RDP 2022-07: The Term Funding Facility: Has It Encouraged Business Lending? 3. The International Context

3.1 Other central bank term funding schemes

A number of other central banks also introduced term funding schemes in response to challenges arising from the outbreak of the COVID-19 pandemic, including the Reserve Bank of New Zealand (RBNZ), the Monetary Authority of Singapore (MAS), and the Sveriges Riksbank.^[5] Additionally, some other schemes have been operating for around 10 years or more, including those introduced by the Bank of Japan (BoJ), the ECB and the Bank of England (BoE) as they sought to impart further stimulus at a time when policy rates were near the effective lower bound and the supply of credit had contracted.

Term funding schemes implemented by other central banks have varied in their design. Low-cost funding is a common feature across all term funding schemes, and the price is often tied to a policy rate. The overall price of the scheme may also incorporate additional fees or discounts that vary depending on certain criteria, and, in some cases, the price has been used to incentivise banks to direct lending to particular sectors of the economy. The BoE, ECB and Riksbank introduced schemes in which the cost of funds available to banks has varied depending on their volume of lending to the private sector.

Banks' allowances under the scheme is another parameter that varied across schemes. Similar to the TFF, some schemes offered banks a fixed initial borrowing allowance based on their total credit outstanding, plus an additional allowance that is contingent on their volume of new lending during a particular period. Other schemes offered more targeted credit support by providing an allowance solely dependent on the volume of a particular type of lending, and banks could only draw on the facility by posting these loans as collateral. In contrast, in the Riksbank's onward lending program banks did not have individual allocations. Instead, banks would bid for funds at a weekly auction until all funds were exhausted.

Since the onset of COVID-19, several term funding schemes made favourable lending conditions contingent on the volume of lending to SMEs, similar to the TFF. Such schemes were introduced by the BoE, Banco de México, and the BoJ. Other schemes have directly complemented government loan programs by linking funding allowances to the use of these programs. This included the US Federal Reserve's liquidity facility, which extended credit to facilities that originated loans under the government's Paycheck Protection Program, with the loans to be used as collateral. The RBNZ and the MAS introduced similar programs where loan sizes depended on the volume of lending under a government loan program.

Many term funding schemes required the provision of collateral by the banks to mitigate financial risks to the central bank. Often, collateral requirements were aligned with a central bank's eligible collateral policy for repo funding through open market operations. Some central banks expanded their collateral requirements as part of their wider COVID-19 monetary policy measures, such as the ECB and the BoE. For example, some central banks accepted a wider range of collateral for these facilities than typically accepted for open market operations. This included the TFF which, as discussed above, allowed banks to post self-securitisation notes for funding (avoiding a drain on their liquid assets or the need to borrow to purchase eligible collateral). As noted above, other central banks allowed banks to post certain types of loans as collateral, such as those made through government loan programs or made to select sectors.

3.2 Literature

This paper contributes to the growing literature studying the effect of term funding schemes implemented around the world and, to the best of our knowledge, is the first empirical study of the effect of the RBA's TFF on bank credit supply in Australia. Few studies to date have examined the effects of central bank term funding schemes throughout the pandemic. In a comprehensive examination of euro area monetary and prudential policies implemented during the pandemic, Altavilla et al (2020) concluded that lending to firms would have been 3 percentage points lower in 2020/21 – 2021/22 in the absence of the ECB's third series of the TLTRO (TLTRO III).^[6] Da Silva et al (2021) also found significant positive effects on credit supply from TLTRO III in the euro area. A couple of studies have examined the effects of policies within countries. For example, in a study using Austrian bank-level data, Kwapil and Rieder (2021) found a positive effect of TLTRO III on Austrian banks' loan supply between July and September 2020.

The US Federal Reserve implemented several targeted lending programs in response to the pandemic, including the Main Street Lending Program (MSLP) and the Paycheck Protection Program Liquidity Facility (PPPLF). Minoiu, Zarutskie and Zlate (2021) found that, while overall take-up was limited, bank participation in the MSLP was associated with relatively less tightening of commercial and industrial lending standards than non-participating banks. Additionally, participating banks were more likely to increase business lending. Anbil, Carlson and Styczynski (2021) found that banks that used the PPPLF extended over twice as many paycheck protection program loans compared to non-participating banks.

Notwithstanding the limited literature on central bank term funding schemes during the pandemic, several papers have examined the effect on banks' lending of the first and second series of the ECB's TLTROs, introduced in 2014 and 2016, respectively. The literature generally found that the first two TLTRO series were effective in increasing loan supply or decreasing interest rates for participating institutions relative to non-participating institutions (e.g. Afonso and Sousa-Leite 2020; Esposito, Fantino and Sung 2020; Andreeva and García-Posada 2021; Benetton and Fantino 2021; Laine 2021). Many of these papers employed difference-in-differences approaches to compare the behaviour of participating and non-participating institutions. These studies also generally used instrumental variables to account for endogeneity issues in the self-selection of banks into accessing central bank funding. Some papers examined both the extensive (i.e. participation in the TLTRO relative to non-participation) and intensive margin effects (i.e. the amount of funding), with mixed results as to the relative importance of these effects.

Literature on the effects of other central banks' funding schemes remains sparse.^[7] Examining the BoE's Funding for Lending Scheme, Havrylchyk (2016) found no effect of sharpened incentives for lending to SMEs on SME credit growth relative to large businesses. Havrylchyk investigated the effectiveness of the incentive mechanisms in the scheme for banks to expand their supply of lending to SMEs by exploiting the BoE's 2013 amendments to the scheme's rules. When the Funding for Lending Scheme was first implemented in 2012, banks' additional allowances were set such that a £1 increase in net lending would grant a bank £1 of additional allowance. However, from April 2013 to 2014, the BoE amended this allowance so that, for every £1 of net lending to SMEs, banks were granted £10 of additional allowance.

3.2.1 Identification challenges in the literature

Empirical papers assessing the effect of term funding schemes on credit face a few common identification challenges. Firstly, banks could decide whether or not to access central bank funding. The self-selection of banks into receiving funding raises an endogeneity issue as banks with certain characteristics (for example, those who would have increased lending anyway) may be more likely to participate than others. Empirical papers often address this issue through an instrumental variable strategy. Various papers studying the effect of the ECB's TLTROs on credit supply used banks' predetermined borrowing allowances or usage in previous funding series as an instrument for TLTRO uptake (including Andreeva and García-Posada (2021); Benetton and Fantino (2021); Kwapil and Rieder (2021) and Laine (2021)). The rationale typically provided for this instrument is that borrowing allowances are set by the ECB and based on the stock of loans prior to the period being studied. For the US Federal Reserve's facility, Minoiu et al (2021) used proxies of bank familiarity with Federal Reserve procedures and facilities to isolate exogenous variation in participation.

Many of these papers found that this self-selection issue would bias the coefficients in an unclear direction; OLS estimates may be biased upwards if banks that accessed the facility would have increased their lending anyway. In contrast, the coefficients may be biased downward if banks that chose to borrow were experiencing worse unobservable funding problems, and therefore may have granted fewer new loans.

The second key challenge is disentangling the effects of credit supply and credit demand. Studies have commonly dealt with this challenge by adding fixed effects for the borrowing firms. In particular, several papers used the Khwaja and Mian (2008) method, which exploits the fact that many firms simultaneously borrow from several banks (Esposito et al 2020; Benetton and Fantino 2021; Minoiu et al 2021). This allows for a comparison of credit growth across different lenders for the same firm. However, this method relies on access to data covering firms that had loans from multiples lenders. Without access to these types of data, Kwapil and Rieder (2021) utilised a bank-specific covariate which takes into account each bank's lending composition and the effect of the COVID-19 pandemic on each sector. Laine (2021) employed a different approach, using country–time fixed effects and adding housing loans as a control variable.

Our paper contributes to the literature by assessing the effects of the RBA's TFF on the amount of business lending. As noted earlier, empirical papers on the effects of central bank term funding schemes during the pandemic remain limited. Furthermore, while some central bank funding schemes have included additional incentives to promote lending to particular groups such as SMEs, very few papers have examined the effects of these specific incentives. Unlike the bulk of existing literature, we specifically examine whether the additional incentives for SME lending provided in the RBA's TFF (in the form of higher funding allowances) were effective in stimulating increased lending to this sector. Our analysis contributes to our understanding of central bank funding schemes and their potential to encourage increased lending.

Footnotes

See Cantú et al (2021) for a database on central banks' monetary policy responses to COVID-19. [5]

TLTRO III was first introduced in 2019 and adapted in 2020 in response to the pandemic. Under TLTRO III, banks could borrow at a rate as low as –1 per cent if they exceed lending targets. [6]

Churm et al (2021) examine the macroeconomic effects of unconventional monetary policies in the United Kingdom, focusing on additional quantitative easing (QE2) and the Funding for Lending Scheme. They find positive effects on GDP and inflation. [7]