Bulletin volume 71 no 2 - the changing transmission mechanism of new zealand monetary policy

The changing transmission mechanism of New Zealand monetary policyRishab Sethi1 This is the second of two Bulletin articles on the transmission mechanism of New Zealand monetary policy. In the first article (Drew and Sethi 2007), we described this mechanism, detailing the process by which changes in the Reserve Bank’s primary monetary policy instrument, the Official Cash Rate (OCR), eventually influence the general level of prices. This article examines how certain aspects of the transmission mechanism have changed over time. Assessing these changes is especially topical given that, in the estimation of some commentators, the most recent period of monetary tightening has witnessed policy that has been less effective at dampening inflation than previously. We briefly review the case for these claims and catalogue evidence from several sources to show that the overall impact of monetary policy on activity and inflation has not obviously weakened, and that some intermediate links in the mechanism may have, in fact, strengthened over the past decade.
to wholesale short-term interest rates” and “from effective In Drew and Sethi (2007), we described the process by mortgage rates to house prices”. In this article, the flow which changes in the Reserve Bank’s primary monetary chart is somewhat less comprehensive, featuring fewer links policy instrument, the Official Cash Rate (OCR), eventually because gaps in available research or data permit conclusions influence the general level of prices. In this article, our on only a subset of the intermediate links identified in our focus turns to assessing how this process, known as the transmission mechanism of monetary policy, may have In figure 1, links that are deemed to have become stronger changed over time. This is an especially topical exercise given over time are mapped in solid lines, weaker ones in dotted that the present economic cycle and monetary response lines, and those that have remained largely unchanged in have exhibited characteristics that have led to some concern dashed lines. Green lines continue to represent the interest about the efficacy of New Zealand monetary policy.
rate channel of the transmission mechanism, blue lines Specifically, we look at several intermediate links in the denote the effect of changes in the exchange rate, and red mechanism, assessing possible changes in either the lines refer to effects related to inflation expectations. The strength of response of one variable to changes in another, numbers next to the links provide easy reference to points in or in the timing with which this response occurs. Consistent with the approach in Drew and Sethi (2007), we organise For each of the links considered, the evidence presented the discussion in this article around a stylised representation in this article is based on either (a) available academic or of the transmission mechanism (Figure 1). In Drew and Sethi internal Reserve Bank research, or (b) the three economic (2007), we identified 17 intermediate links in a detailed models used to study the transmission mechanism in Drew representation of the mechanism, such as “from the OCR et al. (2008). These models are summarised in Box 1.2 Note that the various analyses differ in their data definitions and This article is based on Drew et al. (2008), a Reserve Bank of New Zealand discussion paper, to which sample periods, and an element of judgement is necessary interested readers are referred for the econometric detail when aggregating this evidence. Consequently, we are underlying the results presented in this article. The research was initially prepared for a conference on the conservative in our conclusions here, defaulting to the view business cycle, housing, and the role of policy hosted by The Treasury and the Reserve Bank in December 2007. Drew and Buckle (2008) summarise the main themes from Though the transmission mechanism of New Zealand this conference. I thank Tim Hampton, John McDermott, monetary policy has been widely studied using a range and Tim Ng for valuable comments on an earlier draft of of economic models, there is a paucity of prior research that addresses changes in the mechanism over time.
Reserve Bank of New Zealand: Bulletin, Vol. 71, No. 2, June 2008 Figure 1Changes in the transmission mechanism of New Zealand monetary policy(elements not examined in this article are not shown) that a link is broadly unchanged over time, unless the bulk To set the context, we briefly review the transmission mechanism in section 2, together with an overview of the In this section, we review the transmission mechanism major public causes for concern about the efficacy of recent using the example of a monetary tightening (an increase monetary policy. In section 3, we present evidence on the in the OCR) that is aimed at dampening projected future changing influence of monetary policy on market wholesale inflationary pressure, as in Drew and Sethi (2007). Opposing and retail interest rates, and on the exchange rate. Section 4 effects may generally be expected in the event of a decrease looks at the impact of these financial prices on real activity. in the OCR, though there is likely to be significant variation Section 5 documents changes in the relationship between aggregate activity and inflation, a link commonly known as An unanticipated increase in the OCR tends to result in increases in other wholesale and retail interest rates for both short- and long-term maturities. These interest rate increases reduce the present values of income streams from a variety of assets such as bonds, equities and real estate, Reserve Bank of New Zealand: Bulletin, Vol. 71, No. 2, June 2008 and lead to lower prices for these assets. Debt servicing These three channels – interest rates, exchange rates costs and rewards to saving increase on the back of higher and inflation expectations – comprise the transmission interest rates, and combine with lower asset values to limit mechanism of monetary policy. Drew and Sethi (2007) detail credit available to households and firms. Higher interest the role played by several intermediate variables along each rates, being an increase in the ‘price’ of using money now rather than later, effectively increase the current price of any dollar-denominated expenditure relative to its future cost. Consequently, households and firms face incentives Major features of the current business cycle, to postpone current consumption and investment, reducing current aggregate demand for goods and services. Finally, Since emerging from a brief recession in the late 1990s, if this new level of aggregate demand declines relative to New Zealand has enjoyed both the longest and strongest the economy’s supply capacities, inflation pressures may be uninterrupted expansion in aggregate economic activity in the post-war period. In recent years, the strong growth has been accompanied by an unemployment rate low by historical The second major channel for the transmission of monetary and international standards, inflationary pressures stemming policy is through the exchange rate. An unanticipated from private domestic demand and capacity constraints, increase in the OCR immediately appreciates the New Zealand and increasing rates of public and private investment. There dollar, as higher domestic interest rates attract foreign have been large increases in asset prices, especially for real capital. In theory, the foreign currency price of New Zealand estate between 2001 and 2007. These domestic drivers of dollars should be bid-up to a level such that the expected growth have been supported by international factors such depreciation from that point on is just sufficient to leave an as increasing commodity export prices, low costs of capital, investor indifferent between holding assets denominated in and, until relatively recently, a substantial and increasing New Zealand dollars and in the foreign currency. appetite for risk. The primary response of monetary policy to A higher exchange rate implies a lower New Zealand dollar elevated inflationary pressures is seen in an OCR that is now, price for foreign-produced goods and services. These lower at 8.25 percent, 3.25 percent higher than it was at the start prices are normally passed on through the supply chain, lowering prices of both intermediate goods and final goods Large macroeconomic imbalances have accompanied the such as those measured in the Consumers Price Index (CPI). expansion in the form of record current account deficits, A higher exchange rate also renders export-oriented and substantial declines in the household saving rate, an exchange import-competing firms less competitive, reducing their rate that has been widely considered to be exceptionally and earnings and again dampening overall domestic activity and unjustifiably overvalued, and a concentration of growth in sectors of the economy relatively sheltered from international The final channel through which monetary policy influences prices is through inflation expectations. If households and These patterns in New Zealand’s recent economic firms are convinced of a central bank’s commitment to an performance have prompted public concern about the role inflation target, and of its ability to meet this target, then they are likely to respond to a change in monetary policy Some authors have described a fourth channel in the by adjusting their own expectations of future activity and transmission mechanism relating to credit creation inflation. On anticipating changed economic conditions by banks and financial intermediaries in response to changes in monetary policy settings. See Bernanke et al. in the future, they are likely to modify current levels of 4 Note that real interest rates, measured as the excess consumption and investment, and their strategies for setting of the overnight interbank cash rate over annual CPI inflation, ranged between 3.6 and 3.9 percent between March 2004 and September 2006, and peaked at 6.4 percent in September 2007.
Reserve Bank of New Zealand: Bulletin, Vol. 71, No. 2, June 2008 Table 1The changing transmission mechanism: a summary of findings comparing the strength and timing of intermediate links in the 2000s with the 1990s. From monetary policy to short-term wholesale rates From monetary policy to long-term wholesale rates From monetary policy to effective mortgage rates From monetary policy to the exchange rate of monetary policy and its recent conduct. Our study of products, owing both to lower longer-term wholesale rates changes in the transmission mechanism specifically addresses and to margin compression amongst mortgage lenders on two of these concerns. First, changes in the OCR are thought fixed mortgage rates. Some commentary also suggests the to have provoked disproportionately large changes in the Reserve Bank was unable to convince market participants exchange rate in recent years, a claim based principally on of the underlying strength of inflation pressures in the the low levels of international risk aversion that have been early part of the current tightening cycle, and consequently observed since 2002. There is a risk premium associated changes in the OCR had less influence on longer-term rates with investments in NZD-denominated assets, representing the additional payment required to compensate investors for The remainder of this article presents evidence to show that possible future exchange rate depreciations.5 With reduced the overall impact of policy on activity and inflation has not aversion to bearing risk, investors pay less regard to the obviously weakened, and in some ways, has strengthened possibility of future depreciations, and face added incentives over the past decade relative to the 1990s. Table 1 presents to engage in ‘carry trades’ wherein they borrow in currencies a summary of results discussed below.
with low financing costs to purchase NZD-denominated There is one important caveat to the exercise: the changing transmission mechanism is likely to be due not just to the Some observers have raised another important concern changing impact of one intermediate variable on another, about the recent effects of monetary policy. They perceive but also to the changing nature of the random fluctuations that the transmission of monetary policy from changes in (or shocks) that buffet an economy over time.7 It can be the OCR to changes in domestic activity and inflation has difficult to perfectly account for the relative contributions weakened as a whole. Several reasons have been proposed to changes in the mechanism from these two causes. in support of this conjecture. For example, low international interest rates and increasing capital markets integration The other major concern about the recent impact of monetary policy is not addressed in this article. This may each have limited the ability of the Reserve Bank to relates to the observation that, in a cycle where the bulk independently influence longer-term domestic interest of inflationary pressure has been sourced in the domestic economy, monetary policy has been poorly targeted, with rates. Also, there has been a growing substitution away sectors exposed to the exchange rate bearing the brunt of the burden of reducing inflation.
from floating rate mortgages to longer-term fixed rate For example, some economists attribute the ‘Great Moderation’ – the period of relatively high growth Major reasons for this currency risk premium include: and low inflation in the US since the mid-1980s – to a the relatively small and relatively undiversified nature sequence of beneficial supply shocks (such as improved of the New Zealand economy, and the persistent productivity from increased use of computers or the rise macroeconomic imbalances and low national saving rate of China as an efficient manufacturing base) rather than Reserve Bank of New Zealand: Bulletin, Vol. 71, No. 2, June 2008 inflation, output, interest and exchange rates, and others The economic models in Drew et al. such as indices of climatic conditions.
The third model considered in Drew et al. is small, featuring only five variables. This ‘New Keynesian’ model Many of the conclusions in this paper are based on the is a variant of a type widely used in academic research findings of Drew et al. (2008), a recent Reserve Bank for assessing monetary policy, as it captures many features Discussion Paper. Here, we summarise the main features of the macroeconomic data in a reasonably parsimonious of the three economic models used by the authors to study manner. The relationships between variables are motivated by general equilibrium (or whole-economy) theory, and are The Reserve Bank’s Forecasting and Policy System (FPS) has internally consistent, allowing the authors to form tentative been used as the primary tool for producing the Bank’s conclusions about why the transmission mechanism may published forecasts since 1997. It features a dynamic have changed over time. New Keynesian models can adjustment process that determines how inflation and feature one or more different sources of inflexibility that other variables return to their long-run equilibrium values cause variables to adjust slowly in response to an economic following an economic shock, and is calibrated to match shock, and motivate an important role for expectations of certain characteristics of the New Zealand business cycle.8 future inflation and output in determining the dynamic In turn, the long-run equilibrium is based on a framework of utility-maximising consumers and profit-maximising firms. Although the underlying structure is largely unchanged since the inception of FPS, the dynamic adjustment paths have been heavily adjusted over time. These adjustments, The three models in Drew et al. use different methods to or recalibrations of the model, are prompted by evidence assess changes in the transmission mechanism. Results from both sectoral and aggregate-level research, and by from FPS rely on comparing properties of the different judgement on changes in economic structure, and on the vintages of the model from 1997, 2002 and 2007. In propagation of economic shocks. As such, the comparison contrast, the VAR and New Keynesian models are estimated of different calibrations of FPS over time offers a very on New Zealand data, and the changing intensities of convenient synthesis of the Reserve Bank’s changing views the relationships between variables are captured in the on various macroeconomic relationships.
changing magnitudes of the estimated parameters in the The second model used to study the transmission model equations. Specifically, the VAR uses the method of mechanism in Drew et al. is a vector auto-regression, or rolling regressions wherein the model is initially estimated VAR. The defining characteristic of a standard VAR is that it on data from 1989 (September quarter) to 1996 (June presupposes little formal theory for specifying relationships quarter). Data from additional quarters until the end of the between variables. Instead, it is assumed that the future available sample are successively added, and the model is re- path for a variable depends simply on its own history and estimated on each successively larger sub-sample. Though that of other variables in the model.9 The VAR in Drew rolling regressions yield continuous estimates of changes et al. features core macroeconomic variables such as in economic relationships, these estimates can be volatile, as they are sensitive to individual observations that may or may not be in a given sub-sample. More importantly, the See Black et al. (1997) for detail on the structure of the FPS.
method is backward-looking in that it relies solely on past The precise assumed dependencies between variables will vary from one VAR model to another. Some observations and ignores the information content of future extensions, known as structural VARs, motivate data in estimating a model at a given point in time. contemporaneous relationships from economic theory. See Hamilton (1994) for more on these models.
Reserve Bank of New Zealand: Bulletin, Vol. 71, No. 2, June 2008 A technique known as the Kalman filter provides an sample of data, and is relatively insensitive to individual elegant solution to the problem of estimating relationships observations.10 Drew et al. use this method to estimate that may change over time. It yields smooth estimates of parameter changes, makes use of the full available Accordingly, we base our conclusions on evidence from multiple sources, including from economic models that make different assumptions about the structure of the New Zealand economy and the nature of these shocks. the period 1990 to 2005. He focuses on a sub-sample from 1998 to 2005 to determine whether monetary policy and its effects display similar characteristics in this period as in the full sample. In response to a change in monetary policy, Schmidt-Hebbel finds that though monetary policy has significant bearing on 10-year government bond rates in both the full- and the sub-sample, the differences between Following an actual or expected change in the OCR, financial asset prices, such as interest and exchange rates, are first to respond. It is important that these prices respond suitably Recent ‘event analysis’ at the RBNZ suggests that since the as they, in turn, affect other asset prices, and real activity start of the decade, changes in the OCR have had a significant and inflation at later stages in the monetary transmission impact on wholesale interest rates.11 A hypothetical monetary process. As a recent example, rapidly increasing house prices policy surprise of 1 percent is estimated to prompt one-year contributed to strong growth in household consumption in forward swap rates to increase by 0.9 percent for settlement the five years to the end of 2007, and were considered to one year ahead and by 0.3 percent for settlement five years be important contributors to inflationary pressure over the period. As such, in a series of Monetary Policy Statements 10 Originally devised for applications such as tracking beginning in December 2005, the Reserve Bank was spacecraft in the 1960s, the Kalman filter is widely used increasingly pointed in noting that a slowdown in house in engineering, aeronautics and communications. Since the mid-1990s, it has also become an essential tool in price growth was needed to weaken the case for further economic research for retrieving information from noisy or even partially missing data, for estimating economic interest rate increases. In this section, we look at whether models, and for smoothing time series by incorporating information from past as well as future data.
