The Role of Development and Commercial Banking in Promoting Economic Growth in Mexico: A Sectoral Analysis

Abstract

This study examines the impact of credit extended by both commercial and development banks on exports, GDP, and FDI across different sectors in Mexico. The research lies within the broader context of financial sector development and its role in promoting regional development and economic growth. We use panel data techniques to compare traditional models that include only commercial bank credit with models incorporating development bank credit. Our results emphasize the different and complementary roles of commercial and development banks. The latter is crucial in addressing market failures and supporting long-term, socially and economically beneficial projects. The study concludes that development bank credit should be integrated into economic studies to better understand its contribution to sustainable growth. It suggests that policymakers should coordinate the roles of both banking sectors to foster a balanced and resilient economy.

1. Introduction

Economic development is intrinsically linked to access to finance, particularly credit. While commercial banks play a crucial role in providing credit, development banks offer a unique approach to financing socially beneficial but riskier projects. The main contribution of this paper is to examine the simultaneous impact of development banking alongside commercial banking on various economic dimensions, including foreign direct investments (FDI), exports, and gross domestic product (GDP) in Mexico. This dual perspective provides a more complete understanding of how these distinct sources of credit influence economic outcomes across primary, secondary, and tertiary sectors. By making an explicit connection between each bank type’s objective function—commercial banks’ profit-driven, short-term goals versus development banks’ long-term, socially focused objectives—our study offers insight into how targeted credit strategies can promote sector-specific growth, with wider implications for economic policy and credit allocation in emerging economies.

We explore a dataset that collects credit supplied by commercial and development banks in Mexico from 2007 to 2021, and disaggregated at the state level. By analyzing the relationship between development bank credit and these economic indicators, we contribute to the debate about the role of banks in fostering economic growth and development.

A large body of literature has investigated credit’s importance for economic development, but opinions are mixed on its effective impact. A country’s economic potential depends not only on the availability of resources and technology but also on access to financing. Credit can help growth by enabling firms to invest in new projects, expand their markets, and improve efficiency (Popov and Udell 2012; Grandi and Bozou 2018; Leon 2015; Beck et al. 2004).

However, according to some economists, the role of credit in promoting growth has been exaggerated (Chandavarkar 1992; Fama 1980). The argument is that, if markets are perfect, the amount of credit supplied is optimal. It follows that any exogenous increase in credit would not cause further improvements to a regional economy (see, for instance, Lucas 1972). On the other hand, other authors pointed out that if the market is imperfect, exogenous increases in credit could help counteract market failures and increase economic performance (Stiglitz and Weiss 1981; Stiglitz 2002), therefore determining a significant impact of credit on growth (Levine et al. 2000).

These arguments tend to consider only commercial loans, which are extended from private banks whose objective is to maximize profits. Evidence about the impact of this channel of credit on economic performance has been mixed and inconclusive. Some studies connect credit to high economic growth (King and Levine 1993; Beck et al. 2000; Xu 2000; Rajan and Zingales 1998; Guiso et al. 2004; Musso and Schiavo 2008; Kendall 2012; Hassan et al. 2011). Darrat et al. (2006) investigate the impact of financial-sector development on real economic activity in emerging markets, finding that while financial deepening fosters sustained long-term growth across sectors, its short-term effects are minimal or inconsistent. This implies that for financial development to meaningfully stimulate growth, improvements must be sustained over time. Moyo and Le Roux (2020) point out that the benefits of financial markets need to be compounded with the larger exposure to crises. On the other hand, Christopoulos and Tsionas (2004) argue that government restrictions to the banking system are an obstacle to growth. In Mexico, while Arestis and Demetriades (1999) measure a positive role of credit on economic growth, Venegas-Martínez et al. (2009) and Tinoco-Zermeño et al. (2014) only find a modest long-term impact and not a short-term one.

Given that the influence of credit interacts with the effectiveness of a market, which is intrinsically related to local institutional, cultural, and geographic conditions, researchers should be careful about generalizing findings obtained from particular studies focused on specific realities. Instead, an attempt should be made to expand the scope of these studies to diverse markets and to collect enough perspectives to gather a global understanding of the topic.

For this purpose, researchers should consider that emerging economies often include an alternative channel of credit presented by development banks, whose objective is not exclusively to reach profit maximization but include different social goals. In emerging economies, development banking is crucial to financing socially beneficial but riskier and longer-term projects (Kilpatrick and Williams 2021; Pradhan et al. 2017; Schclarek et al. 2023). These banks are financial institutions that provide capital for economic development projects on a non-commercial basis (Cipoletta Tomassian and Abdo 2022; Tsvirko 2021). They are often established and owned by governments or non-profit organizations to finance projects that would otherwise not be able to get financing from commercial lenders. They can promote private investment opportunities, support socially responsible investing and contribute to providing public goods1 that promote growth but are not necessarily priced by the market (Mazzucato and Penna 2016).

As a consequence, a complete analysis of the impact of credit should include the different roles played by both commercial and development banks in promoting investments, with particular emphasis on their different scopes. Emerging economies offer an advantage for such analysis, as they present more variability in the types of credit available, with stronger differences between the objectives of commercial and development banks. This variability allows for a deeper examination of how different objective functions affect the allocation of credit and its impact on economic growth. By studying these dynamics, we can better understand the interactions between credit type and economic outcomes.

