Ben Bernanke spoke, and the world shuddered. Testifying before Congress, the chairman of the Federal Reserve indicated that the FOMC could start to scale back the pace of its asset purchases “in the next few meetings.” In suggesting that the Fed might start buying less than the current $85 billion of bonds per month as early as September, he pulled the trigger on what Warren Buffet called the shot heard around the world.
Markets around the world went into a tailspin. All of the main G10 equity markets suffered declines on Bernanke’s words. Most notably, the Nikkei 225 tumbled 7%, its largest single day drop since the Tokohu earthquake and tsunami in 2011.
This is remarkable. Bernanke was clear that change to QE would involve buying fewer securities, not ending the programme entirely. In his prepared statement, Bernanke struck a dovish tone, praising extraordinary monetary operations for their role in supporting the real economy and making it clear that the Fed did not want to pull back on stimulus measures prematurely. Continuing Buffet’s analogy, Bernanke might have fired the shot, but the gun was full of blanks.
We did, however, get a preview of how the market might react when monetary conditions do begin to normalise. Three insights are apparent after seeing how financial markets could react if the Fed were to begin to change the direction of monetary policy.
1. There’s no such thing as strictly domestic monetary policy. Formally, the Fed will keep with accommodative monetary policy until the FOMC is sufficiently confident that the unemployment rate will fall below 6.5 per cent. All central banks are bound by their domestic mandates, such as price stability or promoting full employment within their respective borders. Recently the Fed has been focusing on the employment side of its mandate.
The sell-off that met Bernanke’s suggestion of slowing QE underscores the truly global nature of domestic monetary policy decisions. Discussions of monetary policy coming out of central banks around the world often begin with the domestic picture of, say, the labour market and price levels and then turn to international headwinds later. To the extent that international reverberations of domestic monetary policy can have an impact on a central bank’s domestic policy goals, it is indeed within a central bank’s mandate to consider the global implications of monetary policy. In other contexts, such as recent fears over currency wars, the international community can exert significant influence on domestic policy.
So in spite of the Fed’s domestic mandate, the international picture is important as well. If Bernanke just hinting at winding down QE can have such a strong impact on asset prices globally, it’s likely that the Fed doesn’t just look at domestic labour market and price level indicators when deciding on US monetary policy.
2. Asset prices are highly sensitive to stimulus withdrawal. This seems obvious on the face of it, but no one really knew (or knows) how tightening monetary policy will impact asset prices. We have no modern precedent. The last time the Fed raised its benchmark interest rate was in June of 2006, when the target rate was increased from 5 per cent to 5.25 per cent. Interest rates at that level aren’t even conceivable in the current policy environment. We’re in uncharted territory, both with regard to the level and duration of interest rates and the size of the Fed’s balance sheet.
The frenzied downside reaction to Bernanke’s testimony highlights how sensitive global asset prices are to central bank largesse. This is especially true considering the Fed didn’t actually do anything. He hinted at the possibility that the Fed might consider scaling back the pace of QE. Even so, scaling back asset purchases probably wouldn’t alter the path of the Fed’s balance sheet materially. Also, interest rates would still remain pressed up against the zero bound.
The fall in pro-cyclical asset prices probably wasn’t entirely due to Bernanke. China is slowing, and many investors were likely taking some profits off of the recent bull run, particularly in equities. Nonetheless, it’s clear that Bernanke’s comments pushed prices down to some degree. If asset prices were indeed justified by fundamentals, then Bernanke’s comments would have had a more muted effect. The sharp fall in risk assets suggests that central banks are blowing bubbles in the financial markets. The extent to which asset prices are due to bubbles versus fundamentals might be proxied by how steep the drop in prices were in response to Bernanke’s comments. However measured, last week offered some clues about just how dependent on stimulus some assets really are.
3. When monetary normalisation does occur, it will be slow and announced far in advance. Eventually, monetary policy conditions will have to return to normal. Some might argue that future monetary policy will be radically different to pre-2008 trends due to the “new normal” situation in the real economy. I think monetary policy indicators will eventually revert to the long-run averages. Regardless of which camp you might be in, virtually no-one agrees that monetary policy can remain as it has been for the last five years.
This raises the question of how central banks will unwind highly accommodative monetary operations. Like all central bank operations, monetary normalisation will be as much about communication strategy as it will be about the technical details of the policy change.
John Williams of the San Francisco Fed opined that policy movements could be up or down, but that would be out of step with the Fed’s policy since the crisis. The Fed, along with its peers around the world, have tried to blunt the uncertainty associated with being off the monetary map by communicating its policy decisions in advance and even tying policy to some concrete dates or indicator values.
Last week, the Fed saw the effects of announcing the potential of tapering QE months in advance, and they were nasty. Barring a substantial spike in inflation, the Fed might eschew the Volcker approach to monetary tightening and instead slowly communicate to the market that QE will be slowly winding down and, eventually, interest rates will slowly rise. That is, not only will policy be announced far in advance, but policy changes will be gradual. This could allow asset bubbles to deflate prior to the tightening itself, which would hopefully minimise volatility in the markets. It may also allow firms and individuals that hold deflating assets to realise losses more gradually and remain solvent.
This blog has focused on the Fed, but the same arguments apply to the Bank of England or the European Central Bank. In short, the episode surrounding Bernanke’s testimony before Congress gives us a glimpse into what to expect when central banks globally start cooling down the printing presses.
On April 30, India’s Ministry of Civil Aviation allowed airlines to unbundle charges for services including drinks (except drinking water), carriage of sports equipment, seat selection, baggage and so on. While some airlines have implemented additional charges for some services and reduced their free baggage allowance, base fares have not yet declined. The Air Passengers’ Association of India opposed the unbundling decision and worries that the decision would merely be a licence for airlines to indirectly raise fares. Airlines contend that base fares would reduce once they acquire a clearer picture of the potential revenue generating power of ancillary services.
In our article of 3 May 2011 for the Competition Law Insight analysing the Surcharges Supercomplaint by the UK Consumer Forum Which? to the UK Competition Commission (http://www.europe-economics.com/publications/the_surcharges_supercomplaint.pdf), Dr Stefano Ficco and I analysed the practice of airlines charging extra for online booking services in the UK. The Supercomplaint contended that the charges were unfair and anti-consumer. Our analysis at that time is very relevant to the current situation in Indian aviation.
A package of services
In our article, we had written that the product in question was a package of services that enabled the customer to move from one city to another. This includes the price of the ticket as well as that of any ancillary services. There is even an argument to include transport to and from the airport within the package (as evidenced by many airline booking sites offering to book a taxi pick up and drop off with the ticket). It is the price of the complete package that the customer takes into account while making a purchasing decision.
Unbundling and efficiency
Consumer preferences naturally vary when it comes to a bundle of airline services. Some consumers might be flying for a short time, and would not wish to carry too much baggage. Passengers with slight mobility issues (such as the elderly) may place a premium on being able to sit in the first two rows. When the entire bundle is priced at a single price point, airlines are unable to take into account the different willingnesses to pay of different customers for different parts of the package. Unbundling allows service providers to adjust to variations in consumer preferences. Moreover, if certain ancillary services are priced in proportion to cost, this could reduce or eliminate perverse incentives facing customers. For instance, charging for baggage could incentivise passengers to carry only as much as they require, leading to reduced weight and subsequently savings on fuel.
Since the price of the complete package is taken into account by the customer, unbundling would have the effect of allowing providers to provide several, slightly differentiated, products where earlier they could only differentiate based on limited criteria like class of travel. The Indian aviation sector is sufficiently competitive – data from the Directorate General of Civil Aviation shows that the largest carrier has a market share of 30 per cent and four carriers have shares of over 19 per cent – and this would limit the ability of airlines to use pricing of ancillary services to extract any more surplus from customers than they are currently doing.
Theoretically, at least, unbundling should lead to an efficiency gain. Consumers, faced with an incentive structure for every service can calibrate their package according to their preferences, and being able to charge for ancillary services should lead to a reduction in base fare as the price for the total package remains the same, or reduces due to properly incentivised consumer choices.
The death of traditional aviation in India?
Of course, the model described above has been adopted consistently for a long time by low cost carriers in Europe (like easyJet and Ryanair). This has not led to the complete elimination of airlines offering only the package rather than its constituent parts. Two insights from behavioural economics can easily explain this.
First, there is a known disutility associated with making too many choices – a mental transaction cost. For some customers, the extra time and mental energy spent deciding on each of the constituent parts of the package may not be worth the monetary savings in relation to deciding on a package price.
Second, since flying with low cost carriers may be seen as a signal of belonging to a lower economic class, there might be a ‘social status’ premium that some fliers might attach to not flying with a low cost carrier.
These factors, specially the first, should guarantee that traditional all-inclusive tickets will remain a part of Indian aviation.
