Abstract
The real interest rate is a very important variable in the transmission of monetary policy. It features in vast majority of financial and macroeconomic models. Though the theoretical importance of the real interest rate has generated a sizable literature that examines its long-run properties, surprisingly, there does not exist any study that delves into this issue for South Africa. Given this, using quarterly data (1960:Q2-2010:Q4) for South Africa, our paper endeavors to analyze the long-run properties of the ex post real rate by using tests of unit root, cointegration, fractional integration and structural breaks. In addition, we also analyze whether monetary shocks contribute to fluctuations in the real interest rate based on test of structural breaks of the rate of inflation, as well as, Bayesian change point analysis. Based on the tests conducted, we conclude that the South African EPPR can be best viewed as a very persistent but ultimately mean-reverting process. Also, the persistence in the real interest rate can be tentatively considered as a monetary phenomenon.
Similar content being viewed by others
Notes
See Neely and Rapach (2008) for a detailed literature review.
Studies that exists for South Africa only deals with the uncovered interest rate parity condition, or in other words, the interest rate behavior of South Africa relative to other developed or emerging economies. See for example Kahn and Farrell (2002), Kryshko (2006), Lacerda et al. (2010) and de Bruyn et al. (2011).
In February of 2000, the Minister of Finance announced that inflation targeting would be the sole objective of the South African Reserve Bank. Currently, the Reserve Bank’s main monetary policy objective is to maintain CPI inflation between a target-band of 3–6 %, using discretionary changes in the repo rate as its main policy instrument.
See Erdman and Emerson (2007) for a detailed literature review in this regard.
See Sect. 3 for further details.
The steady state real interest rate could also be affected by changes in distortionary tax rates (See Blanchard and Fischer 1989, pp. 56–59).
Although much of the empirical literature analyzes EPRR in this manner, the time-series properties of EPRR can differ from those of the EARR, mainly because of two reasons: First, at short-horizons, the behavior of EPRR could differ from that of the EARR, and; second, some estimation techniques can generate different persistence properties between the EARR and EPRR. The reader is referred to Neely and Rapach (2008) for further details.
This can be found in Neely and Rapach (2008).
The annual population figures are interpolated to quarterly values.
Specifically as far as the data is concerned, the 3-month Treasury constant maturity rate monthly data are converted to quarterly frequency by averaging over the three months comprising a quarter. The annualized CPI inflation rate is based on the seasonally adjusted (at an annual rate) CPI with the base year of 2005. Ex post real interest rate is defined as the three-month Treasury bill rate minus the realized inflation rate in the subsequent quarter. Finally, the annualized consumption growth rate is based on real personal consumption expenditures (base year of 2005) adjusted seasonally at an annual rate. Annualized inflation and consumption growth rates are computed by taking 400 times the first differences of the natural logs of the CPI deflator and consumption.
Given this, we also conducted the Phillips and Perron (1988) [PP], Elliot et al. (1996) [ERS] point-optimal test and the Kwiatkowski et al. (1992) [KPSS] unit root tests on the treasury bill rate. The PP test could not reject the null of unit root even at the 10 % level, while the KPSS test rejected the null of stationarity at 1 % level of significance. However, the ERS point-optimal test rejected the null of unit root even at 1 % level. In addition, we also tested the EPRR based on the PP, ERS point-optimal test and KPSS test. The null of unit root was rejected even at 1 % based on the PP test, while the null of stationarity could not be rejected at 5 % based on the KPSS, but the ERS could not reject the null of unit root at even 10 %. Further, nonlinear unit root test proposed by Kapetanios et al. (2003) [KSS] and the Bayesian unit root test developed by Sims (1988) was also used. Both the KSS and the Bayesian tests rejected the null of unit root. These results are available upon request from the authors.
Following Gregory and Hansen (1996), we also tested for cointegration using residual based tests that allows for regime shifts, but the null of no cointegration could not be rejected even at 10 % level of significance. Similar conclusions were also reached based on both the Bierens (1997) and Breitung (2002) nonparametric cointegration tests, as these tests too failed to reject the null of no cointegration. Note that both these tests allow for nonlinearity of an unknown form in the short-run dynamics of the the two variables. Interestingly, even though no cointegration could be detected, estimates of θ 1 based on the dynamic ordinary least squares (DOLS) or the Johansen (1991) methods are not significantly different from unity. However, this result, perhaps, explains as to why we detect stationarity for the EPRR based on the unit root tests with a prespecified cointegrating vector of (1, −1)\(^{\prime}\). Further, using the nonparametric nonlinear co-trending analysis developed by Bierens (2000), the null hypothesis that there exists one co-trending vector between the nominal Treasury bill and the inflation rate could not be rejected. These results are available upon request from the authors.
Until recently, researchers used models of cointegration that assumed both the cointegrating relationship and short-run dynamics to be linear. But now studies have started to relax these linearity assumptions based on nonlinear cointegration or threshold dynamics, which in turn, allow the cointegrating relationship and mean reversion to be contingent on the current values of the variables. Given this, when we tested for nonlinear cointegration between the 3-month Treasury bill rate and the the inflation rate based on the test proposed by Li and Lee (2010), the null of no cointegration could not be rejected even at the 10 % level of significance. This result is available upon request from the authors. It must be pointed out that, although evidence of threshold behavior could be interesting, these models do not obviate the persistence in the EPRR, since there are still regimes where the real interest rate behaves like an unit root process.
Similar results were obtained for the fractional integration parameter of EPRR using the Geweke and Porter-Hudak (1983) and the Robinson (1992) methods, though the latter estimation strategy produced a relatively lower value of d. In addition, when we used the modified R/S statistic developed by Lo (1991), we confirmed the EPRR to be a long-memory process. These results are available upon request from the authors.
This short-memory behavior of the per capita consumption growth rate was also confirmed by the Geweke and Porter-Hudak (1983) and the Robinson (1992) estimation of d, as well as by the modified R/S statistic which failed to reject the null hypothesis of the per capita consumption growth being a short-memory process. These results are available upon request from the authors.
Note that in Table 3, we see that F(2|1) is rejected meaning that we reject 1 break in favor of 2 breaks. Also, F(3|2) is not significant, implying that we cannot reject 2 breaks in favor of 3 breaks. These results suggest 2 breaks. However, the 10 % confidence interval around the endpoint of the first break, identified at 1974:1, so was so wide that it included the starting point of the sample. In light of this, we decided to choose one break rather than 2 for the EPRR. Further, if we allowed for two breaks, suggesting three regimes, the estimate of the mean EPRR (0.89) under the first regime was insignificant with a t-statistic of 1.42. The details of these results are available upon request from the authors.
In contrast to this evidence of a unique break in the EPRR, the Bai and Perron (1998) methodology obtained no structural break in the mean of per capita consumption growth. These results are available upon request from the authors.
Note that, when we conducted the Zivot and Andrews (1992) unit root test allowing for a break in the intercept, it rejected the null of unit root at the 10 % level of significance for the EPRR. This test identified a break for the series at 1987:Q4. In addition, the Lee and Strazicich (2003) unit root test allowing for 2 breaks, also rejected the null hypothesis of unit root test, with the breaks identified in 1972:Q3 and 1999:Q3.
Refer to Ludi and Ground (2006) for a detail discussion of the history of monetary policy in South Africa.
Using a measure of monetary policy surprise developed by Reid (2009) and Gupta and Reid (forthcoming), we analyzed the impulse response function of the EPRR obtained from an autoregressive distributed lag (ARDL) model following a contractionary monetary policy shock over the period of 2002:Q2-2010:Q4. The results, based on Romer and Romer’s (2004) Monte Carlo methods to generate confidence bands for the impulse response functions, indicated a delayed very short-lived (1 quarter) positive effect on the EPRR, suggesting again a weak monetary explanation of the persistence of the real rate—a result in line with the findings of Amusa et al. (forthcoming). The impulse response functions from the ARDL model is available upon request from the authors. Note, Reid (2009) and Gupta and Reid (forthcoming) used the change in the 3-months Banker’s Acceptance rate on the day after the monetary policy committee makes its statement as a proxy for the surprise component of monetary policy. This event-based data was converted to a quarterly frequency by averaging the shocks over a three month period.
References
Andolfatto D, Hendry S, Moran K (2008) Are inflation expectations rational. J Monet Econ 55(2):406–422
Amusa K, Gupta R, Karolia S, Simo-Kengne BD The long-run impact of inflation in South Africa. J Policy Model (Forthcoming)
Bai J, Perron P (1998) Estimating and testing linear models with multiple structural changes. Econometrica 66(1):47–78
Barro RJ, Sala-i-Martin X (2003) Economic growth, 2nd edn. MIT Press, Cambridge, MA
Barry D, Hartigan JA (1993) A Bayesian analysis for change point problems. J Am Stat Assoc 88:309–319
Baxter M, King RG (1993) Fiscal policy in general equilibrium. Am Econ Rev 83(3):315–334
Bierens HJ (1997) Nonparametric cointegration analysis. J Econ 77:379–404
Bierens HJ (2000) Nonparametric nonlinear cotrending analysis, with an application to interest and inflation in the United States. J Bus Econ Stat 18(3):323–337
Blanchard OJ (1985) Debt, deficits, and finite horizons. J Polit Econ 93(2):223–247
Blanchard OJ, Fischer S (1989) Lectures on macroeconomics. MIT Press, Cambridge, MA
Blough SR (1992) The relationship between power and level for generic unit root tests in finite samples. J Appl Econ 7(3):295–308
Breeden DT (1979) An intertemporal asset pricing model with stochastic consumption and investment opportunities. J Finance Econ 7(3):265–296
Breitung J (2002) Nonparametric tests for unit roots and cointegration. J Econ 108(2):343–363
Bullard JB, Russell SH (2004) How costly is sustained low inflation for the U.S. economy? Fed Reserv Bank St.Louis Rev 86(3):35–67
Cass D (1965) Optimum growth in an aggregate model of capital accumulation. Rev Econ Stud 32(3):233–240
Crowder WJ, Wohar ME (1998) Are tax effects important in the long-run fisher relation? Evidence from the municipal bond market. J Finance 54(1):307–317
Darby MR (1975) The financial and tax effects of monetary policy on interest rates. Econ Inq 13(2):266–276
de Bruyn R, Gupta R, Stander L (2011) Testing the monetary model for exchange rate determination in South Africa: evidence from 101 years of data. University of Pretoria Department of Economics Working Paper 201134
DeJong DN, Nankervis JC, Savin NE, Whiteman CH (1992) The power problems of unit root tests in time series with autoregressive errors. J Econ 53(1-3):323–343
Diamond PA (1965) National debt in a neoclassical growth model. Am Econ Rev 55(5):1126–1150
Dickey DA, Fuller WA (1979) Distribution of the estimators for autoregressive time series with a unit root. J Am Stat Assoc 74(366):427–431
Elliott G, Rothenberg TJ, Stock JH (1996) Efficient tests for an autoregressive unit root. Econometrica 64(4):813–836
Engle RF, Granger CWJ (1987) Co-integration and error correction: representation, estimation, and testing. Econometrica 55(2):251–276
Erdman C, Emerson JW (2007) bcp; an r package for performing a Bayesian analysis of change point problems. J Stat Softw 23(3):1–13
Espinosa-Vega MA, Russell SH (1998a) Can higher inflation reduce real interest rates in the long run? Can J Econ 31(1):92–103
Espinosa-Vega MA, Russell SH (1998b) The long-run real effects of monetary policy: Keynesian predictions from a neoclassical model. Federal Reserve Bank of Atlanta. Working Paper No: 98–6
Faust J (1996) Near observational equivalence and theoretical size problems with unit root tests. Econo Theory 12(4):724–31
Friedman M (1963) Inflation: causes and consequences. Asia Publishing House, New York
Geweke J, Porter-Hudak S (1983) The estimation and application of long memory time series models. J Time Ser Anal 221–238
Granger CWJ (1980) Long memory relationships and the aggregation of dynamic models. J Econ 14(2):227–238
Granger CWJ, Joyeux R (1980) An introduction to long-memory time series models and fractional differencing. J Time Ser Anal 1:15–39
Gregory AW, Hansen BE (1996) Residual-based tests for cointegration in models with regime shifts. J Econ 70(1):99–126
Gupta R, Reid M Macroeconomic surprises and stock returns in South Africa. Stud Econ Finance (forthcoming)
Hamilton JD (1989) A new approach to the economic analysis of nonstationary time series and the business cycle. Econometrica 57(2):357–384
Hansen BE (1999) The grid bootstrap and the autoregressive model. Rev Econ Stat 81(4):594–607
Hansen LP, Singleton KJ (1982) Generalized instrumental variables estimation of nonlinear rational expectations models. Econometrica 50(5):1269–1286
Hansen LP, Singleton KJ (1983) Stochastic consumption, risk aversion, and the temporal behavior of asset returns. J Polit Econ 91(2):249–265
Hosking JRM (1981) Fractional differencing. Biometrika 68(1):165–176
Huizinga J, Mishkin FS (1986) monetary policy regime shifts and the unusual behavior of real interest rates. Carnegie-Rochester Conf Ser Pub Policy 24:231–274
Johansen S (1991) Estimation and hypothesis testing of cointegration vectors in Gaussian vector autoregressive model. Econometrica 59(6):1551–1580
Jouini J, Nouira L (2006) Mean-shifts and long-memory in the U.S. ex post real interest rate. Universit de la Mditerrane, GREQAM, Mimeo
Kahn B, Farrell G (2002) South African real interest rates in comparative perspective: theory and evidence. South African Reserve Bank Occasional Paper 17
Kapetanios G, Shin Y, Snell A (2003) Testing for a unit root in the non-linear star framework. J Econ 112:359–379
Koopmans TC (1965) On the concept of optimal economic growth. In: The economic approach to development planning. Elsevier, Amsterdam, pp 225–300
Kryshko M (2006) Nominal exchange rates and uncovered interest parity: non-parametric co-integration analysis. University of Pennsylvania Department of Economics, Mimeo
Kwiatkowski D, Phillips PCB, Schmidt P, Shin Y (1992) Testing the null hypothesis of stationarity against the alternative of a unit root: how sure are we that economic time series have a unit root. J Econ 54(1-3):159–178
Lacerda M, Fedderke JW, Haines LM (2010) Testing for purchasing power parity and uncovered interest parity in the presence of monetary and exchange rate regime shifts. S Afr J Econ 78(4):363–382
Lee J, Strazicich MC (2003) Minimum LM unit root test with two structural breaks. Rev Econ Stat 85(4):1082–1089
Li J, Lee J (2010) Adl tests for threshold cointegration. J Time Ser Anal 31(4):241–254
Lioui A, Poncet P (2008) Monetary non-neutrality in the Sidrauski model under uncertainty. Econ Lett 100(1):22–26
Lucas RE (1978) Asset prices in an exchange economy. Econometrica 46(6):1429–1445
Ludi K, Ground M (2006) Investigating the bank lending channel in South Africa: a VAR approach. University of Pretoria Department of Economics Working Paper: 200604
Lo AW (1991) Long-term memory in stock market prices. Econometrica 59:1279–1313
Mishkin FS (1981) The real rate of interest: an empirical investigation. Carnegie-Rochester Conf Ser Pub Policy 15:151–200
Mishkin FS (1984) The real interest: rate: a multi-country empirical study. Can J Econ 17(2):283–311
Mundell R (1963) Inflation and real interest. J Polit Econ 71(3):280–283
Neely CJ, Rapach DE (2008) Real interest rate persistence: evidence and implications. Fed Reserv Bank St. Louis Rev 90(6):609–641
Nelson CR, Schwert GW (1997) Short-term interest rates as predictors of inflation: on testing the hypothesis that the real interest rate is constant. Am Econ Rev 67(3):478–486
Ng S, Perron P (2001) Lag length selection and the construction of unit root tests with good size and power. Econometrica 69(6):1519–554
Peláez RF (1995) The Fisher effect: reprise. J Macroecon 17(2):333–346
Perron P (1989) The great crash, the oil price shock and the unit root hypothesis. Econometrica 57(6):1361–1401
Perron P, Rodriguez GH (2001) Residual based tests for cointegration with GLS detrended data. Boston University, Mimeo
Phillips PCB, Perron P (1988) Testing for a unit root in time series regression. Biometrika 75(2):335–346
Rapach DE, Wohar ME (2004) The persistence in international real interest rates. Int J Finance Econ 9(4):339–346
Rapach DE, Wohar ME (2005) Regime changes in international real interest rates: are they a monetary phenomenon. J Money Credit Bank 37(5):887–906
Reid M (2009) The sensitivity of South African inflation expectations to surprises. S Afr J Econ 77(3):414–429
Reis R (2007) The analytics of monetary non-neutrality in the Sidrauski model. Econ Lett 94(1):129–135
Robinson PM (1992) Semi-parametric analysis of long-memory time series. Ann Stat 22:515–539
Romano JP, Wolf M (2001) Subsampling intervals in autoregressive models with linear time trends. Econometrica 69(5):1283–1314
Romer DH (2006) Advanced macroeconomics, 3rd edn. McGraw-Hill, New York
Romer CD, Romer DH (2004) A new measure of monetary shocks: derivation and implications. Am Econ Rev 94(4):1055–1084
Rose AK (1988) Is the real interest rate stable. J Finance 43(5):1095–1112
Samuelson PA (1958) An exact consumption-loan model of interest with or without the social contrivance of money. J Polit Econ 66(6):467–482
Shiller RJ (1979) The volatility of long-term interest rates and expectations models of the term structure. J Polit Econ 87(6):1190–1219
Shimotsu K (2008) Exact local whittle estimation of fractional integration with an unknown mean and time trend. Queen’s University Working Paper
Sims C (1988) Bayesian skeptism on unit root econometrics. Institute for Empirical Macroeconomics, University of Minnesota
Taylor JB (1993) Discretion versus policy rules in practice. Carnegie-Rochester Conf Ser Pub Policy 39:195–214
Tobin J (1965) Money and economic growth. Econometrica 33(4):671–684
Weiss LM (1980) The effects of money supply on economic welfare in the steady state. Econometrica 48(3):565–576
Yao Y (1984) Estimation of a noisy discrete-time step function: bayes and empirical bayes approaches. Annal Stat 12(4):1434–1447
Zivot E, Andrews W (1992) Further evidence on the great crash, the oil price shock and the unit root hypothesis. J Bus Econ Stat 10:251–270
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Das, S., Gupta, R., Kanda, P.T. et al. Real interest rate persistence in South Africa: evidence and implications. Econ Change Restruct 47, 41–62 (2014). https://doi.org/10.1007/s10644-012-9132-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10644-012-9132-5