Exploring Slow Growth in Mexico

Abstract

Mexico’s economic reforms were intended to restart growth after the 1980s debt crisis and lead to a closing of the income gap with the United States and other countries. Since then, the growth failure in Mexico and the lack of convergence with high income countries has created extensive debate but there is no consensus why the reforms failed. Most analysis begins with a growth accounting showing low or even negative growth in total factor productivity since the early 1980s. From there, a variety of explanations are given but without clear consensus about causes or remedies. The usual assumptions of distorted prices and institutional failures assume the need for further reform but without agreement on priorities. This paper puts that analysis into a historical context and asks why Mexico’s growth was rapid and leading to convergence through the middle part of the 20th century, when prices were undoubtedly more distorted than today, and market-based policies were not well supported by the country’s institutions. In place of the current debate about growth in Mexico, the high growth rates through the middle decades of the 20th century and the lack of growth since the reforms of the 1980s and 1990s are explained as the result of deploying the general-purpose technologies (GPTs) of the Second Industrial Revolution but failing to deploy and use the GPTs of the Third Industrial Revolution. This shifts the debate from one of distorted prices and broad institutional weaknesses to a discussion of the policies that enable the widespread adoption of new GPTs.

Introduction: The Growth Puzzle

“In general, we look for a new law by the following process. First, we guess it. Then, we compute — well, don’t laugh, that’s really true. Then we compute the consequences of the guess, to see what, if this is right, if this law that we guessed is right, we see what it would imply. And then we compare those computation results to nature. Or we say, compare to experiment or experience. Compare it directly with observation, to see if it works. If it disagrees with experiment, it’s wrong. And that simple statement is the key to science. It doesn’t make any difference how beautiful your guess is, it doesn’t make any difference how smart you are, who made the guess, or what his name is. If it disagrees with experiment, it’s wrong. That’s all there is to it.”[1] 

Mexico’s experiment with neoliberal reforms in the 1980s and 1990s were a product of the best guesses by economists of the requirements for growth. The reforms remade the Mexican economy (Lustig 1998). They enabled the country’s return to international capital markets, shut the door on the lost decade of the 1980s, attracted a disproportionate share of the world’s foreign direct investment, and brought widespread acclaim to the country and its leaders.[2] Policy reforms followed the recommendations of the World Bank, the International Monetary Fund, the U.S. Treasury Department, and important international think-tanks such as the Institute for International Economics, making the country a leading example of neoliberal economic reform (Lustig 1998; Edwards 1995; Moreno-Brid and Ros 2009). Overall, Mexico’s reformers were relatively cautious and less aggressive than reformers in other parts of Latin America but given its first-mover role, its policy setting gained positive recognition as courageous and innovative (Stallings and Peres 2000).

Mexico unilaterally reduced its tariffs, dismantled most import quotas, privatized hundreds of firms, controlled inflation, and opened its markets to foreign investment, all in the effort to restore growth. In the 1990s it adopted a floating exchange rate and made its central bank independent. When it signed a free trade agreement with the United States and Canada and joined the Organisation for Economic Co-operation and Development (OECD), many proponents of the reforms claimed that Mexico was quickly moving towards becoming a high-income country. While that was propaganda designed to generate political support for the passage of the free trade agreement, the reforms were a serious change of direction in economic policy. Markets became more prominent, while economic interventions by the state were greatly diminished.

Economic growth, however, did not respond. Mexico grew faster in the 1990s than it had in the 1980s, but beating the record of the debt crisis was a low bar. The fact that so many causes of slow growth were suggested should have been a signal that the promoters of economic reforms did not completely understand the requirements for growth. There were too many explanations, some overlapping, some not, and some contradictory. Economists pointed to problems such as dysfunctional credit markets (Krueger and Tornell 1999; Bergoeing, P. Kehoe, T. Kehoe, and Soto 2002; Faal 2005; Kehoe and Ruhl 2010; Hanson 2010, 2012), institutional weaknesses, especially the rule of law and corruption (Acemoglu and Robinson 2012), macroeconomic mistakes (Lustig 2001; Edwards 2010), unintended impacts of tax systems (Dussel 2003; Levy, 2018), distortions in market incentives caused by government policies (McKinsey Global Institute 2014; Levy 2018), the lack of government sponsored industrial policies (Moreno-Brid, Santamaria, and Rivas Valdivia 2005; Moreno Brid and Ros 2009; Ros 2013; Moreno-Brid 2013), a decline in public investment (Núñez Rodríguez 2006; Gutiérrez Cruz, Moreno Brid, and Sánchez Gómez 2021) and too much, or maybe too little, labor market flexibility (Palma 2011).[3] Most analysts would probably agree with Torre Cepeda and Ramos (2015) that disappointing economic growth since the 1980s has multiple causes. Ros (2013 and 2015) catalogs and critiques a large number of additional factors claimed by some economists as the main obstacles to growth, including rigid labor rules, monopolization of key sectors such as electricity, educational quality and quantity, informality, and more.

The failure to understand the requirements for economic growth should not be a surprise. Robert Solow, Nobel laureate in economics and founder of modern growth theory, put it this way in 2007: “But in real life it is very hard to move the permanent growth rate; and when it happens, as perhaps in the USA in the later 1990s, the source can be a bit mysterious even after the fact” (Solow 2007). It is not that nothing can be known, but rather we should be more realistic about the difficulties inherent in designing policies that lead to higher growth.

The goal of this paper is to encourage a different approach to understanding the growth failure, one that does not reject all of neoliberal economics, but rather places the emphasis on the more basic requirement of ensuring that new technologies are deployed instead of emphasizing the barriers to static allocative efficiency. If growth in the global economy has occurred in waves that are tied to specific general purpose technologies (GPTs), then the focus of economic policy and of the analysis of growth failures should be on the obstacles to implementing the new technologies and the many sub-inventions that they give rise to.[4] This shifts the analysis away from a discussion of price distortions or general institutional weaknesses and refocuses it on the specific requirements for implementing the new ideas and knowledge that enabled Mexico to access the technologies that increased global growth during the middle of the 20th century and that have been a source of weakness since the 1980s. The core assumption of this paper is consistent with Solow’s breakthrough model of economic growth, that the intensity of technology usage and diffusion is a key part of the explanation of country productivity differences (Solow 1956, 1957; Gallardo-Albarrán and Inklaar 2021).

The discussion is organized under the following sections:

1. Growth accounting.
2. The quantitative record.
3. Technology deployment, prices, and institutions.
4. Lessons from Mexico’s golden age of growth. 
5. The Third Industrial Revolution.

Growth Accounting — Growth accounting is a common method used by economists to explain country growth records, but it has limitations and does not offer policy solutions. Such solutions typically stem from prevailing economic ideas. For Mexico, growth accounting interpretations were influenced by the Thatcher and Reagan eras, marking a shift to neoliberal policies globally. This included market reforms in China, India, Latin America, and the collapse of the Soviet Union, all emphasizing market reliance over state intervention. This ideological shift brought both necessary corrections and some excesses.

The Quantitative Record — This examination of Mexico’s economic performance spans the 20th century and early 21st century, focusing on growth waves since 1870. Mexico experienced rapid growth from 1932 to 1982 during the Second Industrial Revolution but missed much of the growth benefits of the Third Industrial Revolution from the early 1980s to the present. This analysis highlights the impact of productivity growth and living standards driven by general-purpose technologies (GPTs) during these industrial revolutions.

Technology Deployment, Prices, and Institutions — This section looks more closely at the GPTs that enabled a significant increase in the growth of productivity and income during the Second and Third Industrial Revolutions and hypothesizes their roles in determining the historical pattern of growth in Mexico. 

Lessons From Mexico’s Golden Age of Growth — By surveying the trade and industrial development policies used to promote growth in the middle of the 20th century, before being dismantled in the 1980s, one can analyze the successes and failures and provide useful insights for any reconsideration of current policies.

