Best of Times 2?

In a previous blog I argued that the common perception about developing countries being the beneficiaries of record commodity prices (“the best in generations”) had to be significantly qualified: world real prices for food, beverages and agricultural raw materials do not look that great in real terms; fuels and metals, however, are a very different story, with real prices close to, or above, historical highs. The “best of times” story jointly argues that the world growth rate is at record highs as well. Such benign world environment is credited with the resumption of growth in LAC, but usually the argument comes also with expressions of concern because even during these (supposedly to be) great times, LAC is growing less than other developing regions, including Africa. But is the world really growing much faster than in the past? As in the case of commodities, it depends very much on what indicator you are looking at.


The main reason about the perception of record high growth rates at the world level is the use of IMF data that aggregates country growth rates applying PPP exchange rates to determine the weights of each country in the total. The WEO data base also includes an aggregation using market exchange rates, and the resulting world growth rate is far lower (see next Chart; it shows both growth rates since the 1980s, the period covered by the WEO data base; PPP growth in blue; market exchange rate in dashed red). The reason for that difference is that China and many of the fast growing developing countries have larger shares of world GDP at PPP than at market exchange rates. Notice that I am not saying that an individual country’s growth rate differs when measured in PPP or market exchange measures: the individual growth rate is the same, measured in local currency units and related to some local base period. What changes is the weight in each aggregation. For instance in 2004, China had a share of the world GDP of 5% at market rates (World Bank/WDI) and 13% at PPP rates (IMF/WEO). Therefore, China’s growth rate in that year (10%) added 0.5% or 1.3% to the world growth rate, depending on what share was utilized.


The WDI data base of the World Bank has a longer time series for world growth based on market exchange rates, starting in 1960. In the database for the WEO report the IMF has PPP-based rates since the 1970s. The next Table (with World Bank data for market exchange rates and IMF for PPP rates) shows world growth averages for different growth periods, including the recent one from 2003-2006.


At market rates, world growth during the current growth period is only marginally better than the 1994-2000 period (which had to weather a series of crises in developing countries), and is below any other growth period. The “best of all times” story appears true only when looking at PPP-aggregated growth rates. But LAC growth appears more correlated with world growth aggregated at market exchange rates: the next two Tables show simple regressions between LAC growth rate and world growth rates aggregated at markets exchange rates (first Table), and PPP rates (second Table) (period 1990-2005). It is clear that R2 and R2-adjusted are far larger with world growth aggregated at market rates, and that the statistical significance of the coefficients also differ substantially: significant at better than 10% in the case of market rate aggregation, and not significant even at 20% in the case of PPP aggregation (both equations pass all usual residuals tests; not shown; if anyone is interested I can send them).



Of course, these correlations are only a first approximation to the debate. But there are at least two reasons why it seems reasonable that LAC growth appears more correlated with world growth aggregated at market exchange rates rather than at PPP rates. First, growth impulses are transmitted at market exchange rates and not at PPP rates. Each growth indicator has its specific uses, however: if I am interested about how the world as a whole is doing, but I am not interested about cross-effects, the PPP measure may be fine. But the “best of times” argument is about how world growth is affecting a specific country or subset of them. In that context, transmission mechanisms are crucial, and, therefore, it seems more adequate to use growth measures based on market exchange rates.

Second, for LAC it is obviously not the same if world growth is supported by expanding domestic absorption in the US than if it is based on growing aggregate supply by countries with export-led strategies that may be displacing LAC’s products in the domestic markets of deficit countries (and even in LAC’s internal markets). Market exchange rates give a greater weight to the former type of growth and PPP to the latter one. Of course, at the world level aggregate supply and aggregate demand (except for some statistical error) should balance, but the composition matters for the type of analysis associated with the “best of times” argument. A correlated point is that, for LAC, deceleration in domestic absorption in the US would not necessarily be compensated by growth in aggregate supply in, say, Asia, whose export-led growth may well compound the negative direct effect of a US slowdown on LAC (for instance, even if a LAC country sells metals to that region and does not export many consumer goods to the US, such country, if US demand continues to soften, should expect Asia to look for new markets for their consumer goods, eventually putting pressure on that country’s producers in their own domestic markets). Of course, the impacts on LAC would be different if China and other countries with large trade surpluses of products that are competitive with LAC’s production, decide to stimulate their domestic absorption more in line with their aggregate supply.

