In addition, bribery and corruption themselves also operate on a continuum. In the example cited in the article, the rich landowners don't just remove the property markers—they wait for flooding to wash them away before illegally expanding their claims. In other words, even in a society where official corruption is rampant, there are still societal limits within which corruption is possible.
The impact that blockchains can have is to simplify the tasks that systems of adjudication must undertake in order to enforce property rights and contracts. They also remove individual judgement from most decisions, and make it impossible to alter or destroy records.
I'm not saying that blockchains can entirely replace all human governance systems. Rather, blockchains can make it cheaper and easier to maintain functioning and accessible systems for adjudication. They reduce the surface area for the use of individual judgement, and thereby reduce costs and opportunities for the injection of corrupt influence.
Yes, many of the things that blockchains can do can also be done by centralized database applications—maintaining property records, for example. Those centralized databases still need to be maintained by trusted parties, however—trusted parties who themselves may be subject to corrupt influence.
By moving title records from paper to a centralized database all you are doing is changing who becomes the bribery target—the Postgres DBA instead of the clerk at the records office. With a public blockchain, there is no one you can bribe.