Rupal Patel
Repo markets form part of the plumbing of the financial system. They allow participants to borrow cash against collateral and buy back this collateral at a higher price at the end of the transaction. When there is a blockage in repo it has repercussions on financial markets. Since 2014 there have been significant changes in repo functioning, causing policymakers to question why these changes are happening and what it means for financial stability. Our paper addresses these questions. We find fluctuations in repo were driven by changes in dealers’ supply in the pre-Covid period 2014–18. We subsequently consider the possible role that the introduction of the leverage ratio played in the willingness of intermediaries to respond to demands for cash.
Importance of repo markets
Repo transactions are vital for facilitating the flow of cash and securities around the economy. In doing so, repo markets support a wide range of investment and risk management activities for banks and other financial market participants such as pension funds. Small changes in the cost or availability of repo can therefore have a significant impact on financial markets and the real economy.
Recent developments
While our analysis does not examine changes in repo market during the Covid crisis, we find there was been a sharp deterioration, and then improvement, in the functioning of repo markets secured by sterling UK government bonds, or ‘gilt’ repo (Chart 1) over the period from 2014 to 2018. These changes likely arose due to a combination of both changes in dealers’ supply of – and end users’ demand for – repo market intermediation. Broader gilt market conditions may also have had an impact as they change market participants’ view of the future value of gilts, and hence the cost of repo transactions that use them as collateral.
Chart 1: Three-month gilt repo rates paid by end-users in excess of expectations of policy interest rates(a)
Sources: Bloomberg and Bank calculations.
(a) Expected policy interest rates are measured by three-month indexed swap rates.
Decomposing changes in the functioning of the sterling repo market
We disentangle these factors using a structural vector autoregression (SVAR) model that allows us to see the degree to which different factors have contributed to changes in repo prices and volumes over time. It identifies changes in price and quantity of repo transactions that match a set of assumed directional responses consistent with changes in the supply and demand of market intermediation. See Table 1 for a full set of sign restrictions. SVAR models are used widely in monetary policy literature but to our knowledge, it’s the first time it has been used to examine the functioning of repo markets. We use repo volumes data collected by the Prudential Regulation Authority on 17 banks in the gilt repo market and repo rates data from Bank of England’s Bankstats.
Table 1: Identifying sign restrictions
Chart 2 shows the decomposition of the change in the level of the spread between January 2014 and December 2015 – a period in which repo prices were increasing. The black line represents the changes in the spread compared to its level in January 2014; coloured bars represent the cumulative contribution of supply, demand and gilt market conditions shocks identified by our model.
Chart 2: Decomposition of the change in the cost of repo from 2014 to 2015(a)(b)(c)
Sources: Bloomberg and Bank of England.
(a) Moving quarterly averages.
(b) Expected policy interest rates are measured by three-month swap rates.
(c) The difference between the sum of factors and change in repo spreads is due to the algorithm used in estimating the model.
The bulk of the increase in the spread over this period was driven by a reduction in the supply of dealer repo intermediation (pink bars). One possible explanation may in part be due to the announcement of the UK leverage ratio in 2014 and so banks’ preparations for formal implementation. Repo intermediation increases the size of banks’ balance sheets and so attracts a capital charge under the regulatory leverage ratio. The leverage ratio was introduced by policymakers to, among other things, enhance dealer’s resilience. A potential by-product of the leverage ratio is that it makes it relatively more costly for dealers to carry out low profit margin activities such as intermediation in repo markets at least temporarily, as argued by Kotidis and van Horen (2018). As such, it may have reduced dealers’ willingness to act as repo market intermediaries, or increased the compensation they require for doing so.
That said, some of the increase in the spread is also driven by an increase in the demand for repo intermediation. This might be a result of dealers increasing their trading activity which may have been funded through repo following the end of the Eurozone debt crisis.
We also examine the drivers behind the improvement in repo market functioning between 2016 and 2018, during which the repo spread decreased (black line in Chart 3). During this period the repo spread peaked around June 2016 – around the time of the UK’s EU referendum. We estimate that around 70% of the subsequent improvement in the spread was driven by an increase in the supply of intermediation (purple bars in Chart 2). This increase in supply may follow from dealers’ drive to optimise their balance sheets in light of the introduction of the leverage ratio to manage regulatory costs. Such optimisation includes the increased use of balance sheet netting which enables dealers to provide more repo without paying additional regulatory costs. This netting can be done bilaterally (dealer to dealer) or with a CCP (dealer to CCP).
Chart 3: Decomposition of the change the cost of repo from 2016 to 2018(a)(b)(c)
Sources: Bloomberg and Bank of England calculations.
(a) Moving quarterly averages.
(b) Expected policy interest rates are measured by three-month swap rates.
(c) The difference between the sum of factors and change in repo spreads is due to the algorithm used in estimating the model.
Demand factors play almost no role in explaining the fall in the cost of repo. In contrast, around 20% of the overall reduction in prices arises due to an improvement in broader gilt market conditions.
Impact of regulation
We also use our model to assess the impact of post-crisis regulations and how this may have changed the degree to which the sterling repo market is able to accommodate increases in the demand for cash and collateral. To do so, we compare how the prices and volumes of repo transactions respond to a positive shock to demand for repo market intermediation before and after the introduction of the UK leverage ratio in January 2016.
We run our SVAR model on overnight repo spreads and overnight repo amounts outstanding instead of analysing the market for three-month contracts. This is because end-users are likely to demand shorter maturity repo in stressed market conditions (e.g. to meet margin calls) than term repo. We also split our sample into subsamples that relate to the period before and after the introduction of the UK leverage ratio on 1 January 2016.
We estimate that after the implementation of the UK leverage ratio, repo prices respond more (6 basis points versus 3 basis points) and repo volumes less (£2 billion of transactions outstanding compared to £4 billion) than before the implementation of the leverage ratio (Chart 4). In other words, despite the overall reduction in repo spreads during this period, repo markets are less able to accommodate an increase in demand compared to in the period prior to the introduction of the UK leverage ratio. This is consistent with a more inelastic supply of repo intermediation by regulated market intermediaries.
Chart 4: Sensitivity of overnight repo markets to a positive demand shock before and after the introduction of the UK regulatory leverage ratio(a)(b)(c)
Sources: Bloomberg PRA regulatory returns and Bank calculations.
(a) First week response to a one standard deviation increase in demand for imtermediation.
(b) Period prior to the introduction of the UK regulatory leverage ratio: 2011–15.
(c) Period following the introduction of the UK regulatory leverage ratio: 2016–18.
Finally, we consider how repo prices and quantities respond to increases in demand for repo transactions that are subject to different forms of intermediation. That is transactions that incur regulatory cost under the leverage ratio (non-nettable transactions) and those that are transactions via central counterparties and incur no regulatory costs (CCP nettable transactions).
Finally, we consider how repo prices and quantities respond to increases in demand for repo transactions that are subject to different forms of intermediation. That is transactions that incur regulatory cost under the leverage ratio (non-nettable transactions) and those that are transactions via central counterparties and incur no regulatory costs (CCP nettable transactions).
Chart 5 compares the response of spreads and amounts outstanding to a one standard deviation positive demand shock to CCP-nettable and non-nettable transactions. Non-nettable transaction spreads respond three times as strongly as those that are CCP nettable. In contrast, nettable transaction volumes respond by around a third more than those that are non-nettable. This finding is consistent with the notion that dealers may be less willing to expand their balance sheets to accommodate non-nettable repo transactions because they incur greater cost under the leverage ratio.
Chart 5: Sensitivity of the cost and volume of centrally cleared and bilateral gilt repo transactions to a positive demand shock(a)
Sources: Sterling Markets Dataset and Bank calculations.
(a) First week response to a one standard deviation positive demand shock.
Policy implications
The main finding of this post is that changes to repo prices were largely driven by changes in supply by repo dealers over the period from 2014 to 2018. A host of factors affect supply, one of which may have been the recent introduction of the leverage ratio, which may have increased dealers’ relative costs of intermediating in repo markets, at least temporarily. The leverage ratio was introduced post-crisis by policymakers to enhance dealers’ resilience and thereby help improve the safety of the financial system. This post does not consider these important benefits of the leverage ratio as part of the analysis. Nor does it identify the relative contribution of different drivers of supply including changes in regulation or banks’ risk appetite.
But our results are consistent with the notion that the introduction of the leverage ratio may have played a role in reducing the willingness of intermediaries to respond to demands for cash. And this effect appears to be stronger for transactions that incur regulatory costs. Our findings suggests that nettable centrally cleared repo transactions may benefit repo market resilience. This could support the case for widening access to central clearing.
Rupal Patel works in the Bank’s Macro-Financial Risks Division.
If you want to get in touch, please email us at bankunderground@bankofengland.co.uk or leave a comment below.
Comments will only appear once approved by a moderator, and are only published where a full name is supplied. Bank Underground is a blog for Bank of England staff to share views that challenge — or support — prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.
I don’t understand the premise: “They allow participants to borrow cash against collateral and buy back this collateral at a higher price at the end of the transaction”. Do they sell the asset, and thus need to buy it back, or do they borrow against it, in which case they do not have to buy it back, but merely have to repay the debt secured on it? And, in either case, why should the banking system make special efforts to help people to do this?
I find the implicit assumption in this paper that reduced spreads correspond to “better” repo market functioning to be questionable. Taking that analysis to the extreme, a negative spread would be “better”. I would argue that the changes in spread precisely represent appropriate repo market functioning, and that the range of spreads represented in this analysis had minimal market impact… or to put in another way, anyone who was priced out of the market at these rates was carrying too much risk in the subject asset. In light of the underlying rate, all of these rates represent extremely ready supply of credit and loose monetary policy, which is probably a much greater risk to all of the market participants long term than the change in spreads.
Thank you Robert and Robert for your comments.
Robert Cottrell- You can think of the repo market as a pawn shop for financial market participants such as market-makers and other securities dealers in firms called broker dealers or investment banks. Customers or market participants pledge collateral in return for a cash loan. When the loan is paid back, their collateral is returned to them.
The market plays a key role in facilitating the flow of cash and securities around the financial system, benefiting both financial and non-financial firms. For example, they allow prompt borrowing of securities to prevent settlement fails, and are commonly used to raise cash for use as margin in derivatives transactions. In stressed conditions, repo markets offer banks a route to convert their holdings of liquid assets into cash. If banks are unable to access the repo market they may be forced to sell liquid assets, which could exacerbate stress. Selling assets outright may incur losses from the unwinding of interest rate hedges
Repo also supports liquidity in other markets which contributes to the efficient allocation of capital to the real economy. So a well-functioning repo market is crucial for financial stability and for the efficient transmission of monetary policy.
Robert Seaton- Low spreads don’t necessarily represent a better functioning repo market and very low spreads might be illustrative of too loose credit conditions. Excessive use of repos can facilitate the build-up of leverage and encourage reliance on short term funding. Fluctuations in spreads may not have a large impact on overall market functioning in normal times. But in times of stress it’s important the repo market is accessible to market participants in need of liquidity. In stressed conditions, the evaporation of liquidity can lead to disorderly movements in prices, undermining the confidence of issuers and investors. The analysis does not consider the optimal cost of repo. But market intelligence from market participants suggest a decline in repo market functioning between 2014-2016. You can find more information here: https://www.bis.org/publ/cgfs59.pdf