General Information & Client Services
  • Americas: +1.212.553.1653
  • Asia: +852.3551.3077
  • China: +86.10.6319.6580
  • EMEA: +44.20.7772.5454
  • Japan: +81.3.5408.4100
Media Relations
  • New York: +1.212.553.0376
  • London: +44.20.7772.5456
  • Hong Kong: +852.3758.1350
  • Tokyo: +813.5408.4110
  • Sydney: +61.2.9270.8141
  • Mexico City: +001.888.779.5833
  • Buenos Aires: +0800.666.3506
  • São Paulo: +0800.891.2518

In this article, we analyze the potential effects of upcoming CECL regulations on lenders and explore the impact of CECL under different Moody’s Analytics scenarios. A poorly timed transition could lead to a market-wide liquidity shortage or a crisis in economic activity. We provide suggestions on how the transition to CECL can be managed smoothly for minimal economic impact.

The switch in accounting rules to a current expected credit loss (CECL) framework is intended to increase stability in the financial system and improve liquidity throughout the economic cycle. Under the new framework, firms will begin reserving for potential losses when they first book loans rather than setting aside reserves only after loan performance deteriorates.

However, as with most changes in rules and regulations, what looks reasonable and appropriate in theory may not turn out as such in practice. At a minimum, CECL will lead to front-loading losses relative to the current system. Transitioning from the current system to this new approach may inject some volatility into bank earnings and profitability. CECL also introduces uncertainty into accounting calculations, as economic forecasts are imperfect over long horizons.

In this article, we consider the adoption of CECL with an eye toward assessing its potential benefits – and risks – to the financial system and the broader economy.

Procyclicality Gone Wild

Current accounting rules utilize a “probable and incurred loss” standard which requires lenders to reserve an allowance for loan and lease losses (ALLL) by applying recent performance trends to their outstanding books of business. So, if 10% of loans with certain characteristics have defaulted in the recent past with no recoveries, then lenders should assume the same going forward and add 10% of outstanding balances to their loss reserves. The benefit of this approach is that it is relatively simple to implement and is seemingly objective, as it does not permit the lender to make any rosy assumptions about future performance that would cause it to under-reserve.

But this assessment is not quite correct. Simplicity can come at the cost of accuracy. Lenders need to categorize or cohort their portfolios in order to calculate the historical loss rates to be applied to their current books of business. Just as politicians can influence election outcomes by creatively defining voting districts (i.e., gerrymandering), lenders’ discretion in determining the cohorts or segments of their portfolios could have an impact on computed loss rates. Auditors and regulators may review and challenge lender processes, but some risk remains.

In addition, lenders could influence reported outcomes through the determination of an appropriate loss emergence period. Typically, consumer loans do not default instantaneously. Many borrowers who miss a loan payment are able to catch up and cure before transitioning to a deeper state of delinquency or default.

Under current accounting rules, lenders need to account for this process when assessing the likelihood and severity of losses in their current portfolios. Based on the performance history of their own portfolios, they may determine the average number of months it takes for loans within a given book of business to experience losses. They then look back over recent history for a similar number of months to make their historical loss calculations. Again, while the determination of the emergence period may be largely objective, some discretion in analytical choices can influence results.

Perhaps the biggest criticism of the current process is that it is backward-looking. By restricting default analysis to recent history, loss reserves can become highly procyclical.

Perhaps the biggest criticism of the current process is that it is backward-looking. By restricting the analysis to recent history, loss reserves can become highly procyclical, as shown in Figure 1. Leading up to a recession, loss reserves are low and firms must rapidly add to their ALLL as delinquencies and defaults soar.

Such behavior can exacerbate the recession as lenders are forced to pull back from supplying credit at precisely the time that borrowers and the economy may need credit the most.

Figure 1. Total loss reserves at FDIC-insured institutions
Total loss reserves at FDIC-insured institutions
Sources: FDIC Quarterly Banking Profiles, Moody's Analytics

Lenders also end up over-reserving toward the end of recessions, when realized losses fall as the economy improves. The capital release that follows introduces volatility into the system as lenders flush with capital scramble to deploy it wherever possible, leading to loosened standards and the heightened potential for mal-investment and bubble formation.

