What is more obvious at this stage is that we were too optimistic in some of our country forecasts. The biggest miss occurred in France, where we had been looking for 1.1%Q on the basis of strong industrial production and robust consumer spending dynamics. But the INSEE Institute only reported 0.3%Q for the June-to-September quarter today. We also had a sizable miss in Italy, where we had been forecasting 1.2%Q and the outcome only came to half of that at 0.6%Q. A smaller shortfall relative to our projections occurred in Germany, where we saw 0.7%Q instead of our forecast of 0.9%Q. Meanwhile, we and consensus had been rather too pessimistic on Spain, which contracted one-tenth less than expected at 0.3%Q.
Understanding the country dynamics is important because the latest forecast upgrade was triggered by us discovering a mismatch between our bearish top-down euro area GDP indicator and our bullish bottom-up country GDP indicators (see Euroland Economics: Much Better on Bottom-Up, October 14, 2009). At the time, we decided to side with the bottom-up country estimate as it seemed to track the cyclical dynamics much more closely during the recent crisis. It now turns out that the ‘truth' for the euro area lies somewhere in between. The major miss on some of our country GDP indicators warrants some further investigation as they have been tracking their respective economies well historically. Hence, we would not be surprised to see some upward revision to the GDP data in the future. By the same token, we cannot rule out that some of the monthly indicators that our GDP indicators feed off - notably industrial production and consumer or retail spending - could be revised lower.
Leaving the forecast misses aside and taking stock of the data, we find that Italy, the Netherlands and Belgium - along with smaller countries such as Austria and Portugal - re-emerged from recession in 3Q. Only Spain and Greece still contracted in the third quarter. Among the two large countries that had already started to recover in 2Q - Germany and France - France was able to maintain its growth momentum (expanding at the same 0.3%Q pace) while Germany gained momentum, expanding by 0.7%Q (after 0.4%Q before). After Germany, Italy was the second strongest of the large economies in the third quarter, expanding by 0.6%Q. There is an important divide, though, between both countries as far as the entry into the fourth quarter is concerned. While a 2.7%M jump in September industrial production creates a strong entry point into the fourth quarter in Germany, a 5.3%M plunge in Italy does exactly the opposite. Hence, we expect the German economy to maintain most of its momentum in late 2009 whereas we fear that the Italian economy may fall back towards stall-speed in 4Q.
The relatively little demand detail we have so far (quantitative information from France and the Netherlands and qualitative information from Germany and Spain) suggests that consumer spending contracted after the bulk of the ‘cash for clunker' programmes had run their course. While investment spending continued to contract in France and the Netherlands, it seems to have recovered a bit in Germany. This blurs the Euroland picture and suggests that a stagnation in overall investment spending (on both machinery and equipment, as well as construction) is the most likely outcome. Net exports boosted growth notably in France and the Netherlands. In Germany, the stats office remains ‘stumm' on the contribution of net exports and only mentioned a revival in export demand and a surge in imports. The net effect remains unclear. But as the press release states that the surge in imports led to a build-up of inventories, there is actually a rather bullish story behind the numbers. In our view, the combination of higher imports and higher inventories could signal that German companies are getting ready to step up production further in the present quarter.
In terms of the implications for our full-year estimates, the weaker-than-expected 3Q GDP reports introduce some downside risk to both our 2009 and 2010 forecasts. Part of this downside risk will likely be compensated by some upside risk to our 4Q and possibly also 1Q forecasts. We will thus reserve judgment until we have the full 3Q GDP reports to hand, and can assess where the inventory cycle - which, in our view, is a key driver of the near-term dynamics - is at present. At most there is probably a one- to two-tenth downside risk to our Euroland 2010 GDP forecast of 1.2%, which would still leave us a touch above the European Commission (at 0.8%) but far above the ECB (at 0.2%). Hence, our call for a marked upgrade in the ECB staff projections at the December meeting remains intact.
Given the size of the country-specific surprises, there are no major risks to our German and our Spanish forecasts from the 3Q GDP reports. In both cases, though, we still need to factor the 2010 budget plans into our forecasts. In the case of Germany, additional tax cuts could introduce some upside risks, whereas in Spain tighter fiscal policy from the summer onwards could introduce some downside risk, notably to 2011 estimates. In the case of Italy and France, we remain puzzled by the discrepancies between the GDP data and the monthly indicators, but have to acknowledge some sizable downside risk to our 2010 forecasts.
Appendix: Country Details
Germany: Recovery Gains Momentum
Slightly weaker than expected, the 3Q flash estimate shows that the German economy expanded by a decent 0.7%Q in non-annualised terms between June and September. Being up from 0.4%Q in the previous quarter, the recovery in Europe's largest economy is thus gaining momentum. We had been looking for 0.9%Q while consensus was at 0.8%Q. Somewhat puzzling, the upside risk that our GDP indicator had signalled (giving an estimate of 1.36%Q) have not materialised. Without any details on the demand or the production side to go on until the full report is released on November 24, it is difficult to say anything meaningful about the reasons behind the small shortfall. The impact on our full-year estimate will be negligible though, because 2Q GDP growth revised up from 0.3%Q to 0.4%Q. Hence, we are unlikely to make meaningful changes to our full-year forecasts of -4.9% for this year and 1.7% for next year.
According to the stats office, consumer spending proved a drag on headline GDP growth during 3Q, in line with our forecast. At the same time, investment in machinery and equipment, as well as structures, boosted GDP growth. The former comes rather as a surprise to us, given the low rate of capacity utilisation. The latter likely reflects government-sponsored infrastructure programmes. As expected, export demand also added to overall GDP growth the past quarter. At the same time, imports surged, causing a rebuilding of inventories.
France: Undershooting Expectations
France expanded by a non-annualised 0.3%Q in 3Q, undershooting our and consensus expectations by a large margin. We had been looking for a 1.1%Q gain, while the consensus was at 0.6%Q. This was also below the latest published forecast of the French statistics office, INSEE, of 0.5%Q. Risks for growth in 2009 are now somewhat skewed to the downside: compared with our GDP growth forecast of -1.8%, the 3Q GDP report implies that the French economy may contract by 2.3% this year, all else being equal.
The details reveal that private consumption growth was flat on the quarter and that investment declined at a slightly faster pace (-1.4%Q) than in the previous quarter (-1.2%Q). What is more, French firms continued to reduce their stock of inventories, and they did so at a faster pace. This subtracted about 0.1pp from quarterly GDP growth. The only positive contribution came from foreign trade.
But the chances are that 4Q may turn out to be stronger than expected. Before today's numbers, we had pencilled in a gain of just 0.1%Q. The reason for being more optimistic on the fourth quarter is that France exhibits an unprecedented discrepancy between firms' assessment of demand and their view on the stock of inventories, according to the INSEE survey. This cannot continue indefinitely: either demand catches up and firms finally boost production, or the level of stocks will be viewed less optimistically. At this stage, we believe that there is scope for both a slight improvement in demand and an easing in the pace of destocking.
Italy: Solid Rebound, but Not as Much as Expected
3Q GDP expanded by a non-annualised 0.6%Q, below our and consensus expectations. Still, this means that growth is finally back in Italy. Given the extent of the undershoot relative to our initial prediction of 1.2%Q, risks to our current forecasts of -4.5% in 2009 and 1.2% in 2010 are now skewed to the downside. We remain puzzled by the degree of the miss. The various monthly indicators, ranging from industrial production and orders to the various surveys, all pointed to a more pronounced rebound in GDP. No GDP breakdown is available with the flash estimate. According to the Italian stats office, the expansion in overall GDP - after five consecutive quarterly contractions - was due to a boost from both the industrial and services sectors.
Spain: Pace of Contraction Easing
Spain contracted by a non-annualised 0.3%Q, slightly better than our and consensus expectations of a contraction of 0.4%Q. We think that the Spanish economy is set to shrink by 3.6% this year. Moreover, we expect the economy to contract outright in 2010 too, to the tune of -0.7%, thus underperforming the euro area. No GDP breakdown is available with the flash estimate. According to the Spanish stats office, the contraction in overall GDP - the fifth in a row - was due to another negative contribution from domestic demand, which was partly offset by a strong positive contribution from net trade - primarily because imports fell once again, not because exports accelerated substantially.
