« Proxies for monetary disequilibrium | Main | Thoughts on reading Silvio Gesell on money »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Measuring the Canadian economy in terms of USD gives the impression that more is happening than really is. Owing to the exchange rate, the nominal effects are larger than the real effects, right?

In the case of Greece, you say: back when Greece ran unsustainably large budget deficits, was at least partly "fake", being due to an unsustainably high real exchange rate.

Here you seem to use 'fake' in a different sense to mean that it wouldn't have been unsustainable under a counterfactual flexible exchange rate regime.

To make the two examples comparable, one would have to reconstruct Greece's path from the persepective of a counterfactual Deutschmark. And, had that been in place, the real effects would probably have been much less pronounced, whereas the nominal effects might have shown some similarities to the current situation. The crisis would have been 'fake'. As it stands, it is actually very real, which is why it is such a problem.

So I agree very much with your conclusion but I can't quite wrap my head around your use of the word 'fake'.

Oliver: "So I agree very much with your conclusion but I can't quite wrap my head around your use of the word 'fake'."

Nor can I. That's why I wrote this post.

"Even if Greece had perfectly flexible prices and wages, so fiscal policy had no effect on output and employment..."

Really Nick?
Can you please remind us why in such a world government *spending* has no effect on output or employment?
But I digress......

The single market and common(ish) price level make this data "real."

GDP provides a rough measure of national production, but aggregating production together requires some sort of weighting. That weighting is the overall price level, and the Euro has a more or less shared price level due to the common market.

This is the difference between the EU and the US/Canada.

A drop in the (nominal/real) exchange rate in the 90s for Canada was made up for with changes in the US-dollar-denominated price level. (Equivalently, the Canadian-dollar price level did not change significantly.)

In contrast, the price level of Greece has fallen only slightly whereas the Euro-denominated GDP has fallen a great deal. This translates into less "stuff" for any fixed weighting of stuff, which means that Greek GDP has meaningfully fallen. This would not be the case if somehow Greek prices had fallen in tandem with output, but there is no reasonable way that would happen with the single market.

To micro-size your example, you're ultimately arguing that since a retail clerk and CEO are both employed full time, their individual GDP must be equal. The retail clerk's hours of labour buy very little of the CEO's labour-hours, but that's just terms of trade, yes?

Simon: OK, even in an RBC model, government spending might have supply side effects on output and employment, depending on what particular goods the government bought. I'm setting that aside. Perhaps I should have said "and", instead of "so".

Majro: suppose the Greeks produce exactly the same number of olives, year after year. And the Germans produce exactly the same number of cars, year after year. So both countries have constant GDP, measured in terms of units of domestic output. And suppose the price of olives is the same in Germany and Greece, and the price of cars is the same in Germany and Greece. If Greek consumption basket is biased towards olives, and the German consumption basket is biased towards cars, then a Greek budget deficit (or borrowing from abroad more generally) will increase the price of olives relative to cars. And that will increase Greek GDP relative to German GDP, if both are measured in Euros, or if both are measured on a purchasing power parity basis. It will look like an increase in productivity, even though it isn't.

I don't know if I amin the right ball park at all here but, Greece can only borrow Euro's from another europpean country. If Greece and Germany are the only two that exist for simplicity they can only increase their spending in Euro's if Germany reduces their spending and lends those savings to Greece.

Hi Nick! That's a very interesting point; I'm going to have to spend some time wrapping my own head around it.

Nick, I believe your analysis overlooks the fact that different GDP deflators are used for European countries, implying that a variation in relative prices of olives and cars do not change real GDP values, even it actually changes nominal GDP.
More than olives and cars, however, the important changes are in the non-tradable sectors, where prices have grown more than in tradable sectors (Greece is not that big even in the olive market, which is dominated by Spain and Italy). Again, this did not impact headline real GDP figures, but it increased inflation and reduced real interest rates, contributing to the unsustainable consumption and housing boom. And because of price rigidities, price level in Greece has remained high even after its real GDP has plummeted, leaving a significant wedge between PPP conversion rates in Greece and countries with similar GDP per capita in PPP terms, like Latvia or Lithuania.

I think the word "fake" (from Google) means "a thing that is not genuine; a forgery or sham".

If we apply the word "fake" to a generated statistic such as GDP, are we accusing the creators of the statistic or are we assigning our acceptance of the reality of the statistic?

I think we are "assigning our acceptance of the reality of the statistic". If we followed your example of measuring the Canadian GDP with the American dollar, we would need to add an extra step to the steps required to generate the Canadian GDP statistic. The extra step would be to modify all the data into some version of the American data base. This action wold not make anything untrue. It would only change the reference points and add additional factors into any proposed theory offered to account for variations.

Should such a modified data set be offered, we would certainly be correct to ask why economy data from one economy should be commingled with the data from a second economy? Not that the offered data would be "fake" but that the data presented would be diluted with data unrelated to the data of primary concern.

JEC: Thanks! Yep, I'm trying to wrap my head around it too.

Mik: that sounds right to me. If we use a deflator that measures a price index of the goods the country produces, we should get the right answer. (Though I can't quite get my head around how they handle imported intermediate goods if those prices change). But if we use nominal or purchasing power parity measures of GDP to compare the relative performance of different Eurozone countries, then we are going to see the effects of relative fiscal policy changes affecting GDP even if those fiscal changes had no effect on employment or productivity.

Roger: yep. It's about how we interpret the data, and which particular data series we use for which purpose, rather than Eurostat putting out dodgy numbers. That's why I put "fake" in scare quotes. But in the Eurozone, with a common currency, it's much harder to see the problem. If we compared Canadian to US productivity growth by converting at the exchange rate, it's easier to notice the problem.


