Pension leverage under IFRS and US GAAP – Delta Air Lines

US GAAP and IFRS present the effects of pension leverage differently in financial statements, notably leverage arising from pension fund asset allocation. This complicates the comparison and interpretation of performance measures and valuation multiples.

We use Delta Air Lines to illustrate the positive impact of the US GAAP ‘expected return’ approach on reported profit, including the effect of optimistic return assumptions. If Delta had applied the IFRS ‘net interest’ approach we estimate that a ‘gain’ of $594m would have been excluded from profit and loss and instead reported in OCI.


In our recent article DCF and pensions: Enterprise or equity cash flow we examined the financial and asset allocation leverage created by defined benefit pension schemes1The accounting and analytical challenges (and the IFRS and US GAAP differences) for pension liabilities only apply to defined benefit schemes. For defined contribution plans there is no associated pension leverage, no liability measurement problems or any difficulties with profit and loss presentation. and the implications for cash flow and cost of capital. We illustrated how to deal with these leverage effects consistently in a DCF valuation by providing an interactive model.

Pension leverage also impacts profit measures and hence valuation multiples. The problem for investors is identifying whether this leverage is operating or financing in nature and the amount of pension related expense (or income) included in alternative measures of profit. It is important to make sure that these profit measures are consistent with other valuation inputs, such as enterprise value and the required return. However, this process is complicated by the difference in pension accounting between IFRS and US GAAP.

Delta Air Lines – Pension leverage

Delta has a significant accumulated pensions obligation that has a material impact on the analysis and valuation of the stock. In the most recent annual financial statements, the company reports a gross pension liability of $21.2bn and a further post-employment benefit liability (primarily healthcare costs) of $3.4bn. The combined deficit for these plans is $8.1bn. This is clearly a material matter for investors, considering the current market capitalisation is around $22.3bn.

Delta Air Lines post-employment liabilities and fund assets
Source: Delta Air Lines 2019 financial statements

In the table above you will also notice that Delta reports a zero current service cost for pensions. This indicates that their schemes are closed to further benefit accrual, but this does not mean that the liability is unimportant, or that it has a zero cost. Furthermore, the schemes have important cash flow consequences, as illustrated by the most recent aggregate “employer contributions” to the funds of $1.2bn.

However, it is the pension leverage effects that we focus on here. As we explained in our earlier article, pensions funds create two forms of additional leverage for equity investors.

  • Financial leverage: The pension deficit is, in effect, a debt-like obligation, given that the scheme has a prior charge on company profits. Pension liabilities should be included with other debt in assessing financial leverage.
  • Asset allocation leverage: If pension assets are not invested to produce asset-liability matching (essentially investing in bonds with the same duration as the liability), then investment risk arises. This is mostly borne by the company shareholders and results in a higher level of equity risk.

Both leverage effects are reflected in some of the liability and asset changes reported in the table above.

  • Interest cost: The accretion of the liability due to the unwind of the discounting implicit in the liability measurement.
  • Actuarial loss (gain) on pension liabilities: The change in the liability due to discount rate changes and changes in the cash flow estimates due to, for example, updated longevity assumptions.
  • Actual gain (loss) on plan assets: The change in fair value of the assets held in the pension and other post-employment benefit funds.2The term actuarial gain and losses is also often applied to pension assets as well as liabilities. The actuarial gain or loss for pension assets is the difference between the actual return on plan assets and the accretion amount we describe below.

Unfortunately, these amounts do not directly tell you what the pension leverage is. Part of the interest cost reflects the financial leverage effect arising from the pension deficit. The combined effect of pension remeasurements (comprising the actuarial gains and losses on pension liabilities and the actual gains and losses on pension assets) is, in part, the result of asset allocation leverage, although in this case combined with other pension risks such as longevity risk.

The effect of financial leverage reflects the size of the deficit. Asset allocation leverage is more difficult to interpret because it depends on the underlying investments of the pension funds. For this we need to consider the fund asset disclosures.

Delta Air Lines pension asset allocation disclosures
Source: Delta Air Lines 2019 financial statements

For Delta the pension asset allocation disclosures are not sufficient to accurately assess investment risk, particularly considering the lack of information about the nature of the large investment in hedge funds3If necessary, it is usually possible to find out more about fund asset allocation by examining the financial statements of the funds themselves. We have not done so in this case.. Nor is it sufficient to assess the extent of asset liability matching and interest rate risk, considering the lack of duration information for the bond investments. However, because of the significant investment in equities, it is clear that Delta shareholders are, in effect, invested in a leveraged investment fund in addition to their exposure to the underlying operating business.

We estimate that Delta’s pension leverage increases underlying risk by about 45%

Pension leverage results in a higher volatility (and beta factor) for equity investors compared with the underlying asset risk of the operating business. The impact of this pension leverage depends on the overall capital structure and how risk is shared with other providers of capital. Applying the pension risk model that we used to illustrate how pensions should be included in DCF, and ignoring the effect of other debt finance, we estimate that the overall leverage multiplier due to pensions is presently about 1.45x.

This overall pension leverage effect is the product of our estimated asset allocation risk factor of 1.21x and the financial leverage factor due to the pension deficit of 1.20x (the overall effect is the product of the two). In other words, if the asset beta were 1.0x then (ignoring financial leverage due to debt and lease finance) the equity beta would be 1.45x.4An interesting issue related to pension liabilities and the associated leverage is whether the market fully recognises this risk factor in the pricing of equity securities. The historical beta factor for Delta is about 1.3x, which seems low considering both the financial and pension leverage that affects equity investors. The historical beta factor reflects historical leverage which would be considerably lower than that today, given the fall in stock price and Covid-19 related increase in net debt. However, even after allowing for this, we estimate that the implied asset beta for Delta is only about 0.5x. Of course, observed beta factors need to be treated with caution given the statistical margin of error.

In addition to the analysis of pension risk for the purpose of assessing cost of capital and cash flow (as we did in our previous article) it is important to be aware of how this leverage impacts performance metrics. But for that we need to understand pensions accounting, including the difference between IFRS and US GAAP.

IFRS and US GAAP pension accounting

In some respects, pension accounting under IFRS and US GAAP is very similar. Both measure the defined benefit liability based on a discount rate derived from the yield on high-quality bonds; both include the service cost and related expenses, such as past service costs and the cost of plan settlements, as an operating expense; and both report actuarial gains and losses5Actuarial gains and losses are the remeasurements of the scheme assets and liabilities (other than those reported in profit and loss, such as the accretion amounts described below). This includes the fair value changes for scheme assets and the effects of discount rate changes and changes to the assumptions, such as longevity, that determine the estimated pension payments. in other comprehensive income (OCI) rather than in profit and loss.

The overall expense for pensions that is reported in total comprehensive income6Comprehensive income is the sum of the profit or loss (net income) for the period and the gains and losses reported in other comprehensive income (OCI). is actually very similar under IFRS and US GAAP because the balance sheet amounts are measured in a similar way. But there are important differences in the disaggregation and presentation of the expense. This leads to differences in key performance metrics.

The two most important differences for investors are, firstly, how the interest accretion of pension liabilities and expected return on pension assets are recognised and, secondly, what amount is reported in OCI and the potential impact this has on future profit.

Interest accretion

Pension liabilities are the present value of the expected future payments to retired employees. If a scheme is funded, then the fund assets are also a present value, although in this case the discounting is implicit in the market price rather than a predetermined input. Therefore changes in the measurement of both assets and liabilities during a period include, among other things, the explicit or implicit effects of a discount unwind.

Both IFRS and US GAAP report an accretion amount in profit and loss and separate this from other components of the asset and liability changes which are reported in OCI. The problem for investors is that these accretion amounts differ.

  • IFRS – Net interest approach: Under IFRS, the accretion amount reported in profit and loss only applies to the net deficit (or surplus). It is simply the amount of the deficit multiplied by the liability discount rate. Although the result is the same, you will often find that the explanation and presentation is slightly different where interest is shown for the full liability and this is offset by an interest accretion (at the same liability rate) for the assets.
  • US GAAP – Expected asset return approach: For US GAAP, accretion works separately, and at different rates, for the gross pension liability and for the gross fund assets. The amount reported in profit and loss is the expected rate of return on the asset portfolio applied to the gross asset balance less the discount unwind for the gross liability.

Given the size of the pension schemes for some companies, the effect of this accounting difference can be significant.

Delta Air Lines net pension cost or benefit in profit and loss
Source: Delta Air Lines 2019 financial statements

IFRS net interest approach results in a higher expense than US GAAP

The IFRS net interest approach almost always produces a higher accretion expense. In effect, the amount of expense reported in profit and loss under US GAAP equals the IFRS amount less the expected spread derived from the funded portion of the liability. This assumes that the expected asset return is greater than the liability discount rate, which is invariably the case. Often this produces a net gain for US GAAP reporters, even in cases where the scheme is in deficit – as is the case for Delta.

The remaining changes to pension assets and liabilities arising from their remeasurement at a current value are also gains and losses in the period. These are included in OCI under both approaches, albeit a different amount due to the different accretion.