Walking a tightrope in reining in excess liquidity and in removing monetary policy stimulus. When the last ECB tightening cycle got underway in 2005, it initially caused an outcry amongst European politicians and international organizations, who felt - at the time - that the ECB's move was premature. Yet, the ECB did not waver after both their broad-based analysis of the outlook for price stability and their cross-checking against monetary developments indicated a need to tighten. With the benefit of hindsight, it is clear that the ECB was right. At the present juncture, the macro economic uncertainty, financial market dislocations and foreign exchange repercussions make the exit an even more finely calibrated decision. History shows that even small policy decisions can have a big market impact. Take the fall of 1987 for instance, when a small 10bp hike in the Bundesbank's repo rate caused a transatlantic policy spat. This spat, which aired very publicly in the US media, caused markets to doubt the political commitment to the Louvre Accord, which aimed to stabilise the USD. Concerns about support for the USD sent the US equity market, which already was wobbly in the wake of a hike in capital gains taxes, etc., into a tailspin.
The coming tightening campaign will likely be even trickier, because it has an additional dimension. In addition to the interest rate decision, a range of unconventional measures will need to be unwound. In this note, we look at the exit from these measures. We already discussed the interest rate issue in EuroTower Insights: Tayloring Rates for the Exit, July 29, 2009 and concluded that the ECB will likely keep rates steady until mid-2010. Some unconventional measures might be unwound earlier than that; others will likely remain in place for longer. The non-standard measures, referred to by the ECB as enhanced credit support, are intended to boost liquidity in the financial system and support credit flows to the non-financial sector over and above the interest rate reductions. These measures include
• Switching the refi operations to fixed rate tender with full allotments (i.e., unlimited liquidity),
• Offering supplementary long-term refi operations (LTROs) at 3-, 6- and 12-month maturities,
• Extending the collateral pool down to BBB-,
• Offering foreign exchange liquidity in USD and CHF, and
• Purchasing covered bonds.
At this year's ECB Watchers Conference, ECB President Jean-Claude Trichet and ECB Executive Board Member Juergen Stark outlined how the ECB approach the different dimensions of their policy exit. Essentially, the interest rate decision will be based on the bank's assessment of the risks to price stability. Meanwhile, decisions on enhanced credit support will be taken with a view to financial stability and market functioning. Analytically, it makes sense to separate these two issues. Practically, they are likely to be intertwined though. The ECB stress that price stability takes precedence in all decisions. Should non-standard measures pose a threat to price stability, they will be unwound "promptly and unequivocally". As long as they don't, however, they can be unwound gradually as markets continue to normalise. But if market dislocations continued to impair the transmission mechanism, they could also be left in place for longer.
Most of the ECB's non-standard measures have a sell-by date at which the liquidity will be withdrawn again; unless the bank decides to prolong the programme. According to Juergen Stark and, more recently, Jose Manuel Gonzalez-Paramo, the ECB aims to re-establish a situation where the one-week main refinancing operation (MRO) is the main tool for steering money market rates and where the ECB acts as a ‘rate-taker' in the term-funding money again. The latter would imply switching back to partial allotment or even variable rate tenders. However, the specific steps will depend on the state of the money market. Consider two different scenarios:
• The first scenario is one where the problems in the money markets disappear before any upside risks to price stability emerge. In this case, unconventional measures would be unwound before policy rates are raised. The withdrawal would likely impact money market rates, many of which are below the refi rate at the moment.
• The second scenario is one where upside risks to price stability emerge while the problems in the money market persist and bank funding is still constrained. This could happen if inflation expectations became unhooked, for instance. In this case, the ECB would likely keep enhanced credit support in place and raise its policy rates. Supplying unlimited liquidity to the banking system would likely leave money market rates below the refi rate. In this context, the key question is whether the funding constraints result from dysfunctional markets overall or from specific situations at individual institutions. If the issue was intrinsic to individual institutions, targeted measures outside monetary policy - such as government guarantees, recapitalisations or asset swaps - would seem more appropriate.
Interest rate decision is more straightforward than the non-conventional measures. For the ECB to be willing to raise interest rates, the Council will have to be reasonably confident that the recovery is sustainable. The focus will likely be on the domestic demand recovery, notably corporate investment and consumer spending. A turnaround in the labour market is probably not necessary precondition though (see The Recovery Gains Momentum - But Risks Remain, October 29, 2009). In December 2005, it likely was the monetary analysis that swung the decision towards an early tightening - a decision that was heavily criticised by international organisations and national governments at the time. Our base case is that the ECB will start to raise the refi rate around mid-year. The risk is that the bank decides to wait longer before raising rates though. Two factors could affect the timeline of the tightening, we think: looming credit constraints and rising inflation expectations. The euro is a factor too, but like asset markets, probably less important unless it starts to swing wildly. In any case, the pace of interest rate increases should be gradual. In our view, it will probably take until 2013 before euro area short rates will get close a neutral level.
