Does Financial Deregulation Boost Top Incomes? Evidence From The Big Bang
Brown University
Research Institute of Industrial Economics
Abstract:
This study estimates the impact of financial deregulation on top income shares. Using the novel econometric method of constructing synthetic control groups, we show that the “Big Bang”-deregulations in the United Kingdom in 1986 and Japan 1997-1999 increased the share of pre-tax incomes going to top earners by over 20 percent in the U.K. and over 10 percent in Japan. The effect is strongest in the top five percentiles in the U.K. whereas it is mainly driven by the lower part of the top decile in Japan. The findings are robust to placebo tests, alternative ways to construct synthetic controls and scrutiny of post-treatment trends. Higher earnings among financial sector employees appear to be an important mechanism behind this result.
Does Financial Deregulation Boost Top Incomes? Evidence From The Big Bang – Introduction
In the light of the most recent financial crisis, a discussion has re-emerged about who gains and who loses from governmental policy. One specific line of conflict is the very top one percent of income earners versus the other 99 percent. Even when leaving the turbulent setting of a crisis behind, the question of winners and losers from public policy still stands, especially for policies related to the financial sector. The focus of this study is on the distributional consequences of the vast deregulations of financial and stock markets that helped creating global financial centers in London and Tokyo during the 1980s and 1990s.
Even though the relationship between financial market development and top incomes is of great concern, this question has not received much attention in the previous research literature. Studies of finance and inequality have mainly focused on the low end of the distribution (e.g., Demirguc-Kunt and Levine, 2009; Beck, Demirguc-Kunt and Levine, 2007; Clarke et al., 2006) or on specific groups of high-income earners in the U.S. (Kaplan and Rauh, 2010; Philippon and Reshef, 2012; Nabar and Jerzmanowski, 2013). Roine, Vlachos and Waldenstrom (2009) study top income shares and indicators of financial development but focus on broad correlation patterns over very long-run historical periods.
In this study, we intend to estimate how top income shares are in uenced by financial deregulation. We do this by focusing on two well-known large-scale financial market reforms, the “Big Bangs” in the U.K. in 1986 (Clemons and Weber, 1990) and Japan in 1997{1999 (Toya and Amyx, 2006).1 These Big Bangs incorporated a wide range of reforms that were especially concerned with stock market liberalizations and they were carried out over a quite limited time period. Deregulations did, of course, occur in other countries as well during the 1980s and 1990s, but they were both more gradual and had less of the concentrated and profound impact that is attributed to the Big Bangs.
Several potential channels could explain why financial deregulation drives top income shares. One centers on higher profitability in the financial sector caused by the deregulation, and how the gains are distributed to employees and management of the financial firms in the form of higher remuneration.2 Another possible channel works through higher wealth returns as result of more efficient markets and cheaper investment services after the deregulation, which primarily benefits top income earners with relatively large share of income from capital.
Identifying causal impacts on top income shares has been a key challenge in previous studies of top incomes. Our strategy to overcome this problem is to use the new synthetic control methodology proposed recently by Abadie, Diamond and Heinmuller (2010).3 The idea behind this method is to construct a control group that reflects what would have happened to the income distribution in the treated country, i.e., the U.K. or Japan, in the absence of the financial deregulation. Rather than choosing one or more countries to use as a comparison group (which would be a standard difference-indifference approach), a synthetic control group is created by using a weighted average of countries that are selected based on their similarities with the treated country in terms of the historical evolution of top income shares and of important background characteristics. Compared with other types of comparative case studies, this results in a data-driven approach which provides a better match to the counterfactual trend and also reduces the risk that the results are driven by unobserved shifts in the control group.
The critical assumption of the synthetic control method is that a match in pre-treatment trends between a country and a synthetic control group indicates that they would have continued to follow the same trend, in absence of treatment. This non-parametric choice of a counterfactual has benefits when studying top income shares that may be affected by global or regional trends depending on the institutions of the nation studied. However, it is possible that there are no common trends in top income shares between similar countries and, if so, that the assumption is unlikely to hold. We assess this eventuality by varying the controls used to construct the synthetic control group, and we also use placebo tests which reveal if the estimated effects of treated countries are among the most extreme. In addition, we examine the assumption about post-treatment trends in control variables being parallel between treated and control countries.
See full PDF below.