Global Economic Prospects 2014

Five years after the financial crisis, the world economy is showing signs of bouncing back. Read more ...

Development Prospects

Providing information, analysis, and advice on global trends in the world economy.  Find out more ...

Global Economic Monitor

Now free. Daily and monthly updates on global economic developments and relevant topical issues. Find out more ...

Syndicate content

The Optimal Design of Sovereign Wealth Funds

Jamus Lim's picture

Although much of the news flow regarding oil in recent days has focused on the terrible environmental tragedy off the Gulf of Mexico, another oil-related story has occupied the minds of many market analysts: that of rising oil prices, along with the overall increase in the prices of a whole host of commodities, as the global recovery has taken hold. This has meant that commodity producing nations are once again in a position where they would need to think seriously about how to manage revenue inflows that result from commodity booms.

One strategy for managing such revenue is to establish a stabilization fund, or more generally, a sovereign wealth fund. Chile is perhaps one of the best recent examples of how prudent saving through such funds can help consumption smoothing efforts during periods of economic turbulence, such as during the financial crisis of 2008/09, as well as its more recent experience with earthquake-related reconstruction. But lauding the merits of an SWF in cushioning the impact of volatility takes for granted the fact that the establishment and maintenance of such funds is not at issue, which is clearly not true in the real world.

To assist countries in the design of SWFs, the Fund issued a working paper last year laying out the practical issues associated with establishing SWFs. Beyond the many interesting (and largely academic) issues that have been explored about SWFs in recent times---see especially the important paper by Aizenman and Glick on the determinants and effectiveness of such funds, the monograph by Setser on the implications of reliance on SWF financing, along with the more recent (theoretical) take by Carroll and Jeanne that models the precautionary motive for SWF establishment---the operational questions have been only tangentially addressed. The IMF paper thus provides a firm basis for thinking about SWFs from the perspective of the developing country seeking to establishing such a fund.

The paper distinguishes five main classes of SWFs, consistent with their purported objectives: (1) reserve investment corporations; (2) pension-reserve funds; (3) fiscal stabilization funds; (4) fiscal savings funds; and (5) development funds. The authors do an excellent job of summarizing how a given SWF established along one of these lines can be designed to fulfill these objectives. What is missing from the discussion, however, is a discussion of how the institution itself can be designed to better attain these goals. Such a distinction is analogous to going beyond classifying central banks according to their mandates---which may require them to meet a specific inflation target, maintain price stability, or appropriately manage the inflation-unemployment tradeoff---versus proposing a set of rules that would better allow it to achieve these goals (such as via the appointment of a conservative central banker or one whose contract is tied to inflation).

One key design consideration is whether SWFs can be designed to avoid the resource curse so common in developing country institutions. This is, after all, the expressed purpose of having an SWF to begin with. But having an SWF merely begs the question of what exists to prevent an SWF from being raided by politicians seeking to ensure re-election by passing out goodies? Alternatively, even a benevolent dictator may have an incentive to engage in time-inconsistent SWF management policies if such actions could somehow lead to ex post optimal outcomes.

The traditional manner by which developed countries have attempted to resolve time inconsistencies is to ensure institutional independence. There is evidence that this may be the case with regard to inflation (see figure). The negative relationship between inflation performance and central bank independence is fairly clear, and more formal analysis (using an unbalanced panel of these countries over a 10-year period, and controlling for obvious factors such as GDP) verifies the graphical result.

Notes: Data for 1999 for high-income countries, using CPI data from the IMF IFS and CBI data from Polillo & Guillén (2005), which is based on the Cukierman-Webb-Neyapti methodology.

However, for developing countries, this is largely a red herring, since the institutional framework in these countries is often insufficiently developed to support such a first-best policy (think of how successful developing country central banks have been in preventing the monetization of their fiscal debt). The poorer track record of independent developing country central banks in controlling inflation attest to this fact (see figure); here the relationship is actually slightly positive. While proper econometric analysis does exonerate this damning result somewhat, the broader point is clear: it is both unfair and unreasonable to expect that developing countries have the institutional capacity to sustain a truly independent central bank in the face of legislative or executive pressure.

Notes: Data for 1999 for low and middle-income countries, using CPI data from the IMF IFS and CBI data from Polillo & Guillén (2005), which is based on the Cukierman-Webb-Neyapti methodology.

It's not clear that there is a straightforward solution. However, one (admittedly risky) strategy is to provide such institutional independence by privatizing the SWF. There is a fundamental difference between a public institution that has been granted independence, versus a private institution. For one, the privatization option is far more irreversible. Raids on a privatized SWF would essentially entail the seizure of private assets followed by their nationalization; in contrast, thefts from a nominally independent SWF is mainly a matter of accounting changes made to the books of two relevant government entities.

