About the ERAM
The ERAM combines banks’ interest margins with loan losses and provisions to measure the relationship between these metrics and visualise a large selection of banks, highlighting those that are most likely to show movement in risk-adjusted margins in future periods. The ERAM also shows the potential impact on profitability from such a move, which, in turn, is likely to affect valuation given the strong correlation between profitability and valuation. Furthermore, the findings from the ERAM provide a helpful tool to identify questions to ask banks in order to assess investment opportunities in the sector.
Valuing banks
Traditional cash flow-based models do not work for banks, so absolute valuation models rely more on discounted dividends or distributable earnings as proxies for free cash flow.
However, independent of the model used, banks operating in mature markets have historically been valued predominantly on their profitability, measured as ROE. Profitability determines the ability to generate distributable earnings.
| Exhibit 1 – Price-to-book value vs ROE (FY25) |
 |
| Source: LSEG Data & Analytics |
The R² of 0.426 when comparing profitability (expressed as FY25 ROE) against valuation (expressed as price to end of FY25 book value) across 18 banks suggests a relatively strong relationship between reported profitability and valuation. This is true despite a few banks’ profitability deviating materially from a more normalised ROE due to, for example, merger-related or restructuring costs incurred in the period.
We note that the relationship between profitability and valuation, although still relatively strong, is not as strong as it was when R² values around 70% were not uncommon. This could signal either that profitability is not as strong a valuation driver as it has been or that there are potential investment opportunities in the banking market.
We believe the latter is the case, as we believe the relationship between profitability and valuation seems likely to strengthen from here towards the stronger correlations seen in the past.
Interpretation
The interpretation of the above relationship is that, for each percentage point increase in ROE, banks are valued 7.67bp higher relative to their stated equity. For an ROE of zero (break-even), the market is valuing a bank at 38.52% of its stated equity.
Assuming relatively similar longer-term growth profiles, the relationship between ROE and valuation suggests an implied cost of equity of c 8% for the banks in our group.
Peer group and method
The chosen banks represent a cross-section of European, full-scale prime lenders (we have not included specialised sub-prime lenders). The banks vary in size, including small and large institutions, offer both secured and unsecured lending and all show net losses in their loan losses and provisions P&L line (ie a cost of risk) during the period (FY24 and FY25), so no net loan loss reversals. We use the P&L and balance sheet numbers as reported (ie no adjustments are made to the margin/cost of risk comparison). We use the latest full-year numbers. All numbers, including ROE and valuations, are from LSEG Data & Analytics.
We make only one adjustment to the ROE/price-to-book value comparison in the case of AL Sydbank, where we adjust ROE for the substantial merger-related costs incurred in FY25. We believe that in this case, the market has a ‘clear line of sight’ to a higher future profitability that is largely reflected in the market valuation.
Alternative comparisons of valuation and profitability suggest the stock market in general tends to rely more on as-reported numbers than underlying and adjusted alternative performance measures when valuing banks. Running the valuation against the banks’ ROE targets, where available, indicates that the market is in general not giving credit for these targets – unless there is a clear path to reaching the targets. We have also run the valuation against consensus estimates where these are available and find a similar strength in relationship to that found using the latest full-year numbers.
However, using consensus estimates rather than reported figures results in new questions about how those estimates are derived, such as the number of forecasts used, the potential influence of outliers and the availability of consensus data for smaller banks.
In addition, we note that the accounting standard, IFRS9, now fully embedded in banks’ reporting, requires banks to be forward-looking in their provisioning, making provisions more indicative of expected future loan losses and thereby supporting our comparison of margins with loan losses and provisions.
Valuation considerations
In addition to profitability, both growth and risk are factors in a bank’s valuation.
- Justified price-to-book value = profitability minus growth, divided by risk minus growth
However, differences in growth outlook appear to be a relatively minor factor for the bank valuations in the selected group, which are all operating predominantly in mature banking markets and, hence, have similar structural long-term growth outlooks.
Furthermore, differences in risk remain a relatively modest factor at this point in the economic cycle, where loan losses remain relatively low across lending classes and funding remains readily available at a reasonable price.
Our analysis of the importance of growth, risk and profitability for mature market banks through the economic cycle suggest that, generally, profitability is the dominant valuation driver. Only during and in the immediate aftermath of the global financial crisis did risk (using credit default spreads as an indicator) become the key valuation driver for a couple of years. In addition, leading up to the global financial crisis, growth was briefly the key valuation driver for mature market banks.
The question becomes ‘what drives profitability?’
With profitability being the key driver of European bank valuations, the question becomes what determines it?
A bank’s profitability is determined by three factors:
- the reward for taking credit risk (risk-adjusted margin);
- cost efficiency; and
- capital efficiency.
Combined in the right way, these three factors determine a bank’s ROE.
The reward for taking credit risk
With NII accounting for c 80% of total earnings, a bank’s earnings are, to a large degree, a reward for taking credit risk. However, the risk/reward relationship is often looked at one dimensionally, with each element considered separately.
The two graphs below are examples of how a bank’s net interest margin (NIM) and credit risk cost are often shown.
| Exhibit 2 – NII lending |
 |
| Source: Edison Investment Research and LSEG Data & Analytics |
| Exhibit 3 – Loan losses and provisions/lending |
 |
| Source: Edison Investment Research and LSEG Data & Analytics |
A two-dimensional view
Running the same numbers against each other gives a fuller picture and helps answer questions such as:
- Is there a risk/return relationship? How much of the realised margin is explained by the credit risk taken?
- If there is a relationship, what does it look like and how much extra margin can be expected for extra credit risk?
- What yield can be expected for no loan losses and provisions?
In addition, looking at it this way enables us to identify outliers (ie the banks furthest away from the line).
| Exhibit 4 – NII lending/loan losses and provision lending |
 |
| Source: Edison Investment Research and LSEG Data & Analytics |
Interpretation
Based on full-year results for 18 lenders, we found a clear risk/return line with an R² of 0.6843. The formula for the relationship between margin and loan losses was y = 12.487x + 0.0126 where:
- 12.5bp is the additional basis-point margin for 1bp higher losses and provisions; and
- 0.0126 (=1.3%) is the expected margin for zero loan losses.
So, for each basis point of additional loan losses and provisions, the selected banks have 12.5bp of additional margin and, for no loan losses, they earned a 1.3% yield on the loan book.
Real-life implications
While the above might seem a bit academic, it has real-life implications. Below, we list some of the key benefits of using the ERAM as part of the wider ‘research funnel’ to narrow the large bank universe and find the most appealing investment opportunities.
Finding the investment case
First, we identify banks that are on or close to the ‘best-fit’ line and exclude them from further analysis – in these cases, the credit risk-adjusted return on the loan book does not appear to be part of the investment case, as they broadly earn the margin suggested by the credit risk taken.
We note that in any given period, differences in treasury strategies and/or provisioning policies are likely to cause some deviations in outcomes. Furthermore, significant changes in the loan book over the period could result in ‘timing differences’ that distort the relationship between margin and loan losses.
Our focus is on the outliers, or banks that lie meaningfully above or below the trend line. Those above the line earn margins that exceed what their level of credit risk would imply, while those below the line earn margins that are lower than their credit costs would suggest.
In this case, Paragon, Commerzbank, AL Sydbank and ING (all meaningfully below the line) and Metro Bank, Groenlandsbanken and Laan & Spar (all well above the line) are flagged as warranting further analysis.
Aiming to improve investment performance
When there is a strong relationship between NIM and losses and provisions, suggesting the lending markets are very efficient (ie the margin received to a large degree depends on the risk taken), the outliers are, all else being equal, likely to be pulled towards the line in future periods, giving a basis for anticipating future moves.
A predictive tool
The group’s FY24 numbers showed an R² of 0.64 (ie lower than the 0.68 for FY25), suggesting that risk pricing had become tighter in the period as the banks’ NIMs versus loan losses and provisions were generally closer to the predicted outcome in 2025 than in 2024. Below we show two examples (Bank of Ireland and Fynske Bank) of such predictability, where NIM levels above the line predicted by the loan losses and provisions/lending ratio returned to the line in the subsequent period.
- Bank of Ireland was an outlier in FY24, earning a NIM in excess of that predicted by loan losses and provisions. However, in FY25 it was back on the line, after the margin fell more than the average for the group, while losses and provisions relative to lending increased more than the average.
- Fynske Bank (Denmark) was also an outlier in FY24, earning more NII than suggested by loan losses and provisions booked in the period. However, in FY25, the bank fell close to the line (ie no longer an outlier). Although loan losses and provisions had increased c 10bp against a stable level for the whole group, the NIM had fallen more than 3x as much as the average.
- Groenlandsbanken was another outlier in FY25, with a higher than expected margin for the level of loan losses and provisions. However, in FY24 the bank also sat above the line. We believe this partly reflects its strong market position with a high market share in a relatively small and protected duopoly market. We would therefore not expect a significant pull towards the line in future periods. We note that despite this, Groenlandsbanken is trading at a discount to its peers.
Finding the questions to ask
A high R² not only suggests there is a significant relationship between NIM and loan losses and provisions, it also points to a relatively ‘tight relationship’ (ie that all observations/lenders plot relatively close to the expected level). This means that observations significantly away from the line (ie outliers) warrant further analysis.
In the case of Metro Bank, Paragon, Commerzbank, AL Sydbank and ING, we need to help identify why they are positioned where they are and thus predict if they are likely to move towards the line in future periods. Does their positioning reflect the strength of the franchise or is it a timing effect?
A formula for risk pricing
The slope says something about how much lenders are being paid for taking extra credit risk. For example, for 1bp extra credit cost, lenders received 12.5bp extra margin in FY25.
Looking at the current risk/reward relationship is informative, but tracking movements adds valuable insight. Does the slope steepen or flatten? Does it suggest that the cost of the credit risk part of the lending rate is moving up or down? Do changes in the credit risk versus reward follow credit spreads in the bond market?
We ran the same regression using the same names for FY24 and found the credit line had steepened from 12.3x in FY24, suggesting a slight increase in risk pricing.
We aim to update the ERAM regularly with new numbers, in order to track movements in how banks price credit risk.
A formula for the risk-free rate for banks
The ‘intercept’ (0.0126 or c 1.3% for FY25) is the return on the loan book lenders can expect if they have zero loan losses and provisions.
When we ran the regression for the same group of banks using their FY24 numbers, we found the intercept showing the margin at zero loan losses and provisions at 1.59% for FY24, thus suggesting a c 40bp reduction between 2024 and 2025, which compares to a c 50bp reduction in time weighted average policy rates in the EU and the UK in that period.
Forming an investment narrative
Over time, some of the conclusions from the ERAM, such as the steepness of the line or risk-free rate in a given period, might be surprising. However, they build on reported numbers – so no black boxes – and the findings help inform further analysis and questions to ask and help set a different investment narrative.
A framework for scenario analysis
The ERAM also provides a framework for scenario analysis, enabling us to assess, for example, what effects changes in interest rates, new liquidity or funding rules might have on margins and lenders’ profitability.
What’s next?
Lenders’ profitability and valuations are determined by the risk-adjusted margin, cost efficiency and capital efficiency. We aim to update our ERAM regularly and will publish separate notes on the cost and capital efficiency needed to complete the analysis of the profitability drivers in the coming months.
