Gavyn Davies on the Swiss central bank and why it folded

The SNB balance sheet at the end of December was about 85 per cent of GDP, mostly in foreign currencies, and we do not know whether this has increased markedly during the bout of euro weakness in January. The SNB’s mark-to-market currency losses on Thursday were probably around 13 per cent of GDP (SFr75bn). Paul Meggyesi of JPMorgan says that “the SNB would have been bankrupted by this de-pegging had it not made such a large profit last year”. The SFr38bn profit in 2014 was announced only last week, which is surely not a coincidence.

Many economists believe that balance sheet losses are irrelevant for a central bank, so they should play no role in policy. But the SNB is 45 per cent owned by private shareholders, many of whom are individuals, who receive dividends from the SNB. The rest is owned by the cantons, which have been complaining recently about insufficient cash transfers from the SNB.

This ownership structure contrasts sharply with most other central banks, which are in effect government departments, wholly owned by the treasury and therefore the taxpayer. The Swiss set-up makes the SNB particularly concerned about balance sheet losses, especially since disgruntled citizens can directly force changes in monetary and reserves policy via referendum.

Excellent points, there is more in the FT here.  Here is my post earlier today on whether central banks require capital, financial, political, or otherwise.


Explain to me again how the SNB can lose money buying foreign currency assets (as long as THEY do not bear negative interest) with Franck that cost them nothing to produce. Private sector investor have 45% of that deal and they are complaining. May I please be the SNB for a few weeks?

They lost money not by creating the franc but by buying the euro assets.

I am having trouble following the logic as well. The SNB bought Euros using francs, which it can create in infinite supply. It lost money *only* when and because it abandoned the peg. If I borrow money from you and credibly promise *not* to pay you back, how can I lose unless I abandon my promise to default?

Remember, the SNB was trying to *de*value the franc. I understand why a central bank may be limited in its ability to keep its currency value *above* a floor. But, it seems strange to argue that a bank cannot keep its currency value *below* a ceiling. I guess Krugman was right when he coined the phrase "promise to be irresponsible". Apparently, the SNB had a problem of being too credible, so credible that no one believed that they would be irresponsible.

The SNB did NOT print francs but simply "spent" the cache of francs on deposit.

You do not keep the prices of imported food from inflating by inflating the price of Swiss produced food by printing huge quantities of money.

You need to consider the source of the deposits in the SNB.

To prevent inflation in imports prices, the SNB did not convert deposits make effectively in foreign currency.

You are a Greek citizen with cash in Euro fearful it will be converted overnight into drachma, so you convert your Euros to Swiss F and deposit the Swiss F in a Swiss bank. With hundreds of thousands doing the same, the price of the Swiss F must go up to balance the currency exchange....

...unless the SNB buys Euros and other currency and holds them.

The SNB is doing this to prevent inflation in import prices, but solving the problem by printing money and inflating Swiss prices on goods, services, and assets is even worse.

SNB tried to stop the conversion of Euros to make deposits with currency controls and by charging a fee to have a deposit in the SNB - negative interest, but it could not set the negative interest high enough to reverse the inflow of cash without harming the Swiss economy.

The Swiss just need to accept the fact that lots of people want to own part of the Swiss economy and that means the Swiss get screwed as long as the Swiss are unwilling to be isolationists.

The assumption is that they will eventually have to sell the assets to maintain an inflation target. The larger your exposure, and the riskier the assets, the greater the risk. Would you want to be sitting on a trillion Euros of European debt?

I still think one needs to consider that the initial depreciation didn't accomplish any macro goals -- Switzerland still had deflation. It makes holding those balances hard to defend.

Dude, you gotta stop it with the reading of the tea leaves.

So, basically, the SNB is a poorly-designed institution, since it doesn't answer to the same people it's supposed to serve to.

No country was ever ruined by sound money. The peg was destroying a brand built up over a century. The SNB surely knows that Francs are a more important Swiss product than chocolate or watches -- a pillar of their banking industry. Switzerland is a small country with a big currency; the Franc is a thing to be nurtured carefully and not something that needs to follow the Euro down though its Draghian circles of inferno.

Another point: This "percentage of GDP" nonsense needs to stop. Apples are never any percentage of oranges.

No country was ever ruined by sound money

You may have heard of it, moron.

Good to know the country was ruined.

If you think that the Swiss franc was undervalued and likely to become more so when the Euro becomes undervalued, the policy makes perfect sense.


The SNB lost money with its de-pegging from the perspective of those who own the SNB, some citizens and the cantons. They expect to receive their dividends and transfers in Swiss francs. The value of the foreign assets they hold fell when they de-pegged, thus damaging their balance sheet and making it harder for the SNB to make those payments.

And to those who say that the SNB can just print an infinite amount of CH francs, in principle they can, but as pretty hardened monetarists, they will not do so.

