新台幣與均一定價理論背離: 台幣實質與名目匯率差異的成因 - 政大學術集成

全文

(1)國⽴政治⼤學亞太研究英語碩⼠學位學程 International Master’s Program in Asia-Pacific Studies National Chengchi University. 碩⼠論⽂ Master’s Thesis. 新台幣與均定價理論背離: 台幣實質與名匯率差異的成因 The New Taiwan Dollar against The Law of One Price On the Causes of Taiwan’s Real/Nominal Exchange Rate Spread. Student: Andreas Springer Advisor:. Ching-Chih Lu. 中華民國 104 年 7 ⽉ July 2015.

(2) 新台幣與均⼀一定價理論背離：台幣實質與名⽬目匯率差異的成因 The New Taiwan Dollar against The Law of One Price On the Causes of Taiwan’s Real/Nominal Exchange Rate Spread 研究⽣生：施安德指導教授：盧敬植. Student: Andreas Springer Advisor: Ching-Chih Lu. 國⽴立政治⼤大學. 政治大. 學. ‧ 國. 亞太研究英語碩⼠士學位學程立碩⼠士論⽂文. y. ‧. Nat. er. io. sit. A Thesis. n. Submitted to International Master’s Program in Asia-Pacific a. iv l C n h eStudies ngchi U. National Chengchi University In partial fulfillment of the Requirement For the degree of Master in China Studies. 中華民國 104 年 7 ⽉月 July 2015.

(3) Acknowledgements I want to express my sincere gratitude to my advisor Dr. Ching-Chih Lu. Not only did he provide me with valuable insights and inputs, and helped me overcome organisational and formal obstacles, but also his constant accessibility helped me to progress steadily. I highly appreciate his open-mindedness, guiding me without determining my line of argument, yet investing vast amounts of time assisting me with the details. I am furthermore deeply grateful for the valuable comments and remarks by Dr. PingYin Kuan and Dr. Tseng-Min Wu, who helped me defining my research in the early. 治政大 and other institutions. I also want to thank my former teachers at Chengchi University 立 Many of the ideas I followed evolved from what I once learned from them. stages, and later revising the final version.. ‧ 國. 學. Last but not least I want to express my gratitude towards my parents, who made it possible for me to come to Taiwan, and continued to support me in various ways,. ‧. allowing me to focus on my academic work. . n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v.

(4) Abstract Taiwan’s GDP and wages have remained steady for more than a decade, though the living standard continued to rise. This paradox is induced by the New Taiwan Dollar’s nominal and real exchange rates diverging, an anomaly amongst advanced economies. Investigating ongoing international economic integration, Taiwan’s monetary policy, political situation, and balance of payments unveils the underlying causality. Taiwan could not participate in the most recent wave of globalisation, insulating its price level. In combination with outgoing investments, reserves accumulation, and its traditional business culture and structure, this has created an area of inimitable low. 政治大 needs to either be resolved, or separated from its economic capacity to act. Only then 立 can Taiwan’s economic situation normalise and the real/nominal exchange rate spread prices. In order to prevent distortionary long-term effects, Taiwan’s political situation. ‧ 國. 學. fade, which implies an increase in nominal GDP and wages (measured in USD).. ‧. Keywords: New Taiwan Dollar, Exchange Rate, PPP, Economic Integration, Price. n. al. er. io. sit. y. Nat. Level. i Un. v. 台灣的國內⽣生產⽑毛額及薪資在過去⼗〸十幾年來相當的穩定，但是⽣生活⽔水準卻仍然. Ch. engchi. ⼀一直提升。這種⽭矛盾現象來⾃自於新台幣的名⽬目跟實質匯率背離，在先進國家裡⾯面是相當罕⾒見的。本論⽂文研究了台灣與國際經濟的整合、貨幣政策、政治局勢以及收⽀支平衡帳等因素，並了解他們跟此⽭矛盾之間的潛在因果關係。由於台灣未能充分參與最近⼀一波的全球化浪潮，使得它的價格體系跟國際⽔水準脫勾。⽽而對外投資、外匯累積，以及傳統商業跟⽂文化結構等因素，共同創造了現有獨特的低物價⽔水準環境。為了防⽌止這種⾧長期的價格扭曲影響，台灣的國際政治情勢必須改善，或⾄至少跟經濟⾏行為能⼒力分離。唯有如此，台灣的經濟情況才能恢復正常，讓實質跟名⽬目匯率的差距縮⼩小，（以美⾦金計價的）名⽬目國內⽣生產總值和⼯工資也才能成⾧長。 .

(5) Table of Content 1. Introduction. 1. 2. Fundamental Theories. 7. 2.1 The Law of One Price. 7. 2.2 Purchasing Power Parity. 9. 2.3 Balassa-Samuelson-Effect and Penn-Effect. 13. 2.4 Fisher- and International Fisher-Effect. 15. 2.5 Real / Nominal Exchange Rate Convergence. 16. 政治大 3.1 Inflation, Real Growth and Population Growth 立. 19. 3.2 International Fisher-Effect and the NTD Nominal Exchange Rate. 21. 3.3 Real / Nominal GDP Ratio over Time - Taiwan’s Case. 22. 3. Emergence of Taiwan’s Real / Nominal GDP Spread. 19. ‧ y. 24. sit. ‧ 國. 學. 24. 26. 3.4 International Price Levels. Nat. 3.4.1 All Countries. n. al. er. io. 3.4.2 Advanced Economies 3.4.3 Anomaly Taiwan 3.5 Causes Overview. Ch. engchi. i Un. 4. International Integration. v. 28 32 33. 4.1 Factor Price Equalisation. 33. 4.2 Economic Integration. 37. 4.3 Taiwan’s Isolation. 40. 5. Monetary Policy and Cross-Strait Relations. 46. 5.1 History and Official Foreign Exchange Policy. 46. 5.2 Actual Monetary Policy. 47. 5.3 Taiwan and China Economic Policy. 50.

(6) 6. Balance of Payments. 54. 6.1 Overview. 54. 6.2 Foreign Reserves. 56. 6.3 Current Account. 60. 6.3.1 Overview. 60. 6.3.2 Trade. 60. 6.4 Investment. 65. 6.4.1 Financial Account. 65. 政治大 6.4.3 Inbound Investment Disincentives 立. 65. 6.4.2 Direct Investment Flows. 74. y. 77 79. 7.2 Competition and Efficiency Focus. A. NPV and IRR B. Data. al. n. Appendix. er. io. References. Nat. 8. Conclusion. Ch. engchi. i Un. 73. sit. 7.1 China’s Influence. 73. ‧. ‧ 國. 學. 7. Price Level Considerations. 68. v. 89 89 90.

(7) List of Figures and Tables Table 1 (a) Nominal GDP per capita. 4. Table 1 (b) GDP (PPP) per capita. 5 . Figure 1 Systematic Overview. 6. Figure 2.1 Transaction Costs and Arbitrage Illustration. 8. Figure 3.1 (a) Factors of Taiwan's Nominal GDP Growth. 20. Figure 3.1 (b) Factors of South Korea’s Nominal GDP Growth. 20. Figure 3.2 KRW/TWD Exchange Rate. 22. Figure 3.3 PPP/Nominal GDP Ratio Taiwan and Korea. 23. 治政 Figure 3.4.2 PPP/Nominal GDP Ratio Advanced Economies 大立 Figure 3.4.3 Price Level Time Series. 25. Figure 3.4.1 PPP/Nominal GDP Ratio All Countries. 27 31. ‧ 國. 學. Figure 4.1 International Trade Figure 4.2 FTAs in Asia-Pacific. 35 40. ‧. Table 4.3 (a) Free Trade Agreements. io. sit. Figure 6.1 Balance of Payments. y. Nat. Table 4.3 (b) International Investment Agreements. er. Figure 6.2 Foreign Reserves. n. a l Short Term Debt iv Table 6.2 Foreign Reserves and n Ch U Figure 6.3.1 Current Account Balancee n g c h i. 42 43 55 57 59 61. Figure 6.3.2 (a) Share of China-Exports. 62. Figure 6.3.2 (b) Trade Openness and Total Trade. 63. Figure 6.3.2 (c) Adjusted Trade Openness. 64. Figure 6.4.1 Investment Account. 66. Figure 6.4.2 Foreign Direct Investment. 67. Figure 7.2 Firm Size Comparison. 75. !1.

