ArticlePDF Available

Time-varying risk premia in emerging markets: Explanation by a multi-factor affine term structure model

Authors:

Abstract

From the empirical viewpoint, the Expectation Hypothesis Theory (EHT) of the term structure of interest rates has been extensively tested and rejected for US term structure data. Dai and Singleton [6] show that under the settings of Affine term structure models it is possible that one matches both the historical term structure dynamics and capture an important stylized fact that have contradicted the EHT: Time-varying risk premia. In emerging markets, economic conditions tend to be much less stable than in developed markets. For this reason, if risk premia is dynamic in such markets, intuition would suggest that it is more volatile than in developed markets, implying a stronger statistical rejection of the EHT. In this paper, we verify the robustness of Dai and Singleton's results under these more extreme market conditions. We estimate an arbitrage free Affine Gaussian model for the term structure of swaps in the Brazilian market. We propose an extensive empirical analysis which consists on: defining the optimal number of factors to be used in the model, estimating the model, giving interpretation to the state variables in terms of risk factors, and studying the model implied risk premia. In the end, we propose an application for risk management of interest rates futures portfolios.
A preview of the PDF is not available
... Tab ak and Andrade (2001), Lima and Issler (2003) and Brito et al. (2003) test the ex p ectations hyp othesis for Brazil, whereas Tab ak (2003) and Tab ak and Tab ata (2005) ex amine the resp onse of the term structure to modifications in the targ et of the short-term rate. Almeida (2004) estimates a standard affine term structure model to assess its adeq uacy and ap p licab ility in the Brazilian case. Finally, Silveira (2005) and Matsumara and Moreira (2005) were the first to op enly use macroeconomic variab les in a yield curve model, in order to try to ex p lain the dynamics of the term structure of interest rates in Brazil 1 . ...
... All p arameters are strong ly sig nificant 13 . The standard deviations of the measurement errors for the 1,2, 3, 4, 6 and 12 month yields are, resp ectively, 14, 0, 7, 10, 17 and 95 b asis p oints, slig htly lower than those ob tained in Almeida (2004). The time variation in the risk p remium dep ends g reatly on the yield curve slop e factor, which contains the hig hest coefficients (in module) of matrix λ 1 . ...
... ,Lima and Issler (2003),Brito et al. (2003) andAlmeida (2004) reject, at least p artially, the validity of the ex p ectations hyp othesis and, therefore, I have to sp ecify λ 1 = 0. ...
Chapter
Full-text available
In this chapter I characterize the relationship between macroeconomic variables and the terms structure of interest rates using the recent macro-finance approach adapted to the case of an emerging economy and applying it to Brazil. I find that macro variables help to explain the dynamics of the yield curve in emerging markets, specially in periods of high volatility. Moreover, the notion of great external vulnerability of emerging economies is confirmed by the strong role of the nominal exchange rate change, which explains up to 37% of the variation in yields in Brazil. However, the model does a poor job in forecasting yields during the financial crisis of 2008. This fact seems to be related to the strong fall in international commodity and industrial goods prices (in dollar terms), which limited the passthrough from the strong depreciation of the exchange rate to inflation.
... In our work, we denote B = (b i ) and the matrices, by parsimony, are chosen to be diagonal, therefore denoted by A = diag(a i ) and Σ = diag(σ i ) (i.e. a non-correlated volatility matrix). For more information in the admissibility restrictions in general affine processes, see Dai and Singleton (2000), and for an application of admissible models in the Brazilian market see Almeida (2004). ...
Article
We introduce a dynamical country risk index for emerging economies. The proposal is based on the intensity approach of credit risk, i.e. the default is the first jump of a point process with stochastic intensity. Two different models are used to estimate the yield spread. They differ in the relationship between the default-free instantaneous interest rate process and the intensity process. The dynamics of the interest rates is modeled through a multidimensional affine model, and the Kalman filter with an Expectation-Maximization algorithm is used to calibrate it. The USD interest rates constitute part of the input of the model, while prices of relevant domestic bonds in the emerging market complete the input. For an application, we select the Uruguayan bond market as the emerging economy.
... Tabak & Andrade (2001), Lima & Issler (2003), Brito (2003) (2004) avalia os efeitos de surpresas na política monetária sobre a curva de juros por meio de testes dos impactos das decisões do Comitê de Política Monetária sobre a curva de juros, concluindo que os efeitos de surpresas de políticas monetárias sobre a curva de juros brasileira foram reduzidos com a introdução do RMI, e que os agentes aumentaram seu poder de antecipação das decisões de política monetária. Almeida (2004) estima um modelo afim da curva de juros utilizando variáveis latentes para avaliar sua adequação e aplicabilidade ao caso brasileiro. Já o trabalho de Sekkel & Alves (2005) analisou os efeitos da política monetária e outros choques macroeconômicos sobre a dinâmica da estrutura a termo da taxa de juros no Brasil, tendo como principal resultado que a estrutura a termo fica menos inclinada com choques de política monetária. ...
Article
Full-text available
Under inflation targeting, monetary policy conduction is strongly related with the credibility of the regime. As far as is known, there are no empirical studies that measure the influence of the credibility of the inflation targeting regime on long term interest rate. Hence, based on the assumption that greater credibility reduces the interest rate spread, and based on preliminary evidence that the credibility and the spread are negatively correlated, the paper aims at investigating the influence of credibility on long term interest rates.
Article
Purpose Using a fiscal sentiment indicator, this study aims to verify whether fiscal sentiment affects the yield curve in Brazil. Since policymakers highlight the coordination between monetary and fiscal policies and the importance of fiscal policy to the expectations formation process in inflation targeting regimes, the authors also explore the transmission mechanism through inflation expectations. Hence, the study also analyzes the effect of fiscal sentiment on interest rate swap spreads through the inflation expectations channel. Design/methodology/approach Based on information obtained from official communiqués about fiscal policies issued by the Central Bank of Brazil and the Brazilian Ministry of Finance, the study builds a fiscal sentiment indicator. The econometric strategy to verify whether fiscal sentiment is related to the short tail of the yield curve is based on time series analysis through ordinary least squares and generalized method of moments estimates. In turn, to estimate the transmission mechanism through inflation expectations, the model uses interaction terms between fiscal sentiment and inflation expectations. Findings The results suggest a more optimistic (pessimistic) fiscal sentiment reduces (increases) swap spreads. The findings reveal that improvements in fiscal credibility and a more optimistic fiscal sentiment are able to reduce the positive marginal effect that inflation expectations variations have on interest rate swap spreads. Originality/value This study contributes to the literature, as, to the best of authors’ knowledge, it is the first to analyze the content of the communiqués related to fiscal policy, and based on this content, it extracts the sentiment related to the fiscal environment and analyzes the effect of this sentiment on the yield curve. Besides, different from existing studies that analyze the effect of fiscal backward-looking aspects (such as public debt, budget balance, taxes and public spending) on the yield curve, this study investigates forward-looking aspects related to fiscal policy (such as fiscal credibility and fiscal sentiment).
Technical Report
Full-text available
Se estima la prima por vencimiento a partir de un modelo afín de 4 componentes principales de la estructura a términos de las tasas de interés de los bonos soberanos de Colombia en pesos. Se sigue la metodología propuesta por Adrian et al. (2013) para el periodo comprendido entre enero de 2003 y octubre de 2014. Los resultados obtenidos indican que la prima por término es mayor y más volátil a medida que aumenta el vencimiento. También se observa que esta prima es decreciente en el tiempo, lo cual se puede asociar a las mejores condiciones del mercado de estos títulos, la mayor estabilidad macroeconómica y las mayores condiciones de liquidez a nivel internacional. Adicionalmente, el modelo estimado captura eventos de estrés observados en el mercado.
