assessing price fairness in microfinancerolled-over (the “payday” lending model) is not at all...

Assessing Price Fairness in Microfinance

A framing note to inform the Evolution of the

Client Protection Standards

Lead Author: Daniel Rozas

January 2016

January 2016 | Daniel Rozas | Page

1


1. Background ...................................................................................................................... 2

2. Purpose of this Note ......................................................................................................... 2

3. Introduction ..................................................................................................................... 2

................................................................................................... 4 a. Measuring by induction ................................... 5 b. Changing the Smart Campaign’s responsible pricing framework

1. Preliminary steps ..................................................................................................................................... 5 2. Components not included in pricing assessment .................................................................................... 6 3. Assessing appropriate profit ................................................................................................................... 6 4. OpEx Assessment .................................................................................................................................... 7 5. Concluding the fair pricing assessment ................................................................................................... 8 6. Ongoing maintenance ............................................................................................................................. 9

4. Appendix A – The Client Protection Principles and Standards ...........................................10

5. Appendix B: Developing OpEx Model ...............................................................................11

6. Appendix C: Developing the fair profit assessment ..........................................................14


2


1. Background The Client Protection Standards were created to operationalize where the microfinance industry sets the bar in terms of the minimum behaviors clients should expect from institutions with which they do business. Building off of the seven Client Protection Principles (CPPs), the Client Protection Standards specify what „doing no harm‟ must entail in practice. The current 30 standards (version 1.0) represent the output of several years of industry collaboration and input managed by the Smart Campaign (See Annex A for the complete list of the standards). For the standards released in January 2013 as part of the Client Protection Certification program (version 1.0), the Campaign worked with a Task Force of over 30 experts representing various stakeholders to develop and vet the recommendations. The standards are truly a public good for and by the industry. Standards which reflect social norms and expectations of an evolving industry must be dynamic. In order to incorporate an ever-growing sector and its diversity of products, services and related client-protection risks, the Smart Campaign has begun to work to evolve and improve its standards. One of the key areas under review for the standards 2.0 was Client Protection Principle No.4: Responsible Pricing.

2. Purpose of this Note In the context of the Evolution of Standards of the Smart Campaign and its Responsible Pricing work stream, the Smart Campaign commissioned special research on alternative analysis methods for credit product pricing. This paper aims to provide an alternative to the market based comparative approach used for evaluating pricing for credit products under standards 1.0. The Smart Campaign certification process currently (version 1.0) applies a market test to determine responsible prices: as long as an institution‟s prices are in line with its market peers, its prices are considered responsible. This test was selected rather than other measures because can be applied in a straightforward way. Nevertheless, the Secretariat, Certification Committee and the Smart Campaign Steering Committee have identified challenges linked to a market-based approach in the analysis of responsible prices for the purpose of financial institution certification (SC standard 4.1.1.) including:

Markets are at different levels of maturity and not always functioning well (e.g., high priced markets, monopolistic or oligopolistic environment).

The presence of “market leaders” in certain markets raises the question of whether leading institutions should be held to higher standards in terms of pricing, given their influence on the market as a whole.

There is often difficulty achieving relevance and consistency in the creation of peer groups for pricing comparisons.

The scarcity and poor quality of pricing data makes pricing comparisons challenging.

The purpose of this project is to examine these issues and make recommendations on how to address them in the standard-setting and certification process.

3. Introduction In a business with social objectives, the concept of fair pricing is an integral concern. You can‟t claim to be fulfilling social objectives if you overcharge your customers. Now, the concept of “overcharging” isn‟t actually all that simple. When goods are sold, one could look at the cost of those goods, factor in operating