In Mexico, investment is crucial for commercial participation in domestic and foreign markets and as a driver of economic competitiveness and development at regional and national levels (Delgado et al. 2012). However, many productive agents do not possess sufficient financial resources, forcing them to rely on bank credit to sustain their operations, particularly during challenging periods such as the crises of 2008 and 2020. In a recent study on the impact of commercial credit on Mexico’s regional economies, Flores-Segovia and Torre-Cepeda (2024) find that increases in capital flows through credit to regional sectors for investment in productive projects stimulate productivity, enhance regional competitiveness and drive economic growth. In this article, we extend their framework by incorporating development bank credit to assess the regional impact of credit, focusing on the sectors receiving the credit.

Our analysis reveals that development banks are essential in Mexico, with notable impacts on exports and GDP per capita. Development bank credit, particularly in the primary sector, boosts exports by fostering long-term investments that improve agricultural productivity and align with international standards. In contrast, commercial bank credit appears to have a mixed effect, with a surprisingly negative impact on exports from the primary sector, probably because of the short-term nature of the credit provided.

For GDP per capita, our results indicate that development bank credit to the secondary and tertiary sectors supports real economic growth, emphasizing the importance of long-term, socially oriented investments that generate positive externalities. On the other hand, commercial bank credit seems to be in equilibrium with market demand, with no significant additional impact on GDP, suggesting that the supply of credit in the commercial sector may already be sufficient to meet present investment needs. Development bank credit does not influence FDI significantly. Overall, commercial bank credit seems more aligned with FDI because it is targeted to the profit-driven, short-term financial needs that foreign investors typically have.

Our results emphasize the distinction between commercial and development banking, suggesting that the objective functions of these institutions significantly influence their impact. Some of our findings are better rationalized if we accept that commercial banks focus on short-term profit maximization, offering credit that aligns with market demand and directed at financing less risky, shorter term projects. In well-functioning markets, this channel tends not to stimulate additional economic activity, except in particular cases that need to be contextualized. In contrast, development banks pursue wider social objectives, financing projects that may be ignored by commercial lenders. Their presence in emerging markets like Mexico creates a complementary dynamic that could promote growth in underserved areas. It is important to stress that these properties are not generalizable, as they are intrinsically linked to the particular conditions of each sector and region. A context-specific analysis should be used to identify areas of opportunity where development banking can make a meaningful impact.

We recognize the need for strategic support from government institutions to ensure that development banks receive the necessary resources to target settings that could drive long-term growth, particularly when they generate positive externalities. By adjusting the operation of development credit according to local conditions, policymakers can more effectively create both economic and social value.

2. Materials and Methods

2.1. Theoretical Framework

Different economic theories offer varying perspectives on the role of credit in shaping real economic outcomes. Classical economic models, grounded in the notion of market equilibrium, establish that credit has a neutral impact on long-term growth. In these frameworks, the supply of credit is seen as a response to existing demand, facilitating the allocation of resources without fundamentally altering economic dynamics. This view aligns with the loanable funds theory, which posits that credit markets balance savings and investment, leading to an efficient allocation of resources when markets are free from distortions. Barro (1974) supports the Ricardian equivalence hypothesis, which states that government debt does not affect aggregate demand. While not addressing the topic of banking credit, this implies that consumers anticipate future tax increases to pay off the debt, offsetting the positive impact of government spending on their current consumption.

However, Keynesian (Keynes 1936) and post-Keynesian theories argue that credit can actively influence economic activity, especially in the short run, by stimulating aggregate demand and investment (see, for instance, Bernanke and Gertler 1995). They emphasize the role of credit expansion in reducing unemployment and addressing recessions, with banks acting as intermediaries that can inject liquidity into the economy. In a recent article, Sapriza and Temesvary (2024) suggest that the bank credit channel becomes more effective during economic downturns, amplifying the impact of monetary policy on lending when growth is weak.

While these theories look at the short-term, development economics perspectives highlight the importance of directed credit in fostering long-term structural growth. These theories argue that credit can be a tool for addressing market failures, supporting sectors that are underserved by private lenders due to high risks or long payback periods. Understanding these differing theoretical perspectives is essential for analyzing how commercial and development banks contribute to economic growth in varying institutional and regional contexts.

2.2. Institutional Setting

In the 1930s, Mexican development banking was strongly influenced by Keynesianism. Moreover, around those years, Schumpeter’s (1934) ideas regarding the role of the financial system in economic growth also gained prominence, establishing the existence of development banking as part of the orthodoxy for many years.

In the Mexican economy of the post-war, where various markets were not fully developed, sectors such as infrastructure and certain industrial areas, which required significant financing and had project with long maturation periods, were not served by private financial intermediaries (Ortiz Mena 1998; Marichal 2004). As a consequence, Mexico’s economic growth and industrialization required an appropriate channel for credit.

Some of the first institutions created were the Banco Nacional de Crédito Agrícola in 1926 and the Banco Nacional de Crédito Ejidal in 1935, which were later merged in 1975 into the Banco Nacional de Crédito Rural to avoid duplication.

Mexican modern development banks are majority or fully government-owned financial institutions that allocate long-term resources to support development (Huidobro Ortega 2012). They can be categorized by their primary focus sector. However, it is important to note that these institutions often operate across multiple sectors, leading to overlapping areas of influence.

