Micro-Level Analysis of Mexican Retail Markets and Their Response to Changes in Market Structure and Competition Policies

This paper develops the following price indicators to measure the relative efficiency (functioning) of markets: (a) price dispersion, (b) price volatility, and (c) price tansmission (speed, completeness, and symmetry). The paper uses these indicators to study trends and conditions of the outlet level in retail prices for common commodities sold throughout Mexico. The analysis examines price patterns for each indicator across commodities, regions, and time. The descriptive results indicate that although there is (expected) heterogeneity in the behavior of these indicators across commodities, location variables explain the most variation in the indicators. There are clear and persistent regional- and commodity-specific effects. Thus, the study concludes that Mexico is not one, well-integrated national market. The study tested whether changes in these indicators increased efficiency) have the expected correlation with measures affecting the functioning of markets. It considered changes in competition and entry of large retail stores in the local retail market. These changes affect market efficiency in the way theory would predict. The results suggest that these indicators are good measures of the relative efficiency (functioning) of markets. The findings also suggest that efforts to monitor markets using these indicators may be useful. For example, for policy makers who are concerned about the distributional effects of liberalizing trade, the indicators may predict where price impacts will be felt the most and by whom. In addition, the indicators provide preliminary information about relative competition levels, which may be helpful in saving the time and effort of the competition authorities and possibly making them more effective.


