Managing earnings volatility: Non-GAAP versus OCI

The remeasurement of financial assets and liabilities to current value can be a source of significant earnings volatility. Many companies mitigate this by presenting an adjusted (non-GAAP) performance metric or by reporting some gains and losses in other comprehensive income (OCI).

The insurance sector, and the recently introduced IFRS 17, provides a good example of the communication challenge arising from volatile markets. IFRS reporters are split, largely along geographical lines, between applying the OCI option in IFRS 17 and a non-GAAP approach. Both approaches present challenges for investors.


Companies do not like earnings volatility – most prefer to report a steady increase in profit whenever possible.

If volatility arises from one-off gains and losses (such as the profit on disposal of a subsidiary), companies may present adjusted (non-GAAP or non-IFRS) profit metrics that exclude these items. Although there may be some companies that are unfortunately more inclined to exclude the unusual debits rather than credits, and some that exclude items for somewhat dubious reasons1Stock based compensation is a good example. See our article ‘Dot-com bubble accounting still going strong’., we think investors invariably benefit from having further income and expense disaggregation and additional management determined subtotals2See our article ‘Non-GAAP is more than earnings before bad stuff’.. We also welcome the formalisation of management performance metrics by the IASB, and the related additional disclosures that will be required when IFRS 18 becomes effective in 20273 See our article ‘Why IFRS 18 is good news for investors’..

Volatility caused by financial market effects presents a greater communication challenge

However, if earnings volatility arises from changes in the value of assets and liabilities due to market price movements, effective communication of underlying performance is more challenging. In this case, the gains and losses are recurring but volatile. Simply omitting them entirely leads to incomplete metrics that are unlikely to adequately communicate current performance or provide investors with a sound basis for forecasting. Instead, most insurers disaggregate the value change into an underlying amount that better reflects periodic performance (and which hopefully also has greater predictive value) and a residual market volatility effect. There are essentially two approaches:

  • Expected return based disaggregation using non-GAAP measures
  • Cost accounting based disaggregation using other comprehensive income (OCI)

Both of these are widely used by insurance companies that report under IFRS 17. Volatility caused by financial market effects is often highly significant for insurers considering the relative size of their balance sheets, and that (most) financial assets are measured at fair value and insurance liabilities at updated current values4At least the insurance liability fulfilment value is a current value – the CSM component of insurance liabilities accounted for under the general measurement model is not updated for financial market effects, nor is the liability for remaining coverage of contracts accounted for under the premium allocation approach..   

IFRS 17 has some anti-volatility elements – notably for non-financial estimate changes and the shareholders’ share gains and losses for participating contracts, both of which are deferred in CSM. Additionally, some financial assets may be measured at amortised cost, for which market volatility is largely only found deep in the footnotes. Nevertheless, significant volatility may remain in financial statements. Even if insurers hedge financial market effects, the focus on regulatory capital in hedging may still leave volatility.

Unfortunately for investors, both the above disaggregation approaches are complicated by the mixed measurement applied in financial statements. The expected return disaggregation will likely only apply to part of the asset and liability portfolio, with an element of cost measurement remaining. Likewise, it is not possible to fully apply cost accounting in profit and loss when using the OCI approach. 

In our article, ‘IFRS 17 insurance: Economic versus accounting volatility’ we explain the sources of volatility for insurance companies – we suggest investors unfamiliar with the subject read this before continuing with this article. For an explanation of IFRS 17 and insurance accounting jargon see ‘IFRS 17 insurance: More comparability and new insights’.

The non-GAAP and OCI approaches we describe are simply ways to present additional, and arguably more useful, performance metrics. Reported shareholders’ equity and reported comprehensive income (and reported earnings in the case of the non-GAAP approach) are still likely to be volatile, although these may be managed by the selective use of cost measurement for some financial assets.

Expected return based disaggregation: Non-GAAP earnings

An expected return disaggregation involves analysing the total net financial result (investment income less insurance liability accretion), into what return would be expected (based on values and market rates of return at the beginning of a period) and a residual. This split is not directly available under IFRS 17 (nor is it applied in any other IFRS standard, although there is one use in US GAAP5The expected return approach is applied in pension accounting under US GAAP (and previously in IFRS pension accounting until it was removed several years ago). In US GAAP the expected return on pension fund assets is reported in the income statement and the remaining unexpected portion of the fair value change is recognised in OCI. For more about this topic see our article ‘Pension leverage under IFRS and US GAAP’.), and therefore its application requires a non-GAAP approach.