For Delta the expected return less interest accretion contributes $263m to pre-tax profits. Had the IFRS net interest approach been used this would have been an expense of about $331m, with a pre-tax ‘GAAP’ difference of $594m being the expected spread on the funded portion of the scheme.

Delta Air Lines: Pension interest accretion – IFRS versus US GAAP
Source: Delta Air Lines 2019 financial statements and The Footnotes Analyst estimates

In the US GAAP illustration, we have separated the accretion for the funded and the unfunded (deficit) portions of the net liability. This is merely to illustrate how the overall effect compares to that for IFRS. In practice, and in the extract from Delta’s financial statements above, the net accretion amount is presented as the accretion on the gross liability less the expected return on the pension assets. However, the overall effect is the same.

Our view on accretion

IFRS and US GAAP adopt a different approach to the reporting of pension financial and asset allocation leverage. Under IFRS these two leverage effects are reported in different parts of the performance statement, with the asset allocation impact fully reported in OCI. However, for US GAAP, the effect of asset allocation leverage is split, with the expected gain from investing in risky assets reported in profit and loss. Both are valid approaches to disaggregating the overall pension expense and both have their advantages and disadvantages.

Do not ignore asset allocation risk when interpreting US GAAP earnings

The US GAAP approach separates the benefit and risk of asset allocation leverage. The benefit, as represented by the expected asset return, is reported in profit and loss but the effects of asset allocation on risk is reported in OCI. There is a danger (considering the market’s unhealthy focus on profit and loss) that the gain is factored into investors’ analysis but not the associated risk that is captured in OCI.

Equally there is an issue with the IFRS approach in that the expected return from the asset allocation spread is put in OCI. For most items in OCI the forecast period amount would be zero. However, for pensions there is an expected OCI gain if risk is taken regarding pension assets. There is a danger that investors will ignore this expected gain but take account of the additional risk, such as in setting cost of capital, thereby understating value. The solution is to either forecast comprehensive income rather than earnings, or to ensure that asset allocation risk is also excluded from cost of capital measures and target valuation multiples.

The US GAAP approach is also highly dependent on companies using a realistic measure of the expected return on assets. The IFRS interest amount is much less susceptible to variation and exaggeration.

Delta Air lines – Expected return assumption and its impact

The issue of high expected return assumptions is well illustrated by Delta. Their 8.97% overall expected return for 2019 is one of the highest we have seen. This may be partly explained by a higher than average weighting for equities and hedge funds but, in our view, it does not entirely explain it. Based on our estimates of the overall equity exposure, and assuming that the bond investments are not risk free, we estimate that the assumed return implies a forward equity risk premium (ERP) of about 9%.

A further contribution to Delta’s high expected return (and included in our implied ERP estimate) comes from the company assuming “modest excess return expectations”; in other words that the pension portfolio managers will beat relevant indices. This is certainly plausible but obviously not guaranteed. Including this assumption in calculating the expected return hardly seems prudent.

Delta Airlines expected return assumptions
Source: Delta Air Lines 2019 financial statements

The effect of a high expected asset return assumption is to increase profit and reduce OCI in the current period (although in subsequent periods this may be reversed due to the ‘recycling’ adjustment we explain below).

‘Recycling’ of OCI

The second key difference between pension presentation in IFRS and US GAAP is whether the amounts reported in OCI are subsequently reclassified (or recycled) to profit and loss. Recycling a gain reported in OCI in a previous period simply involves reporting a loss in OCI in the current period and a corresponding gain in profit and loss. Recycling may take place in one accounting period in response to a triggering event or could be an allocation over several periods (which is the case for pensions).

There are differing views regarding the merits of recycling in financial reporting. Some argue that profit and loss measures ‘real’ performance and that gains reported in OCI have not actually been recognised as a profit. Accordingly, some basis must be found so that gains temporarily ‘parked’ in OCI are subsequently reported as part of profit or loss – hence recycling.

However, others regard OCI as merely a section in a larger statement reporting all gains and losses recognised in a period. In this approach profit and loss is, in effect, a subtotal in a more complete performance measure called total comprehensive income (which is the sum of profit and loss and other comprehensive income). With this more expanded view of performance the idea of recycling seems odd as the recycled amounts reported in profit and loss and OCI offset each other.

Disagreement over what OCI represents leads to differences in recycling

A long-standing problem in financial reporting is that there has never been agreement about what OCI represents, what it should be used for, and whether it is part of performance or merely a holding place for ‘quasi’ profits before they become ‘real’ profits in some future period. This is evident in different companies’ financial statements. Some companies clearly label OCI as a component of performance, even to the point of including it on the same page as profit and loss. However, other companies refer to recognising something in OCI as reporting it in equity and not as a gain or loss at all. With this view, the desire for recycling becomes more understandable.

