Insurers’ losses shine light on swaps accounting

An overhaul of US accounting rules promises to iron out wrinkles in the treatment of insurers’ long-duration contracts and accompanying hedges – but the amendments themselves are not without fault, says the industry

The world is getting used to “alternative facts”; at first blush the trend even seems to be catching on in corporate earnings calls.

While the sea of red in US insurers’ annual reports made for uncomfortable reading in February, their respective chiefs painted a rosier picture than the numbers seemed to imply in their accompanying investor communications.

MetLife, for instance, reported a fourth-quarter 2016 net income loss of $2.1 billion. But on a February 2 analyst call, chief executive Steve Kandarian insisted 94% of this was the result of “asymmetrical insurance accounting” and “a non-economic movement” in assets relative to liabilities.

On February 9, Prudential Financial disclosed pre-tax net realised investment losses of $824 million. Prudential chief John Strangfield said the “vast majority” of these were driven by a component of the firm’s variable annuity (VA) reserve it did not consider to be economic.

The pattern repeats elsewhere. Lincoln Financial Group described more than half of its annual $178 million net loss on its VA derivatives portfolio as “a non-economic charge”, while Manulife blamed $330 million of losses on “an accounting mismatch”.

Insurance chiefs understandably want to be clear about the real drivers of profit and loss at their respective firms. However when it comes to their life businesses, they are stymied by the very rules that govern financial statements: US Generally Accepted Accounting Principles (Gaap).

This is the standard adopted by the Securities and Exchange Commission for public financial disclosures. Yet the rules are a poor fit for the insurance industry. Not only do they measure long-duration contract liabilities and the assets used to hedge them differently, injecting volatility into both earnings and shareholders’ equity, but the benefits embedded in popular VAs are valued using two different methodologies, confusing insurers’ income statements and leaking “non-economic” charges into their results.

This makes insurers’ financial statements tricky to decode and, because certain regulatory capital requirements take their cues from Gaap statements, complicates statutory reserving. In some cases it also deters firms from using interest rate derivatives for asset-liability management.

A long-running project by the Financial Accounting Standards Board (FASB), custodians of US Gaap, seeks to remedy these issues. An exposure draft of its Targeted improvements to the accounting for long-duration contracts was published in September 2016 and the board has scheduled a number of roundtable discussions this March to address the industry’s concerns on the way to drafting a final amending text.

Insurers have high hopes for the project, but are wary that a flawed amendment could cause more problems than it solves.

“We are all aligned that we don’t want to have misleading financial results that are not appropriately transparent. But despite some improvements being made here, there are some changes FASB don’t fully get right, and we’ll still have some non-economic volatility,” says the accounting chief at one large US insurer.

Asymmetric warfare

FASB’s proposed improvements take aim at the “asymmetrical” and “non-economic” accounting bemoaned by MetLife’s Kandarian and his peers. At present, long-duration insurance liabilities, including term, whole and universal life policies, as well as certain universal disability and long-term care contracts, are accounted for on an accrual basis.

The contracts are valued using a net premium ratio, which takes the present value of future policy benefits plus maintenance expenses, divided by the present value of future premiums. The cashflow assumptions, including the discount rate used to collapse future obligations into a present day lump sum, are ‘locked in’ on the contract inception date. Essentially, their present values don’t budge in response to market moves.

In contrast, the assets held against these liabilities or used to hedge them are subject to an array of different accounting treatments. Cash instruments, such as bonds, held as available-for-sale, have their valuation changes recorded in other comprehensive income (OCI), a portion of the balance sheet where unrealised gains and losses are reported. Their fluctuations do not hit the net income statement, though in a rising rate environment their fall in value would generate an unrealised loss in OCI that would not be balanced by a corresponding fall in the present value of the liabilities. OCI is a component of shareholders’ equity, so gains and losses here impact on a firm’s net value.

Many of the market risk benefits [of VA liabilities] are not fair-valued right now so if you try to hedge the market component of them, you will have a mismatch

Donald Doran, PwC

Meanwhile, interest rate derivatives, typically used to hedge liability duration, are marked at fair value. Mark-to-market movements filter directly through net income instead of OCI, while the liabilities they hedge remain static.