The Third Industrial Revolution — The paper concludes with some observations on Mexico’s ability to implement the technologies of the Third Industrial Revolution. A major issue is the tension between neoliberal economics and the deployment of new technologies.

Growth Accounting

Analyses of comparative growth rates usually begin with the measurement of productivity growth and then move towards an analysis of the factors behind a country’s performance. To illustrate this point, it is helpful to briefly outline how a growth accounting model unfolds. We define a constant returns to scale (CRTS) Cobb-Douglas production function as 

 

where Y is gross domestic product (GDP), K is capital, L is labor, A is productivity, and 𝛼 is capital’s share of output.[5] With the assumption of CRTS, labor’s share is 1-𝛼. We can differentiate labor by incorporating a human capital component, h, which is a measure of labor quality and will be a factor of increase for the effective labor input. A common way to measure h is with years of schooling. In this case, the production function can be written as

 

Given that the goal is to measure output per worker, we divide both sides by L. The output variable becomes Y/L, or y, which is a measure of labor productivity, and the production function is expressed as in Equation 1, with lower case k and h symbolizing physical capital and human capital per worker.

 

Taking logarithms and differentiating with respect to time: 

 

where ^ or “hat” implies rate of growth. Rearranging:

 

Equation 3 is the standard growth accounting equation showing how the rate of change of productivity (Â) is calculated as the residual and can be broken down into the growth of output per worker, ŷ, and the growth of factor inputs:[6]

 

Since the seminal work of Solow (1956, 1957), the analysis of economic growth has focused on the growth rate of labor productivity, measured as either output per worker or output per hour worked (ŷ), and total factor productivity (TFP), measured as the residual.[7] In Solow’s original model of the U.S. economy, TFP growth accounted for approximately seven-eighths of the growth in GDP per person (Solow 1957). Since then, estimates of the importance of TFP have been reduced, although it remains the most critical and powerful determinant of economic growth in most studies. In general, TFP growth is more important than factor inputs in the determination of growth rates, while factor inputs are more important in the determination of the level of income (Weil 2009, 196–203).[8] Research has also shown that the Solow result and estimates showing TFP as the primary determinant of economic growth may be dependent on the time and place where measurements are made. A recent analysis of long run growth across many countries found that TFP’s dominance in growth accounting is often dependent on an earlier period of capital deepening (Gallardo-Albarrán and Inklaar 2021).[9] This point is worth raising since at least one of the critics of the many analyses of slow growth in Mexico has pointed out that most research focuses on TFP growth but overlooks the low rates of investment and slow capital accumulation (Ros 2013).

TFP can be understood in general terms as a measure of the efficiency in converting capital and labor into output. Alternatively, TFP is the difference between the growth of real output and real factor inputs. Hence it includes a range of influences, including technological changes, organizational effects, economy of scale effects, sectoral effects when workers and capital move off the farm and into the factory, and institutional effects that reduce transaction costs, such as contract enforcement and protection of property rights. Consequently, it is expected that TFP will become the central focus of research on economic growth, and serve as the primary platform for evaluating the effectiveness of Mexican economic reforms.

A major obstacle to this approach stems from the fact — noted earlier — that TFP is calculated as a residual in studies of growth accounting. Consequently, it is a catch-all for everything not directly controlled in the analysis, making its interpretation less than clear. Many economists over many decades have noted that an explanation of TFP is difficult and uncertain (Jorgenson and Griliches,1967; Abramovitz 1993; Easterly and Levine 2001; Hulten 2009; Pradosde la Escosura, Vonyó, and Voskoboynikov 2021). This is one reason why there are so many explanations for slow Mexican growth. The studies showing Mexico’s failure to increase TFP over the last few decades provide no direct insight into the policies that will generate growth. All of the explanations cited at the beginning of this paper come from outside the measurement of TFP, meaning they are either explicit or implicit assumptions about economic variables and their impact on economic growth.

In addition to the absence of any clear policy implication, the problem of viewing TFP as the key to economic growth is compounded by its potential endogeneity: Does TFP determine growth or vice versa?[10] The potential endogeneity of TFP challenges and makes uncertain the idea that growth can be achieved through greater market efficiency. Hence, some critics of the neoliberal policies focused on market reforms have argued that investment and capital deepening are more important and should be prioritized (Ros 2013, 2015).

Despite these limitations, it is worthwhile to look at the quantitative estimates of TFP and other measures of performance. As should be clear from the previous paragraphs, this is to provide an historical perspective on long run growth. This connects conditions in Mexico with conditions in the United States and other countries and creates a comparison between country responses to the new waves of productivity growth produced by new technologies.

The Quantitative Record

To estimate Equation 3, we need measures of the rates of change of y (output per worker), k (capital per worker), and h (human capital per worker). In addition, we need to know a, capital’s share of total output. Currently, there is limited availability of long-run economic data series that are comparable across many countries. The Penn World Table (PWT) is one of the most widely used and cited data sources but it does not have data from before 1960 (Feenstra, Inklaar, and Timmer 2013). The Maddison data set has much longer historical coverage but is limited to real GDP and population (Bolt and van Zanden 2020).[11] The IMF and World Bank datasets have far more variables but, like the PWT, they are limited in time.

The data used for this analysis come from the Long-Term Productivity Database (Bergeaud, Cette, and Lecat v2.6, August 2023).[12] This relatively new database was constructed to measure productivity trends in several high income OECD countries (Bergeaud, Cette, and Lecat, 2016; Prados de la Escosura, Vonyó, and Voskoboynikov 2021). Over the years, its authors have added countries, including Mexico, and continually updated the variables. It currently includes estimates of GDP per capita, labor productivity, total factor productivity, capital intensity, and age of the capital stock. All variables except the age of the capital stock are given in 2010 U.S. dollars, purchasing power parity terms.[13] The estimates for Mexico cover the years 1890–2022.

GDP per capita is based on data from the Maddison Project (Bolt and van Zanden, 2020) and supplemented with Mexico’s national accounts. Labor productivity considers hours of work but not the quality of labor, so quality improvements are a residual and therefore included in the estimate of total factor productivity (TFP). Bergeaud, Cette, and Lecat use a perpetual inventory method to estimate capital inputs, which include buildings and equipment, and assume a depreciation rate of 10% for equipment and 2.5% for buildings. They also assume a constant returns to scale production function with labor and capital shares set to 70% and 30% respectively. These are standard assumptions in growth accounting exercises and their results are robust with respect to the values selected for depreciation, labor’s share, and capital’s share (Bergeaud, Cette, and Lecat 2016).

Both the labor productivity and TFP growth rates are noisy and difficult to interpret when examining the raw data. To provide a clearer picture of the growth trends in different periods, Figures 1 and 2 show the plotted growth rates of each variable smoothed with a 10 period moving average filter (MA10).

Figure 1 — Annual Growth Rate of Labor Productivity, MA10

 

Figure 2 — Annual Growth Rate of TFP, MA10

 

Overall, labor productivity and TFP move together (correlation = 0.958, n =132), with similar periods of positive and negative growth and similar occurrences of peaks and troughs. Both variables were largely positive during the last years of the Porfiriato, through the Revolution and in the 1920s.[14] Both experienced a deep trough in 1932 when they fell by more than 18%. (The declines shown in Figures 1 and 2 are much smaller because the plots are moving averages and the recovery in productivity was quick and strong.) The next 40 years recorded moving average growth rates of between 2% and 4% per year. After that, in the 1970s, productivity growth remained positive, but grew more slowly until it collapsed completely in 1982. Labor productivity growth was -4.05% in that year, and TFP growth was -5.88%. Both measures of productivity remained in negative territory until 1989, when Mexico became the first country to qualify for a Brady Plan partial debt forgiveness. Since 1989, productivity growth has swung back and forth between positive and negative values, but as shown below in the statistical analysis, it has been negative overall.

