Overly processed accounting — who eats this stuff?! | FT Alphaville

Overly processed accounting — who eats this stuff?!

Heavily processed foods are generally unhealthy. Those looking after their waistline may wish to read ingredients lists and follow food journalist Michael Pollan’s advice: “Don’t eat anything that your great-great grandmother would not recognize as food.”

Could the same be said of financial accounts? That is: “Don’t trust anything your grandparents wouldn’t recognise as a justifiable line item.”

While several entries would no doubt fall under this rule of thumb, FT Alphaville has Debit, Credit, and Funding Valuation Adjustments (DVA, CVA, and FVA) in mind for this particular missive. They are the end product of a great deal of over-engineering. When these numbers pop out of internal bank models, they sometimes look like amorphous piles of goo, which makes us wonder: Who eats this stuff?

Equity analysts sure don’t — they keep it on the side of the plate, hardly bothering to take a bite and spit it out. Banks, in their press releases, must continuously repeat that that they are talking about earnings without these numbers. Citi, for example, when reporting third quarter earnings this year:

Citigroup revenues of $19.4 billion, excluding CVA/DVA and the loss on MSSB, were 3% above the prior year period, driven by 5% growth in Citicorp revenues, offset by a 10% decline in Citi Holdings revenues primarily resulting from the ongoing wind down of those assets. Citi Holdings revenues represented 5% of total Citigroup revenues, excluding CVA/DVA and the loss on MSSB.

Citicorp revenues of $17.6 billion in the third quarter 2012 included $(799) million of CVA/DVA reported within Securities and Banking. Excluding CVA/DVA, Citicorp revenues were $18.4 billion, 5% above the prior year period with 15% growth in Securities and Banking revenues and 2% growth in Global Consumer Banking (GCB) revenues, partially offset by a 2% decline in Transaction Services revenues.

Citi Holdings revenues of $(3.7) billion in the third quarter 2012 included $23 million of CVA/DVA and the $4.7 billion pre-tax loss on MSSB. Excluding CVA/DVA and the loss on MSSB, Citi Holdings revenues were $971 million compared to $1.1 billion in the prior year period.

A survey by Ernst & Young of 19 large dealing houses, out last week, further illustrated that banks use diverse and inconsistent methods to calculate and declare numbers like these; there is no standardised approach. This matters because of upcoming accounting changes that will begin applying at the start of 2013. These changes are likely to make bank earnings even more volatile because of these line items — and given the lack of transparency and uniformity in calculating these numbers, to what end?

The E&Y survey found that all respondents record a CVA on their derivatives assets, thereby making an allowance for amounts owed to them that might not be paid due to counterparty default. However, only 13 of those also record a DVA on their derivatives liabilities, effectively discounting the amounts they themselves owe due to their own possible non-performance.

The six holdouts would no doubt argue that DVA is hard to monetise and difficult to hedge. Indeed, not many try hedging DVA, while for CVA it’s more common:

The new reporting requirement, for those who use International Financial Reporting Standards, will force the hand of those banks who object to recording DVAs on their derivatives portfolios. Scheduled for introduction on January 1st, 2013, IFRS 13 contains the effective requirement to report a DVA on derivatives liabilities and to use market data while they’re at it.

Of course, it’s not just derivatives entries that attract this sort of own-credit-risk adjustment that contributes to earnings volatility. Liabilities in the form of own-debt (like bonds) get it too, as and when the “fair value option” is elected. This can be called “own credit adjustment” (OCA) to make it clear what is being discussed, which often doesn’t seem to be the goal of people talking about it, unfortunately.

This OCA versus DVA distinction matters, not least because the methodology can differ even though the underlying principle — that the bank itself may default — is broadly the same. And don’t get us started on how much the netting treatment and disclosure can differ from bank to bank. Concerning methodology, again from the E&Y survey:

Holy lack of internal consistency, Batman!

Maybe we should have showed you this next chart first anyway, as it demonstrates the variation of inputs used to calculate OCA alone:

All 19 E&Y survey respondents report OCA. The accounting firm further notes that:

The magnitude of the OCA reported by a certain number of banks has reinforced questions around the economic relevance of the adjustment.

Economically relevant or not, the volatility wrought by OCA will eventually be withdrawn from the income statement by IFRS 9, whose scheduled time of arrival is 2015. From that point onwards the entry will be written directly into equity, into Other Comprehensive Income. Colour us relieved…

And did we mention that Basel III won’t allow OCA and DVA to be included in core equity tier one capital? Yes, that’s right, even the gnomes of Basel are calling bullshit on these line items. It might have something to do with how the “value” of a bank’s accelerating deterioration in creditworthiness is as easy to capture as rain with a sieve. Put differently, an increasingly high probability of default doesn’t exactly shout “capital cushion”, does it?

CVA, the mirror image of DVA, probably does make sense to grandma on some level. Such adjustments are, after all, a bit like making an allowance for doubtful debts on a loan portfolio. CVA crops up for derivatives where the reporting bank is in-the-money and expecting their counterparty to make good on the contract. CVA is then like a discount taken to reflect the risk that the counterparty might not make good.

As for the gnomes of Basel, a special charge for the volatility of CVA will be applied under the new regulatory regime, i.e. the volatility of the counterparty risk measurement applied to the value of derivatives for which the reporting bank is in-the-money. Still with us?

Arriving then at so-called FVA, or Funding Valuation Adjustment, we’re pretty sure it’s just some quant’s wet dream. It seems a lot like DVA/OCA, only we think FVA is about trying to measure the component about funding oneself (as a bank) that isn’t to do with own credit risk. For example, there might be a liquidity component in there given the part of the bond or the money market that the bank is funding in.

It turns out that the quants arguing about FVA were shouting so loudly at each other that they woke up John Hull and Alan White — the work of whom many a young banker is meant to absorb (mostly by osmosis). The duo spent 1,742 words in the August edition of Risk articulating an intricate argument that we feel can (very) crudely be summarised as: “FVA? Are you smoking crack?”

We additionally note that there was a response to it by a couple of RBS quants that you really should read to get a picture of both sides of the argument.

As for ourselves, we do understand how FVA potentially makes sense from the perspective of internal management. Balance sheet costs money. Resources need to be appropriately allocated within an organisation. But does this valuation, along with DVA and OCA, have a place in financial statements given the lack of disclosure and comparability of the calculations? Shouldn’t accounting rules serve their end-users — by which we do not mean banks, but the equity-investing owners of banks.

Grandma holds some bank stocks and she is not happy.

___________________________________________________________

We also did a video on the topic. Sadly grandma was unable to join us. Next time maybe.

Related links:
DVAs could add to bank earnings volatility – FT
How one bank’s default is the same bank’s gain – FT Alphaville
The truth behind CVA – Euromoney
Spreads and the Citi - FT Alphaville
Sovereign CDS – still not dead yet, regulations helping – FT Alphaville