This procyclicality was evident during the Great Recession and was one of the motivations behind the adoption of the CECL standard. In fact, CECL was initiated by the Financial Crisis Advisory Group (FCAG) and is widely supported by US banking regulators. Figure 2 shows that the increase in the reserve rate in commercial banks lagged the increase of noncurrent loans by several quarters in 2009. Furthermore, the reserve rate declined more slowly than the noncurrent rate in 2012.

Figure 2. Reserve and noncurrent rates for loans and leases at commercial banks
Reserve and noncurrent rates for loans and leases at commercial banks
Sources: FDIC Quarterly Business Report, Moody's Analytics

The divergence between reserve rate and noncurrent rate was even larger for community banks, as shown in Figure 3, although this is largely a function of the higher credit quality of loans at these institutions; note the difference in scale on the y-axes of Figures 2 and 3. The community bank experience is closer to the ideal envisioned under CECL, where reserves are sufficiently high at loan origination and require only small additions when the economy moves into recession.

CECL to the Rescue

Recognizing the flaws in incurred loss accounting, the Financial Accounting Standards Board (FASB) proposed CECL, which requires firms to estimate lifetime expected loan losses starting from the date of inception. In this approach, loss projections are less dependent on recent history and there is less room for individual discretion regarding segmentation and emergence periods.

Figure 3. Reserve and noncurrent rates for loans and leases at community banks
Reserve and noncurrent rates for loans and leases at community banks
Source: Moody's Analytics

However, the new approach is not without its own challenges and potential pitfalls. Many firms may have insufficient data with which to estimate lifetime losses, thereby requiring some supplementation with external sources. Figure 4 provides a list of information necessary to successfully conduct CECL calculations. Even large lenders – which have already gone through Dodd-Frank Act Stress Tests (DFAST) and Comprehensive Capital Analysis and Review (CCAR) regulations and hence are familiar with the types of models needed for CECL – may wish to take advantage of large, industry-level datasets in order to more easily justify the objectivity of their processes.

Given that CECL introduces an element of forecasting to the loss reserving process, auditors and regulators may be justifiably concerned that firms could assume an economic outlook that projects a more favorable – but less realistic – outcome in order to minimize the amount of money they need to set aside. While specific guidance from auditors or regulators has not been issued, we believe that one of two approaches for determining economic scenarios will be acceptable.

Figure 4. Data necessary for CECL compliance
Data necessary for CECL compliance
Source: Moody's Analytics

Under the first approach, lenders would estimate losses on their loans under multiple scenarios: one upside, one downside, and one baseline scenario. In this case, the reported losses under CECL would be derived as a probability-weighted average of the likelihood of each scenario – rather than relying solely on a single econometric model for scenarios.

Alternatively, firms and auditors may wish to adopt a consensus-based approach when determining the economic scenario to use for CECL. That is, they may prefer to average the baseline economic projections of multiple government and professional economic forecasters to create a single consensus scenario of the most likely path for future economic activity and growth. Lenders would then use this scenario to generate their CECL loss numbers with no additional weighting required.

Modeling Options for CECL

As lenders begin to consider the impact of CECL on their accounting processes, they will also consider which models are most appropriate for their situations. A variety of approaches are available, ranging from roll-rate and vintage-cohort models to more sophisticated expected loss and regression models. Lenders that have been through the Federal Reserve’s DFAST or CCAR stress testing process may be tempted to reuse their models for the CECL exercise. Recycling or adapting existing models for CECL would be cost-effective, but there may be some concern that models developed for stress testing may be overly conservative for financial accounting purposes.

CECL models share many of the same characteristics as stress testing models. That is, models for both objectives should account for the life cycle of loans, origination vintage effects, time varying effects, seasonality, and a variety of borrower and loan characteristics. Depending on the particular model specifications, most models developed for stress testing could potentially be used for CECL calculations with few modifications. See Figure 5 for an example of dollar loss rates projected 30 years out for a bank portfolio.

While larger lenders may develop their own internal models for CECL, smaller lenders may be unable to do so, given a lack of historical performance data. Formal development of models including backtesting, sensitivity analysis, documentation, and validation may also prove cost-prohibitive to some lenders. The availability of cohort-level models estimated on industry-wide consumer credit data allows lenders to easily obtain current expected credit loss estimates across all consumer credit products by vintage-cohorts. This permits lenders to generate forecasts by simply merging or looking up projected losses by loan category – even for purchased portfolios for which they have limited performance information. Because industry models have been developed and validated on the entire universe of data, lenders can obtain robust estimates for specific loans on their books.