Three Periods of Productivity Growth

Spanning from the final decade of the 1800s through the second decade of the 2000s, the figures reveal three distinct periods in productivity growth:

1. A rough estimate of the first period would be 1890–1932, during which productivity growth was generally positive, but relatively mild. Growth in both measures of productivity reached a local maximum in 1925 but became more volatile thereafter, swinging between positive and negative values until the major collapse in 1932.  
    
2. The second period is Mexico’s golden age (also known as the Mexican miracle), from 1933 to 1982.[15] Growth was 3.47% per year for labor productivity and 2.11% per year for TFP.  
    
3. The current period started in 1983 and has lasted into the present day.[16] The first years of this period included the worst of the debt crisis — it began in 1982 and involved a long period of failed efforts to emerge from the instability that gripped the country. During the economic reforms initiated in the 1980s and continuing into the 1990s, productivity growth was negative and GDP growth was mediocre at best. The 2000s continued the trends of the 1990s with slightly improved performance in productivity but it still bounced between positive and negative values.

To test the hypothesis that these three periods were distinct and significantly different growth periods, exponential growth rates are estimated using Ordinary Least Squares regression (OLS).[17] The exponential growth equation is:

 

where Yₜ is productivity at time t, 𝛼 and 𝛽 are parameters to be estimated, e is the base of the natural logarithm, and uₜ is a normally distributed independent random error term. Taking the natural logarithm, we have:

 

which is the basic equation to be estimated. To test for significant differences in growth rates during the time periods specified, two dummy variables are introduced along with two interaction terms. The equation becomes:

 

The variables d₁ and d₂ are 0,1 dummy variables. The first variable, d₁, is 1 during the years 1890–1932 and 0 otherwise. The second variable, d₂, is 1 during the years 1983–2022 and 0 otherwise. This makes the golden age years, 1933–82, the control group. The interaction terms, 𝛽₁(d₁)t and 𝛽₂(d₂)t, allow the estimates of the growth rates in the first and third period to differ from each other and from the growth rate during the Golden Age (𝛽₃). The two dummy variables, d₁ and d₂, allow each time period to have its own intercept as well. Table 1 shows the results.

Table 1 — Three Periods of Economic Growth

 

All variables are highly significant, implying that each time period has its own intercept and its own slope and we should reject the hypothesis that the growth rates were the same in each period with only random variation. The test indicates that the rates of growth in each of the three periods varied systematically. The growth rate during the Golden Age was 3.47% per year for labor productivity and 2.11% for TFP (columns 1 and 2). In the period before the golden age, the growth rate of labor productivity was 2.14% (0.0347–0.0133), and TFP was 0.82% (0.0211–0.0129). Positive, but lower. This is not surprising given that the period (1890–1932) encompasses the Mexican Revolution. What may be surprising is that both measures remained mostly positive throughout the period of hostilities, largely due to significant increases in petroleum production (King 1970, 9). More precisely, labor productivity growth was positive every year from 1910 to 1920 and TFP growth was positive except in 1910 and 1911. In the period since the debt crisis began, 1983–2022, estimated labor productivity and TFP growth rates are both negative. Labor productivity’s estimated average growth was -0.57% per year (0.0347–0.0404) and TFP’s was -0.058% per year (0.0211–0.0269).[18]

The growth of labor productivity and TFP are key determinants of GDP growth. Figure 3 is a graph of the growth in real GDP per capita, also smoothed with a 10 period moving average filter. Most of the same patterns seen in the graphs of labor productivity growth and TFP growth are visible here as well: the crisis of the early 1930s, the high and sustained growth rate from the 1940s to the early 1980s, and the weak growth from the 1980s until the end of sample. Table 1, column 3, contains the estimated compound annual growth rates of GDP per person during each of the three periods. The differences for the three periods in Figures 1–3 are significant statistically, regardless which cut-off point is used to distinguish the golden age from the current period.

Figure 3 — Growth Rate of GDP per Capita, MA10

 

Summing up the results of this exercise, we can say that the most significant period of economic growth in Mexico occurred between 1932 and 1982. GDP per capita increased steadily after 1932, except in 1938 and 1940. Thereafter, growth in GDP per capita remained relatively steady until 1982 with only modest declines in 1945, 1953, and 1959. If we compare this record to the 1990s and the two decades of the 21st century, it leaves no doubt which was the period of better economic performance. This is particularly true if we consider either of the two measures of productivity.

Relatively rapid growth in the middle of the 20th century meant that Mexico was converging to the U.S. level of income. Figure 4 shows the ratio of Mexican GDP per person to the United States. Between 1945 and 1982, GDP per capita rose from 29% of the U.S. level to 48%.

Figure 4 — Mexico GDP per Person as Share of US GDP

 

The current period of slow growth in Mexico's economy is puzzling. From the 1930s to the early 1980s Mexico experienced a golden age of economic growth and made significant progress in closing its income gap with the United States, yet economic policy reforms were viewed as necessary. The economic crisis of 1982 discredited the previous policies of import substitution industrialization (ISI), even though the argument that those policies and not some other factors caused the crisis is a gross oversimplification, made during an era when governments (and economists) globally were emphasizing a greater role for markets and a reduced role for governments. A more robust argument about the causes of the debt crisis includes a number of external factors, such as the collapse of oil prices, the global rise in interest rates, the explosive expansion of new forms of international capital flows via direct bank lending, and several macroeconomic policy failures in Mexico that allowed budget deficits and international debt to increase. Mexico’s economic development policies were far from perfect, but they did not cause the debt crisis.

Technology Deployment, Prices, and Institutions

Taking a global view, Gordon (2016), Bergeaud, Cette, and Lecat (2020), and others describe two waves of new GPTs since the late 1800s. Both ushered in a series of related innovations that caused significant increases in the rate of economic growth in every country deploying these technologies. 

The First Wave

The first wave was built on the introduction of electricity, internal combustion engines, and the creation of chemicals for medicine and industry. These enabling technologies led to a large number of sub-inventions that were only possible after the new technologies began to be deployed.[19] The United States was one of the first adopters and that was a central reason why it was able to surpass the U.K.’s economy. One of the key characteristics of the new technologies is that their impacts on economic growth do not show up for several decades: Many were available in the late 1800s, but our best estimates of the U.S. macroeconomic data do not show any significant effects until the 1920s. The reasons for this lag effect are beyond the scope of this paper.

The implementation and use of these technologies and their many sub-inventions continued to increase productivity and material well-being for several decades after they began to be used widely. Electrification, refrigeration, the invention and adoption of new household appliances, personal vehicles, and farm and construction equipment with internal combustion engines, the spread of the interstate highway system, radio, television, telephony, and the relatively rapid development of synthetic materials and new pharmaceuticals, were all part of a wave of new technologies that continued to be deployed into the 1960s and 1970s. 

The Second Wave

As that first wave of productivity increase was exhausted, a second wave of new GPTs began in the 1970s — they were tied to the development of information and communication technologies, and the adoption of computers. Likewise, it did not appear in the macroeconomic record immediately, but a new boost in productivity growth in the U.S. began to be observed in the 1990s.

Whether Mexico’s successful growth from the 1930s through the 1970s was the result of the same or similar factors as those that created rapid growth in the United States, Europe after World War II, and several east Asian countries, is speculative — a more detailed examination of Mexican economic history is required. In the case of the weaker second wave of world productivity growth that began in the 1990s, an obvious hypothesis is that the Mexican economy has not been able to successfully deploy the new technologies associated with information, communication, and computer technologies. If so, Mexico is not alone. Bergeaud, Cette, and Lecat (2016) show a downward productivity trend break in Europe in the 1990s; Islam (2003), using a large panel of countries, estimates as many cases of productivity decline as increase; and Palma (2011) shows that productivity growth in Latin America was uniformly negative in the 1980s and remained so for half the countries in the 1990s and after.