Evan Bogardus, insurance accounting change leader at EY in New York, explains: “Let’s say interest rates decrease. The value of an insurer’s bonds portfolio will increase on paper, though from an economic perspective they will be worried about reinvestment risk. An insurer may go out to hedge that risk with derivatives. That makes economic sense, but when interest rates spike up, as they did somewhat in the fourth quarter of 2016, the value of those hedges will decrease. Rising interest rates decrease bond values – recognised in OCI as available-for-sale debt securities – while traditional life liabilities are not typically impacted, resulting in interest rate volatility in OCI; but then the derivatives are losing value in the actual P&L.”

The 85 basis point jump in the US 10-year Treasury yield following the November presidential election was blamed by MetLife for triggering its Gaap loss, precisely because of this mismatch between long-term liabilities and its hedge portfolio of interest rate derivatives.

John Hele, chief financial officer, said on the company’s earnings call: “We’re better off even with these hedges from an economic balance sheet point of view but you have this noise in the Gaap – so how much do you want to spend money and change hedging to protect Gaap? We are examining various options.”

Then there’s the Gaap treatment of VA liabilities. These typically offer policyholder benefits collectively known as market risk benefits. Right now, some of these benefits are recorded in an insurer’s accounts as derivatives and marked-to-market accordingly. Others are booked on an accrual basis – including guaranteed income and death benefits, as well as some withdrawal benefits.

“Many of these market risk benefits are not fair-valued right now so if you try to hedge the market component of them, you will have a mismatch because the hedging derivatives are going up and down through fair value yet the liability is not,” explains Donald Doran, US national office leader for financial services in the accounting consulting group at PwC in New York.

On the flipside, those VA benefit liabilities that are booked at fair value create accounting noise all of their own thanks to so-called ‘non-performance risk’ (NPR). This is an accounting adjustment that reflects the risk an insurer will not fulfil its obligations to policyholders if it defaults. Somewhat counterintuitively, NPR causes an insurer’s VA benefit liabilities to increase as the company’s credit spread tightens, and shrink it as it widens. In derivatives parlance, this effect is known debit valuation adjustment. Prudential Financial pinned a $1.3 billion investment loss for 2016 on an adverse NPR move which flowed directly into earnings.

In summary, US Gaap instils non-economic volatility into shareholders’ equity and net income alike. These swings “can deter some people from investing”, says the accounting head at a second large US insurer, and means firms have to “spend a lot of time” educating potential investors on their book value. It also incentivises US carriers to publish a variety of non-Gaap measures, further complicating the picture for market participants.

“The industry presents alternative metrics to get rid of what some term ‘accounting noise’. Insurance is a particularly difficult industry to understand and based on only a set of US Gaap financial statements an investor may be hard pressed to figure out if a business will be profitable in the long term,” says EY’s Bogardus.

It is also a factor driving insurers away from synthetic instruments. The first accounting chief confirmed there has been a migration to cash instruments for asset-liability management over recent years, incentivised in part because of the illogical accounting treatment applied to derivatives (see box: Hedge accounting restrictions).

Long-distance target

Small wonder, then, that FASB has been working on an overhaul. What is surprising is just how long it has taken to kick into gear. The Targeted improvements can trace their origins all the way back to 2007, with the publication of a FASB proposal to develop a common standard for insurance contracts.

In 2008 FASB elected to work on the project in partnership with the International Accounting Standards Board (IASB), but the joint venture ended when the broader convergence project between the two accounting standard-setters fell apart.

FASB chose to move forward on its own, in 2013 issuing a proposed Accounting Standards Update, which, following industry feedback, resulted in the exposure draft unveiled last year. The latest proposal attracted 39 comment letters from insurers, accounting firms, industry associations and banks, the majority of which applauded FASB’s efforts while warning that certain flaws in the draft would hamper its professed objective of creating more effective financial disclosures.

“There are a number of issues that we have with FASB’s proposal... we think it needs some fine-tuning in order to get it quite right,” said Prudential’s Strangfield on the insurer’s February 9 earnings call.

Three of the core aims of the exposure draft are designed to fix the accounting mismatches described above. Firstly, FASB wants to “unlock” liability cashflow assumptions which – if properly implemented – could eliminate the equity volatility caused by the mismatch of available-for-sale assets and long-duration liabilities.

Secondly, it wants to standardise accounting of embedded derivatives in variable contracts – to remove the existing disincentive to hedge capital market risk exposures arising from those market risk benefits measured on an accrual basis – and shunt NPR-related movements out of net income and into OCI.