The non-GAAP performance measure includes only the expected asset return and expected liability accretion. The residual fair value change for the assets and the change in liability due to discount rate changes, for example, are excluded.

Insurers that adopt this approach will tend to measure financial assets at fair value through profit and loss6 Although there is no reason why the non-GAAP and OCI approaches cannot be combined (see Axa below), doing so will inevitably lead to more confusion for investors.. Fair value measurement is required for some assets under IFRS 9. However, for those assets where the default measurement is amortised cost or fair value through OCI, insurers wishing to apply an expected return based disaggregation will likely use the fair value option7The IFRS 9 fair value option permits companies to measure any financial asset at fair value through profit and loss if doing so corrects an accounting mismatch or if the asset is managed and performance evaluated on a fair value basis. Both of these reasons can apply to insurers, which means that the fair value option can be applied to virtually any financial assets not already required to be measured at fair value through profit and loss. to measure these assets at fair value through profit and loss instead.

Non-GAAP measures based on expected returns may not be comparable

UK insurer Aviva provides a good example of the non-GAAP approach to managing volatility. The company excludes ‘investment variances and economic assumptions’ from their non-GAAP performance metrics. The other adjustments in the note below are the more usual adjustments found in most sectors – it is the removal of the ‘unexpected’ net financial result that is peculiar to insurers.

Aviva adjusted operating profit

Aviva 2024 financial statements

If we assume (see below) that the remaining amount of Aviva’s net financial result (investment income less insurance liability financial expense) is the market net expected return, we can derive the following disaggregation. Notice that the expected return component is significantly less volatile than the total net financial result, with the variation largely reflecting changes in market interest rates.

Net financial result expected return disaggregation – Aviva

Aviva financial statements

The benefit of this approach is that, if the disaggregation is realistic and unbiased, the expected net financial result should be a good basis for forecasting, although it would need to be adjusted to allow for interest rate changes. Furthermore, the group adjusted operating profit is arguably a more useful measure of performance.

An expected return based disaggregation works well if all financial assets are measured at fair value and the full insurance liability at a current value; the challenge for investors is that neither of these is likely to apply in practice.

On the liability side, the financial expense for the contractual service margin (CSM) component of the insurance liability is based on a locked-in interest rate. This means part of the interest expense is based on what is, in effect, cost accounting and is not a true expected return.

Some assets may be measured at cost to reduce capital volatility

On the asset side, insurers adopting the non-GAAP approach may choose to measure some assets at cost. This is not a free choice – IFRS 9 restricts cost measurement to simple principal and interest debt instruments which are held for collection, meaning there are little or no asset sales prior to maturity. While most financial investments held by insurers are required to be measured at fair value, in practice insurers are likely to have enough flexibility in their asset portfolio to measure at least some at cost.

To the extent that financial assets exceed the fulfilment value of insurance liabilities (the part that is updated for financial market effects), measuring at cost will reduce the volatility of shareholders’ equity. The downside for investors is that the profit and loss disaggregation becomes more complex. The ‘expected asset return and insurance financial expense’ in the above table is actually a combination of current market rates and historical (cost) rates. However, we do not believe the effect of cost measurement is particularly significant for Aviva, which adopts FVPL for almost all its financial assets.

How comparable are non-GAAP measures?

All non-GAAP measures are company specific. While they should be reasonably comparable over time for a single company, there is no requirement for any consistency between companies. The main problem is exactly how companies estimate the expected return, and the assumptions made.

For debt securities, expected return is likely to be the current yield less an allowance for estimated expected credit losses. For some debt securities where contractual cash flows are subject to change (such as sustainability bonds) further estimates may be required to allow for the potential cash flow effects. If some investment cash flows are not contractual, for example infrastructure or property, an estimate is required for the internal rate of return for the asset flows, which may include an expected disposal value. While equity investments by insurers are less common (unless backing participating contracts), they can still form a material part of the overall asset portfolio. For these, the expected return would be based on an assumed equity risk premium.

Lack of comparability is a potential problem for non-GAAP measures

Clearly, to different degrees, all the estimates required to derive an expected return are subjective and likely to differ between companies. Furthermore, the methodology may differ in other ways. For example, the expected return for a bond portfolio could be based on the yield to maturity or, alternatively, reflect a 1-year forward rate in the yield curve.

On the liability side the expected liability accretion is likely to be simply the current liability discount rate applied to the opening liability value. But again, the methodology may differ between companies, including the frequency of the reset and the use of the one-year rate from the yield curve or some other measure of accretion. Liability discount rates themselves differ between companies, something we discussed in our article ‘Insurance company profit and the illiquidity premium’.

