Exiting EMU: not surprisingly, the Maastricht Treaty contains no exit procedures, but bailing out would not be that difficult.

The issues of how a country could withdraw from EMU or how EMU might be totally dissolved are crucial in any serious analysis of the euro's future and its risks as an international investment currency. Unsurprisingly, no insights can be gained from a reread of the Maastricht Treaty. The founders and their functionaries had no inclination to spell out exit provisions. Their hard sell depended on demonstrating the benefits--economic and political--from eternal not provisional monetary union. Most probably, if they thought about it at all, they viewed any reversal as presenting huge technical problems and only possible at frightening cost. From their perspective, a pull-out from EMU would be a serious treaty violation, threatening to throw Europe back into a dark age of competing nationalisms.

In the five years since the euro was launched, speculation on how it might disintegrate or shrink has remained confined to the dark recesses of the market's mind. Yet history is full of treaties being re-negotiated or broken--and so why not the Maastricht Treaty? Under imaginable circumstances in which the European Central Bank became deeply unpopular (major revealed mistakes in the conduct of monetary policy combined with arrogance, aloofness, or even corruption), a reappearance of national monies might actually save the process of European political and economic integration. Surely the remarkable advances in information technology and their application to financial engineering could facilitate a possible reversal of European monetary union at only moderate overall cost.

Consider as a first scenario a present member of medium economic size, say Holland, contemplating withdrawal from monetary union. There are in effect two exit possibilities--phased or sudden. A phased withdrawal could take place in three phases over several years. In the first stage, starting in say January 2006 and ending in December 2007, the Netherlands central bank would stand ready to convert the euro deposits which Dutch banks hold with itself into a new currency, the florin, on a 1:1 basis (and conversely). As yet no banknotes in the new currency would be issued. All new bond issues by the Dutch government would be in florins (rather than euros). The Dutch central bank could obtain on request re-denomination of Dutch government bonds in its portfolio from euros to florins.

The new currency (florins) would be acceptable on a 1:1 basis in settlement of all amounts due in euros to government departments and public utilities. Many businesses, including retail outlets, would most likely also accept florins on the same basis. The government and public sector enterprises might start paying for supplies only in florins. Banks would be obliged (by law) to re-denominate all deposits and loans into florins if requested to do so by Dutch customers. Re-denomination of long-term fixed-rate deposits or loans would take place on the basis of present market value and the setting of a new market interest rate for the residue of the fixed term.

Interest rates on euros and florins for terms up until the end of 2007 would be identical, but for longer terms the rates could differ, depending on expectations about re-born sovereign Dutch monetary policy into Stage 2 and beyond. In the late weeks of 2007 short-term interest rates extending over the year-end (into Stage 2) could become highly volatile and reach large positive or negative levels--depending on expectations regarding the opening level of the pound against the euro. Banks would strive to balance their positions in florins (no open positions of size) by end-year.

At the start of Stage 2 (January 1, 2008), 1:1 convertibility of euros to florins at the Dutch central bank would be suspended indefinitely. A fixed exchange rate system would be introduced for the florin against the euro--say with a fluctuation limit 5 cents in either direction from parity. The announced intention would be for a free float to start at the end of Stage 2 (December 2009). Wages of employees in the public sector and public utilities, all social welfare entitlements (and pensions), and tax bills, would now be denominated and payable in florins. (As a variant, this shift could have occurred already late in Stage 1). All real estate rental agreements between Dutch residents would be re-denominated into florins.

All pricing by the public sector and public utilities would be in florins. (For the first year of Stage 2 the public sector and utilities would still accept euros in settlement of florin invoices, with translation made on the basis of a reference exchange rate for the relevant period). The Netherlands central bank would issue florin banknotes on a 1:1 basis on demand against florin deposits with itself and conversely. Private retail outlets would be obliged (during Stage 2) to quote prices in florins and in euros (dual pricing), using the current "mid-point" exchange rate. Wages in the private sector could be paid in euros or florins, depending on individual agreement (the assumed tendency, except in the export sector perhaps, would be towards florins).