(8) 1. Introduction “Why Is Taiwan Lagging Behind? Over the last decade, South Korea has pulled ahead of its rival Taiwan. Why has South Korea been able to graduate from the ‘Four Little Dragons of Asia,’ achieving a per-capita GDP 20 percent higher than that of Taiwan?” (Lai (2010)), “ titles an article in the Taiwanese Common Wealth Magazine 5 years ago. Since then, not much has changed, similar questions are still asked by the media and are discussed in a variety of forums online. Though Taiwanese tend to have different opinions and hence sometimes disagreeing answers, the great majority is convinced this is a key question, not doubting the underlying assumption: Taiwan did. 治政 line. “Taiwan's low average wage level shows that the 大 country's overall 立 competitiveness is not strong enough” said Quanta Chairman Barry Lam in 2014. fall behind. Complaints about Taiwan’s low and stagnating wage level take the same. ‧ 國. 學. (Focus Taiwan) and the Common Wealth Magazine (Lin and Teng (2014)) claims “Taiwan's youth are leaving in droves to serve as low-wage workers in Singapore”. ‧. attributing it to the low pay in Taiwan, which also causes brain drain in higher skilled. y. Nat. fields.. io. sit. But is this notion true, did Taiwan fall behind? Is leaving Taiwan the right. n. al. er. choice for young Taiwanese to increase their living standard? If not, how could this. i Un. v. generally accepted view come up and remain over such a long period? I will come to. Ch. engchi. the conclusion no, Taiwan did not fall behind, and hence leaving does not increase the living standard in most cases. The misperception came up due to a singularity in the exchange rate of the New Taiwan Dollar and consequentially an abnormally low price level in Taiwan.. Many Taiwanese today are dissatisfied with their country’s economic situation and development. From their perspective - after decades of fast pace growth - their country has started to stagnate. While the East Asian miracle marked the beginning of the rise of four East Asian economies or ‘tigers’ - namely Singapore, Hong Kong, South Korea and Taiwan - only Taiwan’s momentum seems to have faded - or passed on to China. With high growth in the country that is at the same time a cultural !1.

(9) relative - at least a cultural ancestor - and the most dangerous political and military threat, many young Taiwanese are in the ambivalent situation of fearing China and at the same time hoping to find their personal career opportunities associated with its rise. Both supports the impression of Taiwan’s economic inferiority and the notion of ‘falling behind’ relative to its geographic neighbours. The economy of Taiwan is most often compared with South Korea’s. This is reasonable, as both took part in the East Asian Miracle and opposed to Hong Kong and Singapore, both are not city states. Both countries democratised at around the same time - and only after significant economic development has happened, setting them aside from many newly democratised countries, which face the task of economic modernisation only after giving voting rights to their people. Also their economic. 政治大. model was comparable, being inspired by the Flying Geese model (Kojima (2000)). 立. and first functioning as an outsourcing location for manufacturing, later moving up. ‧ 國. 學. the value chain towards the implementation of high tech industries. Today, in both countries semiconductors play an important role, and though both have a different. ‧. approach and hence the industry is shaped in different ways, it is nevertheless an. y. Nat. extraordinary similarity, as Chen and Sewell (1996) show, probably comparable to the. n. al. er. io. of Japan and Germany.. sit. differences and similarities that can be found when looking at the dominant industries. i Un. v. Scitovsky (1985) identifies several key factors that are underlying both. Ch. engchi. Taiwan’s and South Korea’s development:. - Common Chinese tradition and Confucian philosophy which is reflected in • placing high value on education resulting in high levels of human capital • ability and willingness to work hard resulting in a high count of working hours • wage flexibility and employment stability. - Japanese Rule that lead to • fast productivity growth in agriculture • development of infrastructure • reduction of inequality by repossessions of former Japanese assets. - Reduction of Inequality due to land reform/farm-prize support program - Generous economic aid and heavy burdens of military expenditure !2.

(10) - Low inbound FDI Therefore it is not surprising that Taiwan’s recent success - or lack thereof - is frequently derived from a direct comparison with South Korea’s performance. Economic Development is pursued to increase the quality of life for a country’s citizens. Measuring life quality is a very arbitrary process though, especially when comparing countries with different cultural backgrounds, norms and value systems. Economists therefore defined “standard of living” as “the level of wealth, comfort, material goods and necessities available to a certain socioeconomic class in a certain geographic area. The standard of living includes factors such as income, quality and availability of employment, class disparity, poverty rate, quality and affordability of housing, hours of work required to purchase necessities, gross. 政治大. domestic product, inflation rate, number of vacation days per year, affordable (or free). 立. access to quality healthcare, quality and availability of education, life expectancy,. ‧ 國. 學. incidence of disease, cost of goods and services, infrastructure, national economic growth, economic and political stability, political and religious freedom,. ‧. environmental quality, climate and safety.”1 This measure is very complex as it tries to. y. Nat. grasp all aspects of well-being, and as utility2 associated with each factor differs from. sit. person to person, it can surely be assumed they differ between differently socialised. n. al. er. io. groups. Citizens of very religious countries might have no need for religious freedom,. i n U. v. the importance of housing space relative to for instance health care might be lower in. Ch. engchi. less individualist socialised countries, wide-spread infrastructure might be less important for highly urbanised countries etc. Also the discussion about “Asian values” comes into play here and whether or not ‘Asian values’ exist, the academic discussion about them proofed that at least to a certain extent societies place different value on certain aspects of their life (Thompson (2001)). Reducing the measure of living 1. Investopedia. 2. Utility is a concept that tries to quantify wellbeing. Everything that improves an individuals ‘happiness’, be it monetary or non-monetary, can be added or compared this way. Its advantage over more the more common indicator welfare or income derives from the fact, that well-being is not a linear function of income and can differ from person to person. Some might for instance derive an increase in utility buying a more expensive good that was produced under conditions acceptable by the buyer (‘fair trade’). !3.

(11) Table 1 (a) Nominal GDP per capita Per capita GDP (current. Taiwan. Korea. 1990. $8,086. $6,642. 2000. $14,641. $11,948. 2010. $18,488. $22,151. 2014. $21,572. $28,739. prices USD). standard to one of its components, namely gross domestic product (per capita) therefore makes sense for economic considerations, especially due to many other factors being directly connected to GDP - a certain level of infrastructure,. 政治大. employment, political stability etc. is implied in high GDP, while at the same time. 立. high GDP is a necessity for widespread access to quality health care and education.. ‧ 國. 學. According to International Monetary Fund data, in 1990 Taiwan’s GDP per capita was nearly 22 percent higher than South Korea’s, a head start, that it. ‧. maintained into the early 2000s. However, between 2000 and 2010 Taiwan’s GDP per capita fell relative to South Korea’s to only 83% thereof. By 2014, the most recent. y. Nat. sit. number, it diminished to three quarters, or 75% of its neighbours per capita GDP.. er. io. A relative decrease of 47% within only 14 years indicates a clear failure of. al. iv n C within 14 years means a compoundhgrowth e n gofcless h ithanU3% per year. While this might n. economic development - or doesn’t it? Fact is, the move from 14,614$ to 21,573$. be considered as decent for a fully developed country - it actually beats the US’s longterm growth average of just 2% between 1870 and 2005 (Barro (2007)) - for a country under Taiwan’s conditions in that time, it can be associated with stagnation. The numbers shown before (Table 1 (a)) were derived under current prices. Prices though do not necessarily develop in the same pace everywhere in the world. The deflator that is used for the USD, not necessarily deflates the prices in Taiwan and South Korea correctly, however, by measuring GDP in terms of USD, we imply the US deflator. Inflation occurs differently in different regions, there might actually be differences in inflation between different districts of the same city. Using the same deflator for different countries that even have different currencies introduces a. !4.