-

Primary sector: Banco Nacional del Comercio Exterior (BANCOMEXT), Sociedad Hipotecaria Federal (SHF), Financiera Rural (FR);

-

Secondary sector: Nacional Financiera (NAFIN), Banco Nacional de Obras y Servicios Públicos (BANOBRAS), Banco Nacional del Comercio Exterior (BANCOMEXT), Sociedad Hipotecaria Federal (SHF);

-

Tertiary sector: Banco Nacional de Obras y Servicios Públicos (BANOBRAS), Banco del Ahorro Nacional y Servicios Financieros (BANSEFI), Banco Nacional del Ejército, Fuerza Aérea y Armada (BANJERCITO).

2.3. Dataset

We explore a longitudinal credit dataset from commercial and development banks covering each of the three economic sectors. The data were sourced from Mexico’s Central Bank (Banxico) and are publicly available. Links to the dataset and a public repository containing our compact version and code are provided in the Data Availability Statement at the end of this manuscript. We used trimestral information from 2007 to 2021, covering the 32 federal entities of Mexico. The primary dependent variables in our study are Exports, GDP, and FDI. Building on the work of Flores-Segovia and Torre-Cepeda (2024), we include the following explanatory variables as regressors: government expenditures, investments in infrastructure, investments in education, and inflation. We add credit extended by commercial and development banks to each of the three sectors to the regressors. We then normalize all variables (except for GDP per capita) by dividing them by GDP and then taking their logarithms. Because of this transformation, the estimated coefficients will be interpreted as elasticities (see Flores-Segovia and Torre-Cepeda 2024). Finally, all variables were standardized so that their coefficients could be directly compared.

Although the credit dataset is the best available for Mexico and offers a novel perspective, a few considerations must be acknowledged. First, our sample is unbalanced, as information about credit from development banks to the primary sector is unavailable for all federal entities and periods. Although unbalanced panels can introduce issues such as missing data bias, we have carefully examined the nature of the missing data. There is no evidence to suggest the presence of selection bias. Instead, the gaps in data are due to factors such as inconsistencies in reporting and data collection due to confidentiality rather than systematic biases.

To evaluate the impact of using the unbalanced panel, we compared the results from the standard approach, which employs a balanced panel excluding development credit, with those from our model that uses the unbalanced panel, including development credit. Additionally, we compared the unbalanced panel model with one that excludes only the critical variable: development credit to the primary sector. The results and interpretations do not differ significantly across these models. This consistency supports our idea that the missing data are likely random and that the unbalanced panel is justified for the analysis.

The different credit variables in our dataset are correlated, leading to multicollinearity that can introduce additional noise to our residuals and reduce the significance of our estimates. We conducted a regression analysis on the credit allocated to the primary sector by commercial banks and calculated a Variance Inflation Factor (VIF). The results revealed that all forms of credit from commercial banks are highly correlated, which is reasonable considering that commercial banks typically diversify their loan portfolios across different sectors. As a result, shocks affecting a particular bank are likely to be distributed across the various areas of its portfolio.

In contrast, credit to the primary sector from development banks shows a weaker correlation with other variables, hinting that this sector is where the roles of commercial and development banks in Mexico diverge the most. The primary sector’s lower profitability probably causes the market to focus on short-term, less risky projects, which creates a gap that development banks can fill by focusing on longer-term investments.

Going back to the problem of multicollinearity, it is important to recognize that our estimates will exhibit lower significance than the true ones. Although it is often recommended to remove some of the correlated variables to address multicollinearity, doing so does not eliminate the problematic variables. These variables, in fact, enter the error term, potentially violating the strict exogeneity assumption due to omitted variable bias. This problem is more severe than multicollinearity itself. To evaluate the impact of this strategy, we checked for significant changes when removing the most correlated commercial credit variables. Our results showed no significant changes in the estimates or their significance levels, pointing to the fact that the potential omitted variable bias is not severe. However, it also suggests that removing a few variables cannot simply fix multicollinearity.

For these reasons, we kept all credit-related variables in our full model, recognizing that the significance of some estimates might be lower. This approach allows us to preserve the integrity of our analysis while minimizing the risk of introducing more severe biases.

As a last comment, we identified four instances where the foreign direct investment (FDI) values were negative. According to Banxico’s dataset documentation, negative values indicate net outflows of financial resources from Mexico. Due to the lack of context and to the limited number of such cases, we decided to omit them from the FDI analysis.

2.4. Methodology

We employ panel data techniques to address the longitudinal structure of our sample. First, we conducted a Harris–Tzavalis test to detect the potential presence of unit roots in our panel. This test is more appropriate for scenarios where the time dimension is relatively short compared to the number of cross-sectional units, which is consistent with our dataset. The test results allowed us to reject the null hypothesis, eliminating the need to compute first (or second) differences in the critical variables.

It is known that a standard regression will suffer from endogeneity issues because the various federal entities present unobserved, time-invariant heterogeneity. As this heterogeneity enters the error term, it will make the residuals correlated over time and possibly with some of the exogenous variables. For this reason, the use of panel data techniques is required to obtain unbiased estimates. We estimated the models using random and fixed effects and then applied a Hausman test to determine the most appropriate method. Across all models, the Hausman test pointed to the use of fixed effects.

The fixed effects model removes the unobserved, time-invariant heterogeneity by demeaning all variables through the within transformation. Unfortunately, this process excludes all time-invariant variables from the regression, such as regional dummies. On the other hand, it ensures that the results are unbiased. Finally, after selecting the appropriate method, we re-estimated the model with fixed effects using robust standard errors.

