Purchasing Power Parity – New Estimates

Statsitics Canada has released some interesting new obscure research on what constitutes a “purchasing power parity” exchange rate for Canada.  It was summarized in an article in the December Canadian Economic Observer, and is explained more fully in an on-line technical paper by John Baldwin and Ryan Macdonald.

First off, let me remind everyone that the international empirical evidence is clear that PPPs do NOT determine exchange rates.  Given the extent of financialization and international integration of financial markets, exchange rates are determined like other financial variables: by the massive to’s and fro’s of financial flows that reflect the moment-to-moment expectations and moods of financial investors. They buy a currency as an asset (not in order to facilitate a transaction).

This in itself is not to say that exchange rate are determined solely by “speculation.” Although speculative judgments by financial traders determine each movement, those traders themselves must have some underlying conception of what they think the exchange rate “should” be.  But it is certainly true that exchange rates reflect a stock equilibrium in financial markets — not some equilibrium in the real economy — with all the unpredictability and irrationality that implies.

The Bank of International Settlements conducts a triennial survey of forex trading; the most recent was in April 2007 (before the financial crisis).  At that time total trading in the $C averaged $130 billion (US), or $145 billion per day Cdn.  Today that  means the market is probably worth over $50 trillion per year (likely more, given the enhanced level of trading activity since the financial crisis).  The sum total of our annual exports including tourism ($500 billion) and our incoming new foreign direct investment (a few $10’s of billion per year) amounts to barely one percent of that.  Those are the things that require foreigners to purchase $C to facilitate real transactions (with a broadly similar amount of foreign currency required by Canadians to facilitate our purchases from foreigners).  So the vast majority of trading (in the order of 99%) has nothing directly to due with buying or selling real products or capital assets.

That being said, I believe that PPP is a relevant measure (among others) for guessing at the underlying equilibrium value of a currency (and measuring whether the actual exchange rate is “high” or “low”).  And Statistics Canada’s new estimates provide some interesting food for thought…

How might PPP measures affect exchange rate dynamics (given that exchange rates are a financial, not a real, variable)?  PPP may serve to inform financial traders’ expectations; and if the actual exchange rate gets too far out of whack, that could inspire arbitrage behaviour in goods markets (eg. Canadians’ cross-border shopping) that may (or may not) help to “correct” the problem.

The core concept of PPP is that an equilibrium market exchange rate should serve to broadly equalize the comparable cost of purchasing a range of products in different countries (after adjusting for country-specific factors, like taxes, that might affect the price of particular products).  Canada’s PPP has been estimated in recent years by the OECD to be in the low-80-cent U.S. range.  By that token, the $C has been overvalued since mid-2005.  Today, at 96 cents, it’s above its PPP measure by more than 15%.  An overvalued exchange rate makes our products unnaturally expensive to foreign purchasers, and has contributed to the dramatic decline in price-sensitive exports (including manufacturing and tourism) which has indeed been experienced since then.

The new Statistics Canada paper adds a curious twist to that research, by presenting two different conceptions of PPP: one based on production (GDP) and one based on income (GDI — which differs from GDP by virtue of an adjustment for terms of trade).  A GDP-based PPP is weighted according to the prices of the basket of what we produce in Canada.  A GDI-based PPP is weighted according to the prices of the basket of goods we consume in Canada.

Methodological niceties aside, what’s interesting is how these two measures have diverged since 2003 (when the $C began appreciating dramatically).  At that time the PPP by both measures was estimated by Baldwin and Macdonald at around 83 cents (U.S.).  Since then the PPP based on GDP hs declined, to around 81 cents (U.S.).  But the PPP based on GDI has increased substantially — to almost 90 cents (U.S.) today.

What that means is that a dollar at 90 cents allows Canadian consumers to purchase with one $C the same goods as could be bought in the U.S. with 90 cents US.  This measure has increased because to some extent the higher $C means lower import prices (although imperfect pass-through of exchange rate savings by importers blurs this effect somewhat).  By that measure, the statisticians argue, Canadians’ real income is actualy higher relative to the U.S. today than in the past.  This exchange rate-driven terms of trade effect pushed up Canadians’ relative standard of living by 2008 to 92 percent of U.S. levels.

I am still puzzling over how to interpret this.  It is quite different from saying that 90 cents (U.S.) is now an equilbrium exchange rate … because it was the appreciation of the currency that explains the rise in the GDI-based PPP measure in the first place (so that PPP measure cannot, in turn, explain exchange rates — that would be circular reasoning!).  What I think it means is simply that there are some benefits to a higher exchange rate, something we’ve always recognized — for those with the money and inclination to purchase imports (including foreign travel).

In the productive sphere of the economy, however, we’ve gone the other way.  It is here where Canadians must work and produce, in order to have any chance of spending their money (on imports, or on anything else!).  By this measure, the PPP has fallen — probably since the cost of producing stuff in Canada (despite our crisis) has held up more than in the U.S.  Across the whole sphere of productive industries, the $C should be at 81 cents U.S. for our costs to be “fairly” evaluated in international markets.

In some spheres, the exchange rate should be lower — depending on productivity comparisons.  For example, an 80-cent exchange rate equalizes hourly wage costs in manufacturing between Canada and the U.S.  But most Canadian manufacturing sectors (with a couple of exceptions, like auto and peterochemical) have lower productivity than their U.S. counterparts.  That means that to compete within North America, let alone globally, the exchange rate should be below 80 cents.

Of course, we still hear the refrain that “for too long lazy Canadian manufacturers have been sheltered by a low exchange rate.”  By PPP standards, the exchange rate hasn’t been “low” for 5 years now — in fact, measured in the sphere of production (rather than consumption), it’s at least 15% too high, and getting higher.

In sum, this interesting Statistics Canada study highlights the two sides of Canada’s resource boom.  On one hand, it’s harder for us to make and sell stuff to the world; even including our resource exports, our total export sales are way down.  On the other hand, the higher dollar makes it cheaper for us to buy stuff from the rest of the world, so we “feel” richer (for a while).  That imbalance between producltion and consumption can’t continue forever; it is inevitably reflected in growing international debt.