Thursday, May 16, 2013

16/5/2013: On That Impossible Monetary Policy Dilemma

At last, the IMF has published something beefy on the extraordinary (or so-called 'unconventional') monetary policy instruments unrolled by the ECB, BOJ, BOE and the Fed since the start of the crisis in the context of the question I been asking for some time now: What happens when these measures are unwound?

See and


Box 2 in the main paper is worth a special consideration as it covers Potential Costs of Exit to Central Banks. Italics are mine.

Per IMF: "Losses to central bank balance sheets upon exit are likely to stem from a maturity mismatch between assets and liabilities. In normal circumstances, higher interest rates—and thus lower bond prices—would lead to an immediate valuation loss to the central bank. These losses, though, would be fully recouped if assets were held to maturity. [In other words, normally, when CBs exit QE operations, they sell the Government bonds accumulated during the QE. This leads to a rise in supply of Government bonds in the market, raising yields and lowering prices of these bonds, with CBs taking a 'loss' on lower prices basis. In normal cases, CBs tend to accumulate shorter-term Government bonds in greater numbers, so sales and thus price decreases would normally be associated with the front end of maturity profile - meaning with shorter-dated bonds. Lastly, in normal cases of QE, some of the shorter-dated bonds would have matured by the time the CBs begin dumping them in the market, naturally reducing some of the supply glut on exits.]

But current times are not normal. "Two things have changed with the current policy environment: (i) balance sheets have grown enormously, and (ii) assets purchased are much longer-dated on average and will likely not roll-off central bank balance sheets before exit begins."

This means that in current environment, as contrasted by normal unwinding of the QE operations. "…valuation losses will be amplified and become realized losses if central banks sell assets in an attempt to permanently diminish excess reserves. But central banks will not be able to shrink their balance sheets overnight. In the interim, similar losses would arise from paying higher interest rates on reserves (and other liquidity absorbing instruments) than earning on assets held (mostly fixed coupon payments). This would not have been the case in normal times, when there was no need to sell significant amounts of longer-dated bonds and when most central bank liabilities were non-interest bearing (currency in circulation)."

The IMF notes that "the ECB is less exposed to losses from higher interest rates as its assets— primarily loans to banks rather than bond purchases—are of relatively short maturity and yields of its loans to banks are indexed on the policy rate; the ECB is thus not included in the estimates of losses that follow." [Note: this does not mean that the ECB unwinding of extraordinary measures will be painless, but that the current IMF paper is not covering these. In many ways, ECB will face an even bigger problem: withdrawing liquidity supply to the banks that are sick (if not permanently, at least for a long period of time) will risk destabilising the financial system. This cost to ECB will likely be compounded by the fact that unwinding loans to banks will require banks to claw liquidity out of the existent assets in the environment where there is already a drastic shortage of credit supply to the real economy. Lastly, the ECB will also face indirect costs of unwinding its measures that will work through the mechanism similar to the above because European banks used much of ECB's emergency liquidity supply to buy Government bonds. Thus unlike say the BOE, ECB unwinding will lead to banks, not the ECB, selling some of the Government bonds and this will have an adverse impact on the Sovereign yields, despite the fact that the IMF does not estimate such effect in the present paper.]

The chart below "shows the net present value (NPV) estimate of losses in three different scenarios." Here's how to read that chart:

  • "Losses are estimated given today’s balance sheet (no expansion) and the balance sheet that would result from expected purchases to end 2013 (end 2014 for the BOJ, accounting for QQME). 
  • "Losses are estimated while assuming everything else remains unchanged (notably absent capital gains or income from asset holdings)… [so that] no stance is taken as to the precise path and timing of exit. ...These losses—which may be significant even if spread over several years—would impact fiscal balances through reduced profit transfers to government. 
  • "Scenario 1 foresees a limited parallel shift in the yield curve by 100 bps from today’s levels. 
  • "Scenario 2, a more likely case corresponding to a stronger growth scenario requiring a steady normalization of rates, suggests a flatter yield curve, 400 bps higher at the short end and 225 bps at the long. The scenario is similar to the Fed’s tightening from November 1993 to February 1995, which saw one year rates increase by around 400bps. Losses in this case would amount to between 2 percent and 4.3 percent of GDP,  depending on the central bank. 
  • "Scenario 3 is a tail risk scenario, in which policy has to react to a loss of confidence in the currency or in the central bank’s commitment to price stability, or to a severe commodity price shock with second round effects. The short and long ends of the yield curve increase by 600 bps and 375 bps respectively, and losses rise to between 2 percent and 7.5 percent of GDP. 
  • "Scenarios 2 and 3 foresee somewhat smaller hikes for the BOJ, given the persistence of the ZLB.

And now on transmission of the shocks: "The appropriate sequence of policy actions in an eventual exit is relatively clear.