Figure 5. Projected dollar loss rates
Projected dollar loss rates
Source: Equifax, Moody's CreditCycle™

Lenders can also use industry models to produce credit loss estimates of future loan originations to estimate and prepare for CECL’s impact on new loan bookings. Although lenders are strongly encouraged to start preparing for CECL by performing gap and impact analyses, lenders must maintain allowances for incurred loans and leases in compliance with current generally accepted accounting principles (GAAP) until CECL is officially adopted. As a result, prudent lenders will likely increase retained earnings in anticipation of CECL. This will make their Tier 1 capital ratios look impressive in the short-term (to the delight of regulators) while limiting dividend payments and share buy-backs (to the dismay of shareholders).

Timing is Everything

Given our assessment of the current and future states of the loss reserving process, we believe the adoption of the CECL framework is a positive step, both in terms of providing investors with a more accurate assessment of the financial positions of lenders, and in terms of improving the stability of the overall financial system. Firms must adopt the CECL framework by 2019 or 2020, depending on firm size, which allows sufficient time to change internal systems.

But in our opinion, they cannot start the transition soon enough. As described in the previous section, there are multiple moving parts in the process that will require significant time to develop, test, and deploy, especially if a lender has not been through the CCAR/DFAST stress testing processes yet. Given that most lenders will need to increase their loss reserves once CECL takes effect, it would be prudent to run both the current and the new accounting standards in parallel so lenders have ample time to transition.

Adoption of the CECL framework is a positive step, both in terms of providing investors with a more accurate assessment of the financial positions of lenders, and in terms of improving the stability of the overall financial system.

Figures 6 and 7 give a simple example of such an analysis, comparing the current ALLL incurred loss approach to CECL using the following simplifying assumptions:

  1. Assume five-year installment loans with each vintage originating with a $100 balance.
  2. Assume each vintage follows the same pattern of losses over five years (i.e., 10% cumulative loss rate with $1 of loss in the first year, $2 of loss in the second year, $4 of loss in the third year, $3 of loss in the fourth year, and $0 of loss in the fifth year).
  3. Assume perfect foresight in reserving so that each year the lender can perfectly anticipate losses in the following year.
  4. Assume CECL takes effect in 2020.
  5. Assume 0% discount rate for the sake of simplicity.
Figure 6. Example of reserve contributions by origination vintage under the incurred loss approach
Example of reserve contributions by origination vintage under the incurred loss approach
Source: Moody's Analytics
Figure 7. Example of reserve contributions by origination vintage under the CECL approach
Example of reserve contributions by origination vintage under the CECL approach
Source: Moody's Analytics

This simple example illustrates the potentially substantial effect of CECL, as all future losses on existing loans will need to be reserved instantaneously in 2020. In reality, the impact of CECL for each lender will depend on several factors, including:

  • Age and expected remaining life of the loans in a portfolio. For example, the larger the number of new originations at the time of transition, the bigger the impact.
  • Portfolio quality, defined by origination credit score, loan-to-value ratio, debt-to-income ratio, etc. The impact will be lower on higher-quality portfolios.
  • Types of loans in portfolio. Installment versus revolving as future draws could impact loss reserves.
  • Terms of loans. For example, longer-term loans could lead to higher loss reserves under the life-of-loan assumption.
  • Geographic location of loans. Geography will affect the quality of the portfolio. Exposures in stressed areas could have higher loss projections.
  • Current status of loans. Loans that are currently delinquent will have higher loss projections than non-delinquent loans.

Finally, the impact of CECL will depend on the economic conditions at the time of loan origination as well as every subsequent reporting period.

From an economic perspective, the timing of the transition will be critical. CECL front-loads losses, as compared with the current system. As an immediate result, firms will need to significantly increase overall loss reserves from current levels. According to an analysis performed by the Office of the Comptroller of the Currency (OCC), firms may need to increase their ALLL by as much as 30% to 50% over current levels.1 If lenders plan for this eventuality over the next three to four years, the overall impact to both earnings and the economy should be minimal. Firms may retain more of their earnings and report lower profits than they might have previously, but investors will have an understanding that lender profitability will be less volatile in the future.