3

expenses, provide some room for a reasonable profit margin, and – voila! – arrive at a fair price. At least there‟s no confusion as to what constitutes a price. But in microfinance, such confusion is very much present. As Chuck Waterfield and the Microfinance Transparency project have educated us over the years, price depends a great deal on how it‟s presented. A standard formulaic presentation – an APR – is a good metric for comparing the pricing of different products. The trouble is, credit pricing is inescapably tied to the product features. Is the loan small or short-term? Well, it will come with high operating costs, yielding a fair price – one that covers costs and provides a reasonable profit margin – that can easily run into the 100s of percent. But it‟s not just the product features that affect price fairness. So does the business model. Making expensive short-term loans can be reasonable and fair. Making expensive short-term loans that are regularly rolled-over (the “payday” lending model) is not at all fair, since the high short-term loan price is being applied to a loan that has effectively a much longer term. Similarly, allowing flexible repayment (for example, an optional grace period or a payment extension) lowers the effective cost of the loan even as its features remain the same. Lastly, even if one can set parameters for fair pricing of a given product, it is largely impossible to implement that fair pricing in practice. Consider a 25% APR loan that comes with a range of features:

Weekly payment for 30% APR loan

$ weeks

100 300 500 1000

4 251 76 127 254

12 9 26 43 86

36 3 9 15 31

52 2 7 11 22 Though borrowers should and do use APR to compare loans, most will inevitably look at payments to assess how much they can afford to borrow. And the table here is already somewhat counter-intuitive – borrowing $100 for 4 weeks costs almost exactly the same as borrowing $300 for 12 weeks. But ok, perhaps one can do the mental calculations and see that this is in fact an equally-priced loan (repayments include principal, which can be very easily subdivided into periods). But what if the loan were presented with a fair price, i.e. reflecting the higher costs of making short-term / small-sized loans? Here‟s a stylized example:

APR for fair-priced loan

$ weeks

100 300 500 1000

4 107% 89% 81% 70%

12 71% 59% 54% 47%

36 43% 36% 32% 28%

52 36% 30% 27% 23% Source: MIX and MF Transparency; term factors based on work by Emmanuelle Javoy

1 Weekly payment is $25.36, but interest is not visible due to rounding.


4

If one can imagine a perfectly reasoned conversation over a payments table, I struggle to imagine a reasonable conversation over this APR chart. What do you mean my $100, 4-week loan costs four times more than the $500, 52-week loan? You‟re charging me less when I borrow more? And yet, this chart is a pretty good approximation of what often takes place in practice, due to pricing schemes that tack on flat fees and other costs that effectively increase the price of small/short-term loans. Even if in practice, the price is fair – it reasonably accurately reflects the loan costs – such practices are also non-transparent, preventing clients from being able to compare offers from different lenders. There is thus a conflict between fair pricing in its absolute form, where each product‟s price reflects its cost, and the demand for pricing transparency. And indeed, a reasonable fair pricing standard should not require that every product reflect its respective costs. Cross-subsidization within the portfolio is both appropriate and widely practiced, whether knowingly or not. And that implies that measuring price fairness may likewise be more suitable at the portfolio level than for each product. This also has the added advantage of avoiding the complexity of gathering a large set of data to evaluate pricing for multiple product permutations.

a. Measuring by induction Indeed, to assess price fairness, one need not look at price at all. Consider the following accounting equivalence:

For financial institutions that derive their main revenue from loans, the two sides would be largely the same. If we posit that an institution whose operations are reasonably efficient, credit losses are limited, financing costs are market-driven, and profits are fair, then its pricing (as reflected by its income) will likewise be fair. As it turns, assessing fairness on the expense side is both easier and more accurate. Consider each of the expense-side categories: Profit is one of the main causes for consternation for social investors. Among the key charges of microfinance critics is that high prices are fueling exorbitant profits off the backs of the poor. There is no better way to counter this than by demonstrating that these profits are in fact modest. Financial expense: almost no MFI has much control over this. Cost of debt is set by the market. Meanwhile, while deposit-taking institutions can substantially lower their financial expense by raising lower-cost deposits, this comes at the price of higher operating costs to support the new deposit-taking infrastructure and staff. The savings from one are usually fully offset by the expenses from the other.