(12) Table 1 (b) GDP (PPP) per capita GDP (PPP) Per capita. Taiwan. Korea. 1990. $10,316. $7,812. 2000. $21,207. $16,452. 2010. $36,830. $29,825. 2014. $41,539. $33,791. (current prices USD). potentially large error into this comparison and its conclusions, which is the reason why PPP data is usually used to determine the standard of living.. 治政大 such a difference in the better position today than South Korea is. But why is there 立 nominal GDP and GDP (PPP) numbers? How did it happen and will the difference Using numbers adjusted for purchasing power, Taiwan is de facto still in a. ‧ 國. 學. persist, misleading more people about Taiwan’s economic situation in the future? The answer is tied to Taiwan’s political future. If Taiwan gains the ability to. ‧. act as an independent economy3 and utilises this capability to catch up with the. y. Nat. process of international integration that Korea and other advanced and non-advanced. io. sit. economies have undergone during the early 21st century, then Taiwan’s nominal. n. al. er. income is likely to increase and the relative income between Taiwan and Korea should. i n U. v. be about equal in nominal and PPP terms. Otherwise, the difference appears to be. Ch. engchi. stable, and in the long run Taiwan’s GDP (PPP) might even approach its nominal GDP due to a lack of growth potential. Figure 1 provides a systematic overview on how to derive at this conclusion. I will show how Taiwan’s real/nominal GDP spread emerged after introducing some necessary theories in section 2. The emergence will be shown in section 3. Section 4 shows Taiwan’s absence from international integration, which explains why Taiwan’s nominal GDP and GDP (PPP) differ. The following three sections show why Taiwan’s GDP (PPP) is larger than nominal GDP and not vice versa: Section 5 describes Taiwan’s monetary policy and its special relationship with China, section 6 looks into the balance of payments and section 7 will discuss China’s economic influence and 3. De jure political independence is not a necessary condition. !5.

(13) Figure 1 Systematic Overview. 立. 政治大. ‧. ‧ 國. 學 sit. y. Nat. al. er. io. the local business environment. As can be seen, this is a three-step causal research,. v. n. hence the counter-clockwise progression, starting with the phenomenon and moving backwards the causal chain. . Ch. engchi. !6. i Un.

(14) 2. Fundamental Theories This research requires the acknowledgement of a number of well-established theories, before further investigation can be conducted. They build the foundation for findings, further theories and data interpretation later, hence this early section will introduce those underlying concepts.. 2.1 The Law of One Price4 The Law of One Price states that one good sells for the same price at all locations, as. 治政大 without any associated selling at the higher priced one, a trader could realise profits 立 risks. This then drives demand up at the lower priced location and increases supply at otherwise arbitrage opportunities would arise. By buying at the cheaper location and. ‧ 國. 學. the higher priced one. This movement of goods continues until demand and supply meet at the same price at both locations, as only then the trader cannot realise any. ‧. gains anymore by buying and selling simultaneously at both places. Hence, in an. sit. y. Nat. efficient competitive market, there are no arbitrage opportunities, which concludes to the law of one price.. io. n. al. er. As GDP (PPP) is a composite of the price of all goods (ideally, factually it. i Un. v. only looks at the prices of the reference basket) sold in a market and the law of one. Ch. engchi. price applies to each variety of a good, then if the law of one price holds, so does PPP. Differences in PPP must therefore correlate with a deviation from the Law of One Price.. The law holds, as long as two assumptions are met:. - There must be no transaction costs - There must be no trade barriers. 4cf.. Krugman, Obstfeld & Melitz (2012) !7.

(15) Price. Figure 2.1 Transaction Costs and Arbitrage Illustration. 立. 政治大. ‧ 國. Demand Location B Supply Location B Upper Bound. Demand Location C Supply Location C. ‧. Demand Location A Supply Location A Lower Bound. 學. Quantity. y. Nat. Transaction (which includes transportation) costs create a possible price difference,. io. sit. however this price differences in theory should not be larger than the transaction. n. al. er. costs, as only a difference within the price corridor created by the transactions costs is shielded against arbitrage.. Ch. engchi. i Un. v. Figure 2.2 illustrates the mechanic. As long as the transaction costs are at least as high as the difference between the lower and the upper Bound, there is no arbitrage opportunity and the price difference can maintain. Should the transaction costs be lower, then goods from the dotted line market would be transported to the dash lined market, increasing demand at the dotted and increasing supply at the dash lined market, moving their prices closer together until the difference between the upper and lower Bound (the upper bound being the highest, the lower bound being the lowest price) is equal to the transaction costs. Trade barrier include everything, that is preventing goods and services from crossing the border from one market to another. They hinder arbitrageurs from taking advantage of price differences by stopping trade. While some trade barriers are !8.

(16) physical - especially in the service sector which often requires actual attendance for consuming the service - Trade barriers are also often policy driven and are not necessarily bad for an economy - for instance they can ensure quality in what else would become a Market for Lemons5. Frequently they appear in the forms of customs, which, as some other trade barriers, do not so much stop trade completely, as they actually add to transaction costs and hence increase the distance between the upper and lower Bound, allowing greater price variety between different markets.. 2.2 Purchasing Power Parity Purchasing Power Parity is a concept that compares a country’s gross domestic. 政治大. product with the average prices of goods and services within this country. Not. 立. spending money per se, but the consumption of goods and services creates utility for. ‧ 國. 學. the individual. Purchasing Power Parity refers to the amount of goods and services that can be consumed rather than just the income that is available for spending.. ‧. Income adjusted for the price level, i.e. purchasing power, is therefore a better proxy. y. Nat. for well-being than nominal GDP is.. sit. A Purchasing Power Parity comparison can be achieved by generating a. n. al. er. io. theoretical exchange rate under which the average price for goods and services. i Un. v. consumed is exactly the same for both countries, independent of the currency used.. Ch. engchi. More than two countries can be compared by setting the price level of one country as the norm und comparing every country with the normed one. In practise, the country whose price level is used as the norm is the US with a price level of 1 or 100%. As prices usually are more stable in a country's own currency than in foreign currency (market exchange rates can be very volatile and fluctuate by the second, with the possibility of several percentage point movements within few days - unless the rate is fixed, which tends to be less and less common), an income comparison by GDP (PPP) is also more stable than using nominal exchange rates. Fixed exchange rates can divert from the “real” exchange rate even more than floating market rates, as they artificially prevent the market rate from being adjusted by the invisible hand - and 5. cf. Akerlof (1970) !9.

(17) might disguise the real value of a currency (for instance the HKD). If Purchasing Power Parity holds (i.e. the adjusting factor is 1 compared to the price level of the US), then the market exchange rate and the real exchange rate are identical. There exist a variety of sources and methods to calculate the differences in price levels6. The most simplistic one is the “Big Mac Index” published by the British “The Economist”, which compares the price of a Mc Donald’s Big Mac amongst different countries. Though it reflects average price differences surprisingly well, it is not a good Index for further economic analysis due to its focus on just one product. A better one are the comparative price levels published every month by the OECD, which calculates the prices for the same basket of goods and services for each of its members, and sets the price level of the US at 100 for reference.. 政治大. The most all-encompassing approach today is the International Comparison. 立. Program (Ahmad (2003)). It was first introduced by the UN and the University of. ‧ 國. 學. Pennsylvania and led by Irving Kravis (1975), who developed the general methodology of calculating the real income levels amongst different countries. While. ‧. the first round in 1970 only included 10 countries, the number of observes countries. y. Nat. expanded rapidly for the next rounds7. The comparison of countries within different. sit. geographical areas and developmental stages proved difficult, as it wasn’t obvious. n. al. er. io. which items to use for comparison. Also, many countries had a strong interest in being. i Un. v. compared with countries that are similar to themselves, i.e. geographical and cultural. Ch. engchi. proximity, and less interest in being compared with countries they had less in common, therefore they didn’t want the comparison to be influenced by the different. 6. I use the terms “price level”, “real/nominal exchange rate ratio” and “real nominal exchange rate spread” as synonyms. Mathematically, price level and real/nominal exchange rate ratio are inverse, i.e. a ratio 2 implies a price level of 50% (relative to the US). A spread is the difference between the two exchange rates, i.e. a ratio of 2 implies a spread of 100% (based on the nominal exchange rate). Alternating between those terms depending on the context avoids unnecessary abstractions. 7. year: 1970 countries: 10 (World Bank). 1973 16. 1975 34. 1980 60. !10. 1985 64. 1993 117. 2005 146. 2011 199.