Article
In the current low interest rate environment, the possibility of a sudden increase in rates is a potentially serious threat to financial stability. As a result, analyzing interest rate risk (IRR) is critical for financial institutions and supervisory agencies. We propose a new method for generating yield curve scenarios for stress testing banks’ exposure to IRR based on the Nelson-Siegel (1987) yield-curve model. We show that our method produces yield-curve scenarios with a wider variety of slopes and shapes than scenarios generated by the historical and hypothetical methods typically used in the banking industry and proposed in the literature. We stress test the economic value of equity of a bank balance sheet based on Call Report data from a large U.S. bank. We show that our method provides more information about the bank’s exposure to IRR using fewer yield-curve scenarios than the alternative historical and hypothetical methods.
Article
In general, derivatives are not redundant instruments. They usu- ally carry some type of information independent of their underlying assets prices. Connections between derivatives prices and their re- spective underlying assets prices are interesting because they inform the degree of completeness of the market as a whole. In this paper, we propose to test the ability of a multi-factor ane model to price and hedge options, when it has been estimated based only on the un- derlying market data. Adopting the Brazilian ID-Futures market, we estimate how much of the information carried on its corresponding option market (IDI options) is explained by only the sources of un- certainty of the primary market. Our empirical results suggest that the Brazilian fixed income market is incomplete. We conclude then proposing a dynamic term structure model which is a promising can- didate to reconcile theory to the obtained empirical results.
Article
Full-text available
We derive and estimate an affine no-arbitrage model with default risk and macroeconomic state variables to investigate the determinants of the term structure of emerging market external debt for Brazil, Colombia and Mexico. In particular we assess the importance of US macroeconomic factors, country's solvency ratios and latent variables in determining each country's term-structure. Our results indicate that: (i) the single most important variable is the joint latent variable interpreted as international liquidity, being more important in longer yields - it accounts for around 35% of the emerging markets 6 months duration spread movements and around 46% of 10 years duration spreads; (ii) the contribution of US macro variables ranges from 15% of the movements in short spread to 10% in the longer spreads; (iii) solvency variables are relatively more important to explain spreads in Mexico than in Brazil and Colombia; (iv) the idiosyncratic latent factor - interpreted as political risk - is also relevant accounting for approximately 30% of the movements, but its effect is larger on shorter yields.
Article
Full-text available
This paper characterizes the term structure of Treasury bond yields for Brazil, and estimates a Nelson-Siegel Model to reproduce its stylized facts for the period 2004-2010. For this purpose, this paper uses a software developed by Fund staff. In addition, the paper estimates two versions of the Nelson-Siegel Model that incorporates macroeconomic variables with the aim of assessing the dynamic interactions between the yield curve and the macroeconomy.
Article
We construct a model for pricing sovereign debt that accounts for the risks of both default and restructuring, and allows for compensation for illiquidity. Using a new and relatively efficient method, we estimate the model using Russian dollar-denominated bonds. We consider the determinants of the Russian yield spread, the yield differential across different Russian bonds, and the implications for market integration, relative liquidity, relative expected recovery rates, and implied expectations of different default scenarios. THIS PAPER DEVELOPS A MODEL of the termstructure of credit spreads on sovereign bonds that accommodates: (i) Default or repudiation: The sovereign announces that it will stop making payments on its debt; (ii) Restructuring or renegotiation: The sovereign and the lenders ‘‘agree’’ to reduce (or postpone) the remaining payments; and (iii) A‘‘regime switch,’’ such as a change of government or the default of another sovereign bond that changes the perceived risk of future defaults.We build on the framework of Duffie and Singleton (1999), showing that cash flows promised by a sovereign bond can be discounted using a default-adjusted short-termdiscount rate that reflects the m ean arrival rate, and associated losses in market value upon arrival, of each of the aforementioned types of credit events. Since a sovereign credit event is often not a ‘‘liquidation event,’’ the model allows for continued trading (and pricing) of a sovereign instrument through credit events, possibly after writedowns in face value or cash distributions to creditors on credit event dates. Additionally, we accommodate the possibility that bonds issued by the same sovereign, of exactly the same type but possibly of different maturities, may be priced in the market using different discount factors. Reasons for this may include: (i) Bond covenants may not include cross-default clauses that would force, upon the default of one bond, the simultaneous default of other bonds of the same type, but of a different maturity. For various strategic reasons related to internal
Article
Brazil's public debt is large and interest payments weigh dangerously on the government's budget. In 2001 interest expenditure amounted to 7,3 per cent of GDP. On a mark-to-market basis (that is considering the effect of exchange rate depreciation on the value of foreign currency-denominated bonds) interest expenditure reached 9 per cent of GDP. These figures are large, though not unusual in high debt countries: they are comparable to those observed in some European countries (Italy, Greece and Belgium) prior to monetray union. Two factors contribute to this level of debt service: the size of the debt and the average cost of the debt.
Article
This paper presents a consistent and arbitrage-free multifactor model of the term structure of interest rates in which yields at selected fixed maturities follow a parametric muitivariate Markov diffusion process with "stochastic volatility." the yield of any zero-coupon bond is taken to be a maturity-dependent affine combination of the selected "basis" set of yields. We provide necessary and sufficient conditions on the stochastic model for this affine representation. We include numerical techniques for solving the model, as well as numerical techniques for calculating the prices of term-structure derivative prices. the case of jump diffusions is also considered. Copyright 1996 Blackwell Publishers.
Article
This paper presents a regression approach to measuring the information in forward interest rates about time varying premiums and future spot interest rates. Like earlier work, the regressions identify variation in the expected premiums on longer-maturity Treasury bills. The more novel evidence concerns the forecasts of future spot rates in forward rates. The regressions provide evidence that the one-month forward rate has power to predict the spot rate one month ahead. During periods preceding 1974, forward rates have reliable forecast power for one-month spot rates up to five months in the future.
Article
Linear projections of returns on the slope of the yield curve have contradicted the implications of the traditional “expectations theory”. This paper shows that these findings are not puzzling relative to a large class of richer dynamic term structure models. Specifically, we match all the key empirical findings reported by Fama and Bliss ((1987) American Economic Review 77 (4), 680–692) and Campbell and Shiller ((1991) Review of Economic Studies 58, 495–514), among others, within large subclasses of affine and quadratic-Gaussian term structure models. Additionally, we show that certain “risk-premium adjusted” projections of changes in yields on the slope of the yield curve recover the coefficients of unity predicted by the models. Key to this matching are parameterizations of the market prices of risk that let the risk factors affect the market prices of risk directly, and not only through factor volatilities. The risk premiums have a simple form consistent with Fama's findings on the predictability of forward rates, and are also shown to be consistent with interest-rate feedback rules used by a monetary authority in setting monetary policy.
Article
We make two contributions to the study of interest rates. The first is to characterize their dynamics in a new way. We estimate forecasting relations based on one-period changes in forward rates, which are more easily compared than earlier work on yields to the stationary theory of bond pricing. The second is to approximate these dynamics and other salient features of interest rates with an affine model. We show that models with “negative” factors come closer to accounting for the properties of interest rates, including their dynamics, than multifactor Cox-Ingersoll-Ross models.
Article
The known functional form of the conditional characteristic function (CCF) of discretely sampled observations from an affine diffusion is used to develop computationally tractable and asymptotically efficient estimators of the parameters of affine diffusions, and of asset pricing models in which the state vectors follow affine diffusions. Both ‘time-domain’ estimators, based on Fourier inversion of the CCF, and ‘frequency-domain’ estimators, based directly on the CCF, are constructed. A method-of-moments estimator based on the CCF is shown to approximate the efficiency of maximum likelihood for affine diffusion and asset pricing models.