the effect of OCR increases on wholesale and mortgage 11 See Drew and Karagedikli (2008) and Karagedikli interest rates was smaller over this period than previously and Siklos (2008). An event analysis, which attempts to isolate the impact of the surprise component of an observed, and so failed to contribute to an easing of the economic event, is thought to be a more robust way of identifying the impact of monetary policy and other rapid increase in house prices in a timely fashion. We also economic developments on financial market prices than consider the changing impact of monetary policy on the the traditional time-series approaches above, and has become an increasingly popular tool in international 12 Swaps are financial instruments that allow counterparties to manage risk, usually by achieving a better match between their assets and liabilities. A forward swap is an instrument that is negotiated at the present juncture, for settlement at a point in the future, and which matures several periods further into the future. The interest rates on these instruments are benchmark wholesale interest rates.
Reserve Bank of New Zealand: Bulletin, Vol. 71, No. 2, June 2008 ahead.12 Event analysis studies require very high frequency using weekly data, finding that both floating and fixed data; the required intra-day daily data used in the Bank’s mortgage interest rates adjust fully in the long run, but that research is only available from around 2001. However, in formal robustness checks, the authors of these studies do Between 2005 and 2007, the share of outstanding mortgages not find evidence of instabilities in their regressions, as on floating interest rate terms declined from 42 percent to might have been the case if the impact of monetary policy 30 percent. Meanwhile, the value of outstanding mortgages on contracts fixed for two years or more increased from For additional perspective over a longer period, we estimate 11 percent to 32 percent. There seems to be little doubt some simple regressions linking changes in monetary policy that the present cycle witnessed a change in preferences and short- and long-term wholesale interest rates. As these for mortgage borrowing towards the fixed-rate contracts, regressions are estimated over a long sample from 1992, we causing the effective mortgage rate to increase more slowly use changes in interest rates on 90-day bank bills as a proxy for changes in monetary policy settings.13 Generally, we The overall assessment from this evidence is ambiguous: it find that volatility in wholesale interest rates has decreased seems that the response of wholesale and mortgage interest since about 2000, and that there has been no systematic rates is more or less unchanged in recent years when decline in the total impact of changes in monetary policy compared to the 1990s, but that this response occurs more on these interest rates. However, comparing estimates from the past five years with those from 1992 to 2002, it appears that more distant changes in monetary policy settings have exercised greater influence on long-term wholesale interest rates relative to more recent changes. In other words, Drew et al. (2008) study changes in the transmission changes in policy settings now take a little longer to flow mechanism using FPS, the Reserve Bank’s core through to these wholesale interest rates.
model for producing macroeconomic forecasts. As Liu et al. (2007) estimate the degree and speed of witnessed by changing FPS calibrations, Drew et al. report response of retail interest rates to changes in policy- that the Reserve Bank perceives the exchange rate response controlled rates between 1994 and 2004. They find that, to a change in monetary policy to have been stronger and following the introduction of the OCR in 1999, floating more persistent in 2007 than in 2002 or 1997. This is seen interest rates on mortgages and deposits respond more in figure 2, which documents impulse response functions strongly to changes in monetary policy, while fixed interest from the three vintages of FPS – that is, the figure maps the rates do so more slowly and by less. Tripe et al. (2005) also changing response of the exchange rate to a unit increase examine the impact of the introduction of the OCR, and (an impulse) in monetary-policy-controlled interest rates. reach a similar conclusion that the long-term impact of Two smaller models considered by Drew et al. offer wholesale rates on floating mortgage rates increased in the conflicting evidence on changes in this relationship, with post-OCR period, while the impact on fixed mortgage rates the VAR finding support for an unchanged link, and the decreased slightly. Pais (2007) reaches different conclusions New Keynesian model suggesting that the impact of a policy tightening on the exchange rate declined over the 13 The dependent variable in the regressions is the monthly 1990s, but has increased since 2002.14 Exchange rate change in the interest rate on one-year forward swaps for settlement one year ahead (a proxy for short- behaviour is notoriously difficult to pin down accurately, term wholesale interest rates), and the comparably transformed interest rate on one-year forward swaps for and we are inclined to favour evidence from the changing settlement five years ahead (as a proxy for long-term) wholesale interest rates. The regressions are estimated 14 Results from the events analysis research cited earlier using a Kalman filter with time-varying parameters, and indicate that an (unanticipated) 1 percent increase in simple diagnostics indicate that the residuals are well the OCR leads to a 3.5 percent increase in the NZD-USD Reserve Bank of New Zealand: Bulletin, Vol. 71, No. 2, June 2008 FPS calibrations, which aggregates information from a wide some disagreement on changes in this relationship, with range of internal Bank research and from staff judgement as the VAR suggesting that the response of the output gap is mildly weaker, and the New Keynesian model favouring the interpretation that the impact of interest rates on output growth is significant across the entire sample period According to FPS, the response of the exchange (between 1992 and 2007), but strongest towards the end rate to changes in monetary-policy-controlled both in terms of the initial response and the persistence of interest rates has increased over time.
According to FPS, the impact of interest rates on output has become somewhat stronger and a As wholesale and retail interest rates increase in response to tighter monetary policy, they prompt households and Comparing FPS calibrations over time, Drew et al. firms into reassessing their consumption and investment report that a temporary exchange rate appreciation decisions, reducing aggregate demand. In this section, we was thought to have had a large and rapid impact on look at how links from the exchange rate to real activity and export and import demand in 1997, causing overall demand from wholesale and retail interest rates to real activity have pressures to decline quickly too. In the 2002 vintage of FPS, these effects are muted and delayed, becoming somewhat Studying the characteristics of a medium-sized stronger again, though further delayed, in 2007 (as seen in model of the New Zealand economy, Schmidt- figure 4). On the other hand, estimates from the two smaller Hebbel (2006) finds that the output gap – a measure of models suggest that the output response to changes in excess demand – declines in response to higher short-term the exchange rate has been small across the entire sample. interest rates. Interestingly, this decline is small and not Though the effect is estimated to be small by both models, statistically different from zero when the model is estimated they do differ in that the VAR model indicates that a stronger on a sample from 1992 to 2005, but does become larger exchange rate actually generates a very mild positive impact and statistically significant when the sample is restricted to in output from about 2001, while the New Keynesian model reports a more conventional negative response, and one that becomes only marginally stronger over time. From figure 3, we see that successive recalibrations of FPS have increased the restraint that interest rates exercise on aggregate demand, with much of the increase coming from greater sensitivity of business investment to interest 15 According to this model, a 1 percent increase in the real rates. The two smaller models in Drew et al. are again in interest rate now reduces the output gap by around 0.4 percent in the long-term.
Reserve Bank of New Zealand: Bulletin, Vol. 71, No. 2, June 2008 FPS suggests output now responds more slowly FPS suggests pass-through of the exchange rate to changes in the exchange rate than previously. to CPI inflation declined between 1997 and The effect is small and is mostly unchanged in 2002, remaining largely unchanged since.
Figure 6 documents some estimates of long-run responses, and essentially captures the changing magnitude of the The Reserve Bank’s view of the inflation process can be coefficients in the Phillips curve. These long-run responses represented by a Phillips curve – inflation arises when are calculated conditional on the assumption that there is no aggregate output increases beyond an economy’s supply change in monetary policy settings that aims to offset these capacities. Modern versions of the Phillips curve also effects. We see that the New Keynesian model also reports postulate an important role for expectations of future lower overall pass-through for the most part, except in very inflation in determining current inflation.16 More elaborate recent years when it finds for increased pass-through (panel formulations include dependencies on exchange rates, oil (b) of figure 6). According to this model, in the absence of and commodity prices, and trading partner inflation.
a monetary policy response, a 10 percent depreciation in FPS recalibrations suggest that the total impact the exchange rate today eventually leads to a long-term of the exchange rate on CPI inflation, also called increase in inflation of around 0.6 percent, compared with exchange rate pass-through, declined between 1997 and an increase of 1.2 percent in the early 1990s.17 2002, and has remained largely unchanged since. This total Another major conclusion that we can draw from impact can be decomposed into two sub-effects: the impact the New Keynesian model in Drew et al. is that of the exchange rate on import prices and the subsequent inflation became increasingly forward-looking over the pass-through from import prices to CPI inflation. Comparing sample period. The decline in inflation persistence was steady the responses of CPI and non-tradable inflation from FPS, over the 1990s and has levelled off since 2001. Overall, it appears that the latter effect has been dominant. For these results are fairly consistent with FPS recalibrations, example, the CPI inflation responses are fairly similar in the and, all else equal, imply an increase in the effectiveness of 2002 and 2007 vintages of FPS, yet non-tradable inflation troughs lower in the 2007 vintage. Hence, the second sub- 17 Researchers report the stylised facts that countries with credible anti-inflationary policies experience lower pass- effect, from import prices to CPI inflation, is likely to have through to CPI inflation, that pass-through is associated been the dominant contributor to the overall decline in pass- with lower inflation expectations, and that it increases with volatility in inflation and exchange rates (see Campa and Goldberg (2002) and Engel (2002)). Possible microeconomic reasons for declining pass-through 16 Given that true expectations cannot be observed, modern include: a change toward invoicing trade in goods and Phillips curves can be cast in several different forms. services in New Zealand dollars; increased shares of Some include past inflation as a proxy for expected distribution costs in the price of imports, which provide inflation, others include survey measures and many for greater opportunities for margin compression; a macroeconomic models assume that households and changing mix of goods and services imported into New firms form rational expectations that fully incorporate Zealand over time; and increasing use of sophisticated Reserve Bank of New Zealand: Bulletin, Vol. 71, No. 2, June 2008 Some concern has been raised about the effectiveness of recent monetary policy in relation to its influence on the exchange rate, and with respect to the overall impact on activity and inflation. On the basis of evidence presented in this article, especially that from the analysis of three different models of the New Zealand economy, we find that there has been no discernible change in the overall influence of monetary policy on inflation in the present economic cycle. However, we find some evidence of changes over time in the strength and timing of the various relationships that are intermediate in the transmission of the OCR to inflation. First, changes in monetary-policy-linked interest rates are reflected in wholesale interest rates to at least the same degree as in the 1990s, though perhaps with increased delay. Second, even though the exchange rate is estimated to respond more to changes in monetary policy settings, Note: In the New Keynesian model in Drew et al., inflation is specified as a function of the variables in this figure, it is also estimated to have become a somewhat less and we see the changing long-run impact on CPI inflation (assuming that there is no monetary policy response) of (a) consequential determinant of output and inflation in itself. the output gap, (b) the exchange rate, (c) trading partner inflation, (d) oil prices, and (e) commodity prices. Panel In other words, much larger changes in the exchange rate (f) shows the coefficient on lagged inflation.
are now required to effect a given change in output and inflation, and it appears that monetary policy settings are The model also indicates that the relationship indeed prompting these larger changes in the exchange rate. between excess demand and inflation became a Meanwhile, the inflation process has become more forward- little stronger over the 1990s and has stabilised since 2002. looking, with expectations of future inflation becoming Assuming no further response from monetary policy, a 1 more important in determining current inflation.
percent decrease in the output gap today eventually leads In summary, we conclude that there has been some re- to a decrease in inflation of around 0.35 percent. The rise weighting in the relative burden of macroeconomic in this long-term effect is due to an increasing coefficient adjustment borne by the different channels of the on the output gap in the Phillips curve, which has been transmission mechanism. Transmission through interest tempered somewhat by the decline in inflation persistence.
rates remains vital and effective, but eventuates with In summary, the most important development in the Phillips longer lags. The exchange rate channel has become more curve over the past 15 years appears to be the increasing important over time: the exchange rate itself responds to importance of expectations in determining current inflation. a greater degree but, for a given unit of change, generates To translate this observation into a well-known ‘lesson’ for smaller responses in output and inflation than previously. monetary policy: though the increasingly dominant role of Overall, the transmission mechanism remains effective, expectations means that policy can accomplish more with and monetary policy continues to eventually influence the smaller changes to interest rates, it remains crucial to ensure that these expectations remain well anchored to the inflation Reserve Bank of New Zealand: Bulletin, Vol. 71, No. 2, June 2008 Pais, A (2007) “The transmission of interest rate changes in Bernanke, B, M Gertler and S Gilchrist (1999) “The financial the New Zealand economy”, Banks and Bank Systems, 2(3).
accelerator in a quantitative business cycle framework”, Tripe, D, C J McDermott and B Petro (2005) “A test of in Handbook of Macroeconomics, eds J Taylor and M the response to a monetary policy regime change in New Woodford, Vol. 1(C), 1341-1394.
Zealand”, Contemporary Studies in Economics and Financial Black, R, V Cassino, A Drew, E Hansen, A Scott and D Rose (1997) “The forecasting and policy system: The core model”, Schmidt-Hebbel, K (2006) “New Zealand”s monetary Reserve Bank of New Zealand Research Paper, RP43.
and exchange rate policy in international comparison”, in Campa, J M and L Goldberg (2002) “Exchange rate pass- Testing Stabilisation Policy Limits in a Small Open Economy: through into import prices: a micro or macro phenomenon,” Proceedings from a Macroeconomic Policy Forum, eds B National Bureau of Economic Research (NBER) Working Buckle and A Drew, Reserve Bank of New Zealand and the Drew, A and R Buckle (2008) “The business cycle, housing, and the role of policy: Summary of a recent conference held by The Treasury and the Reserve Bank of New Zealand”, Reserve Bank of New Zealand Bulletin¸ 71(1), 40-44.
Drew, A and Ö Karagedikli (2008) “Some benefits of monetary policy transparency in New Zealand”, Reserve Bank of New Zealand Discussion Paper, 2008/01.
Drew, A and R Sethi (2007) “The transmission mechanism of New Zealand monetary policy”, Reserve Bank of New Zealand Bulletin, 70(2), 5-19.
Drew, A, Ö Karagedikli, R Sethi, and C Smith (2008), “Changes in the transmission mechanism of monetary policy in New Zealand”, Reserve Bank of New Zealand Discussion Engel, C (2002) “The responsiveness of consumer prices to exchange rates: A synthesis of some new open-economy macro models,” The Manchester School Supplement.
Hamilton, J (1994), Time Series Analysis, Princeton University Karagedikli, Ö and P Siklos (2008) “Explaining movements in the New Zealand dollar: Central bank communication and the surprise element in monetary policy?”, Reserve Bank of New Zealand Discussion Paper, 2008/02.
Liu, M-H, D Margaritis and A Tourani-Rad (2008) “Monetary policy transparency and the pass-through of interest rates,” Journal of Banking and Finance, 32, 501-11.
Reserve Bank of New Zealand: Bulletin, Vol. 71, No. 2, June 2008

Source: http://reservebank.govt.nz/research_and_publications/reserve_bank_bulletin/2008/2008jun71_2sethi.pdf