Firms may need to increase their ALLL by as much as 30% to 50% over current levels. If lenders plan for this eventuality over the next three to four years, the overall impact to both earnings and the economy should be minimal.

If lenders wait, however, and rush to increase reserves closer to the deadline, it could significantly impact profitability. In a worst-case scenario, the rush could lead to a liquidity crisis as firms hoard funds and drive up the cost of capital in a mad dash to comply with regulations. Such a financial shock would be felt immediately in the real economy as banks reduce lending to both the commercial and household sectors. Economic activity would slow as a result of a credit crunch.

Figure 8 provides some sensitivity analysis around the potential increase in reserve allowances by assuming various impact levels of CECL. We compare reserve amounts from the start of the Great Recession (2007Q4) with those realized at the middle of the recession (2008Q3) and at the end of the recession (2009Q2). We also report the realized allowance for 2010Q2 when reserves hit a historical maximum. Finally, total outstanding reserves as of 2016Q1 were reported to be around $120 billion. Therefore, if CECL went into effect today and the impact was 30%, then FDIC-insured institutions would need to increase their reserves by about $36 billion based on today’s numbers.

Figure 8. Potential increase in reserve allowances for all FDIC-insured institutions, assuming various impact levels of CECL
Potential increase in reserve allowances for all FDIC-insured institutions, assuming various impact levels of CECL
Source: Moody's Analytics

Economic conditions could change by the time CECL takes effect, so we also consider what the reserves could be at a date closer to CECL’s 2020 implementation, given potential changes in the economic environment.

In order to estimate an upper bound on CECL’s impact, suppose that lending standards loosen over the next few years and the economy experiences another Great Recession starting in 2019, just as CECL is scheduled to take effect. Also assume that banks need $251.6 billion in reserves, as they did during the Great Recession, and that the transition to CECL will require a 30% immediate increase in reserves. Banks would need to add another $75.5 billion to their reserve amount.

In other words, a total of $327.1 billion would need to be dedicated to reserves and would therefore be unavailable for lending to consumers. A potential credit crunch as such could be exacerbated if lenders held back new originations to reduce the impact of CECL, as new originations will impact reserves the most. The effect could be increased further if banks have insufficient capital to meet their reserve obligations. This could force the Federal Reserve to intervene and either increase discount window borrowing or lend directly to institutions – as occurred during the last recession.

Allaying some of the concerns around the adoption of CECL is the fact that the US banking system is well-capitalized. According to the FDIC, both commercial banks and savings institutions increased their levels of capital while simultaneously reducing the number of nonperforming loans in their portfolios in the wake of the Great Recession. Capital ratios have increased to record high levels as a result, far exceeding the 8% ratio that defines wellcapitalized institutions according to the Federal Reserve’s Regulatory Capital Guidelines2 (see Figure 9). Utilizing some of this capital to meet CECL obligations would still leave the banking system as a whole adequately capitalized, although some individual institutions would undoubtedly be strained.

Figure 9. Total risk-based capital ratios
Total risk-based capital ratios
Source: FDIC Quarterly Banking Profile

While the Moody’s Analytics baseline economic forecast suggests a much more modest scenario than a severe downturn, such an outcome is not without precedent given the Great Recession. The opaqueness of credit market derivatives, combined with strict market-to-market accounting rules, exacerbated the financial stress caused by the collapse of Lehman Brothers across the financial system. A recession may have been inevitable, considering imbalances introduced by over-investment in the housing sector. But additional flexibility and forbearance in the financial system may have prevented a garden-variety recession from turning into the Great Recession.

Moreover, the regulations brought on by DFAST caused further tightening in credit markets and are cited as one of the reasons the recovery this time around has been one of the slowest. Some of the consumer credit markets such as mortgage and bankcards have started to recover only recently and are seeing a slight loosening in standards. Another tightening in credit caused by another policy change – if not timed correctly – could precipitate an unforeseen chain of events.

The Future is Now

By acting as a countercyclical buffer, CECL holds great potential to improve the stability of banks and the overall financial system, but only if the transition is orderly. Lenders need to start preparing as soon as possible, and regulators need to be ready to adjust to conditions on the ground as the CECL deadline approaches. With the labor market steadily improving and consumer credit losses near record lows, the current environment is ideal for lenders to prepare for the transition. Should the implementation of CECL coincide with a stumble in economic performance, the benefits of transitioning will be muted at best and could trigger a recession at worst. For their own benefit, as well as the benefit of the financial system and the broader economy, all lenders should start preparing for CECL without delay.