Ahead of raising the refi rate itself, the ECB could nudge money market rates higher. With the EONIA overnight rate currently at 35bp, there is some scope for ‘stealth' tightening by the ECB. Even though the ECB currently seems content with EONIA being close to the deposit rate, according to ECB President Trichet in the most recent press briefing, they might want to bring market rates closer to the refi rate before actually hiking it. The convergence between market interest rates and the refi rate can be brought about in several ways.
• First, the ECB could simply bring the deposit rate closer to the refi rate, probably as part of a decision to narrow the corridor defined by the deposit rate and the marginal lending facility. The advantage of this approach is it is simple and it directly affects interest rates. By paying higher rates on deposits, the ECB would incentivise banks to hold liquidity buffers. At the same time, liquidity hoarders would have less incentive to offer funds in the interbank market.
• Second, an indirect approach to nudge market interest rates closer to the key ECB policy rate is to drain liquidity from the money market by conducting reverse refi operations in which the ECB offers to take back excess funds at an interest rate that is somewhere above the deposit rate but still below the refi rate. The ECB regularly engages in such reverse refi operations already. But, typically these are quick tenders with a one-day maturity at the end of the maintenance period when the market tends to be awash with cash after minimum reserve requirements have been met. The advantage of reverse refi operations is that they offer operational flexibility. The disadvantage is that relying on them might require rather frequent market intervention.
• The third option is to drain liquidity by issuing so-called ECB debt certificates. By issuing such certificates, which can have a maturity of up to one year, the ECB would enter new territory. Issuing ECB debt certificates would probably be best suited to a situation where some monetary financial institution have a structural liquidity overhang that they are reluctant to offer in the interbank market. It is difficult to assess from the publicly available data whether the deposit facility is repeatedly used by the same institutions or whether there is rotation in the banks that are depositing their excess funds with the ECB overnight. The ECB has the underlying micro data and, hence, is able to judge whether the excess reserves are structurally tied to the same institutions.
• The fourth option is to switch the operational procedure for some or even all of tenders away from fixed rate with full allotment. A first step would be to switch back to partial allotments. This system, which was in place between January 1999 and June 2000, allows the ECB to fully control the actual refi rate paid by banks and the amount of liquidity added to system. If the bids exceed the liquidity offered by the ECB, the allocation will be pro-rata. In this case banks would be forced to raise additional funds in the interbank market. The main drawback of partial allotment is that banks tend to overbid massively. A more radical step would be to go back to the variable rate tenders that were in operation until October 2008, which by definition only offer partial allotment. In contrast to the fixed rate tenders, however, banks will also need to submit an interest rate bid.
Note that the ECB does not necessarily have to nudge EONIA closer to the refi rate before raising the refi rate itself. But if rising inflation expectations caused concerns, the Governing Council could decide to send a strong signal by hiking the refi rate (and/or the deposit rate) without having previously managed market rates back up to the refi rate. Our base case, however, is that the ECB won't be forced to embark on such drastic action and will instead start its tightening campaign by nudging market rates closer to the refi rate again. As the ECB is likely to first roll back the ultra-long refis, this would initially affect the longer-dated EURIBOR rates, later the short-dated EURIBOR rates and, eventually, the EONIA overnight rate.
ECB is unlikely to switch away from fixed rate tenders with full allotment in the MROs anytime soon. Switching back to the variable rate tenders that were in place between July 2000 and October 2008 or fixed rate tenders with partial allotment that were used before July 2000 would obviate the need to manage market rates towards the policy rate in one full sweeping move. For these tenders to work, though, you need to rely on the interbank market to provide institutions that didn't successfully bid at the refi operation with funding. At the start of a new tightening cycle, this might be a rather courageous assumption that the interbank market will be able to effectively match borrowers and lenders. Hence, we think fixed rate tenders with full allotment will likely remain in place for a while longer for the MRO, while some (or even all) of the LTROs could be amended in the new year. With interbank markets not yet fully functional, the ECB will probably find itself intermediating funding markets for some time. On the one hand, the ECB will have to provide access to unlimited funds. On the other hand, the ECB will have to remove a liquidity overhang elsewhere in the system either through reverse refi operations or through issuance of ECB debt certificates.
