Inflation and the exchange rate

4.28  The exchange rate affects the inflation rate through both the demand and the supply side, and not just through demand. An appreciation of the exchange rate reduces the cost of imports and therefore has a negative effect on the price level and hence on inflation. The effect on inflation is not permanent and will decline steadily over time. This is probably the main channel by which exchange rates affect inflation.

4.29  The fall in the price of imports relative to domestically produced goods and services (and of domestically produced goods with a large import content) will cause a partial substitution of imports for domestic goods and services, thereby reducing the demand for domestic goods and services. The decrease in the price of imports will also have a positive income effect on the demand for all goods and services. In practice, however, the combined substitution and income effects on total demand are likely to be fairly small.

4.30  In contrast, the supply-side effects coming through the higher price of exports may be quite powerful. Due to strong competition in export markets, in general there will be many substitutes for United Kingdom products. Hence, if exporters pass on the exchange rate change by a price rise, the demand for their product will be less, possibly much less. This would result in reduced output and employment and so lead to a fall in aggregate demand. And if exporters price to market, i.e. leave their export price unchanged, their sales revenue in sterling terms falls. With exports becoming less profitable this is likely to lead to a reduction in output once more. Thus the first round supply-side effect may lead to a strong second round demand effect.

4.31  While the benefits of inflation are largely short-term, the harm to exporters may be longer lasting. It has been found that foreign markets once lost can be difficult to regain even if, as expected, the real exchange rate and hence competitiveness, are restored in the long run. To set against this, to some extent the increase in the real exchange rate may simply trigger an over-due correction to an existing lack of competitiveness.

A quantitative assessment

4.32  The Bank of England does not provide any estimate of the relative importance of the different channels through which an interest rate change affects inflation. It only reports an estimate of the total effect. Mervyn King explained why:

"There is no operational reason for wanting to distinguish between the importance of different channels in the sense that when you change interest rates it is the overall impact you are looking at. There may or may not be an effect on the exchange rate, that will depend on circumstances, but you cannot undo it. You cannot not only half a transmission mechanism and not the other half." (Q 1622).

The discussion above suggests that the exchange rate channel will be very important, especially in the short run. An estimate of the relative importance of the different channels can be made from the quarterly macroeconometric model published in "Economic models at the Bank of England"[19]. As the Bank has not released a computer version of the model, and since the model parameters are not fully specified in the published report, the estimates can only be taken as a rough guide to what the full model would produce. (We are obliged to say we are not happy with this state of affairs. When we were promised publication of the suite of models, we expected them to be in such a form that experts could reproduce them and get them to work. In addition, a workable version on disc would be expected to follow fairly rapidly.)

4.33  Two types of simulation are made. Simulation (i) assumes that the short-term interest rate is increased by one percentage point for one quarter, and then returns to its previous level. This implies a temporary shock to the interest rate. Simulation (ii) assumes that the short-term interest rate is increased by one percentage point for one year and that markets fully anticipate this; it then returns to its original level. From the uncovered interest rate parity (UIP) condition, this implies that the exchange rate will instantly appreciate by 4 per cent, and then depreciate by 1 per cent each quarter until after a year it is restored to its original value following the return of interest rates to their original level, where markets assume they will stay indefinitely.

4.34  The results are reported in Table A. The total effect on RPIX inflation and the contributions due to the exchange rate and demand channels are given for each type of simulation. For both simulations the initial impact comes through the exchange rate channel. As time elapses the importance of the demand channel increases. After one year it exceeds that of the exchange rate channel, and within another two years becomes dominant.

4.35   These results go a long way to explain the concerns about the exchange rate expressed in their evidence by witnesses involved in the export sector. This sector bears much of the initial burden of anti-inflation policy. It is only after a year that the other sectors carry the larger burden.

4.36  Whatever the correctness of the Governor's argument that much of the recent strength of sterling is due to exogenous factors, and not United Kingdom monetary policy, the Bank's own model shows that interest rates have a substantial effect on the exchange rate, and this is the main channel in the short run by which monetary policy affects inflation.

4.37  We accept that what we have raised here is a preliminary attempt at clarifying a most complex subject. We hope others will develop this. In addition, it would surely be helpful to both Houses in fulfilling their role of accountability if a joint facility were brought into existence to run these models, and relate them to the practice of policy making.

Techniques of monetary policy

4.38  In our discussion of the background to current monetary policy we noted that the choice of the current techniques of control owed much to the failures of previous methods, and in our discussion of the transmission mechanism we touched on many of the difficulties of successfully implementing the new policy. We also mentioned some of the undesirable costs, such as those arising in the short run from the exchange rate being an important transmission channel. We now consider a number of other issues relating to the implementation of monetary policy using short-term interest rates. It should be pointed out that it is not clear what the best answers to many of these questions are. The MPC is in a learning process, and some of the problems have not yet arisen in practice.

The choice of interest rates as the monetary policy instrument

4.39  The Bank has chosen the repo rate as its instrument. How this works is explained clearly in the MPC's paper on the transmission mechanism. The Bank buys certain assets (mainly gilts, sterling Treasury Bills, United Kingdom foreign currency debt and eligible bank and local authority bills) from a small group of counterparties active in the money market (such as banks, securities dealers and building societies) who agree to buy them back in about a fortnight's time. The repo rate is the rate of interest implied by the difference between the sale and repurchase price. The aim is to provide short-term liquidity to the system, and replaces earlier ways of doing this in which eligible assets were more narrowly defined. It also provides a base-line lending rate to money markets from which other financial assets are priced. A change in the repo rate will rapidly affect other rates.

4.40  The Bank does not, however, have much control over the prices of other financial assets which are affected. In a world without capital controls, these are determined in the global market. It can, of course, choose at what point of the yield curve it wishes to issue bonds and this would provide some control over the United Kingdom term structure, and hence the cost of long-term debt. In practice, however, it has not chosen to do this.

The degree of independence of United Kingdom rates

4.41  The interdependence of world capital markets means that any control the Bank can exert over interest rates is likely to be small and temporary, especially at the long end. At the short end, this control is at the cost of other, possibly unwanted, consequences. Principal among these is the exchange rate. If United Kingdom rates are deliberately kept above foreign rates in order to control inflation then, as we have seen, the exchange rate will appreciate. This is, of course, what is intended and is an important part of the transmission mechanism. Equally, if rates elsewhere are raised then in general United Kingdom rates will need to be raised too to stop the exchange rate depreciating and causing inflation. Thus, contrary to what many people think, the Bank has limited discretion, even over the short rate.

The size, frequency and predictability of interest rate changes

4.42  Should the MPC make a small number of large changes to interest rates, or a large number of small changes in the same direction? And should the MPC signal future changes in advance, or give a hint about this as the Fed does? These are areas that are potentially important to policy making, but where not much is yet known. As in some other aspects, the MPC is still learning about the best approach to these issues.

4.43  In practice, the MPC has chosen to make relatively small changes in interest rates, often in the same direction over several months. On occasion individual members of the MPC have voted for larger increases, only to be out-voted. One explanation for this is a natural caution: by making small changes it might seem that a large error is not being committed. It also allows additional evidence to accrue and as they are less of a surprise to markets interest risk premia are kept lower. On the other hand, the Governor told this committee that each month the MPC changes its previous decisions only if there is new information. This would suggest that there is no attempt to smooth interest rate changes and each change is a "surprise" to markets. If this were literally true then interest rate changes would be equally likely to be up as down, which is clearly not the case. A more prosaic explanation for the policy of making a sequence of small changes in the same direction and not being highly active in moving interest rates one way and then another is that the public is likely to perceive this as the MPC changing its mind. If true, the MPC's caution is understandable. Nonetheless, there are clearly circumstances where a large change that lasts for a short time could be the right decision and the MPC should not shrink from this.

4.44  A move like this would be designed to have a big impact on expectations and hence bring about a more rapid and less costly adjustment of inflation. Mervyn King told this Committee "the question of the size of an interest rate change for us is relatively straightforward. It is a change in interest rates we need to make to get inflation back on track to hit the target. We do not sit down and say by how much do we need to change interest rates in order to hit the target and then a second question do we do three ten basis point changes over the next three months or one thirty point basis change." (Q 243). He added "We move by the amount that we think is justified having assessed the situation. We are quite prepared in the future, we have said this, to move by larger steps than we have moved so far or indeed by smaller ones." (Q 244). Another way in which the MPC may be able to help shape expectations is by talking up or down possible future interest changes by referring to a bias one way or the other as the Fed does. As noted in Appendix 4, which discusses uncovered interest rate parity, both current and expected future interest rate expectations affect the exchange rate, and the further in the future expectations can be influenced, the larger will be the impact on the exchange rate and the smaller the actual interest change need be. Of course, such guidance should appear credible otherwise it will soon cease to be effective. Our judgment is that it is too early to say that the MPC ought to anticipate policy by indicating what it is likely to do in future following on what it has just done.

Rules versus discretion

4.45  Monetary policy as currently carried out is based on the notion that the MPC exercises complete discretion in its choice of interest rates. Theoretically, an alternative would be to make interest rate decisions using pre-determined rules about how interest rates will be changed in response to economic events. (In that case we would not need an MPC!) The rule might, for example, take the form of specifying by how much interest rates will be changed in response to a deviation of inflation from target. A rule could also take into account other objectives such as the growth rate, unemployment and the exchange rate. The academic literature has spent much effort analysing this sort of monetary policy. The Taylor rule, named after John Taylor, an American economist, is the best known example. It says that interest rates will be increased by 1.5 percentage points for each percentage point that inflation exceeds its target and by 0.5 of a percentage point for each percentage point by which GDP growth exceeds its target. Even the Bank in some simulations of its macroeconometric model includes such a rule. To date no monetary authority has adopted rules in preference to discretion; so far it remains just a theoretical possibility.

4.46  One of the main advantages foreseen for rules is that they make monetary policy completely transparent and as predictable as one can get. If interest rate changes were predictable then markets would have anticipated them and already factored them in to decisions. The impact of the change when it was made would then be much smaller. In this case only genuine new information would have much effect. The main disadvantage of rules is the belief held by many that there would inevitably be occasions when the rule would need to be over-ridden and discretion used instead. There would also be much scope for differences of opinion about the relative importance of the weights to be given to the different objectives.

4.47  The theoretical literature generally shows that if the control of inflation is the ultimate objective of policy, then discretion is better than a rule that gives weight to other objectives. Having other objectives, such as an output target, tends to lead to a rate of inflation higher than the target. There also exists simulation evidence based on estimated macro-econometric models. In the report on its economic models, the Bank compares using rules with discretion based on evidence from simulations of its model.

4.48  Despite its theoretical advantages, at this stage the judgement must be that the case for rules is unproven, and the MPC is correct not to use them. It remains to be seen, however, how rule-like their actual behaviour turns out to be. There will soon be enough data to find out.

The role of inflation and other forecasts

4.49  The MPC stressed to the committee how much effort it takes to examine all relevant data. It spends a lot of its time studying data on current inflation and forecasts. It also looks at data on the real economy and financial data. And its coverage is of the whole economy and of regions and sectors. This information is published in the Inflation Report. Mervyn King told this committee that the forecasts of inflation are very helpful in keeping inflation on target because interest rate decisions can be taken earlier than if they waited for the current inflation figure to be published.

4.50  It is clear that the MPC does an admirable job in taking note of so much information. And it helps allay the concerns of many witnesses that the MPC is too narrowly focussed. What is less clear is whether and, if so, how the MPC uses these data to respond to the injunction in the Act to support the other government's objectives on growth and employment, or to make its interest rate choice.

The open letter and how to respond to inflation that is due to supply shocks

4.51  As has been noted above (for example, in paragraphs 4.21-4.23), this is one of the areas where there is the least clarity about how the MPC would react. In its defence the MPC can claim that there has been no instance of such a shock yet, at least one of any size, and there is no need to say what it would do in hypothetical situations. It is nonetheless appropriate for this report to comment on what might be done in such a situation.

4.52  A permanent supply shock, such as an increase in import prices due to a rise in world prices, would be passed on by importers. The more price inelastic the products involved—oil was highly price inelastic in the short run at the time of the 1973 oil price increase—the greater the would be the contractionary effect on the demand for other goods and services.

4.53  In the long run the adjustment to an oil price rise or similar shock has to make room for the increased expenditure on oil and the higher import bill. This involves a fall in the domestic demand for oil in response to the increase in its price relative to substitutes. It also involves a depreciation of the exchange rate. That is required to restore balance of payments equilibrium. It does so by making imports more expensive and stimulating exports. Hence, raising interest rates would send the exchange rate in the wrong direction. In short, by allowing inflation temporarily to exceed target, demand would be re-balanced faster by permitting the required depreciation of the exchange rate. This seems to us to be convincing. It is what the MPC should be proposing in its open letter to the Chancellor, if that were ever required.