Purchasing Power Parity in Forex: What It Can and Cannot Do

A currency can look cheap on a purchasing power parity table and expensive on the screen at the same moment, and that distance can sit there for a decade.

That is not a defect in the data. Purchasing power parity is produced by statistical agencies so the size of economies and the level of prices inside them can be compared in one currency. Nothing in that purpose involves the rate a broker quotes.

What follows sets out what the official figures measure, who produces them, how old the newest ones are, and which measure does the job PPP is usually borrowed to do.

What PPP Actually Measures

The international programme behind the official global figures gives itself two objectives: to produce purchasing power parities and comparable price level indexes for the economies taking part, and to convert measures of gross domestic product into a common currency using those parities.

Both objectives are about comparison. A parity is the divisor that lets output in one country be set beside output in another without the market rate distorting the picture.

A market exchange rate is a different object: a price struck between buyers and sellers, moved by capital flows, rate expectations and hedging demand, settling wherever those forces leave it on the day.

The two are not competing estimates of one thing. One is a statistical instrument for comparing economies; the other is the outcome of trading, and a gap between them carries no instruction.

Absolute and Relative PPP Are Two Different Claims

The absolute version is a claim about a level: convert a sum at the market rate and it should buy the same real basket on either side of the border, which holds only where the two price levels stand in the same proportion as the rate itself.

The relative version is a claim about a change: the rate should move over a period by roughly the difference between the two inflation rates.

PPP is usually presented as one idea. It is two, and the distinction decides what a deviation means. Under the absolute claim any gap is a mispricing awaiting correction. Under the relative claim a constant gap is not a mispricing at all, because it cancels out of every change, and only a shift in the gap says anything.

The absolute claim asks a great deal of the world. It needs prices to level across borders, which requires goods to be tradable, cheap to move, and taxed and regulated alike. A haircut, a rent and a bus fare meet none of those conditions, and together they are a large share of what an economy spends on.

The relative claim escapes that, because whatever fixed wedge those factors create drops out when the comparison runs between two dates rather than two levels. It is the narrower claim and the better supported one, and almost never the one behind a headline announcing a currency is undervalued by some percentage.

Who Computes the Official Figures, and When

The global figures come from the International Comparison Program, managed by the World Bank under the auspices of the United Nations Statistical Commission. At its forty-seventh session, in March 2016, the Commission made it a permanent element of the global statistical programme, which is what allows one cycle to be compared with the next.

The calculation starts far below the level anyone quotes. Participating economies collect prices for items from a common list covering household consumption, government consumption and capital formation. Those prices produce parities for basic headings, the lowest level of the expenditure classification at which national spending weights can be estimated, and are aggregated upward from there.

Three features of that process change how a figure should be read, and none of the most visible explanations of PPP mentions any of them.

First, a parity between two economies is not a two-economy number. The published methodology is explicit that a parity linking two economies depends on the parities each of them holds with every other economy in the comparison, so altering the set of economies included moves the parity for any given pair.

Second, parts of the calculation are not price collection at all. Where collecting prices for a heading would be too costly, too slow or too unreliable, reference parities are used instead. Four basic headings take the market exchange rate directly instead: exports; imports; net purchases abroad; and acquisitions less disposals of valuables. To that extent the market rate sits inside the parity rather than opposite it.

Third, economies that do not take part still receive figures. Their parities are imputed with two jointly estimated regression models, using inputs such as output per capita at market rates, trade shares and the age dependency ratio. For them the number is a model output, not a measurement.

The Vintage Problem: Why the Number Is Older Than It Looks

The single fact that decides whether a parity can be used for anything is its vintage, and it is the fact hardest to find beside the figure.

The most recent global release appeared in May 2024 and reports the 2021 cycle: data for reference year 2021, revised data for 2017, a parity series covering 2018 to 2020, and extrapolated gross domestic product parities for 2022 and 2023.

Read in the second half of 2026, the newest measured benchmark rests on prices collected five years earlier. The years after it are projections carried forward, and the years since are not in the release at all.

Between benchmark years the parities are not collected either: basic heading parities are interpolated and then aggregated in the usual way, which makes those years constructed rather than observed.

So before setting a parity beside a live rate, establish which reference year it belongs to and whether that year was collected, interpolated or extrapolated.

What a PPP Gap Does Not Tell You About Timing

Suppose the vintage checks out and a genuine gap remains. It still says nothing about time, which is the whole of the problem.

A parity is a distance. It carries no half-life, no closing date and no direction of travel, and is equally consistent with a gap that narrows next year, one that widens for five more, and one reflecting a structural difference that never closes.

The costs on the other side of the trade are measured in days. Financing accrues nightly and margin is marked continuously, so a position held over the multi-year horizon a parity implies pays those costs thousands of times before the thesis can be right or wrong.

Anything that moves a rate within a tradable horizon sits elsewhere: in the inflation release, in hawkish and dovish policy, and in the flows behind both.

Where PPP Sits Beside the Measures Professionals Use

A measure does exist that answers something close to the question PPP is usually borrowed for, is maintained by a statistical institution, and is republished constantly.

The Bank for International Settlements publishes effective exchange rates. The nominal index is a geometric trade-weighted average of bilateral rates, and the real index adjusts that for relative consumer prices, so a change in it carries two things at once: the nominal move, and how far this economy’s prices have run ahead of or behind the partners in the basket.

Weights come from manufacturing trade flows, are double-weighted so third-market competition is captured, and are revised on a rolling three-year basis.

The contrast is not that one measure is better; they answer different questions, and a real effective exchange rate forecasts nothing either. It is that one describes this week and the other 2021, and only one is a reasonable companion to a live chart.

Reading a Currency as Overvalued Without PPP

If the underlying question is whether a currency looks stretched, there are ways to frame it that do not rest on a five-year-old benchmark. None of them predicts anything.

The first is to compare a currency with its own history rather than with another currency. A real effective exchange rate index carries a stated base, so a reading can be set against the same currency a decade earlier without any claim about where a fair level sits.

The second is to work from the inflation differential directly, which is the relative version of the parity idea over a period short enough to observe.

The third is to separate a move in one currency from a move in its counterpart. A pair that has travelled a long way may be describing the other side of the quote entirely, which is what the dollar index is useful for, and what currency correlation exposes when two pairs stop moving together.

The fourth is to check whether other markets agree. Intermarket analysis asks whether rates, commodities and equities tell one story or three.

None of this produces a target price or a date, and any framework that appears to produce both from a price comparison is adding a claim the data does not contain.

Who This Page Is Not For

If decisions are made and closed inside a day, a week or a month, purchasing power parity has no bearing on them. It has no resolution at that scale, and the most recent official figure predates the position by years.

If the aim is an entry level, a target or a stop, a parity supplies none of the three. It states a distance without a direction of travel or a schedule.

The readers it does serve are different. Someone comparing the size of two economies, or the cost of living inside them, is using the figure for its designed purpose. And someone reading a research note that quotes a parity now has three questions for it: which reference year, collected or extrapolated, and which economies were in the comparison.

Frequently Asked Questions