Introduction
Economic growth remains the main driver of the well-being of nations and even for the viability of their political systems. However, nations' growth trajectories are facing unexpected challenges in the wake of the recent coronavirus crisis. This calls for a renewed examination of the drivers of economic growth, especially with regard to factors pertinent to the present situation.
Individual nations, depending upon their comparative advantages and focus, are beginning to recognize the supply chain challenges posed by lockdowns in the wake of the coronavirus spread (see, for example, a related news report from U.S.-Mexico: https://www.bloomberg.com/news/articles/2020-04-28/lowly-water-heaters-get-ensnarled-incross-border-covid-19-fight; https://www.businessinsider.com/how-long-grocery-storeshortages-supply-chain-disruptions-could-last-2020-4). Not only are the supply chain disruptions making lives difficult for residents and firms, such disruptions, especially related to the medical and safety equipment, can prolong the current crisis, thus having a compound effect. 1 Even more challenging (and unknown) would be the ability of supply chains to run smoothly and in a coordinated and timely fashion when the lockdowns are finally lifted. Thus, ascertaining the possible impacts of disruptions under alternative scenarios should help planning and recovery, once the uncertainty is resolved. With increased globalization and specialization of production across nations, disruptions even in a single nation that is the source of crucial inputs can have adverse impacts on global supply chains (see https://www.weforum.org/agenda/2020/04/covid-19-pandemic-disrupts-global-value-chains/).
The global supply chain has been undoubtedly disrupted in the past, most recently during the 2008-2009 financial crash (Baldwin and Weder di Mauro (2020)). However, while lessons from the past would be instructive in present times, there are some qualitative differences this time around. At that time, it was more a factor of demand than supply, whereas the current crisis has impacted both supply and demand significantly. Baldwin and Weber di Mauro mention global value chains (GVCs) that require parts of products to be manufactured in different countries before being assembled in another nation.
COVID-19 initially impacted China, which is at the center of many GVCs, disrupting the supply chain. On the demand side, as the lockdown continues and consumers' physical spending is decreasing, global demand continues to fall. Global oil prices are already showing the impact COVID-19 has on supply and demand (Fernandes (2020)). Worldwide oil demand has decreased due to the global lockdown and other variables related to COVID-19, the oil prices have fallen over 50% from the start of 2020, due in part to a surplus of supply. Many companies (https://www.bbc.com/news/business-52226236) are planning to cut back production, to raise and keep prices up as demand continues to fall. This may have long term impacts when the lockdown lifts and supply shortage issues continue. However, the end period if the current crisis is uncertain, which makes planning all the more challenging.
Supply chain performance is crucial to the smooth functioning of economies, and glitches can create bottlenecks with adverse implications for economic productivity and growth. However, there are many dimensions to supply chains and their coordinated, often sequential, functioning is crucial in the smooth on-time delivery of products to consumers and inputs to businesses.
Whereas realization of the reliability and performance of the supply chain is not new, 2 supply chain reliability has attracted renewed attention in the wake of the recent COVID-19 crisis. Governments around the world have been scrambling to ensure adequate supplies of medical equipment, avoid excessive transport disruptions and ensure reliable food supplies. Besides short run inconveniences, supply chain disruptions have longer-term effects on economic growth. For example, when tires are not delivered to manufacturers on time, trucks do not get built or are not usable, so the investment in the other dimensions of truck production (chassis, instrumentation, etc.) become unproductive.
Key questions addressed in this research are the following: • What is the impact of supply chain disruptions on economic growth?
• Is the impact of the different dimensions of supply chain performance on economic growth similar? • What is the projected impact of supply chain disruptions-e.g. to those related to COVID-19-on economic growth?
The findings will be instructive, especially with the worldwide supply chain challenges faced in the wake of unexpected developments posed by the coronavirus crises. This paper uses data over a recent decade for over 130 nations to examine the relative effects of the different aspects of supply chain logistics on economic growth. Placing the empirical framework in a standard growth model, results show that improvements in the aggregate supply chain logistics performance yield positive growth dividends. These dividends are also present in most of the disaggregated dimensions of logistics, with the performance of the timeliness of shipments having the greatest positive impact on growth. A simulation exercise discussed some supply chain disruption scenarios and their growth implications that could provide useful in nations planning for the challenges posed by the COVID-19 and other unexpected crises.
The structure of the rest of the paper includes the literature and the model in the next section, followed by data and estimation, results, and conclusions.
The importance of economic growth has garnered a lot of interest by economic scholars over the years, with different studies examining the effectiveness of various drivers of growth across countries (see Barro and Sala-i-Martin (2003), Gollop et al. (1987), Jones (2016), Levine and Renelt (1992), Mankiw et al. (1992)). However, within this body of research, the impact of supply chain performance, especially along its various dimensions, on economic growth has not been considered. The present research aims to fill this gap. Equally important, we provide some estimates of the impacts of unexpected supply chain disruptions on economic growth. This exercise gains added importance in the wake of the current COVID-19 crisis that has strained supply chains worldwide.
Given the recent (and unexpected) nature of the COVID-19 crisis, there are very few papers discussing its economic implications-see Baldwin and Weder di Mauro (2020) and Eichelberger et al. (2020). Due to the relatively recent nature of the coronavirus crisis, almost all of the related research is qualitative in nature (due to a lack of access to relevant data). The present study adds by bringing hard evidence to bear in the context of international economic growth.
The consideration of the different supply chain dimensions also enables us to take account of the sequential nature of some of the aspects of the supply chain. If it turns out that the individual aspects of logistics performance have differing growth consequences, i.e., with some positive and others negative growth impacts, then overall economic growth would not go up (and in fact could go down), unless all supply chain dimensions improve in a coordinated and timely manner. Viewed from a different angle, the individual stages of the supply chain can potentially pose hold-up issues, where individual bottlenecks could detail a nation's growth. The potential for hold-up would depend upon related technical and logistical complementarities (Goel (2009)). Therefore, an identification of such potential bottlenecks would be instructive for policymakers.
Beyond the technical dimensions of logistical bottlenecks, there may be political obstacles related to the will and financing to remove such obstacles trajectory (see Werlin (1991)). The logistics performance literature notes the importance of lower international trade frictions brought on by better logistics performance (Hausman et al. (2005), Nordås et al. (2006). The multilateral coordination challenges across nations have also been discussed in the context of China's Belt and Road (B&R) initiative (https://www.cfr.org/backgrounder/chinas-massive-beltand-road-initiative). The B&R initiative is a massive multi-country initiative aimed at streamlining transportation and improving logistical bottlenecks in global supply chains. The present work can also be tied to decision-making under uncertaintyhow do the uncertainties related to timely delivery or the quality of shipments impact economic growth?
An overview of the different bodies of work shows that the aspect studied is somewhat unique and the tie to the coronavirus emergency adds special importance. Our formal estimation model follows.

