Addressing the Global Dollar Shortage: More Swap Lines? A New Fed Repo Facility for Central Banks? More IMF Lending?
The dollar is in some ways a global currency. The BIS estimates that banks outside the United States have over $13 trillion in dollar denominated assets. Just as no part of the global economy will be insulated from the impact of the virus, no major part of the global economy will be insulated if dollar funding markets break down. And there is a growing body of work that suggest a strong dollar hurts most of the global economy, thanks to widespread financial use of the dollar outside of the United States.
This global use of the dollar isn’t a new issue—the need to assure access to hard currency has long led emerging economies to hold significant stashes of dollars as part of their central banks’ reserves, and it was also an issue during the 08-09 global financial crisis. European banks in particular were running large dollar balance sheets at the time—borrowing short-term dollars in the market (or swapping euros for dollars) and then buying long-term U.S. mortgage backed securities. Basically, Europe’s banks were running U.S. shadow banks, raising wholesale funds to buy longer-term U.S. securities even though they lacked a large base of dollar deposits…
When access to wholesale market funding for risky balance sheets dried up, the Fed responded in two ways:
First, it created facilities that allowed all banks (with appropriate collateral) to get access to Fed financing. These facilities auctioned off financing—and European banks were big buyers in the auction. “Branches and agencies of foreign banks” were eligible. European banks also were heavy users of the Fed’s commercial paper financing facilities.
Second, the Fed provided swap lines to the ECB, the Bank of Japan, the Bank of England, the Swiss National Bank, and a few other major central banks. These countries generally didn’t have a lot of dollars in their foreign exchange reserves—at least not in relation to the size of the balance sheets of “their” banks. So to provide a dollar lender of last resort to their institutions, they had to borrow dollars from the Fed. Central banks typically borrow from each other through swaps—the Fed provides say the ECB with dollars, and the ECB provides the fed with euros (which serve as collateral in a sense). The transaction has no exchange rate risk as it unwinds at a fixed price—it is just a mechanism in this case for borrowing dollars. Those receiving dollars in turn could provide a dollar lender of last resort to their financial institutions.
A select number of emerging economies received access to modest swap lines as well.
To be clear, European banks were at the time a big part of the U.S. financial system—they were raising funds in the United States and globally to buy U.S. securities, and the securitization process underpinned a lot of U.S. domestic lending. The United States wasn’t acting altruistically here—it was acting to stabilize its own financial system. The equity capital that backed a portion of U.S. credit happened to be European at the time. The United States didn’t want European owned shadow banks that were big players in the U.S. market to fail just because they couldn’t borrow dollars in the market and couldn’t “roll” their short-term financing (shadow banks like banks borrow short-term to buy longer-term and less liquid assets).
Today’s problem (see Jon Turek of Cheap Convexity for a good overview) is both similar and different.
Similar, in that there are financial institutions around the world not just in the United States that operate in dollars and potentially need access to dollar credit amid the corona virus shock.
And different in that the shock this time originated outside the financial system. A pandemic is a true exogenous shock.
So what to do?
One option is to broaden access to the dollar swap lines—
Another way to provide dollar liquidity to those economies that already have substantial dollar reserves (like say China, or for that matter Taiwan) is to provide a Federal Reserve repo facility for central banks, so that these central banks can obtain cash by borrowing against the collateral provided by their Treasury securities. This would provide a way for these central banks to provide dollars to their banks without having to sell their Treasuries in the market.
The difference here isn’t huge—a swap effectively is a loan against foreign currency collateral (but with no exchange rate risk unless the counterparty defaults). A repo is a loan against Treasury collateral. But if there are any political concerns about providing China a swap, lending against Treasuries is an easy alternative—worst comes to worse, the United States ends up buying back some of its own debt.
Another is to have the IMF provide emergency lending funds—if its backup lines of defense are activated, Mark Sobel (the former U.S. Executive Director) believes the IMF has close to $700 billion in lending capacity (around $275 billion from members' quota contributions, $185 billion from the New Arrangement to Borrow, and another $350 billion in commitments from existing bilateral credit lines). The World Bank can also lend to help national export credit agencies and the like.
And the IMF has the ability to provide all countries with more reserves (though the biggest countries get the most reserve increase) through what’s called an SDR allocation.
The right answer is probably a mix of all four (and no doubt there are other options)—
Extending the swap lines to a broader set of countries—all of the G-10, and probably the four emerging economies that received them last time makes sense (It is in the United States national interest to help Mexico and to help stabilize the peso; Korea is a close ally and its insurers and banks have a hedging need though one that Korea could meet out of its own substantial reserves; Singapore does't really need the funds given its massive reserves and Brazil isn’t as strong as it was in 2008 but this isn’t the time for fine gradations). The swap line though are traditionally available to help countries support their own banks dollar funding needs—not for outright intervention in the FX market.
It isn’t obvious to me though that China needs a swap line, despite the arguments made by Claire Jones at FT Alphaville—the government of China’s foreign currency reserves substantially exceed China’s foreign currency debts. Countries with large reserves should be using those reserves now.
I can though see why it is neither in China’s interest nor the United States' interest for China (and other countries with large reserves) to sell their Treasury portfolio right now. The Fed—for good reason—wants U.S. rates to be low.
As a result, it might be in everyone’s interest if the Fed let countries with lots of reserves—typically held as Treasury and Agency securities—borrow dollar cash against those securities. Central banks could get their own repo facility at the Fed (repo is a form of secured borrowing in effect).
It is just an idea—but this is a time for creativity.
That though isn’t enough for countries that neither have large reserves of their own nor are financially strong enough to quality for a swap line. Many of these countries have heavily domestically dollarized banking systems—realistically, they will need to meet domestic liquidity needs mostly in their own currency. But the IMF should be creative in finding ways to support these countries.
And I have no objections to an SDR allocation—it isn’t a huge sum for most countries (it won’t cover the foreign exchange needs of those with dollarized domestic banking systems) and it doesn't necessary go to the countries with the greatest need, but this is a "whatever it takes" moment, and a time for solidarity.
As background, there are a range of dollar borrowing (and hedging) needs in the global economy outside the United States—this isn’t an exhaustive list, but I hope it provides some clues.
1) Dollar funding for global banks
This was at the center of the 2008 crisis—big banks, especially big European banks, were borrowing dollars on a range of “funding” markets (issuing commercial paper, taking in big deposits, swapping euros for dollars) to buy U.S. asset backed securities. And those same markets are used to support these banks other dollar-based activities—project finance and trade finance in particular. (See Hyun Shin and his colleagues)
The European banks have pulled back a bit since then—but Japanese banks, and I think Canadian banks, have stepped up. See this BIS paper.
And then there is the special case of the two Japanese policy banks—Postbank and Norinchukin. Both have large dollar investment portfolios (Norinchukin mostly holds CLOs).
It should be noted though that a portion of the cross-currency hedging need of the Japanese banks is tied to their trust accounts -- e.g. it isn't actually a balace sheet need of the bank so much as a need of their clients. See p. 70 and figure V-5-8 of the BoJ's Financial System Report.
2) Hedging needs of Asian insurers
This is one of the big “non-bank” sources of demand for dollars—
Think of it this way. A Japanese insurers that didn’t like the yields available on Japanese government bonds could buy higher yielding U.S. securities. But it would then be exposed to exchange rate risk—its portfolio’s value would swing with the yen dollar. The solution was to “hedge” a portion of the portfolio, by swapping yen for dollars. The hedges are generally done on a fairly short-term basis—think 3 months, and the underlying portfolio tends to be longer-term bonds.
That economically looks a lot like borrowing dollars for 3 months in the short-term funding market to buy long-term corporate bonds. But there is one important difference. If the insurers cannot rollover their hedges, they aren’t necessarily required to sell the underlying portfolio—they could operate with an unmatched currency book for a while. And the Japanese and Taiwanese governments could provide their insurers with “insurance” against big swings and a degree of regulatory forbearance.
The numbers here are big—Japanese insurers have something like $900 billion in foreign securities (25 percent of their assets), and not all are hedged. Taiwanese insurers are heading toward $600b in foreign securities if you count the bonds held by "domestic" bond ETFs that raise funds from the lifers to invest abroad (more). But in both cases, the “home” government has substantial capacity to help. Taiwan especially, as its true reserves are about $600 billion, or around 100 percent of GDP.
3) The “China” complex—China is big, and even though its foreign currency debt isn’t enormous relative to the size of its economy, it is a large absolute sum—$1.3 trillion. That total may leave out some funds that Chinese companies have borrowed through their offshore subsidiaries.
Let’s estimate—based on the balance of payments data—that half of that debt has been taken out by the state banks and the like, and thus is implicitly backed by the government. Some portion is from state owned companies. Another fraction has been issued by big private firms (including big real estate companies like Evergrande) that may or may not be backed by the government.
But China has $3 trillion in reserves, including over a trillion of Treasuries (the real total is more like $1.2 trillion because of the Treasuries China has custodied through Belgium) and at least $200 billion of Agencies. It like has more dollars squirreled away in less visible parts of the market—SAFE is widely thought to have been swapping dollars for yen (hence the rise in China’s JGB holdings). So China in aggregate has enough liquid dollar assets to meet the demands for payment on the foreign currency debt of its banks and firms if it wanted to—at least in normal times.
Right now, though, it might be better if the Fed let China borrow against its Treasuries than to sell its Treasuries— enlighted self interest and all, as the Treasury market is having to absorb a lot of forced selling. Hence my idea for an "official" repo facility at the Federal Reserve for foreign central banks.
4) The “oil” or “quasi-sovereign” borrowers
The biggest corporate dollar debtors in the emerging world are companies like Petrobras, Pemex, and now Aramco. Rosneft and Gazprom were too before the sanctions. These type of firms, and their smaller counterparts, account for much though obviously not all of emerging market corporate cross border foreign exchange exposure. In some cases this should be viewed as hidden sovereign borrowing in my view—it isn’t the most immediate concern though, as these companies generally have issued long-term bonds that aren’t coming due for a while.
5) Tax plays
Turkey’s exporting firms supposedly liked getting paid at a subsidiary in Europe and then lending a portion of the proceeds back to their Turkish parent. I was a bit suspicious but cross border lending to Turkish firms has been resilient over the last two years.
6) Domestically dollarized banking systems—
Turkey’s banks have borrowed from abroad to provide foreign exchange that they swap into lira—whether in the market or (now) with the Central Bank of the Republic of Turkey.
They also have taken in a lot of foreign currency denominated dollar deposits.
Lebanon’s banking system is even more dollarized.
As domestic dollar deposits exceed foreign currency reserves, it is hard to see how in a crisis like this these domestic deposits aren’t settled in the local currency. That though risks putting more pressure on the exchange rate.
There is no single solution here—as there is no single source of need for dollars. But there are different tools that could help address different aspects of this problem. More countries should get access to the Fed's swap lines—but there are other sources of dollars available too, including the accumulated foreign exchange reserves of some key countries.
One last note. This is a work-in-progress. I intend to update this blog post as additional information becomes available. I rushed this post out for obvious reasons.