3. Results

First, we report our results on correlations between credit-related variables.

Table 1. Matrix of correlations.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
(1) BComercSecPrim	1.000
(2) BdeDSecPrim	0.255	1.000
(3) BComercSecSec	0.509	0.181	1.000
(4) BdeDSecSec	0.448	0.192	0.955	1.000
(5) BComercSecTer	0.516	0.194	0.975	0.972	1.000
(6) BdeDSecTer	0.406	0.127	0.899	0.952	0.947	1.000

illustrates how various measures of credit move jointly. Development credit to the primary sector is the least correlated variable. However, the remaining measures present high degrees of covariation. As a consequence, our regression estimates will suffer from noise and appear less significant than they probably would be. On the other hand, the estimates will not be biased.

Next, we study the impact of different sources of credit on three measures of economic performance. We present the results from these three analyses separately, and then we will comment on them.

3.1. Export

In the first analysis, we focus on the impact of credit on normalized (logarithmic) exports. The model is as follows:

$${\log \left({\displaystyle \frac{Exports}{GDP}}\right)}_{it}={\alpha }_i+X_{it}^{\prime }\beta +{DummyRegion}_i^{\prime }\theta +{Credit}_{it}^{\prime }\gamma +{\mu }_{it},$$

(1)

where ${\alpha }_i$ represents unobserved heterogeneity, and ${\mu }_{it}$ is an idiosyncratic error term. As mentioned earlier, matrix X contains explanatory variables such as infrastructure, public expenditure, investments in education, and inflation. In the original runs, we controlled for regional differences; however, these variables disappear after taking fixed effects (nevertheless, they are essential when running the random effects alternative for comparison). Credit contains loans from commercial and development banks, to each of the three productive sectors, for a total of six variables. More specifically, the variables of commercial credit are BComercSecPrim, BComercSecSec and BComercSecTer, corresponding to the primary, secondary, and tertiary sectors, respectively. The main contribution of our study is the introduction of credit variables from development banks: BdeDSecPrim, BdeDSecSec, and BdeDSecTer. We also add a variable to control for the 2008–2009 financial crisis.

We initially ran a Harris–Tzavalis unit-root test on the dependent variable and found it stationary. At a p-value of 0.0000, we could reject the null hypothesis that the data contained a unit root. We then estimate the model using both fixed and random effects. To pick the correct option, we ran a Hausman Test. The test returned a p-value of 0.0465, indicating the need to use fixed effects.

The results of the fixed effect estimates are presented in

Table 2. Impact of credit on exports.

	(1) logExport	(2) logExport	(3) logExport
Public Expenditure	−0.014	0.048	0.005
	(0.027)	(0.037)	(0.032) 2
Infrastructure	0.362 ***	0.569 **	0.246 ***
	(0.113)	(0.205)	(0.089)
Education	−0.09	−0.008	−0.106
	(0.071)	(0.151)	(0.071)
Inflation	0.004	0.01	0.002
	(0.01)	(0.014)	(0.012)
dummyCrisis	−0.094 ***	−0.065	−0.086 **
	(0.033)	(0.038)	(0.032)
BComercSecPrim	−0.02	−0.252 **	−0.059
	(0.076)	(0.091)	(0.07)
BComercSecSec	−0.012	−0.025	0.13
	(0.099)	(0.111)	(0.088)
BComercSecTer	0.484 ***	0.475 **	0.478 ***
	(0.103)	(0.176)	(0.098)
BdeDSecPrim		0.117 ***
		(0.035)
BdeDSecSec		−0.111	−0.052
		(0.072)	(0.053)
BdeDSecTer		0.003	−0.001
		(0.034)	(0.029)
Observations	438	86	315
R-squared (Within)	0.855	0.935	0.891

² Robust standard errors are in parentheses. *** p < 0.01, ** p < 0.05.

. We compare the traditional models, which include only measures of commercial credit, with a model that also incorporates development credit. The findings for the traditional model are shown in the first column, while our extended model is detailed in column 2. As mentioned earlier, while our dataset is mostly balanced, one critical explanatory variable—credit from development banks to the primary sector—exhibited some missing observations across certain federal entities and periods. To address the potential impact of these missing data, we compared the results obtained from the full unbalanced dataset (column 2) with those derived from a balanced dataset where the critical variable was excluded (these outcomes are shown in column 3). The similarity in results between these two approaches confirms that the missing observations are probably random and do not introduce bias into our results. This consistency reinforces our analysis’s validity regardless of the dataset’s unbalancedness.

Fixed-effect models control for individual-specific heterogeneity by including a dummy variable for each individual. This means that the model explains the variation in the dependent variable within individuals, not between individuals. For this reason, we report only the within R-squared, which provides a better measure of the model’s explanatory power in the context of these models, as it focuses on the most interesting variation to researchers.

At first glance, we can immediately see that including credit from development banks improves the model, emphasizing this channel’s critical role in promoting exports from the primary sector. It suggests that development banking is a significant factor in favoring regional growth in Mexico and should not be ignored in future studies. In the next paragraphs, we discuss the impact of each significant variable separately.

Public expenditure on infrastructure influence exports positively. This is intuitive, as better-connected locations favor trade, allowing products to be transported at lower costs than in other areas. The financial crises of 2008 and 2009 negatively impacted exports, but their significance was somewhat obfuscated by the extra noise introduced by multicollinearity following the inclusion of development banking credit variables.