Model
The general form of the estimated growth model, with subscripts i and t, respectively denoting a country and a year, is the following: The dependent variable is the real per capita annual GDP growth rate across the 130-odd nations in our sample. As Table 1 shows, the average per capita growth rate over 2007-2017 was 1.5 percent, with substantial variation across individual nations.
Using insights from the growth literature, we include a number of regressors to explain economic growth, with the main focus (and novelty) lying in the consideration of various aspects of supply chain logistics. How does supply chain performance impact economic growth?
Turning to the main focus then, we first consider an aggregate index of supply chain logistics performance (SClogistics). Improved logistics performance, ceteris paribus, would lead to higher economic growth. However, individual dimensions of logistics might impact growth differently. Accordingly, we consider six sub-components or aspects, dealing with customs performance (SCcustoms), infrastructure (SCinfrastructure), shipments (SCshipments), logistics quality (SCquality), the tracking of shipments (SCtracking), and the timeliness of delivery (SCtimeliness)see Table 1 details about how these were specifically measured.
As alluded to above, the impact of COVID-19 on global supply chains has also directly impacted the very system treating it. Before the pandemic, about half of the global face masks were produced in China (Ranney et al. (2020)). Due to the virus, the production was halted for a time, resulting in decreased exports and global face mask (and ventilators) shortages. Other supply chain disruptions have resulted in a shortage of other Personal Protective Equipment (PPE) such as gloves, gowns, and hand sanitizer for healthcare workers on the frontline (see Rowan and Laffey (2020) for evidence from Ireland). This decreased supply and increased demand has led to increased infections and deaths among healthcare workers when they are severely needed. To help increase the supply, governments are looking towards other industries using similar equipment, such as artists and construction workers and non-essential services (such as elective surgeries) have been suspended.
Not only are the individual aspects of supply chain logistics qualitatively different, many of those are sequentially dependent. For instance, a good infrastructure would be an important precursor to the performance of subsequent shipments, tracking and timely delivery. On the other hand, the performance of customs might not matter very much for nations that are largely selfreliant.
The usual controls as drivers of economic growth include physical capital investment (INVEST), labor force growth (EmpGR), and labor quality (HumanCap), (see Romer (1990)). 3 The consideration of capital and labor inputs can be viewed in the context of an aggregate production function of nations. These are expected to have a positive impact on growth and are routinely used in the related literature (see Barro and Sala-i-Martin (2003), Jones (2016)). In addition, to consider the possibility that the level (or base) of economic prosperity dictates how fast a nation might grow economically, we include lagged GDP as a regressor. It is generally the case that after a certain level of prosperity, nations find it hard to maintain high rates of economic growth (so-called convergence).
Furthermore, land locked nations like Austria, Mongolia, Nepal, Switzerland, etc. face special logistical challenges (see Arvis et al. (2007), Eyob and Kahsai (2019)), and we include a dummy variable, LandLocked, to identify such nations. Landlocked nations have to rely on their neighbors for access to sea routes, and, depending on a landlocked nation's geography, even building airfields in some landlocked nations might be challenging (e.g., Bhutan, Bolivia).
Whereas aspects of the infrastructure are included in the SCinfrastructure index, we also consider direct measures of the infrastructure, including the rail and air networks (RAIL and AIR, respectively), internet diffusion (INTERNET) and shipping connectivity of a nation (ShippingINDX), (Table 5). A greater diffusion of the internet in a nation, for example, should facilitate the tracking of shipments. 4 Finally, a nation's colonial past might dictate the development of logistics and their impacts of growth and this is accounted for by including a dummy variable for nations with colonial past (COLONY), (Table 6). For instance, the colonizers were instrumental in the early development of railroads and the postal system in nations like India; yet, they favored the dismantling of native cottage (textile) industries in favor of mass-produced goods from the industrial revolution that could be imported. Thus, the overall impacts on growth are unclear a priori.
The following section discusses the data used and the estimation procedure(s) employed to estimate equation (1).