In the neoliberal era of economic theorizing, getting prices and institutions right are viewed as the key task for policy makers. Consequently, if economic growth is slow, analysis tends to focus on the microeconomic functioning of markets, transaction costs, and the incentives created for individuals and firms. It is uncertain, however, whether the use of GPTs require the types of microeconomic reforms that are at the core of neoliberal policy prescriptions. Many, if not most, economists would probably agree that the reforms prescribed are generally desirable policies under most circumstances. Nevertheless, many countries have achieved high rates of growth without getting all prices right, and without solving all their problems of institutional failure or weakness. The assumption that policies to promote microeconomic efficiency are a requirement for the deployment and usage of new technologies overlooks many case studies of countries where technological catch-up has been successful in economic environments with distorted prices and less than ideal institutions. China is the outstanding example, but other countries in East Asia have also managed to create rapid economic growth while having a variety of inefficient markets, government distortions, weak rule of law, and a lack of intellectual property protections. And Mexico during its golden age of growth is another example. The experiences of these countries do not invalidate the idea that the reforms were necessary in Mexico, but they do raise questions about the necessary effects of market distortions and weak institutions — they suggest that the discussion of price distortions and institutional weaknesses is a misdirection, or at least a misunderstanding of the most necessary changes for restoring growth. It is conceivable that specific reforms, for example in the labor market, would make that market more efficient but have little effect on overall growth.[20]

Lessons From Mexico’s Golden Age of Growth

The decades of high growth in Mexico during the middle of the 20th century may hold important lessons for restoring growth, even if the requirements for growth have changed in the intervening years. Consequently, it is worth considering what was different in that era, and what current economic thought says about those differences. Some of the policies of the period from the 1930s through the 1970s were successful, while others created weaknesses that became problematic during the debt crisis. Rather than a wholesale rejection of all the policies implemented during the high growth decades, a more careful analysis of the successes and failures is useful for any reconsideration of current policies.

In general terms, the two most dramatic differences between the policies of the mid-century and the neoliberal period were the use of 1) industrial policies, and 2) trade restrictions, which we will examine in turn. 

Industrial Policies

Industrial policies were embodied in a series of laws designed to promote “new” and “necessary” industries.[21] Industries were granted fiscal incentives such as tax rebates, reduced duties on imports of capital equipment, and favorable credit terms. The national development bank, Nacional Financiera, was created in 1934 to support industrial development and became particularly focused on infrastructure after World War II (Moreno-Brid and Ros 2009; Cardenas 2022). In addition, the Fund for Industrial Equipment (FONEI) was started in 1971 to help finance capital imports with subsidies for technological adaptation and innovation.

Trade Restrictions

Fiscal incentives were important but the primary policy tool for industrial promotion was trade protection (Moreno-Brid and Ros 2009, 98). New measures included the implementation of import licenses (1944) and their application to more and more products through the 1950s and the 1960s (Wallace 1980). Changes to the import tariff code were also extensive, such as the 1930 increases in import duties, the 1935 customs law, and significant tariff revisions in 1947. Trade protection in Mexico during the ISI period has been portrayed as a move towards a closed economy, but imports and exports continued to grow throughout that time. A better way to understand Mexico’s trade policy is as a case of managed trade. Mexico did not seek to close its economy. Trade continued to grow but policymakers sought to manage the goods and services that were traded.

Support for These Policies

These efforts at structural transformation through industrial promotion and trade policy were supported by many efforts to fill information gaps and to raise the level of education and training:

• Annual publication of lists of imports that could be produced domestically were offered and state banks analyzed and made available studies of the potential profitability of new industries.
   
• ISI policies also played an active role in helping to create Mexico’s national innovation system (Di Maio 2009). A notable example is the National Council for Science and Technology (CONACYT), founded in 1970 with the mission to promote science education and to offer small research grants.[22]
   
• Public research institutes were created in specific fields, including the National Institute for Nuclear Research (ININ), Electrical Research Institute (IIE, today it continues as INEEL, the National Institute of Electricity and Clean Energy), Mexican Institute of Water Technology (IMTA), and the Mexican Petroleum Institute (IMP). While CONACYT’s role was broadly focused on educational development, the specific field institutes were meant to adapt new technologies, either of national origin or borrowed from abroad. All of these helped deploy the technologies of the Second Industrial Revolution.

The goal of industrial policies is to aid in the structural transformation of an economy (Stiglitz and Greenwald 2014, 379). By that standard, Mexico’s mid-century policies were relatively successful. Across the decades of high growth, the structural transformation of the economy through an increase in the volume of manufacturing output was significant and steady (Table 2). Also shown are two indicators representing two core elements of the Second Industrial Revolution, electricity and the internal combustion engine. Mexico’s industrial policies were not uniformly successful, but overall, they led to the widespread adoption of the same technologies that were responsible for a large share of economic growth in the United States, Western Europe, Japan, and elsewhere during the middle decades of the 20th century.

Table 2 — Average Annual Rates of Growth

 

While economic development was foremost in the thinking of policymakers, it was not the only motive behind their desire to transform the country’s economic structure. Economic independence from the United States was a serious concern given the history of conflict, the disputes over water on the border, and the continued interest by U.S. groups in acquiring Baja California. The desire to strengthen the relatively less populated northern border region resulted in the creation of water infrastructure as part of an effort to persuade farmers to relocate and led to support for border industry through the creation of the National Border Program (PRONAF) in 1961. Water development through irrigation systems and dams was successful in creating significant agricultural growth in the north, but PRONAF accomplished very little.[23]

Two other goals were important in addition to economic transformation and independence: 1) protection of the balance of payments through controls on imports and promotion of exports, and 2) the generation of revenue via tariffs, import licenses, and other financial measures such as export taxes. 

Balance of payments deficits were a persistent problem throughout the period and policymakers made several attempts to promote exports. These covered many different actions and were one of the more complex features of commercial policy. They included the various laws to support industries designated as new or necessary, offered potential exemptions from import controls on capital imports and intermediate inputs, tax exemptions, credit lines from banks and the Fund for the Promotion of Exports of Manufactured Products (FOMEX), and the direct promotion of industry by government, sometimes working alone and other times in collaboration with private partners (Wallace 1980; Moreno Brid and Ros 2009). Overall, however, exports grew more slowly than desired, and more slowly than world exports.[24]

The failure of export promotion increased the vulnerability of the economy to a sudden stop crisis.[25] While the crisis that began in 1982 is usually viewed as a debt crisis, it had elements of several types of financial crises, and from a trade perspective it can also be seen as a sudden stop crisis. The ever-larger trade imbalances that developed throughout the 1970s and the increasing dependence on external financing made the country vulnerable to a sudden re-evaluation by foreign lenders. Prior to the announcement of default in August 1982, many observers at home and abroad had decided that their prospects for good returns had declined and they stopped risking assets on the Mexican economy. Over-borrowing by the Mexican government was a key part of the equation, and the emphasis on government debt rather than trade imbalances is appropriate. Nevertheless, trade imbalances played an important role and were one reason why the debt crisis lasted so long.[26]

Lessons Learned

What should be learned from Mexico’s use of industrial policies and its period of high rates of growth? One clear lesson is that new technologies require a midwife. The list of externalities that block their deployment is well known, and in hindsight, we can see several features of Mexico’s industrial and commercial policies that were directed at overcoming the barriers to adoption.

• Risk — Because they are risky, banks are often hesitant to lend. Hence, Nacional Financiera and other development banks were created to provide finance and to fill some information gaps.
   
• Infrastructure — Often new infrastructure is needed, which has coordination problems and significant positive externalities. A well-known characteristic of positive externalities is that the firm that makes the investment cannot obtain all of the benefits — they are external to the investing entity. In an economy that relies on markets alone, provision of goods such as infrastructure will be undersupplied since the investment decision will only consider the private benefits and not take into account the public ones. Consequently, the government built roads and electrical generation plants.
   
• Learning and new methods — Learning and the adoption of new methods are usually needed, both of which raise the level of risk and uncertainty about the outcome. In response, development banks studied potential areas of growth and made their findings public, national research institutes were created in key fields, and a variety of subsidies for national firms and protection from more advanced foreign firms were given. 