Credit provided by development banks to the primary sector positively affects exports. Agricultural products are typically produced by small farms operating at a subsistence level. Because this sector often lacks substantial profits or collateral, it is better targeted by development banks, which play a crucial role in improving regional growth. Interestingly, credit from commercial banks to the primary sector seems to work against exports. Although this may seem puzzling, we can rationalize it as follows.

Development bank loans are aimed at long-term investments that improve productivity and sustainability in agriculture. It could include financing for advanced agricultural technologies, infrastructure development (such as irrigation systems), or sustainable farming practices that help farmers move away from subsistence or low-value crops. These investments can promote efficiency, higher yields, and better quality produce, making the agricultural sector more competitive in international markets and increasing exports. Additionally, development banks might provide technical assistance or subsidies encouraging farmers to improve practices aligned with international standards, all factors that contribute to their export potential.

In contrast, commercial banks may provide credit primarily for short-term operational needs, such as purchasing inputs (seeds, fertilizers) or working capital. While this is important for immediate production needs, it might not contribute significantly to long-term improvements in productivity or competitiveness. It could lead to focus on immediate cash crops or quick turnover activities less aligned with export markets. In some cases, this type of credit might even lead farmers to take on more debt without a proportional increase in their ability to repay, potentially leading to financial stress and a focus on local markets. These farmers will then prioritize short-term financial obligations over making the types of investments needed to maintain or increase export levels. It is important to stress that this behavior seems to fit Mexico’s regional economy but cannot be generalized. For instance, Peng et al. (2020) document a significant role of commercial credit on regional agricultural growth in China.

The net impact of credit on export to the primary sector is negative (the variables were standardized before the analysis, which makes the coefficients comparable), probably due to the larger presence of commercial credit over development.

None of these types of credit affect the secondary sector, suggesting that this market presents fewer problems.

Finally, exports from the tertiary sector are positively boosted by commercial banks, but development ones do not significantly impact them. The tertiary sector in Mexico includes several activities appropriate for export. Tourism and hospitality are considered exported as they represent capital flows from foreign sources to domestic ones. In addition to that, call centers and customer support businesses, IT services, film and television, music, air and maritime transport, and higher education provide services to the Latin American or US market. As all these businesses fit well in the portfolio of commercial banks, thanks to their revenue streams and lower capital requirements, it is no surprise that they are less affected by development banks. Interestingly, the significant impact of commercial credit points to the fact that this sector is probably not in equilibrium with the money supply market and that expansive credit policies benefit it.

3.2. Foreign Direct Investment

In the second model, we explore the impact of credit on determining foreign direct investment (FDI) as component of economic growth2. We estimate the following model:

$${\log \left({\displaystyle \frac{FDI}{GDP}}\right)}_{it}={\alpha }_i+X_{it}^{\prime }\beta +{DummyRegion}_i^{\prime }\theta +{Credit}_{it}^{\prime }\gamma +{\mu }_{it}.$$

(2)

We conducted an initial Harris–Tzavalis unit-root test on FDI, and at a p-value of 0.0000, we rejected the null hypothesis that the data contained a unit root. The Hausman Test produced a p-value of 0.0037, pointing once more to using fixed effects.

The results of our estimations are reported in

Table 3. Impact of credit on foreign direct investments.

	(1) logFDI	(2) logFDI	(3) logFDI
Public Expenditure	−0.018	0.018	−0.003
	(0.057)	(0.053)	(0.064) 3
Infrastructure	0.702 ***	0.745 **	0.441 **
	(0.193)	(0.266)	(0.186)
Education	−0.296 ***	−0.444 **	−0.429 ***
	(0.094)	(0.174)	(0.091)
Inflation	0.015	0.078 **	0.014
	(0.02)	(0.029)	(0.017)
dummyCrisis	−0.328 ***	−0.277 ***	−0.41 ***
	(0.076)	(0.084)	(0.075)
BComercSecPrim	0.122	0.284 **	−0.045
	(0.126)	(0.107)	(0.122)
BComercSecSec	−0.033	−0.229	0.252
	(0.138)	(0.265)	(0.166)
BComercSecTer	0.398 ***	0.378	0.447 ***
	(0.122)	(0.273)	(0.118)
BdeDSecPrim		0.163
		(0.103)
BdeDSecSec		−0.097	−0.118 **
		(0.106)	(0.045)
BdeDSecTer		−0.013	0.044
		(0.078)	(0.046)
Observations	465	90	332
R-squared (Within)	0.663	0.830	0.732

³ Robust standard errors are in parentheses. *** p < 0.01, ** p < 0.05.

:

We observe that investments in infrastructure promote FDI, while economic crises tend to discourage foreign investments. Interestingly, the level of education has a negative impact on FDI. It is a new and counterintuitive result that does not have an easy explanation. One possible rationale could be that increases in education expenditure may raise concerns among foreign investors about the sustainability of Mexico’s fiscal policies, even after accounting for ordinary public expenditure. Moreover, the benefits of investing in education will probably materialize only in the distant future, which does not align with the relatively shorter-term horizons of foreign companies. These are, however, speculative explanations, and further research is necessary to understand this effect fully.

