It is not news that Latin America’s exports depend heavily on commodities. Particularly, The Pacific Alliance (TPA) countries exports are also concentrated in commodities, hence have a significant exposure to price volatility of these products, affecting direct and indirectly key macroeconomic variables such as economic growth, inflation, interest rates, etc. According to World Bank data, Chilean exports of copper represent 55% of total exports; in Peru copper and gold together are 41.3%; meanwhile in Colombia exports of oil and coal stand by 51.5%. The story in Mexico is a bit different, exports are much more diversified and maybe less reliant on commodities. The importance of these products in total Mexican exports rounds 18%. Nevertheless, commodities exports are important because of fiscal revenues, the Federal Government of Mexico gets 20% of total income from oil production.
Several authors have studied the effects of commodity shocks in Latin American economies. For example, Gruss (2014) has estimated the effects of changes in commodity prices – driven by Chinese demand – on economic growth of Latin American countries. The results show that Latin American business cycles are heavily reliant on commodity prices. In response to a drop of 1% of China’s GDP, and consequently a drop in commodity prices of 3% (even 8% in some countries), there is a downfall of 0.5% on average in Latin American countries’ GDP.
Because of its trade and investment linkages, it seems interesting to decompose the contribution of commodity shocks, not just price shocks but also commodity sector productivity shocks, in the performance of TPA countries in order to study its relevance for this group. Here, I use the Model for Analysis and Simulations (MAS) developed by Medina and Soto (2007) and developed and adapted by Cerda R and Gonzales-Carrasco L (2017), and tailored for the Chilean economy. The MAS is a Dynamic Stochastic General Equilibrium model and its structure is considered applicable for other economies of the TPA1. For each country, parameters of the model were partly calibrated, and others estimated with Bayesian methods2.
Commodity price shocks
As a first step, I describe the effects of commodity price shocks . For the Chilean, Colombian and Mexican cases I found that one standard deviation shock in the commodities price index induced an expansion of production, a real exchange rate appreciation and a reduction in the nominal interest rate. The shock causes a downturn in inflation that is directly related with a fall in imports prices due to real appreciation. These first round effects on inflation generate a further decline in the nominal interest rate. The expansive effect on production lasts around two years, and then its magnitude diminishes significantly because of the real exchange rate appreciation and the consequently lost in competitiveness. The effects on inflation and on the nominal interest rate last until a year after the shock, even though effects over these variables are small.
In a similar exercise we can obtain a different lesson from the case of Peru. For this country, the model was adjusted by removing the fiscal rule from the baseline model because Peru doesn’t have an explicit structural budget balance rule, that corrects by extraordinary commodities and cyclical revenues, in the same fashion of the other TPA countries. Then the fiscal policy is more expansive in response to commodity shocks. In fact, when we compare the magnitude of first round effects of the commodity price shock on production, I found that Peru has the strongest response among Alliance members (GDP rises about 1.5% in deviation from its steady state versus 0.05%, 0.02% and 0.2% for Chile, Colombia and Mexico, respectively).
The growth in government revenues due to an increase in commodity prices translates in a direct expansion of aggregate demand via government spending (through standard Keynesian channels). But this expansion varies in magnitude depending if there is a fiscal rule that controls the procyclicality of government expenditures or not. As mentioned above, in Peru the effects are greater than in other countries because of its procyclicality in expenditure. This last effect pushes up prices and consequently the nominal interest rate, and as a second round effect, I found that after one year after the shock the growth rate in Peru will be below its steady state value, and just after two years will return to its steady state. This translates into more volatility of Peru’s growth rate.
Commodity production shocks
In the case of commodity production shocks the results are similar across countries, not just in form but also in magnitude. As a first round effect, an increase in commodity production generates pressure over inflation and consequently over the nominal interest rate. These two last effects in addition with a first round of real appreciation, have a negative effect over economic growth. I found that around the fifth quarter after a commodity production shock, economic growth falls below its steady state. Once again, Peru seems to be the most volatile country in this group, but the magnitude of the deviation from its steady state doesn’t differ significantly in absolute terms respect to others (0.015% versus 0.0086%, 0.0084% and 0.012% for Chile, Colombia and Mexico, respectively).
Contribution of shocks
Even though the previous analysis sheds some light about the vulnerabilities faced by Pacific Alliance countries respect to both price and production shocks3, now I turn to the historical decomposition of the growth rate of each country in order to analyze the contribution of commodities price shocks. It’s important to mention that the model has fifteen different shocks that were divided in four groups: 1) supply4 , 2) demand5, 3) external6, 4) commodity7. Figure 1 shows the historical decomposition of growth rate of real GDP for TPA countries. The first finding from this exercise is the importance of the commodity shocks to explain the ups and downs of GDP. In the last years the effect of commodity related shocks has had heterogeneous impacts amongst TPA countries.
On the one hand, Chile shows negative contributions in the last quarters of the sample due to the pronounced drop in copper prices. These shocks account for 45.4% of the growth rate on average between 2013Q1 – 2016Q3. On the other hand, the effect of the commodity related shocks in Peru are quite different than the Chilean case. Even though copper is the most important commodity in Peru’s exports bundle, commodities production is more diversified. Copper represents 42% of commodities exports in Peru, gold accounts for 31%, oil 7% and the 20% remaining is divided between other like zinc and lead. The results show that since 2013 gold, zinc and lead have registered increases, and these commodities represents 51% of total commodities exports. Thus diversification in commodities production explains why commodity shocks had opposite effect on historical growth rates in Peru.
Figure 1:Historical decomposition of quarterly GDP growth rates (y-o-y) of TPA Countries. 2013Q1 -2016Q3
Finally, there exist other transmission channels for the remaining countries. For Mexico, the ups and downs are heterogeneous because the behavior in oil price was erratic since the end of 2014. Despite the fact that government expenditures aren’t as procyclical as in the case of Peru, revenues coming from commodity production are more important for the government because oil production is a state monopoly. Then the intake for government is much more important than the other members. The positive shock in commodities implies that the effect in fiscal revenues is larger than in other country members, then the government expenditure has an important aggregate demand impact. For the Colombian case the changes in commodities prices and production don’t have a significant weight in economic growth as in the other countries. Oil and carbon together represents just 7.3% of the GDP, compared with the production of commodities as share of GDP in the other countries (14% in Peru, 12% in Chile and 10.1% in Mexico).
Overall, the effect of commodity shocks on TPA countries is mixed, because of diversification of exports and due to institutional differences. It’s important to emphasize the MAS might not be to stylized to capture all factors that explain the historical decomposition in each TPA country. Nevertheless, I consider that this first exercise is a good approximation to understand the effects of commodities shocks in these economies.
 All Pacific Alliance members are commodity exporter economies. Additionally, they have similar monetary and fiscal policy arrangements. They all follow an inflation targeting scheme. They have a similar fiscal rule, except for Peru. Also, they are countries with high trade openness, with a trade openness index of 54,2% vis-à-vis average 42% for the rest of Latin American countries ↩
 The exogenous shocks of the model have a AR (1) process, the Bayesian calculations were made for the persistency parameters and for the standard deviation of the white noise innovations of the shock processes. ↩
 In this case, the standard deviation shock is specific for each country, since each country have an specific commodities prices index. ↩
 Transitory and permanent productivity shocks, commodities production shocks, investment adjustment costs, depreciation shocks and labor supply shock. ↩
 Preference shock, government expenditure shock and monetary policy shock. ↩
 Oil Price shock, foreign production shock, foreign interest rate shock, foreign inflation shock and imports price shock. ↩
 Commodity price and production shocks. ↩