(18) price and expenditure structures of countries who are not part of their geo-cultural group8. As a consequence, the ICP focused more on regional groups starting in 1980, which means countries are first categorised into a region and compared amongst each other, without any influence from out-of-region data. National agencies organise the price surveys, regional agencies monitor the price gathering and conduct the comparison at the regional level. World-wide PPP data is generated by the Global Coordinating Agency based on Regional PPPs. The Global Coordination Agency was located at the UN until 1985 and at the World Bank after that. The scope of the Program increased tremendously over time and now includes more than 199 countries/territories.. 政治大. Over the years, the methodology has been changed several times. A major step. 立. was the 2005 release, which responded to the criticism the ICP faced for its earlier. ‧ 國. 學. publications, like ‘the resources employed in the PPP program are minuscule, and this is reflected in the quality of the results’ (Castles (1997)), and ‘those conclusions put. ‧. very briefly that ICP is a programme worth keeping but that its current condition, if. y. Nat. little is done about it in terms of credibility, quality of output and survival prospects,. sit. is poor’ (Ryten (1998).. n. al. er. io. The 2005 publication focuses on the basic heading level, as Hill and Hill. i Un. v. (2009 p.192-93) describe, ’the lowest level of aggregation at which expenditure. Ch. engchi. weights are available. The basic heading price indexes provide building blocks from which the overall comparison is constructed. If these building blocks are biased or otherwise flawed, then everything that builds on them will be likewise tainted.’ It also introduced the structured product description method which improved comparability. For instance, 20 different varieties of rice were identified and used in East Asia for price comparison, according to the Asian Development Bank. However, as different varieties are consumed in different quantities from country to country, it is difficult to assess the right weight for every variety in the basket. Comparability trades in for representativity. Representativity means a good or service is representative for the basic heading it is categorised in, i.e. it has significant market share. Unfortunately the 8. For instance South East Asia or Sub-Saharan Africa !11.

(19) actual quantity of goods sold is only collected at the basic heading level, consequently which good or service is representative must be decided by local authorities. A generally applied assumption is the one that a good with a relatively low price probably has a larger market share than a relatively higher priced one, therefore are low priced items more likely to be representative. However, the assessment turned out to be very difficult to make for the 2005 ICP, especially in Asia and Africa. For every country, national average prices are calculated as an arithmetic average based on a larger number of quotations, that should be representative for the country as a whole. It it reasonable to to give lower weight to items with a higher price standard deviation, however this has not been done before 2014. The ICP 2011 Report drastically reevaluated the GDP (PPP) of many,. 政治大. especially relatively poor, countries, as Deaton and Aten (2014) show. However,. 立. besides their relative share of world GDP (PPP), it did not change much for most. ‧ 國. 學. developed countries, including Taiwan.. It should also be noted, that any item, that only has a listed price in one. ‧. country, has no influence on PPP as it lacks comparability.. y. Nat. Furthermore it is important to note that, even though the result of the PPP estimation. sit. is usually presented in USD, the US is not the benchmark other countries have to. n. al. er. io. compare to - as can be seen in the model below, the point of comparison is an. i Un. v. international or regional average price. Therefore PPP is base-invariant.. Ch. engchi. Though not the only method applied by the IMF today, the Geary-Khamis method models the fundamental concept of PPP calculations very well: International average price9:. Therefore a country’s purchasing power parity:. 9. Deaton & Heston (2008) !12.

(20) Definitions: = international average price good i !. = price good i in country j !. = quantity produced good i in country j !. !. = purchasing power parity country j. 學. ‧ 國. !. 政治大 = world production 立 good i The methods referred to by the ICP are GEKS, Jevons-GEKS*, Jevons-GEKS*(S),. ‧. CPD, CPRD, and weighted CPS methods. Whether or not those calculations are giving a biased picture, is a difficult question to answer. As PPP data seems plausible. Nat. sit. y. for most countries, with one exception that will be discussed later, they can be. er. io. assumed to be a good-enough approximation for the purpose of this research.. al. This ICP approach is also the basis for the data provided by the International. n. iv n C Monetary Fund: “The ICP provideshpurchasing powerU e n g c h i parities for a benchmark year (2011 in the most recent release). Outside the benchmark year, the IMF follows the. standard practice of extrapolating purchasing power parities using relative inflation rates based on GDP deflators.” (World Economic Outlook Update July 2014) As both OECD and World Bank Group do not publish data for Taiwan, IMF data will be used for all PPP considerations within this paper.. 2.3 Balassa-Samuelson-Effect and Penn-Effect The Balassa-Samuelson-Effect explains how under the law of one price there still can be different price levels in different countries (or territories). Technically it describes two effects (the Balassa- and the Samuelson-Effect) that both lead to the Penn-Effect. !13.

(21) Both effects assume that developed and non-developed countries have different levels of productivity for tradeable goods - which is plausible as the output per hour worked is the single most important variable for a country’s prosperity and hence marking the difference between developed and non-developed countries. Due to international competition and the law of one price, the price for tradeables is approximately the same in every country independent of its productivity (Kravis and Lipsey (1988)). Ballassa (1964) further assumes that for non-tradeable goods and services there are no significant differences in productivity. As many of those are labor-intense and productivity can hardly be increased, for instance restaurants, hairdresser etc., this assumption can be justified. However, the sector producing tradeable goods is the one defining the wage-level measured in international currency. Due to labor mobility -. 政治大. workers can decide in which sector to work, thereby equalising the wages for both. 立. sectors - the tradeable goods sector wages are adopted by the nontradeable sector. As. ‧ 國. 學. a result, less developed countries can produce tradeables at the same price more developed countries can, but non-tradeables are produced more cheaply (as their. ‧. relative price in the form of hours worked is lower). Since tradables are priced the. y. Nat. same, non-tradeables are priced lower, the average price level is lower. The existence. sit. of trade-barriers hence creates non-tradeable goods and services which then create. n. al. er. io. different price levels resulting in differences in purchasing power.. i n U. v. Samuelson (1964) builds up on Ballassa, adding that developing countries. Ch. engchi. increase their productivity foremost in the tradeables sector. The increase in income in the tradeables sector (increase in productivity while prices are international and constant) increases the prices of non-tradeable goods, resulting in fast developing countries facing a relative higher inflation. While the Ballassa-Samuelson-Effect provides the theory, the Penn-Effect is the data-based economic finding that the GDP per capita by market exchange rates exaggerates the differences in the living standard between different countries.10 Today, both are often used as synonyms.. 10. cf. Heston, Summers & Aten (2012) !14.

(22) 2.4 Fisher- and International Fisher-Effect11 Different countries with a different level of development often have different currencies with different monetary policy and different inflation rates. This however does not affect relative Purchasing Power Parity, as can be derived from the FisherEffect. The Fisher-Effect - named after Irving Fisher - describes the interest rate adjusting to the inflation rate - and vice versa. This relationship is proportional as investors add the expected inflation towards the real interest they want to gain in order to keep the return on investment constant.. !. 政治大 = change in interest rate 立 !. = change in inflation rate. ‧ 國. !. ‧. Therefore:. 學. !. !. = nominal interest rate !. = inflation rate. Ch. !. = real interest rate. n. al. er. io. sit. y. Nat. !. engchi. i Un. v. As long as inflation and interest rate are fairly low, their product can be neglected, leading to the well-known Fisher-equation:. !. 11. cf. Mishkin (2006) !15.

(23) The International Fisher-Effect translates the original Fisher-Effect into an openeconomy model. The two assumptions are that the Fisher-effect is given in both countries and that real interest rates are the same for both countries - which is plausible as the interest rate is the price for capital and according to the law of one price, a difference in real interest rates would lead to arbitrage opportunities.. !. !. !. 政治大 = nominal interest rate 立 !. = foreign country. ‧. ‧ 國. = home country. 學. !. n. al. er. io. sit. y. Nat. !. !. = exchange rate C h !. = period. engchi. i Un. v. This implies, a difference in inflation leads to a change in the exchange rate. As a result, inflation is counteracted by currency depreciation and leads to no relative change in income between the two countries. Inflation differentials do hence neither improve nor worsen the relative standard of living.. 2.5 Real / Nominal Exchange Rate Convergence. !16.

(24) The relationship between the nominal exchange rate and the exchange rate underlying PPP considerations is given by the law of one price. By now there is a vast amount of empirical research published about this relationship. There are findings rejecting a long-run equilibrium, for instance Taylor (1988) comes to the conclusion ‘in particular, we are unable to reject the hypothesis of noncointegration of the exchange rate and relative prices for any of the countries concerned. Far from finding a stable, long-run proportionality between exchange rates and relative prices, our results therefore suggest that they tend to drift apart without bound.’ However, most publications are in support of the PPP hypothesis. Abuaf and Jorion (1990) find ‘deviations from PPP, while substantial in the short run, appear to take about three years to be reduced in half’, which ‘follow from more powerful. 政治大. estimation techniques, applied in a multilateral framework’. Frankel and Rose (1996). 立. ‘re-examine deviaions from PPP using a panel of 150 countries and 45 annual post. ‧ 國. 學. WWII observations’ and find ‘strong evidence of mean-reversion that is similar to that from long time-series. PPP deviations are eroded at a rate of approximately 15%. ‧. annually, i.e., their half-life is around four years.’ Lothian (1997) uses ‘panel data for. y. Nat. the United States and 22 other OECD countries for the current float’ and shows. sit. ‘evidence that despite substantial short-term perturbations, purchasing power parity. n. al. er. io. actually performed much better than commonly believed.’ Recently, the most. i Un. v. predominant hypothesis assumes, that the half-life of deviations lies in the so-called. Ch. engchi. Rogoff’s (1996) puzzle interval, which is 3 to 5 years.. Decisive for the question of the validity of PPP / nominal exchange rate convergence is where the boundaries are set, i.e. which deviation is small enough to be explained by market imperfections and hence is able to resist the convergence in the long run. Since the first hike of real exchange rate research in the 1970’s and 80’s, PPP calculations have become significantly more sophisticated, narrowing the gap between nominal and real exchange rate, as Taylor and Taylor (2004) show. In fact, a lot of the difficulties clearly arise from the non-reliability of the data, especially pre-2005-data and data from non OECD countries, and sometimes also the limited time span available for time series analysis. The problem hence lies in the implementation, not in the model. Nevertheless, PPP is probably not a good tool to. !17.

(25) forecast exact exchange rates - not in the short run and not in the long run - but the general assumption proofed to be right: Nominal exchange rate gravitates towards the real exchange rate and the greater the deviation, the stronger the depreciative/ appreciative pressure in absolute terms. Figure 3.4.2 (later in the text) presents visual proof of the clustering of the real/nominal exchange rate ratio. . 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. !18. i Un. v.

(26) 3. Emergence of Taiwan’s Real / Nominal GDP Spread As seen in the introduction, Taiwan’s Nominal GDP fell behind Korea’s, while this was not true for its PPP adjusted GDP, which is due to the Fisher-Effect - and therefore relative purchasing power parity - not occurring. While all other developing countries’ price level converged over the last 20 years, Taiwan’s did not, making it an anomaly - the only one over that period.. 3.1 Inflation, Real Growth and Population Growth. 治政 Formula for real GDP growth is fairly straight forward:大立. Important for judging a countries economic progress is it’s real GDP growth. The. ‧. ‧ 國. 學. !. io. sit. y. Nat. !. = Real GDP Growth !. = Nominal GDP growth. n. al. er. !. Ch. engchi. i Un. v. approximately:. !. Nominal growth therefore can be split into real growth and inflation. Another important factor is population growth. Though not very large, any difference can be directly added or subtracted to/from GDP growth (approximately, the actual calculation is the same as for inflation adjustment) when talking about per capita numbers. As all growth data is exponential, over many years this effect has. !19.

(27) Figure 3.1 (b) Factors of South Korea’s Nominal GDP Growth 14%. 1.01%. 12% 10%. 6.21% 0.83%. 8%. 0.48% 0.52%. 6%. 3.97%. 3.34% 3.04%. 4%. 0.54%. 6.81%. 1.55%. 2%. 3.94%. 立. 1996-2000. 2001-2005. 2006-2010. 2.43% 2011-2013. 學. ‧ 國. 1991-1995. 4.23% 政治大 3.57%. ‧. Population Growth Average Inflation Average GDP Growth Average (adjusted for population growth). al. n 0.96%. 10% 8%. sit. io. 12%. er. 14%. y. Nat. Figure 3.1 (a) Factors of Taiwan's Nominal GDP Growth. Ch. engchi. i n U. v. 3.76% 0.85%. 6% 4%. 1.41% 6.53%. 2%. 1991-1995. 5.08%. 1996-2000. 0.36% 0.49% 0.68%. 1.19%. 3.55%. 3.93%. 2001-2005. 2006-2010. Population Growth Average Inflation Average GDP Growth Average (adjusted for population growth) !20. 0.29% 0.83% 2.40% 2011-2013.

(28) significant impact on the living standard. Not that population growth necessarily affects the living standard negatively, but failing to utilise the increasing headcount towards increasing output means sharing a fixed amount of resources by more people. Therefore, published GDP growth data should always be adjusted for population growth - unless otherwise stated - just like it is always adjusted for inflation.12 Doing the split for Taiwan and Korea (Figure 3.1 (a) and 3.1 (b)) uncovers the fact that both grew by about the same rate in real terms,while Korea’s inflation was consistently higher, as was its population growth after 2005. While the difference in population growth is not connected to the difference in GDP (PPP) to nominal GDP ratio, it added about 1.6% to Korea’s relative to Taiwan’s GDP in the period after 2005, which is about two thirds of the recent yearly growth average. In theory,. 政治大. inflation is taken account for by the Fisher-effect - as shown before - and therefore has. 立. no longtime effect on the PPP ratio.. ‧ 國. 學. 3.2 International Fisher-Effect and the NTD Nominal Exchange Rate. ‧. y. Nat. Comparing the actual exchange rate for KRW/TWD with the one calculated by the. n. al. er. io. exchange rate, that is more stable than predicted.. sit. International Fisher-effect based on the actual exchange rate for 1990 shows an. i Un. v. As Figure 3.2 shows, both financial crises had a large, however temporary,. Ch. engchi. impact on the exchange rate. While being highly volatile, in its nature more volatile than its purchasing power parity adjusted counterpart, the nominal exchange rate ignoring crisis effects - maintained close to the expected rate under Fisher until the early 2000s, after which they disperse. For 2013 the difference between the nominal and the expected exchange rate lies at 18% (base is current nominal exchange rate). The divergence proves to remain starting at any point between 1990 and 2000. Since the exchange rate between the New Taiwan Dollar and the Korean Won, and between the New Taiwan Dollar and the US Dollar during that period is plausible (under PPP considerations, as will be shown later), the trend appears to be robust.. 12. Resistance towards doing so might stem from political considerations. !21.

(29) Figure 3.2 KRW/TWD Exchange Rate 45. 40. 35. 30. 立. 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012. 學. Exchange rate KRW/TWD (yearly avg) Rate adjusted for Fisher-effect Asian Financial Crisis World Financial Crisis. ‧. ‧ 國. 25. 政治大. sit. y. Nat. The exchange rate towards special drawing rights would be a more neutral. io. er. measure due to it being a composition of major currencies, however inflation is difficult to assess for special drawing rights and differences in monetary and fiscal. n. al. Ch. i n U. v. policy, each with different lags and implications, within the countries taking part in. engchi. constituting the SDRs, reduce the expressiveness of its exchange rate. As under Purchasing Power Parity the Korean Won today is still on the edge of being undervalued - it’s GDP (PPP) to nominal GDP ratio is the third highest among the developed countries, with both higher rated countries being poorer13 - the Korean Won is a good proxy and a very conservative point of comparison.. 3.3 Real / Nominal GDP Ratio over Time - Taiwan’s Case. 13. See Appendix for individual country ratios. !22.

(30) Figure 3.3 PPP/Nominal GDP Ratio Taiwan and Korea 2.2. 50,000. 2 40,000. 1.8 1.6. 30,000. 1.4 20,000. 1.2 1 0.8. 立. 政治大. 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 2010 2013. 學. Ratio Korea Korea GDP (PPP) per capita (current) Taiwan GDP (PPP) per capita (current) Ratio Taiwan. ‧. ‧ 國. 0.6. 10,000. sit. y. Nat. During the early 1980s Taiwan’s PPP/nominal GDP per capita ratio increased and. er. io. maxed out at 1.87 in 1985 (Figure 3.3). This seems quite high, however during these days Taiwan was far behind the developed world, hence this ratio does make sense,. n. al. Ch. i n U. v. furthermore it moved parallel to South Korea’s ratio. After that, both country’s ratio. engchi. declined, until around 1995, where it began to increase again. South Korea’s spike in 1998 can be attributed to the Asian Financial Crisis and it’s effect on the exchange rate of the South Korean Won. In US-Dollar Korea’s GDP faced one of the most sudden declines in world history, however the real living standard was affected less strongly (but still severely). After 2000 Taiwan’s ratio started to deviate from Korea’s as it continued to increase, while the latter, which showed a pattern familiar to the Asian Financial Crisis around the time of the World Financial Crisis, followed an overall slightly decreasing trend. Korea’s ratio meets the Balassa-Samuelson-effect, while Taiwan’s doesn’t, as it’s trend should have been vice versa and nominal GDP should have grown significantly faster than GDP (PPP), since an increase in the former creates inflationary pressure on non-tradeables and hence increases the price !23.

(31) level resulting in slower growth of the latter. While for PPP/nominal GDP considerations deviations are seen frequently - as exchange rates are changing nearinstant, while prices are slow to adapt and momentary price levels are always unknown (their assessment has a time lag) - they are usually only temporary. Arbitrage does not only need a price difference, but it also requires the acknowledgement of the difference and in practise also time to be set up efficiently. What is most surprising about Taiwan’s ratio therefore is not only how high it is, but also how long it already persists on such a high level. From a time series perspective Taiwan’s current situation is not an singularity, but the result of a continuous trend.. 3.4 International Price Levels. 立. 3.4.1 All Countries. 政治大. ‧ 國. 學. Figure 3.4.1 shows all countries the IMF Economic Outlook 2014 provides data for.. ‧. Unfortunately, many of them are estimates, also many are not reflecting 2013 but. y. Nat. previous years. The data used is the most recent (up to 2013) for each country.. io. sit. It is arguable, whether to use a countries nominal GDP or GDP (PPP) for the x-Axis,. n. al. er. i.e. the independent variable. In literature, usually GDP is used as a proxy for the. i Un. v. standard of living and motivation to increase GDP / to generate growth is derived. Ch. engchi. from its relationship to the general well-being. As shown earlier however, GDP (PPP) is a much better proxy for well-being as it actually measures the amount of goods and services that can be consumed. Nominal GDP though is much easier to calculate and a significantly older account, hence its dominance in academic and other literature. Looking into PPP and using it to generate the ratio variable, it would be unreasonable to do the step back and use nominal GDP instead of an PPP adjustment, as PPP is a direct reflection of the amount of goods and services a country produces, while nominal GDP depends on PPP and exchange rate. The countries are scattered a lot, with a ratio up to about 4 (Myanmar at 3.90) and a minimum value of about 0.6 (Norway at 0.64). While there is a generally decreasing trend, it is not too obvious and low income countries face a huge range. !24.

(32) Figure 3.4.1 PPP/Nominal GDP Ratio All Countries 4. 3. 2. 1. 立. 25,000. 50,000. 75,000. 100,000. 125,000. 150,000. 學. ‧ 國. 0. 政治大. (from less then 1 to almost 4). To calculate a trend line, I will use a logarithmic least-. ‧. square regression, as neither the Balassa-Samuelson-Effect nor later literature suggests any specific relationship, i.e. exponential, power, standard etc., while. Nat. sit. y. logarithmic provides the best fit and is most plausible, as it normalises the differences. al. er. io. in base values14. The x-axis is displayed in linear style in order to not distorted the. n. visibility of countries’ relative living standard differences, hence the regression line’s curve.. Ch. engchi. i Un. v. The resulting equation is:. y = -0.2998ln(x)+4.6122 (s.e.). (.3565). R2 = 0.2799. (.3294). While there is significant literature on the mean-reversion of real exchange rate (see previous chaper), there aren’t many findings showing what causes them in the first place, except for various economic shocks (e.g. Sager (2006)). While Aizenman 14I.e.. the change from 1 to 0.9 does not equal the change from 0.1 to 0, but from 0.1 to 0.09 etc. !25.

(33) (1983) shows that a higher price flexibility reduces deviations, it remains unknown to what extent deviations can maintain in the long-run and what causes these long-term deviations (besides the Ballassa-Samuelson-Effect). It is therefore - at the current state of knowledge - not possible to the describe a model that explains the deviations found in Figure 3.4.1, i.e. the prediction quality cannot be improved by relying on established theory.. 3.4.2 Advanced Economies There are a variety of explicable reasons however to exclude a number of countries. For one, most rentier states, like most Arabian oil-exporting countries, show abnormal. 政治大. ratios due to their governmental income relying only on ex- and imports rather then. 立. on income or value-added taxation, therefore having an artificially low price level.. ‧ 國. 學. Furthermore, their income does not derive from productivity, but from selling their resource stock, meaning wages do not need to rise, as labour is not a key determined. ‧. of the country’s GDP. Also any kind of price-fixation obviously alters the relative. y. Nat. price level. Small-countries and underdeveloped countries with a large wealth and. sit. income inequality can also provide non-tradeables cheaper than what would be. n. al. er. io. expected by the average income level as the mean and median income disparity is. i Un. v. larger, implying that most productivity benefits are grasped by the upper class and do. Ch. engchi. not translate into increasing wages properly. The same goes for countries with a large share of foreign workers that are paid significantly below the average local income level. The necessity of at least a managed exchange rate float is apparent, so is political stability, and public order and safety. Regular and in-depth surveys are necessary to acquire reliable data. As mentioned before, PPP calculations have been criticised a lot concerning their quality and as with most economic data, a countries developmental stage strongly determines the quality of its data collection. As the Regional Offices heavily rely on the National Agencies, their capability translate into the quality of the data output at the Regional and Worldwide level. Furthermore it seems plausible that less developed countries face a greater share of goods traded that is not standardised, which makes is difficult or impossible to assess an international. !26.

(34) Figure 3.4.2 PPP/Nominal GDP Ratio Advanced Economies 2 1.75 1.5 1.25 1 0.75 0.5 0.25. 立. 10,000. 20,000. 30,000. 40,000. 50,000. 60,000. 70,000. 學. ‧ 國. 0. 政治大. average price for them, hence they get ignored for PPP considerations. A lower level. ‧. of urbanisation/centralisation creates difficulties collecting prices in a way that they are representative for the whole nation (as weighting is important). Last but not least. Nat. sit. y. all mentioned theory underlies the efficient market hypothesis, an assumption that. er. io. cannot be made for all countries.. al. It is arguable which country falls under which restriction. For instance, no. n. iv n C country has a perfectly efficient market. to prevent any suspicion of h e nIn gorder hi U c manipulation, I will refer to a commonly accepted grouping, which is all countries which are defined as developed by the IMF. As there is no data provided by the IMF for San Marino, i will also exclude it. Finally I will exclude city states as they face extraordinary circumstances like usually not having any industrial sector, no rural population and infrastructure, no resources and agricultural sector and special opportunities concerning taxation and migrant workers, which frequently makes them outliers when looking at a variety of indicators. In this case, the specific reason for each case are:. - Hong Kong: De facto fixed exchange rate, which means the spread between real and nominal exchange rate is not defined by the market.. !27.