Inflation is positively correlated with FDI. High inflation often results in the depreciation of the local currency—in this case, the Mexican peso. This depreciation can make Mexican assets more affordable for foreign investors who hold capital in stronger currencies (e.g., U.S. dollars, euros), creating opportunities to acquire assets, real estate, or businesses at lower costs. This potential for lower-cost acquisitions could increase FDI as investors seek to take advantage of favorable exchange rates.

We now focus on the sectors that attract FDI to Mexico. The agribusiness sector, particularly in food processing and beverage production, has been attracting FDI, with companies like Nestlé, Coca-Cola, and PepsiCo investing in production facilities.

Mexico’s automotive manufacturing sector attracts substantial FDI from companies such as Ford, General Motors, Volkswagen, and Toyota. The electronics industry, producing televisions, computers, and other consumer electronics, has seen strong investments from companies like LG, Samsung, and Sony. In addition to that, the aerospace sector has experienced growing FDI, with firms like Bombardier, Safran, and Honeywell building manufacturing facilities.

FDI has also flowed into Mexico’s telecommunications sector, with major investments from companies like AT&T and América Móvil. Global retail giants, like Walmart, Costco, and Carrefour, have invested heavily in Mexico’s retail sector, creating extensive networks of stores and distribution centers across the country. Amazon has also invested in logistics, warehouses, and technology infrastructure to support its operations.

Mexico’s tourism sector, especially in popular destinations like Cancún, Los Cabos, and Riviera Maya, has attracted FDI in the construction and operation of hotels, resorts, and related infrastructure. Foreign investors are also actively investing in real estate development for both commercial and residential properties in these tourist spots.

Our results indicate that while development credit improves the model’s fit, it does not significantly impact FDI. Foreign investors are primarily driven by profit, and FDI in Mexico is more closely related to commercial bank credit because these institutions are better aligned with the profit-driven, commercially focused nature of FDI projects. Commercial banks offer the financial products, risk management capabilities, and sectoral expertise sought by foreign investors, making them the preferred choice for financing FDI ventures. In contrast, though essential for national development, development banks typically focus on projects with broader social and economic impacts. These projects may not align with the immediate commercial objectives of FDI and often generate positive externalities not captured by the market. As a result, they may be less appealing to foreign investors.

3.3. GDP per Capita

In our last model, we measure the impact of credit directly on (the log of) GDP per capita. Following Flores-Segovia and Torre-Cepeda (2024), we augment the model by including the (normalized and logarithmic versions of) exports and the lagged value of GDP per capita among the explanatory variables.

The Harris–Tzavalis unit-root test produced a p-value of 0.0000, leading us to reject the null hypothesis of a unit root. We computed the Hausman Test to select the appropriate panel data technique. The test had a p-value of 0.0000, dictating again the use of fixed effects.

Our estimates are collected in

Table 4. Impact of credit on GDP per capita.

	(1) logGDPpc	(2) logGDPpc	(3) logGDPpc
logGDPperCapitan1	0.033	−0.019	0.009
	(0.049)	(0.026)	(0.045) 4
logExport	0.162	−0.121	−0.04
	(0.164)	(0.185)	(0.223)
Public Expenditure	0.025	0.101	0.03
	(0.029)	(0.073)	(0.038)
Infrastructure	−0.159	−0.145	0.074
	(0.145)	(0.156)	(0.222)
Education	0.441 ***	0.541 ***	0.582 ***
	(0.068)	(0.169)	(0.088)
LInflation	0.066 ***	0.008	0.066 ***
	(0.012)	(0.024)	(0.014)
dummyCrisis	−0.087 **	0.03	−0.028
	(0.038)	(0.057)	(0.035)
BComercSecPrim	−0.013	0.121	0.127
	(0.144)	(0.164)	(0.132)
BComercSecSec	−0.172	−0.001	−0.321 *
	(0.11)	(0.142)	(0.171)
BComercSecTer	0.131	−0.086	0.157
	(0.122)	(0.173)	(0.159)
BdeDSecPrim		−0.016
		(0.072)
BdeDSecSec		0.135 *	−0.03
		(0.074)	(0.074)
BdeDSecTer		0.129 **	0.006
		(0.049)	(0.027)
Observations	468	90	332
R-squared (Within)	0.386	0.651	0.4

⁴ Robust standard errors are in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

:

Education positively affects GDP per capita, which is consistent with theories on human capital (Lucas 1988). The regression that omits development banking shows that the coefficient of inflation is significantly positive, an effect that disappears when the additional variables are included. Rather than interpreting this as a causal relationship, we should view it as a correlation. Periods of high growth, or regions that grow faster, tend to experience higher inflation.

The commercial banking sector is in equilibrium with the market, as none of the coefficients for credit provided by it are statistically significant. This result is compatible with the strand of literature that argues that the supply of credit is generally in equilibrium with the demand necessary to finance private projects. When the commercial banking sector reaches a level where the demand for credit is fully satisfied, additional credit may not find productive use. Businesses that require credit already have access to it, while those that do not borrow may lack viable projects or perceive insufficient returns to justify the cost of borrowing. In this case, additional credit would not lead to increased investment or production and thus would not contribute to GDP growth. When the credit market is saturated, additional credit supplied by commercial banks may neutralize its effect on GDP. Instead of stimulating new investment, it might simply increase reserves, savings, or speculative investments that do not contribute to economic activity.

However, a different picture emerges when we control for credit granted by development banks. When these institutions offer credit to the tertiary and secondary sectors (the latter being significant only at the 10% level, which might be acceptable if we consider the additional noise from multicollinearity), they give an impulse to real economic growth. The reason probably lies in the long-term horizon and social objectives of development banking, which supports projects contributing to the economy through externalities, such as public goods not priced by the private market.

3.4. Robustness Analysis

We expand the scope of our analysis to investigate if there are particular interactions between credit and expenditure in infrastructure. This is achieved by creating additional explanatory variables which include the product of a particular measure of credit and infrastructure. We label these new variables as Interaction_Priv# if they include commercial bank loans, and Interaction_Dev# if they refer to development bank credit. The results of the extended regression are portrayed in

Table 5. Introduction of interaction terms in the analysis of exports.

	(4) logExports
Interaction_Priv1	−0.005 5
	(0.039)
Interaction_Priv2	−0.121 **
	(0.053)
Interaction_Priv3	0.112 *
	(0.058)
Interaction_Dev1	0.033 **
	(0.015)
Interaction_Dev2	0.043
	(0.04)
Interaction_Dev3	0.112 **
	(0.04)
Observations	86
R-squared (Within)	0.959

⁵ Robust standard errors are in parentheses ** p < 0.05, * p < 0.1.

. As this is just a supplementary analysis, for reasons of space, we do not report again all results from the regression. The introduction of interaction terms does not change the qualitative outcomes measure before but introduces more interesting effects.

To maintain interpretability, we only ran the fixed effect version of the regression. The interaction terms improve the fit of the panel regression. Moreover, they add richness to our analysis. Commercial credit to the secondary sector interacts negatively with expenditure in infrastructure (Interaction_Priv2); that is, as infrastructure spending increases, the positive impact of commercial credit to the secondary sector on Exports diminishes. When infrastructure is low, commercial credit directed toward the secondary sector might have a relatively strong positive impact on economic outcomes. However, as infrastructure improves, the marginal benefit of this credit declines. This could happen because the infrastructure improvements themselves reduce the need for additional financing from commercial banks, making their role less critical.

On the other hand, credit from development banks directed to the secondary and tertiary sectors interacts positively with infrastructure. When infrastructure spending is higher, development credit to the primary and tertiary sectors becomes more effective in boosting economic outcomes. In the primary sector, better infrastructure (e.g., transportation, water supply) may enhance agricultural productivity or the ability to export agricultural products, making development loans more effective. On the other hand, improved infrastructure (e.g., digital connectivity, roads, or services) could create a more supportive environment for businesses in the service sector. Thus, development bank credit helps those businesses take advantage of the improved infrastructure, amplifying their economic contribution.

Next, we explore the negative impact that Education appeared to have on FDI. We generate new interactions terms between this variable and each form of credit. The new variables are labeled InteractionEd_Priv# and InteractionEd_Dev#. Our results are reported in

Table 6. Introduction of interaction terms in the analysis of FDI.

	(4) logFDI
InteractionEd_Priv1	0.111 6
	(0.196)
InteractionEd_Priv 2	−0.169
	(0.208)
InteractionEd_Priv 3	0.241 ***
	(0.083)
InteractionEd_Dev1	−0.069
	(0.059)
InteractionEd_Dev2	−0.068
	(0.175)
InteractionEd_Dev3	−0.02
	(0.127)
Observations	90
R-squared	0.856

⁶ Robust standard errors are in parentheses *** p < 0.01.

.

The positive interaction term between education and private credit to the tertiary sector suggests that when private credit is available to businesses and services within this sector, the negative impact of education on FDI is mitigated or reversed. This indicates that foreign investors might see more value in regions where education investments are complemented by easy financing for the service industry. Educated workers, combined with robust financing options for businesses in the tertiary sector (like services, retail, or high-value services), make the region more appealing to foreign investors.

While education alone may not directly attract FDI, potentially because the benefits of education investment are more long term, when education is paired with strong support from private banks in the tertiary sector, it likely creates a favorable environment that foreign investors find appealing. This combination might signal a skilled labor force and a dynamic service sector supported by finance, which aligns with the interests of many foreign investors looking for developed markets with growth potential in services.

4. Discussion

The primary aim of this study was to investigate the distinct economic impacts of credit provided by commercial and development banks on the primary, secondary, and tertiary sectors of the Mexican economy. In the specific, we tried to understand how the differing objective functions of these banks—profit maximization for commercial banks and broader social development goals for development banks—translate into contributions to economic growth. Traditional studies often focus exclusively on private credit, missing the different and complementary role development banks play in the economy. By introducing this additional channel, our analysis provides a better understanding of how credit influences economic outcomes, particularly in the Mexican agricultural sector.

Few studies explicitly focus on how these differences impact economic outcomes through a direct comparison across the primary, secondary, and tertiary sectors. We take a new angle connecting these differences back to the banks’ objective functions—commercial banks with profit maximization and shorter time horizons versus development banks with long-term, socially focused goals. This perspective not only deepens the understanding of how credit affects sectoral growth differently but also clarifies why certain sectors may benefit more from one type of bank credit than the other. By making this explicit connection, we emphasize an often implicit intuition in the literature.

It is also essential to recognize that the complementarity between commercial and development credit and its sectoral impacts is inherently specific to the economic and institutional context of individual countries and regions. Different sectors require targeted financial support designed for the unique social, economic, and regulatory conditions in which they operate. For instance, credit strategies that stimulate agricultural growth in Mexico may not produce similar outcomes in other emerging markets with different agricultural frameworks or institutional dynamics. Thus, this study contributes to a particularized understanding of the interaction between credit type and sectoral growth, providing a framework that can be adapted and extended to different national and regional contexts.

To address this objective, we conducted an empirical analysis using a longitudinal dataset that includes credit from both channels across the three productive sectors. This comparative approach allowed us to isolate the effects of each credit source on economic outcomes. We also compared results between a balanced dataset (excluding a critical unbalanced variable) and the full unbalanced dataset, to ensure robustness of our findings and to validate that missing data did not significantly bias our results.

When focusing on Exports, we observed that the agricultural sector and the tertiary one require structurally different financing approaches. While commercial banks do not stimulate the primary sector, probably due to its low returns and long nature of investment projects, development banks appear to significantly boost its exports. Because of their different roles, these institutions can focus on long-term, riskier projects, which can significantly improve the productive potential of their targets. These results are consistent with studies that indicate commercial banks often limit lending to high-return projects (e.g., secondary and tertiary sectors) and bypass agriculture due to its risk and long-term horizon (see Ghosh and Moon 2010; La Porta et al. 2002).

On the other hand, FDI is best supported by private banks. This occurs because their objectives are more aligned. Foreign investors are more attracted by ventures that offer less risky financial returns, avoiding social projects that generally produce unpriced externalities. Being unable to appropriate such externalities leads foreign capital to seek alternatives. Our results align with studies that emphasize how foreign investment prioritizes low-risk, high-return projects that align with private bank objectives (Alfaro et al. 2004). These findings underscore the often-contrasting incentives between FDI inflows and development-oriented lending, particularly in markets with externalities that may otherwise deter private capital (Rodrik 2008).

Finally, we observe that Mexican regional wealth, measured by the GDP per capita, is boosted by development credit but not by a commercial one. Private banks appear to be in equilibrium with the market, catering to the existing demand and not generating a new one. Conversely, the positive influence of development banks hints at the presence of socially useful projects with positive externalities, such as public goods. These projects fail to be addressed by the market, whether because of local institutional conditions or because of more general inefficiencies. For this reason, it is essential to have an established alternative channel with a different purpose.

International comparisons highlight similar dynamics across various emerging economies. In Nigeria, for instance, commercial bank credit has been shown to significantly affect both short- and long-term growth in the manufacturing and agricultural sectors, underscoring the role of targeted credit in these key areas (Eburajolo and Aisien 2019). Similarly, in Jordan, a long-term relationship was observed between bank credit to multiple sectors and economic growth (Ananzeh 2016). In Brazil, Pereira (2018) reveal that public banks have higher direct economic impacts and positive spatial spillovers compared to private banks, and Barboza et al. (2023) show that development bank loans can increase investment, exports, employment, and GDP, particularly for small- and medium-sized enterprises.

Compared to previous studies that predominantly focus on private or commercial credit alone, our findings emphasize the complementary roles of both credit types across economic sectors. Although private credit has a proven but limited role in addressing market-driven growth opportunities, development banks, by financing projects with high externalities, fill critical gaps left by commercial banks. This dual perspective, therefore, contributes new insights into the literature on credit’s impact across sectors and regions (Demirgüç-Kunt and Maksimovic 1998; Beck et al. 2010). By illustrating how these institutions operate in a synergistic yet independent way, we can draw new implications for policymakers seeking balanced, sustainable economic growth through targeted credit strategies.

5. Conclusions

The main contribution of our study was to address the distinct roles of commercial and development banks in impacting economic outcomes in Mexico. Our results reveal that the different objectives of these institutions lead to different impacts on different sectors of the economy. Development banks, by supporting projects that may carry higher risks or longer time horizons, play a critical role in driving growth in areas that are underfinanced by commercial banks, especially in the primary sector.

Our analysis affirms the importance of acknowledging these differences when analyzing the impact of credit on economic growth. While commercial bank credit aligns closely with market demand and tends to reinforce existing economic structures, development bank credit offers an alternative mechanism that can address market gaps and support sectors with high potential but lower immediate returns.

From a policy perspective, we believe that a coordinated approach is necessary to take full advantage of the potential of both types of banks. Governments should consider aligning development credit with strategic national goals, such as infrastructure development or agricultural modernization to maximize long-term gains, and ensuring that development banks receive adequate resources to support projects that generate positive externalities.

Because of its context-specific nature, the impact of credit from development banks cannot be generalized across different regions and sectors without careful understanding of local conditions. Each region has unique needs, risk profiles, and growth potential. The success of interventions depends on the ability to adapt strategies to these specificities, addressing local market failures and leveraging opportunities that may not align with traditional profit-maximizing approaches of commercial banks.

To maximize the effectiveness of development banking, it is critical that policymakers adopt a targeted approach, understanding local institutional and cultural barriers, and identifying sectors and regions where development credit can produce the largest impact. This means conducting detailed analyses of regional economies to pinpoint areas where public investments can generate significant long-term returns, particularly in projects that create positive externalities. By tailoring the operations of development banks to the specific conditions of local markets, policymakers can more effectively create economic and social value.

Future research should explore how these dynamics play out in other emerging markets and under different institutional frameworks, extending the analysis to see whether the results hold in varying economic contexts. Such studies would further illuminate the complex interplay between different types of credit and economic growth, providing valuable insights for both researchers and policymakers.