Monday, January 31, 2011

Could an Obscure Loophole Send the Euro the Way of the Ruble?

Could an obscure loophole known as emergency liquidity assistance (ELA) lead to the collapse of the euro area, much as the post-Soviet ruble area collapsed in 1991-1993? Some people seem to think so. The Irish Independent says that use of ELA by the Irish central bank amounts to "printing its own money." Tracy Alloway, writing on, emphasizes the secret, hush-hush nature of ELA operations. One blogger goes so far as to speak of hyperinflation. Is there really something fishy going on? And what does ELA have to do with the ruble?

Last summer I wrote a post about the breakup of the short-lived ruble area that existed among the 15 successor states of the Soviet Union from 1991 to 1993. The ruble area had several flaws that made its demise inevitable. One of them was the fact that the Central Bank of Russia, as the lead institution within the currency area, retained less than full control over monetary policy. Although it had a monopoly on the issue of paper currency, it allowed national central banks almost unlimited latitude to create bank credit. The national central banks could use bank credit to finance their own government deficits and make loans to state-owned industries. The inflationary consequences of doing so were spread among all 15 ruble area members, creating a serious free-rider problem.

In the earlier post, I saw fiscal policy free riders as a real threat to the euro, as they had been to the ruble, but at the time, I did not see the possibility of a spillover into monetary policy. Now it appears that things are not quite so simple, as a post on Open Europe Blog first called to my attention.

Tracking down the story down led to a research note by Willem Buiter that gives a detailed explanation of emergency liquidity assistance  in the euro area. To make a long story short, it turns out the the national central banks (NCBs) of euro-area members retain lender-of-last-resort powers in at least partial independence from the European Central Bank (ECB). For example, the Irish central bank can loan liquid reserves to Anglo Irish Bank, taking securities as collateral. That causes total reserves of the euro-area banking system increase. If those reserves are used as a basis for further loans, the euro money stock increases. Under the right conditions, the increase in the money stock could cause inflation throughout the euro area. Add in the fact that ELAs are, as a matter of policy, conducted behind a veil of secrecy and it becomes tempting to make a leap to a parallel with the ruble.

But hold on a minute. The whole story is overblown. What is going on with euro ELAs is far different from what brought down the ruble area--so far, at least.

The first big difference is that the ECB has the tools it needs to override any monetary impact of ELAs by national central banks. For one thing, the ECB can use any of its usual policy instruments to withdraw an equivalent quantity of reserves from the euro banking system, thereby sterilizing the effect of ELAs. In addition, although NCBs do not have to report ELAs to the public, they do have to report them to the ECB, which can veto them by a two-thirds vote of its board. True, as Buiter points out, the ECB may choose not to sterilize or veto the monetary effects of ELAs, but presumably, that would either be because it regards their effects as minimal, or because it considers the resulting monetary expansion to be consistent with euro-wide monetary objectives.

The second difference lies in the motivation for liquidity creation by national central banks. In the case of the euro, ELAs are supposed to be made for lender-of-last-resort purposes. As such, they are subject to rules first outlined by Walter Bagehot a century and a half ago. Such loans must be made only to institutions that are illiquid but not insolvent, they must be made on good collateral, and they must be made at a penalty interest rate. Although ELAs made under these conditions do add reserves to the banking system, their purpose is not to bring about an expansion of the money stock. Instead, it is to prevent an undesired contraction that might occur if a solvent but illiquid bank were forced to close its doors. Far from having inflationary potential, properly executed lender-of-last-resort operations contribute to the goal of price stability.

It is true that central banks are sometimes tempted to fudge on matters like collateral and solvency. Some observers suspect the Irish central bank of doing so in some of its recent ELAs. Still, if a national central bank were to violate the lender-of-last-resort principles too egregiously, ECB rules would allow it to veto the operation

The monetary manipulations of NCBs in the post-Soviet ruble area were very different. The post-Soviet NCBs began life as the republic branches of Gosbank, the central bank of the USSR. Gosbank was a monobank that combined central banking and commercial banking functions. As such, it held retail deposits and made loans both to the government and to state-owned industries. When Gosbank broke up, at least some of the NCBs in the post-Soviet ruble area continued those monobank practices. They extended credits directly to local governments and made generous loans to often-failing state enterprises. Those operations were inherently inflationary. When local enterprises used the funds borrowed from their local NCBs to make purchases from firms in other former Soviet republics, and those payments were cleared through the banking system, the inflationary pressures spread throughout the ruble area. The ability to finance national government and state enterprise operations in this way created a free-rider problem on an altogether different scale than anything associated with ELAs in the euro area.

In short, there is no real parallel between emergency liquidity operations as currently conducted by NCBs in the euro area and the inflationary machinations of NCBs in the ruble area. One has to stretch the imagination to see how ELAs could  pose an inflationary threat to the euro. Something like one or all of the following would have to happen:
  • The NCBs of the euro's biggest members would have to get involved. Buiter estimates that Ireland's central bank has extended ELA credits equal to 31 percent of the country's GDP, but Ireland is a small country. If Germany were to act on the same scale, then ELAs might start to matter.
  • ELAs would have to break through the limits set for lender-of-last resort functions in a big way. Ireland may have put a toe over the line in treating banks like Anglo-American as solvent or by accepting less than gilt-edged collateral, but something more than that would be needed. We might start to be worried, for example, if NCBs were to make loans to banks in the guise of liquidity assistance, and then require the banks to use the proceeds to purchase local sovereign bonds on terms better than those offered by the open market.
  • If several NCBs combined to form a blocking coalition of one third plus one on the ECB board to prevent veto of excessive ELA activity, then monetary expansion inconsistent with ECB policy objectives would become possible.
At present, there is no sign that any item on the above list is happening or about to happen. The euro does have its problems, but emergency liquidity assistance operations do not yet add greatly to them.

Follow this link to view or download an updated slideshow on the breakup of the ruble area with comparisons to the present condition of the euro.


  1. Typo alert: you mean Anglo-Irish (the insolvent bank) rather than Anglo-American.

    Very good post. ELA concern seems overblown. First usage in recent history was Germany lending to Hypo Real Estate in 2008. Truly emergency. One question that is not clear is whether the resulting deposits at CB of Ireland are transferrable outside Ireland. If not, then this is not really a net liquidity add except in the Irish banking system, and we should only concern ourselves in the net position of Irish banks relative to the ECB. They currently take down around EUR 140bn (ish).

  2. If Euro country get panic on currency problem, how much more in the third world countries? If Irish Central Bank use ELA for printing its own money, maybe some other European countries can do also.