25. (2019) Forecast ranked tailored equity portfolios, (with Cord Stern). Journal of In­ter­national Financial Markets, Institutions and Money, (In Press) doi.org/10.1016/j.intfin.2019.101138. Show Ab­stract

    We use a dynamic model averaging (DMA) approach to construct forecasts of individual equity returns for a large cross-section of stocks contained in the SP500, FTSE100, DAX30, CAC40 and SPX30 headline indices, taking value, momen­tum, and quality factors as predictor variables. Fixing the set of ‘forgetting factors’ in the DMA prediction framework, we show that highly significant return forecasts relative to the historic average benchmark are obtained for 173 (281) individual equities at the 1% (5%) level, from a total of 895 stocks. These statistical forecast improvements also translate into considerable economic gains, producing out-of-sample R2 values above 5% (10%) for 283 (166) of the 895 individual stocks. Equally weighted long only portfolios constructed from a ranking of the best 25% forecasts in each headline index can generate sizable returns in excess of a passive investment strategy in that index itself, even when transaction costs and risk taking are accounted for.

24. (2019) Identification and Estimation issues in Exponential Smooth Transition Autoregressive Models, Oxford Bulletin of Eco­nom­ics and Statistics, 81(3), 667-685. (Online Appendix, 5.19MB). Show Abstract

    Exponential smooth transition auto­re­gres­sive (ESTAR) models are widely used in the inter­national finance literature, particularly for the modelling of real exchange rates. We show that the exponential function is ill-suited as a regime weight­ing function be­cause of two undesirable properties. Firstly, it can be well approximated by a quadratic function in the threshold variable whenever the transition function parameter γ, which governs the shape of the function, is ‘small’. This leads to an identification problem with respect to the transition function para­meter and the slope vector, as both enter as a product into the conditional mean of the model. Secondly, the exponential regime weight­ing function can behave like an indicator function (or dummy variable) for very large values of the transition function parameter γ. This has the effect of ‘spuriously overfitting’ a small number of ob­ser­va­tions around the lo­cation parameter µ. We show that both of these effects lead to estimation problems in ESTAR models. We illustrate this by means of an empirical replication of a widely cited study, as well as a simulation exercise.

23. (2017) The Role of Jumps and Leverage in Forecasting Volatility in International Equity Markets, (with Katja Gisler) Journal of In­ter­national Money and Finance, 79(December), 1-19. (Online Appendix, 1.85MB). Show Abstract

    We analyse the importance of jumps and the leverage effect on forecasts of realized volatility in a large cross-section of 18 international equity markets, using daily realized measures data from the Oxford-Man Realized Library, and two widely employed empirical models for realized volatility that allow for jumps and leverage. Our out-of-sample forecast evaluation results show that the separation of realized volatility into a continuous and a discontinuous (jump) component is im­por­tant for the S&P 500, but of rather limited value for the remaining 17 international equity markets that we analyse. Only for 6 equity markets are significant and sizable forecast improvements realized at the one-step-ahead horizon, which, nevertheless, deteriorate quickly and abruptly as the pre­diction horizon increases. The inclusion of the leverage effect, on the other hand, has a much larger impact on all 18 international equity mar­kets. Forecast gains are not only highly significant, but also sizeable, with gains re­maining significant for forecast horizons of up to one month ahead.

22. (2017) Macroeconomic Factors and Equity Premium Predictability, (with Martin Tischhauser) International Review of Economics and Finance, 51(September), 621-644. (Data, 1.07MB). Show Abstract

    Neely et al. (2014) have recently demonstrated how to efficiently combine information from a set of popular technical indicators together with the standard Goyal and Welch (2008) predictor variables widely used in the equity premium forecasting literature to improve out- of-sample forecasts of the equity premium using a small number of principal components. We show that forecasts of the equity premium can be further improved by, first, incorporating broader macroeconomic data into the information set, second, improving the selection of the most re­le­vant factors and combining the most re­le­vant factors by means of a forecast combination regression, and third, imposing theoretically motivated positivity constraints on the forecasts of the equity premium. Applying standard out-of-sample forecast evaluation tests, we find that in particular our proposed forecast combination approach, which combines forecasts of the most re­le­vant Neely et al. (2014) and macro­economic factors and further imposes positivity constraints on the equity premium forecasts, generates statistically significant and eco­no­mi­cally sizable improvements over the best per­forming model of Neely et al. (2014). Out-of-sample R2 values can be as high as 1.75%, with (annualised) gains in certainty equivalent re­turns of up to 3.35%, relative to the ALL factors forecasts of Neely et al. (2014).

