Are Non-Microenterprise Loans Less Resilient to Domestic Shocks? Lessons from the 2009 Economic Recession
Before the 2009 financial crisis, it was believed that microfinance institutions (MFIs) were highly resilient to domestic macroeconomic shocks, in particular to contractions of gross domestic product (Krauss and Walter, 2006; Gonzalez, 2007; Ahlin, Lin and Maio, forthcoming). Bangladesh and Bolivia have been used as classic examples of how the microfinance sector has survived regional and national macroeconomic crises and recovered faster than the rest of the financial system. Recent studies indicate that the sector may not be as resilient as previously thought, due to an increase in the share of domestic formal-sector lending by MFIs (Wagner, 2010; Di Bella, 2011). However, this hypothesis has not yet been validated empirically.
This paper explores the formalization of microcredit through two different processes. One process is the formalization of MFI portfolios by decreasing the share of loans to microenterprises, presumably, the most informal clients of MFIs. Repayment capacity does not depend on a salary for microentrepreneurs, and when unexpected negative shocks affect them, it may be easier for them to find alternative sources for repayment. In contrast, repayment of non-microenterprise loans may depend on salaries, or very inflexible sources of repayment, like small and medium enterprises. From this perspective, the problem is not formalization per se, but the increase in the correlation of repayment capacity with the general economic situation.
The second formalization process is the expansion of microcredit in economies with high percentages of salaried workers and few microentrepreneurs. The assumption behind this is that, in countries with few entrepreneurs, microenterprise loans will be more correlated with domestic macroeconomic performance, as it will be harder for MFIs to find microentrepreneur clients with the same levels of informality and flexibility as in countries with large shares of informal workers.
Analysis using these specifications for formalization confirms that the microcredit sector was less resilient during the 2009 economic recession in comparison with previous years, and that this lack of resilience was in part due to an increase in the share of non-microenterprise lending (e.g. lending for consumption, education, SMEs and mortgages), combined with an expansion of microcredit in countries with more formal economies (as measured by the percentage of salaried workers).
In addition, in order to measure the level of formalization, we define an MFI formalization index by multiplying the percentage of gross loan portfolio for non-microenterprise purposes (from MIX Market product line data) by the percentage of salaried workers at the country level (from World Development Indicators).
Impact of the 2009 crisis on MFIs
The impact of the recent financial crisis was not homogenous across developing countries, as some countries are less integrated with the global economy than others. In particular, Eastern Europe and Central Asia (ECA) and Latin America and the Caribbean (LAC) were the most affected by the 2009 economic recession, while South Asia was almost untouched at the macro level. Of the 58 developing countries with more than four MFIs reporting to MIX in the period 2008-2009, 17 experienced a contraction in gross domestic product (GDP) in 2009. Of these, 15 were located in ECA and LAC, as shown in Figure 1. By gross national income (GNI) per capita, the effects of the recent crisis are more visible. Deceleration in growth is shown in Figure 1, with ECA and East Asia and the Pacific (EAP) suffering the most. In contrast to all other regions, none of the five South Asian countries in the sample experienced an economic contraction in 2009, or a reduction in their average growth in comparison with 2008.
Figure 1. 2009 Economic Recession in Developing Countries by GDP and GNI
Consistent with the impact of the 2009 economic crisis, the biggest deteriorations in portfolio quality were observed in Eastern Europe and Central Asia (ECA), East Asia and Pacific (EAP), and Latin American and the Caribbean (LAC). In addition, it is worth highlighting that MFIs in Sub-Saharan Africa and South Asia only experienced a modest deterioration in portfolio quality, as shown in Figure 2.
Figure 2. Median Risk by Region
Portfolio quality was also less resilient to changes in GDP in 2009. In particular, observed levels of portfolio quality in 2009 were worse than in previous years, for the same levels of GDP growth. In addition, the average PAR30 before and after 2009 was very similar for MFIs operating in countries with positive growth. However, there was an increase in the average level of risk in 2009 for MFIs operating in countries with negative growth rates, as shown in Figure 3.
Figure 3. PAR30: Before and After 2009
Figure 4. Simulated Risk
In order to measure the level of formalization, or integration of MFIs with the formal economy, a formalization index is defined by multiplying the percentage of gross loan portfolio for non-microenterprise purposes (from MIX Market product line data) by the percentage of salaried workers at the country level (from World Development Indicators).
Formalization Index = 100-(% of microenterprise loans * % of country self-employed workers/100)
% microenterprise loans = microenterprise loans in USD / total loan portfolio in USD
% of country self-employed workers = 100-“Waged and salaried workers, total (% of total employed),” most recent figure from World Development Indicators.
Non-microenterprise loans include consumption loans, education loans, housing loans, and small and medium enterprise (SME) loans. By this definition, an MFI with 50 percent of their portfolio in non-microenterprises operating in a country with 50 percent salary workers will have an index level of 25, while an MFI with the same portfolio operating in a country with 100% salaried workers will have an index level of 50.
The formalization index is based on the following assumptions:
- Repayment of non-microenterprise loans (consumption, education, mortgages, SME, etc.) is more dependent on salaries than repayment of microenterprise loans (which are, presumably, repaid through microenterprise/household income).
- During recessions, the repayment capacity of salaried workers is affected more than that of informal workers (typically associated with microenterprise loans), as salaried workers lose income when they become unemployed and formal job searches will be more difficult when the economy is growing very slowly or contracting.
- Informal workers are more accustomed to dealing with economic hardship, and may have multiple sources of repayment when crisis hit or may be more creative in finding new opportunities when formal economies go in recession. This does not imply that informal workers are immune to shocks, but they are more resilient than salaried workers (Gonzalez, 2008).
