Where Angels Prey

Where Angels Prey is a novel by Ramesh S Arunachalam. Please refer to www.whereangelsprey.com for more information

Friday, October 22, 2010

PAR and its Limitations


The loan portfolio outstanding is the largest income-generating asset for a micro-finance institution[1] (MFI) and the quality of this asset is therefore very critical to its survival. Thus, it is imperative that judgments about the quality of an MFI’s portfolio are made in a (reliable) manner, so as to accurately portray the level of default or credit risk. While the extant literature suggests that one of the best measures of asset quality is Portfolio at Risk (called ‘PAR’), practical situations reveal that this measure also suffers several limitations, which, if unaddressed, could result in an inaccurate portrayal of the default/credit risk. In this note, an attempt is made to highlight events and methodological aspects that distort the ‘PAR’ measure and propose appropriate solutions so that PAR can be used to accurately identify and control the credit risk.


Portfolio at Risk Defined

Portfolio at Risk (‘PAR’) is a percentage (%) measure[2] and it denotes the “proportion of an MFI’s total gross outstanding loan portfolio that is at default/credit risk. The formula for ‘PAR’ is…

Sum of Unpaid Principal Balance of All Loans with Payments Past Due (1 to 365 Days and more)
Total Gross Outstanding Loan Portfolio (Sum of Principal Outstanding of All Loans)

Three aspects about PAR deserve mention: 1) It is a stock measure and reflects the default/credit risk as at a given date; 2) It attempts to measure the default/credit risk by extrapolating past client behaviour into the future; and 3) Specifically, its estimation of the default/credit risk is based on one critical question - as on a given date, if each and every delinquent borrower, were to completely default, then how much (money) is the MFI likely to loose? PAR therefore provides a very pessimistic[3] estimate of the default/credit risk as it assumes that today’s (delinquent) behaviour by clients would be prevalent in the future[4] as well.

Aged Portfolio at Risk

‘Aged PAR’ is a variant measure of ‘PAR’ and is often calculated. In micro-finance, examples of ‘Aged PAR’ categories include: PAR > 0 days (which is the same as normal PAR given earlier), PAR > 30 days, PAR> 60 Days and so on. The formula for PAR > 30 Days is:

Sum of Unpaid Principal Balance of All Loans with Payments Past Due (31 to 365 Days and more)
Total Gross Outstanding Loan Portfolio (Sum of Principal Outstanding of All Loans)

This leads to the following generalizations for ‘Aged PAR’:

PAR > 0 Days
= PAR 1-30 Days + PAR 31-60 days…+ PAR > 365 Days
PAR > 30 Days
= PAR 31-60 Days + PAR 61-90 days…+ PAR > 365 Days
PAR > 60 Days
= PAR 61-90 Days + PAR 91-120 days…+ PAR > 365 Days
PAR > 90 Days
= PAR 91-120 Days + PAR 121-180days…
+ PAR > 365 Days
PAR > 180 Days
= PAR 181-365 Days + PAR > 365 Days
PAR > 365 Days
= PAR > 365 Days

Interpreting PAR

PAR interpretations generally concern three basic aspects: 1) Absolute values of PAR; 2) Trends in absolute values of PAR over time; and 3) Aged absolute values of PAR and associated trends.

Regarding absolute values of PAR, smaller values imply a lower level of default/credit risk. Thus, PAR of 3% is always better than PAR of 10%. With regard to trends, decreasing absolute values of PAR (in comparison to the last reference period) imply a reduction in the default/credit risk. In contrast, increasing absolute values of PAR (in comparison to the last reference period) imply an escalation in the default/credit risk.

Thus, an MFI whose reported PAR was 10% at the end of the last period is said to have reduced risk in its portfolio if the reported PAR at the end of this period is 6% - this decreasing trend in absolute values of PAR by 4% typically implies a reduction in the default/credit risk. The same principles apply to the aged values of PAR with lower absolute values and decreasing trends implying lesser risk and higher absolute values and increasing trends suggesting higher levels of credit/default risk in the MFI’s portfolio.

However, there is one additional aspect to consider while interpreting the Aged PAR measures. Cetrus Paribus, the same absolute value of PAR in higher age categories implies a higher level of default/credit risk – i.e., PAR 181-365 Days = 5% implies a higher level of default/credit risk rather than PAR 1-30 days = 5%.  This is because in the case of the former, the clients have exhibited risky behavior (by having past dues) for greater than 180 days whereas in the latter case, they have had past dues for at most 30 days, which can (perhaps) be paid back and/or recovered more easily because they represent fewer skipped payments[5].

Operational Events that Distort PAR

As we noted earlier, a decreasing Portfolio at Risk (PAR) is positive. However, such a trend can be misleading because, a lower (PAR) ratio can in fact be obtained by decreasing the numerator and/or increasing the denominator. This can happen in several ways: (1) Rescheduling – whereby, risky and past due loans are made current. This decreases the numerator of PAR without affecting the denominator; (2) Refinancing – where, risky and past due loans are first rescheduled and extra amounts are given to the same clients. This decreases the numerator of PAR while simultaneously increasing the denominator; (3) Write-offs – where, risky and past due loans are written off from the portfolio. This decreases the numerator and denominator of PAR; and (4) Fresh Loan Disbursements[6] – where, new loans are disbursed to clients. This increases the denominator of the PAR measure.


