Abstracts by Topic

“An Econometric Characterization of Business Cycle Dynamics with Factor Structure and Regime Switches,” International Economic Review, Vol. 39, No. 4, 969-96, 1998. (Download: Working Paper or IER)

 Abstract: A dynamic factor model with regime switching is proposed as an empirical characterization of business cycles. The approach integrates the idea of comovements among macroeconomic variables and asymmetries of business cycle expansions and contractions. The first is captured with an unobservable dynamic factor and the second by allowing the factor to switch regimes. The model is estimated by maximizing its likelihood function and the empirical results indicate that the combination of these two features leads to a successful representation of the data relative to extant literature. This holds for within and out-of-sample and for both revised and real time data.   

“Dating Business Cycle Turning Points in Real Time,” with James D. Hamilton, “Nonlinear Time Series Analysis of Business Cycles,” ed. Van Dijk, Milas, and Rothman, Elsevier’s Contributions to Economic Analysis series, 1-54, 2006. (Download DatingBC)

 Abstract:This paper discusses formal quantitative algorithms that can be used to identify business cycle turning points. An intuitive, graphical derivation of these algorithms is presented along with a description of how they can be implemented making very minimal distributional assumptions. We also provide the intuition and detailed description of these algorithms for both simple parametric univariate inference as well as latent-variable multiple-indicator inference using a state-space Markov-switching approach. We illustrate the promise of this approach by reconstructing the inferences that would have been generated if parameters had to be estimated and inferences drawn based on data as they were originally released at each historical date. Our recommendation is that one should wait until one extra quarter of GDP growth is reported or one extra fmonmonth of the monthly indicators released before making a call of a business cycle turning point. We introduce two new measures for dating business cycle turning points, which we call the “quarterly real-time GDP-based recession probability index” and the “monthly real-time multiple-indicator recession probability index” that incorporate these principles. Both indexes perform quite well in simulation with real-time data bases. We also discuss some of the potential complicating factors one might want to consider for such an analysis, such as the reduced volatility of output growth rates since 1984 and the changing cyclical behavior of employment. Although such refinements can improve the inference, we nevertheless recommend the simpler specifications which perform very well historically and may be more robust for recognizing future business cycle turning points of unknown character.

“A Comparison of the Real-Time Performance of Business Cycle Dating Methods,” with Jeremy Piger, Journal of Business Economics and Statistics, Vol. 26, No. 1,  42-49, 2008. (Download Repec or BC_RealTime)

Abstract: This paper evaluates the ability of formal rules to establish U.S. business cycle turning point dates in real time. We consider two approaches, a nonparametric algorithm and a parametric Markov-switching dynamic-factor model. In order to accurately assess the real-time performance of these rules, we construct a new unrevised "real-time" data set of employment, industrial production, manufacturing and trade sales, and personal income. We then apply the rules to this data set to simulate the accuracy and timeliness with which they would have identified the NBER business cycle chronology had they been used in real time for the past 30 years. Both approaches accurately identified the NBER dated turning points in the sample in real time, with no instances of false positives. Further, both approaches, and especially the Markov-switching model, yielded significant improvement over the NBER in the speed with which business cycle troughs were identified. In addition to suggesting that business cycle dating rules are an informative tool to use alongside the traditional NBER analysis, these results provide formal evidence regarding the speed with which macroeconomic data reveals information about new business cycle phases.

“Identifying Business Cycle Turning Points in Real Time,” with Jeremy Piger, Federal Reserve Bank of Saint Louis Review, March/April, 47-62, 2003. (Download Repec or TPrealtime) 

Abstract: In this paper we take it as given that the NBER correctly identifies the dates of business cycle turning points. We then evaluate the real-time performance of the Markov-switching model in replicating the NBER’s business cycle dates. We apply the model to two datasets, growth in quarterly real gross domestic product (GDP) and growth in monthly economywide employment. We first confirm the result found elsewhere that the model is able to replicate the historical NBER business cycle dates fairly closely when estimated using all available data. Second, we evaluate the real-time performance of the model at dating business cycles over the past 40 years; this is accomplished by estimating the model on recursively increasing samples of data and evaluating the evidence for a new turning point at the end of each sample. 

"Business Cycle Monitoring with Structural Changes,” with Simon Potter, International Journal of Forecasting, Vol. 6, No. 4, 777-793, 2010. (Download Repec or MonitorBreaks).

Abstract: This paper examines the predictive content of coincident variables for monitoring U.S. recessions in the presence of instabilities. We propose several specifications of a probit model for classifying phases of the business cycle. We find strong evidence in favor of the ones that allow for the possibility that the economy has experienced recurrent breaks. The recession probabilities of these models provide a clearer classification of the business cycle into expansion and recession periods, and superior performance in the ability to correctly call recessions and to avoid false recession signals. Overall, the sensitivity, specificity, and accuracy of these models are far superior as well as their ability to timely signal recessions. The results indicate the importance of considering recurrent breaks for monitoring business cycles

“Increased Stabilization and the G7 Business Cycles” with Fang Dong, in Business Fluctuations and Cycles, ed. T. Nagakawa, 265-283, 2007. (Download StableG7 or here) 

Abstract: This paper models the G7 business cycle using a common factor model, which is used to investigate increased stabilization and its impact on business cycle phases. We find strong evidence of a decline in volatility in each of the G7 countries. We also find a break towards stability in their common business cycle. This reduction in volatility implies that recessions will be significantly less frequent in the future compared to the historical track.

“International Business Cycles: G7 and OECD Countries,” with Chengxuan Yu; Economic Review, Federal Reserve Bank of Atlanta, First Quarter, Vol. 91 No. 1, 43-54, 2006. (Download Repec or Bc_G7OECD)

Abstract: The progressive globalization of markets has sparked a worldwide interest in using economic indicators to analyze cyclical fluctuations. Governments and the private sector seeking to conduct their activities in light of both national and international economic conditions could benefit from international indicators that serve as a warning system to detect recessions in major economic partners and in industrialized countries as a whole. This article constructs just such a warning system. Using a Markov-switching dynamic factor model with a self-adjusting variable-bandwidth filter, we construct business cycle indicators for G7 countries and for an aggregate measure of output by twenty-nine member countries of the Organisation for Economic Co-operation and Development (OECD). The model yields probabilities of the current business cycle phase for each G7 country and for the aggregate OECD and G7 output measures and reveals a common cycle underlying the OECD countries that characterizes an international business cycle. The proposed filter sorts out minor contractions and estimates only major economic recessions and expansions, thereby minimizing the occurrence of false turning points. This feature is especially important for central banks that may want to adjust monetary policy only in the event of major recessions affecting a broad set of economic sectors.

"Maturing Capitalism and Stabilization: International Evidence,” with G. Popli in Journal of Business and Economics, Vol. 1, No. 12, 5-22, 2003. (Download or JBE)

Abstract: Recent literature has found that the U.S. business cycle has experienced a substantial decrease in volatility since the mid-1980s. An increased stability of business cycles has important policy implications since it affects the frequency, duration, and probabilities of future recessions and expansions. The findings are that the increased stabilization is widespread across many sectors of the U.S. economy. However, most authors have considered this as a recent phenomenon particular to the U.S., which narrows the search for potential causes. In this paper we go one step further and investigate whether this recent change is unique to the U.S. and a phenomenon particular to the 1980s alone or if this is part of a long run trend in volatility shared by several countries.  In particular, we examine whether maturing capitalism has engendered a continuous stabilization of business cycles in eleven industrialized countries over time. We do not try to quantify changes in volatility pre and post-War, which could be compromised by differences in the quality of the data.  Instead, we focus on examining structural changes in the long run trend of volatility in these countries.  Recursive stabilization tests are applied to examine breaks in the volatility of production in these countries, assuming that their dates are unknown. We find strong evidence of multiple structural breaks leading to more stability in these countries over time, and that the recent decrease in U.S. output volatility is part of a broader long-term trend shared by all industrialized countries studied. Since these breaks tend to be clustered for groups of countries, this makes it easier to investigate major common historical experiences that may explain changes in volatility.

"Markov Switching in Disaggregate Unemployment Rates,” with C.  Juhn and S. Potter, Empirical Economics, Vol. 27, No.2, 205-232, 2002.(Download Repec or MSunemployment). Reprint in: Advances in Markov Switching Models, ed. J.D. Hamilton and B. Raj. Studies in Empirical Economics. Physica-Verlag, 61-  88, 2002.

Abstract: We develop a dynamic factor model with Markov switching to examine secular and business cycle fluctuations in the U.S. unemployment rates. We extract the common dynamics amongst unemployment rates disaggregated for 7 age groups. The framework allows analysis of the contribution of demographic factors to secular changes in unemployment rates. In addition, it allows examination of the separate contribution of changes due to asymmetric business cycle fluctuations. We find strong evidence in favor of the common factor and of the switching between high and low unemployment rate regimes. We also find that demographic adjustments can account for a great deal of secular changes in the unemployment rates, particularly the abrupt increase in the 1970s and 1980s and the subsequent decrease in the last 18 years.

"Employment and the Business Cycle," with Jeremy Piger, Manchester School, Vol 81, S2, 16-42, 2013. (Download Repec, Working Paper or Manchester )

Abstract: This paper investigates the differences in the cyclical dynamics in employment on non‐agricultural payroll (ENAP) and total civilian employment (TCE), and the implications for monitoring US business cycles in real time. We find that employment measures have diverged considerably around the last three recessions and subsequent recoveries. This significantly impacts identification of turning points. Models that use TCE are more in line with the National Bureau of Economic Research (NBER) recession dating, and deliver faster call of troughs in real time, whereas models that include ENAP series yield delays in signaling troughs, especially the most recent ones.

“Sunspots, Animal Spirits, and Economic Fluctuations,” with J.T. Guo, Macroeconomic Dynamics, Vol. 7, No. 1, February 2003. (Download SSRN or sunspots)

Abstract: Multiple-equilibria macroeconomic models suggest that consumers and investors' perceptions about the state of the economy may be important independent factors for business cycles. In this paper, we examine empirically the interrelations between waves of optimism and pessimism and subsequent economic fluctuations. We focus on the behavior of non-fundamental movements in the consumer sentiment index, as a proxy for consumers' sunspots, and in the business formation index, representing investors' animal spirits, around economic turning points. We find that bearish consumers and entrepreneurs were present before the onset of some U.S. economic downturns, sometimes even when the fundamentals were all very strong. In particular, our analysis shows that self-fulfilling pessimism may have played a nontrivial role for the 1969-70, the 1973-75, and the 1981-82 recessions. The results are robust to a range of alternative linear and nonlinear specifications. Our evidence provides some empirical support for the role of non-fundamental rational expectations in economic fluctuations.

“The Brazilian Business Cycle and Growth Cycle,” Brazilian Economic Journal (Revista Brasileira de Economia), Vol. 56 nº 1, 75-106, 2002. (Download BrazilBCGC or Repec )

Abstract: This paper uses several procedures to date and analyse the Brazilian business and growth cycles. In particular, a Markov switching model is fitted to quarterly and annual real production data. The smoothed probabilities of the Markov states are used as predictive rules to define different phases of cyclical fluctuations of real Brazilian economic activity. The results are compared with different non-parametric rules. All methods implemented yield similar dating and reveal asymmetries across the different states of the Brazilian business and growth cycles, in which slowdowns and recessions are short and abrupt, while high growth phases and expansions are longer and less steep. The resulting dating of the Brazilian economic cycles can be used as a reference point for construction and evaluation of the predictive performance of coincident, leading, or lagging indicators of economic activity. In addition, the filtered probabilities obtained from the Markov switching model allow early recognition of the transition to a new business cycle phase, which can be used, for example, for evaluation of the adequate strength and timing of countercyclical policies, for reassessment of projected sales or profits by businesses and investors, or for monitoring of inflation pressures.

“Recent Changes in the U.S. Business Cycle,” with S. Potter, Manchester School, Vol. 69, No. 5, 481-508, 2001. (Download Repec , Working Paper or ChangesBC). 

Abstract: The U.S. business cycle expansion that started in March 1991 was the longest on record (as of year 2000). This paper uses statistical techniques to examine whether that expansion was a one-time unique event or whether its length is a result of a change in the stability of the U.S. economy. Bayesian methods are used to estimate a common factor model that allows for structural breaks in the dynamics of a wide range of macroeconomic variables. We find strong evidence that a reduction in volatility is common to the series examined. Further, the reduction in volatility implies that future expansions will be considerably longer than the historical average.

"International Stock Market Linkages," with B.Y. Chen, Handbook of Global Financial Market: Transformations, Dependence, and Risk Spillovers. Ed. S. Boubaker and D.K. Nguyen, World Scientific Publishing, March 2019. (Download )

Abstract: This paper investigates international stock market dynamics and their linkages. It uses factor models to extract stock market indicators from common cyclical stock components of industrialized countries, emerging markets, the BRICT, and global stock markets. We find that the stock market indicators for these groups are correlated with each other and with the global market factor. The BRICT display the highest average stock return and are the least correlated with the others. The stock return indicators as well as the global stock market factor show close relationship with economic downturns, entering in bear phases around the beginning of recessions, and in bull phases mid-way through recessions, anticipating future economic recovery. We also find that the stock return indicators are more persistent and, therefore, more predictable than the stock market of individual countries. We study international linkages across these stock market groups through impulse response analysis and find that economic development levels play an important role in shock propagation. In particular, all stock market indicators respond positively to global factor shocks, with the least reactive group being the BRICT, and the most responsive being the emerging markets. Interestingly, the BRICT respond negatively to positive shocks to industrialized countries stock markets, indicating that the BRICT may have a role in hedging risk. 