Potential pitfalls – search costs
There have been recent media reports to the effect that the Indian Ministry of Civil Aviation is likely to ask airlines to increase transparency regarding the services they would charge extra for and is considering capping the number of ‘privileged’ seats that may be reserved on payment of a fee. This is closely related to an issue we raised in our article for the Competition Law Insight in May 2011 – that of search costs. The issue, in a nutshell, is that the total package price is not known to the consumer a priori. The customer has to search for the prices of ancillary services, which are often revealed only at the end of a lengthy booking procedure in order to learn the final price. The time and energy invested in learning the price entails a cost. I quote from our 2011 article:
“In economics, search costs are known to detract from competitive forces – effective competition relies on consumers being able to compare all package prices, but with costs involved in learning package prices, consumers are less likely to want to search for information on all package prices from all airlines, hindering their ability to compare. This gives airlines market power as it allows them to charge more than they would be able to in the presence of effective competition. Naturally, the higher the search costs, the lower the willingness of consumers to carry out an additional search, and the higher the market power enjoyed by airlines.”
The Ministry of Civil Aviation is correct to ask for greater transparency – publishing a list of ancillary services along with their prices on the home pages of websites (for example) would significantly reduce the costs faced by potential customers in learning the final package price.
Whether or not unbundling will ultimately benefit the Indian consumer and whether base prices do eventually come down are things that time will tell. The Ministry of Civil Aviation can, however, increase the likelihood of this move being successful by pre-empting attempts to sabotage the move by artificially increasing search costs. In this regard, transparency and ease of access to information requirements would play a major role.
One area where the Eurozone did manage to achieve convergence is in government borrowing costs. The Maastricht Treaty, economists argued, had the effect of minimising currency risk premiums among future Eurozone countries over time, and effectively disappeared with the introduction of the euro. Countries enjoyed investor faith in the euro, which was viewed as a hard currency. If this is the case, then it stands to reason that if investors were to lose faith in the euro as a hard currency, or that economic fundamentals had not come into line sufficiently to minimise currency or default risk, then borrowing costs should rise. Unsurprisingly, this is what happened.
Figure 1: Annual average yield on 10-year sovereign bonds, 1995-2012
Source: United Nations, Bloomberg, Europe Economics
In fact, those countries that benefitted the most from joining the euro are precisely those countries that are now seeing their yields rise. In order, Greece, Italy, Spain, and Portugal saw their borrowing costs fall the most between 1995 and 1999. Finland, France, Germany, and the Netherlands also witnessed yields on their sovereign debt come down, but less than the GIPS group.
Core countries benefitted from the reduction in risk premiums brought on by the euro, but it also appears their fundamentals were more in step with what investors prefer, as yields on their debt since 2008 have fallen as well. That, or they’re the “cleanest dirty shirts” for investors that can’t get out of their euro exposure.
GIPS countries, on the other hand, borrowed at lower rates almost exclusively due to the risk premiums effect of the euro. As Eurozone members, they allowed their fundamentals to go haywire and drift substantially in the wrong direction away from the Eurozone average. Now they are, literally, paying for it.
Figure 2: Change in annual average yield on 10-year sovereign bonds
Source: United Nations, Bloomberg, Europe Economics
The hope of the Maastricht Treaty was that economic management principles imposed on countries before joining the Eurozone would minimise differences among national economic structures, allowing for a well-functioning currency union. Instead, differences persisted, but were largely papered over — by euros. The sovereign debt crisis exposed lingering differences and began to pull yields of distressed countries towards their pre-euro levels. Rates on GIPS sovereign debt have fallen recently, thanks in part to economic reforms, and in part to the Draghi Put. But you can’t fight fundamentals forever. In the long-term, distressed Eurozone sovereigns, if they wish to remain in the single currency area, will need to continue to bring their economic fundamentals in line with their non-distressed peers to borrow at affordable rates.
The UK and US labour markets have both been in a dreadful state for sometime now. Unemployment rates on both sides of the Atlantic are nowhere near their pre-recession levels and, even though both markets are showing some signs of improvement, many commentators have become resigned to a "new normal" of higher (non-inflationary) unemployment.
One way of analysing patterns in labour market dynamics is to look at the Beveridge Curve, which plots the unemployment rate against the job vacancy rate (i.e. the number of job openings over the size of the labour force). Beveridge Curves are typically downward-sloping. When jobs are plentiful, unemployment tends to be low; when jobs are scarce, it’s harder to find employment.
The Great Recession pulled the UK and US labour markets from the high vacancy, low unemployment end of the Beveridge Curve to the low vacancy, high unemployment end. The Beveridge Curve for the year June 2008 – June 2009 shows a distinct migration across unemployment-vacancy plot.
