What You Export Matters: Does it Really?

In 2007, Hausmann, Hwang and Rodrik (HHR) demonstrated that export specialization patterns have important implications for economic growth. The authors developed an indicator of income level linked to the country’s exports they called EXPY and showed that higher values of the indicator lead to higher subsequent economic growth. The present paper tests whether HHR’s conclusions are valid even in times of economic crisis and rising prices of primary commodities, using data from 2004-2013. We show that, in the aggregate, higher values of EXPY are connected with faster economic growth. However, the relationship is much more statistically significant in countries that focus heavily on exporting primary commodities than in other countries. This implies that the rising prices of primary commodities in the last decade have altered the traditional link between export sophistication and economic growth. As a result, we argue that EXPY is not a good predictor of future economic performance when the prices of primary commodities are unstable. Policy makers must be aware that, while what countries export is important, it is equally important when they export it: in times of stable prices of primary commodities, a focus on the export of sophisticated goods generates higher economic growth in the future. In times of rising prices of primary commodities, however, the effects can be exactly the opposite.


Introduction
The question of why some countries are rich and others are poor has been present in the economic literature for decades. Multiple theories have been devel-tion patterns on the basis of its export structure and can also be interpreted as a measure of the quality of the country's export basket. HHR proved that "countries that latch on to a set of goods that are placed higher on this quality spectrum tend to perform better" (p. 24). Their results are based on data for the period 1962-2003. The aim of the present paper is to test the link between export quality and growth using the most recent data. We come from the assumption that dot.com crisis, the impact of the World Trade Center attacks and the great financial and economic crisis have significantly changed the global economic environment (Lipkova, 2012). As a result, it is doubtful whether "old conclusions" are still valid. We show that events of the last 15 years have altered the traditional link between export sophistication and economic growth and that EXPY is not a good predictor of future economic performance when the prices of primary commodities are unstable. Moreover, increasing quality of exports does not guarantee rising terms of trade.
In addition to theoretical contributions, our research has important policy implications. In demonstrating that higher export productivity does not necessarily accelerate growth and enhance the terms of trade, we show that countries can improve their economic performance even when focusing on primary commodities. In times of economic crisis, their exports appear to be more advantageous than the export of goods placed higher on the quality spectrum.
The remainder of the paper is organized as follows.
Section 2 presents the idea behind the EXPY indicator and offers a brief literature review. Section 3 details and justifies the methods applied in this paper.
Section 4 tests the hypothesis that higher productivity of exports leads to higher economic growth. In section 5, a link between productivity of exports and net barter terms of trade is explored. The final section concludes.

Literature review -the EXPY concept
In a widely cited paper, HHR (2007) introduced a measure of productivity level associated with a country's exports that they called EXPY. The basic idea behind the concept goes back two decades to Michaely (1984), and it captures the average level of income generated globally by the commodities the country exports. High values of EXPY indicate that the country's export patterns are dominated by high-income products, whereas low values mean that exports are dominated by lowincome goods.
To calculate EXPY, a PRODY index must be constructed first. PRODY is "a weighted average of the per capita GDPs of countries exporting a given product, and thus represents the income level associated with that product" (p. 9). For each product k it equals where the numerator is the value-share of the commodity in country j's overall export basket and the denominator aggregates the value-shares across all countries exporting the good. Following HHR, both x and X in the equation stand for exports, the difference being that X represents the total exports of a country j, while x is the country's exports of a product k. Y stands for per capita GDP. A similar index was independently developed by Lall, Weiss and Zhang (2006), but never gained wide use.
In 2013, the product items with the highest associated income levels were plastic waste, parings and scrap, watches, clocks, fur skins and jewelry (table 1).
On the other end of the list, crude fertilizers, nickel ores, tobacco and natural abrasives had the lowest PRODY values.
The EXPY of a country is defined as a weighted average of the PRODYs for that country, where the weights are the value shares of the products in the country's total exports:  (Minondo, 2010b). He showed that there is a positive link between an exports' productivity and growth at a regional level. Saadi (2012)  higher export productivity is found to increase subsequent economic growth, it should also be the case that it enhances the terms of trade.
Other notable recent studies include Nyarko's (2013) application of the model to Sub-Saharan Africa, Bernatonyte's (2011) analysis of the export productivity of the Baltic nations, Weiss's (2010) discussion of how changes in trade structure affect growth and Jarreau and Poncet's (2011) study of the regional variation in export sophistication in China.
While the EXPY indicator has become widely used, it is not without critics; it rests on the uncertain assumption that more advanced countries produce sophisticated goods and conceals diversity in the quality and subtypes of goods (Wang, Wei & Wong, 2010). "what countries export may be very different from what they actually contribute to the production process;" hence, the index fails to capture trade with processing goods. This criticism has to be kept in mind when drawing conclusions.

Methodology
Our research builds on HHR (2007) and Saadi (2012) and tests their conclusions in the period 2004-2013.
We examine the links between export quality (proxied by EXPY) and economic growth and between export quality and net barter terms of trade using pooled ordinary least squares regression analysis with time-specific effects and heteroskedasticity-corrected OLS. The

EXPY and terms of trade
An important indicator of a country's trade performance is the terms of trade index, a ratio of export prices to import prices. An improvement of a country's terms of trade means that it is able to exchange the same amount of exports for a higher amount of imports. It can be expected that rising export sophistication should be linked to rising terms of trade. Saadi (2012), however, has shown that increase in the sophistication of developing countries' exports is accompanied by a deterioration of their terms of trade.
He offers several explanations of this counterintuitive result, including excess production capacity, technological upgrading without simultaneous advances      (Table 4). It appears that the majority of the countries with negative correlation are highincome developed nations. On the contrary, countries with a high positive correlation include numerous primary commodities dependent developing countries, which contradicts Saadi's results.
The preliminary analysis suggests that there are important differences in the effect of export sophistication on terms of trade among different groups of countries. It appears that these differences may be connected to the countries' per capita incomes and their development status. We will test this claim using pooled OLS and heteroskedasticity-corrected OLS regression analyses with time specific effects. Following Saadi (2012), in addition to the log of EXPY, we will also use log of imports divided by GDP as a control variable. This is based on the empirical assumption that countries with a higher share of imports on GDP (and hence higher trade openness) are more dependent on foreign suppliers and their demand is inelastic to price changes. As a result, a negative sign of the coefficient is expected.
Regressions using data for all countries and territories show a statistically significant positive link between EXPY and terms of trade (Table 4,    Our test has shown that, on the aggregate level, the link is still valid even in the current period of global economic crisis and ensuing changes. The relationship is stronger and much more statistically significant in countries that focus heavily on the export of primary commodities than in other countries. Similar results have been achieved when regressing the net barter terms of trade on export sophistication. While primary-commodities exporting countries and mediumincome nations display a positive relationship between the variables, it is negative for non-resource-based low-income countries, effectively meaning that a rise in export sophistication deteriorates their terms of trade. This is a paradoxical result that can be explained by the rapid increase in prices of primary commodities in the period studied and by the fact that an absolute increase in export sophistication does not mean that relative export sophistication has increased as well. As a result, it appears that EXPY is not a good predictor of future economic performance when the prices of primary commodities are unstable.
To conclude, what is the solution to the question asked in the title of this paper? Our answer is very dif- matter what you export; it mainly matters when you export it. No absolute truth exists and the crucial task of policy makers is to make the right export decisions at the right time.