"A Dynamic Factor Model of the Yield Curve Components as a Predictor of the Economy", with Z. Senyuz, International Journal of Forecasting. Vol. 32, April-June 2016, 324-343, 2016. (Working Paper or IJF )

Abstract: In this paper, we propose an econometric model of the joint dynamic relationship between the Treasury yield curve components and the economy, for predicting business cycle turning points. The nonlinear multivariate dynamic factor model takes into account not only the popular slope, but also information extracted from the level and curvature of the yield curve, and from macroeconomic variables. We investigate the interrelationship between the phases of cyclical fluctuations in yield curve components and the phases of the business cycle. The results indicate a strong interrelationship between the yield curve and the economy. The proposed model has substantial incremental predictive value relative to alternative specifications. This result holds both in-sample and out-of-sample, using revised and real time unrevised data

"What Does Financial Volatility Tell Us About Macroeconomic Fluctuations?", with Z. Senyuz, Z. and E. Yoldas, Journal of Economic Dynamics and Control. Vol. 52, March, 340–360, 2015. (Download Working Paper or JEDC)

Abstract: We provide an extensive analysis of the predictive ability of financial volatility for economic activity. We consider monthly measures of realized and implied volatility from the stock and bond markets. In a dynamic factor framework, we extract the common long-run component of volatility that is likely to be linked to economic fundamentals. Based on powerful in-sample predictive ability tests, we find that the stock volatility measures and the common factor significantly improve macroeconomic forecasts of conventional financial indicators, especially over short horizons. A real-time out of sample assessment yields similar conclusions under the assumption of noisy revisions in macroeconomic data. In a nonlinear extension of the dynamic factor model, we identify two distinct volatility regimes, and show that the high-volatility regime provides early signals of the Great Recession, which was associated with severe financial distress and credit disintermediation.

“Nonlinear Risk,” with S. Potter, Macroeconomic Dynamics, Vol. 5, No. 4, 621-646, 2001. (Download Repec or Nonlinearrisk)

Abstract: This paper proposes a flexible framework for analyzing the joint time series properties of the level and volatility of expected excess stock returns. An unobservable dynamic factor is constructed as a nonlinear proxy for the market risk premia with its first moment and conditional volatility driven by a latent Markov variable. The model allows for the possibility that the risk-return relationship may not be constant across the Markov states or over time. We find a distinct business cycle pattern in the conditional expectation and variance of the monthly value-weighted excess return. Typically, the conditional mean decreases a couple of months before or at the peak of expansions, and increases before the end of recessions. On the other hand, the conditional volatility rises considerably during economic recessions. With respect to the contemporaneous risk-return dynamics, we find an overall significantly negative relationship. However, their correlation is not stable, but instead varies according to the stage of the business cycle. In particular, around the beginning of recessions, volatility increase substantially, reflecting great uncertainty associated with these periods, while expected returns decrease, anticipating a decline in earnings. Thus, around economic peaks there is a negative relationship between conditional expectation and variance. However, toward the end of a recession, expected returns are at its highest value as an anticipation of the economic recovery, and volatility is still very high in anticipation of the end of the contraction. That is, the risk-return relation is positive around business cycle troughs. This time-varying behavior also holds for non-contemporaneous correlations of these two conditional moments.

“Stock Market Fluctuations and the Business Cycle,” Journal of Economic and Social Measurement, Vol. 25, No. 3, 235-258, 1998/1999.(Download SSRN or SMBC)

 Abstract: This paper examines the dynamic relationship between stock market movements and business cycles at the monthly frequency. Given the forward-looking behavior of stock market investors, it explores the possibility of using fluctuations in the stock market to forecast business cycle turning points using promptly available financial variables. The model generates predictions of business cycle turning points, using economic variables, and anticipation of these predicted turns, using financial variables. The empirical analysis is divided in two parts: first, it studies the historical track record of the interaction between the stock market indicator, the business cycle indicator, and the NBER-dated recessions. From this analysis, it is found that bear markets are closely associated with low-growth phases and economic recessions. Second, it implements an out-of-sample real time analysis of the performance of the stock market indicator in predicting economic recessions. The recursive real time forecasting performance of the stock market indicator is compared with the unrevised Composite Leading Indicator (CLI), computed by the Conference Board. The proposed indicator presents several advantages over the CLI. First, the financial indicator is less noisy than the CLI, which makes it easier to use it as a tool for anticipating turning points. Second, since financial market participants continuously update their expectations about the state of the economy as new information becomes available on a daily basis, over the course of the month movements in the financial indicator reflect revisions in these market perceptions. That is, the indicator can be computed at the end of each month reflecting updated information for that month. The unrevised CLI is also released at the end of the month, but it reflects information from the previous month. In addition, revisions are incorporated in the CLI index with a much lower frequency. Third, even though the CLI contains stock prices as one of its components, predictions of turning points using the stock market factor lead predictions using the unrevised CLI. Thus, the stock market indicator extracted from the model is a leading indicator of the state of the business cycle, and has been shown to be an effective tool to anticipate economic turning points in real time.

“Coincident and Leading Indicators of the Stock Market,” with S. Potter, Journal of Empirical Finance, Vol. 7, No. 1, 87-111, 2000. (Download Repec or CoinleadSM)

Abstract: In this paper we have two goals: first, we want to represent monthly stock market fluctuations by constructing a non-linear coincident financial indicator. The indicator is constructed as an unobservable factor whose first moment and conditional volatility are driven by a two-state Markov variable. It can be interpreted as the investors’ real-time belief about the state of financial conditions. Second, we want to explore an approach in which investors may use their perceptions of the state of the economy to form forecasts of financial market conditions and possibly of excess returns. To investigate this, we build leading indicators as forecasts of the estimated coincident financial index. The leading indicators yield better within and out-of-sample performance in forecasting, not only the state of the stock market but also of excess stock returns, as compared with the performance obtained using linear methods that have been proposed in the existing literature.

"Monetary Policy Regimes and the Stock Market," with Chuanlei Sun, in Business Cycles in Economics: Types, Challenges and Impacts on Monetary Policies. Ed. Jason Hsu, Nova Science Chapter 6, 87-1116, 2014. (Download )

Abstract: This paper studies the relationship between monetary policy and the stock market across business cycle recessions and expansions. A nonlinear two-state Markov switching model is used to obtain regimes in interest rate cycles and in stock market cycles. The model identifies tight and loose monetary policy phases, and bear and bull stock market regimes. Turning points for each cycle are established and their lead-lag relationship is examined and contrasted with NBER recessions. We also examine the linear predictive relationship between interest rates and stock returns. The results indicate that there are strong linkages among interest rate cycles, stock market cycles, and business cycles. We find that, generally, the stock market enters a bear market phase at around the same time as monetary policy enters a tight phase, which is associated with the onset of economic recessions. We also find that future stock returns are substantially impacted by information on monetary policy regimes (i.e. loose or tight). On the other hand, interest rates are most influenced by future inflation and recessions rather than by changes in the stock market. 

“Coincident and Leading Indicators of the Stock Market,” with S. Potter, Journal of Empirical Finance, Vol. 7, No. 1, 87-111, 2000.(Download Repec or CoinleadSM)

Abstract: In this paper we have two goals: first, we want to represent monthly stock market fluctuations by constructing a non-linear coincident financial indicator. The indicator is constructed as an unobservable factor whose first moment and conditional volatility are driven by a two-state Markov variable. It can be interpreted as the investors’ real-time belief about the state of financial conditions. Second, we want to explore an approach in which investors may use their perceptions of the state of the economy to form forecasts of financial market conditions and possibly of excess returns. To investigate this, we build leading indicators as forecasts of the estimated coincident financial index. The leading indicators yield better within and out-of-sample performance in forecasting, not only the state of the stock market but also of excess stock returns, as compared with the performance obtained using linear methods that have been proposed in the existing literature.

