RDP 2024-03: Demand in the Repo Market: Indirect Perspectives from Open Market Operations from 2006 to 2020 5. Elasticity of Demand and What It Can Tell Us

We define elasticity as the percentage change in the quantity of liquidity demanded in response to a 1 per cent change in the spread between the repo cut-off rate (the lowest rate accepted in the auction) and OIS. We consider the demand response to percentage changes in the spread, rather than absolute changes in basis points, because the magnitude of any pricing shock is likely to be proportional to the level of the repo spread. For example, if the repo spread was narrow and stable, a 2.5 basis point increase from 5 basis points to 7.5 basis points would be regarded as a significant event (i.e. a 50 per cent increase). However, if the repo spread was 50 basis points, an increase to 52.5 basis points would most likely be considered as less significant. Using an empirical framework, we estimate the price elasticity of demand for liquidity obtained under reverse repo in open market operations. We produce a quarterly estimate for repo elasticity, with the model controlling for fixed effects due to the repo term and specific dealing days (such as quarter-end; for details see Appendix B).

Knowing the price sensitivity of demand is informative because it helps explain market conditions and the likely price response to a given increase in liquidity supplied. To elicit the same percentage decrease in the repo spread, the Reserve Bank does not need to supply as much liquidity when demand is inelastic as it would when demand is elastic. Generalising this finding from open market operations to the broader market implies that participants have reason to be more or less willing to accept cash lent under repo for a given fall in the price. This finding opens up avenues for further research to determine the drivers of such changes in behaviour.

5.1 The evolution of elasticity

As was the case with our visual representation of the data, our estimate confirms that elasticity varies considerably over time and is positively correlated with the repo spread (Figure 9). Between 2009 and 2014, the repo spread averaged around 3 basis points. Demand was relatively inelastic, with a 1 per cent decrease in the repo spread associated with an average increase in quantity demanded of just 0.5 per cent. This result can also be used to estimate the percentage increase in supply that would be needed to narrow the repo spread by 1 per cent (since the supply curve was vertical over this period). Our findings imply that in response to a 1 per cent increase in the daily supply of liquidity, the repo spread would have narrowed by 2 per cent. Over the period, the average quantity supplied in a given day was about $1.3 billion. To narrow the spread by 1 per cent would therefore have required an increase in daily supply of only around $6.5 million.^[14]

Between 2016 and 2019, the estimated demand curve became notably more elastic. At the time the repo spread was also much higher than in the earlier period, averaging around 26 basis points. In this period, a 1 per cent increase in the daily supply of liquidity would have narrowed the repo spread by just 0.3 per cent. Given that the average supply of liquidity each day was $2,098 million, a 1 per cent narrowing of the repo spread would have required an increase in daily supply of $61 million. Compared with 2009 to 2014, this is about ten times the increase in daily liquidity needed to achieve a similar percentage narrowing in the repo spread. Extrapolating this, a rough estimation of the increase in supply required to bring the average repo rate back down to a zero spread in a single day during this period is around $6.1 billion. This is significantly more than in the 2009 to 2014 sample. When the repo spread peaked at around 50 basis points in 2018, the amounts in question would likely have been much larger. In other words, between 2016 and 2019 the increase in supply necessary to narrow the repo spread was significantly larger than in 2009 to 2014.

Although there are other factors that also would have influenced these outcomes, it is hardly surprising that when prices are higher due to rising and more elastic demand, more supply than usual is required to meet that demand in order for prices to fall back to their starting point.

5.2 The COVID-19 period

After peaking in 2018, the repo spread began to narrow and elasticity trended lower. This narrowing of the repo spread accelerated after the Reserve Bank's response to the pandemic, where the supply of liquidity was increased significantly. The relatively low elasticity of demand leading into the crisis may have also increased the impact of these liquidity injections on repo spreads.

The price elasticity of demand cannot be estimated from April 2020 onwards, because the repo rate over this period has been constant (aside from the change in policy rates in November 2020). Auction participants at that time typically bid at the single fixed repo rate. Therefore, there was no longer a range of repo spreads over which elasticity can be estimated. The demand curve effectively became close to perfectly elastic (flat). Notably though, demand was truncated at relatively low quantities given ample system cash since March 2020.

Footnote

Our results are best interpreted using small changes from the point at which the elasticity is calculated. In our specification we assume that the demand curve is convex to the origin; an assumption justified by the data. With a convex demand curve, a simple linear extrapolation would underestimate the required increase in daily supply. [14]