Inequality and Corruption: The Role of Land Reform in Korea, Taiwan, and the Philippines
University of California, San Diego
This paper explores how inequality increases corruption via electoral clientelism, bureaucratic patronage, and elite capture of policy process through a comparative historical analysis of South Korea, Taiwan, and the Philippines that shared similar conditions at the time of independence. It finds that success and failure of land reform, which was little affected by corruption but largely determined by exogenous factors such as external communist threats and US pressures for reform, produced different levels of inequality, which in turn influenced subsequent levels of corruption through capture and clientelism. In the Philippines, failed land reform maintained high inequality and domination of the landed elite in both politics and economy, which led to persistent political clientelism, increasing patronage in bureaucracy, and policy capture by the powerful elite. In contrast, successful land reform in South Korea and Taiwan dissolved the landed class and produced egalitarian socio-economic structure, which helped to maintain state autonomy, contain clientelism, promote meritocratic bureaucracy, and develop programmatic politics over time.
Keywords: inequality; corruption; clientelism; capture; land reform; Korea, Taiwan, and the Philippines
When Korea, Taiwan and the Philippines gained independence after World War II, they were all similarly poor; if anything, the Philippines was ahead of the other two countries in terms of per capita income as well as educational attainment. Since that time, the remarkable economic growth of South Korea (hereafter Korea) and Taiwan during the late 20th century has often been contrasted with the dismal performance of the Philippines. The Philippines is also known for its high level of corruption and crony capitalism, while Korea and Taiwan have been widely recognized for good governance. Why have Korea and Taiwan developed into rich countries with good governance, while the Philippines remains relatively poor and with poor governance?
The international development community emphasizes the importance of good governance, including control of corruption, for successful development. Empirical studies have found a negative effect of corruption on economic growth (Mauro 1995). The Philippines has been called a predatory state characterized by crony capitalism, even “booty capitalism” (Hutchcroft 1998). Korea and Taiwan have been praised by many scholars as a model developmental state with a competent and relatively uncorrupt bureaucracy (Amsden 1989; Evans 1995; Haggard 1990a; Johnson 1987; Wade 1990). A variety of evidence shows that, at least since the early 1980s and probably from a much earlier period, the Philippines has suffered from a very high level of corruption, while Korea and Taiwan have maintained much lower levels of corruption than the Philippines.
Thus, differences in governance, including control of corruption, may explain the divergent developmental paths of Korea and Taiwan on the one hand and the Philippines on the other hand. But what, then, explains corruption? Although cross-national studies have identified various possible causes of corruption such as economic development, democracy, religion, ethnic heterogeneity, trade openness, and income inequality, these studies have been plagued by well-known endogeneity problems (Treisman 2007; You & Khagram 2005). It is not easy to solve the endogeneity problems through cross-national study due to lack of sufficient longitudinal data and plausible instruments for endogenous variables.
This paper takes a different approach. It explores the causal relationship between inequality and corruption through a comparative historical analysis of Korea, Taiwan, and the Philippines. Comparing historical sequences can be a powerful tool for analysis of causal pathways (Rueschemeyer & Stephens 1997). The three countries are ideal comparison cases because they share a number of similar initial conditions such as levels of economic development. They all experienced colonial rule before World War II, and were all heavily supported by the United States during the Cold War era. They all had democratic transitions in the late 1980s. Yet they had quite different corruption outcomes. Thus, these countries satisfy the requirements of the ‘most-similar cases’ (Gerring 2007: 131-134).
I argue that different levels of inequality caused different levels of corruption in these countries. I propose two broad causal mechanisms from inequality to corruption: capture and clientelism. I provide reasons why high inequality leads to state capture by the powerful elite on the one hand and clientelistic politics on the other hand. I approach the problem in two steps: some simple tests of the plausibility of alternative explanations; and an historical analysis that addresses the potential endogeneity of inequality to corruption. My process tracing of the three countries not only establishes causal direction from inequality to corruption but also corroborates the proposed causal mechanisms.