2

Credit losses are essentially what separates microfinance from subprime lending. The role of the former is to identify capable clients and lend amounts that they can repay. The subprime model also targets the poor, they make much more limited (if any) efforts to evaluate the borrowers, and instead rely on higher price

2 Rozas D., Exploring the Business Models Behind Microsavings, Financial Access Initiative, Nov 2015

Operating expenses

Credit losses

Profit

Financial expenses

Portfolio Yield

Expenses Income


5

markups to make up for the high resulting default rates. Of course even responsible lenders may experience higher losses due to external factors (economic stress, political unrest, natural disaster, etc.), but persistently high credit losses are rare for well-run MFIs. Operating expense is nearly always the largest component of an MFIs‟ costs – an unavoidable byproduct of focusing on small loans. Indeed, for many institutions, operating expense can easily consume half or more of their total expense. And yet, unlike with the other three factors, setting standards or benchmarking operating expenses is exceptionally difficult – they are inevitably tied up with the specific institutional model – where they operate, whom they serve and with what sort of loans, the scope of their deposit operations, if any, and many other factors. Yet any effort to assess fair pricing by evaluating the institution‟s expenses requires developing a suitable benchmark for operating expense. Indeed, doing so is necessary to be able to credibly answer the question of whether or not an MFI is efficient.

b. Changing the Smart Campaign’s responsible pricing framework Until now, assessing responsible pricing has entailed comparing an MFI‟s offerings with those of its competitors. That comes with a number of challenges – developing consistent peer groups is difficult, and the methodology cannot distinguish cases situations where the peer group was composed of inefficient or excessively profitable institutions, making the peer group comparison problematic. The proposed framework relies on a new approach: assessment by induction. Rather than assessing price directly, it aims to assess the expense components. Specifically, it focuses on two: profit and operating expense, given that financial expense is nearly always outside the MFIs‟ control, and credit losses are already assessed as part of avoiding overindebtedness. Below is an explanation of the framework and underlying methodology.

1. Preliminary steps

1.1. Required metrics

Indicator MIX Market definition

ROA Return on assets

OpEx Operating expense over assets

FinEx Financial expense over assets

Credit losses Provision for loan impairment

Yield Yield on Gross Portfolio (nominal), recalculated as ratio over average assets

Assets Assets

Voluntary deposits Total deposits – compulsory deposits

Deposits to loans Voluntary deposits to loans ratio

Average loan balance Average outstanding balance

GNI per capita GNI per capita, Atlas method (current US$) (World Bank)

Rural population density Rural population / total land area (km2)

1.2. Adjust for compulsory deposits Compulsory deposits can substantially distort pricing and other financial metrics. Consider a hypothetical case – if to lend $100, and MFI requires $50 as compulsory deposit, it is in effect lending only $50. Yet its loan portfolio and balance sheet will reflect the larger value. An APR pricing methodology nets out compulsory deposits from the loan amount. For the same reason, financial metrics must be similarly adjusted.


6

The calculations are as follows:

Adjusted loan portfolio = loan portfolio – compulsory deposits

Adjusted assets = assets – compulsory deposits

Recalculate Yield, FinEx, ROA, Credit Loss Ratio, OpEx, Avg Loan Balance, Deposits to Loans ratio,

and using the adjusted values for loan portfolio and assets. These recalculated values should be

used for all subsequent calculations.

For multi-year averages (for example, 3yr avg ROA), first recalculate the values for each year, then

calculate the multi-year average.

1.3. Test data validity In some cases, MFIs may have significant income from sources other than its loan portfolio. In other cases, financial operations (for example, portfolio securitizations) may alter the balance sheet, converting what otherwise would be loan income into income from asset transfers. These and other situations can break down the validity of the inductive test, since either the expenses or income comparison no longer reflect lending operations. To insure that the test is assessing the correct metrics, it‟s important to compare that the two sides of the accounting equivalence – the four components on the expense side must reasonably equal the portfolio yield on the income side:

ROA + OpEx + FinEx + Credit Loss ≈ Yield If the two sides are within a 5% margin of each other, then one can proceed to subsequent tests. If they are outside this threshold, it‟s important to evaluate the source of the discrepancy and make the necessary adjustments, as appropriate. For example, in the case of an institution that sells or securitizes large portions of its loan portfolio, either off-balance sheet assets need to be included (affecting both portfolio and assets, along with other metrics) or income from the portfolio sale needs to be excluded. Similarly, if an institution has significant non-loan income (for example, from a bond portfolio, as is common for many banks), then this income should be excluded from the yield calculation. The guiding principle should be that loan-related income and cost of associated activities are included on both sides of the equation.