Gradually through Stage 2 households would dispose of their euro banknotes to a large extent in exchange for florin banknotes--in anticipation of Stage 3 when mandatory requirements for dual pricing would come to an end (and florins would be the conventional denominator for most retail trans actions). In effect, the euro banknotes being disposed of would be credited to euro deposits and sold in the currency markets for florins. The euro banknotes from Holland would end up as the source of additional deposits held by member banks with the central banks still remaining in monetary union and so adding to the amount of high-powered money circulating in that truncated area.

The ECB would be obliged--in the absence of an overriding agreement between the Netherlands and the European Union--to mop up the excess monetary base by selling bonds out of its portfolio. Hence the seigniorage enjoyed by European governments still in monetary union would decrease. In practice, Holland would surely agree (as part of an exit deal) to buy back the flotsam of notes (which would have found their way into central bank vaults in the truncated union as a counterpart to deposit growth). Technically, the Dutch government would issue (florin) bonds to its central bank and sell the florin proceeds in the foreign exchange markets for euros with which to buy the notes (from the central banks still in EMU). The Dutch government would present these (the redeemed notes) to the Dutch central bank in cancellation of euro-denominated Dutch state loans still in its portfolio.

At the start of Stage 3, the florin would be floated freely. The Dutch central bank would almost certainly regard the florin-euro exchange rate as a key variable in the setting of monetary policy. The whole proceedings from start to finish would be smoothest where there were strong and widespread expectations that the florin would trade at around parity with the euro both during Stage 2 and beyond. During Stage 1, Holland might still remain formally a member of European Monetary Union (participating in meetings, etc.). Into Stage 2, Holland would have made a definitive exit, with the euro's continuing use in the Netherlands being as a parallel currency.

The staged withdrawal is not suitable where the purpose of exiting EMU is to achieve an immediate substantial change in the Dutch price level relative to that in the euro area (as would be the case if the aim were to escape a deflationary depression). A sudden withdrawal would be effected by the Dutch government freezing euro deposits held by Dutch banks with the Netherlands central bank (except for that part which represents reserves against non-resident euro deposits with Dutch banks). Dutch banks could no longer transfer resident euro deposits (now called "Dutch euros") on a 1:1 basis into foreign euro deposits. The 1:1 convertibility of euro banknotes into resident euro deposits (or conversely) at Dutch banks would be suspended.

Dutch euros could be used freely to settle euro-denominated accounts between Dutch residents (for example, purchases and sellers of goods and services) and to repay euro loans by Dutch banks to Dutch residents. There would be a new category of euro deposit, termed "free," which Dutch residents could open and funds from these would be transferable on a 1:1 basis into non-resident euro deposits and foreign euro deposits. Euro banknotes would remain convertible on a 1:1 basis against free euro deposits (but not Dutch euro deposits). An exchange market would start up between Dutch euros and free euros (or foreign euros).

In the case of Dutch euros falling in value versus the free euro (as where the purpose of leaving EMU is to escape deflation), euro banknotes would go to a premium against Dutch euros (where the premium would be determined by the exchange rate between Dutch and free euros). Dutch residents in obtaining cash from automatic teller machines would be levied extra charges based on the rate of premium. In the retail economy, dual pricing would apply--one price for payment by check or card and a lower price for payment by cash.

Eventually, the Dutch euro deposits would be re-denominated as florin deposits and euro loans by Dutch banks to Dutch residents as florin loans. Newly printed florin banknotes would be obtainable on a 1:1 basis against the florin deposits. Dutch residents would freely determine when to dispose of their euro banknote holdings and acquire florin banknotes. As this process got underway, central banks in the euro-system would find themselves with a flotsam of notes as in the staged withdrawal process, and this would be dealt with similarly (under a seigniorage agreement between Holland and the EU).

But unlike for the staged withdrawal, the switch to the new currency would have occurred largely by decrees. These would not apply to non-residents of Holland. Thus, bank deposits and other debts outstanding to non-residents in euros for example would not in general be subject to forced conversion. Dutch government bonds, however, and other bonds issued in the domestic market, would be payable in Dutch euros and later florins, even if held by non-residents.

Exit by sudden decree would create windfall losses or gains most particularly for Dutch financial intermediaries. Dutch resident deposits in euros with Dutch banks and Dutch bank loans to Dutch residents would not have been in balance except by chance at the time of exiting EMU. If deposits are well below loans and the Dutch euro falls immediately against the euro, then the banks would experience net loss. There is also the question of interest rate exposure.

Long-term interest rates might jump immediately on Dutch euros and inflict loss on banks' holdings of long-maturity loans at fixed interest rates. The loss would be diminished by the presence of long-maturity fixed-rate liabilities also transformed into Dutch florins (for example Dutch bank bonds or long-term deposits/savings plans outstanding). Some prior interest rate hedges, however, would be ineffective. For example, euro interest rate swaps entered into originally to protect profits against unexpected euro rate rises would not be helpful in

the case of a change in monetary regime, on the assumption that the decrees did not include a re-denomination of swap contracts.

As part of the exit plan, the Dutch government would compensate Dutch financial intermediaries for windfall losses in excess of a stipulated percentage of capital. There might also be a windfall profits tax on more fortunate institutions. It is likely that the financial intermediaries would also have suffered some loss in the inevitable period of speculation ahead of the exit from EMU. At that time, Dutch banks would have scrambled to retain Dutch resident euro deposits fleeing for safety into foreign euro accounts (where they could not be forcibly transformed into Dutch euros).

The premium interest rate that the banks would have then offered might have been offset only in part by loan rate surcharges passable on to some Dutch borrowers. Indeed, rising losses for the banks during the period of speculation on exit might force an emergency withdrawal, which in principle could be reversed later if indeed the democratic process brought a decision in favor of continued membership in EMU. In practice, reversal would depend on the goodwill of Holland's EMU partners. At best they would accept Holland back as a full EMU member almost immediately with Dutch euros re-converted into ordinary euros at a 1:1 rate.

The diplomacy of withdrawal for a medium- or small-size country is fundamentally different from that for e largest country, Germany. Many if not all of the other countries might well decide that they would exit together with Germany rather than remain part of a rump monetary union without Europe's largest economy. Diplomatic negotiations would be fully multilateral, with possibly several countries taking lead roles. In particular, pre-EMU history suggests that the Netherlands and Belgium would follow Germany and anchor themselves to a re-incarnated Deutsche mark. France might seek to salvage a Latin Monetary Union (France, Italy and Spain) extended to include Austria and the new EU countries in Central Europe.

Suppose negotiations ended in a general dissolution of EMU. There would have to be a multilateral agreement about how to deal with euro banknotes outstanding and non-resident euro deposits and loans in each member banking system (on the assumption that decrees re-creating national monies apply only to resident clients).

The simplest procedure would be to re-incarnate the ecu (the basket currency in which EU monies were weighted roughly according to economic importance and which was converted into euros on a 1:1 basis at the start of EMU). The new ecu, however, would consist just of the new monies of the ex-members (not, for example, pounds sterling). Euro banknotes would not be convertible on a 1:1 basis into bank deposits in the new national monies but into ecu deposits. The ex-member governments would join together to redeem on demand the euro-banknote circulation--each supplying their own national currency in the appropriate proportions. Non-resident euro deposits and loans in each banking system would be converted 1:1 from euros into ecus.

Of course, a serious exit bid by Germany could be the catalyst to agreement on major reforms of EMU rather than to total disintegration. As part of the diplomatic brinkmanship, France and Germany might threaten to exit jointly and form a new monetary union. The euro is in practice like a solution that can be distilled into a range of national components and smaller union currencies. The distillation process imposes economic and financial costs--greater when quick than slow--but these are not obviously always larger than the economic benefits gained over the long run (for the separating country or countries). There is no egg that has to be unscrambled.

No Way Possible?

History is full of treaties being re-negotiated or broken--and so why not the Maastricht Treaty? Under imaginable circumstances in which the European Central Bank became deeply unpopular (major revealed mistakes in the conduct of monetary policy combined with arrogance, aloofness, or even corruption), a reappearance of national monies might actually save the process of European political and economic integration.

Brendan Brown is director and head of research, Tokyo-Mitsubishi International plc., London, and author of Euro on Trial: To Reform or Split Up (Palgrave, March 2004).