It is well known that the Philippines had a very high level of inequality in income and wealth, but Korea and Taiwan maintained unusually equal distribution of income and wealth over the postwar period. Examination of simple correlations between corruption and various possible causes of corruption shows that inequality is the only one that is correctly correlated with corruption across the three countries. However, the direction of causality might have run the other way: from corruption to inequality. While some cross-national studies found a significant effect of inequality on corruption (Easterly 2007; You and Khagram 2005), other studies have found causal effect of corruption on inequality (Gupta, Davoodi, & Alonso-Terme 2002; Li, Xu, & Zou 2000). Hence, the identification of causal direction requires that different levels of inequality across the three countries be at least to some extent exogenous to corruption.
In fact, the differences in inequality across the three countries are to some extent exogenous and can be traced to the success of land reform in Korea and Taiwan and its failure in the Philippines. If the course of land reform had been determined by corruption, then endogeneity problems would remain. However, the external determinants of land reform permit a natural quasi-experiment. I find that the success (in Korea and Taiwan) and failure (in the Philippines) of land reform was primarily determined by exogenous factors such as communist threat from North Korea and mainland China as well as US pressures for reform. The success and failure of reform produced starkly different levels of inequality, which in turn influenced corruption through capture and clientelism.
This paper is organized as follows. First, I propose causal mechanisms through which inequality increases corruption. Second, I assess various pieces of available evidence on the relative levels of corruption in the three countries. Then, I examine various possible explanations for different levels of corruption across these countries and single out inequality as a plausible explanation. In the subsequent sections, I explore how different levels of inequality were produced by the success and failure of land reform, and how different levels of inequality have influenced subsequent levels of corruption through capture and clientelism. The final section summarizes my findings and concludes with research and policy implications.
WHY DOES INEQUALITY INCREASE CORRUPTION?
Corruption is commonly defined as “misuse of public office for private gain” (Rose-Ackerman 2008). Corruption takes many forms such as political and bureaucratic corruption, depending on the types of public officials-elected or appointed- that are involved in corrupt acts, and petty and grand corruption, depending on the magnitude of corrupt transactions. Corruption is, by definition, illegal, although one can consider some legal acts and practices corrupt such as political influence of big money and conflict of interests.
Why does inequality increase corruption? I argue that inequality increases corruption by encouraging capture and clientelism. Capture means that the state or a specific government agency has lost autonomy and serves the special interests of the elite or those that are supposed to be regulated. Although state capture can occur without illegal corruption, for example through legal campaign contributions, capture by the private interests often involves illegal exchange of governmental favors and illegal campaign contributions or personal bribery. In clientelistic political systems, politicians (patron) and voters (client) exchange votes for particularistic benefits. Clientelism is not necessarily illegal or corrupt, but it often involves vote buying and provision of patronage jobs in the public sector in violation of merit rules.
Inequality leads to capture of politics and bureaucracy, which increases political and bureaucratic corruption. Inequality also encourages clientelism, which involves vote buying and patronage appointments in the bureaucracy, and the former increases political corruption and the latter bureaucratic corruption. Figure 1 outlines the causal mechanisms.
[Figure 1 about here]
First, higher inequality would lead to higher redistributive pressures, which would in turn give the wealthy incentives to capture the state to defend their interests (You and Khagram 2005). In the absence of capture, high inequality should lead to high levels of redistribution, as for example in the Meltzer-Richard model (Meltzer & Richard 1981). But in highly-unequal settings, the wealthy have strong incentives to buy political influence and employ corruption to minimize taxation and redistribution. Capture can occur at the highest level of policy-making process as well as at the policy implementation process, encouraging grand political corruption as well as bureaucratic corruption.