2. Components not included in pricing assessment Both FinEx and Credit losses are excluded from this portion of the assessment. In the case of FinEx, the vast majority of financial institutions have little to no influence on the cost of debt, making assessment largely irrelevant. Assessment of pricing of savings products, over which FIs have direct control, may be added at some point in the future, if it‟s deemed relevant. Credit losses are already fully addressed under the rubric of Prevention of Overindebtedness, and there is no reason to duplicate it here.

3. Assessing appropriate profit Assessing appropriate profit is a two-step process, first a quantitative benchmarking, and in the event profit is found to exceed the maximum threshold, a qualitative assessment of the use of excess profits.

3.1. Profit benchmarking A simple distribution of profit levels shows that half of MFIs have three-year average ROA levels below 3%, while only a quarter of MFIs exceed 5.7% (see Appendix C for detail). These delineations are the basis of the proposed guidelines.


7

To benchmark profits, calculate a 3-year average ROA (use ROA adjusted for compulsory deposits for each of the years), then apply the following test:

ROA Assessment action

<1% Assess institutional sustainability

1-3% Normal range

3-6% Elevated range

>6% High range

If ROA is in the normal range, no further profit assessment is required. Move to OpEx assessment. In the event ROA falls in either the elevated or high range (i.e. above 3%), proceed to next step.

3.2. Evaluation of elevated / high profits A high ROA does not necessarily constitute excessive profit. There may be many reasons why profits would fall in this range that are perfectly consistent with appropriate pricing, including (but not limited to) the following:

Profits diverted to external entity (ex: affiliate NGO) that provides services that are important for

clients (ex: non-financial services)

Profits shared with clients

High inflationary environment

Grow client base with limited access to outside equity

Build up equity and strengthen FI

Early stage institutions

Subject to regulation that increases earnings requirements (e.g. high reserve requirements, etc.)

Profitability inflated by donations, subsidies or other temporary or short-term events

High country risk necessitates an additional cushion to protect against adverse events

These and other reasons need to be evaluated on a case-by-case basis, with final decision made by the certification committee. The primary goal here is to assess whether the high profits are either 1) a result of the context in which the MFI is operating (inflation, high country risk, etc.), or 2) they are used in ways that benefit the clients in some way. If high profits mainly benefit shareholders above the levels justified by the operating context (e.g. after accounting for inflation, country risk, etc.), then the MFI‟s profitability should be seen as inconsistent with appropriate pricing, and the MFI should not be certified with one exception: in the event profits are in the elevated (but not high) range and are assessed to be primarily benefiting shareholders beyond levels reasonable given inflation/country risk, but the MFIs OpEx is at least 3% below its expected level (see below), then the MFI‟s profits can be said to be offset by higher efficiency, and it remains eligible for certification.

4. OpEx Assessment As with profit assessment, OpEx is a two-step process, consisting of a quantitative benchmark, and if necessary, followed by a qualitative assessment.


8

4.1. OpEx benchmarking Assessing operating expense is done by following a linear regression model that estimates the expected OpEx given key factors of the MFI and its operating context (for full regression results, see Appendix B). The model is as follows (subject to change, based on testing on an updated MIX and raters dataset): Expected OpEx = 0.7928661 + [GNI per capita] * 0.0000144 + ln([Average outstanding balance]) * -0.053171 + [Voluntary deposits to loans] * 0.0156794 + ln([Assets]) * -0.0185402 + [Rural Population Density] * -0.0002098 Maximum OpEx Level = Expected OpEx + .065,

where .065 is the tolerance level at the 85th percentile of MFIs, with only 15% of MFIs exceeding the

Maximum OpEx Level. If the actual OpEx is < Maximum OpEx Level, then the institution‟s OpEx is seen as consistent with appropriate pricing. No further analysis is required. If actual OpEx is greater than this level, then proceed to qualitative assessment.

