Embed Size (px)
Citation preview
Very preliminary, please do not quote
Greece, the Eurozone, and ECB Monetary Policy
Mats Persson,
Institute for International Economic Studies, Stockholm University, S-106 91 Stockholm, Sweden
E-mail: [email protected]
1. Introduction
The debt crisis has now plagued the eurozone for more than half a decade, and the literature
dealing with it is enormous. The economics profession has not yet reached any consensus on
how to look upon, and interpret, what has happened during these dramatic years. In this lecture, I
will discuss a few points where academic economists are nevertheless beginning to agree, and a
few points where they are not.
Despite its title, the lecture does not primarily deal with Greece per se, but with the problems
facing the Eurozone when one or several member states run into debt problems. Greece is just an
illuminating example; although the basic causes of the debts and deficits are different, Spain or
Portugal or Ireland may equally well serve as examples.
Let us start with a very brief recapitulation of what happened. In early 2010, it was discovered
that the Greek government budget deficit was substantially larger, and the level of government
debt substantially higher, than reported by the Greek government’s statistical agency. This
revelation caused the interest rates on Greek government bonds to soar and, in the spring of
2010, Greece could not roll over its debt at what was considered “reasonable” rates of interest.1
Félix Neubergh Lecture, held at University of Gothenburg on November 4, 2015. 1 In Section 3.2 below I will discuss what “reasonable” interest rates might be.
2
The European Union (via its ad hoc support funds EFSF, EFSM and ESM), the European
Central Bank and the International Monetary Fund stepped in and lent money to the Greek
government at considerably lower interest rates that the market demanded. Thus, Greece could
continue to roll over its debt as it came due. This procedure continued over the years to follow,
and today (in late 2015) the total support to Greece from the EU funds and the IMF amounts to
around €185 billion. Moreover, the ECB has provided €130 billion in the form of loans to Greek
banks.2
In the meantime, doubts about the finances of other euro countries drove up the interest rates of
their government bonds. In order to prevent a so-called “sudden stop” – i.e., that the borrowing
costs became so high that borrowing would be unsustainable – Spain, Portugal, and Ireland also
received support. The total amount lent to those countries ended up at €187 billion in the form of
direct loans.3
This brief summary of what actually happened is not controversial; I think all economists agree
with my description. In the next section, I will give a more detailed description of what actually
happened. For brevity, I will use the label “the Troika” for the eurozone (EZ) support funds, the
ECB and the IMF.
2. Five facts about the support measures
To academic economists who have been studying the debt crisis, the five facts that I am going to
mention in this section are hardly controversial; they will fit nicely into simple description in the
previous section of what actually happened.4 However, these facts are not widely known among
policymakers and journalists – and even among non-specialized scholars, they are surprisingly
unknown. Since they are important for the interpretation of the events, and for the policy
conclusions one may draw from these events, they deserve a somewhat longer discussion.
2 Sources for these figures, and for the corresponding figures for Spain, Portugal, and Ireland are the web pages of ESFS, EFSM, ESM and IMF (accessed 2015-10-28). 3 Cyprus, Hungary, Latvia and Romania have also received support, but at much lower levels. In the sequel, when I refer to “the crisis countries”, I abstract from those. 4 For a representative sample of such economists, see, for instance, the contributions in Baldwin and Giavazzi (2015).
3
Fact 1: The interest rate on Greek bonds was relatively high even prior to the crisis.
When discussing the crisis with non-specialists, one often hears the statement that “The financial
markets were malfunctioning. Prior to the crisis, Greece could borrow at the same interest as
Germany”. This statement is plainly wrong. In Figure 1, I show the interest rate spread between
Greek and German government bonds from 2005 to 2009 (for interest spreads during later years,
see Figure 3 below). We see that already in 2005-2007, the Greek bond had a premium, as
compared to the German bond, of around 25 basis points. This spread then increased and reached
50 basis points in early 2008, and 100 basis points half a year later. In 2009, the spread
fluctuated around 200 basis points.5
Figure 1: Interest spreads of 10-year government bonds for a number of European
countries against Germany, 2005-2009. Percent. Source: Eurostat.
One may, of course, say that the spread was too small, given that the Greek public finances were
in such a bad shape. Or one may argue that the spread was too large, given that the true state of
5 The dramatic development of Irish interest rates in 2008-2009 is due to the local banking crisis, where the Irish government had to recapitalize the major banks – banks that were almost too large for even a country with a fairly moderate government debt (cf. Table 1 below).
‐0,50
0,00
0,50
1,00
1,50
2,00
2,50
3,00
0 12 24 36 48 60
Ireland Greece Spain Italy Portugal
2005 2006 2007 2008 2009
4
the Greek public finances was hidden from the market due to the machinations of the Greek
government and its national statistical agency. In any case, the spread was not zero; for whatever
reason, the market placed some kind of risk premium on the interest rate of Greek government
bonds in the years prior to the crisis, and that premium grew particularly large in 2009.
Fact 2: The money contributed by the Troika did not go to the Greeks.
In the public debate during the crisis years, we often heard statements to the extent that “out of
solidarity, we must help the Greeks”.6 The truth is that most of the support loans were used to
save the banks that had invested in Greek bonds from losses. The mechanism is simple. Before
the crisis, Greek bonds had a relatively high yield (cf. Figure 1) which made them attractive to a
large spectrum of investors. But a high yield is normally a rational compensation for risk – and
when a Greek default seemed imminent in the spring of 2010, those bonds fell in value. If
Greece had defaulted, the investors would have made huge losses.
Instead, at each instant when Greek government debt should be serviced, the Troika provided
new loans. This means that the old creditors got their money back, at the full value of their
claims, and the debt was rolled over by shifting it to the EZ support funds and to the IMF, which
mainly means the EZ taxpayers (the loans provided by the IMF came from a broader source,
namely the taxpayers of all countries – EZ and non-EZ – that are members of the IMF). Thus, the
money did not go to the Greeks, but to the banks and other institutions that had lent money to
Greece in the first place; the bad risk that materialized in early 2010 did not harm the original
lenders, but was absorbed by the taxpayers behind the Troika.