Acemoglu and Robinson (2008) proposed a model of “captured democracy,” in which de jure political power of citizens is offset by de facto political power of the elite based on lobbying, bribery, and use of extralegal force. Although they did not discuss the role of inequality in this version of their model, “captured democracy” is more likely at higher levels of inequality because of higher stakes and greater expected returns for the elite from controlling politics. Ziblatt(2009) found that land inequality led to capture of local institutions by landed elites, which in turn led to electoral fraud in the late 19th-century Germany.
Second, as Robinson and Verdier (2013) argued, clientelism becomes an attractive political strategy in situations of high inequality. Under high levels of inequality, the wealthy elite will not want programmatic politics to develop, because programmatic competition is likely to strengthen leftist parties that will jeopardize their interests. The rich have incentives to develop clientelistic politics as an alternative. Also, high inequality means a large proportion of poor people, and the poor are vulnerable to clientelism (Scott 1972). The poor may be better off by developing programmatic politics, but they usually suffer from collective action problems. It makes perfect sense for poor individuals to engage in clientelistic exchange of their votes with particularistic benefits such as cash, gifts, and patronage jobs in the public sector even if it is collectively irrational.
Vote buying itself is not only a form of petty electoral corruption, but it also increases large scale political corruption by encouraging politicians to solicit illicit campaign contributions. Moreover, clientelism is likely to protect corrupt politicians, because clientelism converts principal-agent relationship between voters and politicians into patron-client relationship. While programmatic competition will enable voters (principal) to hold corrupt politicians (agent) accountable, under clientelistic competition voters (client) will not be able to punish corrupt politicians (patron).
Clientelism typically increases patronage appointments in the public sector, and the prevalence of patronage in the bureaucracy will increase bureaucratic corruption. Those bureaucrats who have entered the bureaucracy via patronage appointments are less likely to be committed to public interest than those who have become civil servants via open competition on merit. The former are more likely to try to get promotions via patronage or by bribing their bureaucratic bosses and political patrons. They are also more vulnerable to corrupt demands from their political patrons.
Capture and clientelism may be the main reasons why many democracies, especially young democracies, fail to control corruption. These problems are likely to be more serious in countries with higher levels of economic inequality. Also, the effect of inequality on corruption is likely to be stronger in democracies and authoritarian regimes with elections than in hard-authoritarian regimes, because competitive elections and democratic policy-making processes are more vulnerable to clientelism and capture.
3. RELATIVE LEVELS OF CORRUPTION IN KOREA, TAIWAN, AND THE PHILIPPINES
It is not easy to compare corruption across countries. Corruption, by its nature, is conducted secretly, and the probability of exposure will be different in different countries. Most empirical studies of corruption rely on measures of perceived levels of corruption such as Transparency International’s Corruption Perceptions Index (CPI) and Kaufmann, Kraay, and Mastruzzi’s (2010) Control of Corruption scores (CC). The CPI and CC are aggregate indices based on a variety of surveys of mostly business people and ratings by country experts. Because perceptions are subjective and prone to bias, there must be substantial measurement error. Moreover, the definitions of corruption employed by various cross-national measures of perceived corruption are often vague and conceptually different from one another, but most of the measures are highly correlated with one another (Lambsdorff 2008).
Although TI has been publishing CPI annually since 1995, it has also published “historical” CPI data for the periods of 1980-85 and 1988-92. Table 1 shows the trends of CPI scores and ranks for Korea, Taiwan, and the Philippines from the early 1980s to the present. CPI can vary between 0 (most corrupt) and 10 (least corrupt). The table indicates that Taiwan’s CPI scores have been the highest among the three countries throughout the entire period, those of the Philippines the lowest, and those of Korea in between. Out of 54 countries for which the historical CPI scores are available, Taiwan has consistently ranked among the top performers within developing countries, the Philippines among the worst, and Korea in between, albeit with a trend of gradual improvement.
