The ECB March 12th package reconsidered
The package of ECB measures that its President presented on March 12th sort of disappeared from public attention, being overshadowed by the much more forceful action decided just 6 days later in the Pandemic Emergency Purchase Programme. Still, this post argues that the March 12th package, while eventually being insufficient, included significant measures that should help credit growth. This is particularly the case for the measures decided by the supervisory function of the ECB, which received little attention both in the press conference and in the ensuing comments.
The March 12th package consisted of four components:
- The reduction by 25 basis points of the rate on the TLTRO refinancing, starting in June, to – 75 basis points, conditional on banks maintaining their lending levels between April 1st 2020 and March 31st 2021,
- The “subsidy” that the ECB pays to banks borrowing from it under the TLTRO, coming from the fact that banks can borrow at -75 basis points and invest on the ECB deposit facility at – 50 basis points,
- The reduction of the regulatory capital, and other subsidiary easing measures, decided by the semidetached part of the ECB responsible for banking supervision,
- An increase of 120 billion of the QE purchases until the end of the year.
The difficulty in assessing the package is twofold:
- With the exception of the rate reduction on TLTRO, the impulse from the ECB actions is not obvious,
- It would be useful to have a synthetic indication of the overall impact of the program, expressed in a variable which it is easy to interpret.
This post deals with these two difficulties, concentrating on the first three components of the program, while not addressing the QE increase. Any effect of the latter should, in any case, be assessed together with that of the new PEPP.
The first step to come to an overall assessment of the ECB package is to estimate the impulse coming from the three measures on which we concentrate.
The impulse from the reduction of the rate on the TLTRO is easy to estimate, since it can just be assimilated to a generalised rate reduction.
The estimate of the impulse from the subsidy by the ECB starts from the amount that banks can borrow under the TLTRO, 50 per cent of the stock of eligible loans 1)Total of lending to non-financial corporations and to households excluding mortgages (ECB Statistical Data Warehouse ) on February 28th 2019, i.e. €2904bn. If banks would fully exploit this opportunity, the subsidy would be 7.5 billion euro in the year between April 2020 and March 2021. Just for comparison, this would be about 6 per cent of the annual 2018 profit of the euro-area banking systems, 127 billion.2)https://sdw.ecb.europa.eu/quickview.do?SERIES_KEY=117.BSI.M.U2.N.A.T00.A.1.Z5.0000.Z01.E
Finally, quantifying the impulse coming from the supervisory side is the most difficult exercise. One can figure out that:
- Operating temporarily below the capital conservation buffer (CCB) would relieve 2.5% of capital requirement,
- Reducing fully Pillar 2 guidance, could lower the capital requirement by another 1.5%,
- If national macroprudential authorities heeded the ECB advice and relaxed the countercyclical capital buffer (CCyB), there could be another 1.1 % of reduction (the total actual buffer of 11.7% – 10.6 prescribed CET1).
While the reduction coming from the CCB is certain, the reductions coming from Pillar 2 guidance and the CCyB are more difficult to estimate, as the figures are not the same for all banks and countries. Indeed, not all countries have a CCyB. Still, adding the three components would give a maximum of 5.1 % relief, which looks very large.
In addition, the ECB added three complementary measures: allowing banks to fulfil the Pillar 2 requirement (P2R), equal to 2.1%, with Additional Tier 1 or Tier 2, making the Liquidity Cover Ratio less stringent and giving more flexibility to supervisors to adjust Non Performing Loans treatment to bank specific circumstances.
In aggregate, if we assume that the easing coming from the three above complementary measures offset the uncertainty about the reduction coming from Pillar 2 guidance and the countercyclical capital buffer, an overall estimate of a reduction of 5 per cent of capital requirement looks plausible. In the quantification of the overall ECB package carried out below, a more conservative estimate of 4 per cent is also used.
The second step to come to an overall assessment of the ECB March 12th package consists in bringing the three measures to a common and easy to interpret denominator, bank credit: what is the total estimated impact of the three impulses mentioned above on bank credit?
In order to carry out this exercise we need elasticities, or rather semi-elasticities, of bank credit growth with respect to the three components of the package we are considering.
Firstly, the effect of the reduction by 25 basis points of TLTRO credit depends on the elasticity of credit growth to a reduction of interest rates. In the Appendix, we report some selected estimates of this semi-elasticity. The literature does not give a univocal result and the variation between studies is substantial. Furthermore, the effect can depend on the state of the economy and the level of interest rates. Overall, however, a review suggests an elasticity between 1 and 2, thus the 25 basis points cut could result in a growth in credit between 0.25 and 0.5 percent over the 12 months following the rate cut. The relatively small effect should not be surprising, as the original impulse is comparatively low. Indeed, while rate cuts lower than 25 basis points have become customary in the current low rate environment, central banks in the past considered quarter point changes in their policy rates as the minimum level worth considering: a ¼ point change was at the limit of practical significance.
Secondly, when looking at the elasticity of bank lending to capital relief measures, it is worth noting that in most cases the opposite relationship is estimated: how an increase of regulatory capital reduces bank credit. One has to assume symmetry to derive what would be the effect of capital relief. Even with this symmetry assumption, a reading of the selected literature, as summarized in the Appendix, concludes that there are heterogenous results.
In order to have a first and necessarily approximated value of the impact on bank lending by capital relief regulations, we take, as a rule of thumb, 2 different values of elasticities for our exercise: 0.5 and 2. These numbers should not be considered as a precise estimate of the impact for any European country and any kind of lending contract, but rather represent an approximate average of all the possible situations.
This would give an effect on credit growth, depending, on the assumed impulse and the two different elasticities, of between 2.0 and 10.0 per cent. The range between the lower and the higher estimate is very wide, but both deserve the adjective of large. For comparison purposes, it can be recalled that bank credit growth has averaged 1.3 percent since 2010. 3)http://sdw.ecb.europa.eu/quickview.do?SERIES_KEY=117.BSI.M.U2.Y.U.AT2.A.1.U2.2000.Z01.E https://www.ecb.europa.eu/stats/ecb_statistics/key_euro_area_indicators/html/index.en.html
An estimate of the effect of the subsidy from the ECB to banks on their credit expansion is dealt with in the Appendix. The estimated elasticities give an impact on bank lending between 0.1% and 0.2%.
The total effect on bank credit of the three measures announced by the ECB in March can be estimated between 2.26% and 10.7%, depending on assumptions, as it can be seen in table 1.
The large effect, even taking the lower range of the estimates, should be taken more as a potential than as a firm consequence of the measures. Indeed, it could be even considered as a form of forbearance, since COVID-19 is making borrowers riskier. For the potential additional lending to become reality, fiscal measures of the kind proposed by Demertzis et. al., Constancio and Leipold are critical.4)https://www.bruegel.org/2020/03/economic-response-coronavirus/. A. Leipold, ‘Whatever it takes’ Why urgent fiscal policy action is key to Eurozone success’, The Lisbon Council, 9 March 2020. Vitor Constancio, in a 4 March Twitter thread, while agreeing that the brunt of the action should be taken by fiscal policy, argued that central banks have a role to play in ensuring appropriate liquidity.
Table 1. estimated credit growth effect of the measures in the ECB March package.
|Elasticity||Impact on Bank Lending|
|Lower Value||Higher Value||Lower Value||Higher Value|
Within the total in table 1, the most sizable contribution is exactly the one that received least attention in the press conference by the ECB President: the easing of bank capital charges. Indeed, she devoted few words to this issue in the introductory statement: “We also welcome the decisions taken by the ECB’s Supervisory Board, which are detailed in a separate press release published earlier today.” And in the questions and answers part of the press conference there was no emphasis at all on this measure that was, after all, approved by the ECB Governing Council. It would probably have been beneficial if, on this occasion, the ECB President had invited the Chair of the ECB Supervisory Board to participate to the press conference and explain the measure his Board had taken. While the ECB organizational structure is a bit awkward, with the supervisory function legally subordinated to the Governing Council but de facto independent, such an invitation would have not broken any regulation, but only precedent.
The criticism received in the economic press for the lack of a reduction of the deposit facility rate does not look justified for two reasons. First, there was a reduction of the rate on the TLTRO, which is arguably more relevant, in current conditions, than that on the deposit facility. Second, as the above reasoning shows, the effect of a 10 or even 20 basis points reduction of the deposit facility rate would have had small effects.
Another comment applies here: the ECB Supervisory Board understood that a macroprudential measure was needed and indeed it took it, moving well beyond the narrow microprudential scope within which some would like to constrain its activity. It was a pity that the message of the strong measure it decided did not go through during the press conference.
Of course, the need to come up with other measurees just a few days after the March 12th package shows that, even if stronger than assessed by observers, that package was not strong enough, given the mayhem affecting the euro area and indeed the global economy. One could also have hoped more innovative measures: the purchase of bank loans or a Derivative Market Program5)https://www.bruegel.org/2020/03/the-case-for-a-derivative-market-programme/ Reference to the post on DMP. have been proposed as possible innovations in the ECB armory. Still the package constitutes a significant measure by the ECB.
Literature about the policy rate and Bank Lending Elasticity
|Selected Literature||Sample Used||Estimated of the impact of 1 ppt change in the policy rate on bank lending in ppt (in 12 months)|
|Kakes and Sturm (2001)||Germany||0 – 2.5|
|Borio and Gambacorta (2017)||14 advanced economies||1.5|
|Gambacorta and Marqués-Ibáñez (2011)||EU & US||1.6 – 2.00|
|Imbierowicz, Löffler and Vogel (2019)||Germany||0*|
|Loupias, Savignac and Sevestre (2002)||France||2.5 – 3|
|Bernanke and Blinder (1992)||US||1.5 – 2|
* ‘We find that changes in the monetary policy stance, tightening as well as easing ones, are positively related to euro area lending rates but are not related to loan growth’.
It should be noted that the different studies employ slightly different measures for credit growth. Furthermore, some of the studies divide lending into categories based on the type of financial entity. The results provided on the table are a necessary approximation of the full results in the papers.
Literature about Capital Requirement and Bank Lending Elasticity
|Selected Literature||Sample Used||Estimated Impact of 1 ppt change in Capital Requirement on Bank Lending in ppt|
|Bernanke and Lown (1991)||US||2.00-3.00|
|Berrospide and Edge (2010)||US||0.145|
|Carlsson et al (2013)||US||0.13-2.00|
|Francis and Osborne (2012)||UK||0.060|
|Bridges et al (2014)||UK||0.410|
|Aiyar et al. (2015)||UK||6.000|
|Mesonnier and Monks (2015)||Europe||1.200|
|De Jonghe et al. (2016)||Belgium||0.13-1.80|
|Gambacorta and Mistrulli (2004)||Italy||0.744|
|Brun et Al. (2013)||France||5.000|
|Lebonne and Lame (2014)||France||0.91-1.11|
Each study has a different approach and they not only change the used sample, but also the regulatory framework and the definition of bank lending and the way they measure capital requirements. Furthermore, generally, the studies have more detailed results that consider the sector, the composition of the bank assets (Tier I, II and III) and further details. The results provided in the table are a necessary approximation of the full results presented in the papers. For our assumption and dataset, the studies that could be considered more similar are: Carlsson et al (2013); Bridges et al. (2014); and Lebonne and Lame (2014). Those suggest us as a reliable value of elasticity between 0.4 and 1.1., which one can round to 0.5 and 1.0 to avoid an undue sense of precision.
Literature about Central Bank Subsidy and Bank Lending Elasticity
|Selected Literature||Sample Used||Estimated Impact of 1 ppt change in Central Bank Subsidy on Bank Lending in ppt|
|Craig and Thomson (2001)||US||0.24-0.29|
|Craig and Thomson (2003)||US||0.20-0.30|
|Selected Literature||Sample Used||Estimated Impact of 1% change in Income to Asset Ratio on Bank Lending|
|Cohen and Scatigna (2014)||EU||25.91|
Notice that the two first studies are quite old and both about U.S. bank system. Furthermore, both consider only the case of small and medium banks that receive a subsidy from Fed.
Cohen and Scatigna6)BIS Working Papers No 443. Banks and capital requirements: channels of adjustment. Benjamin H Cohen and Michela Scatigna. March 2014. estimate for Europe a parameter of 25.91 for a 1 per cent increase in the income to asset ratio. Given the increase of this ratio from 0.400 to 0.409 % (0.009% change) because of the ECB subsidy of 7.5 billion €, the effect on credit growth would be 0.23%.
This post was written by Francesco Papadia and Marta Domínguez-Jiménez and Leonardo Cadamuro
|↩ 1.||Total of lending to non-financial corporations and to households excluding mortgages (ECB Statistical Data Warehouse|
|↩ 3.||http://sdw.ecb.europa.eu/quickview.do?SERIES_KEY=117.BSI.M.U2.Y.U.AT2.A.1.U2.2000.Z01.E https://www.ecb.europa.eu/stats/ecb_statistics/key_euro_area_indicators/html/index.en.html|
|↩ 4.||https://www.bruegel.org/2020/03/economic-response-coronavirus/. A. Leipold, ‘Whatever it takes’ Why urgent fiscal policy action is key to Eurozone success’, The Lisbon Council, 9 March 2020. Vitor Constancio, in a 4 March Twitter thread, while agreeing that the brunt of the action should be taken by fiscal policy, argued that central banks have a role to play in ensuring appropriate liquidity.|
|↩ 5.||https://www.bruegel.org/2020/03/the-case-for-a-derivative-market-programme/ Reference to the post on DMP.|
|↩ 6.||BIS Working Papers No 443. Banks and capital requirements: channels of adjustment. Benjamin H Cohen and Michela Scatigna. March 2014.|