RDP 9703: The Implementation of Monetary Policy in Australia 4. The Impact on Financial Markets

The move to a transparent approach to implementing policy has reduced volatility in short-term interest rates and resulted in faster pass-through of policy changes to lending and deposit rates of financial intermediaries.

4.1 Volatility in the Overnight Interest Rate

The panels in Figure 1 show daily observations of the overnight interest rate for the period since the mid 1980s. The shaded areas in the pre-announcement period show the informal operating levels that guided market operations.

Figure 2 shows a measure of cash rate volatility over the same period.^[15]

The main points to note from these two figures are as follows:

• First, volatility tended to decline progressively in the pre-announcement period, as operating techniques were gradually refined after the start of deregulation.^[16] By late 1989, just prior to formally announcing changes in policy, daily volatility had declined to around 25 basis points, compared with around 100 basis points in the mid 1980s.
• Second, after the move to formally announcing policy changes, volatility fell noticeably further. Currently, it is rare for the cash rate to deviate by more than a few of basis points from the announced target. One of the reasons for this appears to be that market participants assume that the Bank will keep rates very close to the target rate, so that the demand curve for settlement balances tends to be very elastic around the announced rate. The tighter range for the overnight rate has not required the Bank to be more active in managing the supply of reserves. Indeed, the opposite is the case. In the pre-announcement period, around one in five daily market operations were more than one-quarter of the size of banks' settlement balances.^[17] In contrast, in the post-announcement period, just one in 10 daily operations are associated with transactions of this magnitude.

4.2 Influence on Other Short-term Interest Rates

Short-term market yields are, of course, closely related to expectations of the future level of the cash rate. With the Bank announcing the target cash rate, and daily volatility in the cash rate much reduced, one would expect a commensurate decline in volatility in short-term yields, with yields unlikely to move significantly unless views about the future direction of monetary policy change.

The three panels in Figure 3 show, respectively, the distribution of daily movements in 30-day bill yields, 90-day bill yields and 10-year bond yields, divided into the pre- and post-announcement periods.^[18] For bill yields, movements in the pre- and post-announcement periods are strikingly different. In the pre-announcement period, yields moved by more than 10 basis points on almost one in two days. In contrast, the distribution of the daily change in short-term rates in the post-announcement period tends to be much more clustered around zero. More than 90 per cent of changes are 10 basis points or less, with rates rarely moving by more than 25 basis points on any given day. The strong implication is that much of the noise in movements in short-term rates has been removed by increasing the transparency of monetary-policy actions. As expected, the change in procedures has not had a significant impact on the distribution of daily movements in bond yields. These yields are influenced more by broad domestic and international economic developments rather than changes in cash rates.

4.3 Impact on Futures Trading

With relatively deep and liquid futures markets in Australia, many investors find it more cost-effective to take on or unwind interest-rate exposures in the futures markets rather than buying or selling securities outright. Turnover of 90-day bill futures, for instance, is more than tenfold the turnover in physical bank bills. We might expect, therefore, that the announcement of changes in policy may initially show up in the level of activity on futures exchanges.

Figure 4 shows the average volume of 90-day bill futures and 10-year bond futures contracts traded around the dates of changes in policy. Volumes have been adjusted to take account of the natural growth of the market, by expressing daily turnover as a ratio of the average daily turnover in that year. A figure around one, therefore, implies that turnover was normal on that day, numbers above/below one indicate the extent to which turnover was above/below average.

In the period when announcements were not made, there was little impact upon trading on the day when the Bank moved to implement a change of policy. This is not surprising, given that it took time for market participants to assess whether the movement in the cash rate represented a change to a new operating level or day-to-day noise around an unchanged target. Over the course of the next few days, however, it typically became apparent to the market that policy had, in fact, been changed. The market gradually adjusted, as more participants acknowledged that a change had taken place. Activity on the futures market rose, peaking at around 40 to 50 per cent above average four days after the Bank moved to implement the policy adjustment. The adjustment continued over the course of a couple of weeks.

In contrast, portfolio balancing occurs at a much faster pace now that policy changes are announced. Turnover surges on the day of the announcement. Trading in bill contracts is about 2.5 times larger than average, with turnover in bond contracts almost 1.8 times higher than average. There is some residual rebalancing on the day following the announcement, probably reflecting offshore investors adjusting their exposures. Turnover quickly settles to around average levels thereafter.

Figure 5 shows the behaviour of the physical 30-day bill yield around the time of a change in monetary policy.^[19] In the post-announcement era, expectations of a change in policy have typically built up in the month prior to the rate change, so that about half the adjustment was typically priced into 30-day bill yields before the announcement; the change is fully priced in on the day of the adjustment. One reason why markets now better anticipate policy changes is because, by making announcements, the Bank over time highlights the issues which are important in setting monetary policy. Market participants are therefore in a position to assess data in the economy as they become available and thereby make assessments about policy settings. In the pre-announcement period, the adjustment path was very different. The market typically priced in only a small part of the change in policy prior to its implementation. It then went on to continue adjusting short-term yields in the two or three weeks following the policy change, usually overshooting. In other words, the adjustment phase was slower and much less well-defined.

4.4 Pass-through into Lending and Deposit Rates

In Australia, a high proportion of banking products have traditionally been priced off short-term interest rates. Most loans for housing and for business have variable interest rates, which are set by banks on the basis of the level of short-term interest rates. With policy changes now more transparent, their flow-through to deposit and lending rates could be expected to be faster. Table 4 shows the time lapse between a change in the cash rate and the movement in the 3-month deposit rate, the mortgage rate and the business loan indicator rate for each episode of policy easing and tightening over the past decade. It shows the average time it has taken banks to bring about a change in deposit and lending rates in response to changes in monetary policy.^[20]

The lag with which banks altered their deposit and lending rates fell substantially in the second half of the 1980s, and was reduced further in the post-announcement period. In the mid 1980s, banks took about one month to adjust deposit rates and up to four to five months to adjust lending rates in response to tightenings of policy. The lags were roughly halved in the tightening cycle in 1988–89 and halved again in the 1994 phase. Changes to deposit and lending rates are now moved within one to four weeks of the change in cash rates compared with upwards of 20 weeks in the pre-announcement era.^[21]

Another change is in the way banks pass through rate changes. In the pre-announcement period, business rates tended to be changed in small amounts over a long period, with changes occurring in up to 10 steps. In contrast, recent changes to cash rates have tended to be fully passed through in one step.

There are many other variables, such as the structure of banks' deposits, the riskiness of bank lending and the degree of competition in banking, that affect the extent and the speed by which changes to monetary policy are reflected in commercial banks' borrowing and lending rates. The gradual decline in banks' deposits which paid a low interest rate and increased competition from non-bank institutions in the housing market would have contributed to the faster pass-through of monetary policy changes. Nevertheless, the unambiguous signal that policy has altered has lead to a much greater awareness amongst banks' customers of the extent and the speed by which changes in policy show up in changes in banks' interest rates.

Footnotes

For the period prior to making policy announcements, the observations represent the standard deviation of the actual daily cash rate from the centre of the informal trading band; after January 1990, volatility is measured using differences in the actual and announced cash rate target. [15]

Significant changes included: the start of dealing in repurchase agreements (repos) with the authorised dealers; allowing authorised dealers to deal in repos with others in the market; increasing the flexibility of the terms on which the Bank deals in repos; and the use of foreign exchange swaps. [16]

Prior to July 1996, we have used banks' loans to authorised dealers as the measure of reserve balances. [17]

The distribution for movements in 180-day bill rates is virtually the same as other bill yields. [18]

Again, in the pre-announcement period, the mid-point of the band is taken as the appropriate reference point. [19]

An earlier paper by Lowe (1995) looked at the extent of the pass-through for the period from 1986 to 1994. He estimated that the pass-through of changes in the cash rate to 3-month money market rates, housing mortgage rate and business indicator rates were 0.97, 0.65 and 0.89, respectively, over the long run. [20]

These figures refer to the time taken for the change in rates to become effective. Lags in announcement of rate changes are significantly shorter. [21]