There is one exception to this. In the spring of 2012, there was a general agreement among
European politicians that the private lenders had gotten away too easily. Thus, a “restructuring”
took place of Greek debt not yet held by the Troika. The restructuring meant a lengthening of the 6 Such statements abound. Here are just two random examples: In February 2012, the President and the Vice President of the European Parliament, Martin Schultz and Roberta Angelilli, express “solidarity with the Greek people” in connection with the decision to launch the 2012 bailout package: http://www.eppgroup.eu/press-release/Europeans-show-solidarity-with-the-Greeks In January 2015, Spanish Finance Minister Luis de Guindos said that “Thanks to this financing, which Greece could not get from financial markets, it was able to maintain all of its public services… to pay its doctors, its police, its retirees, thanks to this solidarity”: http://news.yahoo.com/spain-reminds-greeks-owe-europe-222240064.html
5
maturity of the debt, together with a reduction in the interest payments – a kind of “orderly
default”. The total fall in market value caused by this restructuring has been calculated to around
60 percent. However, note that this fall in value only applied to the Greek debt that was still in
private hands, which was about XX percent of outstanding debt at that time; Greek bonds held
by the Troika were not affected. 7
All things considered, recent estimates say that about 90 percent of the total support loans seem
to have gone to the banks and the other private holders of Greek government bonds, while about
10 percent seem to have gone to the Greek government for alleviating part of the budget deficit.8
However, this coarse estimate, only refers to the loans to Greece from the EZ support funds and
the IMF. The incidence of the ECB actions is much more difficult to evaluate. By purchasing
Greek bonds in the secondary market, and by lending money to the Greek banks against
collateral consisting of Greek government bonds, it is clear that the ECB actions benefited both
banks (by bidding up the prices of bonds in the bank portfolios) and the Greek government (by
making it easier to borrow in the market). How much each of these parties gained has, however,
to the best of my knowledge not been calculated. This is, in fact, an interesting research topic.
In any case, it is not correct to say that the EZ taxpayers have given hundreds of billions of euros
to the Greeks – a more correct figure would probably be tens of billions, while hundreds of
billions went to the banks. Or, rather, to the owners of the banks.
Fact 3: If the objective was to save the banks, this policy was an inefficient and expensive way to
do it.
The mundane objective of saving the banks may be far away from the noble objective of
“showing solidarity with the Greeks” mentioned above – but it is still a respectable objective. In
the spring of 2010, the U.S. savings and loans crisis and the bankruptcy of Lehman Brothers in
September 2008 were still in fresh memory, and everybody feared a new round of financial
crisis. If Greece had written down its debt in 2010, a solution now favored by many economists
7 For a detailed account of the 2012 restructuring, see Zettelmeyer et al. (2013). 8 See, e.g., The Guardian, July 29 2015, and the New York Times, July 30 2015. For another attempt to estimate where the money went, see Mouzakis (2015).
6
who have written about the crisis,9 it is true that several banks would have been at a risk. Exactly
which banks were in danger depends on how the exposure to Greek bonds was distributed at the
time, and how much the Greek government would have chosen to write down its debt. But
certainly most Greek banks (because of the home bias that always seems to exist in banking),
and probably a handful of French and German banks (because of the calculated risk stimulated
by relatively high yields on Greek bonds prior to the crisis – cf. Figure 1 above) would have
risked bankruptcy if there had been an orderly Greek default in 2010.
In any event, as the debt turned sour, the method chosen by the EU and the IMF was to take over
the bad debt from all investors: from banks that were so heavily exposed that they would
otherwise have risked bankruptcy, and from other investors that would not constitute a threat to
the financial system. This latter group consisted of banks that were less exposed and would only
have made moderate losses without risking bankruptcy, other financial institutions (like
insurance companies), hedge funds, industrial corporations, and private individuals.
Saving everybody from losses is, by definition, much more costly than saving only the banks that
would otherwise have risked bankruptcy. Ideally, only those banks should have been saved, and
in the same way as that used in a number of cases earlier: if the old owners cannot recapitalize
their bank, the government guarantees the survival of the bank by injecting new equity capital
and thereby takes over the bank. The old owners would thus lose their bank. After a few years, as
things have calmed down, the government sells out the bank again. This is the method used in
the case of, e.g., the Irish banks as well as the British RBS in 2008-09, and a few Scandinavian
banks in the banking crisis of the 1990s. In fact, after having resold the banks to the private
sector, the government might not have lost one cent – or at least only a fairly modest amount.10
9 For examples, see Frankel (2015) and most of the other contributors to the Baldwin and Giavazzi (2015) volume. 10 The resale value of the Swedish banks nationalized in the crisis of the 1990s turned out to be larger than the amount originally used to recapitalize the banks. But since the resale took place several years after the original capital injection, there was a (modest) net cost for the taxpayers in terms of discounted present value; see Barr and Pierrou (2015).
7
One may argue that nationalization (and later resale) of banks was not politically feasible. This
may be true for the U.S. – but hardly for Europe in 2010. In any case, this issue was never
brought up during the crisis.11
I agree that when the crisis hit the markets in early 2010, everybody was totally unprepared, and
many important decisions had to be made in a rush. In that tumult, the choices of the policy-
makers turned out to systematically benefit the bank owners. Hopefully, discussing the matter in
retrospect, like we are doing today, will help us to be better prepared next time a crisis erupts.
Fact 4: The support is a violation of the Maastricht Treaty.
The “constitution” of the EZ, i.e., the Maastricht Treaty of 1992, has two articles that were
considered as safeguards against bailing out reckless governments and banks. Article 123
prohibits ECB from lending money to a member state, and Article 125 prohibits EU institutions
and/or member states from assuming the liabilities of another member state.
But when the Greek crisis occurred in 2010, clever lawyers discovered ways to circumvent the
two paragraphs. As for Article 123, ECB did not lend money directly to Greece, but bought
Greek bonds in the secondary market, thereby driving up the price (i.e., driving down the interest
rate) so as to make it cheaper for Greece to issue new bonds. It also lent money to Greek banks,
against collateral in the form of Greek bonds, thereby making it easier for the banks to buy new
bonds.
As for Article 125, no EU institution, and no other member state, formally “assumed the
liabilities” of Greece. Instead, the EU institutions, together with IMF, lent money to Greece to
make it possible for that country to pay back its debt to the banks. Those loans were financed by
the support funds EFSF, EFSM and ESM borrowing money in the financial markets, backed by
guarantees from the member states. While this action did not violate the letter of the Maastricht
11 Ardagna and Caselli (2014) analyze the political economy of the Greek bailout. They conclude that a bailout in the form it took was a political equilibrium if the other alternative was a “disorderly” default of Greece, but that the equilibrium was Pareto-inferior to other policy measures. They do not discuss whether a disorderly default was the only feasible alternative at the time.
8
Treaty, it was certainly a violation of the spirit of the Treaty. That was an important lesson; even
articles that look iron-clad to a non-lawyer can obviously be revoked if the stakes are high
enough, and if the political will to do so exists.
Fact 5: The support measures have created animosity among the EZ members.
By shifting most of the Greek sovereign debt (to date, ¾ of the debt) from private lenders to EU
institutions and to the IMF, a complicated political situation has arisen.12 If all debt were held by
private banks, a Greek default would have caused substantial distress to some of these banks, and
Greece would have been blocked from borrowing for a decade or two. But that would have been
a purely commercial problem, not a political one.
In contrast, a Greek default today would have political repercussions. The politicians behind the
Troika cannot possibly, without losing face in front of the voters, let Greece write down its debt.
If that happened, it would be considered by the voters as merely a confirmation of what they
already suspect, namely that “we have given away hundreds of billions to the reckless Greeks”.
And the politicians are unlikely to defend themselves by saying that “we have not given away the
money to the Greeks, but to reckless bank owners” (cf. Fact 2 above). Further, the demands from
the Troika on Greece, in the form of austerity measures and structural reform, can be considered
by the Greeks as acts of unfriendly meddling with a sovereign state’s own business. Therefore,
the Germans are hated in Greece today, and the Greeks are hated in Germany. We have all seen
pictures from demonstrations in Athens where Angela Merkel is displayed in Nazi uniform. The
notion of the euro as a peace project has been quite tarnished.
If Greece had made an orderly default in 2010, there would, of course, have been problems. A
country that is blocked from international credit markets is in a difficult situation, as is
demonstrated by Argentine’s default in 2001. It would have to run a balanced budget each year,
or borrow domestically if that is possible. But the Germans would probably not have been
blamed for that, since the austerity measures would then have been decided by the country’s own
12 Just for the record, all banks that hold Greek debt are not privately-owned; some are state-owned, others are municipal, or cooperatively-owned. For brevity, I nevertheless use the term “private” for all non-Troika lenders.
9
government, rather than having been forced upon it from abroad. By moving the Greek debt
problem from the commercial arena to the political arena, the politicians created a large but non-
tangible cost for the support measures, in the form of increased animosity among EZ members.
The five facts presented above are hardly controversial. In the following sections, I will discuss
some policy conclusions that have been drawn from the events of 2010-2015 and that are today
subject to dispute among the economists. To a large extent, these policy conclusions depend on
how one looks at what actually happened.
3. Multiple equilibria: The theoretical concept
In addition to the earlier arguments for motivating support loans in 2010-2015, namely that we
must “show solidarity with the Greeks” and that we must “save the banks”, a third argument is
now emerging.
The argument is largely theoretical – although it does, of course, have practical implications. It
can be shown that, in a model of rational investors, expectations about a future default can
generate multiple solutions to the price (and thereby the interest rate) of a bond. Assume that the
probability of default is low – say, virtually zero. Then, the price of the bond is equal to the
discounted value of the future coupon payments. If instead the probability of default is high, the
expected value of future coupon payments is correspondingly lower, as is the price of the bond.
Assume now that a country issues government bonds. If the market considers the probability of
default as low, the effective interest on these bonds will be low. If, by contrast, the market
regards the probability of default as high, the effective rate of interest will be high; in fact, the
cost of borrowing may then be so high that the country’s stock of debt, which has to be rolled
over at regular intervals, is not sustainable. Thus, the probability of default becomes an
immediate reality, and the market’s expectation of default becomes a self-fulfilling expectation.
13
13 The model is spelled out in Calvo (1988) and Cole and Kehoe (2000).
10
Thus, there are two possible rational-expectations equilibria for such a country. For a given
initial level of government debt, there is one low-rate and one high-rate equilibrium, as depicted
in Figure 2 (there is also one intermediate equilibrium, but it can be shown that this rate is
unstable, so we will probably never observe it in reality).
Figure 2: Multiple equilibria in the bond market
Multiple equilibra arise in many contexts, not only in models of financial markets. For instance,
in models of crime there are often two equilibria: one “good” low-crime equilibrium and one
“bad” high-crime equilibrium.14 The policy problem is then to ascertain that if the economy has
ended up in the bad equilibrium (with high crime, or a high interest rate) it will move to the good
one (with low crime, or a low interest rate).
However, there is a difficult empirical problem associated with this largely theoretical literature.