While there is a lot of similarity regarding the use of OCI under IFRS and US GAAP, there are also important differences, one of which relates to pensions.

  • Pension recycling in IFRS: There is none. When pension actuarial gains and losses are reported in OCI, IFRS considers them recognised and that no subsequent adjustments are required.
  • Pension recycling in US GAAP: The pension gains and losses initially reported in OCI are recycled to profit and loss. They are allocated to profit and loss over the expected remaining service lives of employees, although only to the extent that the cumulative absolute amount in OCI is greater than 10% of the gross pension obligation (or gross assets if higher). If the accumulated gain or loss is within that ‘corridor’ then no recycling occurs.

For Delta the cumulative amount of pension losses previously reported in OCI (less that amount that has already been recycled) is $8,765m. The recycled loss (a gain in OCI) in the most recent period is $291m, which is included in the overall net pension expense in profit and loss. You should not include this recycled amount in any metrics used for analysis or valuation.

Our view on recycling of pension related OCI

We think recycled pension gains and losses are meaningless. If a pension fund deficit increases in a particular period due to, for example, low returns on pension assets then we believe that this is a loss that should be reported in that same period and only in that period. Recognising the loss in OCI rather than profit and loss may have some merit in that it separates a volatile and quite different gain or loss from other profit and loss items, but, in our view, the act of recycling adds nothing except causing confusion.7We are not saying that all recycling adjustments are meaningless. In some cases recycling is necessary, such as to achieve the objective of cash flow hedge accounting.

So how would Delta Air Lines look under IFRS?

We have recalculated the Delta pension related amounts using an IFRS approach, taking into account both the difference in measuring the accretion amount and the lack of recycling of gains and losses reported in OCI.

We estimate that the overall pension expense measured under IFRS would be $410m compared with $144m under US GAAP. The difference of $266m comprises two offsetting amounts, a higher interest expense under IFRS of $594m and the absence of the recycled actuarial loss of $328m.

Delta Air Lines pension expense – IFRS versus US GAAP
Source: Delta Air Lines 2019 financial statements and The Footnotes Analyst estimates. We have assumed that the healthcare service cost, the pension settlement expense and prior service credits are the same under IFRS and US GAAP. This may not be the case in practice, although generally any difference would be small in comparison with the accretion and OCI recycling differences we highlight.

Performance measures and valuation multiples

In our previous article we explained that, if pension leverage effects are dealt with consistently in free cash flow, the discount rate and the enterprise value bridge adjustment, then a correct DCF value can be obtained. However, our preference is an enterprise cash flow approach with the exclusion of asset allocation leverage from both the discount rate and free cash flow.

Consistency is also important for valuation multiples, both consistency between IFRS and US GAAP reporters (if you are making such comparisons) and consistency in how multiple differences are interpreted, considering other value drivers.

  • Consistency in accounting: Do not compare price earnings ratios for IFRS and US GAAP reporters if pensions are significant without making adjustments to earnings to ensure comparability. The easiest way to do this to convert US GAAP to IFRS, as we illustrate above.
  • Consistency in analysis: If the effect of asset allocation leverage (the asset return / liability accretion spread) is included in performance measures then make sure that the related risk is included in the interpretation of the resulting multiples. Higher asset allocation risk and gains means a lower deserved multiple. 

In our view the best way to achieve consistency on both counts is to apply an enterprise value approach to your analysis and to ensure that all pension leverage effects are excluded from performance measures. This means focusing on profit after deducting only the service cost, thereby avoiding the GAAP difference regarding the accretion expense and OCI recycling. It also means using an enterprise value that includes the net pension deficit or surplus. There may still be comparability issues to look out for, such as differences in discount rate, but these are much less likely to cause problems. 

Insights for investors

  • Pension accounting under IFRS and US GAAP differs. This can affect the comparability of key profit metrics of companies with material defined benefit pension liabilities.
  • The pension accretion amount under US GAAP includes the expected gain from asset allocation leverage. Under IFRS, all financial effects of asset allocation risks are reported in OCI.
  • The net interest accretion reported under IFRS results in lower profit compared with US GAAP.
  • The pension recycling adjustment in US GAAP is meaningless – exclude it from all performance metrics.
  • If global comparability is needed, and pensions are material, we suggest restating US GAAP to IFRS – restate the accretion gain or loss to the IFRS net interest expense and exclude the recycling gain or loss.
  • An enterprise value approach is the best way to ensure that pension leverage is consistently dealt with in performance metrics, valuation multiples and other components of equity valuation, including cost of capital.

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 more articles select subject tags:


    Or see suggested related articles below: