Quarterly Update December 31, 2013 Issue 20 Special Interview 9 Governance Meetings

9th Executive Board Meeting, Washington, DC (October 30, 2013)

9th ICP Executive Board Meeting sults with a special emphasis on the economic and 2011 ICP results. consistency of the results of the largest econo- The 9th meeting of the ICP 2011 Executive Special Task Force Meeting mies. Board took place on October 30, 2013, in The Special Task Force, created by the ICP Washington, DC to address the critical stage Results Review Group Meeting Executive Board, met from November 21-22, of finalizing the computation process and The Results Review Group, made up of sev- 2013 in Washington, DC to discuss and thor- publishing results. Meeting discussions relat- eral prominent Technical Advisory Group oughly review the quality assurance issues ed to the ICP 2011 progress report; communi- experts, was formed to provide adept review surrounding the results of economically large cation strategy; revision policy; and evalua- of the ICP 2011 results and to ensure the countries. Specifically, the meeting focused tion of the ICP 2011 round. highest quality is achieved. The experts of the on item representativity, the treatment of The Board stressed the critical stage of re- Results Review Group met on November 19, outliers, as well as housing and construction viewing estimates and finalizing the computa- 2013 in Washington, DC. approaches. tion process. Preliminary results were com- The meeting was attended by the World Bank PPP Computation Task Force Meeting puted by the PPP Computation Task Force Senior Vice President and Chief Economist. through parallel and independent processes. The PPP Computation Task Force met in The meeting specifically covered poverty The metadata underlying all computation Washington, DC from December 2-3, 2013 to issues arising from the current set of prelimi- steps is being documented for transparency ensure that their parallel and independent nary results; the comparison of 2011 bench- purposes. processes of global computation converged to mark results with 2011 extrapolated PPPs and the same results. They thus harmonized their For quality assurance, the Board formed a volume measures in the World Development approaches and methods and reviewed the Special Task Force to review the global re- Indicators; and the consistency between 2005 preliminary results generated through the fourth round of computation of results. The sessions were productive and led to an agree- ment on detailed guidelines for the subse- quent rounds of computation and a strict time- table to produce the results. ICP Blogs

Updates including the upcoming release of the ICP 2011 results, as well as information on updating the PPP time series published in the World Development Indicators will be provided via the Open Data blog website

http://blogs.worldbank.org/opendata.

Special Task Force Meeting, Washington, DC (November 21-22, 2013) ICP Quarterly Update — December 31, 2013 — Issue No 20 Page 2 Special Interview with Alan Heston, professor emeritus in the Department of Economics at the University of Pennsylvania

Q1. Considering the evolution of the ICP, Linking issues include choices of comparable which you have witnessed firsthand since items to price across some or all countries in the ICP’s first year in 1968, do you find the different regions. Before 1980, all country methodological improvements to be head- price data were at the Global Office (at the ing in the right direction? UNSO or Penn University) and were checked in one place. The cost of this was that often it The broad framework of the ICP that was was not feasible to interact with countries on formulated around 1960 by Irving Kravis and whether their submitted prices matched prod- Milton Gilbert for binary comparisons in the uct descriptions that were often too generic. OEEC (predecessor of the OECD) has been In contrast to earlier ICPs, growth of the in- retained. An important component of their Professor Heston at the WSS Seminar, ternet and the face-to-face validation within work was to estimate PPPs from the expendi- Washington, DC (November 6, 2013) regions have very much improved the price ture side of national accounts. A major expan- comparisons. In 2005, a set of 18 ring coun- so that the ICP benchmarks and the ex- sion of their vision was to develop a multilat- tries were used to link the regions analogous trapolation of PPPs could be more con- eral framework by modifying the Geary pro- to what was available to link regions in 1980. sistent? posal for cross-country agricultural price Because of questions raised by the 2005 link- indexes as supplemented by Salem Khamis. At this stage two changes, both of which have ing through Ring countries, the approach in The appeal of this Geary-Khamis aggregation been developed by the EU, can be made to 2011 was to use a core price list from which technique was its additivity across time and improve extrapolations between benchmark all countries would price as many items as space at a time when non-additive indexes ICP results. The first is to carry out extrapola- were available in their country. It is too early were less commonly used. Other non-additive tions at the most detailed level provided by to assess the overall success of the core list aggregations were also reported in early ICPs countries for their time to time price indexes. approach. What is clear is the heterogeneity and were regularly used later in the EU and The second is to reduce the time between within and across regions will continue to OECD. benchmark comparisons by undertaking some pose issues for future ICPs. new surveys each year if possible, or finding An important change in the ICP was moving It has been recognized from the inception of readily available prices that allow for some to a regional organization of the comparisons the ICP that the conversion of net exports by updating of PPPs, perhaps for summary ex- in the 1980 ICP. This has posed many meth- the exchange rate is an ad-hoc procedure penditure headings. odological problems that are gradually being adopted mainly because it is a major task to resolved. The biggest improvement in the ICP The easiest improvement in extrapolation of estimate PPPs for exports and imports. As has been in the methods and quality of price GDPs is to add together separate updates of world trade has grown along with the number comparison and validation occurring at the domestic absorption and exports and imports. of participating countries where exports plus regional and global level. In terms of aggrega- This is especially important because simple imports in relation to GDP exceeds 50%, it is tion, moving to other methods like GEKS extrapolation at the GDP level ignores chang- more pressing to find a better method of han- (Gini, Eltetö, Köves, and Szulc) have been es in the terms of trade that for some coun- dling trade. useful, particularly in linking the regions. In tries is very large even from year to year. the 1980 ICP a method was developed to Estimation of PPPs from the production side Q4. What is it about PPPs that make you aggregate to world GDP that could be in turn of GDP is more difficult than from the ex- happy to work on the topic? As one of the divided between the regions, and further allo- penditure side because one needs prices or "fathers" of PPPs, how would you define cated to countries within a region to preserve unit values for both finished goods and raw your legacy? ‘fixity’ of regional relationships. This is the materials and other inputs. The cost of not CAR method that is being used in the 2011 including the production side is that compari- The UK made in kind contributions to the ICP. sons of productivity by sector are not possible 1975 ICP in the form of the time of retired unless very questionable assumptions are statisticians. One of these was Angus Fell When the ICP was initiated, several of the made about mapping expenditure PPPs to who by chance I had met 15 years earlier in a initial ten participating countries did not have PPPs for productive sectors. Even expanding Bombay bar on route to assist in the tabula- detailed breakdowns of GDP from the ex- price collection within the present ICP frame- tion of the Malayan census. I met Angus penditure side and this early attempt to im- work to a number of raw materials, chemi- again in the 1980s in Hong Kong where I was prove the national accounts has increasingly cals, and the like would be helpful. These visiting the statistical office and he was a become a focus of subsequent ICP rounds. additional prices when combined with the consultant. After work in a pub, Angus asked The ICP has been important agent in helping many construction items and wage data al- me what was the latest with the Mighty Gen- countries move their national accounts from ready collected by the ICP would provide a ghis Khan method, which, after initial sur- the 1968 to the 1993 SNA. rough basis for providing PPPs for many prise, realized this was his term for the Geary- Q2. What do you propose should be done productive sectors. Khamis method. I looked at him and asked to further improve the ICP going forward, whether when I was retired 15 years I would Q3. As one of the founders of the Penn especially regarding the consistency be- still be thinking about ICP problems. He took World tables, what would you recommend tween results of two different benchmarks? another drink and just smiled and his smile ICP Quarterly Update — December 31, 2013 — Issue No 20 Page 3 Special Interview cont’d told it all in terms of whatever it is that keeps one involved with ICP type problems. As to legacy, mine is to have participated in a major intellectual shift in perspective on the structure of the world economy. By way of background, Paul Samuelson informally drew up a list of earlier economists who predated creation of the Nobel prize in economics. African Statistics Day, Washington, DC (November 18, 2013) Gustav Cassel was on the list for a seminal movements between a significant group of African Statistics Day: Promoting Quality article published in 1918 in the Economic countries that greatly dwarf in size the trade Data for Africa’s Progress Journal, whose then editor was Keynes. There in tangible goods and services. The large Cassel said, “I propose to call this parity The World Bank Africa Statistics Practice increase in short-term capital movements and “the purchasing power parity”. As long as Group celebrated African Statistics Day in rise in the share of non-tradeables in GDP has anything like free movement of merchandise Washington, DC on November 18, 2013 with given way to a very different world than and a somewhat comprehensive trade be- the theme “Promoting Quality Statistics for 1918. These developments help account for a tween two countries take place, the actual Africa’s Progress”. One of the event's parallel persistent ICP finding that price levels across rate of exchange cannot deviate very much sessions was on the "International Compari- participating countries rise with the level of from this purchasing power parity." son Program for measuring Africa’s pro- per capita GDP. Large short-term capital gress". Nada Hamadeh, ICP Team Leader, Cassel was writing at a time when many movements can be much more important in made a presentation on the ICP and its appli- countries were returning to the gold standard influencing exchange rates than changes in cation within Africa, the upcoming release of after its suspension during World War I, and PPPs often leading to differences in price the results of the 2011 round of the program, his advice to the Swedish government was to level for countries at the same level of per and the way ahead. set the value of the Kroner near its then pur- capita GDP for a number of years. chasing power. The context of Cassel’s mes- What has been gratifying for all of us in- sage became lost and there emerged the so- volved with ICP work since the early rounds called Cassel Doctrine, that Samuelson is that PPPs are now common in textbooks in termed “a misleading and pretentious doc- international trade and development and have trine.” The international trade and finance been more generally accepted in the media, in group of economists moved from Cassel’s research and by national governments and insight to the “law of one price”, a very pow- international institutions. Quoting Keynes, erful assumption on which to base trade theo- “The real difficulty in changing any enter- ries and models. prise lies not in developing new ideas, but in Thus, Cassel’s PPP doctrine entered interna- escaping from the old ones.” A major legacy tional trade textbooks as the PPP doctrine of of the ICP has been to move thinking away Cassel in its absolute one price version and from the Cassel doctrine to the more useful the relative version. The relative version said task of measuring PPPs that originally in- that PPPs and relative prices will move to- spired Cassel. gether in the short and long run. Like Samuel- son, Kravis knew that in the world we live in, Seminars the law of one price rarely held, and the rela- Washington Statistical Society Seminar tive version of the PPP doctrine may hold in the long run but the deviations in the short run The Washington Statistical Society (WSS) Michel Mouyelo-Katoula at the WSS Semi- could exist for many years. The PPP doctrine held a seminar titled “Measuring the Real nar, Washington, DC (November 6, 2013) informed many of the economists recruited to Size of the World Economy-Methodology the World Bank and particularly the IMF. and Challenges” at the Bureau of Labor Sta- International Comparison Program (ICP) The initial ICP did not gain friends at these tistics on November 6, 2013 in Washington Development Economics Data Group institutions by labeling the inverse of what we DC. Michel Mouyelo-Katoula, ICP Global (DECDG) now call the price level as the exchange rate Manager, Fred Vogel, ICP Technical Adviso- The World Bank, 1818 H Street, NW deviation index. ry Group co-Chair, and Alan Heston, Profes- Washington, D.C. 20433 USA sor at the University of Pennsylvania, briefed Ironically, when the fixed exchange rate sys- Phone: 1 800 590 1906 participants on the conceptual framework of tem broke down in the early 1970s, the law of (from inside the US) the ICP and presented an overview of the one price for expected return from short term +1 202 473 3930 statistical methods used to estimate Purchas- investments has been slowly approximated in (from outside the US) ing Power Parities, changes made from the the international financial sector. With the IT Email: [email protected] ICP 2005 round, and the possible impact on and internet revolution, it has become possi- the final ICP 2011 results. Web: www.worldbank.org/data/icp ble for financial markets to generate capital