Assume that we observe a crime rate (i.e., a percentage of the population engaged in crime) or a
government borrowing cost of, say, ten percent. How do we know that this is a bad equilibrium,
and that by suitable policy measures we can attain a good equilibrium with only five percent
14 Cf. Persson and Siven (2006).
Government
debt
Interest rate
11
crime or a five-percent borrowing cost? The ten percent we observe might actually be the low-
crime equilibrium, and by changing some parameter in the model, the system might jump to a
high-crime equilibrium of, say, twenty percent.15 In fact, we do not even know whether there are
multiple equilibria in reality. Although the figure I have drawn shows an S-shaped curve with
two stable and one unstable intersection, the curve may equally well be more stretched out and
only intersect the vertical line at one (stable) point.
While the theoretical results are unambiguous – multiple equilibria can occur – the empirical
results are, to the best of my knowledge, nonexistent. We simple do not know what kind of
equilibrium we are observing in reality; working out empirical methods for identifying the
character of a real-world equilibrium is an important topic for future research.
4. Multiple equilibria: The role of monetary policy
The notion of multiple equilibria, some of which are “good” and some are “bad”, thus raises
difficult empirical questions. Nevertheless, one may argue that in connection with sovereign
debt, the problem is simpler than in other contexts (like crime). In fact, we have cases where a
country has enjoyed a relatively low borrowing cost for a long time – and then, all of a sudden
and for no obvious reason, the interest rate skyrockets. One may then argue that this rise has
been caused by the country being pushed by some external and irrelevant factor from a good to a
bad equilibrium.
The comparison to the economics of crime is instructive. There we do not have much material
for comparison; observations are few, and if one country has a higher crime rate than another,
there are always many differences between the countries that one may point at without having to
invoke the theory of multiple equilibria. By contrast, in financial markets there is a much larger
number of observations, both across time and across countries, and the existence of multiple
equilibra may seem as good an explanation for country differences as any other.
15 Note that we cannot rely on comparative-statics results (for instance, that at the low-crime equilibrium, crime is decreasing in the severity of punishment while at the high-crime equilibrium, crime is decreasing). With two stable equilibria, as I have illustrated in the model in Figure 2, both these equilibria will have similar comparative-statics properties.
12
In fact, this is what I alluded to when I said in the Introduction that many commentators
considered the high Greek interest rates in early 2010 to be “unreasonable”. By unreasonable,
one would then mean that Greece suddenly has been pushed from a good to a bad equilibrium.
Such an argument obviously disregards the notion that the interest rate soars “for no obvious
reason”. In fact, the revelation in early 2010 of the faked Greek national accounts may be
regarded as an obvious reason.
Let us now look at the interest rates of some of the other crisis countries from 2010 on (Figure
3). We see that the Greek rates continue to increase to “unreasonable” levels in 2011 and 2012.
As for the interest rates of Spain, Portugal and Italy, there might initially have been fears of
manipulations of the national accounts in those countries too, but it was soon found out that their
national accounts were reliable – and nothing dramatic happened to their interest rates during
2010. Their spreads against the German Bund does, however, increase over the years, to a
maximum of 5 percent in 2012 for Spain and Italy, and to a maximum of 10-12 percent in 2011
for Ireland and Portugal.
Figure 3: Interest spreads of 10-year government bonds for a number of European
countries against Germany, 2010-2014. Percent. Source: Eurostat.
0
5
10
15
20
25
30
0 12 24 36 48 60
Ireland Greece Spain Italy Portugal
2010 2011 2012 2013 2014
13
Disregarding the development in specific countries, we may nevertheless discuss the theoretical
model. Therefore, assume that we somehow know that the present interest rate has been caused
by self-fulfilling, pessimistic expectations and that there is another, lower interest rate just
around the corner. This means that we have abstracted from the practical problem of
distinguishing between illiquidity and insolvency. Distinguishing between these two concepts
may be difficult in practice; I will return to this issue in section 6 below.
We thus assume that we somehow know that a country is basically solvent, but that it for some
reason has ended up in a high-interest equilibrium, with such high borrowing costs that it would
become insolvent. There is then room for the central bank to get things right. By buying
government bonds in the market, the central bank drives up the price of the bonds, thereby
driving down the interest rate – and, ideally, the “good” equilibrium can be thus attained. 16
There is a term for such a policy: it says that the central bank acts as the government’s “lender of
last resort”. In fact, many economists claim that the ECB functioned successfully as such a
lender during the EZ crisis.17
The above mainly being a theoretical argument, there is a theoretical counter-argument. Assume
that a country is in an expectations-driven, bad equilibrium. In the ideal case, an intervention by
the central bank would not affect inflation; one can show that the intervention will be
automatically sterilized. But since we are in a world of maybe not entirely justified expectations,
observing the central bank taking actions in the bond market might fuel fears of the bank
monetizing government spending and thereby creating inflation. The country might then move
from a situation with a bad equilibrium with self-fulfilling expectations about default to another
bad equilibrium with self-fulfilling expectations about inflation.
This model soon becomes very complicated and it is not obvious that one equilibrium is better
than the other. The scope for the central bank acting as a successful lender of last resort to the
sovereign is therefore rather difficult to analyze theoretically – and without a clear-cut theoretical
16 See Corsetti and Dedola (2013) for the details of this model. 17 For references, see section 6 below.
14
model, it is probably impossible to analyze empirically. There may be cases where the central
bank will improve the situation and there may be cases where it will make it worse. It all
depends on the model parameters – and we have very little knowledge about the true values of
those parameters. At the same time, there are large sums at stake among the private banks that
own sovereign debt, and one may be certain that those banks – via their lobbying organizations –
will swiftly present evidence in favor of the solution that benefits them the most.
In particular, there is one argument that might be pursued: the above model, by Corsetti and
Dedola (2013), applies to one individual country with its own central bank. In a monetary union,
with many countries and one central bank for all of them, the argument is simpler. If ECB
functions as a lender of last resort to a small member state like Greece, the effect on the union’s
inflation rate will be negligible. Even if we were sure that the ECB were monetizing Greek
spending, that would hardly affect overall inflation. Thus, a bad rational-expectations
equilibrium with high inflation is very unlikely to emerge (although we cannot dismiss it
altogether, since the model deals with expectations that are unfoundeed, but that can nevertheless
emerge and push the whole economy to another equilibrium). If we choose to abstract from
multiple equilibria in terms of inflation, but retain the notion of multiple equilibria in terms of
default risk, the policy conclusion is clear. The ECB should intervene when a small member
country has ended up in a high-interest equilibrium – still assuming, of course, that the ECB can
distinguish between illiquidity and insolvency.
5. Does the “multiple equilibria model” help us understand what happened?
How does the model of the previous sections help us understand what happened during the
crisis? Remember that the bad equilibrium can emerge for a country that is, in every relevant
aspect, identical to a country in a good equilibrium. The empirical question now is whether the
crisis countries were in every relevant aspect identical to the non-crisis countries. In Table 1, we
can see the budget deficits for all EU countries during the period 2009-2014.
15
Table 1: Budget deficits in the EU, 2009-2015
General budget deficit, percent of GDP
2009 2010 2011 2012 2013 2014 2015
Austria 5.3 4.4 2.6 2.2 1.3 2.7 n.a.
Belgium 5.4 4.0 4.1 4.1 2.9 3.1 n.a.
Bulgaria 4.1 3.2 2.0 0.6 0.8 5.8 n.a.
Croatia 5.8 5.9 7.8 5.3 5.4 5.6 n.a.
Cyprus 5.5 4.8 5.7 5.8 4.9 8.9 n.a.
Czech Rep. 5.5 4.4 2.7 4.0 1.3 1.9 n.a.
Denmark 2.8 2.7 2.1 3.6 1.3 -1.5 n.a.
Estonia 2.2 -0.2 -1.2 0.3 0.1 -0.7 n.a.
Finland 3.5 2.6 1.0 2.1 2.5 3.3 n.a.
France 7.2 6.8 5.1 4.8 4.1 3.9 n.a.
Germany 3.2 4.2 1.0 0.1 0.1 -0.3 n.a.
Greece 15.2 11.2 10.2 8.8 12.4 3.6 n.a.
Hungary 4.6 4.5 5.5 2.3 2.5 2.5 n.a.
Ireland 13.8 32.3 12.5 8.0 5.7 3.9 n.a.
Italy 5.3 4.2 3.5 3.0 2.9 3.0 n.a.
Latvia 9.1 8.5 3.4 0.8 0.9 1.5 n.a.
Lithuania 9.1 6.9 8.9 3.1 2.6 0.7 n.a.
Luxemb. 0.5 0.5 -0.5 -0.2 -0.7 -1.4 n.a.
Malta 3.3 3.2 2.6 3.6 2.6 2.1 n.a.
Netherlands 5.4 5.0 4.3 3.9 2.4 2.4 n.a.
Poland 7.3 7.5 4.9 3.7 4.0 3.3 n.a.
Portugal 9.8 11.2 7.4 5.7 4.8 7.2 n.a.
Romania 9.1 6.9 5.4 3.2 2.2 1.4 n.a.
Slovakia 7.9 7.5 4.1 4.2 2.6 2.8 n.a.
Slovenia 5.9 5.6 6.6 4.1 15.0 5.0 n.a.
Spain 11.0 9.4 9.5 10.4 6.9 5.9 n.a.
Sweden 0.7 0.0 0.1 0.9 1.4 1.7 n.a.
U.K. 10.8 9.7 7.7 8.3 5.7 5.7 n.a.
Source: Eurostat, accessed October 27 2015
16
Obviously, if multiple equilibria existed, they did not hit countries in an arbitrary fashion. Let us
first look at the entries of the year immediately preceding the crisis, i.e., 2009. We see that the
crisis countries were similar in one respect: they all had government budget deficits amounting to
around 10 percent of GDP the year preceding the crisis.18 The only other countries with deficits
of a similar order of magnitude were Latvia, Lithuania, Romania, and the United Kingdom – but
those countries had much smaller stocks of accumulated debt (cf. the debt data in Table 2).19
As for the stock of government debt, we see from Table 2 that Greece and Italy were in a
different league from the other crisis countries in 2009. Spain, Portugal, and Ireland all had
deficits amounting to 60-80 percent of GDP – which was high compared to most EU countries,
but not so bad compared to the United Kingdom, France, Germany, and Belgium. Obviously,
when assessing the sustainability of government debt, stock and flow data for only one year is
not sufficient. What really matters is also the long-term history of the country, plus the political
viability of budgetary discipline. This involves a great deal of subjective assessment, and a full
discussion of these factors is beyond the scope of the present paper.
In addition to the outbreak of the crisis in 2010, there is one episode that deserves more attention.
That is the summer of 2012, when the ECB more explicitly claimed that it was its role to move
the bond market from a bad to a good equilibrium. The next section is devoted to that episode.
18 One exception is Italy, which did not have a very large deficit in 2009 (5.3% of GDP), but which, on the other hand, had a very large debt (112.5% of GDP). Italy never received any support loans, but benefited from the ECB interventions in the bond market. 19 In fact, Lithuania and Romania also received support loans in those years.
17
Table 2: Gross government debt in the EU, 2009-2015
Gross government debt, percent of GDP
2009 2010 2011 2012 2013 2014 2015
Austria 79.7 82.4 82.2 81.6 80.8 84.2 n.a.
Belgium 99.5 99.6 102.2 104.1 105.1 106.7 n.a.
Bulgaria 13.7 15.5 15.3 17.6 18.0 27.0 n.a.
Croatia 48.1 57.0 63.7 69.2 80.8 85.1 n.a.
Cyprus 53.9 56.3 65.8 79.3 102.5 108.2 n.a.
Czech Rep. 34.1 38.2 39.9 44.7 45.2 42.7 n.a.
Denmark 40.4 42.9 46.4 45.6 45.0 45.1 n.a.
Estonia 7.0 6.6 5.9 9.5 9.9 10.4 n.a.
Finland 41.7 47.1 48.5 52.9 55.6 59.3 n.a.
France 79.0 81.7 85.2 89.6 92.3 95.6 n.a.
Germany 72.5 81.0 78.4 79.7 77.4 74.9 n.a.
Greece 126.7 146.2 172.0 159.4 177.0 178.6 n.a.
Hungary 78.0 80.6 80.8 78.3 76.8 76.2 n.a.
Ireland 61.8 86.8 109.3 120.2 120.0 107.5 n.a.
Italy 112.5 115.3 116.4 123.2 128.8 132.3 n.a.
Latvia 36.6 47.5 42.8 41.4 39.1 40.6 n.a.
Lithuania 29.0 36.2 37.2 39.8 38.8 40.7 n.a.
Luxemb. 15.5 19.6 19.2 22.1 23.4 23.0 n.a.
Malta 67.8 67.6 69.8 67.6 69.6 68.3 n.a.
Netherlands 56.5 59.0 61.7 66.4 67.9 68.2 n.a.
Poland 49.8 53.3 54.4 54.0 55.9 50.4 n.a.
Portugal 83.6 96.2 111.4 126.2 129.0 130.2 n.a.
Romania 23.2 29.9 34.2 37.4 38.0 39.9 n.a.
Slovakia 36.0 40.8 43.3 51.9 54.6 53.5 n.a.
Slovenia 34.5 38.2 46.4 53.7 70.8 80.8 n.a.
Spain 52.7 60.1 69.5 85.4 93.7 99.3 n.a.
Sweden 40.4 37.6 36.9 37.2 39.8 44.9 n.a.
U.K. 65.7 76.6 81.8 85.3 86.2 88.2 n.a.
Source: Eurostat, accessed October 27 2015
18
6. The events in the summer of 2012
The actions of ECB in the summer of 2012 can be seen as a case when the theoretical reasoning
of the previous sections was put into practice. On July 26, ECB President Mario Draghi stated
that “the ECB is ready to do whatever it takes to preserve the euro”. Immediately after that
statement, ECB launched its so-called Outright Monetary Transactions programme, signaling its
readiness to buy massive amounts of bonds of fragile countries.
The behavior of the Spanish and Italian interest rates during that dramatic summer has been seen
as an indication that these rates were actually expectations-driven in the sense of the multiple-
equilibria model, and that ECB managed to curb these expectations. Given the scale of Figures 1
and 2, it is difficult to see in detail how these rates moved. Therefore, I have drawn them in a
separate diagram, in Figure 4. 20
Figure 4: Interest spreads of 10-year government bonds for Spain and Italy against
Germany, 2010-2014. Percent. Source: Eurostat.
20 Véron (2015, p 15) has a detailed high-frequency data diagram of Spanish and Italian interest rates, with a vertical line at July 26, 2012 (the date of Mr Draghi’s “whatever it takes” speech). That diagram gives a more vivid, day-to-day picture of the events.
0
1
2
3
4
5
6
0 12 24 36 48 60
Spain Italy
2010 2011 2012 2013 2014
19
We see that both the Spanish and the Italian rates peaked in July, 2012. And then they started to
fall almost monotonically. Now, is that a convincing evidence of the multiple-equilibria model?
That the interest rates react to new information – in this case that ECB would buy large amounts
of bonds in 2012 – is hardly surprising. Such price movements are not by themselves proof that
there are multiple equilibria around; even with a single equilibrium, a major player, announcing a
new policy, can, of course, affect that equilibrium. Nevertheless, many writers have regarded the
behavior of the Spanish and Italian rates during the summer of 2012 as evidence in favor of the
multiple-equilibria model.21
As we see from Table 1, Spain had a long history of large budget deficits up to 2012, with no
tendency for the deficits to decrease. That Spanish interest rates rose in 2011 and 2012 is
therefore not a particularly convincing example of a solvent, but illiquid, country suffering from
a “bad equilibrium”. The same seems to hold for Italy. Table 3 shows that Italy started out with a
very high stock of government debt, and although the deficits were only moderate (according to
Table 2, around 3-5 percent of GDP), there is no tendency for any budget surpluses, and thus no
tendency for the debt to fall. In fact, Italy’s sovereign debt increased from 112.5 to 123.3 percent
of GDP during the period up to Mr Draghi’s speech in the summer of 2012.
Admittedly, the fall in Spanish and Italian interest rates in late 2012 may be attributed to the
launching of ECB’s Outright Monetary Transactions (OMT) programme. As pointed out above,
this is not surprising; one does not need any model of multiple equilibria to understand that a big
player buying virtually unlimited amounts of sovereign bonds will drive down their interest rates.
But it seems unconvincing to attribute the rise in those interest rates prior to Mr Draghi’s speech
as the result of multiple equilibria playing their sinister game. Considering the development of
Spanish and Italian finances during the period 2009 to 2012 (Tables 1 and 2), one does not get
the picture of basically solvent countries that for some inexplicable reason suffer from temporary
illiquidity.
21 See, for instance, Corsetti (2015), De Grauwe (2015), and Véron (2015).
20
This dismal view of the Spanish and Italian public finances is aggravated when we look at the
figures in Tables 1 and 2 for those countries after 2012. The Spanish debt has increased to 99.3
percent of GDP in 2014, and the Italian debt has increased to 132.3 percent. These are not debt
levels that we would like to associate with basically solvent countries, suffering from a
temporary bout of illiquidity. While many EZ countries have improved their debt situation, either
because of higher GDP growth or because of a budget surplus, Italy and Spain do not belong to
that group.22
Quite independently of the multiple-equilibria reasoning, there is another rationale for
intervention. If a country is hit by an asymmetric shock, and that shock is so large that the
country itself cannot handle it by the policy instruments it has at its disposal, there is a case for
risk-sharing among countries. This risk-sharing could take the form of either rescue loans,
administered by the EU funds, or by the ECB driving down the interest rate on the country’s
bonds. And this holds regardless of whether the shock makes the country insolvent (like in the
multiple-equilibria model) or is merely costly, without completely wrecking the public finances.
This argument is different from the lender-of-last-resort argument discussed earlier. In fact, it is
an argument about insurance, and it is the one behind the calls for a “fiscal union” raised in the
aftermath of the Greek crisis. But even if the fiscal union is no longer on the political agenda,
insurance in various forms is. To get this into perspective, I will discuss it under its own heading.
7. The insurance argument
The advantage of insurance is risk-sharing. But there is also a disadvantage: moral hazard.
Normally, there is always a trade-off between the two. If it is possible to design an insurance
contract striking such a trade-off between risk sharing and moral hazard that the advantages
dominate, then insurance is warranted. Otherwise it is not. 23
22 Incidentally, France and Portugal – countries that are also sometimes referred to as examples of the multiple-equilibria model at work – also show a deteriorated financial situation. 23 Needless to say, the benefits of insurance, and the problems of moral hazard, are not limited to countries that are members of a currency union. Even non-union countries might be in need of insurance – but since countries in a currency union have a closer cooperation, and the individual union members have abstained from their own monetary policy, the concept of risk-sharing seems more natural there.
21
The Maastricht Treaty of 1992 took a clear stand on this issue: no insurance is warranted. In fact,
fiscal-union arrangements in the form of loans, transfers and central-bank actions like the OMT
should not be permitted, according to Articles 123 and 125. Is there today any new information
available that might make us reach another conclusion than that of the Maastricht Treaty? In
other words, have recent events demonstrated that moral hazard is less of a problem than the
architects of the Treaty thought 23 years ago?
The Greek debt crisis in 2010 was mainly self-inflicted, and thus constituted an example of
moral hazard. 24 Although it is hard to dismiss the multiple equilibria-argument empirically, that
argument seems less reasonable for Greece than the alternative view. Most economists today
recognize that Greece was not merely illiquid, but rather likely insolvent, in 2010. 25
As for the other crisis countries, things are more complicated. One may argue that the interest
hikes on Spanish, Portuguese and Irish government bonds during 2010 and 2011 (cf. Figure 3)
may be “contagion” from Greece, and therefore not the result of moral hazard.
This calls for a more thorough discussion of moral hazard. One might say that the large budget
deficits of Ireland and Spain were not self-inflicted. Instead, they were exogenous, resulting from
the attempts by these countries to rescue their banks. These banks had run into serious trouble
because of housing bubbles that burst during the international recession in 2008-2009 – and that
recession was certainly exogenous to Ireland and Spain.
But such reasoning defies the background of their banking crises, namely the failure of national
banking supervision authorities to prevent reckless lending to the real estate sector. With a more
prudent supervision, the Irish and Spanish banking crises would have been much milder, and
would not have wrecked the public finances of these two countries. Now, it is always easy to say
ex post what a government should have done. I do not want to moralize over the actions (or non-
24 Here we disregard the rather subtle philosophical question of whether the actions of a corrupt and incompetent government are beyond the control of the voters in a democracy, and thus the voters should be insured against such a government. 25 Cf. the contributions to the Baldwin and Giavazzi (2015) volume.
22
actions) of the Irish and Spanish governments,26 but I do want to point out that the view of the
banking crises as exogenous shocks to the Irish and Spanish public finances can be questioned.
Incidentally, this also holds for some of the smaller crisis countries not discussed here, for
instance, Cyprus.
When looking at the figures in Tables 1 and 2, and considering the background actions and non-
actions of the governments in question, it is hard to see that moral hazard can be altogether
dismissed as an explanation. Thus, the question is whether new information in favor of insurance
really has been forthcoming. Such information should show that the fear of moral hazard among
the architects of the Maastricht Treaty was unfounded. I find it hard to see that this is the case.
Thus, the policy conclusion of the Maastricht Treaty, namely that no insurance is warranted, has
not been obviously proved wrong by the events of 2010-2015. Note, however, that it has not
been proven right, either. Those who want to maintain the Maastricht view can claim that all
interest hikes in the crisis countries in the period 2010-2015 are the result of moral hazard. On
the other hand, those who want to discard the “no-bailout principle” can claim that the crisis
countries (except Greece) were simply illiquid, that they were innocently either hit by real-estate
bubbles and/or pushed to a bad equilibrium, and that no moral hazard has been at work in 2010-
2015.
The matter is thus complicated and can probably not be definitely resolved. Discussing the
evidence over and over again is, however, necessary – if not for other reasons than that it will
make us intellectually more prepared when the next crisis erupts. If we are more prepared next
time than we were in 2010, we will probably be better at handling the skillful lobby groups that
try to affect policy decisions in their favor.
26 Strictly speaking, bank supervision in Spain and Ireland was the task not of the government, but of the Banca de España and the Bank of Ireland.
23
8. Concluding comments
There are many more matters concerning the euro crisis that warrant a discussion – for instance,
the so-called “doomed loop” and the use of risk weights in the portfolios of the banks,27 the need
for higher capital requirements in banks,28 and the new banking union that is being implemented
right now.29 All these issues are interesting and important, but trying to squeeze them into my
lecture would lead too far. I would like to end with some reflections on the new role assumed by
the ECB when launching the OMT in the summer of 2012.
When taking the stand that Italy and Spain were solvent but illiquid, the ECB took a great risk.
The budget situation of these countries does not seem to have improved after the OMT.
Therefore, it may turn out that the ECB has wasted money on failing countries instead of just
helping illiquid ones to pass a temporary problem.30
By launching the OMT programme, the ECB also opened up for other countries to ask for help,
claiming that they are not insolvent, either – or at least not more insolvent than Spain and Italy.
This may be a source of conflict within the eurozone in coming years, in particular if the debt
situation does not improve. A particularly interesting case is that of France, which is not only an
important country in its own right, but also a country that quickly approaches an indebtedness
level of 100 percent (Table 2).
One may guess that the ECB is following the development of Spain and Italy, as well as of
France and Portugal, with great interest. A recovery may improve the debt/GDP ratio – but a
recovery will also create demand for government spending on social services that have been
neglected during the years of austerity. A spell of inflation may help reduce the real value of the
debt, and we may guess that the ECB will not be too eager to curb future inflation in the
eurozone. An inflation rate somewhat above the target of two percent will reduce the risk of a
27 Cf. Gros (2013). 28 Cf. Admati and Hellwig (2013). 29 Cf. Véron (2015). 30 Every year, the ECB pays a dividend to the treasuries of EZ countries. That dividend basically comes from the seignorage, and capital losses on bonds reduce the amount the ECB can pay in dividends. Thus, such capital losses are ultimately borne by the tax payers of the EZ countries.
24
major embarrassment to an ECB that has claimed that it is able to distinguish between
insolvency and illiquidity.
Table 3: Gross government debt in
the entire euro zone (19 countries), 2006-2015
Year Debt, percent of GDP
2006 67.3
2007 64.9
2008 68.5
2009 78.3
2010 83.8
2011 86.0
2012 89.3
2013 91.1
2014 92.1
2015 n.a.
Source: Eurostat, accessed 2015-10-27
This issue becomes even more intriguing when one looks at the debt/GDP ratio of the entire
Eurozone – see Table 4. During the last decade, the indebtedness has increased almost
monotonically from 67.3 to 92.3 percent of GDP. And this in an area where, according to both
the Maastricht Treaty of 1992 and the Growth and Stability Pact of 1997, the critical level for
sustainability is 60 percent – a level that not even Germany has been able to keep.
The prestige invested by the ECB in its ability to distinguish between insolvency and illiquidity,
and the difficulties of the EZ countries to keep debt at sustainable levels, might have implications
for what inflation rates the financial markets will expect during the next decade. Thus,
complicated mechanisms of inflation and default, of the kind discussed in section 4 above, might
be set in motion, and we may witness higher nominal interest rates in the years to come –
accommodated or not accommodated by the ECB allowing higher inflation.
25
However, let me end in a positive tone. Everything does not look entirely dark. First, it is not so
surprising that indebtedness has increased since 2006. The last decade has not been a normal
period, but a period of repeated crises and prolonged recession. Once we are back to more
normal times, we may witness higher GDP growth rates and improved public finances. This does
not, however, mean that all individual countries will be out of danger. The fact that so many
countries had failed to build up reserves during the good years, and thus were so vulnerable
when the crises started in 2007, is a bad sign.
Second, many of the EZ countries have not only witnessed austerity and contraction, but also
structural reform. This holds for Greece, Spain, Portugal, Ireland and Italy and may lead to
higher growth rates in the future. Still the lack of reform in France is problematic – the question
is whether just a “normal” business-cycle recovery will be sufficient for French public debt to
fall as a percentage of GDP.
26
References
Admati, A. and M. Hellwig (2013), The Bankers’ New Clothes: What's Wrong with Banking and What to Do about It, Princeton University Press, Princeton, N.J.
Ardagna, S. and F. Caselli (2014), “The Political Economy of the Greek Debt Crisis: A Tale of Two Bailots”, American Economic Journal: Macroeconomics, Vol. 6 No.4, pp 291-323. Baldwin, R. and F. Giavazzi, eds. (2015), The Eurozone Crisis – A Consensus View of the Causes and a Few Possible Solutions, CEPR Press, London (can be downloaded from www.cepr.org) Barr, D. and H. Pierrou (2015), “Vad blev notan för 1990-talets bankstöd?”, Ekonomisk Debatt Vol. 43 No. 5, pp 41-54. Bénassy-Quéré, A. (2015). ”Maastricht flaws and remedies”, in Baldwin and Giavazzi (eds.) (2015). Calvo, G. (1988), ”Servicing the Public Debt: The Role of Expectations”, American Economic Review, Vol. 78 No. 4, pp 647-661. Cole, H.L. and T. Kehoe (2000), “Self-Fulfilling Debt Crises”, Review of Economic Studies, Vol. 67 No. 1, pp 91-116. Corsetti, G. (2015), “Roots of the Eurozone crisis: Incomplete development and imperfect credibility of institutions”, in Baldwin and Giavazzi, eds. (2015). Corsetti, G. and L. Dedola (2013), ”The Mystery of the Printing Press – Monetary Policy and Self-fulfilling Debt Crises”, CEPR Discussion Paper 9358. De Grauwe, P. (2015), “Design failures of the Eurozone”, in Baldwin and Giavazzi, eds. (2015). Frankel, J. (2015), “Causes of Eurozone crises”, in Baldwin and Giavazzi (2015). Gros, D. (2013), “Banking Union with a Sovereign Virus”, CEPS Policy Brief No. 289, March 2013. The Guardian 2015-07-29: “Where did the Greek bailout money go?”
27
Liikanen, E. et al. (2012), High-level Expert Group on reforming the structure of the EU banking sector, EU Commission, Brussels. Mouzakis, Y. (2015), “Where did all the money go?”, MacroPolis: http://www.macropolis.gr/?i=portal.en.the-agora.2080 The New York Times 2015-07-30: “A Flow of Funds, Out of Greece”.
Persson, M. and C.-H. Siven (2006), “Incentive and Incarceration Effects in a General Equilibrium Model of Crime”, Journal of economic behavior & organization, Vol. 59 No. 2, pp 214-229.
Véron, N. (2015), Europe’s Radical Banking Union, Bruegel Essay and Lecture Series, Brussels (can be downloaded from www.bruegel.org).
Zettelmeyer, J., C. Trebesch and M. Gulati (2013), “The Greek debt restructuring: an autopsy”, Economic Policy, July 2013, pp 513-563.