21. (2017) Measuring the Output Gap in Switzerland with Linear Opinion Pools, (with Oliver Müller) Economic Modelling, 64(August), 163-171. (Data, 100KB). Show Abstract

    We use the recently proposed linear opinion pool methodology of Garratt et al. (2014) to construct real-time ensemble nowcast densities of the out­put gap for Switzerland over an out-of-sample period from 2003:Q1 to 2015:Q4. The model space consists of a large number of bivariate VAR specifications for inflation and the output gap, with each specification using a different estimate of the output gap, lag order in the VAR, and structural break information. The ensemble now­cast densities for the output gap are constructed by combining the predictive densities of the individual VAR specifications, weighted by their ability to provide accurate density forecasts for inflation. The overall performance of the linear opinion pool is assessed by its real-time output gap nowcasts and by the size of the ex post revisions to the output gap nowcasts. We find that the linear opinion pool does not produce any more accurate density or point forecasts of in­fla­tion than a number of simple univariate bench­mark models that condition on the same structural break information. Further, the linear opinion pool’s real-time estimate of the output gap is no more robust to ex post re visions than the real-time estimates of the individual univariate output gaps. The fact that Swiss GDP price de­flator data are subject to large revisions, complicates the measurement and forecasting of inflation.

20. (2016) The Term Structure of Interest Rates in an estimated New Keynesian Policy Model, (with Philipp Lentner) Journal of Macroeconomics, 50(December), 126-150. Show Abstract

    We jointly estimate a New Keynesian Policy Model with a Gaussian affine no-arbitrage specification of the term structure of interest rates, and assess how important inflation, output and monetary policy shocks are as sources of fluctuations in interest rates and the term premium. We work with observable pricing factors and utilize the computationally convenient normalization of Joslin et al. (2013b). This allows us to estimate the model without needing to restrict the parameters driving the market prices of risk. Using data for the U.S. from 1962:Q1 to 2014:Q2, we find that inflation and the output gap account for around 80% of the unconditional forecast error variance of bond yields at the short and medium end of the term structure, while monetary policy shocks account for around 20%. Bond yields respond to macroeconomic shocks only gradually, peaking after about 4 quarters. This is due to sizable monetary policy inertia estimates in our model. At the peak of the response, inflation shocks increase bond yields by more than one-to-one, and output shocks by less than one-to-one, which is consistent with a Taylor type monetary policy rule. Our term premium estimate is strongly counter-cyclical and can capture salient features of the term structure that constitute a puzzle in the expectations hypothesis.

19. (2016) Global Equity Market Volatility Spillovers: A Broader Role for the United States, (with Katja Gisler) International Journal of Forecasting, 32(4), 1317–1339. Show Abstract

    Rapach et al. (2013) have recently shown that U.S. equity market returns carry valuable in­for­ma­tion to improve return forecasts in global equity markets. In this study, we extend the work of Rapach et al. (2013) and examine if U.S. based equity market in­for­ma­tion can be used to improve realized volatility forecasts in a large cross-section of international equity markets. We use volatility data for the U.S. and 17 foreign equity markets from the Oxford Man Institute’s realized library and augment for each foreign equity market our benchmark HAR model with U.S. equity market volatility in­for­ma­tion. We show that U.S. equity market vola­tility information substantially improves out-of-sample forecasts of realized volatility in all 17 foreign equity markets that we consider. Forecast gains are not only highly significant, but produce out-of-sample R2 values between 4.56% and 14.48%, with 12 of these being greater than 10%. The improvements in out-of-sample forecasts remain statistically significant for horizons up to 1 month ahead. A substantial part of the predictive gains are driven by forward looking volatility as captured by the VIX.