But, until that happens, it seems more adequate to look at world growth aggregated at market exchange rate if you want to discuss its impact on LAC. And using the proper indicator, world growth does not seem that great compared to the previous decades.

So far I looked at commodity prices (previous blog) and world growth (this one). The “best of times” argument also mentions very low interest rates as part of the triangle of good news. Indeed, they are low in real terms when compared to the 1980s and 1990s, but not against the levels of the 1960s and 1970s. So, in this regard, the whole argument depends on the period covered to define “the best of times.” But still, this is an area where, along with prices for fuels and metals, the story about a very supportive world environment has more credibility.

All in all, my conclusion is that the hype about the “best of times” story for LAC requires no trivial qualifications, both in the variables, countries, and time periods considered. That would give us a better starting point to understand the current performance of the region.

6 Responses to "Best of Times 2?"

  1. Anonymous   July 3, 2007 at 7:56 pm

    Very interesting arguments. But do you expect a protein price shock? Many in the markets do. And a variety of food commodity prices have recently sharply increased

  2. Henry Smith   July 4, 2007 at 6:45 am

    When you say world growht is lower than the 1960s do you control for productivity?

  3. JPN   July 4, 2007 at 6:51 am

    The t-student from both regressions is not signicant. The R2 are too small, in both cases. Why are you presenting them to boost your case given that they are not fostering your argument? Even if one is worst than other, both are bad pieces of evidence.

  4. Vitoria Saddi   July 4, 2007 at 7:02 am

    Eugenio, Like your argument about the exchange rate instead of the PPP, as well the evidence. Important to look at the data. Question: do you think that there is more room for an increase in commodities?

  5. Guest   July 5, 2007 at 9:43 am

    Thanks for your comments  Anonymous and Vitoria: real prices for food products (including beef, butter, and soybeans) during this last world growth cycle have been below their previous peaks (always in real terms), even looking at the latest data included in the IMF/IFS database. That may well change in the future (see the study by OECD/FAO,3343,en_2649_201185_38893378_1_1_1_1,00.html)  Henry Smith: I am not sure what do you mean by controlling by productivity. All growth measures include a component of total factor productivity.   JPN: first my argument; then the econometrics.  My argument: I am trying to qualify the view that these are “the best of times” and that is the reason why LAC is growing. I looked at 2 growth measures. One measure (PPP-aggregated) seems to support the view that these are “the best of times.” The other (aggregated at market exchange rates, MER), does not show particularly high rates in this world growth cycle. The (very) simple regressions show that the MER-aggregated measure is clearly more correlated with LAC growth than the PPP-aggregated measure, and I try to explain why you should not use the PPP variable to define the world environment for LAC. I am not trying to have a full-blown explanation of LAC growth (you need many other economic and non-economic variables to do that).   Econometrics: the usual levels of significance are 1%, 5% and 10%. The t-statistic of the MRE-aggregated world growth is significant at between 5% and 10% (7.8% exactly), while the value for the PPP-aggregated measure is not significant even at 20%. This supports my conclusion about the relative relevance of those two variables.   Whether the R2 is too small or not has to be assessed against the type of variables and regressions considered. World growth is an I(0) variable, with relatively high volatility. The regression uses a single variable, and is run on an annual basis (not 3 or 5 year aggregates as in most of the growth regressions) without lags (or dummies). In that case you could argue that explaining about 20% of the variance is not that low. In any case my argument, as explained above, is a relative one, and the regressions support that (relative) point.

  6. Eugenio Diaz Bonilla   July 5, 2007 at 3:20 pm

    I did not realize I signed as Guest the previous comment. I wanted to use my name. Sorry