[Table 1 about here]
Figure 2 shows the trends in Control of Corruption scores, for Korea, Taiwan, and the Philippines. CC is a standardized score with the mean of zero and the standard deviation of one, and a higher CC value represents a lower level of corruption. The trend of corruption control (CCI) for the Philippines is a deteriorating one, starting from about 0.2 standard deviation below the world mean in 1996 and moving downward to about 0.8 standard deviation below the mean in the recent years. Taiwan’s CCI shows a slight improvement from 1996 to 2004, followed by some deterioration until 2008 and a rebound since then, indicating a trend from about 0.6 standard deviation above the world mean in 1996 to over 0.8 in 2004 to 0.5 in 2008 to 0.9 in 2011. Korea’s CCI score improved from about 0.2 in 1996 to 0.6 in 2005, but it has slightly declined to 0.45 in 2011. The trends of CCI for three countries are roughly similar to those of CPI, with Korea and Taiwan making improvements but with the Philippines showing deterioration in control of corruption.
[Figure 2 about here]
It is often more useful to look at specific indicators of corruption that are comparable across countries and across time than composite indices of corruption such as CPI and CC. The Executive Opinion Survey of the World Economic Forum has been asking businessmen to choose the biggest obstacle for doing business in each country. In the Philippines, an average of 22.9 per cent of businessmen chose corruption as the biggest problem between 2003 and 2011. In contrast, only 2.4 per cent of Taiwanese businessmen and 5.5 per cent of Korean businessmen did so.
TI’s Global Corruption Barometer Survey has been asking the general public if the respondents or their family members have paid a bribe to public officials during the past year. This survey question is likely to represent the relative frequency of petty bribery. In the Philippines, 17.5 per cent of the respondents on average (between 9 and 32 per cent each year, from 2004 to 2010) admitted paying a bribe to officials, while in Taiwan and Korea the averages were 3.3 per cent and 2.9 per cent, respectively.
Both surveys suggest that the Philippines is among the most corrupt countries in Asia. In terms of business executives’ perceptions on corruption, Taiwan and Korea are slightly behind Singapore and Japan but substantially better than most Asian countries. In terms of petty bribery, Taiwan and Korea are among the best along with Japan, Singapore, and Hong Kong.
All the evidence presented above consistently indicates that Korea has been slightly more corrupt than Taiwan but much less corrupt than the Philippines at least since around 1980. Moreover, these differences antedate the transition to democratic rule. During the period of 1980-85, all the three countries were under authoritarian regimes. CPI 1980-85 finds that the authoritarian regime of Marcos (1972-1986) was perceived to be extremely corrupt, while that of Chiang Ching-kuo (1975-1987) was relatively clean and that of Chun Doo-hwan (1980-87) in between the other two.
Unfortunately, there are no available quantitative measures of corruption comparable across these countries for an earlier period. However, there is evidence that there was rampant corruption in all the three countries in the early post-independence period. Corruption and looting were so common under the Chinese military government after the liberation from Japanese rule in 1945 that Taiwanese revolted in a popular uprising, known as the “February 28 Incident,” in 1947. It is well known that corruption was widespread under Korea’s first president Syngman Rhee (1948-1960), who had to resign as president after large student demonstrations against the election rigging and corruption of the regime, known as the “April 19 Student Revolution,” in 1960.
This suggests that Korea and Taiwan made substantial improvement in control of corruption over time, while the Philippines failed to make improvement or even deteriorated throughout the post-independence period. It is clear that both Korea and Taiwan have been significantly less corrupt than the Philippines at least since the early 1980s, and probably from much earlier, although they were all similarly corrupt at the time of independence. Also, Korea and Taiwan have made further improvement over the last decades after the democratic transition, but the Philippines has shown the opposite trend. The main purpose of this paper is to ask what caused this difference.
4. POSSIBLE EXPLANATIONS
Why have Korea and Taiwan been much less corrupt than the Philippines? Although inequality of income and wealth is closely correlated with the relative levels of corruption in these countries, it is worthwhile to consider the explanatory force of several other factors that have been advanced in the literature, including economic development, democracy, government intervention, Protestantism, Confucianism, and ethnic homogeneity. None appears to provide a convincing explanation of the three cases in question.