Data
The data used to estimate the model is an unbalanced panel of 136 countries observed from 2007 to 2017- Table 1 includes variable definitions, summary statistics, and sources. The time frame of our sample is constrained by the logistics variables that are available biannually from 2007 to 2018. The main dependent variable is a measure of total annual economic growth calculated as the log difference of real GDP per capita. The average growth rate (from 2007 to 2017) in the sample is 1.5%, ranging from a low of -9.5% (Yemen) to a high of 7.6% (Zimbabwe).
The main explanatory variable is an index capturing the overall supply chain logistics performance. This index is constructed using several dimensions of supply-chain logistics based on timeliness, tracking and tracing, logistics quality and competence, international shipments, infrastructure and customs (see https://lpi.worldbank.org/about for details; also see Arvis et al. (2018)).
The average supply-chain logistics index (SClogistics) is 2.85 (on a five-point scale; higher values denoting better performance), with Germany showing the highest supply-chain performance (4.23) and Afghanistan (1.21) the lowest. Likewise, supply-chain performance varies across each individual dimension. For example, in terms of the timeliness dimension of supply chain performance (SCtimeliness), Luxembourg (4.8) scores the highest while Afghanistan and Somalia (1.4) both score the lowest. Strategic, political, and geographic factors, along with the degree of globalization will dictate how the supply chains in individual nations would fare in the wake of the coronavirus crisiseven in nations that are largely unaffected directly by the disease. However, the unknown scope and end times of the virus spread make decisions, especially related to long term investments, challenging. Figure 1 shows the relation between EconGR and SClogistics across the sample countries. This figure provides an exposition of the correlation (-0.06 and statistically insignificant) between the two variables given in Table 2. Interestingly, improvements in supply chain performance precede improvements in economic growth and this is clearly evident around the time following the recent financial crises. The econometric analysis below will formally test this relationship.
Details about the data used, including summary statistics, variable definitions and data sources are provided in Table 1. Table 2 provide pairwise correlations between the key variables of interest. Notably, not all aspects of logistics performance share a positive and significant correlation with economic growth and the formal analysis below will test the strength of these relations in the context of other potential influences on economic growth.

Estimation
To estimate the baseline models for the overall and each dimension of supply-chain logistics performance we use OLS and control for regional specific effects and time effects. Furthermore, we gauge the statistical significance of individual coefficient estimates using robust standard errors. However, there is the potential for simultaneity between economic growth and supplychain logistics performance, thus we account for this endogeneity using two-stage least squares (2SLS) estimation in Table 4; also see Section 4.2. The results section follows. Table 3 reports results from our baseline regressions, with both the aggregate logistics index and its different components highlighted.

Baseline models
The results show that improvements in overall logistics performance boost economic growththe resulting coefficient on SClogistics is positive and statistically significant. In terms of economic significance, the elasticity estimate suggests that economic growth would increase by approximately 14.4% following a 10% improvement in supply chain performance.
Upon disaggregating, supply chain logistics across different, qualitatively distinct and somewhat sequential, dimensions show that not all have positive and statistically effects on economic growth. Specifically, infrastructure (SCinfrastructure) and shipments (SCshipments) dimensions are positive but statistically insignificant. This may be due to the longer-term payoffs from infrastructure improvements, while shipments could be in inventory. On the other hand, the timeliness of shipments turns out be strongly statistically significant. This makes sense, since the timeliness of shipments enables planning for delivery and use, and that does have positive growth dividends. Also, if deliveries are delayed and not on time, consequent payments for goods and services are delayed, then workers and suppliers are not paid on time and so on. This ripple effect understandably would have adverse growth consequences. Quantitatively, a ten percent increase in the performance of supply chain timeliness would boost economic growth by about 17.6%.
Regarding the usual drivers of growth, the results are in line with the literaturehigher investment, and both the quantity and quality of labor boost economic growth. The complementarity between the quantity and the quality of labor has obvious public policy implications. As expected, more prosperous nations grow somewhat slowly, suggesting convergence in the rates of growth. Finally, economic growth in landlocked nations was no different from other nations. 5

Robustness check accounting for potential simultaneity
There is the possibility that there might be reverse feedbacks from growth to logistics performancehigher growth nations have greater resources devoted to logistics and infrastructure, and thus would likely have better performance. Accordingly, Table 4 considers the different aspects of supply chain logistics to be endogenous, thus we instrument each logistics measure using a weighted average of neighboring countries' supply-chain logistics performance (averaged from 2007 to 2010). 6 This spatial lag is constructed by calculating the inverse distance between each country i and j, 1/dij, where d is the geographic distance between country i and country j's centroid. The spatial lag thus is able to consider a nation's relations beyond its immediate neighbors. In addition, we consider the length of a country's land border (in kilometers) as an additional instrument.
Both instruments (assumed to be exogenous) are highly correlated with supply-chain logistics performance as evidenced by the rejection of the underidentification and weak identification tests reported in Table 4. Furthermore, the Hansen J test for overidentification fails to reject the validity of the instruments-instrument diagnostic tests and results are reported at the bottom of Table 4.
The results show that the baseline results are valid. Importantly, the aggregate and (most) individual dimensions of supply chain performance contribute positively to economic growth. Indeed, the impact is more than twice as much as the baseline estimates, however, less precisely estimated. Again, the growth dividends from improvements in the timeliness performance are substantial (as is also the case with the performance of shipments (SCshipments)).
The results for investment and labor and lagged GDP are qualitatively similar to Table 3. Further support of the baseline models is shown by the endogeneity test results that suggest that supplychain logistics is exogenous. Additionally, there seems somewhat greater support for higher growth in landlocked nations. 7 We now consider some further dimensions to shed additional light.

Additional consideration 1: Impact of infrastructure dimensions
Infrastructure is an important cog in the logistics wheel, and it is considered in the indices considered in Tables 3-4. As an alternative consideration and as a robustness check, in Table 5, we consider direct measures of infrastructure, including a nation's rail network (RAIL), air network (AIR), internet (INTERNET), and shipping (ShippingINDX). These direct measures of infrastructure and logistics capabilities might capture impacts on growth that the aggregate indices are unable to uncover.
The results show that the coefficients on RAIL, AIR, and INTERNET are statistically insignificant, but better shipping connectivity had positive growth effects. This suggests that the logistics measures are considering broader dimensions of logistics that these individual measures are not capturing. Interestingly, growth was higher in landlocked nations, even though shipping connectivity for them (at least directly) did not matter. The other findings were supportive of earlier results.

Additional consideration 2: Impact of colonial legacy
As a further consideration, we consider the effect of colonial legacy. A nation's colonial past shapes the institutional legacy, which would impact growth and also shape the development of infrastructure (Goel and Saunoris (2016)). Accordingly, in Table 6 we consider a variation of Table 3 with a dummy variable (COLONY) identifying nations that were former colonies. A little more than four-fifths of the sample nations were former colonies ( Table 1).
The resulting coefficient on COLONY was statistically insignificant in all cases, and the other results support those reported in Table 3. Thus, economic growth in former colonies was no different from that in other nations. 8 The importance of aggregate supply chain performance (SClogistics) and the relative superiority of timeliness (SCtimeliness) is reinforced (as in Table 3). Next, we consider some simulated scenarios to see how unexpected supply chain disruptions might impact economic growth.

Simulated impact of supply chain disruptions
In this section we conduct simulations of supply-chain disruptions and their impact on economic growth (see Ivanov (2020) also). To carry-out these simulations we consider five somewhat heterogeneous countries for a basis of comparison. We then consider a 10%, 20%, and 30% 7 It could be the case that the nature of outputs of the landlocked nations is such that their geography is less of an issue (e.g., computer software). 8 One reason for the lack of significance of colonial effect might be that nations became free from colonial rules in the past and the years of independence enabled them to overcome some of the drawbacks from colonial rule. decline in supply-chain logistics performance. 9 For each scenario, we trace out their corresponding impact on economic growth based on the elasticity 1.44 estimated in Model 3.1 of Table 3.
The results show a significant impact on economic growth. For example, in Germany a 30% shock or disruption in supply-chain logistics performance (4.12 to 2.88) would yield a decline in economic growth from the sample average of 1.3% to 0.8%. On the other hand, a smaller disruption of 10% would lower the German growth to 1.1%. Indeed, these are likely to underestimate the true effect given the obvious spillovers from supply chain disruptions from one country to another (see Hausman et al. (2005)).
The corresponding drops for the other nations are shown in Table 7. 10 Overall, we see that supply chain disruptions have significant adverse growth effects and these disruptions have become significant in many cases with the current pandemic. 11 Supply chain performance is a crucial cog in the growth wheel. The concluding section follows.