 Aviva provides a relatively comprehensive explanation of how they determine their expected asset return. Nevertheless, investors wishing to compare companies will still likely struggle to identify whether Aviva’s methodology is more aggressive or more prudent than its peers.

Aviva expected asset return and liability accretion explanation

Aviva 2024 financial statements

Our view on non-GAAP for insurers

Despite the potential comparability problems, and the potential confusion due to partial measurement at cost, we think an expected return based disaggregation of the net financial result of insurance companies is useful for investors. We prefer this approach to the cost based disaggregation using OCI described below.

However, we also think that companies could go further with their disaggregation and related explanations. We would like to see more explanation and supporting data for the main drivers of the expected net financial result, such as the contribution of the expected spread between asset returns and insurance fulfilment value accretion. If this spread arises because of risk taking by the insurer, the value effect of that component of adjusted earnings may be very different from others, and potentially zero. Identifying this based on the disclosures we have seen thus far is challenging.

Cost based disaggregation: Segregate the volatility in OCI

IFRS 17 itself provides an alternative way to reduce earnings volatility – the option to present the effects of changes in financial inputs on the current value of insurance liabilities in OCI. This is combined with the use of OCI (and possibly cost accounting) for financial investments. The result of this is that cost accounting is largely applied in measuring the net financial result in profit and loss, with residual value changes recognised outside net income in OCI.

The OCI approach also presents challenges; not least because not all volatility ends up in OCI, particularly on the asset side where companies are forced to use fair value through profit and loss for some investments. Investors also need to be aware that there are different methods of determining what is reported in OCI and when, and if, it is recycled. Confusingly, there are two different OCI methods for the asset side and three for insurance liabilities.

Click on the title below to understand more about each of the asset and insurance liability OCI methods, and the extent to which they are consistent with cost accounting.

Understanding OCI for financial assets and insurance liabilities

The OCI option applied to insurance contracts adds significant complexity to insurance financial statements. The calculations themselves are inherently complex, and the OCI approach does not have to be applied consistently – companies can choose to apply it to only selected portfolios.

Another problem with the OCI method of reducing earnings volatility is that the application on the asset and liability side may not be consistent. Indeed, the OCI option can create accounting mismatches and potentially lead to earnings management.

OCI also creates asset-liability mismatches and earnings management opportunities

Most non-participating insurance contracts are backed by investments in bonds which, if the OCI option is chosen for the insurance liabilities, are likely to be measured at FVOCI with a cost-based measurement applied in profit and loss. If these assets are sold, any fair value gains or losses that have accumulated in OCI are recycled to profit and loss. However, there is no equivalent recycling applied to the accumulated OCI balance in respect of the related insurance liability. Although bond assets and insurance liability may be economically matched, an accounting mismatch arises if the assets are sold, even if the proceeds are reinvested in similar matching assets. Not only does this one-sided recycling distort the profit and loss result in the period of sale, it also creates an opposite effect in subsequent periods.

Sales of FVOCI debt instruments may partly undo the volatility mitigation from using OCI in the first place. In addition, management can potentially structure the timing of asset sales to manage earnings. The problem for investors is that it is usually impossible to identify the impact of these mismatches, with the insurance net financial result (and therefore insurance earnings) difficult to interpret.

French insurer Axa is a good example of a company that uses the OCI option in IFRS 17 and measures many of its financial assets at FVOCI. The note below is the analysis of the financial income and expense showing the different components and which are reported in profit and loss versus OCI.

Axa insurance net financial result disclosure

Axa financial statements 2024

We think it is very challenging for investors to understand and analyse the above note. There is nothing wrong with what Axa presents, it is just that the accounting is inherently complex and made even more so by the (partial) disaggregation between profit and loss and OCI. IFRS 17 requires that companies provide an explanation of the net financial result and the relationship between the income and expense components, but even with this we think investors will struggle with this component of earnings.

In the table below we summarise the amounts reported by Axa in profit and loss and OCI. The use of OCI does achieve a much less volatile net financial result in profit and loss. Indeed, it is less volatile than that for Aviva because much of it comprises interest income and expense at locked-in rates that only change gradually as market rates fluctuate.

Axa net financial result OCI based disaggregation

Axa financial statements 2024 and 2023

It is also worth noting that Axa presents adjusted performance metrics – we show the two adjustments related to the net financial result above. The first is the effect we mentioned about a mismatch arising when FVOCI assets are sold and the accumulated OCI amount recycled. The second is the removal of fair value changes because some assets must still be measured at FVPL.