18. (2016) Heterogeneous Agents, the Financial Crisis and Exchange Rate Predictability, (with Gion Donat Piras) Journal of In­ter­na­tio­nal Money and Finance, 60(February), 313–359. (Data, 1.71MB). Show Abstract

    We construct an empirical heterogeneous agent model which optimally combines forecasts from fundamentalist and chartist agents and evaluates its out-of- sample forecast performance using daily data covering an overall period from January 1999 to June 2014 for six of the most widely traded currencies. We use daily financial data such as level, slope and curvature yield curve factors, equity prices, as well as risk aversion and global trade activity measures in the fundamentalist agent’s predictor set to obtain a proxy for the market’s view on the state of the macroeconomy. Chartist agents rely upon standard momentum, moving average and relative strength index technical indicators in their predictor set. Individual agent specific forecasts are constructed using a flexible dynamic model averaging framework and are then aggregated into a model combined forecast using a forecast combination regression. We show that our empirical heterogeneous agent model produces statistically significant and sizable forecast improvements over a random walk benchmark, reaching out-of- sample R2 values of 1.41, 1.07, 0.99 and 0.74 percent at the daily one-step ahead horizon for 4 out of the 6 currencies that we consider. Forecast gains remain significant for horizons up to three-days ahead. The forecast improvements are largely realised before and around the time of the Lehman Brothers collapse. We show further that our model combined forecasts produce economic value to a mean variance investor, yielding annualized Sharpe ratios of around 0.89 and annualized performance fees in excess of 460 basis points.

17. (2015) Forecasting Copper Prices with Dynamic Averaging and Selection Models, (with Carlo Moretto) North American Journal of Economics and Finance, 33(1), 1–38. (Online Appendix, 276KB). Show Abstract

    We use data from the London Metal Exchange (LME) to forecast monthly copper returns using the recently proposed dynamic model averaging and selection (DMA/DMS) framework, which incorporates time varying parameters as well as model averaging and selection into one unifying framework. Using a total of 18 predictor variables that include traditional fundamental indicators such as excess demand, inventories and the convenience yield, as well as indicators related to global risk appetite, momentum, the term spread, and various other financial series, we show that there exists a considerable predictive component in copper returns. Covering an out-of-sample period from May 2002 to June 2014 and employing standard statistical evaluation criteria we show that the out-of-sample R2 (relative to a random walk benchmark) can be as high as 18.5 percent for the DMA framework. Time series plots of the cumulative mean squared forecast errors and time varying coefficients show further that firstly, a large part of the improvement in the forecasts is realised during the peak of the financial crisis period at the end of 2008, and secondly that the importance of the most relevant predictor variables has changed substantially over the out-of-sample period. The coefficients of the SP500, the VIX, the yield spread, the TED spread, industrial production and the convenience yield predictors are most heavily affected, with the TED spread and yield spread coefficients even changing signs over this period. Our predictability results remain valid for forecast horizons up to 6 months ahead, but are weaker and smaller than at the one month horizon.

16. (2015) Measuring Fund Style, Performance and Activity: A New Style Profiling Approach, (with Robert Hill and Jon Eggins) Accounting and Finance, 55(1), 29–55. Show Abstract

    We construct new measures of fund style, performance and activity from linear com­bi­na­tions of off-the-shelf stock-market indices. A fund's benchmark portfolio is a linear combination of two or more reference portfolios that in a least-squares sense most closely approximates the fund's portfolio. The resulting linear combination scalar is itself a measure of fund style, and the distance between a fund and its benchmark is a measure of fund activity. Our approach has a number of advantages over existing cha­ra­cter­is­tic matching methods. We illustrate our approach using a data set of US institutional funds.

