Skip to content
Licensed Unlicensed Requires Authentication Published by De Gruyter June 28, 2019

Constrained interest rates and changing dynamics at the zero lower bound

Gregor Bäurle, Daniel Kaufmann ORCID logo, Sylvia Kaufmann and Rodney Strachan

Abstract

The interaction of macroeconomic variables may change as nominal short-term interest rates approach zero. In this paper, we propose to capture these changing dynamics with a state-switching parameter model which explicitly takes into account that the interest rate might be constrained near the zero lower bound by using a Tobit model. The probability of state transitions is affected by the lagged level of the interest rate. The endogenous specification of the state indicator permits dynamic conditional forecasts of the state and the system variables. We use Bayesian methods to estimate the model and to derive the forecast densities. In an application to Swiss data, we evaluate state-dependent impulse-responses to a risk premium shock identified with sign-restrictions. We provide an estimate of the latent rate, i.e. the rate lower than the constraint on the interest rate level which would be state- and model-consistent. Additionally, we discuss scenario-based forecasts and evaluate the probability of exiting the ZLB region. In terms of log predictive scores and the Bayesian information criterion, the model outperforms a model substituting switching with stochastic volatility and another including intercept switching only combined with stochastic volatility.

JEL Classification: C3; E3

Acknowledgement

We thank for constructive comments of the Editor, Associate Editor and two anonymous referees. Rodney Strachan acknowledges the support from the Australian Research Council, Discovery Project DP180102373.

A Appendix

Distributional properties of censored and uncensored variables

Given the normality assumption on εt, model (1) defines a joint normal distribution for the variables yt=[y1t,y2t], where y2t gathers the uncensored variables.

(16)[y1ty2t]|Xt,It,θN([m1Itm2It],[Σ11,ItΣ12,ItΣ21,ItΣ22,It])

where θ={βk,Σk,γr,γ|k=0,1} represents the model parameters and miIt=Xitβi,It and Σij,It are obtained by gathering the corresponding rows in (1) and by partitioning accordingly the moment matrices. This allows the expression of the joint observation density f(yt) as the product of a marginal and a conditional density, f(yt|)=f(y1t|y2t,)f(y2t|), where:

(17)f(y2t|)=N(m2It,Σ22,It)=N(X2tβ2It,Σ22,It)
(18)f(y1t|2|)=f(y1t|y2t,)=N(m1It|2,M1It|2)

with

m1It|2=m1It+Σ12,ItΣ22,It1(y2tm2It)M1It|2=Σ11,ItΣ12,ItΣ22,It1Σ21,It

The factoring of f(yt|) partitions the joint distribution into two parts and allows the implementation of a normal regression model for the unconstrained variables and a conditional censored normal regression model for the constrained variables

(19)[y~1t|2y2t]|Xt,It,θN([m1It|2m2It],[M1It|200Σ22,It])1(y~1tb)

B Appendix

Bayesian framework

B.1 Likelihood

Define the number Nj, j = 1, 2, which indicates the number of, respectively, censored and uncensored variables.

Conditional on I and using (19), the data likelihood can be factorized

(20)f(y~|X,I,θ)=t=p+1Tf(y~t|Xt,βIt,ΣIt)1(y~1tb)
(21)=t=p+1Tf(y~1t|y2t,X1t,β1It,Σ11,It)1(y~1tb)f(y2t|X2t,β2It,Σ22,It)

From (17), the period t density contribution is multivariate normal for y2t

(22)f(y2t|X2t,β2It,Σ22,It)=(2π)N2/2|Σ22,It|1/2×exp{12(y2tX2tβ2It)Σ22,It1(y2tX2tβ2It)}

and the period t contribution of censored variables is

(23)f(y~1t|y2t,X1t,β1It,Σ11,It)1(y~1tb)=Φ(M1It|21/2(bm1It|2))1(y~1t=b)×|M1It|2|1/2ϕ(M1It|21/2(y~1tm1It|2))1(y~1t>b)

where Φ(M1It|21/2(bm1It|2))1(y~1t=b) equals

bN1b1|M1It|2|1/2ϕ(M1It|21/2(y~1tm1It|2))dy~11,tdy~1N1,t,

and ϕ denotes the pdf (see (22)) of the standard (multivariate) normal distribution.

The likelihood of the complete data factorizes

(24)f(y|X,I,θ)=t=p+1Tf(y1t|y2t,X1t,β1It,Σ11,It)f(y2t|X2t,β2It,Σ22,It)

where the moments of the marginal and condition normal observation densities are given in, respectively, (17) and (18).

B.2 Prior distributions

To complete the Bayesian setup, we specify the prior density of the state indicator I:

(25)π(I|r,γ,γr)=t=p+1Tπ(It|rt1,γ,γr)

The prior for the censored variables is assumed to be diffuse, π(y1)1(y1b). We might also work with a proper prior distribution, restricted to the latent area, π(y1)N(0,κI)1(y1b) with κ some real number.

Finally, we assume independent priors for the model parameters:

(26)π(θ)=π(γ,γr)k=01π(βk)π(Σk)

The prior for (γr, γ) includes a state-identifying restriction and additional information on the threshold level, see (10).

The priors on βk are independent normal, with variance structure implied by Minnesota priors, π(βk)=N(v_,V_k). The vector v_ is of dimension N(Np + 1), see (1). Given that we estimate a VAR in levels, we center the first own autoregressive lag at 1 and all other coefficients at zero, v_={v_l|l=1,,N(Np+1)}, with

(v_l,V_k,ll)={1l=(j1)(Np+1)+(j+1),j=1,,N0otherwise

We specify the corresponding elements in V_k as

Var(Bkl,ij)={0.01/l2i=j0.25(0.01/l2)(σki2/σkj2)ij,i,j=1,,N

for k = 0,1 and l = 1, …, p. The state-specific variances in the scale factor (σki2/σkj2) are equal to the variance of residuals of univariate state-specific autoregressions in which states are predefined as I_t=1 if the libor ≤ 1%. For the intercepts, we work with diffuse priors, Var(μki)=5.

For Σk, we assume an inverse Wishart prior distribution IW(s_,S_k) with degrees of freedom s_=N+2 and scale S_k with diagonal elements S_k,ii=σki2.

B.3 Posterior distributions

To obtain draws from the posterior

π(ϑ|y~)f(y|X,I,θ)π(I|r,θ)π(y1)π(θ)

we sample iteratively from the posterior of

  1. 1.

    the state indicator, π(I|y,X,r,θ). Given that there is no state dependence in the state probabilities, we are able to sample the states simultaneously. We update the period t prior odds P(It=1)/P(It=0)=exp(γrrt1+γ) to obtain the posterior odds

    P(It=1|)/P(It=0|)=f(yt|Xt,β1,Σ1)exp(γrrt1+γ)f(yt|Xt,β0,Σ0),t=p+1,,T

    We sample Tp uniform random variables Ut, and set It = 1 if

    P(It=1|)/(P(It=0|)+P(It=1|))U
  2. 2.

    the censored variables, π(y1|y2,X,I,θ)1(y1b). Conditional on I and the observed variables, and given a diffuse prior, the moments of the posterior normal distribution π(y1|) are given by (18). We sample from this distribution truncated to the region y1b.

  3. 3.

    the parameters of the state distribution, π(γ|r,I)1(γr<0)1(γrγ_γγrγ¯). First, we introduce two layers of data augmentation, which renders the non-linear, non-normal model into a linear-normal model for the parameters (Frühwirth-Schnatter and Frühwirth, 2010):

    1. We express the state distribution in relative terms, i.e. as difference between the latent state utilities

      ϖt=I1tuI0tu=γrrt1+γ+ϵt,ϵtLogistic

      where

      I1tu=γrrt1+γ+ν1t, and I0tu=ν0twith νkt i.i.d. Type I EV
    2. We approximate the Logistic distribution by a mixture of normals with M components, R=(R1,,RT). Conditional on the latent relative state utilities ϖ and the components, we obtain a normal posterior distribution, N(g, G) with moments:

      G=(G01+t=p+1TZtZt/smt2)1g=G(G01g0+t=p+1TZtϖt/smt2)

      where Zt=[rt1,1] and smt2=sm2 is the variance of the mixture components Rt, see Table 2 in Frühwirth-Schnatter and Frühwirth (2010).

    To implement the restrictions on 𝜸 according to (6), we partition the posterior appropriately:

    π(γr,γ|)N([g1g2],[G11G12G21G22])

    Then we first sample γr,(mc) from N(g1,G11)1(γr<0) and then sample γ from the truncated conditional posterior (Robert, 2009; Botev, 2017):

    γ|γr=γr,(mc)N(g2c,G2c)1(γrγ_γγrγ¯)

    with moments

    g2c=g2G21G111(γr,(mc)g2)G2c=G22G21G111G12
  4. 4.

    the rest of the parameters, π(θγ|X,y1,I). Conditional on I and the augmented data y, the model in (1) is linear. The posterior distribution of the regression parameters and of the error variances are then, respectively, normal and inverse Wishart, the moments of which can be derived in the usual way.

C Appendix

Addtional results

C.1 Duration of minimum 2 periods

Figure 17: Duration of minimum 2 periods.Left-hand: Observed interest rate and model-based estimate of $\boldsymbol{y^{*}_{1}} < b$y1∗<b (red); the black line is the median, the areas decreasing in shades correspond to, respectively, the 25%, the 50% and the 80% interval of highest posterior density; mean posterior probability of state 1 (yellow). Right-hand: histogram of ${y^{*}_{1t}} < b$y1t∗<b for $t=2014.00,2014.25,2014.50$t=2014.00,2014.25,2014.50, i.e. 2014 first through third quarter; the second and third numbers under each histogram refer to, respectively, $P\left(y^{*}_{1t} < \min_{t}\{y_{1t}\}\right)$P(y1t∗<mint{y1t}) and $\textrm{median}\left(y^{*}_{1t}\right)$median(y1t∗).

Figure 17:

Duration of minimum 2 periods.

Left-hand: Observed interest rate and model-based estimate of y1<b (red); the black line is the median, the areas decreasing in shades correspond to, respectively, the 25%, the 50% and the 80% interval of highest posterior density; mean posterior probability of state 1 (yellow). Right-hand: histogram of y1t<b for t=2014.00,2014.25,2014.50, i.e. 2014 first through third quarter; the second and third numbers under each histogram refer to, respectively, P(y1t<mint{y1t}) and median(y1t).

C.2 Additional results of the Full model

Figure 18: Scatter plots: State It = 0 (x-axis) against It = 1 (y-axis) parameter draws along with the 45% line. (A) Intercept and lag 1. (B) lag 2.
Figure 18: Scatter plots: State It = 0 (x-axis) against It = 1 (y-axis) parameter draws along with the 45% line. (A) Intercept and lag 1. (B) lag 2.

Figure 18:

Scatter plots: State It = 0 (x-axis) against It = 1 (y-axis) parameter draws along with the 45% line. (A) Intercept and lag 1. (B) lag 2.

C.3 Additional results of the SV model

Figure 19: Switching VAR with SV. Scatter plots: State It = 0 (x-axis) against It = 1 (y-axis) parameter draws along with the 45% line. (A) Intercept and lag 1. (B) lag 2.
Figure 19: Switching VAR with SV. Scatter plots: State It = 0 (x-axis) against It = 1 (y-axis) parameter draws along with the 45% line. (A) Intercept and lag 1. (B) lag 2.

Figure 19:

Switching VAR with SV. Scatter plots: State It = 0 (x-axis) against It = 1 (y-axis) parameter draws along with the 45% line. (A) Intercept and lag 1. (B) lag 2.

References

Arias, J. E., J. F. Rubio-Ramírez, and D. F. Waggoner. 2018. “Inference Based on Structural Vector Autoregressions Identified with Sign and Zero Restrictions: Theory and Applications.” Econometrica 86: 685–720.10.3982/ECTA14468Search in Google Scholar

Auerbach, A. J., and M. Obstfeld. 2005. “The Case for Open-Market Purchases in a Liquidity Trap.” The American Economic Review 95: 110–137.10.3386/w9814Search in Google Scholar

Auerbach, A. J., and Y. Gorodnichenko. 2012. “Measuring the Output Responses to Fiscal Policy.” American Economic Journal: Economic Policy 4: 1–27.10.3386/w16311Search in Google Scholar

Bańbura, M., D. Giannone, and L. Reichlin. 2010. “Large Bayesian Vector Auto Regressions.” Journal of Applied Econometrics 25: 71–92.10.1002/jae.1137Search in Google Scholar

Bauer, M. D., and G. D. Rudebusch. 2016. “Monetary Policy Expectations at the Zero Lower Bound.” Journal of Money, Credit and Banking 48: 1439–1465.10.1111/jmcb.12338Search in Google Scholar

Baumeister, C., and L. Benati. 2013. “Unconventional Monetary Policy and the Great Recession: Estimating the Macroeconomic Effects of a Spread Compression at the Zero Lower Bound.” International Journal of Central Banking 9: 165–212.Search in Google Scholar

Bäurle, G., and D. Kaufmann. 2018. “Measuring Exchange Rate, Price, and Output Dynamics at the Effective Lower Bound.” Oxford Bulletin of Economics and Statistics 80: 1243–1266.10.1111/obes.12260Search in Google Scholar

Belmonte, M., G. Koop, and D. Korobilis. 2014. “Hierarchical Shrinkage in Time-Varying Parameter Models.” Journal of Forecasting 33: 80–94.10.1002/for.2276Search in Google Scholar

Benhabib, J., S. Schmitt-Grohe, and M. Uribe. 2002. “Avoiding Liquidity Traps.” Journal of Political Economy 110: 535–563.10.1086/339713Search in Google Scholar

Botev, Z. 2017. “The Normal Law Under Linear Restrictions: Simulation and Estimation via Minimax Tilting.” Journal of the Royal Statistical Society, Series B 79: 125–148.10.1111/rssb.12162Search in Google Scholar

Caggiano, G., E. Castelnuovo, and G. Pellegrino. 2017. “Estimating the Real Effects of Uncertainty Shocks at the Zero Lower Bound.” European Economic Review 100: 257–272.10.1016/j.euroecorev.2017.08.008Search in Google Scholar

Chan, J. C. C., and E. Eisenstat. 2018. “Bayesian Model Comparison for Time-Varying Parameter VARs with Stochastic Volatility.” Journal of Applied Econometrics 33: 509–532.10.1002/jae.2617Search in Google Scholar

Chib, S. 1992. “Bayesian Analysis in the Tobit Censored Regression Model.” Journal of Econometrics 51: 79–99.10.1016/0304-4076(92)90030-USearch in Google Scholar

Christensen, J. H., and G. D. Rudebusch. 2016. “Modeling Yields at the Zero Lower Bound: Are Shadow Rates the Solution?” In Dynamic Factor Models (Advances in Econometrics), edited by E. Hillebrand and S. J. Koopman, Vol. 35, 75–125. Howard House, Wagon Lane, Bingley BD16 1WA, UK: Emerald Group Publishing Limited.10.1108/S0731-905320150000035003Search in Google Scholar

Clark, T. E. 2011. “Real-Time Density Forecasts from Bayesian Vector Autoregressions with Stochastic Volatility.” Journal of Business & Economic Statistics 29: 327–341.10.1198/jbes.2010.09248Search in Google Scholar

Debortoli, D., J. Galí, and L. Gambetti. 2018. “On the Empirical (Ir)relevance of the Zero Lower Bound Constraint.” Economics Working Papers 1594, Department of Economics and Business, Universitat Pompeu Fabra.10.3386/w25820Search in Google Scholar

Diebold, F. X., F. Schorfheide, and M. Shin. 2017. “Real-Time Forecast Evaluation of DSGE Models with Stochastic Volatility.” In Theoretical and Financial Econometrics: Essays in Honor of C. Gourieroux, edited by S. Darolles, E. Renault, and A. Monfort, 322–332. Journal of Econometrics 201, Special issue.10.1016/j.jeconom.2017.08.011Search in Google Scholar

Doan, T., R. Litterman, and C. A. Sims. 1984. “Forecasting and Conditional Projection Using Realistic Prior Distributions.” Econometric Reviews 3: 1–100.10.3386/w1202Search in Google Scholar

Eggertsson, G. B., and M. Woodford. 2003. “The Zero Bound on Interest Rates and Optimal Monetary Policy.” Brookings Papers on Economic Activity 34: 139–235.10.1353/eca.2003.0010Search in Google Scholar

Eggertsson, G. B., M. Woodford, T. Einarsson, and E. M. Leeper. 2004. “Optimal Monetary and Fiscal Policy in a Liquidity Trap (with Comments).” NBER International Seminar on Macroeconomics 1: 75–144.10.1086/653955Search in Google Scholar

Frühwirth-Schnatter, S., and R. Frühwirth. 2010. “Data Augmentation and MCMC for Binary and Multinomial Logit Models.” In Statistical Modelling and Regression Structures – Festschrift in Honour of Ludwig Fahrmeir, edited by T. Kneib and G. Tutz, 111–132. Heidelberg: Physica-Verlag.10.1007/978-3-7908-2413-1_7Search in Google Scholar

Galí, J. and T. Monacelli. 2005. “Monetary Policy and Exchange Rate Volatility in a Small Open Economy.” Review of Economic Studies 72: 707–734.10.3386/w8905Search in Google Scholar

Gust, C. J., J. D. Lopez-Salido, M. E. Smith, and E. Herbst. 2017. “The Empirical Implications of the Interest-Rate Lower Bound.” American Economic Review 107: 1971–2006.10.1257/aer.20121437Search in Google Scholar

Iwata, S., and S. Wu. 2006. “Estimating Monetary Policy Effects when Interest Rates are Close to Zero.” Journal of Monetary Economics 53: 1395–1408.10.1016/j.jmoneco.2005.05.009Search in Google Scholar

Jordan, T. 2009. “Die Geldpolitik der Schweizerischen Nationalbank in stürmischen Zeiten.” Speech on 19 March, Swiss National Bank.Search in Google Scholar

Kaufmann, S. 2015. “K-state Switching Models with Time-Varying Transition Distributions – Does Loan Growth Signal Stronger Effects of Variables on Inflation?” Journal of Econometrics 187: 82–94.10.1016/j.jeconom.2015.02.001Search in Google Scholar

Kimura, T., and J. Nakajima. 2016. “Identifying Conventional and Unconventional Monetary Policy Shocks: A Latent Threshold Approach.” The B.E. Journal of Macroeconomics 16: 277–300.10.1515/bejm-2015-0074Search in Google Scholar

Krippner, L. 2015. Zero Lower Bound Term Structure Modeling. New York: Palgrave Macmillan.10.1057/9781137401823Search in Google Scholar

Lopes, H. F., and E. Salazar. 2005. “Bayesian Model Uncertainty in Smooth Transition Autoregressions.” Journal of Time Series Analysis 27: 99–117.10.1111/j.1467-9892.2005.00455.xSearch in Google Scholar

Miyao, R. 2002. “The Effects of Monetary Policy in Japan.” Journal of Money Credit and Banking 34: 376–392.10.1353/mcb.2002.0044Search in Google Scholar

Nakajima, J. 2011. “Monetary Policy Transmission Under Zero Interest Rates: An Extended Time-Varying Parameter Vector Autoregression Approach.” The B.E. Journal of Macroeconomics 11: 1–24.10.2202/1935-1690.2323Search in Google Scholar

Plante, M., A. W. Richter, and N. A. Throckmorton. 2018. “The Zero Lower Bound and Endogenous Uncertainty.” The Economic Journal 128: 1730–1757.10.1111/ecoj.12445Search in Google Scholar

Reifschneider, D., and J. C. Williams. 2000. “Three Lessons for Monetary Policy in a Low-Inflation Era.” Journal of Money, Credit and Banking 32: 936–966.10.2307/2601151Search in Google Scholar

Robert, C. P. 2009. “Simulation of Truncated Normal Variables.” arxiv:0907.4010v1 [stat.co], LSTA, Université Pierre et Marie Curie, Paris.Search in Google Scholar

Schenkelberg, H., and S. Watzka. 2013. “Real Effects of Quantitative Easing at the Zero Lower Bound: Structural VAR-Based Evidence from Japan.” Journal of International Money and Finance 33: 327–357.10.1016/j.jimonfin.2012.11.020Search in Google Scholar

SNB. 2003. “Swiss National Bank Lowers the Target Range for the Three-Month Libor Rate by 0.5 Percentage Points to 0%–0.75%.” Press release on 6 March, Swiss National Bank.Search in Google Scholar

SNB. 2009. “Swiss National Bank Takes Decisive Action to Forcefully Relax Monetary Conditions.” Monetary policy assessment on 12 March, Swiss National Bank.Search in Google Scholar

SNB. 2011. “Die Nationalbank Trifft Massnahmen Gegen den Starken Franken.” Press release on 3 August, Swiss National Bank.Search in Google Scholar

Swanson, E. T., and J. C. Williams. 2014a. “Measuring the Effect of the Zero Lower Bound on Medium- and Longer-Term Interest Rates.” The American Economic Review 104: 3154–3185.10.3386/w20486Search in Google Scholar

Swanson, E. T., and J. C. Williams. 2014b. “Measuring the Effect of the Zero Lower Bound on Yields and Exchange Rates in the U.K. and Germany.” Journal of International Economics 92: S2–S21.10.1016/j.jinteco.2013.11.009Search in Google Scholar

Teräsvirta, T., and H. M. Anderson. 1992. “Characterizing Nonlinearities in Business Cycles Using Smooth Transition Autoregressive Models.” Journal of Applied Econometrics 7: S119–S136.10.1002/jae.3950070509Search in Google Scholar

Woodford, M. 2003. Interest and Prices: Foundations of a Theory of Monetary Policy. Princeton: Princeton University Press.Search in Google Scholar

Wright, J. H. 2012. “What does Monetary Policy do to Long-Term Interest Rates at the Zero Lower Bound?” The Economic Journal 122: F447–F466.10.3386/w17154Search in Google Scholar

Wu, J. S., and F. D. Xia. 2016. “Measuring the Macroeconomic Impact of Monetary Policy at the Zero Lower Bound.” Journal of Money Credit and Banking 48: 253–291.10.3386/w20117Search in Google Scholar


Supplementary Material

The online version of this article offers supplementary material (DOI: https://doi.org/10.1515/snde-2017-0098).


Published Online: 2019-06-28

©2020 Walter de Gruyter GmbH, Berlin/Boston