The publication of the Bank of Italy (BdI) balance sheet for March provides the first glimpse of how the Italian economy is navigating the COVID-19 Crisis through the lens of the official sector. However, there are several moving parts—APP and PEPP purchases, VLTRO access, and USD swap facilities—which makes interpreting the balance sheet trickier than usual. Still, March 2020 will go down as one of the most remarkable in the history of the euroarea as witnessed through BdI balance sheet changes.
Let’s walk through the interesting balance sheet components.
Italy’s TARGET2 debit increased EUR107bn to approach a record high of EUR492bn. However, part of the increase is due to increased euro claims on the ECB as part of the FX swap accessed by Italian banks (via the BdI.) Indeed, Italian banks accessed about EUR12bn of USD swaps in March—about 9% of euroarea-wide EUR137bn. The actual “underlying” change in TARGET2 debit was therefore only EUR95bn, to reach EUR480bn. This is the highest (adjusted) TARGET2 debit since May 2019 (EUR485bn). As a result, most of the improvement in Italy’s TARGET2 debit over the past 12 months—due to improved sentiment, non-resident inflows, and the ECB’s tiering system—was undone last month.
Even adjusting for swap access, the change in TARGET2 debit in Italy in March was the largest on record. Indeed, the previous largest increase was in March 2012, when Italian banks, equipped with the proceeds of Draghi’s VLTRO in the months leading to “whatever it takes” accumulated BTPs as non-residents sold—temporarily holding down Italian yields. At that time, the TARGET2 debit increased EUR75bn—that is, EUR20bn less than last month.
March was indeed a historic month.
How was this non-swap related TARGET2 debit increase facilitated? Reporting in BdI monthly statement allows us to identify four main sources and a residual, as summarized in the table below which compares last month to March 2012. There are two main weaknesses with the monthly reporting, since it doesn’t: (i) decompose into government and “other” non-MFI deposits; and (ii) report assets held for monetary policy purposes, instead showing domestic assets held including valuation changes. Still, the table below gives a sense of the drivers of Italy’s “balance of payments” deficit, in the old school sense of the change in the official balance.
The largest component part of the change in the swap-adjusted TARGET2 debit (EUR44½bn or 47%) was access to LTRO funding by Italian banks. This cheap funding (likely -50bps) through bridge LTROs was announced on March 12. It is not yet clear whether, as in 2012, domestic banks used this to fund a carry trade and buy BTPs. But Italian banks require only EUR112bn to cover their required plus tiered reserves (EUR16*7) and total (current account and deposit facility) deposits at end-March were exactly this amount. Banks unwound the EUR13bn in surplus liquidity above this amount during the month of March (14% of the TARGET2 change). Summing the deposit change and VLTRO access, therefore, banks financed EUR57bn, or 61% of the TARGET2 adjustment.
Against this, currency in circulation increased EUR5.6bn. Unfortunately, this accounting is also complicated to read. This is Italy’s (weighted) share of the total increase in currency in circulation across the euroarea, meaning the aggregate system-wide increase for the month is likely about EUR35bn. This will be one of the largest monthly increases on record, though less than the EUR48bn increase in October 2008. Since we cannot, at this stage, back out the exact number for Italy, this is only an approximation. But it implies that at least part of the reduction in banks’ deposits with BdI was needed to facilitate the withdrawal of cash-in-hand.
In addition, EUR25½bn (27%) of the swap-adjusted TARGET2 debit increase can be explained by a drawdown in deposits by government and other sectors. While it is not clear what role “other” plays here—as in non-monetary financial intermediaries—at the end of February these deposits were split between government (EUR45.4bn) and other (EUR25.6bn) which suggests the former took much of the strain. If so, an increase in net issuance in April will be needed to re-build government deposits.
Allowing for these means of facilitating the growing TARGET2 debit, we are left with a residual. For last month the residual—which must include BdI purchases of domestic assets—was about EUR17½bn (19%).
Why can’t we simply read off the increase in domestic claims? Because domestic assets held include valuation adjustments and increased only EUR1bn during the month. But we do know BdI marked down it’s revaluation accounts by EUR14.6bn in March, which would reflect a selloff in Italian assets during the month. In addition, we also know from ECB reporting that BTP purchases of EUR11.9bn happened through PSPP during March. If the revaluation account captures BTPs alone, then total flow purchases for the month would be EUR15.6bn (EUR14.6bn revaluation plus EUR1bn increase). With the PSPP flow of EUR11.9bn, this implies EUR3.7bn in Pandemic Emergency Purchase Program (PEPP) purchases. Since euroarea-wide PEPP flows were EUR15.4bn, this would imply Italy enjoyed 24% of the PEPP purchases into month end, lower than the PSPP purchase share for March of 35%. But both are larger than adjusted capital key of about 16%, however.
And if this EUR15.6bn in asset purchases is indeed accurate, then our residual in the above table is itself over-stated by EUR2bn—which could reflect an overstatement of the actual change in currency in circulation compared to Italy’s share in the aggregate.
Only with time will we be able to complete this decomposition. END.
The General Theorist 
One thought on “Banca d’Italia balance sheet: March 2020”