Thirdly, it wants to “improve the effectiveness of the required disclosures” and ensure financial statements provide “decision-useful” information.

The method of unlocking long-term liabilities attracted a great deal of industry comment. FASB’s idea is to unlock both the cashflow and discount rate assumptions underpinning contract valuations each quarter.

The discount rate assumptions correspond to the investment yield used to translate future policy benefits into a dollar amount today. The cashflow assumptions reflect non-economic effects such as mortality, morbidity and longevity experience. As the discount rate typically has the greatest impact on liability valuations, it is this change that has attracted stakeholders’ attention.

FASB proposes the discount rate be updated using a “high-quality fixed-income instrument yield”. This has been interpreted by commentators to mean an AA-rated bond portfolio – which not everyone is happy about.

FASB is endeavouring to tether itself to an AA discount rate because it would be transparent, it would be consistent, but it wouldn’t match up neatly with the assets

Accounting head at a large US insurer

A number of comment letters argue the actual investment yield of the asset portfolio held against the liabilities should be used instead. Four insurers – Prudential Financial, MetLife, New York Life, and Manulife – produced a joint report informed by a self-directed field test that laid out the issues.

Their field test concluded that an AA-yield “does not provide an adequate illiquidity premium for non-participating contracts” and showed that in choppy markets the spreads on AA instruments would diverge dramatically from those of the actual instruments used to back the liabilities.

This in turn would “lead to a large decrease in US Gaap equity or even negative equity” and send “a false signal” to investors and regulators who may use US Gaap financial statements as a measure of insurance company solvency, the insurers wrote.

They instead recommended the use of a single A-rated yield curve, which they say is closer to the typical investment portfolio used to back non-participating contracts.

“When we price a liability we focus on using a discount rate that aligns nicely and neatly with the investment portfolio that is going to be used to defease those liabilities. FASB is endeavouring to tether itself to an AA discount rate because it would be transparent, it would be consistent, but it wouldn’t match up neatly with the assets. So you are still going to have a fair amount of breakage that will reside in equity,” says the first accounting head.

If an inappropriate discount rate is selected insurers could be deterred from underwriting certain businesses, they say.

The choice of discount rate is one problem; the application of the adjustment to the net premium ratio is another. FASB endorses a retrospective approach where the updated cashflow assumptions are calculated and applied as though they were in effect at the time of each contract’s inception. This is to help determine an appropriate “catch-up adjustment” for the present reporting period and avoid a huge swing in liability valuations in the first set of accounts following the adoption of the new rules.

But stakeholders say this would load unmanageable operational burdens on insurers. Primerica, a top-tier insurer, wrote in its comment letter to FASB that collating contract cashflows from years – if not decades – in the past would be “extremely data intensive and require a level of information technology expertise and capacity that many insurance companies could not maintain in house”.

The challenges would be acute for small and medium-sized entities in particular. “In principle they agree that unlocking assumptions makes a lot of sense. It’s not so much the theory behind doing it, it’s having the systems and personnel to do it. At some of these [small insurers] the information you have is very sketchy – some of it is literally on pieces of paper. For some closed blocks of business it would be near impossible to go back and do a retrospective adoption,” says Ken Hugendubler, Pennsylvania-based insurance industry practice leader at Baker Tilly, an accounting and advisory firm.

Most comment letters to FASB instead favour a prospective application which they say would be simpler to apply and cheaper to implement. “With the prospective approach, an insurer says, ‘I’ll not change today’s balance sheet but start factoring in the updated assumptions going forward so eventually we’ll end up with a net premium ratio accurately reflecting these assumptions’,” explains PWC’s Doran.

Not fair for all

While the unlocking proposals have attracted the most heat, FASB’s mission to account all market risk benefits embedded in VAs at fair value has also run into opposition.

“Here’s a case where you have a significant volume of activity happening and at least three different methods being used for the accounting, so FASB is appropriately bringing consistency to the picture. It’s trying to accomplish that by putting everybody on to one model. We’re happy with this outcome other than for GMDBs [guaranteed minimum death benefits]. This is one benefit that we don’t think fits with our mark-to-market model and everyone seems to agree with that,” says the accounting head at the first insurer.

A GMDB, the argument goes, does not resemble a pure-play market risk benefit as it is triggered by an insured event: the policyholder’s death. As such the benefit should be recognised as an insurance risk, rather than a market risk, and carved out of the fair value treatment. The four life insurers that conducted the field test argue as much in their comment letter. Ironically, considering the purpose of the update in the first place, fair value treatment of GMDBs could add unjustified noise to the income statement.

“Most companies do not hedge their GMDBs and so putting them in fair value would cause volatility as insurers would have a liability being fair-valued and wouldn’t have anything to offset it. Many companies have made the decision that mortality risk is something they understand and do not choose to hedge given the costs of maintaining derivatives positions,” says Doran.

If you are not hedging today, it’s complicated and it can be expensive, so you have to really want to do it

Evan Bogardus, EY

Elsewhere, the four insurers’ field test revealed that applying the market risk benefit changes resulted in “higher and more volatile segment reserves” for VAs than under the existing guidance – primarily due to the switch of guaranteed minimum income benefits to the fair value treatment rather than death benefits.

This underlines how the rule change could spur a shift in hedging activity by life firms. Doran says there are certainly companies that have elected not to hedge market risk benefits in the past because it would create an accounting mismatch they would have to explain to investors. The second accounting chief says the rule change creates a “better symmetry” between VA liabilities and derivatives hedges – perhaps incentivising hedging.

EY’s Bogardus doubts the improvements will spark a wholesale change: “If you are not hedging today, it’s complicated and it can be expensive, so you have to really want to do it. If the new accounting is considered more reflective of economics, that’s something worth considering, but I wouldn’t call it the only determining factor,” he says.

Financial disclosures and hedging behaviour may not be the only aspects of the insurance business impacted by the accounting changes. They could also have repercussions for the largest US insurers’ capital adequacy. At the state level, statutory accounting for calculating regulatory capital is governed by state supervisors, who co-ordinate transpositions of US Gaap into their own rules under the auspices of the National Association of Insurance Commissioners. They can choose to adopt, adopt with modification, or reject FASB changes as they see fit.

However, Gaap does provide the bedrock for global capital standards due to be rolled out for internationally active US firms. These global players are expected to be subject to an insurance capital standard under development by the International Association of Insurance Supervisors (IAIS). The very largest – MetLife, Prudential Financial and AIG – will also be subject to an additional higher loss absorbency requirement as part of their designation as global systemically important insurers.

Both capital requirements are being honed with reference to Gaap valuations. Therefore how Gaap measures shareholders’ equity, recognised as prime regulatory capital, is of crucial importance.

“We don’t want to have misleading results that are not appropriately transparent to the users who count on them. Despite some improvements, you’ll still have some non-economic volatility [because of the exposure draft], mostly in equity. Among the concerns is that you could have regulatory regimes that climb on to a new standard like this and say, ‘This is how we should measure an insurance company’s capital’. We’re sitting here thinking: you’re using a figure that is not really representative of our capital,” says the accounting head at the first insurer.

Just as in the world of politics, the insurance industry may have to learn to live with the idea of alternative facts, despite FASB’s best efforts.

Hedge accounting restrictions

MetLife’s net loss for the fourth quarter of 2016 included $3.2 billion in net derivatives losses “reflecting changes in interest rates, foreign currencies and equity markets”. The insurer blamed this hit on the way in which gains and losses on derivatives are accounted for differently from the risks being hedged.

Obtaining hedge accounting treatment for these various derivatives positions would be the obvious cure to this earnings volatility, as it would permit a liability and its hedging asset to be reported as a single accounting entry, meaning opposing moves would be offset.

But Gaap conditions make this difficult, as the accounting chief at one large US insurer explains: “It’s a low percentage [of derivatives that qualify for hedge accounting]. The rules require a fair amount of specificity around the exposure you are hedging and by that I mean tagging specific derivatives to specific assets or liabilities as opposed to what we view as portfolio hedging, where we look at an entire block of liabilities or entire portfolio of assets. The hedge accounting rules don’t let you achieve hedge accounting at a portfolio level like that.”

He adds that 10–20% of the company’s derivatives qualify for hedge accounting. The commensurate figure for MetLife, disclosed in its earnings call, was 15%.

As FASB’s Targeted improvements are focused on the liability side of the balance sheet, the current mismatches attributable to portfolio hedging are not addressed. Curiously, nor is the issue discussed in a separate FASB accounting standards update proposal on hedge accounting.

The American Council of Life Insurers made its displeasure known in a November 2016 letter to the board in which it argued the narrow applicability of hedge accounting “causes earnings distortion” and results in “misleading financial statements”.