An answer to the second question above involves answering the first question. In short the use of pyramid structures is a way for government to attempt to commit to a policy of non-intervention without going the whole way to privatisation. A pyramid is a (partial) way of separating firms from political interference.
The development of organisational pyramids give governments more credibility in committing to nonintervention than simply a policy prescription that calls for increased delegation of decision rights to managers of SOEs. This is because the complex organisational structure of pyramids increases the government's cost of obtaining sufficiently timely information to interfere in the day-to-day operations of the firm, and thus prevents an intervening government from imitating a pyramidal strategy. The added complexity of the pyramid structure reduces the "bosses" access to the information needed to intervene in the day-to-day operations of a firm at the bottom of the pyramid and thus protects the managers of that firm from the interference of government officials leaving the managers more exposed to market forces.
Such a organisational structure will not be as effective as actual privatisation - the discipline of bankruptcy, for example, will be weaker under any form of government ownership than it would be under private ownership - but is likely to be more efficient than direct control by government ministers or officials. In situation where privatisation is difficult or impossible for political reasons a pyramid structure may be a second-best solution.
To test this idea that pyramids act to insulate managers from political interference, Fan, Wong and Zhang (2013) focus on the case of China. The Chinese setting, they argue, has several important advantages for this purpose. First, in China, state assets and their controlling ownership are not freely transferable across firm boundaries. As a result, state owners almost always possess 100% of the equity ownership of a pyramid’s firms, which precludes equity financing from serving as the primary reason for a pyramidal structure. Further, because state owners are unable to use outright sales as a means to transfer decision rights in a firm to a third party, as is typical in a market economy, a transfer of decision rights that increases autonomy can be achieved only via a mechanism—such as a pyramid—which is short of an actual transfer of ownership. Second, China’s market economy is young. This allows us to investigate corporate pyramids near their inception, during a time when decentralization has been a chief part of the country's market reforms. Third, China’s markets and geographic regions provide sufficient variation in institutional settings to facilitate measurement of controlling owners' (i.e., local governments’) incentives to decentralise decision-making power to firm managers through pyramidal organisational structures.
Their empirical evidence is consistent with the hypothesis outlined above. In particular, using a comprehensive sample of 742 initial public offering (IPO) firms, majority owned by local governments in China, they find that local governments form more extensive pyramids when they have incentives to lower firm political costs arising from interference. The government's weaker incentives to intervene induce it to credibly transfer decision rights to management through pyramids. While the reduction of SOE political costs leads to more decentralisation of power through pyramid formation, adding more layers could increase agency costs, as the higher information costs make monitoring more difficult. Their empirical results show that more extensive pyramidal structures are associated with stronger legal or market discipline on firm managers, indicating that stronger institutions reduce agency problems arising from empowering management. Further supporting the conjecture that pyramids insulate firms from government intervention, they find a significant positive association between the number of pyramidal layers and the extent of firm managerial professionalism, employment efficiency, total factor productivity, and profitability. In additional analyses, they survey government officials in charge of state asset management and find evidence that managers of the firms who are part of pyramid structures have more decision rights than managers of the firms who are directly linked to the governments.
Thus pyramids achieve some of the depoliticisation aims of privatisation in cases where privatisation itself is, at least in the short term, impossible.
- Fan, Joseph P. H., T. J. Wong, and Tianyu Zhang (2013). 'Institutions and Organizational Structure: The Case of State-Owned Corporate Pyramids', Journal of Law Economics and Organisation 29(6): 1217-1252.