Methodological Deficiencies that Impact PAR

Several methodological aspects also affect PAR and could result in an inaccurate portrayal of the default/credit risk prevalent in an MFI’s loan portfolio. These include: (1) Improper rules for classifying a loan as past due – sometimes, a loan may not be classified as past due even if the concerned installments have not been paid by the due dates. In some cases, due dates may themselves be variable. In still others, an official grace period[7] may prevent a risky (several installment skipped) loan from being classified as past due. Finally, classification of loans as past due may be based on ‘principal amounts’ that are past due rather than all amounts (including interest) that need to be considered; (2) Inappropriate sequence for appropriating client repayments – whereby, client repayments are first appropriated[8] towards principal and later towards interest; and (3) Incorrect method of ageing of past due loans – whereby, the installment method is used for ageing of past due loans. As has been demonstrated[9], the installment method of ageing understates age of a past due loan (if the concerned loan is beyond its loan term) and overstates age of a past due loan (if it is within its loan term). All of above methodological deficiencies have an impact on PAR as they reduce the numerator.

Implications for Credit Risk Management

Thus, as discussed, several events such as re-scheduling, refinancing, loan write-offs and fresh loan disbursements and methodological aspects including improper rules for classifying a loan as past due, inappropriate sequence for appropriating client repayments and incorrect method of ageing of past due loans, could result in a lower PAR ratio, while the default/credit risk may still be quite high.  While these situations have several implications for effective management of credit/default risk in an MFI’s portfolio, the most important among these is the fact that credit/default risk should be promptly recognized (so that it can be dealt with in the first place). This means that accurate, reliable and timely information should be available regarding the loan portfolio.

Among other things, this mandates the use of a standard best practices methodology in estimating the PAR measure. This would entail incorporation of several standard best practices features into the manual or automated MIS used by the MFI: (1) fixed and a priori terms for loans including repayment frequency, installment amounts and due dates, loan tenure etc. with provision for changes under special circumstances[10] along with prompt disclosures; (2) timely recognition of a loan as past due, even if a fraction of the interest and principal amounts due/past due are not paid by the said due date; (3) disclosure of unduly long grace periods and making appropriate adjustments to PAR for the same; (4) use of a correct sequence for appropriating client repayments whereby interest amounts are credited first and principal portions thereafter; (5) use of the best practices method for ageing past due loans whereby age of a past due loan is equal to the difference in the number of days between date at which PAR is being calculated and the date of the earliest unpaid past due amount[11]; and (6) disclosures of any loan rescheduling, refinancing, write-offs and fresh loan disbursements to delinquent clients (greening) as well as those for which repayments are yet to begin and making appropriate adjustments to PAR for all of these situations.

Only if all of these standardized best practices are implemented, can MFIs clearly understand the credit/default risks in their loan portfolio, set acceptable limits on these risks, and most importantly take the steps necessary to monitor and control these risks. Senior management of MFIs must therefore accept the responsibility to implement these best practices with regard to PAR and also ensure their consistent use in the operational situation. Otherwise, they run the biggest risk of delinquency, ‘the hidden beast’, manifesting itself suddenly and exploding to the detriment of their institution.

[1] This is the case with most MFIs, which tend to have a larger portfolio in loan as compared other financial products
[2] This is the standardised CGAP definition and is called the normal PAR.
[3] Arrears Rate is an optimistic default/credit risk measure as it uses the delinquent amount rather than the entire loan outstanding in the numerator.
[4] That is why the entire loan outstanding of delinquent clients is represented in the numerator of the formula given above.
[5] The loan repayment frequency (i.e., weekly, monthly etc) will have a significant difference in analysing this aspect. For example, in case of weekly payment loans, those deemed to be just 31 days past due could in fact be 4 payments skipped where as in the case of monthly repayment loans, they would be just 1 payment skipped.
[6] Sudden and large disbursements of loans could mask the actual default risk very significantly. Also, in an MFI that is fast expanding in terms of loan disbursements, the same applies. And when the repayment period for these fresh loans are yet to begin, the problem is really exacerbated. Fresh loans to delinquent clients also cause such problems.
[7] An MFI had a product where clients had to repay in 46 weekly instalments spread over 52 weeks. So, while MFI reported 0% PAR, as high as 23% of the clients with outstanding loans had skipped 4/5/6 weekly payments. In reality, PAR > 1 Day exceeded 38% when all past due loans (even those with 1 skipped instalment) were considered.
[8] Such a sequence of appropriation may reduce the principal amounts overdue and hence, PAR but more importantly, result in lower income for the MFI because of lost/delayed interest payments and thereby affect its sustainability and survival
[9] Please refer to Micro-Finance Capacity Builder, Volume 1, Issue 1, (2002) for a complete discussion of the best practices method of ageing and limitations in using the instalment method of ageing
[10] In case of natural calamities like Tsunami, earth quake etc
[11] Additional conditions would include consideration of all amounts due (principal, interest etc) as well as use of the correct method of client repayment appropriation.

No comments:

Post a Comment