"Leading Indicators of Country Risk and Currency Crises “ The Asian Experience,” with Fang Dong , Economic Review, Federal Reserve Bank of Atlanta, First Quarter, Vol. 89, No. 1, 26-37, 2004. (Download Repec or CountryRisk)

Abstract: Most emerging capital markets in recent years adopted a system that narrowly pegs their currencies’ exchange rates to the U.S. dollar. While such a system has a number of advantages, it makes a country vulnerable to shocks in mobile international capital markets and can lead to reactive strategies that can drive the country into a currency crisis and inflationary recession. ; This article aims to construct an early warning system for international currency crises using financial variables reflecting investors’ expectations and banking distress, which are highly sensitive to changes in the economic environment. The authors use a dynamic factor model that switches between two regimes—representing periods of relative calmness and periods prone to currency crises - to construct leading indicators of country risk and currency crises. ; The method is applied to evaluate the model’s in-sample and out-of-sample performance in anticipating currency crises in the last two decades in Thailand, Indonesia, and Korea. The model successfully produces early signals of these crises, particularly the most severe one, which occurred in 1997. ; The study’s success in signaling future currency crises in real time demonstrates that the model’s “country risk” indicators can be informative tools that allow central banks to take preemptive counterpolicy measures to avoid a crisis or mitigate its severity.

"Leading Indicators of the Capital Goods Industry in Brazil," with Igor Morais, Brazilian Review of Econometrics, Vol. 31, No. 1, 137-171, 2011.  Recipient of Honor Award from the National Confederation of Industries of Brazil  - Confederacao Nacional das Industrias, CNI. (Download LeadingCapital)  

Abstract: The goal of this paper is to build leading indicators to predict the capital goods business cycle in Brazil. We propose a probit model with autoregressive dynamics consisting of series with predictive power to anticipate contractions in this sector. The model is especially suitable for this sector as it includes information about the characteristics of cycle phases and their duration in the estimation of recession probability, and adapts well to the high volatility of the capital goods business cycle. The results indicate that the dynamic probit model has a better forecasting performance than the simple model in several aspects, both within and out of sample and in real time. 

“A Monthly Indicator of Brazilian GDP,” in Brazilian Review of Econometrics Vol. 21, No. 1, 1-15, 2001 – Recipient of Honor Award from the Brazilian Econometric Society. (Download Repec or BrazilIndicatorGDP)

Abstract: This paper constructs an indicator of Brazilian GDP at the monthly frequency. The peculiar instability and abrupt changes of regimes in the dynamic behavior of the Brazilian business cycle were explicitly modeled within nonlinear frameworks. In particular, a Markov switching dynamic factor model was used to combine several macroeconomic variables that display simultaneous comovements with aggregate economic activity. The model generates as output a monthly indicator of the Brazilian GDP and real time probabilities of the current phase of the Brazilian business cycle. The monthly indicator shows a remarkable historical conformity with cyclical movements of GDP. In addition, the estimated filtered probabilities predict all recessions in sample and out-of-sample. The ability of the indicator in linear forecasting growth rates of GDP is also examined. The estimated indicator displays a better in-sample and out-of-sample predictive performance in forecasting growth rates of real GDP, compared to a linear autoregressive model for GDP. These results suggest that the estimated monthly indicator can be used to forecast GDP and to monitor the state of the Brazilian economy in real time.

“Leading Indicators of the Brazilian Inflation,” Economic Research and Planning (Pesquisa e Planejamento Economico), Vol. 31, 1, 323-354, 2001 (Download Repec - English or BrazilLeadingInflation-Portuguese).

Abstract: The goal of this project is to construct leading indicators that anticipate inflation cycle turning points on a real time monitoring basis. As a first step, turning points of the IPCA inflation are determined using a periodic stochastic Markov switching model. These turning points are the event timing that the leading indicators should anticipate. A dynamic factor model is then used to extract common cyclical movements in a set of variables that display predictive content for inflation. The leading indicators are designed to serve as practical tools to assist real-time monitoring of monetary policy on a month-to-month basis. Thus, the indicators are built and ranked according to their out-of-sample forecasting performance. The leading indicators are found to be an informative tool for signaling future phases of the inflation cycle out-of-sample, even in real time when only preliminary and unrevised data are available.

"The Future of Oil: Geology versus Technology",  with Benes, J., Kamenik, O., Kumhof, M., Laxton, D., Mursula, S., Selody, J., International Journal of Forecasting, vol. 31:1, 207-221, 2014. Recipient of the 2015-2016 Outstanding Paper Award  from the International Journal of Forecasting. (Download or IJF)

Abstract: We discuss and reconcile the geological and economic/technological views concerning the future of world oil production and prices, and present a nonlinear econometric model of the world oil market that encompasses both views. The model performs far better than existing empirical models in forecasting oil prices and oil output out-of-sample. Its point forecast is for a near doubling of the real price of oil over the coming decade, though the error bands are wide, reflecting sharply differing judgments on the ultimately recoverable reserves, and on future price elasticities of oil demand and supply.

"vOILatility: Forecasting Oil Prices under Uncertainty," mimeo, University of California Riverside, 2018. 

Abstract: "Historically, oil prices are subject to sudden jumps as well as smoother changes due to changes in supply and demand. Although linear models may capture some of the dynamics in between the jumps in-sample, they fail to represent and predict nonlinearities underlying the market out-of-sample, real time. Some of the abrupt changes in oil price dynamics were due to OPEC decisions in the 1970s-1980s. Recent developments such as shifts to new technology or cooperation of Russia and OPEC can potential engender new structural breaks in the oil market dynamics, with the possibility of markedly different results in out of sample real time forecasts. Models and methods that take into account instability/breaks might substantially improve forecasts. Nonlinear models reveal additional information compared to frameworks that take into account only average linear effect of one series on another. This paper proposes a model specifically designed to forecast oil prices taking into account potential nonlinearities and nonstationarities. The autoregressive multivariate mixed frequency model has probabilities of structural breaks in the mean and volatility of oil price as a function of several variables including: indicators of potential sudden changes in oil supply/price (news on OPEC, Russia’s oil policy and changes in inventories), indicator of economywide demand and oil consumption in the largest consumers and importers of oil, indicator of recent technology shifts, indicator of changes in risk. Preliminary results indicate that the model provides accurate real-time forecast of oil price remarkably superior to forecasts from alternative linear frameworks."

"Incomplete Price Adjustment and Inflation Persistence," with Insu Kim. (Formerly circulated as "Microfoundations of Inflation Persistence in the New Keynesian Phillips Curve"). Forthcoming, Journal of Money, Credit, and Banking (Donwload Repec) 

Abstract:This paper proposes a sticky inflation model in which inflation persistence is endogenously generated from the optimizing behavior of forward-looking firms. Although firms change prices periodically, their ability to fully adjust them in response to changes in economic conditions is assumed to be constrained due to the presence of managerial and customer costs of price adjustment. In essence, the model assumes that price stickiness arises from a combination of staggered contracts as in Calvo (1983) as well as quadratic adjustment cost as in Rotemberg (1982). We estimate the model using Bayesian techniques. Our findings strongly support both sources of price stickinessin the U.S. data. The model performs well in matching microeconomic evidence on price setting, particularly regarding the size and frequency of price changes. The paper also shows how incomplete price adjustments in a staggered price contracts model limit the contribution of expectations to inflation dynamics: it generates the delayed response of inflation to demand and monetary shocks,and the observed correlation between inflation and economic activity.