The level of economic development (per capita income) has been found by many empirical studies to have significant effects on corruption; in short, more developed countries are generally less corrupt (Lambsdorff 2005; Treisman 2007). Figure 3 indicates, however, the Philippines was initially more developed than either Korea or Taiwan; it was only surpassed by them in the late 1960s. Considering that the Philippines’ higher corruption appears to have started much earlier than 1980, it is more likely that different levels of corruption explain the variations in economic growth in these countries rather than the other way around (Kaufmann and Kraay 2002).
[Figure 3 about here]
Although democracy provides checks against corruption, democratic elections are not invariably a cure for corruption (Rose-Ackerman 1999). Montinola and Jackman (2002) demonstrate that partial democratization may increase corruption, but that once past a threshold, democracy inhibits corruption. Treisman (2007) also finds long-established democracies are significantly less corrupt. On the other hand, some scholars and politicians, notably Lee Kwan Yew, former Prime Minister of Singapore, have argued that authoritarian regimes are superior to democracies in controlling corruption and achieving economic growth (Zakaria 1994).
Both Korea and the Philippines initially had some degree of formal democracy until 1972 (except for 1961-63 in Korea), and had dramatic democratic transitions in 1986 (the Philippines) and 1987 (Korea) through “people power” movements. Taiwan had an authoritarian regime for a long time and experienced a gradual democratic transition during the late 1980s. Among the three countries, Taiwan used to be the least democratic until the early 1990s, but also the least corrupt. This seems to support the argument of Lee Kwan Yew. However, corruption has decreased since the democratic transition, especially since the mid-1990s, in Korea and Taiwan. Also, Marcos’s Martial Law regime (1972-86) is known as a period of the most egregious corruption in the history of the Philippines. Thus, there is no simple relationship between democracy and corruption among these three countries.
The degree of government intervention in the economy, which is often measured as the size of government, or the proportion of government expenditure over GDP, is also regarded as a cause of corruption, because it can create rents and encourage rent-seeking activities. The share of government in real GDP was lower in the Philippines, with an average of 12.7 per cent between 1953 and 1970, than in the other two countries, with an average of 20.9 per cent in Korea and 25.2 per cent in Taiwan during the same period, according to data from Heston, Summers, and Aten (2009). However, the level of corruption was not lower, but higher in the Philippines.1
Ethno-linguistic fractionalization has been found to be positively correlated with corruption, although its significance disappears after per capita income and latitude controls are added (Mauro 1995; La Porta et al. 1999). Ethno-linguistic diversity may partly explain high level of corruption in the Philippines, considering that Chinese-Filipino businessmen tended to bribe government officials and provide politicians with illicit campaign contributions to avoid discriminatory treatment in doing business (Wurfel 1988: 57-58). However, Korea has an extremely high level of homogeneity both ethnically and linguistically, but corruption used to be somewhat higher in Korea than in ethnically and linguistically heterogeneous Taiwan. Thus the ethno-linguistic story does not seem to work well either.
Protestantism has been found by many cross-country studies to be associated significantly with less corruption (La Porta et al. 1999; Paldam 2001). On the other hand, “Confucian familism” has often been accused of fostering patrimony, nepotism, social distrust, and bribes or gift exchanges (Fukuyama 1995; Kim 1999). But Korea has had a much larger Protestant population than Taiwan, and Korea used to be slightly more corrupt than Taiwan. Confucian tradition is very strong in Korea and Taiwan, while it is absent in the Philippines, but Korea and Taiwan are much less corrupt than the Philippines. There is no obvious link between religion and corruption in the three cases.
Income inequality is the one of the several variables reviewed here that is correctly correlated with the relative levels of corruption among these three countries. Figure 4 presents the trends of income inequality (Gini index) in the three countries. It is clear that income inequality in the Philippines has been much higher than that in Korea and Taiwan at least since around 1960. Between Korea and Taiwan, Korea seems to have had slightly higher level of inequality, especially during the 1970s and 1980s.