The Penn World Table (PWT) is a set of national-accounts data developed and maintained by scholars at the University of Pennsylvania to measure real GDP (per capita) from the corresponding relative price levels across countries and over time.[1] Successive updates have added countries (currently almost 190), years (1950-2009), demographic data, and capital-stock estimates.[2]
A common practice for comparing GDPs across countries has been conversion to a common currency by multiplication of exchange rates, as though one price — measured by the common-currency exchange rates — thereby prevails across countries, that is, purchasing power parity of each currency. By contrast, PWT estimates employ detailed price indexes within each country for different output categories relative to those of the common-currency country, regardless of whether the output is traded internationally (say, computers) or not (say, haircuts). In so doing, it thereby "corrects" GDP to better approximate purchasing power parity.
An empirical finding documented extensively by PWT is the "Penn effect" of exchange-rate measures for real GDP measured by exchange-rate conversions significantly overstating output differences between high- and low-income countries on average.