So the SNB isn't a central bank, but a group of old Swiss men with savings accounts that occasionally meddle in forex trading, trying to manipulate the market in their favor. Good to know, and quite the safe haven.

Probably Nazis too.

Even putting the unconventional ownership of the SNB aside, the idea that balance sheet losses are "irrelevant" for a central bank is a fallacy, and an alarmingly persistent one at that.

Holding monetary policy fixed, the marginal liabilities of a central bank must pay the market interest rate. A central bank's ability to create money at will isn't much better in this regard than the US Treasury's ability to issue T-Bills at will. Of course, the central bank controls the short-term market interest rate -- but this control is used to achieve macroeconomic objectives, and if it's used to manipulate the cost of central bank liabilities instead, some of the traditional objectives must be sacrificed. After all, the Fed could make phenomenal profits by holding its current portfolio to maturity and refusing to raise interest on reserves above 0.25% for the next decade, but it won't -- because that would involve keeping market interest rates near zero too, and (excluding the possibility of secular stagnation) if continued for long enough this would create an inflationary boom.

Of course, undercapitalized central banks can usually survive, but this is because they have a very valuable implicit asset that doesn't appear on the official balance sheet: a perpetual monopoly on the issue of zero-yield paper currency, which offers a healthy net interest margin on part of the portfolio. This stream of future income is typically enough to overcome any losses, but that doesn't make the losses illusory: as the central bank retains income to recapitalize itself, it will distribute less to the government and shareholders. With a big loss, this might take some time -- and with a sufficiently big loss, things might never return to normal. Holdings of Swiss paper currency are about 10% of GDP, which at an average nominal rate of 3% provides an annual flow net interest income equal to 0.3% of GDP. This isn't that large compared to a capital loss of 30% or 40% of GDP, which the SNB could easily have suffered had it continued the fight for another year or two before giving in.

Yes, there are other ways for central banks to make their liabilities cheaper: for instance, they can raise reserve requirements and pay low or zero interest on those reserves. But this is ultimately just a tax on certain forms of financial intermediation -- not necessarily a very efficient tax, or the optimal way for a state to raise revenue at the margin. And in its ability to levy this tax, the central bank is just a weaker version of the central government, which can levy a much broader range of taxes. No one claims that the ability to tax renders conventional fiscal losses or debt irrelevant; it's puzzling that similar claims are so often made about central banks. Ultimately, the simple reality is that a loss is a loss, both for central banks and for any other branch of government -- and while it's possible to get confused about this in all the meandering rhetoric about balance sheets, you can't miss it in any full accounting.

While I'm sympathetic to many of the SNB critics, I think they suffer from a lack of imagination about what the peg would mean going forward.

Most importantly, it's possible that in the long run, the Swiss real exchange rate simply needs to appreciate. Switzerland's current account surplus is immense and has averaged over 10% of GDP in the last decade. There are various reasons why the headline figure may be an overstatement, but by any measure the balance is high -- and it is putting Switzerland, which already has one of the highest net international investment positions in the world relative to GDP, on track for even greater accumulation. At some point, this will probably need to reverse itself, and this reversal will probably involve a change in relative prices. (The notion that a reversal can come solely through a change in Swiss consumption demand is what Krugman is fond of calling the "fallacy of immaculate transfer".) In fact, this shift in relative prices may be quite sizable: the SNB president once defended the peg by arguing that the Swiss current account is little affected by the exchange rate, but of course if this is true it implies that a larger relative price change is necessary to achieve any given shift in quantities -- with proportionately larger consequences for the SNB's balance sheet.

Now, am I certain that Swiss real exchange rate must eventually appreciate? Absolutely not. But it's a plausible outcome, with a probability well above zero, and if you're the SNB that means you need to consider it.

Imagine the following scenario. In a parallel universe, the SNB continues to avidly defend the peg for several years, even as the need for real appreciation becomes increasingly clear. Its foreign exchange portfolio continues growing, to the point where it reaches 200% or 300% of GDP. The awkwardness of the SNB's position gradually dawns on voters and policymakers: if it gives up now, it will incur massive losses, of magnitude possibly greater than the government's entire preexisting debt.

On the other hand, if it maintains the nominal peg, then the adjustment must take place through nominal prices instead; a 30% real appreciation, spread over 10 years, would imply 3% additional inflation each year. For a central bank that traditionally aims for price stability, this is an awkward situation indeed; can it delicately engineer a temporary spurt of inflation followed by a return to normal? Not to mention that a gradual real appreciation is unnatural: if the expected path of nominal interest rates increases in line with the expected path of inflation, then markets will produce an *instant* nominal appreciation once the policy becomes known. The only way to avoid this is to keep the path of expected nominal rates low, allowing inflation to push down expected real rates. (And in the process keeping the SNB's net interest margin positive; otherwise, if the cumulative 30pp greater inflation is reflected in higher nominal rates, then the SNB's cumulative financing costs will be 30pp higher too over the course of the adjustment, producing a loss equivalent to the capital loss from an immediate revaluation. There's no easy way out!)