4.2. Assess appropriateness of high OpEx The OpEx model, while robust, is not perfect. The elements it captures are the major drivers of variances operating expense: loan size, GNI per capita, level of voluntary deposits, institutional scale, and the rural population density of the country in which it operates. However, there may be reasons for having a higher OpEx that are specific to the MFI‟s operations and that could never be captured by any model. These include (but are not limited to) the following:

MFI operating in a low-security environment, requiring significant spending on non-standard security

costs

MFI is serving particularly difficult-to-reach clients, in a way that is not representative of the national

rural population density. The model is by necessity limited to national density averages, even as

these may vary greatly within different subnational regions.

MFI serving an exceptionally under-privileged population, requiring add-on services (youth, disabled,

etc.)

MFI is operating non-financial programs that are useful to clients

The final assessment should be made based on the level of excess OpEx and the scale of the operations (e.g. security, difficult outreach, etc.) not captured by the OpEx model.

5. Concluding the fair pricing assessment The final assessment is based on meeting the criteria for both appropriate profit and appropriate OpEx. However, for institutions that exceed one of these two metrics, without sufficient justification, the certifier can take into consideration the other component to judge whether it may to sufficient degree offset the excess in the other (i.e., in the example mentioned above, an institution with elevated profits, but a lower than expected OpEx that sufficiently offsets that profit excess, can still be eligible for certification). The same approach can work when the two metrics are reversed, i.e. higher OpEx, but lower ROA can still offset each other. That said, the offsets are limited, and should in principle not be eligible to offset more than 3% of excess. An institution with (unjustified) profits above 6% should not be eligible for certification, regardless of its OpEx level. Similarly, an institution with an unjustified OpEx level that is more than 3% above the Maximum OpEx Level would not be eligible for certification, even if its profits are zero.


9

6. Ongoing maintenance As mentioned earlier, both OpEx and ROA assessment models are based on data that is not adjusted for compulsory deposits. It‟s important that both sets of calculations be revised to account for this, and the models and bands be updated, if necessary. From time to time, the models and the bands will need to be revisited, to insure that they continue to be based on data that is representative of the sector. It would be sufficient for these revisions to be done bi- or tri-annually. The OpEx model itself could also be revisited at a similar interval, to evaluate whether improvements can be made, including change in predictor variables if needed.


10

4. Appendix A: The Client Protection Principles and Standards

1 - - Client Protection Principle 1: Appropriate Product Design and Delivery Channels 1 1 - The FI designs products that are appropriate to client needs and do no harm

1 2 - The FI seeks client feedback for product design and delivery

1 3 - The FI does not use aggressive sales techniques

2 - - Client Protection Principle 2: Prevention of Over-indebtedness 2 1 - The FI conducts appropriate client repayment capacity analysis before disbursing a loan

2 2 - The FI incentivizes quality loans

2 3 - The FI uses credit bureau and competitor data, as feasible in local context

2 4 - The FI Management and Board is aware of and concerned about the risk of over-indebtedness

2 5 - The FI's internal audit department monitors that policies to prevent over-indebtedness are applied

2 6 - The FI avoids dangerous commercial practices (i.e., avoids combining loan products to meet the same need, or restricting the loan use; sets prudent limits to allow for the renewal of a loan in case of early repayment; sets guidelines for appropriate rescheduling policies)

3 - - Client Protection Principle 3: Transparency 3 1 - The FI fully discloses cost and non-cost information

3 2 - The FI communicates proactively with clients in a way that clients can easily understand

