Pass the Salter

How does a sudden stop and internal devaluation affect the economy? A post dedicated to the  memory of Wilfred Salter

What are the effects when an economy that uses a foreign currency (the euro) and is “living beyond its means” gets suddenly cut off from its creditors? This is, by and large, the situation being faced by Greece, Portugal and Spain, and what to do about it is one of the most fundamentally important questions facing Europe’s policy makers. What I hope to show is that an internal devaluation (with wage contracts demoninated in euros) is more likely to cause a banking strees as an external devaluation, if not more so.

To walk through this problem you need models developed by the would-have-won-the-Nobel-prize Australian economist Salter (and brought to the attention to the world by the writings of Max Corden – another of Australia’s great economists). Unfortunately, Torrens Hume is no Max Corden, so the discussion here will not be as eloquent as what you would find in his masterful Economic Policy, Exchange Rates and the International System (1994).
Back in 1959, figured out that to answer a question like ours, a really useful abstraction is to separate goods into traded and non-traded goods and to account for the important role that relative prices of these two goods can play in an economy. Salter, by the way, has been all but forgotten by his peers. His masterful elaboration “Internal and external balance: the role of price and expenditure effects” is cited only 419 times according to Google Scholar, but his model is THE standard model that lies at the heart of all sophisticated open-economy models used and developed in thousands of economics papers. One reason that Salter is forgotten and did not win the Nobel Prize is that he died at the age of 34 (hence would-have-won, not should-have-won). Anyway, we digress.


Here is the Salter 1959 model : The vertical axis measures tradeable goods – call them “goods” and on the horizontal axis are non-tradeable things, which we’ll call “services.” The curved line running from A to B is the transformation curve or “production possibilities frontier.” It shows the combinations of each good that can be produced by the economy in question where resources are fully employed as traded goods are transformed into non-traded goods by shifting workers and other resources out of the production of goods and into services. Note: if there is unemployment, then production would be somewhere inside the transformation curve. The slope of the curve tells us about prices faced by producers – if you think of the whole economy as being run by a big company, then the amount of goods that it gives up to produce an extra unit of services is the price of that service.
Of course the abstraction is not prefect: there are some too-heavy-to-ship goods, and some traded services (especially in the internet age), but in general it’s a neat categorisation because (traded) goods prices will generally be set on world markets, while (non-traded) services will be set in the domestic market.

There is also a line lablelled with a “1” on the diagram. It shows how many goods and services a country can afford to buy, given the nations income (the value of G&S it actually produces) plus any money it borrows from abroad. The slope of line 1 tells you about the price of services relative to goods faced by consumers (a steeper line means people in the economy must give up more goods to buy extra services (in other words, more expensive services). Since line 1 lies outside the transformation curve, this economy will be spending more than the value of goods it produces and will have to borrow from abroad to do so.

So now let’s see how this internal devaluation works. To start with, suppose the economy was consuming at point E. When there is full employment in the economy, the consumption of services must equal the production of services. Therefore the economy will be producing at a point along the transformation curve directly below point E at point F. Since the value of goods that the economy is buying is greater than what it produces, it must be borrowing from the rest of the world. The distance EF is the measure of how much is borrows in terms of traded goods. This distance is also a measure of the value of net imports of traded goods.

O’kay, so what happens when the world cuts up the visa card and this country can no longer borrow? With no foreign cash to finance spending, the economy’s budget is cut, spending falls (and line 1 shifts in, we’ll discuss be how far it shifts it in a moment). Suppose for a moment that prices, which remember are in Euro, are left unaffected. With prices unchanged the economy still wants to produce at A, but demand for non-traded services is falling (the demand for traded goods doesn’t change, the economy just buys less from the world market). So we are left with an excess supply of non-traded services. This drives down the price of these goods relative to the price of traded goods, which are set in the world market.