Conclusions
Reliable supply chains are crucial to productivity and economic growth of nations. The reliance of nations on global supply networks has increased with greater globalization (Arvis et al. (2018)). The performance and reliability of global supply chains have gained greater recognition in the face of the challenges posed by the COVID-19 pandemic (see https://www.weforum.org/agenda/2020/04/covid-19-pandemic-disrupts-global-value-chains/).
Despite the recognition of its importance, formal research on the contribution of supply chain logistics, including the relative impacts across different logistics dimensions, is less. This paper uses data over a recent decade for over 130 nations to examine the relative effects of the different aspects of supply chain logistics on economic growth. Figure 1 shows no clear pattern of a relation between the aggregate logistics performance and cross-country economic growth.
Our results show that improvements in the aggregate supply chain logistics performance yield positive growth dividends. These dividends are also present in most of the disaggregated dimensions of logistics, with the performance of the timeliness of shipments have the greatest positive impact on growth. The importance of timeliness makes sense when one thinks that payments for transactions are fully made upon timely delivery. In the absence of timeliness, the subsequent production stages would be adversely affected, with negative growth consequences.
The growth impacts of investment, labor growth, labor quality, are positive and in line with the extant literature. The main findings of the paper are robust to the allowance for bi-directional causality between supply chain logistics and economic growth. However, nations' colonial legacy did not matter and the impacts of being landlocked were somewhat mixed.
Many of these drivers for economic growth are adversely impacted by COVID-19, resulting in negative growth dividends due to the disruption of supply chains. In addition to direct supply shortages (i.e. less exports), the less impacted countries will have greater difficulty obtaining supplies to manufacture goods (Baldwin and Weder di Mauro (2020)). The decreased demand (due to recession and caution by consumers) will further increase the economic decline. In other words, there will be a decrease in investment, labor growth, and other supply chain logistics. To combat this, countries may look to diversify their overseas suppliers as well as encourage return of business to national soil (although this is currently undermined by the travel restrictions), (Baldwin and Weder di Mauro (2020)). As the pandemic subsides, a close eye must be kept on the synchronous activity of multiple components of the supply chain because if, for example, if the shipments are on time but customs are still strict, there can still be continued disruption. Quantitatively, using the elasticities from Table 3, a ten percent improvement in, say logistics quality (SCquality) or logistics tracking (SCtracking), with a concurrent and corresponding decline in logistics timeliness (SCtimeliness), would result in an overall decline in the economic growth.
Governments around the world have been paying attention to improving supply chain performance. The coronavirus crisis has brought these issues to the forefront and added a sense of urgency to government intervention. In the United States, the government has implemented The Defense Production Act (https://www.fema.gov/media-library/assets/documents/15666) to deal with supply constrains related to COVID-19. However, the realization of where to focus limited government resources (in the United States and elsewhere) to ensure sufficient supplies of essentials and lower the impacts of economic growth does not seem to be there. In this regard, the results of this study can provide some guidance.
Implementation of improvements in logistics performance is difficult and time-consuming, given the complexities involved (see Arvis et al. (2018)). 12 In the case of the current pandemic, the supply chain performance would contingent on the opening and closing times (Ivanov (2020)), and with the closing time being uncertain. In addition to the shortages and on-time delivery of products, other impacts of COVID-19 could include less product diversity for consumers and greater monopoly for those companies that can continue to produce products (see Goel (2009) for a theoretical model regarding the impact of technological complementarities on market power). This would then pose challenges for antitrust regulators in the intermediate run. The impact of COVID-19 could be further extrapolated to other major supply disruptions in the future, resulting in readiness for future events.

SClogistics
Overall supply chain logistics performance index (LPI) measured as an unweighted average of the six sub-indicators including timeliness, tracking and tracing, logistics quality and competence, international shipments, infrastructure, and customs. The index is rated from 1 = "very low" to 5 = "very high".

Figure 1 Economic growth and supply-chain logistics
Note: Data based on averages of all countriessee Table 1 for details.