3 3 - The FI uses a variety of disclosure mechanisms

3 4 - The FI leaves adequate time for client review and discloses at multiple times

3 5 - The FI provides accurate and timely account information

4 - - Client Protection Principle 4: Responsible Pricing 4 1 - The FI offers market-based, non-discriminatory pricing

4 2 - The FI’s efficiency is in line with its peers

4 3 - The FI does not charge excessive fees

5 - - Client Protection Principle 5: Fair and Respectful Treatment of Clients 5 1 - The FI culture raises awareness and concern about fair and responsible treatment of clients

5 2 - The FI has defined in specific detail what it considers to be appropriate debt collection practices

5 3 - The FI's HR policies (recruitment, training) are aligned around fair and responsible treatment of clients

5 4 - The FI implements policies to promote ethics and prevent fraud

5 5 - In selection and treatment of clients, the FI does not discriminate inappropriately against certain categories of clients

5 6 - In-house and 3rd party collections staff are expected to follow the same practices as the FI staff

5 7 - The FI informs clients of their rights

6 - - Client Protection Principle 6: Privacy of Client Data 6 1 - The FI has a privacy policy and appropriate technology systems

6 2 - The FI informs clients about when and how their data is shared and gets their consent

7 - - Client Protection Principle 7: Mechanisms for Complaints Resolution 7 1 - The FI's clients are aware of how to submit complaints

7 2 - The FI's staff is trained to handle complaints

7 3 - The FI's complaints resolution system is active and effective

7 4 - The FI uses client feedback to improve practices and products


11

5. Appendix B: Developing OpEx Model The model is built on a multi-factor regression, using MIX Market data:

Panel of MFIs reporting during 2006-2013 (inclusive)

284 MFIs, 45 countries, 2,124 observations

MFIs selected according to data quality (required fields populated), and have at least 6 out of the 8

years reported

When applied to the dataset, the model is able to explain 40% of variation in OpEx. More importantly, taking an average of 2011-2013 reported values, 85% of reported OpEx values do not exceed the predicted OpEx value by more than .063 points. Among 38 certified MFIs, only 2 exceed this level. However, note that these figures have not been adjusted for compulsory deposits. Those adjustments may increase the number of observed outliers.

Difference between Actual and Predicted OpEx values (average of 2011-13 figures)


12

The current model is conservative – most reported OpEx values are lower than the predicted value, though MFIs that have very high OpEx, even if fewer in number, have a substantially greater deviation from the model. This is apparent in the below chart:

The actual regression output is as follows:

A number of other indicators were also evaluated as part of this analysis, with some having substantial value in predicting OpEx (for example, if an institution is a cooperative). However, these have not been included in the model, since they cannot reasonably be judged as criteria for assessing fair pricing. Thus, while it is reasonable to expect large institutions to have economies of scale and thus have more efficient operations (hence the inclusion of the Assets indicator), it is not reasonable to expect institutions to have higher efficiency because they happen to be cooperatives. And if they do have lower than expected OpEx, then certainly their pricing would be judged efficient, when compared to other institutions. Likewise, an institution with a low number of borrowers per staff member can be expected to be less efficient, but should not then be held to a lower bar because of that choice. For this reasons the variable is excluded from the OpEx model.

20%

35%

62%

87% 93%

-0.3

7

-0.3

3

-0.3

1

-0.2

8

-0.2

6

-0.2

4

-0.2

2

-0.2

-0.1

8

-0.1

6

-0.1

4

-0.1

2

-0.1

-0.0

8

-0.0

6

-0.0

4

-0.0

2 0

0.0

2

0.0

4

0.0

6

0.0

8

0.1

0.1

2

0.1

4

0.1

6

0.1

8

0.2

0.2

2

0.2

4

0.2

6

0.2

8

0.3

0.3

3

0.3

6

0.3

9

0.4

6

0.6

1.6

4

Distribution of observations for difference between actual & predicted OpEx

_cons .7928661 .0212149 37.37 0.000 .7512617 .8344704

ruralpopulationdensity -.0002098 .0000177 -11.82 0.000 -.0002446 -.000175

logAssets -.0185402 .0012669 -14.63 0.000 -.0210247 -.0160557

depositstoloans .0156794 .0053141 2.95 0.003 .005258 .0261008

gnipercapita .0000144 8.63e-07 16.70 0.000 .0000127 .0000161

logALB -.053171 .0022791 -23.33 0.000 -.0576405 -.0487015

operatingexpenseassets Coef. Std. Err. t P>|t| [95% Conf. Interval]

Total 27.7236238 2103 .013182893 Root MSE = .0886

Adj R-squared = 0.4046

Residual 16.4678897 2098 .007849328 R-squared = 0.4060

Model 11.2557341 5 2.25114682 Prob > F = 0.0000

F( 5, 2098) = 286.79

Source SS df MS Number of obs = 2104


13

A list of these indicators is described below:

Variable* Explanation Included in Model

Effect on price

LN (average outstanding balance)

Core measure of “price curve” as described by C. Waterfield. Here used in natural log form, to closely approximate the curve

Y ↓

GNI per capita Proxy for operating cost. Higher-income countries mean higher expenses for same-sized loans.

Y ↑

Voluntary deposits-to-loans Deposit-taking institutions have significant deposit-related expenses. These are included in the model to avoid penalizing them unfairly.

Y ↑

LN (assets)

Large institutions can (and should) realize economies of scale. This metric allows the model to account for institution size. Value expressed in log form to reflect diminishing returns.

Y ↓

Rural population density

All else equal, more densely populated countries are easier to serve than sparsely-populated ones. Calculated as (total rural population / total land area).

Y ↑

Borrowers per staff member Standard efficiency metric N ↓

Avg Salary / GNI per capita Higher salary levels relative to national averages mean higher costs.

N ↑

Portfolio / Assets MFIs with high asset utilization face lower operating costs

N ↓

Write-off ratio High writeoffs often mean high costs from collections

N ↑

Degree of competition (HHI) No significant impact on OpEx. N n/a

Market share

Moderately significant and positive, implies

possible monopoly effect (high market share

implies less competition to push efficiency

improvement)

N ↑

Credit bureau quality

(Microscope)

Small downward effect on OpEx, but data not

available for all countries N ↓

MFI is a Cooperative

Large & negative effect on OpEx (-11.8%).

Coops appear more efficient than other MFI

types. Not included in model, since this is simply

an observation, not a useful predictive factor.

N ↓

MFI is a for-profit Small, but significant effect on OpEx (2.4%). N ↑

* Unless otherwise noted, variables are as defined in the MIX Glossary.


14

6. Appendix C: Developing the fair profit assessment The fair pricing model is largely based on prior work developed by Chuck Waterfield.

3 However, the main

change is to express the profit level over assets (ROA) rather than equity (ROE), since the former is more directly comparable to income figures. By contrast equity returns can be heavily affected by changes in the institution‟s financial leverage – a factor that is wholly unrelated to loan pricing. A simple distribution of profit levels shows that half of MFIs have three-year average ROA levels below 3%, while only a quarter of MFIs exceed 5.7%. These delineations are the basis of the proposed guidelines. Unlike with OpEx distribution, profit distribution of 38 certified MFIs largely follow the broader group, with exactly half (19) showing ROA below 3%, and another 10 below 6%, and the final group of 9 with an ROA exceeding 6%, in two cases with a large margin.

Distribution of ROA (average of 2011-13 figures)

3 Waterfield, C., Growth, Profit & Compensation in Microfinance: How much is too much?,

MFTransparency.org, Sep 2012

16%

34%

49%

61% 69%

76% 80%

85% 89%

93% 96% 97% 98% 99%

0% 2% 4% 6% 8% 10% 12% 14% 16% 18% 20% 22% 25% 28% 32% 37% 52%

Distribution of ROA (positive only)

assessing price fairness in microfinancerolled-over (the “payday” lending model) is not at all...

Documents