(35) - Singapore: The large share of foreign migrant workers in low paying jobs implies that sectors with low productivity growth do not need to compete for employees with the high growth sectors, therefore the assumptions for the Balassa-SamuelsonEffect are not given. In a sense, ‘non-tradeables’ are cheaply imported by Singapore, which makes it a special case and disturbs comparability with other countries.. - Luxembourg: As a city state in the center of the European Union it faces a situation similar to Singapore. Low geographic extension in combination with free labor and goods movement and the use of the same currency as neighbouring countries, many of the goods consumed within country borders are likely to not be bought within the country and many jobs are likely to be given to workers living outside. 政治大. the country, including jobs in the ‘non-tradeable’ sector.. 立. Doing so leaves 32 countries (Figure 3.4.2) with a GDP (PPP) per capita. ‧ 國. 學. between roughly 20,000 and 65.000 and a ratio between 0.64 (Norway) and 1.99 (Taiwan). The logarithmic trend function becomes:. io. y. al. n. 3.4.3 Anomaly Taiwan. sit. (1.7794). er. (.1693). Nat. (s.e.). ‧. R2 = 0.3925.. y = -0.7452ln(x)+8.8753. Ch. engchi. i Un. v. If we now remove Taiwan from the sample, the function becomes:. y = -0.8145ln(x)+9.569502 (s.e.). (.1001). R2 = 0.6955. (1.0515). It is worth mentioning that the next largest outlier is the US with a ratio of 1 (as defined) and a predicted ratio of roughly 0.75. This is 2013 data, looking at the exchange rate shift of the US dollar at the end of 2014 and early 2015, we can expect the US now having a roughly 20 percent higher price level compared to most other currencies then at 2013. Therefore the US has recently moved much closer towards. !28.

(36) the ratio that is predicted by the model, hence supporting the purchasing power parity hypothesis and making Taiwan an even more extreme outlier by increasing the distance between the difference of Taiwan’s predicted and actual ratio and the country with the second largest difference, and significantly increasing R2 (to roughly 0.74). Therefore, not only visually, but also statistically, Taiwan is the only strong outlier. It is however possible, that those random deviations temporarily occur over time. I already showed that Taiwan’s ratio grew over time, it is hence not a random shortterm/temporary movement or shock. Nevertheless it makes sense to make a time series analysis in order to see whether or not other deviations with comparable impact occurred. Unfortunately, four major effects blur any look into the past of PPP ratios:. 政治大. Firstly, as the calculations change, so does their outcome. With regular revisions of. 立. how prices are assessed and compared between countries and regions, data between. ‧ 國. 學. years becomes less comparable the larger the time difference. The obvious reason for those calculations to be changed is that they are improved. This might be put into. ‧. question at a detail by detail inspection - parts are controversial15 - but on a larger. y. Nat. scale the tendency is unambiguous. Consequently we can expect more errors in the. sit. price level estimations in the past, than today. Therefore, slightly higher deviations are. n. al. er. io. to be anticipated, and we have to bear that in mind when comparing that data to Taiwan’s situation today.. Ch. engchi. i n U. v. Secondly, the definition of which country is developed and which is not, is quite arbitrary. That counts for today as it does for the past, however defining a year for which a country started to be a developed country is a difficult step to make, and one that no matter what definition is used, will always be impossible to defend. I will therefore stand back from doing so and consider all countries that the IMF defines as developed today, as developed countries for the whole time series. This is an approach shared by the IMF, which also does not give definite dates, but only the categories, as the IMF does not imply a country became developed the moment or year the IMF 15. Especially the recent change in 2011, which resulted in a huge boost of PPP income for emerging economies relative to developed economies. This is another reason why this research focuses in developed countries as those are only marginally affected by this change. !29.

(37) recognises the fact it now is. Hence, the further back our observation goes back, the more countries will show characteristics that we associate with a lower level of development, as they have been catching up during most of the period, and therefore the data becomes less reliable. Thirdly, PPP calculations start with gathering huge amounts of information on which goods are sold how frequently and for what price. This requires a high level of collaboration between ICP staff and local authorities. It is not unlikely to assume a notable learning curve here, as management practises might strongly differ due to different sociocultural backgrounds, and while the private sector is fast to adapt best practises, the public sector - which is the one involved here - is notorious for slow adaption and more legacy work practises. Also on the individual employee level we. 政治大. can assume a learning curve, with its effect stronger the further we go back in time.. 立. Fourthly, a significant number of countries has not been part of the First World. ‧ 國. 學. during the period of the Cold War. As we know, Communist countries usually do not provide an efficient market, the price mechanism is distorted and in some instances. ‧. not existent at all, which means the assumptions we need for the law of one price are. y. Nat. not met. It also is plausible to assume a time lag, as the reforms of former Soviet. n. al. er. io. sit. Block countries took many years and their effects took many years more.. i Un. v. On order to assess whether there have been other anomalies besides Taiwan, I will. Ch. engchi. compare a time series analysis conducted with Taiwan and one without Taiwan. Using an ordinary least square regression in the form of16:. !. !. 16. cf. Agresti and Finlay (2008) !30.

(38) Figure 3.4.3 Price Level Time Series 3. 1.00. 0.75 2 0.50 1 0.25. 立. 政治大. Beta (excl.Taiwan). R^2 (incl. Taiwan). 學. ‧ 國. 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 R^2 (excl. Taiwan). ‧. Starting in 1990, R2 starts off very low and Beta (excl. Taiwan) jumps from extremely. sit. y. Nat. low to extremely high and finally to a plausible level (Figure 3.4.3). This is due to bad. io. er. data quality and the step by step addition of the former Soviet controlled Eastern European states. Hence it makes sense to only interpret the findings from about 1995. n. al. Ch. i n U. v. onwards. As can be seen, until 2003, R2 and R2 including Taiwan are very close, but. engchi. then spread permanently. This is caused by Taiwan’s increasing ratio at that time. Also this is the only significant drop in R2, implying another outlier like Taiwan has never occurred during that period, i.e. Taiwan’s ratio relative to its income level is unparalleled taking into account the current price diversity within the developed world. Interestingly, Beta shows a declining trend. As the function is logarithmic, the implications of this change are more formidable than they appear on first sight. Also, the period of strong decline in Beta between 2002 and 2008 has its timing conspicuously close to the decline in R2 including Taiwan (roughly 2003 to 2008). This concludes Taiwan became an anomaly at around the time when the price differences between the other countries became smaller and the spread between !31.

(39) Taiwan’s actual ratio and the expected ratio stabilised at around the time price level differences between the other countries stabilised too.. 3.5 Causes Overview After having seen that, and investigated how, the next step is to ask why Taiwan’s price level diverted from those of the other advanced economies. This puzzle can be solved by looking into a number of factors that are connected, but work parallel on different levels. First, in section 4 I will look into on-going international integration in which Taiwan did not take part as extensively as the other advanced economies. Next,. 政治大. Taiwan’s monetary policy needs to be taken into account, in its official and actual. 立. version, which shows significant intervention and sterilisation efforts. I will do so in. ‧ 國. 學. section 5, in which I will also take a look at Taiwan’s special relationship with China. This relationship turns out to set the foundation for many of the other forces causing. ‧. the emergence of Taiwan’s exchange rate spread. With this in mind, the balance of. y. Nat. payments can be interpreted in section 6 and will give a reasonable answer to the. sit. aforementioned question. Finally, in section 7 I will discuss other price level. n. al. er. io. considerations, for instance Taiwan’s competitive national environment. . Ch. engchi. !32. i n U. v.