"Real-Time Indicator of Weekly Inflation with a Mixed-Frequency Unobserved Component Model with Stochastic Volatility," with Mingyuan Jia, mimeo, University of California Riverside.

Abstract: This paper builds a coincident indicator of inflation at the weekly frequency. We propose a mixed-frequency unobserved component model in which the common permanent and transitory inflation components have time-varying stochastic volatilities. The key aspect of the model is its flexibility to describe the changing inflation over time and to accommodate distinct time series properties across price indices sampled at mixed frequencies. The model is estimated using Bayesian Gibbs Sampler and data on weekly commodity inflation, monthly consumer inflation, expenditures inflation, and quarterly GDP deflator inflation. The empirical results show that the weekly inflation index closely matches monthly consumer and expenditure inflation. Additionally, an alternative measure of high frequency trend inflation is proposed and estimated.

"The Credit-Card Services Augmented Divisia Monetary Aggregates," with W. Barnett, D. Leiva-Leon, and L. Su. Revised/Resubmitted. (Download)

Abstract: While credit cards provide transaction services, they have never been included in measures of the money supply.  The reason is accounting conventions, which do not permit adding liabilities to assets. However, index number theory measures service flows and is based on aggregation theory, not accounting. We derive the theory needed to measure the joint services of credit cards and money, where the transaction services of credit cards are deferred payment services not provided by money. We also propose and evaluate two aggregate measures of their joint services. One is based on microeconomic structural aggregation theory, providing an aggregated variable within the macroeconomy with a wide range of potential applications. The other is a credit-card-augmented aggregate, optimized as an indicator that captures the contributions of monetary and credit card services as a nowcasting indicator of nominal GDP.  Our structural credit-card augmented aggregates are now available monthly from the Center for Financial Stability and soon will be available to Bloomberg Terminal users.  Our indicator optimized credit-card augmented aggregates will be available from the Center for Financial Stability following completion of this research.

"Asset Prices and Optimal Monetary Policy," with S. d'Addona and V. Kakar. Submitted.(Download)

Abstract: This paper proposes a model that incorporates macro-.finance linkages in a production-based economy to study optimal monetary policy in the presence of asset price volatility, while generating realistic asset prices consistent with the historical equity premium and risk-free rate. In particular, we propose a Dynamic Stochastic General Equilibrium (DSGE) model that incorporates .financial frictions, recursive preferences, long-run productivity risk with convex adjustment costs in a production-based economy. We also consider a central bank reaction function and optimal monetary policy rules. Our main result suggests that optimal monetary policy should react to asset price misalignments over and above the inflation and output outlook in order to achieve greater macroeconomic and .nancial stability. This paper contributes to the current debate on how central bankers ought to respond to asset price volatility in the context of an overall strategy for monetary policy.

"Assessment of Hybrid Phillips Curve Specifications,' with J. Hur and I. Kim, Economics Letters, Vol. 156, 53-57, 2017. ( Download or EL) 

Abstract: Rudd and Whelan (2006) document evidence that the first-difference of inflation negatively depends on its own lag, and highlight that sticky price models emphasizing the role of firms’ forward-looking pricing behavior cannot be reconciled with the stylized fact. We show that the puzzling negative dependence of the first-difference of inflation on its own lag is consistent with the prediction of the hybrid New Keynesian Phillips Curve (NKPC) with lags of inflation, whereas, as it is argued, it is inconsistent with the prediction of both the purely forward-looking NKPC and its hybrid variant with a lag of inflation. Our theoretical results show that the negative dependence appears only when firms’ forward-looking pricing behavior is relatively more important than backward-looking behavior in determining inflation dynamics.

"Quantifying the Monetary Transmission Mechanism: A Mixed-Frequency Factor-Augmented Vector Autoregressive Regression Approach," with Z. Zhao, Working Paper, University of California Riverside.

Abstract: This paper studies the monetary transmission mechanism in the U.S. It proposes a mixed-frequency version of the factor-augmented vector autoregressive regression (FAVAR) model, which is used to construct a coincident index to measure the monetary transmission mechanism. The model divides the transmission of changes in monetary policy to the economy into three stages according to the timing and order of the impact. Indicators of each stage are measured and identified using different data frequencies: fast-moving variables (stage 1, asset returns at the weekly frequency), intermediate moving variables (stage 2, credit market data at the monthly frequency), and slow-moving variables (stage 3, macroeconomic variables at the quarterly frequency). The resulting coincident index exhibits leading signal for all recessions in the sample period and provides implications on the dynamics of the monetary transmission mechanism. The proposed coincident index also indicates that monetary transmission mechanism is changing over time.

"Monetary Policy Regimes and the Stock Market," with Chuanlei Sun, in Business Cycles in Economics: Types, Challenges and Impacts on Monetary Policies. Ed. Jason Hsu, Nova Science Chapter 6, 87-1116, 2014. (Download )

Abstract: This paper studies the relationship between monetary policy and the stock market across business cycle recessions and expansions. A nonlinear two-state Markov switching model is used to obtain regimes in interest rate cycles and in stock market cycles. The model identifies tight and loose monetary policy phases, and bear and bull stock market regimes. Turning points for each cycle are established and their lead-lag relationship is examined and contrasted with NBER recessions. We also examine the linear predictive relationship between interest rates and stock returns. The results indicate that there are strong linkages among interest rate cycles, stock market cycles, and business cycles. We find that, generally, the stock market enters a bear market phase at around the same time as monetary policy enters a tight phase, which is associated with the onset of economic recessions. We also find that future stock returns are substantially impacted by information on monetary policy regimes (i.e. loose or tight). On the other hand, interest rates are most influenced by future inflation and recessions rather than by changes in the stock market. 

"Nonlinear Relationship Between Permanent and Transitory Components of Monetary Aggregates and the Economy," B. Jones, R. Anderson, Econometrics Review, vol. 34:1-2, 228-254, 2014. (Working Paper or ER)

Abstract: This paper uses several methods to study the interrelationship among Divisia monetary aggregates, prices, and income, allowing for nonstationary, nonlinearities, asymmetries, and time-varying relationships among the series. We propose a multivariate regime switching unobserved components model to obtain transitory and permanent components for each series, allowing for potential recurrent and structural changes in their dynamics. Each component follows distinct two-state Markov processes representing low or high phases. Since the lead-lag relationship between the phases can vary over time, rather than pre-imposing a structure to their linkages, the proposed flexible framework enables us to study their specific lead-lag relationship over each one of their cycles and over each U.S. recession in the last 40 years. The decomposition of the series into permanent and transitory components reveals striking results. First, we find a strong nonlinear association between the components of money and prices—all low phases of the transitory component of prices were preceded by tight transitory and permanent money phases. We also find that most recessions were preceded by tight money phases (its cyclical and permanent components) and high transitory price phases (with the exception of the 2001 and 2009-2010 recessions). In addition, all recessions were associated with a decrease in transitory and permanent income.  

"Measurement Error in Monetary Aggregates: A Markov Switching Factor Approach,” with William Barnett and Heather Tierney, Macroeconomic Dynamics, Vol 13, 381-412, 2009. (Download Repec or MeasErrMoney) 

Abstract: This paper compares the different dynamics of the simple sum monetary aggregates and the Divisia monetary aggregate indexes over time, over the business cycle, and across high and low inflation and interest rate phases. Although traditional comparisons of the series sometimes suggest that simple sum and Divisia monetary aggregates share similar dynamics, there are important differences during certain periods, such as around turning points. These differences cannot be evaluated by their average behavior. We use a factor model with regime switching. The model separates out the common movements underlying the monetary aggregate indexes, summarized in the dynamic factor, from individual variations in each individual series, captured by the idiosyncratic terms. The idiosyncratic terms and the measurement errors reveal where the monetary indexes differ. We find several new results. In general, the idiosyncratic terms for both the simple sum aggregates and the Divisia indexes display a business cycle pattern, especially since 1980. They generally rise around the end of high interest rate phases — a couple of quarters before the beginning of recessions — and fall during recessions to subsequently converge to their average in the beginning of expansions. We find that the major differences between the simple sum aggregates and Divisia indexes occur around the beginnings and ends of economic recessions, and during some high interestrate phases. We note the inferences' policy relevance, which is particularly dramatic at the broadest (M3) level of aggregation. Indeed, as Belongia (1996) has observed in this regard, "measurement matters."

"How Better Monetary Statistics Could Have Signaled the Financial Crisis," with William Barnett, Journal of Econometrics,Vol. 161, No. 1, 6-23, 2011. (Download Repec or JE) 

Abstract: This paper explores the disconnect of Federal Reserve data from index number theory. A consequence could have been the decreased-systemic-risk misperceptions that contributed to excess risk-taking prior to the housing bust. We find that most recessions in the past 50 years were preceded by more contractionary monetary policy than indicated by simple-sum monetary data. Divisia monetary aggregate growth rates were generally lower than simple-sum aggregate growth rates in the period preceding the Great Moderation, and higher since the mid 1980s. Monetary policy was more contractionary than likely intended before the 2001 recession and more expansionary than likely intended during the subsequent recovery.

“Real Time Changes in Monetary Policy: A Nonparametric Approach,” in Proceedings of the 27th International Forecasting Symposium, New York, June 2007. (Download).

Abstract: This paper investigates potential changes in monetary policy over the last decades using a nonparametric vector autoregression model. In the proposed model, the conditional mean and variance are time-dependent and estimated using a nonparametric local linear method, which allows for different forms of nonlinearity, conditional heteroskedasticity, and non-normality. Our results suggest that there have been gradual and abrupt changes in the variances of shocks, in the monetary transmission mechanism, and in the Fed’s reaction function. The response of output was strongest during Volcker’s disinflationary period and has since been slowly decreasing over time. There have been some abrupt changes in the response of inflation, especially in the early 1980s, but we can not conclude that it is weaker now than in previous periods. Finally, we find significant evidence that policy was passive during some parts of Burn’s period, and active during Volcker’s disinflationary period and Greenspan’s period. However, we find that the uncovered behavior of the parameters is more complex than general conclusions suggest, since they display considerable nonlinearities over time. A particular appeal of the recursive estimation of the proposed VAR-ARCH is the detection of discrete local deviations as well as more gradual ones, without smoothing the timing or magnitude of the changes.

"Forecasting Output," with Simon Potter, in Handbook of Economic Forecasting, vol. 2, ed. A. Timmermann and G. Elliott, Elsevier/North Holland, 1-56, 2013. (Download Working Paper or Handbook) 

Abstract: This chapter surveys the recent literature on output forecasting, and examines the real-time forecasting ability of several models for U.S. output growth. In particular, it evaluates the accuracy of short-term forecasts of linear and nonlinear structural and reduced-form models, and judgmental forecasts of output growth. Our emphasis is on using solely the information that was available at the time the forecast was being made, in order to reproduce the forecasting problem facing forecasters in real-time. We find that there is a large difference in forecast performance across business cycle phases. In particular, it is much harder to forecast output growth during recessions than during expansions. Simple linear and nonlinear autoregressive models have the best accuracy in forecasting output growth during expansions, although the dynamic stochastic general equilibrium model and the vector autoregressive model with financial variables do relatively well. On the other hand, we find that most models do poorly in forecasting output growth during recessions. The autoregressive model based on the nonlinear dynamic factor model that takes into account asymmetries between expansions and recessions displays the best real time forecast accuracy during recessions. Even though the Blue Chip forecasts are comparable, the dynamic factor Markov switching model has better accuracy, particularly with respect to the timing and depth of output fall during recessions in real time. The results suggest that there are large gains in considering separate forecasting models for normal times and models especially designed for periods of abrupt changes, such as during recessions and financial crises.

"Real-Time Nowcasting of Nominal GDP with Structural Breaks", with W. Barnett and D. Leiva-Leon, Journal of Econometrics. Vol. 191, April, 312-324, 2016. (Download or JE) 

Abstract: This paper provides early assessments of current U.S. Nominal GDP growth, which has been considered as a potential new monetary policy target. The nowcasts are computed using the exact amount of information that policy makers have available at the time predictions are made. However, real time information arrives at different frequencies and asynchronously, which poses the challenge of mixed frequencies, missing data, and ragged edges. This paper proposes a multivariate state space model that not only takes into account asynchronous information inflow it also allows for potential parameter instability (DYMIBREAK). We use small scale confirmatory factor analysis in which the candidate variables are selected based on their ability to forecast nominal GDP. The model is fully estimated in one step using a nonlinear Kalman filter, which is applied to obtain simultaneously both optimal inferences on the dynamic factor and parameters. Differently from principal component analysis, the proposed factor model captures the comovement rather than the variance underlying the variables. We compare the predictive ability of the model with other univariate and multivariate specifications. The results indicate that the proposed model containing information on real economic activity, inflation, interest rates, and Divisia monetary aggregates produces the most accurate real time nowcasts of nominal GDP growth. 

"Mortgage Default Risk: New Evidence from Internet Search Queries", with S. Gabriel and C. Lutz, Journal of Urban Economics. Vol. 96, November, 91-111, 2016. (Download) 

Abstract: We use Google search query data to develop a broad-based and real-time index of mortgage default risk. Unlike established indicators, our Mortgage Default Risk Index (MDRI) directly reflects households’concerns regarding their risk of mortgage default. The MDRI predicts housing returns, mortgage delinquency indicators, and subprime credit default swaps. These results persist both in- and out-of-sample and at multiple data frequencies. Together, research findings suggest internet search queries yield valuable new insights into household mortgage default risk.

"Forecasting Recessions Using the Yield Curve,” with S. Potter, Journal of Forecasting, 24, 2, 77-103, 2005. (Download  Repec or ForecYield)

Abstract: We compare forecasts of recessions using four different specifications of the probit model: a time-invariant conditionally independent version, a business cycle specific conditionally independent model, a time-invariant probit with autocorrelated errors, and a business cycle specific probit with autocorrelated errors. ; The more sophisticated versions of the model take into account some of the potential underlying causes of the documented predictive instability of the yield curve. We find strong evidence in favor of the more sophisticated specification, which allows for multiple breakpoints across business cycles and autocorrelation. We also develop a new approach to the construction of real time forecasting of recession probabilities.

“Predicting Recessions: Evidence from the Yield Curve in the Presence of Structural Breaks,” with S. Potter, Economics Letters, Vol. 77, No. 2, 245-253, 2002. (Download Working Paper, Repec or PredicBreaks)

Abstract: We use a probit model of the term structure to examine the stability of recession forecasts under the presence of a structural break. We find strong evidence of a break, but with very uncertain location, which affects considerably recession predictions.

"Nonstationarities and Markov Switching Models," with Y. Su, in Recent Advances in Estimating Nonlinear Models, Springer, 123-148, 2013. (Download) 

Abstract: This paper proposes a flexible model that allows for recent changes observed in the US business cycle in the last six decades. It proposes a Markov switching model with three Markov processes to characterize the dynamics of US output fluctuations. We consider the possibility that both the mean and the variance of growth rates of real GDP can have short run fluctuations in addition to the possibility of a long run permanent break. We find that, differently from several alternative specifications in the literature, the proposed flexible framework successfully represents all business cycle phases, including the Great Recession. In addition, we find that the volatility of US output fluctuations has both a long run pattern, characterized by a structural break in 1984, as well as business cycle dynamics, in which periods of high uncertainty are associated with NBER recessions.

"Predicting Recessions in Brazil," with I. Morais, Latin American Meetings of the Econometric Society and Latin American and Caribbean Economic Association Meetings, Rio de Janeiro, Brazil, 2008. (Download )

Abstract: This paper constructs leading indicators that anticipate turning points of the Brazilian business cycle. We propose a time-varying autoregressive probit model, which is composed of several economic series that display predictive power to anticipate the beginning or end of recessions. The Brazilian economy is characterized by several different policy regimes and instabilities that have potentially engendered breaks in its dynamics.The extended probit model is especially suited for this economy, since it takes into consideration parameter change across each cycle in addition to phase duration in the estimation of recession probabilities. We find that the extended probit model exhibits superior predictive performance over the standard probit model in several dimensions both in-sample and in an out-of-sample real time exercise.

"vOILatility: Forecasting Oil Prices under Uncertainty," mimeo, University of California Riverside, 2018.

Abstract: "Historically, oil prices are subject to sudden jumps as well as smoother changes due to changes in supply and demand. Although linear models may capture some of the dynamics in between the jumps in-sample, they fail to represent and predict nonlinearities underlying the market out-of-sample, real time. Some of the abrupt changes in oil price dynamics were due to OPEC decisions in the 1970s-1980s. Recent developments such as shifts to new technology or cooperation of Russia and OPEC can potential engender new structural breaks in the oil market dynamics, with the possibility of markedly different results in out of sample real time forecasts. Models and methods that take into account instability/breaks might substantially improve forecasts. Nonlinear models reveal additional information compared to frameworks that take into account only average linear effect of one series on another. This paper proposes a model specifically designed to forecast oil prices taking into account potential nonlinearities and nonstationarities. The autoregressive multivariate mixed frequency model has probabilities of structural breaks in the mean and volatility of oil price as a function of several variables including: indicators of potential sudden changes in oil supply/price (news on OPEC, Russia’s oil policy and changes in inventories), indicator of economywide demand and oil consumption in the largest consumers and importers of oil, indicator of recent technology shifts, indicator of changes in risk. Preliminary results indicate that the model provides accurate real-time forecast of oil price remarkably superior to forecasts from alternative linear frameworks."

Predicting Default Risk of Small Business Loans with Big Data" with Hien Nguyen.

Abstract: This paper applies machine learning methods on big data sets on firm characteristics, bank balance sheets and loan information to study the default risk of loans to small businesses under the Small Business Administration (SBA) 7(a) loan guarantee program. We find that loan age is the most important predictor of loan default for all periods: before, during and after the 2008 financial crisis. Bank balance sheet variables-bank capital and bank assets-follow loan age in ranking for before crisis and during crisis periods. However, after the crisis firm characteristics, earnings-to-assets and debt-to-assets, surpass bank variables to be the most important predictors after loan age. The results show that due to major reforms in the banking industry after the most recent financial crisis, the quality of bank balance sheets is improved. Bank characteristics, therefore, are less crucial in determining the quality of loans after 2008.

"New Class of Volatility –SVR and LSTM Models," with Igor Morais.

Abstract: The use of neural networks and machine learning for solving complex nonlinear problems has become more promising with greater availability of data and powerful algorithms. One of these options is the use of Deep Learning and SVR-Support Vector Regression for analysis of financial market time series. This paper makes use of these two techniques to estimate the daily volatility of the S & P500, comparing its prediction results with the traditional deterministic models of the GARCH family and of stochastic volatility. The major contributions of this study are related to the applied methods, with emphasis on the implementation of different kernel functions in SVR models and of different activation functions in the use of LSTM in Deep Learning. The results indicate that even in the absence of information on parameters that are obtained with parametric models, the fact is that these new techniques are more efficient in predicting volatility for different crisis scenarios.

“The Brazilian Business Cycle and Growth Cycle,” Brazilian Economic Journal (Revista Brasileira de Economia), Vol. 56 nº 1, 75-106, 2002. (Download BrazilBCGC or Repec )

Abstract: This paper uses several procedures to date and analyses the Brazilian business and growth cycles. In particular, a Markov switching model is fitted to quarterly and annual real production data. The smoothed probabilities of the Markov states are used as predictive rules to define different phases of cyclical fluctuations of real Brazilian economic activity. The results are compared with different non-parametric rules. All methods implemented yield similar dating and reveal asymmetries across the different states of the Brazilian business and growth cycles, in which slowdowns and recessions are short and abrupt, while high growth phases and expansions are longer and less steep. The resulting dating of the Brazilian economic cycles can be used as a reference point for construction and evaluation of the predictive performance of coincident, leading, or lagging indicators of economic activity. In addition, the filtered probabilities obtained from the Markov switching model allow early recognition of the transition to a new business cycle phase, which can be used, for example, for evaluation of the adequate strength and timing of countercyclical policies, for reassessment of projected sales or profits by businesses and investors, or for monitoring of inflation pressures.

"Leading Indicators of the Capital Goods Industry in Brazil," with Igor Morais, Brazilian Review of Econometrics, Vol. 31, No. 1, 137-171, 2011.  Recipient of Honor Award from the National Confederation of Industries of Brazil  - Confederacao Nacional das Industrias, CNI (Download LeadingCapital)  

Abstract: The goal of this paper is to build leading indicators to predict the capital goods business cycle in Brazil. We propose a probit model with autoregressive dynamics consisting of series with predictive power to anticipate contractions in this sector. The model is especially suitable for this sector as it includes information about the characteristics of cycle phases and their duration in the estimation of recession probability, and adapts well to the high volatility of the capital goods business cycle. The results indicate that the dynamic probit model has a better forecasting performance than the simple probit model in several aspects, both within and out of sample and in real time. 

“A Monthly Indicator of Brazilian GDP,” in Brazilian Review of Econometrics Vol. 21, No. 1, 1-15, 2001 – Recipient of Honor Award from the Brazilian Econometric Society. (Download BrazilIndicatorGDP)

Abstract: This paper constructs an indicator of Brazilian GDP at the monthly frequency. The peculiar instability and abrupt changes of regimes in the dynamic behavior of the Brazilian business cycle were explicitly modeled within nonlinear frameworks. In particular, a Markov switching dynamic factor model was used to combine several macroeconomic variables that display simultaneous comovements with aggregate economic activity. The model generates as output a monthly indicator of the Brazilian GDP and real time probabilities of the current phase of the Brazilian business cycle. The monthly indicator shows a remarkable historical conformity with cyclical movements of GDP. In addition, the estimated filtered probabilities predict all recessions in sample and out-of-sample. The ability of the indicator in linear forecasting growth rates of GDP is also examined. The estimated indicator displays a better in-sample and out-of-sample predictive performance in forecasting growth rates of real GDP, compared to a linear autoregressive model for GDP. These results suggest that the estimated monthly indicator can be used to forecast GDP and to monitor the state of the Brazilian economy in real time.

“Leading Indicators of the Brazilian Inflation,” Economic Research and Planning (Pesquisa e Planejamento Economico), Vol. 31, 1, 323-354, 2001. (Download Repec - English or BrazilLeadingInflation-Portuguese).

Abstract: The goal of this project is to construct leading indicators that anticipate inflation cycle turning points on a real time monitoring basis. As a first step, turning points of the IPCA inflation are determined using a periodic stochastic Markov switching model. These turning points are the event timing that the leading indicators should anticipate. A dynamic factor model is then used to extract common cyclical movements in a set of variables that display predictive content for inflation. The leading indicators are designed to serve as practical tools to assist real-time monitoring of monetary policy on a month-to-month basis. Thus, the indicators are built and ranked according to their out-of-sample forecasting performance. The leading indicators are found to be an informative tool for signaling future phases of the inflation cycle out-of-sample, even in real time when only preliminary and unrevised data are available.

"Trend-Cycle Decomposition of the Brazilian GDP: New Facts for the period between 1947 and 2012."  with L.S. Lopes and J.E. Lima. Proceedings of the 45th National Economic Meeting (ANPEC). 2017, Natal, RN, Brazil. (Download)

Abstract: We provide information about the Brazilian business cycle from 1947 to 2012, by suggesting a new method that averages over a variety of HP-filters and creates a set of facts which are more related to CODACE dates of expansions and recessions. The main findings are that Brazilian business cycle is asymmetric, with expansions lasting longer than recessions; the long term trend presented a noticeable flatter slope after the 1980s, thus, real long-term growth rate decreased by 50%, from 8% per year, between 1947 and 1980, to 4% per year after that; and, output volatility decreased after 1996-1997, when a statistically significant structural break occurred.

“Leading Indicators of Recession for the Brazilian Economy,” with Jose A. B. da Silva, in Proceedings of the XXVI Meetings of the Brazilian Econometric Society, 2004, Joao Pessoa, Paraiba, Brazil. (Portuguese Download)

Abstract: O objetivo deste artigo é a construção de indicadores que antecipem o início dos ciclos econômicos brasileiros. Como primeiro passo, os pontos de mudanças dos ciclos econômicos foram determinados usando-se um modelo de fator dinâmico com mudança de Markov. Estes pontos são os eventos que os indicadores devem antecipar. Os modelos probit são então utilizados para extrair indicadores antecedentes de recessão de um grupo de variáveis que possuem poder de previsão com relação ao PIB. Os indicadores antecedentes são construídos para servir como um instrumento prático de auxílio à política monetária com base mensal. Assim, os indicadores são construídos e classificados de acordo com o seu poder de previsão dentro e fora de amostra. Os resultados empíricos revelam que os indicadores antecedentes resultantes constituem-se em um instrumento informativo para sinalizar fases futura do ciclo econômico brasileiro.

"The End of the Brazilian Big Inflation: Lessons to Monetary Policy from a Standard New Keynesian Model," with L.S. Lopes and J.E. Lima, Empirical Economics, 2017, 1-31. (Download) 

Abstract: The paper analyzes economic stabilization in Brazil in the context of a New Keynesian model estimated with Bayesian techniques. Dataset covers the period 1975–2012. Our methodology is based on tests for multiple structural breaks at unknown dates and counterfactual exercises. The results show that inflation and output volatility present an inverted U-shape pattern, peaking at the 1985–1994 sample. Changes in the monetary policy stance and milder shocks accounted for the reduced inflationary volatility (about 50% each, in some specifications). However, some assumptions indicated that a sharp decline in the Phillips curve slope was also important for controlling inflation. Concerning to output, the sole explanation for its volatility fall seemed to be smaller shocks. Therefore, we conclude that a mix of the “good luck” and “good policy” hypotheses mainly originated the current period of increased stability in the country.

"Chronology and Prediction of Business Cycles in Minas Gerais" ("Cronologia e Previsão de Ciclos Econômicos de Minas Gerais"). Cadernos BDMG. February, 7-44, 2013. (Portuguese Download)

Abstract: Esse artigo utiliza modelos probabilísticos de fronteira para obter uma cronologia dos ciclos econômicos de Minas Gerais, e para construir indicadores coincidentes e antecedentes da economia mineira. O modelo de fator dinâmico com mudanças de regime de Markov é utilizado para representar os movimentos cíclicos e determinar o começo e fim das fases de recessão e expansão em Minas Gerais. Esse modelo gera um indicador coincidente da economia mineira e probabilidades de recessões e expansões, as quais podem ser utilizadas para analisar e monitorar as diferentes características das fases dos ciclos econômicos. O artigo também propõe a construção de indicadores antecedentes, utilizando um modelo probit dinâmico. O modelo gera não somente probabilidades de recessões futuras, como também probabilidades de continuação ou interrupção de uma fase do ciclo. Esse enfoque também permite avaliação do grau de incerteza ou precisão dessas probabilidades. O indicador coincidente estimado apresenta uma conformidade histórica notável com movimentos cíclicos do PIB mineiro, com relação a sua volatilidade, duração das fases e timing de seus pontos de mudança. Quanto aos indicadores antecedentes, as probabilidades obtidas do modelo probit prevêem todas as recessões na amostra considerada, com uma antecedência de um a dois trimestres. Além disso, a identificação de recessões futuras é nítida uma vez que asprobabilidades aumentam acima de 80% antes de todas as recessões mineiras, e não produzem sinais falsos. O modelo probit também é estimado em tempo real fora de amostra para o período recente.

"Forecasting Brazilian Output and its Turning Points in the Presence of Breaks: A Comparison of Linear and Nonlinear Models,” with E. Lima, G. Domingues, and B. Vasquez, Economic Studies (Estudos Economicos), Vol. 36, No.1, 5-46, 2006. (Download BR_TPBreak) 

Abstract: This paper compares the forecasting performance of linear and nonlinear models under the presence of structural breaks for the Brazilian real GDP growth. The Markov-switching models proposed by Hamilton (1989) and its generalized version proposed by Lam (1991) are applied to quarterly GDP from 1975:1 to 2000:2 allowing for breaks at the Collor Plans. The probabilities of recessions are used to analyze the Brazilian business cycle. The ability of each model in forecasting out-of-sample the growth rates of GDP is examined. The forecasting ability of the two models is also compared with linear specifications. The authors find that nonlinear models display the best forecasting performance and that specifications including the presence of structural breaks are important in obtaining a representation of the Brazilian business cycle.