  • "A tightening cycle would begin with some forward guidance provided by the central bank on the timing and pace of interest rate hikes. [At which point bond markets will also start repricing forward Government paper, leading to bond markets prices drops and mounting paper losses on the assets side of CBs balance sheets]
  • "It would then be followed by higher short-term interest rates, guided over a first (likely lengthy) period by central bank floor rates (which can be hiked at any time, independently of the level of reserves) until excess reserves are substantially removed. [So shorter rates will rise first, implying that shorter-term interbank funding costs will also rise, leading to a rise in lone rates disproportionately for banks reliant on short interbank loans - guess where will Irish banks be by then if the 'reforms' we have for them in mind succeed?
  • "Term open market operations (“reverse repos” or other liquidity absorbing instruments) would be used to drain excess reserves initially; outright asset sales would likely be more difficult in the early part of the transition, until the price of longer-term assets had adjusted. Higher reserve requirements (remunerated or unremunerated) could also be employed." [All of which mean that whatever credit supply to private sector would have been before the unwinding starts, it will become even more constrained and costlier to obtain once the unwinding begins.]
  • For a kicker to that last comment: "The transmission of policy, though, is likely to somewhat bumpy in the tightening cycle associated with exit. Reduced competition for funding in the presence of substantial excess reserve balances tends to weaken the transmission mechanism. Though higher rates paid on reserves and other liquidity absorbing instruments should generally increase other short-term market rates (for example, unsecured interbank rates, repo rates, commercial paper rates), there may be some slippage, with market rates lagging. This could occur because of market segmentation, with cash rich lenders not able to benefit from the central bank’s official deposit rate, or lack of arbitrage in a hardly operating money market flush with liquidity. Also, there may be limits as to how much liquidity the central bank can absorb at reasonable rates, since banks would face capital charges and leverage ratio constraints against repo lending." [But none of these effects - generally acting to reduce immediate pressure of CB unwinding of QE measures - apply to the Irish banks and will unlikely apply to the ECB case in general precisely because the banks own balance sheets will be directly impacted by the ECB unwinding.]
  • "There is also a risk that even if policy rates are raised gradually, longer-term yields could increase sharply. While central banks should be able to manage expectations of the pace of bond sales and rise in future short-term rates—at least for the coming 2 to 3 years—through enhanced forward guidance and more solid communication channels, they have less control over the term premium component of long-term rates (the return required to bear interest rate risk) and over longer-term expectations. These could jump because leveraged investors could “run for the door” in the hope of locking in profits, because of expected reverse portfolio rebalancing effects from bond sales, uncertainty over inflation prospects or because of fiscal policy, financial stability or other macro risks emerging at the time of exit. To the extent a rise in long-term rates triggers cross-border flows, exchange rate volatility is bound to increase, further complicating policy decisions." [All of which means two things: (a) any and all institutions holding 'sticky' (e.g. mandated) positions in G7 bonds will be hammered by speculative and book-profit exits (guess what these institutions are? right: pension funds and insurance companies and banks who 'hold to maturity' G7-linked risky bonds - e.g peripheral euro area bonds), and (b) long-term interest rates will rise and can rise in a 'jump fashion' - abruptly and significantly (and guess what determines the cost of mortgages and existent not-fixed rate loans?).]

And so we do  it forget the ECB plight, here's what the technical note had to say about Frankfurt's dilemma:  "The ECB faces relatively little direct interest rate risk, as the bulk of its loan assets are linked to its short-term policy rate. However, it may be difficult for the ECB to shrink its balance sheet, as those commercial banks currently borrowing from the ECB may not easily be able to repay loans on maturity. The ECB could use other instruments to drain surplus liquidity, but could then face some loss of net income as the yield on liquidity-draining open market operations (OMOs) could exceed the rate earned on lending, assuming a positively-sloped yield curve, if draining operations were of a longer maturity." [I would evoke the 'No Sh*t, Sherlock" clause here: who could have thought Euro area's commercial banks "may not easily be able to repay loans on maturity". I mean they are beaming with health and are full of good loans they can call in to cover an ECB unwind call… right?]

Obviously, not the IMF as it does cover the 'geographic' divergence in unwinding risks: "But the ECB potentially faces credit risk on its lending to the banking system for financial stability purposes. In a “benign” scenario, where monetary tightening is a response to higher inflation resulting from economic growth, non-performing loans should fall and bank balance sheets should improve. But even then, some areas of the eurozone may lag in economic recovery. Banks in such areas could come under further pressure in a rising rate environment: weak banks may not be able to pass on to weak customers the rising costs of financing their balance sheets." [No prize for guessing which 'areas' the IMF has in mind for being whacked the hardest with ECB unwinding measures.]

So would you like to take the centre-case scenario at 1/2 Fed impact measure for ECB costs and apply to Ireland's case? Ok - we are guessing here, but it will be close to:

  1. Euro area-wide impact of -1.0-1.5% GDP shaved off with most impact absorbed by the peripheral states; and
  2. Yields rises of ca 200-220bps on longer term paper, which will automatically translate into massive losses on banks balancesheets (and all balancesheets for institutions holding Government bonds). 
  3. The impact of (2) will be more severe for peripheral countries via 2 channels: normal premium channel on peripheral bonds compared to Bunds and via margins hikes on loans by the banks to compensate for losses sustained on bonds.
  4. Net result? Try mortgages rates rising over time by, say 300bps? or 350bps? You say 'extreme'? Not really - per crisis historical ECB repo rate averages at 3.10% which is 260bps higher than current repo rate... 
Ooopsy... as some would say. Have a nice day paying that 30 year mortgage on negative equity home in Co Meath (or Dublin 4 for that matter).

1 comment:

Unknown said...

The IMF fears are unfounded because excess bank reserves can only be mopped up by sales of short-dated paper (banks eschew duration risk) whilst all long-dated assets will be held to maturity. Perhaps the authors don't realise this?