For analysis of the impact of the 2009 crisis, the 2008 formalization index was multiplied by the level of contraction in GDP for 2009 to create a shock index:
Shock Index 2009 = Formalization Index 2008 * % GDP contraction in 2009
For example, an MFI with a formalization index of 50 in a country where GDP dropped 5% has a shock index of 2.5, while a similar MFI in another country where GDP dropped 10% has a shock index of 5. All MFIs in countries with positive GDP growth in 2009 have a shock index of 0, regardless of their level of formalization.
The analysis of the shock index suggests that, on average, every 1 point difference in the shock index is associated with a 0.31 point difference in total risk (portfolio at risk > 30 days + write-off ratio). The higher the shock, the worse the portfolio quality, as shown in Figure 5. In other words, the impact of the shock is proportional to the level of formalization of the MFI and the contraction in GDP. These results can be clearly visualized through the performance of MFIs in Mexico and Nicaragua in 2009, as shown in the following figure.
Figure 5. Formalization Index and Portfolio Quality for Mexico and Nicaragua
For comparison, the average relationship between total risk and GDP growth, both excluding 2009 data and including 2009 data, are presented in Figure 6. By this simulation, the sector was more resilient before 2009 as can be seen by the flat slope of the curve. However, after including 2009 data, the sector becomes less resilient (the slope increases), meaning that higher contractions in GDP are associated with higher levels of risk.
To simulate the impact of formalization, two hypothetical scenarios are presented: for MFIs with no formalization and for MFIs with 100 percent formalization. Still, the level of formalization does not explain the whole increase in resilience experienced by the sector in 2009 because the 0 percent formalization curve has still a steeper slope than the average curve excluding 2009 data. However, note how MFIs with very high levels of formalization will be even less resilient (Figure 6).
Figure 6. Simulated Results by Level of Formalization
The most important new finding of this analysis is that the higher the level of formalization, or integration of microfinance loan portfolios with the domestic economy, the worse the decline in portfolio quality during economic recessions. However, during good times (positive growth in GDP), there is no difference in performance associated with the level of formalization of MFIs.
In addition, this analysis finds that risk levels improve with moderate levels of inflation, contrary to conventional wisdom. This could indicate that borrowers are able to transfer the inflationary increase in prices to their clients, and that they are thus able to accommodate any increase in interest rates from MFIs due to inflation (most likely coming from an increase in cost of funds and salaries). The econometric results from this analysis also confirm previous findings that market saturation is one of the most important drivers of portfolio quality, and that the type of growth (extensive versus intensive) is more relevant for the discussion of over-indebtedness than the absolute levels of growth.
The risk profile of MFIs depends on the risk profile of their clients, and the relationship between their repayment capacity and the domestic and global economy. Microcredit originally targeted primarily informal microentrepreneurs in developing countries, but recently has expanded into salaried workers in more developed countries. The findings from this paper suggests that the definition of microcredit is more than pure semantics, as an in-depth understanding of MFIs’ credit products and clients is necessary in order to understand their level of risk, and their similarities to and differences from traditional credit providers.
Many MFIs target informal microentrepreneurs mainly with the goal of increasing depth of outreach to vulnerable and credit-constrained populations. However, the results from this paper suggest that by doing so, these MFIs may also be more resilient to economic recessions.
In addition, the findings from this paper reinforce the idea from previous research, that appropriate lending technologies (loan sizes, screening mechanism, interest rates, terms, credit officer training and incentives, etc) are the most important area to focus on in order to prevent overindebtedness.
At this preliminary stage, more analysis is necessary to understand the full trade-offs and synergies between the formalization of the microfinance industry, including the development of better indicators for the informality and flexibility of borrowers and their risk management strategies when facing recessions. The evidence supports the argument that microfinance has been more resilient than the rest of the financial system because microfinance clients are “different” than the clients of formal-sector financial institutions. This suggests that the factors that differentiate microfinance clients are their flexibility to react to financial crises due to the informal and short-term nature of their economic activities, their diversified economic activities, the flexible supply of household labor, and the incentives to maintain access to credit from current MFIs when access to credit from alternative sources is scarce or expensive.
This result is further relevant for MFIs that provide traditional commercial lending products, such as consumer finance to salaried workers or cash-flow based small business lending, where the clients may have less diversified activities and fewer alternatives to deal with economic crises. MFIs that rely more on consumer or small-business lending appear to be more exposed to financial crises than MFIs that based their lending on the traditional microfinance lending methodologies like village banking, solidarity groups or individual microenterprise loans. This result has important implications for those trying to expand credit for SMEs through microfinance providers.
 Considering the effect on portfolio quality of other variables including: MFI’s size and age, inflation, and market saturation.
 The most recent estimate for salary workers was used under the assumption that this indicator is relatively stable overtime.
 Note that this is not equal to: % of non-microenterprise loans * % of country salary workers/100
 There is anecdotal evidence suggesting this hypothesis, but the proposed formalization index is the first formal attempt to try to measure this effect. For instance, Syed Moshin Ahmed of the Pakistan Microfinance Network commented that “anecdotal evidence suggests that MF clients who are economically active have been either positive impacted in the first generation effect of inflation or remained neutral. The only group that has been affected is salaried urban class and people who have taken loan for consumption.” Martin Holtmann from International Finance Corporation (IFC) makes a similar point regarding institutions focus on consumer lending or SMEs (Littlefield, 2008).
 The assumption is that formalization only matters when there is economic recession (GDP is contracting), but it is irrelevant when the whole economy is growing.