15. (2014) Equilibrium Credit: The Reference Point for Macroprudential Supervisors, (with Martin Melecky) Journal of Banking and Finance, 41(4), 135–154. (Addon Material, 249KB). Show Abstract

    Equilibrium credit is an important concept be­cause it helps to identify excessive credit provision in an economy. This paper proposes a structural approach to determine equilibrium credit which is based on the long-run through-the-cycle transaction demand for credit. Using a panel data set consisting of 49 high and middle-income countries from 1980 2010, we show that there exists considerable variation in the cross-country estimates of the income and price elasticities of credit and that the unit elasticity restriction implicitly imposed by the credit-to-GDP ratio is strongly rejected by the data. This suggests that the credit-to-GDP ratio is not appropriate to measure equilibrium credit. We show further that the cross-sectional variation in the income and price elasticities of credit can be related to a set of relevant economic, financial and institutional development indicators of a country. The main determinants that explain the cross-sectional variation in the income and price elasticities are financial depth, access to financial services, use of capital markets, efficiency and funding of domestic banks, central bank indepen­dence, the degree of supervisory inte­gration, and the experience of a financial crisis. As an empirical illustration, we compute equilibrium credit and credit gaps for eleven new EU member states using our structural framework and com­pare it to credit gaps based on the Basel III approach.

14. (2013) Macroprudential Stress Testing of Credit Risk: A Practical Approach for Policy Makers, (with Martin Melecky) Journal of Financial Stability, 9(3), 347–370. Show Abstract

    Drawing on the lessons from the global financial crisis and especially from its impact on the banking systems of Eastern Europe, the paper proposes a new practical approach to ma­cro­pru­dential stress testing. The proposed approach incorporates: (i) macroeconomic stress scenarios generated from both a country specific statistical model and historical cross-country crises ex­pe­ri­ence; (ii) indirect credit risk due to foreign currency exposures of unhedged borrowers; (iii) varying underwriting practices across banks and their asset classes based on their relative ag­gres­sive­ness of lending; (iv) higher correlations be­tween the probability of default and the loss given default during stress periods; (v) a negative effect of lending concentration and residual loan ma­tu­ri­ty on unexpected losses; and (vi) the use of an economic risk weighted capital adequacy ratio as the relevant outcome indicator to measure the resilience of banks to materialising credit risk. We apply the proposed approach to a set of Eastern European banks and discuss the results.

13. (2012) Understanding forecast failure of ESTAR models of real exchange rates, Empirical Economics, 34(1), 399-426. (Data, 56KB). Show Abstract

    The forecast performance of the empirical ESTAR model of Taylor, Peel and Sarno (2001) is ex­amined for 4 bilateral real exchange rate series over an out-of-sample evaluation period of nearly 12 years. Point as well as density forecasts are constructed, considering forecast horizons of 1 to 22 steps head. The study finds that no forecast gains over a simple AR(1) specification exist at any of the forecast horizons that are considered, regardless of whether point or density forecasts are utilised in the evaluation. Non-parametric methods are used in conjunction with simulation techniques to learn about the models and their forecasts. It is shown graphically that the non­linearity in the point forecasts of the ESTAR model decreases as the forecast horizon increases. The non-parametric methods show also that the multiple steps ahead forecast densities are normal looking with no signs of bi-modality, skewness or kurtosis. Overall, there seems little to be gained from using an ESTAR specification over a simple AR(1) model.

12. (2010) The impact of ECB monetary policy decisions and communication on the yield curve, (with Claus Brand and Jarkko Turunen) Journal of the European Economic Association, 8(6), 1266–1298. Show Abstract

    We use intraday changes in money market rates to construct indicators of news about monetary policy stemming separately from policy de­ci­sions and from official communication of the ECB, and study their impact on the yield curve. We show that communication may lead to sub­stantial re­vi­sions in expectations of monetary policy and at the same time exerts a significant impact on interest rates at longer maturities. Thereby, the maturity response pattern to com­mu­ni­cation is hump-shaped, while that of policy decisions is downward sloping.