This relative price change is important (this was really a key element of Salter’s contribution to the literature at the time). If all prices, including wages can adjust, then this economy will adjust to the fall in services prices by producing more traded goods. If we assume that people buy goods and services in roughly fixed proportions, the economy ends up producing at G, and it consumes at G too, no longer borrowing from the rest of the world. It income and expenditures are shown by the line labelled with a 2.
Let’s take a second to think through what happened in the adjustment to get from producing at F to G. Spending fell, which led to a fall in demand for services, which, in turn, reduced the price of services. This led to (and we skipped over this bit) a fall in the profitability of the services sector and a fall in the demand for labour to produce services, which reduce wages. But the fall in the wage made it more profitable to produce traded goods (where prices were unchanged), and cause that industry to expand while the services sector contracted. That’s how the economy moved from F to G where it is now living within its means.

But the economy rarely adjusts so effortlessly. One reason is that wages often don’t adjust as they should – this is the “Keynesian” assumption that nominal wages are fixed in euro terms (because the country doesn’t have its own currency to price in). Now let’s see what happens in the model.

Start again back in the original “living beyond their means equilibrium” and cut up the visa card once again. Now, as the price of non-traded services falls, the demand for labour in that sector will also fall, just as before. The demand for labour in the traded goods sector remains unchanged, because the world price of traded goods is fixed. The fall in the demand for labour means that some workers in the non-traded services sectors will lose their jobs, and because wages don’t fall, they won’t be able to find jobs in the traded goods sector. In other words, with nominal wages fixed, the fall in services prices starts creating unemployment.

With fewer workers employed in the services sector, the production of non traded services will fall. This reduces the excess supply of services, which will ease the downward pressure. Eventually, when the prices have fallen enough, so that enough workers have lost their jobs, the excess supply of services is eliminated. The size of the adjustment in this “Keynesian” case is quite large because there is a vicious circle at work. As people lose their jobs, the country’s national income falls (so the line showing national income is shifting in closer to the origin, which shows where income is 0). The fall in income reduces the demand for services even further than the initial reduction in foreign financing. It is this mechanism, the vicious circle (known as the Keynesian multiplier), that makes the internal devaluation process so terrible. The problem is even worse than the standard Keynesian model because services prices fall in response to the excess supply, creating even more unemployment than a one good model would predict.

Where does the process stop? If we make a simplifying assumption that the economy always consumes goods and services in a fixed proportion (along the ray O H G F), then it stops when the output of services has fallen by the distance FH.
What is more, profits in this equilibrium are lower than in the frictionless equilibrium, meaning that firms face greater challenges paying back their debts. We can see this by imagining what would happen if wages could fall. If wages were to fall, as they did in the case when wages were fully flexible, service sector profits would start to rise and not just because labour is cheaper. As wages fall, so too does unemployment, national income starts to rise and as it does, the economy demands more services, the price of services is bid up and the combination of the lower unemployment and higher prices means that more services are produced. So profits rise because services prices are higher, services output is higher and wages are lower.
The fall in the wage also induces an increase in the production of traded goods too as goods producing workers also hire some of the unemployed. Profits therefore also rise in the traded sector, because the traded goods sector firms are also selling more and paying lower wages. So that sector is better off too. In fact, as output and incomes rise the economy moves from producing (and consuming at H) to G, the equilibrium we described in the frictionless case.
This overly long post is going to stop here. It’s already been too long. But what Torrens hopes you got out of this is simple. Suppose that prices and wages are set in euros (the foreign currency) and there is a sudden reduction in the supply of credit to the economy. Then euro services prices will start to be bid down and if wages are fixed in euro terms, this will cause unemployment, a fall in national output. This is not controversial. But it is important.

The next post will argue that the introduction of a new currency can be used to effectively rewrite wage contracts so that they are paid in terms of domestic currency. If this is done, then unemployment could be alleviated, and many of the problems associated with the internal devaluation will be eliminated. Indeed profits (in euro terms) will rise, and this will alleviate stress in the banking system, not worsen it. Anyway, that is for next time.

2 thoughts on “Pass the Salter

Leave a Reply