Although Mexican policymakers may not have been familiar with the vocabulary of 21st century economics, they implicitly understood many of the problems of externalities and the ways they cause under-investment or no investment in new technologies.

What policymakers did not get right were the incentives for expanding outside Mexico. This is perhaps the greatest difference between Mexican and East Asian industrial promotion policies. Some export promotion measures were tried, such as export financing and exemptions from export taxes, but they had limited success in growing the country’s exports. Every president from Manuel Ávila Camacho (1940–46) to Jose López Portillo (1976–82) averaged six years of trade deficits during their term of office. As noted, this made the country vulnerable to a sudden stop crisis when foreign lenders re-evaluated, but it also meant that there were fewer firms competitive in export markets. When oil prices fell and interest rates rose in the early 1980s, there were not enough firms capable of earning the dollars needed to escape from the debt crisis. As a result, the crisis was prolonged and caused greater damage.

The Third Industrial Revolution

The economic history of the information and technology (ICT) revolution, sometimes called the Third Industrial Revolution, is well-known. As with electricity — which was not widely deployed for several decades after it became technically feasible — advances such as computers, wireless communications, the internet, shipping containers, and cell phones did not appear in the macroeconomic data on growth and productivity until after considerable time had passed. The lag between discovery and deployment was true for many innovations of the Second Industrial Revolution and is a characteristic of decentralized market-based economies (Bresnahan and Trajtenberg 1995).

Mexico’s economy is much more decentralized and market-based than it was in the middle of the 20th century. One interpretation of Mexico’s low productivity growth over the last decades is that it has nearly completely missed the Third Industrial Revolution — at least so far. It is not alone: It is joined by most of the rest of Latin America, parts of the eurozone, and other countries across the globe where the benefits from the adoption of ICT technology are not visible in the macroeconomic data (Hofman and Valderrama 2021; Bergeaud, Cette, and Lecat 2016; Palma 2011). This does not imply that countries lack smart phones: Instead, it indicates that, on average, the adoption of new technologies has not made people more productive at work. This could be due to limited technology adoption and/or insufficient learning and innovation. 

While it must be acknowledged that measurements of productivity and growth may underestimate the benefits of new technologies that enhance our quality of life without being reflected in GDP, the data are clear that the Mexican economy has not found ways to combine its capital and labor more efficiently at work. The limited growth it has experienced is mainly due to more inputs, rather than a better use of inputs.

This has led to lengthy discussions about the rule of law, property rights, contract enforcement, credit markets, and all the other topics mentioned at the beginning of the paper. What the lessons from the Second Industrial Revolution seem to imply, however, is that a more direct way to understand the failure to converge to a high-income level of productivity is through an analysis of the reasons for the lagging deployment and usage of GPTs. In general terms, this has three core elements that need to be examined:

1. The national innovation system.
2. Education and training of the labor force.
3. Policies that support the efforts of firms to adopt new technologies.

Deployment of New Technology

Looking at the Mexico’s 2020 census of population and housing, only 0.8% of Mexican homes do not have electricity. That implies that at least one part of the Second Industrial Revolution has been accomplished.[27] But 62.2% of homes have no computer, laptop, or tablet, and 47.7% have no internet connection. While cell phones are available in 87.8% of households, and some households can access the internet via their phones, the lack of the more robust operating systems and greater processing power offered by computers, laptops, and tablets, combined with limited internet connectivity, leads to reduced learning about the newer technologies of the Third Industrial Revolution, and ultimately less innovation (INEGI 2021).

When we consider businesses and the deployment of new technology, a similar pattern emerges. Bergeaud, Cette and Lecat (2023) estimated the relationship of capital intensity in a cross section of countries. The ratio of capital assets per hour worked was more than three times greater in the United States. At the end of the period of high growth in Mexico (1982), the ratio was 2:1, implying that capital assets at work have fallen further behind the United States’ level since the period of neoliberal reforms. In a study of productivity across Mexican industry, Torre Cepeda and Ramos (2015) show that the distribution of capital equipment is concentrated, skewed towards larger firms, and reflects the highly unequal distributional characteristic of productivity across Mexican firms.

Research and Development Spending

Capital intensity is but one indicator of lagging technology. Another is research and development (R&D) spending. R&D as a share of GDP is less in Mexico than in all Western European, Asian and North American OECD countries and approximately equivalent to the Latin American members of the organization (Costa Rica, Colombia, and Chile). In 2021, Mexico’s R&D spending was one-tenth the level of average OECD countries (0.28% of GDP versus 2.718% in the OECD) and has been trending downwards since 2014 (RICYT 2024a; OECD 2024). Notably, Mexican businesses provided only 17% of the total R&D funding (2021) compared to 66% in the United States (RICYT 2024b). Consequently, problems of business innovation, experimentation, and novel ways to apply new technologies receive less attention than in many other countries. A parallel problem is the relative lack of researchers. Mexico has 1.16 per 1,000 people in the labor force compared to 2.19 in Latin America and the Caribbean overall, and 3.99 per 1,000 in Brazil (2018) (ECLAC 2022). The lack of R&D funding together with the limits on available human capital are obstacles to innovation and investment in new areas and make deployment of new technologies more difficult.

Other Issues

Even without those obstacles, a wide range of externalities are inherent to new businesses and industries. Many of these cause market outcomes to be suboptimal, discourage new investment, and pose significant problems for the introduction of new technologies. For example, first-mover externalities ensure that new businesses will not be able to obtain the full benefits of their actions which, in turn, leads to underinvestment. Some ways in which this manifests:

• Coordination problems arise when the success of an initiative depends on the actions of others who do not know about the opportunity.
   
• Information asymmetries can reduce investment in new technologies because creditors do not understand the potential and are reluctant to lend.
   
• R&D creates information externalities in the form of benefits that are not obtainable by the entity that does the research, causing R&D spending to fall below its socially optimal level.
   
• Public goods, including knowledge and learning, receive less investment in a purely market driven economy. 

These factors explain why middle- and low-income economies struggle to catch up with high-income countries when they rely on the market to develop and implement new technology.

Conclusion

This paper has argued that the reforms in Mexico since the 1980s may be focused on the wrong things. While getting the prices right for market efficiency and strengthening institutions to lower transaction costs are important goals, it is unlikely that they are the main obstacles to higher growth. Both Mexico’s growth record in the mid-20th century and the contemporary record of several mainly East Asian countries demonstrate that high growth is possible even when markets are operating suboptimally and institutions are relatively weak.

What has not been addressed is the role of inequality. This is a complicated topic — and beyond the scope of this paper — but it likely plays a role in Mexico’s failure to benefit from the Third Industrial Revolution. To what extent, for example, is the lack of STEM training and education a result of the country’s meagre revenue collection and the unwillingness of Mexico’s elite to share the benefits of economic growth? Are the low levels of R&D by private enterprises and the relative scarcity of researchers a result of firm’s preferences for rent seeking behaviors over the risks and challenges of innovation?[28] Although these questions cannot be addressed here, they are worth raising to show that there are many layers to a growth failure, just as there are for successful economic growth.

The main assumption of this paper is that the primary determinant of productivity growth is the introduction of new technologies. The historical record shows two waves of productivity growth during the 20th and early 21st centuries, and economic historians have tied those waves to specific new GPTs. Implicitly, it is assumed that failure to deploy and use the new technologies leads to stagnation in productivity and incomes. But productivity-enhancing innovations have a distinctive characteristic that shapes the ability of national economies to simply acquire and deploy them: They do not self-deploy. Economic conditions may be ready to receive them, but the well-known and well-understood range of externalities associated with new technologies guarantees that, without specific policies to promote innovation, the invisible hand of market incentives cannot grasp the opportunities presented.