11. (2008) An estimated New Keynesian Policy Model for Australia, (with Martin Melecky) The Economic Record, 84(264), 1-16. (Additional Notes, 259KB). Show Abstract

    An open economy New Keynesian policy model for Australia is estimated in this study. We in­ves­ti­gate how important external shocks are as a source of macroeconomic fluctuations when compared to domestic ones. The results of our analysis suggest that the Australian business cycle and domestic inflation are most affected by domestic demand and supply shocks, respectively. However, domestic output also appears to be strongly affected by foreign demand shocks, and domestic inflation by exchange rate shocks. Domestic variables do not seem to be significantly affected by foreign supply and monetary policy shocks.

10. (2006) Bootstrap causality tests of the relationship between the equity markets of the U.S. and other developed countries: Pre- and post-September 11, (with Eduardo Roca and Abdulnasser Hatemi-J), Journal of Applied Business Research, 22(3), 65-74. Show Abstract

    We analyse the causal relationship between the equity markets of the US and those of the UK, Japan, Germany, France, Canada and Australia based on leveraged bootstrap approach de­veloped by Hacker and Hatemi-J (2005). This method overcomes problems of non nor­malities and ARCH effects in the data. Using weekly MSCI price indices, we focus our investigation on the period 1998 to 2005 which we divided into two sub-periods to take into account the potential structural break arising from September 11. Our results show that before September 11, there was bi-directional causality between the US and Japan and between the US and Germany. In addition, there was also a uni-directional causality from the US to Canada and from the US to France. After September 11, the only causality was a uni­di­rec­tion­al one from the US to Japan and from the UK to the US. Thus, after September 11, the US Granger-caused a fewer number of markets. This could imply that after September 11, the other markets became more efficient in responding to information transmitted from the US market.

9. (2005) The extent and stability of long-run relationships between stock prices: Evidence from the U.S., the U.K. and Australia, (with Eduardo Roca), Investment Management and Financial Innovations, 2(4), 80-94. Show Abstract

   We examine whether a significant long-run relationship between the US, UK and Australian stock market prices exists and whether this relationship is temporally stable, based on the use of cointegration methodology as applied to the period of 1984 to 2001. We. find a stationary long-run relationship between the UK and US, but not between the Australian and US, and Australian and UK, markets. Our results further reveal that the UK-US relationship was impacted by two structural shocks - the October 1987 stock market crash and the 1993-1994 US bond market crash. This relationship was stable during the period before the October 1987 crash and remained to be so after this period until it was disrupted by the 1993-1994 US bond market crash. During the period starting after 1994 up to 2001, this relationship between the UK and US markets, however, ceased to exist. Thus, for investors with UK and US stocks in their portfolios, there was no need for re-balancing of their portfolio during the periods of 1984 to 1993 but this had to be done after 1994.

8. (2002) Equity market price interdependence between Australia and the Asian Tigers, (with Eduardo Roca), International Journal of Business Studies, 10(2), 61-74. Show Abstract

   The study investigates the extent and structure of long-term and short-term price interaction between the equity markets of Australia and the Asian Tigers - Hong Kong, Korea, Singapore and Taiwan, taking into account the Asian financial crisis. It applies cointegration and generalised variance decomposition and impulse response analyses using MSCI price index data. No sig­ni­fi­cant long-term relationship between Australia and the Asian Tigers is found both before and after the Asian crisis. No significant short-term re­la­tion­ship is also found during the period before the crisis. However, after the crisis, the study finds Australia to be significantly interdependent with Hong Kong and Singapore.

Policy Papers and Book Reviews

7. (2019) The Riksbank's method for stress testing banks' capital, (with Jieying Li, Peter van Santen, Peter Wallin, and Jakob Winstrand), Sveriges Riksbank Staff Memo.

6. (2017) Appendix D - Structural Estimates of the Probability of a Banking Crisis at Different Levels of Capital in: "Appropriate capital ratios in major Swedish banks – new perspectives", (with Markus Andersson), Sveriges Riksbank Staff Memo.

5. (2016) “Superforecasting: The Art and Science of Prediction”, by Philip Tetlock and Dan Gardner invited book review. Risks, 4(3), 1-5.

Central Bank Working Papers

4. (2017) Identification and Estimation issues in Exponential Smooth Transition Autoregressive Models. Riksbank Working Paper No. 344, Sveriges Riksbank, 2017. (Online Appendix, 5.19MB). Slides. Show Abstract

   Exponential smooth transition autoregressive (ESTAR) models are widely used in the in­ter­na­tio­nal finance literature, particularly for the mod­elling of real exchange rates. We show that the exponential function is ill-suited as a regime weighting function because of two undesirable properties. Firstly, it can be well approximated by a quadratic function in the threshold variable whenever the transition function parameter γ, which governs the shape of the function, is ‘small’. This leads to an identification problem with respect to the transition function parameter and the slope vector, as both enter as a product into the conditional mean of the model. Secondly, the exponential regime weighting function can behave like an indicator function (or dummy variable) for very large values of the transition function parameter γ. This has the effect of ‘spuriously overfitting’ a small number of ob­ser­va­tions around the location parameter µ. We show that both of these effects lead to estimation problems in ESTAR models. We illustrate this by means of an empirical replication of a widely cited study, as well as a simulation exercise.

3. (2013) Equilibrium Credit: The Reference Point for Macroprudential Supervisiors, (with Martin Melecky) World Bank Policy Research Working Paper Series No. 6358, The World Bank, 2013. Show Abstract

   Equilibrium credit is an important concept as it helps identify excessive credit provision. This paper proposes a two-stage approach to de­ter­mine equilibrium credit. The two stages allow us to study changes in the demand for credit due to varying levels of economic, financial and in­sti­tu­tio­nal development of a country. Using a panel of high- and middle-income countries over the period 1980-2010, we provide empirical evidence that the credit-to-GDP ratio is in­ap­pro­pri­ate to measure equilibrium credit. The reason for this is that such an approach ignores heterogeneity in the parameters that determine equilibrium credit across countries due to different stages of economic development. The main drivers of this heterogeneity are financial depth, access to financial services, use of capital markets, efficiency and funding of domestic banks, central bank independence, the degree of supervisory in­te­gra­tion, and experience of a financial crisis. Also, countries in Europe and Central Asia show a slower adjustment of credit to its long-run equilibrium compared to other regions of the world.

2. (2012) Macroprudential Stress Testing of Credit Risk: A Practical Approach for Policy Makers, (with Martin Melecky) World Bank Policy Research Working Paper Series No. 5936, The World Bank, 2012. Show Abstract

   Drawing on the lessons from the global financial crisis and especially from its impact on the banking systems of Eastern Europe, the paper proposes a new practical approach to macro­pru­den­tial stress testing. The proposed approach incorporates: (i) macroeconomic stress scenarios generated from both a country specific statistical model and historical cross-country crises ex­pe­ri­ence; (ii) indirect credit risk due to foreign cur­ren­cy exposures of unhedged borrowers; (iii) varying underwriting practices across banks and their asset classes based on their relative ag­gres­sive­ness of lending; (iv) higher correlations be­tween the probability of default and the loss given default during stress periods; (v) a negative effect of lending concentration and residual loan ma­tu­ri­ty on unexpected losses; and (vi) the use of an economic risk weighted capital adequacy ratio as the relevant outcome indicator to measure the resilience of banks to materialising credit risk. We apply the proposed approach to a set of Eastern European banks and discuss the results.

1. (2006) The impact of ECB monetary policy decisions and communication on the yield curve (with Claus Brand and Jarkko Turunen), ECB Working Paper No. 657. Show Abstract

   We analyse high-frequency changes in the euro area money market yield curve on dates when the ECB regularly sets and communicates decisions on policy interest rates to construct different in­di­ca­tors of monetary policy news relating to policy decisions and to central bank com­mu­ni­ca­tion. The indicators show that ECB com­mu­ni­ca­tion during the press conference may result in significant changes in market expectations of the path of monetary policy. Furthermore, our results suggest that these changes have a significant and sizeable impact on medium to long-term interest rates.