Heed the Weakest Link : A Model of Interdependent Institutions
all depend on each other . No one wants a butcher with no idea of hygiene , a careless brewer or a thieving baker .
2 . Economic growth is very complex and thus are very hard to influence . Improvements in one particular institution may not have any detectable impact in aggregate economic statistics ( as in Solow ’ s famous paradox that “ We see computers everywhere except in the productivity statistics ”). In this paper , we give one explanation for why macroeconomic growth regressions typically cannot account for the majority of variance ( i . e . the associated R 2 statistics are not particularly high ). On a more practical level , we also give a justification for using log transformations for variables in regressions . ( This is because a log transformation converts a multiplicative specification into an additive specification , and multivariate regressions are based on an additive specification framework .)
3 . Due to the interdependent nature of institutions , policies , such as those associated with the “ Washington Consensus ” that focus disproportionately on the development of a narrow range of institutions without paying due attention to the other institutions , will probably not succeed in the long run . For example , although a little bit of financial deregulation may be a good thing , nevertheless prolonged financial deregulation may be ineffective or harmful ( Cassidy , 2008 ). Investment in large transport infrastructure projects could be initially effective , but there is a risk that governments may devote too many resources to continued investment in infrastructure projects of decreasing utility ( Ansar , Flyvbjerg , Budzier , & Lunn , 2016 ; Fernandez-Villaverde , Garicano , & Santos , 2013 ). Austerity and fiscal tightening may have its advantages , but pursuing austerity excessively may be ineffective ( Corsetti , 2012 ). Instead , the representation here suggests that governments would do better to aim for a coherent set of institutions , where each institution is at a comparable level of development , and then to identify their weakest institutions and try to improve these . Indeed , instead of standard off-the-shelf one-size-fitsall recommendations , such as financial deregulation or austerity , appropriate policy recommendations will probably differ by country and by time period . Relatedly , a nation should not become obsessed with improving the same institutions over time . The more time a nation spends improving a specific institution , the lower will be the marginal benefits of further improvements in this direction . The same policy should not be pushed too far . Policies that worked well in the past may be ineffective in the future . It might , therefore , be beneficial to periodically reevaluate the direction of the thrust of economic policy , in the context of 5-year or 10-year plans , for example .
4 . Our model implies that the world technological frontier should be defined in terms of its ability to produce complex technologies ( e . g . high-speed trains , high-quality public transport , hybrid cars , modern housing ), and not in terms of narrowly-focused technolo-
59