One of the world’s ironies is that even as John Snow putters on about how currency values should be set by market forces, the market -- or at least the private market -- is now playing a comparatively small role in defining the dollar’s value (and in providing the capital flows the US needs). True, the dollar’s value v. the Euro is set in the private market, to the current chagrin of Germany’s finance minister. But the dollar’s value v. the Chinese renminbi (yuan) surely is not set in private markets -- and the US is likely to import as much from China as the Eurozone this year. The dollar’s value v. the ruble is not really set entirely in the market, given the Bank of Russia’s reserve accumulation. Its value v. the yen? Certainly not set by the private market this spring, when the Japanese government spent $140 billion defending the dollar, and expectations of MOF (Ministry of Finance) intervention still cast a shadow over the market.
The US has not intervened to support a strong dollar during the Bush Administration (intervention requires selling the US government’s stock of yens or euros to buy dollars), but other countries sure as hell have intervened to keep their currencies weak against the dollar (buying dollars with their yen, renminbi or rubles). John Berry is right: the US may not have a strong dollar policy, at least not in the sense of using policy tools to keep the dollar strong. But lots of other countries clearly do have a strong dollar policy: they are spending real money to keep their currencies weak. Intervention is intervention: the dollar’s value is being managed, just not by the United States.
The interesting question is whether some countries currently intervening to keep their currencies weak (and the dollar strong) are tiring of that game -- as it implies both the risk of importing inflation and growing holdings of risky dollar assets. Russia, for example, seems to be interested in holding more euros and fewer dollars. Korea intervened last week, but it may be having second thoughts. The other interesting question is whether some countries (or economic regions) not currently intervening to keep their currencies weak are tiring of watching their currencies strengthen against the dollar, and their economies slow -- afterall, a stronger euro means weaker European exports and a weaker Eurozone economy. Eurozone finance ministers --if not the ECB -- seem to be at least contemplating getting into the currency management game ...
Intervention clearly played a big role supporting the dollar in 2003 and 2004. I increasingly suspect it will play a big role going forward as well. It is hard to go from borrowing $400 billion plus from the world’s central banks annually to zero. The key question may be whether central bank intervention is done to support the current overvalued dollar (as China is doing now), or whether it is done to support an "orderly" depreciation of the dollar -- effectively providing the US with the financing it needs (at the rates the US needs to avoid a sharp slowdown) as it starts to adjust.
More details follow, but a warning first: the following dicussion is rather long, even by Nouriel Roubini/ Brad Setser blog standards.Roubini and some of his non-Setser coauthors developed an interesting model to explain the impact of large player -- which they took to be the IMF -- on financial markets in emerging economies. Their model is based on two observations. First, the IMF does not provide all the financing an emerging market needs in a crisis. The IMF can be overwhelmed if all external creditors and investors -- or if all domestic creditors and investors -- decide to leave. Second, the presence of IMF financing can change the perceptions of other actors in the markets, and thus still have an impact. The US Treasury/ IMF bailout of Mexico worked. Roubini, Corsetti and Guimaraes drew on the global games literature to develop a more sophisticated -- you might even say nuanced -- view of the impact of a large player like the IMF on the market for emerging market debt.
Take the following example. Suppose a country needs to raise $50 billion annually to roll over its domestic and external debt. Suppose its domestic and external private creditors think the private markets will only make $40 billion -- max -- available. That could generate a panic which leads everyone to get out before the financing shortfall leads to big asset price falls/ sharp falls in the currency/ other nasty things. The country might not even by able to get anything close to $40 billion, at least not until after its financial markets have tanked. No one wants to buy high and then sell low. Now suppose the IMF makes $20 billion available. Creditors now think, hey, the country has all the financing it needs, the IMF is putting up $20 and we think the markets will be good for at least $30 billion, so let’s stay in. That oversimplifies a bit -- I am leaving out the role of improved policies in the debtor country, for example. But the basic idea is simple: the presence of a large player willing to put money on the table can influence the behavior of other players in the market.
Does this kind of global games model describe the current market for the dollar -- the biggest market of them all? It is at least interesting to explore the idea.
In 2003, the US obtained $440 billion of central bank financing(according to the New York Fed’s take on the BIS reserve data). That number would have been $480 but for the use of $40 billion of Chinese reserves to recapitalize two state banks, but since the state banks did not to my knowledge, sell the dollars they were given by the central bank, let’s add that to the net financing. The US total financing need in 2003 was around $730 billion. $530 for a current account deficit. $200 to fund US external equity investments (FDI investment by US firms, investment in foreign stock markets by US citizens -- since US outward investment far exceeded inward investment, there was a large net outflow of equity investment). So the US needed to raise $250 billion ($730 billion-$480 billion) abroad from private investors -- or to put it differently, the combination of private inflows of $250 billion (into debt) and private outflows of $200 billion (into equity) provided the $50 billion in net private financing the US needed to fund its current account deficit, given $480 billion of official reserve accumulation.
It we assume that all of the new reserves accumulated by China and Japan were invested in dollar assets (a big assumption, but it has to be broadly true -- globally well over 80% of all the 03 increase in reserves went into the dollar), China provided @$160 billion of the $480 billion in the world’s dollar reserves, Japan provided @$190 billion and everyone else provided @$130 billion. So you could say China and Japan together are sort of like the IMF of this market, lending some but not all of the money the US needs to cover its ongoing financing need. Or to put it in the neutral language of the NY Fed’s Higgins and Klitgaard: "Central bank reserve purchases were large enough to be a key factor in defining the allocation of global capital flows."
The global games model suggests that some of that $250 billion in private inflows might not have been available if private investors had not thought central banks would provide $480 billion. Why lend to the US in dollars if the US cannot raise all the money it needs to support the dollar’s current value and the dollar is about to fall?
Does central bank reserve financing still influence the allocation of global capital flows? My bet is yes. Total central bank financing this year is gonna be big -- probably in the 2003 range. Non Japan Asia was adding about $20 billion a month to its reserves in the first half of the year, and added $30 billion in September. Forecast that out and you get at least $240-50 billion for the year. Japan bought $140 billion to defend the yen earlier this year. That gets you close to $400 billion without looking at the oil exporters/ the Latins. Note this an estimate of total reserve buildup, not dollar reserve buildup, but another $450 in financing from the increase in the world’s dollar denominated central bank reserves is not implausible.
Of course, the 04 US financing need is gonna be even bigger than it was in 2003. The current account deficit is going to be about $650 billion (maybe bigger), and based on data from the first half, net equity outflows of $200 billion are likely --so the total US financing need is around $850 billion. A huge number, but, alas, also a real number. A $450 billion dollar reserve buildup would not be enough to finance the $650 billion current account deficit, let alone finance both the current account deficit and net equity outflows. The US also needs a fair amount of debt financing from non-resident private investors to support the current equilibrium.
That is particularly true now that the BoJ is out of the market -- the BOJ alone provided more than a 1/4 of the total financing the US needed in 03 and was supplying even more of the total financing in the first bit of 2004.
How would hedge funds betting against the dollar impact this kind of game/ this market?
Suppose hedge funds are starting to borrow dollars from US banks and then sell the dollars for euros or yens or won, to the tune of $100 billion. That is a new capital outflow. So the US capital outflow would rise from $200 billion to $300 billion. The total 04 net US financing need (adding in hedge fund and equity outflows) would rise from $850 billion in 04 to $950 billion.
Now, if those currently investing in the dollar knew that the BoJ would respond to hedge fund induced outflows by buying an additional $100 billion in treasuries, they might be happy to continue buying the dollar themselves (and to make everything more complicated, if the hedge funds thought the BOJ would do the same thing, they might not want to make the bet against the dollar). But if they were not sure, then they might worry that the US would not have all the financing it needs to fund both the speculative outflow from hedge funds, the FDI outflow and the current account deficit. In those circumstances, existing private investors in the US might reevaluate their willingness to supply the US with the $400 billion or so in net debt financing they are likely providing the US right now. Indeed, the Roubini et al model of the IMF’s market impact is based on an earlier Roubini et al model of how hedge funds influence the dyanamics of a small market.
Of course, if existing private creditors stop providing financing to the US, the dollar would fall. It would fall until the weak dollar reduced the United States’ need for external financing to the reduced amount of financing that is available, or just until higher dollar interest rates/ lower US asset values made dollar assets attractive to private investors again.
This sense is part of what lies behind the quote in the Roubini/ Setser US external imbalances paper that there is a risk that foreign buying could turn to selling (or at least no new buying) quite quickly -- the risk of net buying becoming net selling is perhaps not an enormous risk, but it is a risk that would be silly to ignore. Doubts about the willingness of the central banks to supply the kind of financing the US needs could lead to doubts among the private investors currently lending to the US ...
To continue to get large inflows of private funds into US debt markets, the US may need for the world’s central banks to continue to provide large net inflows -- actually, in my view, to continue to get net private flows, the US will need both continued support from central banks and to show that it is serious about reducing its overall borrowing need over time (read controlling the deficit -- Treasury Secretary Snow’s comments on the deficit currently have about as much credibility as his commitment to a strong dollar).
If the US could make up for a $100 billion short-fall in reserve financing from central banks in the interest rate on treasuries rises by 50 b.p., there might now be much cause for concern -- small changes in US rates would induce large changes in private capital flows. But if much higher rates are needed to attract private flows if central bank financing falls off -- or if central bank financing simply fails to increase as fast at the US need for financing is increasing -- there is cause for real concern.