What's a generalised devaluation?
Devaluations were a concept that used to fascinate me as a teenager (OK, maybe I was an odd teenager - but in my defence this was before I had a girlfriend). The topic hasn't been as prominent recently, except in the debate on the Chinese-US exchange rate.
But Krugman's comment started me wondering just why it should be so hard to achieve a cut in prices these days. After all, wages and prices are just nominal values; yes, there are costs to changing them but they're not that high with modern technology. The British government just raised VAT by 2.5%, changing most prices across the UK overnight, and we barely noticed. In April everybody's effective wages will rise a little, with an increase in the personal income tax allowance, and nobody will notice that either.
So could the Spanish government simply pass a law stating that all prices and wages are cut in half from tomorrow onwards? On the surface, there are two clear problems with this - but when we look deeper we'll find that they reveal a much more fundamental issue.
First, this would not be a neutral change. Monetary wealth would double in value, enriching people with money at the expense of those with non-monetary assets (or no assets, or money liabilities). If you have €100,000 in the bank you'll suddenly be twice as rich as you were yesterday, and if you have nothing in the bank you'll gain nothing.
Second, it would make many existing contracts (such as leases on apartments) unaffordable, since wages would be halved with rents kept the same. To solve this, the same rule would have to apply to contracts that already exist - all nominal amounts on any goods or service contracts are divided by two as well. Another problem now: people may have borrowed money to buy an apartment which they rent out. If the rent is halved, the landlord can no longer afford to make the mortgage repayments.
So all debt service payments - and the value of all debts - would have to be reduced by the same amount too. Now you get into real trouble, because a French or German investor who's lent money to a Spaniard is hardly going to accept being repaid only half their money. Almost certainly they could take Spain to the European Court and prevent this rule from being enforced.
Of course, this dilemma is exactly what we see in an traditional devaluation - just a little more unwieldy, and thus more obvious. Countries with their own floating currencies find it more expensive to raise loans denominated in local currency than foreign, because of the risk that the foreign investor's loan can be devalued by the local government's fiat. While if they have sold loans in foreign currency and then decide to devalue, they simply have to accept that their loans will become more expensive to service and take up a higher share of their income. That might be worth it if it makes the difference between the risk of being unemployed and having no income at all, versus retaining some competitiveness and keeping your job.
Indeed, looking down from the national level you'll find that the same principle applies for an individual. Imagine a person who's taken on a mortgage when their salary is $100,000 and finds that they're no longer competitive at that wage. They get to choose between unemployment and a likely loan default, or reducing their wage to get a new job but lowering their standard of living to keep paying the loan. The difference is that a "personal" devaluation doesn't allow you to reduce prices, only wages, meaning that you need to cut consumption more than in the national devaluation case. Against that, the personal option leaves you free to choose your own policy independent of your compatriots, and should allow you to borrow more cheaply in your own currency because the lender doesn't have to price in a devaluation option. They are, however, pricing in a default risk instead - which might turn out to be equivalent to the "personal devaluation" risk.
Abstracting one step further, this becomes a question of how an investment is denominated - or more clearly, in what units are you promising to repay someone? Loans by definition are denominated in fixed currency units; equity investments, by comparison, are denominated in a share of the real resources controlled by the investee; and presumably there could be other types of denomination if it suited both parties. For example a contract could be denominated in a commodity such as oil (this is what a future oil purchase contract is). And turning around that apartment lease, if we remove the distinction of "buyer" and "seller", it's a series of future exchange contracts where one party pays in units of "the right to use this apartment for a month" and the other pays in units of "euro". Ultimately the choice of denomination comes down to what things have a predictable value and to whom.
How do those units become predictable? The value of a currency is a function of the shared beliefs of hundreds of millions of people; although the beliefs of some, such as the head of the central bank or the finance ministers, have more influence than others. Equities have a value controlled by a few thousands or tens of thousands of people; or in the case of a private company, perhaps four or ten individuals. Oil's value arises from the habitual usage of billions of people and the habitual production of about a hundred oil producers (countries and companies). And the value of that apartment might seem predictable, but if you lose your job, have to move to another city, or all your neighbours move out and the local shops close down, its value will change substantially.
Your choice of denomination of contract will be a function of which of these groups you're in, how your future income and expenditure are correlated with those of the groups, and which groups you think are predictable in their future behaviour and beliefs.
Currency devaluations matter because they are a function of the inertia of millions of people who have put their common faith in one shared illusion, and have reflected this in their choice of long-term contract instruments. They have settled on a common converged strategy and that settled choice in itself creates predictability. That predictability arises because people are not as flexible as a theoretical market; because people don't want to recalculate prices and values all the time; because transaction costs (such as the cost of moving house or employing a new worker) create holdout problems which are finessed with long-term contracts.
Floating currencies are a way of drawing certain lines around this flexibility. There isn't, therefore, a pure theoretical case for or against them - the question is which group's inertia and converged strategy you want to participate in. Do you throw in your lot with the rest of the British people, the rest of Europe, or do you only trust yourself?
And what would be the equivalent of a devaluation in differently-denominated contracts? A change in the oil price would qualify; it would simply be a "narrower" devaluation, as it would affect a smaller number of transactions than a currency devaluation. Paradoxically, a narrower devaluation is actually more noticeable; if only 10% of prices change, we will see them move relative to the other 90%; while if all prices move at once (except for those of imported goods) it's invisible.
A reduction in property values affects the likelihood that contracts denominated in property will be fulfilled; the same effect that a devaluation of the pound has on sterling earners. The comparison is especially apt when you look at how people pay their foreign debts. Pound earners borrowing in euros must convert their devalued pounds to euros to make payment; property owners must convert their properties into pounds by renting them out. If rents are sticky, this won't be too risky; if they can be devalued, it will.
So knowing which values are flexible and which are fixed, and where the inertia comes from, turns out to be really important. First because the various inertias and nominal rigidities are what cause recessions; and more universally, because they are solutions to some of the coordination problems at the core of economics. What is predictable and what is knowable, and by whom? Questions that govern every transaction we make with another person, but that we rarely remember to ask.
Asking them will help us understand what assumptions are implicit in the nominal definitions of a transaction. We can get a generalised picture of the risks of our contracts changing in value. And by learning which collective assumptions control those values, we'll know what kinds of devaluations to look out for.