Footnotes

1 Curry, Thomas J. “Remarks by Thomas J. Curry, Comptroller of the Currency, Before the AICPA Banking Conference, Washington, D.C.“ September 16, 2013.

2 Federal Reserve System. “Federal Register, Vol. 78, No. 198.“ October 11, 2013.

As Published In:
Related Insights

When Good Data Happen to Good People: Boosting Productivity with High-Quality Data

In this article, we show the mechanisms through which data quality and productivity interact, and how investments in data quality can offer productivity gains.

Reconsidering Risk Management, Governance, and Stress Testing

This article discusses areas such as capital stress testing where simplification of regulations could improve the flow of credit while protecting the financial system.

July 2017 WebPage Dr. Deniz Tudor

When Good Data Happen to Good People: Boosting Productivity with High-Quality Data

With ever-increasing requirements for a higher quantity and quality of analytical output, the need to boost productivity in risk management has become more acute. In pursuing these productivity gains, we have observed that investments in data quality can offer dramatic improvements and typically pay for themselves.

CECL Quantification: Retail Portfolios

In this webinar, our experts discuss the important considerations in the modeling and implementation of the CECL standard for retail portfolios. Learn more about loan-level modeling approaches that can be used to forecast credit losses for retail portfolios and how to leverage existing risk measurement practices.

CECL Quantification: Retail Portfolios Webinar Slides

In this presentation from our webinar, our experts discuss the important considerations in the modeling and implementation of the CECL standard for retail portfolios. Learn more about loan-level modeling approaches that can be used to forecast credit losses for retail portfolios and how to leverage existing risk measurement practices.

CECL Quantification: Retail Portfolios Webinar Slides

In this webinar, our experts discuss the important considerations in the modeling and implementation of the CECL standard for retail portfolios. Learn more about loan-level modeling approaches that can be used to forecast credit losses for retail portfolios and how to leverage existing risk measurement practices.

Expanding and Regionalizing the Federal Reserve CCAR Scenarios

Moody's Analytics will expand each of the Fed's scenarios to produce forecasts for the full set of more than 1,500 variables found in its own macroeconomic forecasts.

March 2017 Pdf Celia Chen, Dr. Cristian deRitis, Ed Friedman, Adam Kamins

CCAR and the Paradox of Lower Losses

Now that bank finances and risk management are stronger and data are more reliable, we focus on the quantitative accuracy of the stress tests themselves and their reasonableness vis-a-vis the Great Recession experience in this article.

April 2014 Pdf Dr. Cristian deRitis, Mustafa Akcay, Jeffrey Hollander

Stress Testing Webinar Series: Macroeconomic Conditional Loss Forecasting Presentation

In this Moody's Analytics webinar, Thomas Day and other Moody's Analytics experts discuss Macroeconomic Conditional Loss Forecasting. Given the criticality of loss estimation, and the need for different models by asset class, we cover loss estimation for Retail Exposures (non-mortgage), Structured Portfolios, Wholesale C&I (non-public), and Wholesale (public).

October 2013 Pdf Thomas Day, Dr. Cristian deRitis, Luis Amador

The Moody's CreditCycle Approach to Loan Loss Modeling

This whitepaper goes in-depth into the Moody's CreditCycle approach to loan loss modeling.

CECL's Implications for Bank Profitability, System Stability, and Economic Growth

In this article, we analyze the potential effects of upcoming CECL regulations on lenders and explore the impact of CECL under different Moody's Analytics scenarios. A poorly timed transition could lead to a market-wide liquidity shortage or a crisis in economic activity. We provide suggestions on how the transition to CECL can be managed smoothly for minimal economic impact.

Expected Consumer Credit Losses (ECCL) Service

Through econometric models based on Equifax industry performance data, Moody’s Analytics can quickly deliver net present value (NPV) of forecasts under the CECL standard for “reasonable and supportable” economic scenarios. Ideal for small consumer lending portfolios or as a benchmark for large consumer lending portfolios, we provide a dashboard plus documentation summarizing the results, assumptions, and model methodology. Our service can help you effectively analyze your portfolio for all consumer credit product lines.

WebPage Dr. Deniz Tudor