(40) 4. International Integration International trade increased tremendously since the end of the Cold War, so did the amount of bilateral trade and investment agreements (and some large free trade areas). This can explain the convergence of price levels and why Taiwan did not take part: The bilateral agreements it signed are fewer in numbers and lesser in significance.. 4.1 Factor Price Equalisation As seen in Figure 3.4.3, the price difference amongst advanced economies has. 治政大 technologies and thereby than those at the technological frontier by adapting foreign 立 increasing total factor productivity - which also was true for Taiwan (Drysale and decreased over time. For one, because less developed countries tend to grow faster. ‧ 國. 學. Huang (1997)). This effect is well documented, and though it is usually applied to emerging market economies, it can be applied to many economies within the group of. ‧. countries this research focuses on. Dowrick and Nguyen (1989) show the catch up. sit. y. Nat. effect for OECD countries in the 1950-85 period, a process that was still in continuation for many of the sampled countries at the end of the last century, and for. io. n. al. er. some still is today. This then, ceteris paribus, including the Balassa-Samuelson-Effect,. i Un. v. leads to more narrow price levels. This effect seems to be the stronger, the higher the. Ch. engchi. level of integration among the countries, as Rapacki and Próchniak (2014) show for the European Union. Furthermore, the change in the regression Beta (the change of real/nominal exchange rate per unit increase in (logarithmic) GDP (PPP) per capita, prior section) implies that not all of the price level convergence can be due to income level convergence, a very large part derives from a change in the function that describes the correlation between the income and price level, i.e. per capita GDP income had a weaker effect over time on the real / nominal exchange rate ratio. To interpret this, the factor price equalisation theorem provides some valuable insight. It is in its nature very similar to the law of one price, though more differentiated and more specific in its limitations. It was first suggested by Heckscher !33.

(41) (1919), and though he wasn’t perfectly right - as factor price equalisation does not work under his assumptions, which were immobile factors and perfectly mobile products - he gave the inspiration for a core element of international economics, which was later extended by Ohlin (1933) and Samuelson (1948). The most talked about condition of the theorem today is the so-called lens condition (Deardorff (1994)). As a result, the theorem becomes more unlikely, as the count of factors and commodities introduced increases (Hansen and Keiding (2008)). Whether or not there is perfect or near perfect factor price equalisation is hence still discussable. “Economists who believe that markets work well (or better than any alternative) believe the general argument that free trade will tend to equalise the return to similar factors internationally. Economists who are suspicious of markets. 政治大. will not trust free trade to create a more equitable international distribution of. 立. income” (Rassekh and Thompson (1993)). It offers however one very good. ‧ 國. 學. explanation of why the relationship between real/nominal exchange rate ratio and (logarithmic) GDP (PPP) per capita changed. Assuming the real exchange rate is. ‧. solely determined by a countries productivity, a change in the real / nominal exchange. y. Nat. rate ratio under a given purchasing power income between two countries translates. sit. into a change in relative price levels, i.e. under fixed PPP income difference, a. n. al. er. io. decrease in ratio difference means a decrease in price level difference. Furthermore,. i Un. v. factor price equalisation provides a connection between an increase in income for. Ch. engchi. countries behind the technological frontier, i.e. lower income countries with lower productivity, and a decrease in the price level of countries closer to the technological frontier, i.e. high income countries with higher productivity. Outsourcing of labour intensive production increases wages in the former, and lowers production costs for goods imported into the latter (relative to local production). It can hence explain both, the higher growth and catch-up of countries with lower per capita income and related narrowing of the price level, and the price level converging due to lesser dependency on GDP (PPP) per capita. Factor price equalisation is a sufficient, however not necessary condition for price convergence. If all factors are priced the same and the technology to produce a certain good is the same everywhere, then the resulting minimum production cost will. !34.

(42) Figure 4.1 International Trade $30t $25t $20t $15t $10t $5t. 立. 政治大. 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012. ‧ 國. 學. Merchandise. Services. ‧. be the same. Relative differences in factor prices can however also coexist with equal product prices, as the factor price differences can cancel each other out, different. y. Nat. al. er. io. advantage can create market entry barriers, and so on.. sit. production methods or economics of scope can be applied, sunk costs / first mover. iv n C Technological Catch-Up), all of them international trade in connection with a h erelyn on gchi U. n. Whichever explanation is used (Law of One Price, Factor Price Equalisation,. reduction of trade barriers.. Figure 4.1 displays the dramatic growth of international trade over the last two decades, from 4.3 trillion USD in 1990 to 23.5 trillion USD in 2013. 2009 shows a dramatic slump due to the World Financial Crisis, however international trade recovered quickly. Trade probably not only directly affects prices by producers/ consumers undercutting/outbidding competition, but also due to the price pressure that potentially arises. For instance, a supplier or group of suppliers might deliberately decrease their prices to a point where they make still profit, however a new market participant could not recover his required upfront investment.. !35.

(43) Therefore a decrease in actual trade might not instantly lead to a change in the price level, which is an explanation for the dent in trade in 2009 not correlating with a significant change in the relationship between the real/nominal exchange rate ratio and GDP (PPP) per capita (Beta) for the same year. Furthermore an increase in total market size due to foreign expansion can offset a decrease in local prices, which might become necessary due to a lower purchasing power of the foreign population and the danger of re-imports. Last but not least, international trade can increase market efficiency, convert monopolies to oligopolies and oligopolies to atomistic markets as well as break up established price collusions and corruption. This becomes even more true when talking about financial markets.. 政治大. Thomas Friedman, columnist for the New York Times and three times Pulitzer. 立. Prize winner, became well-known in the field of political sciences with his ‘Golden. ‧ 國. 學. Arches Theory of Conflict Prevention’. In the same book, ‘The Lexus and the Olive Tree’, he uses the following metaphor:. ‧. As the world becomes narrower due to political cooperation and technological. y. Nat. advance, countries now face a new opportunity to change to the economic fast lane:. sit. They can open themselves up to the electronic herd, which is international capital.. n. al. er. io. However, they have to offer something to attract the herd and once it doesn't feel. i Un. v. comfortable anymore, it just leaves, usually with a domino affect, a part of the herd. Ch. engchi. gets scared, the rest just follows blindly - as happened during the Asian Financial Crisis. There are short-horn cattle, i.e. international traders, and long-horn cattle, i.e. multinational corporations, and though the one moves faster than the other, both tend to move into the same direction. In order to become attractive though, a country needs to fasten the ‘Golden Straightjacket’ and update its ‘Operation System’, which means introducing modern western economic and political systems. There now is both: a role model on what to change, and an incentive. Not only will that converge income levels, but it also will improve market efficiency and equalise production circumstances and hence functions, marginalising inter-country differences, including price levels, ceteris paribus.. !36.

(44) 4.2 Economic Integration Economic integration aims at reducing or eliminating trade barriers, benefitting producers and consumers and therefore overall welfare. Prakash and Hart (2000) describe two approaches to assess the level of economic integration. The first way “examines levels of institutional convergence or harmonisation across countries. The underlying argument is that countries can potentially integrate if minimal institutional obstacles to cross-border economic flows are present.” These obstacles manifest in the form of tariffs and non-tariff barriers as well as capital and investment controls. “The second approach focuses on the outcomes of integration”, Cheung, Yiu and Chow (2008) label them “quantity-based”. The most widely used indicators are trade. 政治大. openness and the flow of foreign direct investment. The rest of this section focuses on. 立. the first approach, trade openness and FDI will be referred to in section 6, where also. ‧ 國. 學. other components of the balance of payments will be analysed.. Balassa (2013) differentiates the following stages of economic integration (in. ‧. increasing order):. y. Nat. - Common Market. er. al. n. - Customs Union. io. - Free Trade Area. sit. - Preferential Trading Area. Ch. engchi. i Un. v. - Economic Union - Economic and Monetary Union - Complete Economic Integration. Most countries are still far away from the later stages, these tend to be rather the exception than the rule today. One notable outlier, today’s most progressed form of economic integration, is the European Union, despite its recent crisis, debt-wise and in regards to increasing nationalist tendencies in many countries, especially the United Kingdom, creates new questions whether this level of integration is maintainable. The speed of future regional and global integration is therefore hard to predict.. !37.