The ECB could offer only partial allotment on the supplementary six-month LTROs though. This way, the ECB gets back into the driver's seat as far as steering the amount of term funding allocated to the market. By bidding higher interest rates than the refi rate itself, the market would also be gradually weaning itself off the ECB's liquidity drip and move towards a new market equilibrium. As the ECB approaches the start of its tightening cycle, it will likely be reluctant to commit funds for a year or even six months at the current refi rate. However, in the case of variable rate tenders, the ECB also doesn't fully control the effective refi rates paid by banks, and individual bids at very high interest rates could get the market speculating (again) about stress in the system. Here a fixed rate tender with partial allotment could be a sensible intermediate step.
December one-year refi likely to be the last one offered at the refi rate, we think. At this stage, we deem a small spread above the refi rate unlikely for this tender. Such a move would likely be seen by the market as the (premature) start of the ECB tightening campaign. Not only is it unlikely that there is already a consensus on the Council on the need to raise rates next year, but also there have been no attempts by the ECB to prepare markets for such a step. At the November press conference, ECB President Trichet clearly did not want to dispel market expectations for the December tender being the last one. On all its supplementary LTROs, the ECB will need to decide whether they will offer additional ones in the new year. Clearly, market participants are eagerly waiting for a new LTRO calendar being announced. Otherwise, the operational system would just default back to the weekly MRO with a weekly maturity and the standard monthly LTRO with a three-month maturity in 2010.
Another decision the ECB must make next year concerns the collateral pool. The already broad collateral pool was widened further in October 2008, when the credit rating of A- was extended down to BBB- for all assets except asset-backed securities (ABS). The ECB will accept this broader range of collateral until the end of 2010, it announced when it launched its one-year tenders in May. An announcement on collateral should come relatively soon though. Historically, changes to the collateral regime have had very long-lead times, e.g., consolidation of tier one and tier two assets. Here some grandfathering rules were applied too, which stipulated that newly issued paper wasn't eligible anymore. In 2010, the ECB will also conduct one of its regular biannual reviews of the risk control measures. The last review in September 2008 led to a range of technical refinements, reflecting a re-assessment of market and liquidity conditions, the use of eligible assets by counterparties and financial innovation. These refinements included additional valuation haircuts for ABS and uncovered bank bonds and the prohibition of ABS issued by closely linked entities. In February 2009, additional transparency requirements were put in place for ABS credit ratings. Unfortunately, high frequency data on collateral use are not publicly available. But the latest available data show substitution away from government bonds towards ABS, etc., at the end of 2008. At some point, however, the ECB will need to narrow the collateral pool again, especially with a view of helping funding markets, such as ABS, to reopen again.
In addition, the ECB will likely complete its covered-bond-buying programme. Currently standing at around EUR22 billion, the purchase programme is slightly ahead of schedule. It is expected to be completed by June 2010, when the ceiling of EUR60 billion should be reached (about 5% of outstanding covered bonds that are eligible). The programme could possibly be extended, but given its success in bringing covered bond spreads down that might not be necessary. Crucial in the covered-bond-buying programme was lending support for a specific market segment the ECB deems to be key to raising funds for financials rather than to expand central bank balance sheet. Hence, the covered-bond-buying programme is not part of the ECB's QE strategy but part of its enhanced credit support (see EuroTower Insights: Not Quite QE, May 13, 2009). Thus, we expect the ECB to hold the covered bonds to maturity.
Bottom line - What a roadmap for the ECB's policy exit could look like: Some operational decisions depend on information that only the ECB has access to. Others are dependent of market developments. With these caveats in mind, we believe that the sequencing of the ECB's policy exit next year could look like this: First, reduce liquidity offered in ultra-long refi tenders (e.g., one-year and increasingly also six-month tenders). Second, wean banks of the full allotment liquidity drip by switching back to partial allotment and possibly to variable rate tenders for some LTROs. Third, bring EONIA closer to refi rate via reverse refis and/or issuing certificates. Fourth, start to hike interest rates gradually. Fifth, switch MRO tenders away from full allotment. In terms of the timeline, we would expect the active management of money market rates back towards the refi rate to start in the spring, in time for a first rate hike around mid-year.
Read Full Article »