See page 51, Macroeconomics 4th Canadian edition, by yours truly.

To me, in the Eurozone, it seems easier to see the problems. The common currency puts all players on the same footing. When they are all on the same footing, it is easier to see the resource differences, the differences in laws, and the differences in social standards.

For example, the high Greek GDP over the last few years (until about 2010) seems (to me) to be based on government deficit spending. Lacking the resources of Germany, the social commitment to work, and differences in pension standards, the Greek economy became dependent upon the deficit of the Greek government to finance the observed GDP. Stated another way, the Greek GDP would have been less had the Greek government NOT had a deficit.

I see exchange rates as predictions placed into economic reality. For example, when Japan announced that it was going to devalue the yen, the exchange rate immediately changed. The actual event had not had time to bring effect. The exchange rate changed because of predictions the the Japan central bank would deliver on it's promise.

I think exchange rates heavily depend upon the ability of the two currencies to preserve buying value. This because money earned by foreign exchange is likely to be spent considerably in the future. This is aggravated by the individual character of foreign sales where one years sales may finance next years production.

Thanks Steve. I think I have a copy in my office. I will take a look after the weekend.

Roger: "...the Greek economy became dependent upon the deficit of the Greek government to finance the observed GDP."

Suppose Greece produces 100 olives, which it sells to Germany at $1 each, and borrows the extra $900 to buy 100 cars from Germany at $10 each. Greek GDP is 100 olives, or $100. Not 100 cars or $1,000. Greek GDP is (German) expenditure on Greek-produced goods, not expenditure by Greeks on (German) goods.


I am assuming that GDP = private consumer spending plus government spending plus (exports less imports).

Both private consumer spending and government spending are increased by any government deficit, as is the level of imports (a increasingly negative component when government has a deficit).

So I do not understand your comment "Greek GDP is (German) expenditure on Greek-produced goods, not expenditure by Greeks on (German) goods."

Roger: OK. I misunderstood you. I thought you were confusing GDP with Domestic Absorption, because you said "to finance". It doesn't need any finance to buy what you produce.

Thanks Nick. I should have been more clear by saying something like "deficit of the Greek government to "partly" finance the observed GDP.

Roger: average Greeks work 7hr/week more than Germans and retire at almost the same age as Britons. We need to focus on other memes.

Nick, you don't mean fake. You mean there is a potentiallt large and unquantified systematic uncertainty in the GDP arising from the analysis technique.

How is that for a catchy title?

Ugh. I post on my phone. Forgive my typos please.

Great post. It's remarkable how badly nominal GDP per capita can mislead.

My favorite example is this: suppose that there are two types of goods, "tradable" and "nontradable". The former is an endowment good with zero trade costs, uniform across all countries, and the latter cannot be traded at all (and we don't care how it's produced). Suppose further that each country has to run a zero trade balance, so that its consumption of tradable goods equals its endowment. And suppose that consumer preferences over tradable and nontradable goods are Cobb-Douglas, with constant expenditure shares alpha and 1-alpha, respectively.

In this world, countries' dollar GDP is determined entirely by their tradable endowment divided by alpha. (By "dollar GDP" I mean nominal GDP converted into a single, fixed country's currency.) Why? Since the tradable good is uniform, its price in dollars is the same across all countries - so the tradable component of dollar GDP is pinned down by the endowment. And then the nontradable component of dollar GDP is just (1-alpha)/alpha times the tradable component by the Cobb Douglas assumption. Adding up, overall dollar GDP is 1/alpha times the tradable component.

So in this world, dollar GDP is unaffected by how terrible (or good) the nontradable sector is. You could have an abysmal housing stock, pitiful restaurants, and the worst barbers in the world, and you'd have the same dollar GDP as in a mirror universe where the construction industry was so productive that everyone lived in a mansion.

Now, this is obviously a special case, but you can modify it in natural ways - e.g. define the tradable endowment so that it includes "terms-of-trade" shocks, or allow for capital flows. Regardless, from the Cobb-Douglas consumption assumption we know that dollar GDP still equals your tradable consumption divided by alpha - and that consumption is determined by the value of your tradable endowment and your capital flows over time. In this world, countries can use international capital markets to smooth consumption in the face of endowment shocks, but there is also a new source of volatility, from shocks to those very same capital markets.

(Even more frightening: changing consumption preferences in one plausible direction, by setting the elasticity of substitution between nontradables and tradables below 1, has actively perverse effects. Then, the less productive your nontradable sector, the higher your dollar GDP!)

Matt: Thanks! Great to see you back (though I was watching your take on Piketty with great interest).

That's a lovely simple little model, that captures well what I'm trying to talk about here. I'm looking at the effect of adding capital flows.

Gotta go.

I do not understand. Is GDP not measured in PPP terms? I know for a fact that for international comparisons local GDP is actually measured in US dollars - of course using Purchasing Power Parity adjustment.

So for instance whole Slovakia which is part of Eurozone has GDP of only about $18,400 nominally compared to Greece's $21,600 they have higher GDP (PPP): $ 28k for Slovakia compared to $26K for Greece. Is this what you wanted to hear?

I agree with the general logic here that Eurozone GDP was inflated in the past. I would put it like this: Due to the unsustainable specialization of high productivity core (Germany, Netherlands) in the tradable sector other countries were allowed to (unsustainably) specialize in the non-tradable sector where their productivity is relatively high (see Balassa-Samuelson). This allows a sort of ricardian efficiency gain which should inietially boost growth and then reverse when imbalances are corrected.

I had outlined this possible mechanism in my blog in 2012 (it is in German but there is a translation function on the right hand side):

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad