On the salience of discrete (discreet?) events

[Warning: I’m not a monetary economist, so the following is mostly pulled from a conversation I had with a fixed-income analyst. Caveat emptor! ]

Paul Krugman has a blog*  on why having your own currency is so useful in debt crises. Basically, you can inflate away debts rather than default or have write-downs. The major premise of why this is better, on average is psychological – it’s a question of reputation and events, rather than pure economics. Bear with me, it gets interesting…

Start with the world having only one currency. Purchasing power inflation may be location specific, but monetary inflation is singular. No one can inflate their way out of debt they can’t afford. Defaults, or write-downs, are the only answer, unless you want to hurt all people with inflation, rather than just bond-holders with default (which just seems mean!). So if there is one global currency, all investors care about is default risk.

If a country creates its own currency, it can manipulate (i) exchange rates and (ii) inflation. Both of those are risk factors in addition to default which investors have to bear. Investors will likely require a higher return because of them. Why would a country go off the global currency except to devalue investments tactically? Note that this argument was used when the USA debated going off the gold standard.

Krugman is arguing this new country can decrease true defaults by manipulating (i) and (ii), but in a specific way. The first possible way (A) is purely economic – it comes from devaluing my investment. They either (a1) print money (and hurt all users of the currency, predominantly it’s citizens), or (a2) change the exchange rate (hard to maintain, but possible). Note that bond-holders have still lost money – their bonds are worth less. However, a default event did not occur.

To understand the second way, you have to understand that defaults are based on the fact that we care disproportionately about discrete events. A lot. If someone wins or loses, it doesn’t matter by how much. If a person or country defaults on their loans, we don’t care that they could have paid back 80% of the value – their reputation, because of that one event, will be shot. They defaulted. No-one will remember a hair-cut. The ratings agencies predict, and CDS contracts are defined, on discrete events. Not inflation/exchange-rate adjusted returns. So this discrete-event, all/nothing model is built into the financial system.

A complete default (write-off) of  8% of debt is better in many ways than 10% inflation. But there is little ability for the system to recognize this as an option and reward debtors for taking it.

Therefore, as a debtor, if you’re going to default, “go big”. Don’t just write-down that which you can’t afford. Write-off as much as possible. So the system basically assesses a default as an all-or-nothing game. The default event happens (and it’s huge), or it doesn’t. This means that the country printing money is simply decreasing the probability of the default event. They are doing the exact same thing, but playing the probabilities rather than the magnitude of defaults.

So what Krugman is arguing is that having your own currency lets you play the probabilities, which gives you breathing room to fix your fiscal situation when cost of debt goes up, and to not have a default event. It means you can flirt with the Rubicon, but not cross it. And knowing you can flirt with it gives investors a bit more comfort to invest in your bonds.

I think the question is, is there an actual probability-adjusted improvement to the system here? If I can only play probabilities, and have finances which would lead me to default 10% of the time, do I actually default less than I would if I could write-off 10% of my debt? The loss of value, on average, in both circumstances might be the same, but there is a lot more fear and positioning around the default point when it’s binary. Because the default point is so salient, involves reputation risk, and bifurcates the return of capital highly.

If fear is more salient around a point than a continuum (likely), then it works, sure. But it means the financial system goes all crazy around that point, becoming like an all-or-nothing game itself. Leverage, capital requirement ratios, and investment criteria based on all-or-nothing probabilities propagate throughout the system, and as humans we systematically mis-weight extreme probability events. I’m not sure that encourages more stability at all.

In other words, is the world better off that bonds are bonds (and default probabilities are all that matter), or would it be better off if bonds were more like equities (and what mattered was return-of-capital ratio)?


* Two problems with his baby-sitting example:

1) In the group, they magically increase actual wealth/GBP simply by putting more money in the system. Not how the real world works, right?

2) There is no pricing mechanism at all, because all goods are exactly the same price. I think babysitting on Saturday night is worth two Tuesday night sits.

3) There is no return on savings (investment) to increase the wealth of the group. Which is kind of the whole point of financial investment….

I see an experiment to be run…

Leave a Reply

Your email address will not be published. Required fields are marked *