Our view on the use of OCI by insurers

We think an OCI based disaggregation is more confusing for investors, produces less useful measures of performance, and has less predictive value when compared with the expected return based disaggregation provided by Aviva. We also observe that OCI does not remove the need for non-GAAP adjustments to deal with volatility that cannot be allocated to OCI and that consequently a non-GAAP approach is likely to be applied as well.

Application of non-GAAP and OCI

The lack of comparability caused by companies adopting differing policies regarding the use of OCI and non-GAAP is partially mitigated by a strong correlation in approach by jurisdiction. The chart below is taken from a publication by KPMG.

Use of OCI by jurisdiction

KPMG publication: Insurers’ 2024 annual financial statements – Real-time IFRS 17.

*IFIE stands for ‘insurance financial income and expense’ or the insurance component of the net financial result.

Notice that, like Aviva, the vast majority of UK companies covered in the KPMG survey do not apply the OCI option, whereas it is applied by 100% of French insurers, including Axa. The reason that less than 100% of French companies apply OCI to most assets is because, for some, the majority of assets back contracts accounted for under the variable fee approach where OCI is less relevant.

The jurisdictional split between fair value and cost is largely due to the history of insurance accounting and the types of insurance products issued, including the type of profit sharing used in participating contracts. We hope that one day investors will benefit from greater comparability and a shift away from the complexities of OCI.

 Impact on equity valuation

The challenges in analysing the net financial result of insurance companies, including the lack of comparability resulting from the different approaches used by companies, raises the question of how this contribution to earnings should be reflected in equity valuation. You could focus on forecast earnings inclusive of a forecast net financial result. However, in our view it is better to focus on the balance sheet.

Focus on the balance sheet (fair) value of the embedded leveraged investment fund

Insurance companies are, in effect, an operating business – the selling of insurance risk coverage – plus a leveraged investment fund. Focusing on earnings that combines these two elements into a single performance metric, to which a multiple, or some form of residual income or DCF valuation is applied, may not be the best way to estimate value.

There is no reason why the two components of insurance companies cannot be valued separately with a focus on the balance sheet value of the outstanding insurance obligation (excluding any unearned insurance profit) and the fair value of financial investments. There are several precedents for this in equity analysis:

  • Investment funds: For closed end investment funds that do not have an attached operating business, investors will focus on the fair value of the underlying assets for valuing the fund, with any debt liabilities measured at fair value. While the historical change in fair value and the income produced by the assets is of interest to investors in terms of performance, it would rarely be directly relevant for valuation.
  • Pension liabilities: An operating business with an attached defined benefit pension fund would usually be evaluated by separating the pension from the rest of the business and including the balance sheet surplus or deficit (possibly with adjustment) in the overall valuation.
  • Real-estate companies: A property investment company is usually valued based on the fair value of the property portfolio, potentially with separate consideration for property development activities, but again the main focus is the balance sheet.

Understanding the composition of earnings and adjusted earnings is important, but the contribution of investment income and insurance liability interest expense may be better viewed from a balance sheet perspective. Insurance companies would be valued at net asset value (with adjustments for deferred insurance profit and assets valued at amortised cost), plus a separate valuation of the operating result from the provision of insurance services.

Insights for investors

  • Most insurance companies mitigate profit volatility caused by financial market effects by applying the OCI option in IFRS 17, and measuring assets at FVOCI or amortised cost where possible, or by presenting adjusted non-GAAP metrics.
  • Insurance adjusted non-GAAP profit is generally based on an expected asset return and insurance liability accretion that reflects current interest rates. Interest rate changes still have a current period effect, but changes in value are excluded from earnings.
  • The application of the OCI option in IFRS 17 produces a form of (partial) cost-based earnings. However, the accounting is complex, with three different calculations, depending on the type of insurance contract, and two different applications of OCI for (some of) the related financial assets.
  • The OCI approach may not eliminate all value effects caused by financial market changes; residual amounts are potentially dealt with by non-GAAP adjustments.
  • The OCI approach may also create accounting mismatches such as when FVOCI debt instruments are sold. Watch out for potential earnings management due to selective asset sales.
  • For the valuation of insurance companies consider whether it is better to focus on the underlying financial assets and outstanding insurance liabilities rather than their respective contribution to (adjusted) earnings.

Don’t miss new articles published by The Footnotes Analyst – subscribe here:

none
Loading

No advertising, no spam, just new articles. Unsubscribe anytime.

Send us a question or comment about this article:

    For related articles select subject tags or see the suggestions below:

    Print or save article as a pdf: