European Central Bank: Three-Year Long-Term Refinancing Operations

Because of the massive size of the program, the three-year LTROs have been quite heavily scrutinized by academics, journalists, and politicians alike. Generally speaking, institutions that were most acutely distressed during the sovereign debt crisis tended to use the facilities the most. The ECB stated that banks that were more risky tended to bid more at the LTROs, and that even if private markets were able to provide some funding to banks, there was still a heavy incentive to borrow from the facility. However, they stated that “by improving funding conditions, the ECB may have prevented the disorderly shedding of assets, which would have placed certain financial market sectors under pressure” as well as emphasizing a belief that the introduction of these programs contained spillover effects to larger markets (ECB Monthly Bulletin, March 2012).

This rationale was very clear because, while there was not significant growth in private credit or bank lending relative to the size of the program, it was effective in halting the freefall and ultimately stabilizing these two channels. Additionally, spillover effects from the financial sector to the real economy, which was one of the ECB’s chief concerns when designing the program, appeared to have been significantly lessened owing to the increase in system-wide liquidity.

One study by the National Bank of Belgium estimated that Euro-area output and inflation were significantly impacted and would have been more than 1% lower without the operations (Boeckx et al. 2017). Demand for sovereign debt was also positively impacted, despite not being a primary objective of the program. In Spain, Italy, and Ireland, sovereign debt yields decreased around the time of the announcement, with the largest decrease being 51 basis points for Spanish two-year bonds (Krishnamurthy et al. 2017).

Despite the seemingly robust impact on sovereign debt markets and macroeconomic aggregates, the results were mixed for bank lending, which was the primary area the ECB was trying to affect. In a paper published in November 2018, Stine Louise Daetz and her coauthors emphasized that, while these programs appeared to have halted the decline in corporate investment, there was no significant increase in lending either. Distressed lenders did not lend more heavily. However, companies that were able to obtain new loans from lenders who received LTRO financing increased their investment. Finally, the authors provided an interesting counterfactual analysis, explaining that non-Eurozone companies that did not have direct access to LTRO financing decreased their investment more than those in the Eurozone (Daetz et al. 2018).

Some analysts argued that the LTROs did not adequately address their primary goals. Italian and Spanish bond yields, while positively impacted by the first round of three-year LTROs, still continued to increase in the later parts of 2012. Spain specifically saw a precipitous increase in its 10-year bond yield, from 5% to 7.6% from March 2012 to July 2012 (Sia Partners 2012). The deteriorating quality of some sovereign debt was also an issue both at the ECB and in periphery banks, which had bought the debt in bulk. For the ECB, a default by any of these periphery countries, however unlikely, would seriously damage the quality of some of the assets on the ECB’s balance sheet. For periphery banks, Sia Partners used Fitch downgrading Santander and BBVA to BBB+ shortly after doing the same to Spain’s credit rating a week prior as an example, suggesting that many of these banks were now tightly tied to their sovereigns. The Economist echoed similar sentiment, arguing that the austerity measures that Spain had instituted were actually harming its return to economic growth, and that the acute phase of the crisis was still quite close (“Draghi strikes back II”).

One evaluation written by Jean Pisani-Ferry and Guntram Wolff called these operations the Eurozone equivalent of QE, but that it did not have the same impact as its sister program. This was predicated on the fact that approximately €700 billion was put into the ECB deposit facility rather than lent out, suggesting that the ECB had become a large substitute in the interbank lending market.^FSee Figure 2 for more information. Additionally, the authors mentioned significant asymmetry in the behavior of banks in the northern part of the Eurozone versus the southern part, explaining that southern European banks are much more heavily reliant on refinancing operations and still have decreased the amount of lending done, contrary to institutions in northern Europe. This asymmetry also extended to the behavior of highly-rated (AAA) issuers, which saw a minimal impact on the yield curve, compared to financially constrained issuers. The authors finished by discussing that the lack of a banking and fiscal union and substantial differences amongst Euro members contributed to massive monetary policy instruments like the LTROs not being as effective (Pisani-Ferry and Wolff 2012).

Figure 3: Key ECB Interest Rates (%)

Source: ECB Statistical Data Warehouse

In an analysis of the French banking sector, the Bank of France explained that most of the monetary policy transmission that one would expect from such a substantial operation happened after the first LTRO auction in December. The impact of the second auction was much more muted. Demand for the facility was highest in banks that were more liquidity constrained, and large, corporate borrowers tended to benefit the most from participating. Additionally, firms that had a previous history of strong profitability saw a positive impact on their supply of credit from lenders, suggesting that the policies did not lead to “zombie lending.” Smaller firms, which were more reliant on bank funding and consequently more vulnerable to a credit crunch, saw a positive effect, but not to the same degree that large and “intermediate” borrowers did. The top 1% of borrowers had an increased credit supply that was 63% larger than the average firm used in the 24-bank sample. The most important finding to come out of this study was that, according to the authors’ base estimate, every €1 billion in liquidity borrowed from the facility is associated with a credit increase of only €186 million (Andrade et al. 2017).

The Federal Reserve’s analysis of the operations found that private credit absent the LTROs would have contracted 2% more, at 5.6% compared to 3.6% with the facilities. This result and the subsequent restoration of private credit was driven by lenders that were more exposed to the credit crunch. On the other hand, less-exposed lenders tended to use the attractive, low-rate funding to purchase higher-yield sovereign bonds, often from peripheral countries. In the case of Italy, which had established a government guarantee program, the authors found that many banks, regardless of the extent of their exposure, would self-select into the government guarantee program and then pledge those same securities as collateral. While more liquidity-constrained banks used the guarantee more, larger, less-affected banks more regularly exhibited the “reach-for-yield” behaviors described above. Specifically, the Fed found that, for every Euro borrowed, exposed and less-exposed banks invested €0.44 and €0.83 in government bonds, respectively. More-exposed banks invested €0.13 of every Euro in private credit (Carpinelli and Crosignani 2017).

One finding that appears to be mostly constant throughout various analyses of the program is that more liquidity-constrained and stressed institutions, often located in “riskier” peripheral countries, used the facilities disproportionately more. Miguel García-Posada and Marcos Marchetti discussed the bank lending channel of the monetary policy transmission of LTROs extensively, saying that the operations had “a positive, moderate-sized effect” on the supply of bank credit to firms. They estimated that annual credit growth directly attributable to the LTROs ranged from 0.8% to 1%, and that SMEs benefitted more from this credit growth because they had fewer options to raise funding that larger firms could tap. Large firms, especially those with strong lending relationships with financial institutions, were not significantly affected at all because, as the authors explain, “relationship lending is a more stable source of credit than transaction lending” (García-Posada and Marchetti 2016).

Thus, the evaluations suggested that the three-year LTROs had a mixed impact. This was due in large part to many institutions opting to keep substantial amounts of borrowed funds at the ECB, a lack of significant upward movement in major macroeconomic fundamentals, and other noneconomic factors, such as the structure of the Eurozone itself. Despite these endogenous and exogenous weaknesses, the most favorable finding was that the hit that funding markets had taken as a result of the stress induced by the sovereign debt crisis had been weathered. The precipitous falls in various macroeconomic and financial aggregates slowed, though their recovery would be slow despite continued ECB intervention.

Banks that were hit the hardest by the credit crunch tended to use the facility the most, and their less-affected counterparts still used the facilities but did little to drive the restoration of private credit. Less-exposed lenders opted instead to invest cheap central bank financing in high-yield sovereign bonds. The overall effect on the bank lending channel and, specifically, private credit, was relatively small compared to the size of the program. However, the facilities appeared to halt or considerably dampen the freezing of the Euro-area money market and prevented a credit crunch from leaking into the real economy. In a press release following the allotment of the first three-year LTRO, Mario Draghi stated that, “. . . we do think that this decision has at least prevented a credit contraction that would have been more serious, far more serious (Draghi and Constâncio 2012).”