Yes, the central bank certainly can implement low real interest rates -- in part, this is how it will generate inflation in the first place. But a cumulative, fully anticipated 30 percentage point shortfall in real interest rates pushes the limits of what monetary policy can do, at least without possible disorder along the way. At the very least, it won't be easy to maintain predictable paths for output and inflation. The real distortions, nominal instability and sudden redistribution away from savers could do plenty of damage to both the economy and the SNB's reputation.

Instead, the SNB decided that it was better to forthrightly acknowledge that the ceiling on the franc was a failed experiment, cutting its losses early rather than saddling the country with a decade or two of monetary unease. It's easy to criticize them now, because we'll never get to see the possibly devastating counterfactual -- but with some imagination, we can get a decent idea.

If this bullsh*t is what passes for "smart" around here, then you people need smarter friends.

Shades of Christmas Past: The Swissies want to avoid imported inflation. The underlying cause of that expected inflation is not of any import to the Swissies. Other countries have been there and done that. So a balance sheet goes down the tubes; a balance sheet is not an income statement.

What the hell is wrong with flexible exchange rates? Are we all victims of the IMF's strive to survive? When will we ever learn?

Comments, on average, are better than the original post.

>>But the SNB is 45 per cent owned by private shareholders

So what? The Fed is 100% owned by private shareholders. But we don't worry about them having an undue influence on Fed policy.

Matt rognlie comments on this issue are incredibly well written.

This is the first time i found a detailed and senseful explanation of why a central bank balance sheet loss does matter even if it can print money at will.

Do you have a site where i can follow your well-thought & well-written opinion? this is top quality stuff.


"which have been complaining recently about insufficient cash transfers from the SNB"
So now with the loss wiping out any and all profits (and probably more), they will be jumping with joy on their future cash transfers?

See, my point isn't so much they loosened the peg. It would, at a certain time, come. It's more on how they did it. Yes, it was going to be disruptive. If they hinted on it, there would be even more CHF positioning, so, yes, hinting may be problematic. But I still believe it could have been done more orderly, if you wanted to do it.

SNB introduced negative rate to make it more orderly. Ok, how about if they sequenced the events like this:
1) introduce 5% negative interest on balances over 10m CHF.
Most likely, this will shake out many many "long CHF" positions, leading to Franc depreciation. I'd say we stop around 1.30+ (-0.25% move resulted in about 0.01 move in Dec. I suspect that the 20time bigger move would not have move 20 times as much); Note that SNB can sell its 400+bn EUR assets into the spike, both realising its EUR profits and making the move more orderly. Yes, there will be a lot of bad-way positioning, but there always would be, and I suspect it's the last thing SNB cared about. As I mention, SNB could well made it more orderly by liquidating its EUR assets into the move;
2) as CHF settles at 1.30 (as I'd expect), you say the floor is now irrelevant and can be removed, possibly moving the negaive rate to -6% at the same time, just to make sure there's no surge.
3) as the markets settle a bit again, you reduce the interest rate say to -4% and see what it does to the Franc. If nothing much, you keep quietly reducing the rate until about -2% - 1.5%

Moreover, if there was need to intervene on the way down, realise that under 1) you reduced your balance sheet which would give them ability to better manage the way down;
The way they did it now is just dumb.

Voila, less volatile markets, floor is gone, SNB made profit, Swiss politicians happy. What's not to like?

Overall, and also in relation to the previous post on the CB and politicians: the problem was that SNB didn't do anything for very long time. It's not just you put the peg in, and then go and sit back. You need to keep the market on its toes. See Czech National Bank. They put in the floor, then they put it up unpredictably. They had to intervene very very little, with almost no balance sheet increase. Few people were willing to take the risk of another unexpected floor raise (and quite a few people are prepared to bet on it going up).

SNB could have lifted the floor to 1.22 when CHF was at 1.25, and then again in a few months to 1.25 Likely CHF would be now 1.30. Also, SNB could have quietly liquidate some EUR assets while CHF was hovering around 1.25 (100m a day would do zilch to markets..), instead of just sitting on it.

Yes, you could get an explosive unwind, but that's a different story. From SNB's balance sheet perspective, they would have been in a much better position.

In short, I believe that even with political and other presures on SNB, the job could have been handled much much better.

Given SNB still purporting "floor is a pillar" as late as Jan 6, they either lied at the time (wouldn't that count as misleading the markets? Can you sue ? :) ), or panicked, as this doesn't smack of a thought-out move.

In other words, lets' move from "peg or not to peg" to "did they think about WTF they were doing"? and I suspect the answer is "nein".

Comments for this post are closed