China’s state banks have long got a lot of attention. Bank—and shadow bank—lending drives China’s current economy, and the state banks, counting the banks owned by provincial and local authorities as well as the big five state commercial banks, account for almost all of China’s banking system.
Turkey’s state banks are now also attracting a lot of attention. Rightly so. They are only about a third (maybe a bit less) of Turkey’s banking system. But they expanded their lending rapidly in the first quarter.
That helped (temporarily?) stabilize Turkey’s growth. It was a backdoor stimulus ahead of Turkey’s recent, not-yet-over, municipal elections (and there was likely a bit of front door stimulus as well).
Their lending also meant that Turkey’s current account balance stopped improving, and that, together with the government's desire to keep the lira up ahead of the election, in turn helped create a bit of pressure on Turkey’s reserves. The central bank's end March disclosure essentially confirmed the basic point made by the Financial Times in their blockbuster story.
While the impact of this increase in lending on Turkey’s economy has gotten a reasonable amount of attention, its impact on the balance sheet of the banks themselves hasn’t gotten as much attention.
I suspect of course that many people assume that the loss rate on the latest round of lending will be significant. But if the losses are in the public banks, they can always be recapitalized—Turkey isn’t a part of the EU, so it isn’t constrained by the EU’s state aid rules or its Bank Resolution and Recovery Directive.
But there is likely a more immediate source of strain—the state banks limited access to deposit financing. In lira.
A bit of background. Turkey’s banking system has a high loan to deposit ratio in Turkish lira (See chart IV.2.4 in the Turkish central bank's Financial Stability Report, as well as my blog post from last August). That’s in part because the banks hold their equity in lira, even though much of their balance sheet is in foreign currency. But it also reflects the central bank’s willingness to let the banks meet a part of their reserve requirement in lira by holding foreign currency at the central bank (the so called reserve option mechanism) and the ability of the banks to in effect borrow lira internationally if they post dollar or euro collateral through the swaps market ("Turkish banks ... have a net swap position of around 217 billion liras ($39 billion) as of end February"). Historically, at least, the private banks and foreign banks have been more aggressive than the state banks in lending out lira—in part because they have had more access to (I suspect) the swaps market.
But, well, the private banks have pulled back a bit on their lira lending after Turkey came under market pressure last August.
While the state banks juiced their lira lending at the start of the year. And that increase came without any growth in the state banks' lira deposits.
As a result, the state banks now have a higher lira loan to deposit ratio than the private banks. That's a shift: historically the state banks have had a more conservative funding ratio. At 143 percent of their lira loan book (end March data), their lira loan to deposit ratio looks absolutely high too.
Even with a clear need for lira, Turkey’s banks—for mysterious reasons, though many suspect a bit of regulatory pressure—haven’t been competing aggressively for lira deposits. Lira deposit rates fell to 20.5 percent (just over inflation, and well under the central bank’s policy rate) earlier this year.
And Turkey’s depositors have responded by shifting out of lira (relatively speaking) and into foreign currency.
I don’t think closing this gap entirely through swap financing from the market is an option. Turkey’s central bank has made it hard for foreigners to borrow lira in the swap market (to deter speculation) and that in turn has made foreigners more reluctant to lend lira to Turkish banks through the swaps market.
And no doubt the risk of U.S. financial sanctions as Turkey's relationship with the United States deteriorates hasn’t helped either. Cutting the state banks off from foreign financing is something that the United States knows how to do (see Russia, state banks—like VTB).
That all leaves the state banks in need of lira.
Of course, the central bank can always supply lira against appropriate collateral.
And Turkey’s state can always provide the state banks with the needed collateral. Among other things, the state can always give the state banks lira bonds—that’s how you do a recapitalization—and the banks can then borrow against the collateral provided by the recapitalization bonds.
But providing a lot of lira to the state banks so that they can juice their lending (in part by lending at below market rates) will fuel speculation against the lira.
Most analysis of the central bank’s use of swaps to increase its reserves has focused on the desire of the Turkish central bank to report higher (net) reserves and thus more “free” cash than it really has. But the swaps also help the banks use their spare foreign currency deposits to fund their lira lending. The central bank's explanatory box on this in their recent inflation report was actually quite useful; in the swap the banks provide the central bank with foreign currency to boost the central banks reserves, and receive lira—and for now the banks have plenty of foreign currency thanks to the shift in resident deposits even as cross border foreign currency funding has fallen.*
Makes for an interesting situation.
Turkey’s depositors know how to shift their domestic deposits toward foreign currency. With new electoral uncertainty, it isn't clear that the recent $20 billion shift toward foreign currency deposits has run its course. Turkey’s firms (particularly the exporters) likely can build up cash offshore if they really want to, though they now have to skirt a requirement that they repatriate their export earnings within 6 months…
While the central bank can always provide lira to the banks, if it does so too freely, it will only add to pressure on the lira.
Call it a twin (balance sheet) crisis if you want, one of both lira funding and corporate foreign currency exposure.
* The banks have more foreign currency debt coming due than they have liquid foreign currency assets. But so long as a relatively high fraction of those maturing debts are rolled over, this theoretical vulnerability isn't an acute vulnerability, and the banks can use their foreign exchange liquidity buffer to borrow lira short-term from the central bank.
Remember that the banks have a lot foreign currency denominated liabilities (both domestic foreign currency deposits and cross border loans and bonds) and a mix of very liquid (deposits abroad, reserves at the central bank) and very illiquid (loans to companies that cannot repay quickly, if at all) foreign currency assets. They have a very complicated aggregate balance sheet.