Appendix — Period Estimates Using 1933–88 as the Control Group

Acknowledgments

This paper was first presented to the weekly seminar of Fellows at the Center for U.S.-Mexican Studies in the school of Global Policy Studies at the University of California San Diego. The author would like to acknowledge and thank Ciro Muyama, Marta Cebollada, Sandra Rebok, Guadalupe Chavez, Alejandro Diaz-Bautista, Cordelia Rizzo, Antonella Bandiera, Rebecca Bell-Martin, and John McNeece for their comments, along with Catheryn Camacho Bolanos and Rafael Fernández de Castro for their support.

References

Abramovitz, Moses. 1993. “The Search for the Sources of Growth: Areas of Ignorance, Old and New.” The Journal of Economic History 53, no. 2 (June): 217–43. https://doi.org/10.1017/S0022050700012882.

Acemoglu, Daron, and James A. Robinson. 2012. Why Nations Fail: The Origins of Power, Prosperity, and Poverty. New York: Crown Publishers.

Bergeaud, Antonin, Gilbert Cette, and Rémy Lecat. 2016. “Productivity Trends in Advanced Countries between 1890 and 2012.” Review of Income and Wealth 62, no. 3 (September): 420–44. https://doi.org/10.1111/roiw.12185.

Bergeaud, Antonin, Gilbert Cette, and Rémy Lecat. 2020. “Convergence of GDP per Capita in Advanced Countries over the Twentieth Century.” Empirical Economics 59, no. 5 (November): 2509–26, https://doi.org/10.1007/s00181-019-01761-x.

Bergeaud, Antonin, Gilbert Cette, and Rémy Lecat. 2023. “Long-Term Productivity Database, V.2.6.” http://www.longtermproductivity.com/download.html.

Bergoeing, Raphael, Patrick J. Kehoe, Timothy J. Kehoe, and Raimundo Soto. 2002. “A Decade Lost and Found: Mexico and Chile in the 1980s.” Review of Economic Dynamics 5, no. 1 (January): 166–205. https://doi.org/10.1006/redy.2001.0150.

Bolt, Jutta, and Jan Luiten van Zanden. 2020. “The Maddison Project: Maddison Style Estimates of the Evolution of the World Economy. A New 2020 Update.” Maddison-Project Working Paper WP-15, University of Groningen, October, https://www.rug.nl/ggdc/historicaldevelopment/maddison/publications/wp15.pdf.

Bresnahan, Timothy F., and Manuel Trajtenberg. 1995. “General Purpose Technologies ‘Engines of Growth’?” Journal of Econometrics 65, no. 1 (January): 83–108. https://doi.org/10.1016/0304-4076(94)01598-T.

Calvo, Guillermo A. 1998. “Capital Flows and Capital-Market Crises: The Simple Economics of Sudden Stops.” Journal of Applied Economics 1, no.1 (November): 35–54. https://doi.org/10.1080/15140326.1998.12040516.

Cardenas, Enrique. 2022. The Mexican Economy. New York: Columbia University Press.

Cerutti, Mario, and Araceli Almaraz, eds. 2013. Algodón en el Norte de México (1920–1970): Impactos Regionales de un Cultivo Estratégico. [Cotton in Northern Mexico (1920–1970): Regional Impacts of a Strategic Crop]. Tijuana: El Colegio de la Frontera Norte.

Crafts, Nicolas, and Kevin Hjortshøj O’Rourke. 2014. “Twentieth Century Growth.” In Handbook of Economic Growth, vol. 2, edited by Philippe Aghion and Steven N. Durlauf, 263–346. Amsterdam: Elsevier. https://doi.org/10.1016/B978-0-444-53538-2.00006-X.

Crafts, Nicolas, and Pieter Woltjer. 2021. “Growth Accounting in Economic History: Findings, Lessons and New Directions.” Journal of Economic Surveys 35, no. 3 (July): 670–96. https://doi.org/10.1111/joes.12348.

Di Maio, Michele. 2009. “Industrial Policies in Developing Countries: History and Perspectives.” In Industrial Policy and Development: The Political Economy of Capabilities Accumulation, edited by Mario Cimoli, Giovanni Dosi, and Joseph E. Stiglitz, 107–43. New York: Oxford University Press. 

Dussel Peters, Enrique. 2003. “Ser Maquila O No Ser Maquila, ¿Es Esa la Pregunta?”. Revista de Comercio Exterior 53, no. 4 (April): 328–36.

Easterly, William, and Ross Levine. 2001. “What Have We Learned from a Decade of Empirical Research on Growth? It’s Not Factor Accumulation: Stylized Facts and Growth Models.” The World Bank Economic Review 15, no. 2 (August): 177–219. https://doi.org/10.1093/wber/15.2.177.

ECLAC (Economic Commission for Latin America and the Caribbean). 2022. Innovation for Development: The Key to a Transformative Recovery in Latin America and the Caribbean. LC/CCITIC.3/3. Santiago, CL.https://hdl.handle.net/11362/47795.

Edwards, Sebastian. 1995. Crisis and Reform in Latin America: From Despair to Hope. Washington, DC: World Bank.

Edwards, Sebastian. 2010. Left Behind: Latin America and the False Promise of Populism. Chicago: University of Chicago Press.

Faal, Ebrima. 2005. “GDP Growth, Potential Output, and Output Gaps in Mexico.” IMF Working Paper WP/05/93. Washington, DC: International Monetary Fund. https://www.imf.org/external/pubs/ft/wp/2005/wp0593.pdf.

Feenstra, Robert C., Robert Inklaar, and Marcel Timmer. 2013. “The Next Generation of the Penn World Table.” Working paper, Groningen Growth and Development Centre, University of Groningen, the Netherlands, April. https://www.rug.nl/ggdc/docs/the_next_generation_of_the_penn_world_table2013.pdf.

Ferguson, Roger W., and William L. Wascher. 2004. “Distinguished Lecture on Economics in Government: Lessons from Past Productivity Booms.” Journal of Economic Perspectives 18, no. 2 (Spring): 3–28. https://doi.org/10.1257/0895330041371286.

Gallardo-Albarrán, Daniel, and Robert Inklaar. 2021. “The Role of Capital and Productivity in Accounting for Income Differences since 1913.” Journal of Economic Surveys 35, no. 3 (July): 952–74. https://doi.org/10.1111/joes.12374.

Gordon, Robert J. 2016. The Rise and Fall of American Growth. Princeton, NJ: Princeton University Press.

Gutiérrez Cruz, Francisco Salvador, Juan Carlos Moreno Brid, and Joaquín Sánchez Gómez. 2021. “Inversión Pública Y Privada en México: ¿Motores Complementarios Del Crecimiento Económico?” [Public and Private Investment in Mexico: Complementary Drivers of Economic Growth?] El Trimestre Económico 88, no. 352 (October–December): 1043–71. https://doi.org/10.20430/ete.v88i352.1357.

Hanson, Gordon H. 2010. “Why Isn’t Mexico Rich?” Journal of Economic Literature 48, no. 4 (December): 987–1004. https://doi.org/10.1257/jel.48.4.987.

Hanson, Gordon H. 2012. Understanding Mexico’s Economic Underperformance. Washington DC: Migration Policy Institute. https://www.migrationpolicy.org/research/mexico-economic-underperformance.

Hofman, André A., and Patricio Valderrama. 2021. “Long Run Economic Growth Performance in Latin America 1820–2016.” Journal of Economic Surveys 35, no. 3 (July): 833–69. https://doi.org/10.1111/joes.12383.

Hulten, Charles R. 2009. “Growth Accounting.” National Bureau of Economic Research Working Paper 15341, Cambridge, MA, September. http://www.nber.org/papers/w15341.

INEGI. 2021. “Censo de Población y Vivienda 2020.” https://www.inegi.org.mx/programas/ccpv/2020/.

INEGI. 2015. “Estadísticas Históricas de México 2014.” https://www.inegi.org.mx/app/biblioteca/ficha.html?upc=702825058203.

