period, does it mean we will have one value? or variate values of years?

Posted by Vinland | Report as abusive
 

Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/
  •