Figure 1 UK Beveridge Curve: Jan 2002 - Feb 2013
Source: Office of National Statistics, Bloomberg, Europe Economics
In the UK, higher unemployment has been associated with lower vacancies, as theory would predict.
Figure 2 US Beveridge Curve: Jan 2002 - Mar 2013
Source: Bureau of Labor Statistics, Bloomberg, Europe Economics
In the US, however, a given level of the vacancy rate today implies a much higher unemployment rate today than it did before the recession. For example, between January 2002 and May 2008, a vacancy rate of 2.5 per cent was associated with an unemployment rate of around 5.5 to 6 per cent. From July 2009 onwards, that same vacancy rate was realised where unemployment sat around 8 to 8.5 per cent. In other words, the Beveridge Curve has shifted outwards.
There are a number of reasons why this can happen, and no one reason can explain the entire shift. One argument is that there is a structural mismatch between jobseekers and employers in the US. It might be that the workforce simply doesn’t have the skills necessary to fill the vacant jobs. If skills aren’t the problem, there could be a geographical separation between where the workers are and where the jobs are. The relative illiquidity of the post-recession housing market, for instance, could inhibit labour mobility.
Alternatively, the shift could be the result of hold-outs in the labour market. On the labour supply side, workers, who have been able to claim unemployment benefits for a longer period of time, might be holding out to find a better job. On the labour demand side, employers could be delaying hiring decisions in hopes that a better candidate will apply or until growth expectations rise. Workers who are chronically unemployed could find that their skills — actual or perceived — have deteriorated to such a point that employers are unwilling to give them a chance. A Boston Fed paper argues this is what we’ve seen recently in the US, where, they argue, the Beveridge Curve has shifted outwards only for the chronically unemployed.
Estimated Beveridge Curves using data from the pre- and post-recession observed Beveridge Curves suggest that the US Beveridge Curve has indeed shifted to the right. It also flattened, indicating additional frictions in the labour market.
Figure 3 Pre- and post-recession estimated linear UK and US Beveridge Curves
Source: Bureau of Labor Statistics, Office of National Statistics, Bloomberg, Europe Economics
This is markedly different from what we’ve seen in the UK over a similar time period. Instead of shifting outward, the post-recession estimated UK Beveridge Curve has actually shifted inward, which should indicate an improvement in labour market conditions. With unemployment stubbornly lingering around the 8 per cent mark since early 2009, however, it would be hard to argue that there have been improvements. Nevertheless, patterns in the current, post-recession relationship between unemployment and job vacancies don’t seem to have undergone the structural shift we see in the US.
Instead, the primary driver of patterns in the UK Beveridge Curve is a lack of job vacancies to absorb unemployed workers or workers returning to the labour market.
The UK and US labour markets bottomed out in the summer of 2009. Since then, UK job vacancies have picked up modestly, growing by roughly 15 per cent to date. This pales in comparison to the US, which has seen job vacancies increase by around 74 per cent over the same period.
Figure 4 Job vacancies in the UK and the US, Jan 2002 - Feb 2013
Source: Bureau of Labor Statistics, Office of National Statistics, Bloomberg, Europe Economics
So while the UK and US continue to struggle with unemployment, the underlying causes are distinct. In the US, there are plenty of vacancies, but a breakdown between the process of matching employers with employees has kept them unfilled. The anti-clockwise movement in the US Beveridge Curve is typical of post-recession labour market dynamics, and may even point to a future recovery.
In the UK, there has been less of a recent structural shift in labour market dynamics, though evidence suggests there has been in some sectors, notably technology. Of course, we won’t really know if there’s been a shift in the UK Beveridge Curve until the vacancy rate begins ticking upwards, allowing for a comparison of unemployment rates for the same level of the vacancy rate. Once this happens, we might see that matching in the UK labour market has broken down as well, as the IMF has suggested. Nevertheless, the more immediate problem facing the UK — apart from the productivity puzzle, which demands a separate blog post — is not how to slot workers into open jobs, but how to create those job openings in the first place.
A lot of ink has been spilled in recent years on what was once an obscure back-office settlements mechanism. TARGET2, the Eurozone’s real-time interbank payments system, has been labelled a means to finance current account deficits, a passage for capital flight, and a way for non-Eurozone banks to minimise redenomination risk in the event of a break-up of the euro. Actually, all three are true. Identifying the root cause of changes in TARGET2 balances requires a case-by-case analysis.
So how does TARGET2 work? Cross-border financial flows within the Eurozone are intermediated by national central banks, whose transactions are conducted via the ECB. TARGET2 acts as the settlement platform for these flows. This helps maintain the balance of payments identity:
Current Account + Capital Account + Official Settlements Balance ≡ 0
Banks in countries running current account deficits find themselves short of capital due to having transferred customer deposits on its balance sheet to banks in countries with current account surpluses. Under “normal” times, banks in current account deficit countries can recapitalise themselves by turning to the interbank lending market, effectively borrowing back the capital transferred in the deposit transaction. This settles the balance of payments.
In times of financial distress, banks might find borrowing from the interbank market too expensive. Furthermore, the ability to satisfy the balance of payments identity might become strained due to capital flight from distressed countries.
This is where TARGET2 balances come in. Banks in current account deficit countries can post collateral at their domestic central bank, which then transfers the banks cash from the ECB. Sober Look has a simple diagram of this process. TARGET2 liabilities come in via the “Official Settlements Balance” in the balance of payments identity.
Capital has been flowing out of peripheral European banking systems over the past year. GIIPS bank balance sheets shrank 3.2% from March 2012 to March 2013, while Cypriot bank balance sheets were over 8% smaller. Tellingly, in the 11 months from March 2012 to February 2013, GIIPS banking systems contracted 3.3% and Cypriot banking systems only 3.1%. This means the Cypriot banking system diminished a full 5% — more than it had over the past 11 months — during March’s financial meltdown. As steep a downturn as this is, the drop would likely have been far steeper without the strict capital control Cyprus imposed on investors.
Figure 1. Total size of MFI balance sheets, Mar 2012 - Mar 2013
Source: Central Bank of Cyprus, Banco de Portugal, Central Bank of Ireland, Banca d’Italia, Bank of Greece, Banco de España, Bloomberg, Europe Economics.
Given the deteriorating capital accounts of Cyprus and the GIIPS countries, their balance of payments can be settled in two ways: cutting the current account or recapitalising with TARGET2 funds.
The traditional route out of a balance of payments crisis is to slash the current account deficit. To some extent, peripheral Europe has been taking this route. Since the end of 2008, Spain and Portugal have taken steps to bring down their sizable current account deficits. Relative to its peers, Italy has managed its current account deficit fairly well; its current account deficit as a per cent of GDP in the last quarter of 2012 was in line with its medium-term average. Cyprus, however, breaks from this pattern. True, Cyprus’ current account is nowhere near its recent peak of 20% of GDP, but its consolidation has been rocky. Furthermore, the Cypriot current account deficit has been ticking steadily upward, both as a per cent of GDP and in raw numbers, since the middle of last year. Before the euro, Cyprus would have had to dip into its reserves.
Figure 2. Current account deficit as a per cent of GDP, 2006 Q1 – 2012 Q4
Source: Eurostat, Europe Economics
But under the monetary union, Cyprus can instead tap the TARGET2 system, incurring liabilities to the Central Bank of Cyprus and, by extension, the Eurosystem. Starting around the middle of last year, TARGET2 liabilities as a per cent of total banking assets began to fall in both Cyprus and the GIIPS countries. They continued to fall in GIIPS countries in the first three months of 2013, but rose, both as a per cent of bank assets and in levels, in Cyprus.
Figure 3. TARGET2 liabilities as a per cent of MFI assets, Mar 2012 - Mar 2013
Source: Central Bank of Cyprus, Banco de Portugal, Central Bank of Ireland, Banca d’Italia, Bank of Greece, Banco de España, Bloomberg, Europe Economics
Over the past 9 months or so, GIIPS countries have responded to capital outflows by reducing their current account deficits, while Cyprus has become increasingly reliant on TARGET2 support to keep its banks capitalised. That’s not to say GIIPS countries didn’t or now don't need TARGET2 largesse to keep their banking systems afloat. They did, and do. But patterns in their TARGET2 liabilities and current account deficits might be a template for what we can expect in Cyprus over the coming months.
One thing is certain: the Cypriot capital account is not going to surge upward any time soon. Going forward, then, Cyprus can either reduce its current account deficit, or become more reliant on TARGET2 capital. It’s impossible for Cyprus to adjust the current account fast enough to match the speed of capital outflows, even if capital controls will last until September. In the short-term, I expect Cypriot TARGET2 liabilities, both in levels and as a per cent of banking assets, to continue their ascent. Over the medium-term, barring extraordinary action from the ECB, Cyprus will probably bring down its current account, as other GIIPS countries have done since their banking crises.