Islam, Nazrul. 2003. “Productivity Dynamics in a Large Sample of Countries: A Panel Study.” Review of Income and Wealth 49, no. 2 (June): 247–72. http://www.roiw.org/2003/247.pdf.

Jorgenson, Dale W., and Zvi Griliches. 1967. “The Explanation of Productivity Change.” The Review of Economic Studies 34, no. 3 (July): 249–83. https://doi.org/10.2307/2296675.

Kehoe, Timothy J., and Kim J. Ruhl. 2010. “Why Have Economic Reforms in Mexico Not Generated Growth?” Journal of Economic Literature 48, no. 4 (December): 1005–27. https://doi.org/10.1257/jel.48.4.1005.

King, Timothy. 1970. Mexico: Industrialization and Trade Policies since 1940 (Industry and Trade in Some Developing Countries). New York: Oxford University Press.

Krueger, Anne, and Aaron Tornell. 1999. “The Role of Bank Restructuring in Recovering from Crises: 1995–98.” National Bureau of Economic Research Working Paper 7042, Cambridge, MA, March. https://doi.org/10.3386/w7042.

Levy, Santiago. 2018. Under-Rewarded Efforts: The Elusive Quest for Prosperity in Mexico. Washington, DC: Inter-American Development Bank.

Lipsey, Richard G., and Kelvin Lancaster. 1956. “The General Theory of Second Best.” The Review of Economic Studies 24, no. 1 (July) 11–32, https://doi.org/10.2307/2296233.

Lustig, Nora. 1998. Mexico: The Remaking of an Economy. Washington DC: Brookings Institution Press.

Lustig, Nora. 2001. “Life Is Not Easy: Mexico’s Quest for Stability and Growth.” Journal of Economic Perspectives 15, no. 1 (Winter): 85–106. https://doi.org/10.1257/jep.15.1.85.

McKinsey Global Institute. 2014. A Tale of Two Mexicos: Growth and Prosperity in a Two Speed Economy. https://www.mckinsey.com/featured-insights/americas/a-tale-of-two-mexicos.

Moreno-Brid, Juan Carlos. 2013. “Industrial Policy: A Missing Link in Mexico’s Quest for Export-Led Growth.” Latin American Policy 4, no. 2 (October): 216–37. https://doi.org/10.1111/lamp.12015.

Moreno-Brid, Juan Carlos, and Jaime Ros. 2009. Development and Growth in the Mexican Economy: An Historical Perspective. New York: Oxford University Press.

Moreno-Brid, Juan Carlos, Jesús Santamaria, and Juan Carlos Rivas Valdivia. 2005. “Industrialization and Economic Growth in Mexico after NAFTA: The Road Travelled.” Development and Change 36, no. 6 (December): 1095–119. https://doi.org/10.1111/j.0012-155X.2005.00451.x.

Núñez Rodríguez, Gaspar. 2006. “Inversión Pública Y Crecimiento Económico en México. Un Enfoque de Contabilidad Del Crecimiento.” [Public Investment and Economic Growth in Mexico. A Growth Accounting Approach.] Perfiles Latinoamericanos: Revista de La Facultad Latinoamericana de Ciencias Sociales, Sede México 13, no. 27 (January): 11–32. https://doi.org/10.18504/pl1327-011-2006.

Organisation for Economic Co-operation and Development. 2024. “Gross Domestic Spending on R&D.” OECD Data. https://www.oecd.org/en/data/indicators/gross-domestic-spending-on-r-d.html.

Palma, José Gabriel. 2011. “Why Has Productivity Growth Stagnated in Most Latin American Countries since the Neo-Liberal Reforms?” In The Oxford Handbook of Latin American Economics, edited by José Antonio Ocampo and Jaime Ros, 568–607. New York: Oxford University Press. http://143.107.92.173/ccpeae/minicursos/2019/Textos%20Minicurso%20Onda%20Rosa%20USP/Palma-2011-Why%20has%20productivity%20growth%20stagnat.pdf.

Prados de la Escosura, Leandro, Tamás Vonyó, and Ilya B. Voskoboynikov. 2021. “Accounting for Growth in History.” Journal of Economic Surveys 35, no. 3 (July): 655–69. https://doi.org/10.1111/joes.12421.

Reinhart, Carmen M., and Kenneth S. Rogoff. 2009. This Time Is Different: Eight Centuries of Financial Folly. Princeton, NJ: Princeton University Press.

RICYT (Red de Indicadores de Ciencia y Tecnología). 2024a. “Gross Domestic Expenditures on R&D as a Percentage of GDP.”  Accessed August 2, 2024.https://app.ricyt.org/ui/v3/comparative.html?indicator=GASTOxPBI&start_year=2012&end_year=2021.

RICYT (Red de Indicadores de Ciencia y Tecnología). 2024b. “R&D Expenditures by Source of Funds.” Accessed August 2, 2024. https://app.ricyt.org/ui/v3/comparative.html?indicator=GASIDSFPER&start_year=2012&end_year=2021. 

Ros, Jaime. 2013. Algunas Tesis Equivocadas sobre el Estancamiento Económico de México [Some Wrong Theses about Mexico's Economic Stagnation]. College of Mexico and National Autonomous University of Mexico.

Ros, Jaime. 2015. ¿Como Salir de la Trampa Del Lento Crecimiento Y Alta Desigualdad? [How to Get out of the Trap of Slow Growth and High Inequality?].College of Mexico and National Autonomous University of Mexico.

Torre Cepeda, Leonardo E., and Luis Fernando Colunga Ramos. 2015. “Patterns of TFP Growth in Mexico: 1991–2011.” North American Journal of Economics and Finance 34 (November): 398–420.

Solow, Robert. 1956. “A Contribution to the Theory of Growth.” Quarterly Journal of Economics 70, no. 1 (February): 65–94.

Solow, Robert. 1957. “Technical Change and Aggregate Production Function.” Review of Economics and Statistics 39, no. 3 (August): 312–20. 

Solow, Robert. 2007. “The Last 50 Years in Growth Theory and the Next 10.” Oxford Review of Economic Policy 23, no. 1 (Spring): 3–14.

Stallings, Barbara, and Wilson Peres. 2000. Growth, Employment, and Equity: The Impact of the Economic Reforms in Latin America and the Caribbean. Washington, DC: Brookings Institution Press.

Stiglitz, Joseph E., and Bruce C. Greenwald. 2014. Creating a Learning Society: A New Approach to Growth, Development, and Social Progress. New York: Columbia University Press. 

Wallace, Robert Bruce. 1980. “Policies of Protection in Mexico.” In Protection and Economic Development in Mexico edited by Adriaan ten Kate and Robert Bruce Wallace,18–88. New York: St. Martin's Press.

Weil, David. 2009. Economic Growth. 2nd ed. London: Pearson Education. World Bank. (1987). World Debt Tables, 1986–87 Edition. Washington DC: The World Bank.

Notes

[1] Richard Feynman, “Seeking New Laws,” Messenger Lectures series on “The Character of Physical Law,” Cornell University, November 9, 1964, https://jamesclear.com/great-speeches/seeking-new-laws-by-richard-feynman.

[2] On the “lost decade” see Jocelyn Sims and Jessie Romero, “Latin American Debt Crisis of the 1980s,” Federal Reserve History, November 22, 2013, https://www.federalreservehistory.org/essays/latin-american-debt-crisis.

[3] Standard critiques of Mexico’s growth failure hypothesize that the labor market is too rigid due to over regulation. José Gabriel Palma (2011) argues the contrary, that the high elasticity of demand for labor allows employers to easily acquire and dismiss workers, holds wages down, and incentivizes a low wage strategy in which Mexico does not have a comparative advantage.

[4] General purpose technologies (GPTs) are ones that have multiple uses and lead to a wide range of applications and sub-inventions.  Examples include steam power and major improvements in steel manufacturing during the late 18th and 19th centuries, electricity and internal combustion engines in the late 19th century through most of the 20th, and information and communication technologies after approximately 1970.

[5] Constant returns to scale is the idea that an equal increase in all inputs leads to an proportionately equal increase in output. For example, if labor and capital inputs are doubled, then output also doubles.

[6] The residual is the part of the output increase that cannot be explained by measured increases in labor or capital inputs. It is generally thought to represent technological changes not captured by increases in the amount of capital used, or a number of other potential factors that increase production efficiency.

[7] Measurements of labor and capital inputs have undergone several advances. Current methods favor estimates of labor and capital services over counts of workers or hours, or cumulated investment values. See Charles R. Hulten (2009) and Nicolas Crafts and Pieter Woltjer (2021) for a discussion of the changes over time to the measurement of labor and capital inputs.

[8] In the growth model, capital and labor are the two explicitly given inputs. Economists refer to them as “factors” or “factor inputs” because they are the factors used in production.

[9] Capital deepening refers to an increase in the amount of physical capital per worker.  For example, capital deepening occurs when construction workers switch from using shovels to mechanized earth moving equipment, or banks replace bank tellers with ATM machines.

[10] Endogeneity refers to changes that are inherent in a process as opposed to changes that occur separately. In the case given, the question is whether the processes that lead to an increase in income also cause an increase in TFP, in which case TFP is said to be endogenous. Alternatively, TFP may not be caused by income growth but instead be the result of other changes in the economy that cause a given amount of capital and labor to be used more efficiently.

[11] See “Maddison Project Database 2020,” Groningen Growth and Development Centre, Faculty of Economics and Business, University of Groningen, last modified May 23, 2022, https://www.rug.nl/ggdc/historicaldevelopment/maddison/releases/maddison-project-database-2020.

[12] See “Long-Term Productivity Database,” accessed November 15, 2023, http://www.longtermproductivity.com/.

[13] Purchasing power parity provides comparable estimates of the real purchasing power of income or GDP in different countries. PPP takes into account differences in prices so that the actual basket of goods that are available to someone in one country can be compared to the basket of goods available to someone in another country. For example, the International Monetary Fund (IMF) estimated that Mexican GDP per person in 2023 was 13,642 when measured in U.S. dollars using market exchange rates. However, that level of GDP per person could buy a basket of goods in Mexico that would have cost $24,980 if purchased in the United States, Mexico’s PPP GDP per person is the second number. See IMF, “World Economic Outlook Database,” April 2024, accessed August 1, 2024, https://www.imf.org/en/Publications/WEO/weo-database/2024/April.

[14] “Porfiriato” refers to the period of Mexican history when Porfirio Díaz was president (1876–80; 1884–1911) or controlled the presidency (1880–84). The Mexican Revolution began in 1910 and is generally described as ending in 1917, although several years of conflict continued.

[15] The term “Mexican Miracle” does not have a precise set of dates. It always includes the 1940s, 1950s, and 1960s and sometimes the 1970s are included. I am including both the 1930s and the 1970s because they were also decades of rapid growth even though national policies varied.

[16] I have set the start date for the current period as 1983 because that is the year that national policies began a fundamental change. Some readers may want to attribute the debt crisis, 1982–89, to the previous period. It does not matter, as I show in the growth estimates below and in the appendix.

[17] OLS, or ordinary least squares, is a standard statistical method for fitting a straight line through a scatter plot of data points. In this case, the scatter plot has the growth rate on the vertical axis and the year on the horizontal. OLS estimates the slope and intercept of the best line through the data points, where the slope is the estimate of the rate of increase or increase in the line and the intercept is the point where it crosses the vertical axis. OLS provides estimated values for the slope and intercept by finding the values that minimize the distances from the line to the actual data points.

[18] It is conceivable that the economic turbulence of the 1980s is driving this result, so a second set of estimates were made with a different cut-off point for the Golden Age. If the debt crisis was a result of long-run policies implemented through the years of high growth, as some critics allege, then 1983–88 should be included in those years. Recall that 1989 was the end of the debt crisis and while reforms were attempted throughout the 1980s, the constraints imposed by the crisis limited the ability of policymakers to implement them. In addition, despite the relatively unchecked power of Mexican presidents to make policy before the democratic reforms of the 1990s, dramatic changes in economic policy still required time to build consensus. Consequently, an alternative view is that 1989 was the year when Mexico’s free market reforms were no longer hampered by the constraints imposed by the crisis. It was also several years after a relatively broad political consensus on the need for a new development model was reached. To test whether including the bulk of the 1980s into the current period is driving its negative growth pattern, a second set of estimates of average annual compound growth breaks the sample into the same first period, 1890–1932, a longer second period, 1933–88 which includes the debt crisis years, and a post-debt crisis third period, 1989–2022. The results (shown in the appendix) indicate that it does not matter whether the debt crisis years are included in the high growth period or the neoliberal period; the differences in growth rates between the two periods continue to be significantly different and negative in the current period.

[19] General purpose technologies are usually considered enabling technologies in that they are not designed to solve specific problems, but their creation leads to a large number of sub-inventions that have that purpose.

[20] Complications arise because we live in an imperfect world in which making the labor market or other markets more flexible and reducing regulations might actually worsen overall conditions (Lipsey and Lancaster, 1956). Note that Palma (2011) argues Mexico’s labor market is too flexible. High employment elasticities imply that firms can hire and fire easily, filling gaps in labor needs or shedding surplus labor. This reduces pressure on firms to hold on to workers and to pay better wages and is consistent with a low wage strategy. Mexico, however, is not competitive against countries where wages are much lower and the supply of unskilled labor is relatively greater.

[21] A new industry was one where all products were imported and a necessary industry was one that supplied some part of the domestic market but less than 80% These laws had various names, such as the Transformation Industry Law (1941), the Law of Promotion of Transformation Industries (1946), and the Law of Promotion of New and Necessary Industries (1955).

[22] CONACYT was a transformation of the National Institute for Scientific Research which was created in 1960 to promote science education. In 2023 it became the National Council for the Humanities, Science, and Technology or CONAHCYT.

[23] The Mexicali Valley, for example, became one of the largest producers of cotton in the world, and responsible for a significant share of Mexico’s export revenues. It also eliminated the raw material bottleneck that was holding back the textile industry (Cerutti and Almaraz, 2013).

[24] The simple average of the annual growth of Mexican exports was -1% in the 1950s, 4.2% in the 1960s, and 14% in the 1970s (INEGI 2015, Table 17.1). By comparison, world exports grew 6.9, 9.0, and 20.7% in the 1950s, 1960s, and 1970s, respectively (Reinhart and Rogoff, 2009; data downloaded from https://scholar.harvard.edu/rogoff/time-different—data-files).

[25] A sudden stop crisis occurs when there is a sudden cessation of foreign capital inflows that are necessary for financing a trade deficit. See Guillermo A. Calvo (1998).

[26] At the time of the Latin American Debt Crisis, debt levels in a number of other noncrisis countries exceeded the level in Mexico and other Latin American defaulters. For example, in 1980, the ratio of debt-to-GDP was 32% in Mexico, and 49% in South Korea. But the ratio of debt-to-exports was 131% in Korea and 233% in Mexico (World Bank 1987). 

[27] Nevertheless, there do seem to be bottlenecks since expanding an existing electrical connection can be slow and expensive. It is conceivable that the GPTs of the Second Industrial Revolution were deployed sufficiently to raise productivity up until around 1980, but not completely, and gaps remains. This is a hypothesis that requires a deeper level of historical analysis.

[28] Rent seeking refers to efforts by firms or individuals (including government officials) to gain a larger share of the economy for themselves without attempting to increase the total amount available. For example, when firms lobby for subsidies or protection, they use resources on the effort that become unavailable for expanding output.

 

 

This material may be quoted or reproduced without prior permission, provided appropriate credit is given to the author and Rice University’s Baker Institute for Public Policy. The views expressed herein are those of the individual author(s), and do not necessarily represent the views of Rice University’s Baker Institute for Public Policy.