Whilst the cynical me would suggest there is something in it for each bank to post forecasts, Volcker and Frank-Dodd killed prop trading, and therefore banks running their own book. Well theoretically at least.
I suspect some banks still essentially run a P&L book on the premise of managing risk, but not to the extent where they will move markets.
So I don’t believe that Citi is buying and JPM are selling (in very crude terms), and are using research to help their prop trading strategy.
Research departments at banks are run quite differently, with genuinely different thinking emerging from banks. Sometimes, confusingly different departments within the same bank have radically differing opinions which can make banks look somewhat stupid with differing outlooks published days apart with wildly varying predictions.
Putting aside my own views, JPM are specialists in Fixed Income (Bonds) with from my perspective an unparalleled perspective. They also managed to avoid much of the 2008 fallout by having limited exposure to CDS’s and maintaining what they term a “fortress balance sheet” - which is essentially very high value and liquid assets and a high capital to risk ratio. Put simply they managed risk well and whilst other banks faced oblivion, JPM had sufficient capital to both stay afloat and buy assets from other banks despite writing off 5.5bn. Citi on the other hand needed a bailout of approx 50bn.
On balance, I would go with JPM’s estimate, because in my opinion they have the processes and rigour to manage risk at a global market level, with thinking that goes beyond looking at a single asset.
As a side note when talking on risk management, Citi “accidentally” credited USD81 trillion to a single customer earlier this year. They were supposed to send USD280 (yes two hundered and eighty dollars). Their (Citi’s) total assets are only just over USD2bn. And this is not the first time that the basics (like crediting customers with the right amount of money…) were found significantly wanting at Citi.