But how would a privatized SWF be responsible to its original mandate, which is to serve as a stabilization fund for the country? Through a contractual arrangement. Just as utilities companies are obliged to continue to provide their services after they have been privatized, a privatized SWF will be obligated to fulfill their contractually-agreed mandates (whatever those may be). But being private entities, they will be generally freer from political interference. The privatized SWF will continue to receive capital injections from the government when the latter experiences windfalls, because the government knows that when the circumstances meet the pre-defined criteria, the SWF will release the funds for the government's use. In effect, the government is tying its own hands, but in a manner possibly stronger than when it does so with a quasi-government institution such as an independent SWF, since there is now an additional check and balance---the judicial system---that will need to be violated in order to override the independent choices of the SWF.

I'm not suggesting that this solution will definitively work for every country trying to set up an SWF. After all, no commitment device will work if the government has the intent, ex ante, to break the commitment. But, if the issue is mainly one of time inconsistency, then perhaps privatization may be a better solution for developing country SWFs trying to overcome the standard institutional problems so common in the developing world.

Comments

Submitted by Ashby Monk on
Nice post, Jamus. I'm intrigued by the notion of privatizing the SWF to facilitate commitment to the mandate over the long-term. In a sense, this is close (but not quite) to what Canada's CPPIB has done. Still, while interesting in theory, I'm skeptical "privatized SWFs" will ever be roaming the world...

Submitted by Jamus Lim on
Thanks, Ashby. Like you, I am skeptical that we will see privatized SWFs any time soon. The thinking behind the post, however, was not so much to say that we will see of proliferation of these things; rather, it is to stimulate thinking about how time-inconsistency issues can be overcome, should a sovereign actually wish to precommit to not raiding an SWF. That said, I am not sure that a privatized SWF is entirely an abuse of terms. After all, couldn't we have said the same about private utilities companies (and other so-called natural monopolies)? Of course, the main advantage in privatization of previously public utilities is to garner efficiency advantages; while there may be efficiency gains from a privatized SWF, my view is that such gains are secondary to resolving the commitment problem. And we are in fact already seeing some SWFs practicing a form of (limited) privatization, when they allocate some of their funds to external managers, such as Blackstone and Goldman.

Submitted by Rien Huizer on
Mr Lim, you appear to support the position that (generic?) SWFs are inherently difficult to design, if not inherently unfeasible (unless one is satisfied with a very weak version), by pointing at implementation difficulties that may require rather unorthodox solutions. Incidentally, maybe SWFs, unlike Central Banks, do not belong to the category of institutions that can be assessed adequately using "econometric methods" . Intuition combined with guile and the ability to impersonate idealists are the characteristics of successful politicians ("SPs") in any regime and in some regimes politicians carry SWF tools. Very few politicians are in a position where they would submit to time consistency regimes, unless those regimes are so entrenched that changing them would be impracticable. Hence it is unlikely that with a relatively new policy instrument, time consistency would be afforded more than rhetorical attention, plus perhaps more serious attention focusing on precisely the prevention of time consistency enhancing instruments emerging. However, I think your idea might be more feasible if the "private" (I guess any entity not under the control of the government would do) owner/manager/guardian of an ideal SWF would be, say, the World Bank or an affiliate (let's call it WBSWFServices. In the nineteenth century, an SWF for say, Australia would have been established by the benevolent Crown and her enlightened servants, and outside the control of local potentates. Philosophically, discourses involving time consistent economic devices tend to struggle with the twin imperatives of regime design that have dominated since WWII: democracy and national self determination. Useful to keep in mind that the first SWFs were designed by the Colonial Office of the UK.

Submitted by Jamus Lim on
Thanks for the comment. The suggestion is premised, of course, on a willingness by politicians to want to credibly commit but face time inconsistency problems (which in turn assumes that they do in fact wish to commit to begin with). While one can certainly be skeptical about the underlying motivations of politicians, wanting to credibly commit to not raiding an SWF may actually be entirely consistent with politician self-interest: For example, a vote-maximizing politician may recognize that stable, consistent economic growth may improve their electoral chances, and to the extent that having a well-funded SWF available for consumption smoothing would certainly aid in that effort; alternatively, even if the political leader is interested only in maximizing their own economic rent, having a non-politically directed SWF would maximize SWF income that can then be channeled toward this purpose. That said, there is a potential problem with entrusting SWFs to the care of an external manager (or "guardian," as you call it) that is not a private entity. In this case, the guardian may itself be subject to nonpecuniary pressures, which may distort their decisionmaking. Such pressures are likely to be greater for entities whose stakeholders and shareholders extend beyond those of the principal (in this case, the sovereign nation) and agent (the privatized SWF). While we can quibble about how much this is likely to matter in practice, it is hard to argue that private fund managers do not have a strong incentive to make money. One additional, minor, clarification: I based the arguments that I made in the post on a game-theoretic analysis, not on econometric work. While it may be true that data limitations have made SWFs less amenable to econometric analysis, this is by no means impossible. In fact, the post refers to the Aizenman and Glick (2008) paper, which is econometrically based; another recent econometric contribution is that of Bortolotti, Fotak, Megginson and Miracky (2009).

Add new comment