End Of Recession Research Articles

Economic Recession Prediction Using Deep Neural Network

We investigate the effectiveness of different machine learning methodologies in predicting economic cycles. We identify the deep learning methodology of BiLSTM with autoencoder as the most accurate model to forecast the beginning and end of economic recessions in the United States. We adopt commonly available macro and market-condition features to compare the ability of different machine learning models to generate good predictions both in-sample and out-of-sample. The proposed model is flexible and dynamic when both predictive variables and model coefficients vary over time. It provided good out-of-sample predictions for the past two recessions and early warning about the COVID-19 recession.

Business cycle duration dependence and foreign recessions

AbstractThis study revisits business cycle duration dependence in G7 countries by controlling for foreign recessions, defined as the number of other G7 countries in a recession. Estimates from regime switching logit models show that the monthly likelihood of ending an expansion roughly doubles for every extra G7 country in recession, but the end of foreign recessions do not affect the ending of recessions. They also show that recessions are duration dependent in all G7 countries, but expansions only in the United States and Germany. The economic importance of foreign recessions and duration in driving business cycle phase changes vary across countries.

The rise of US earnings inequality: Does the cycle drive the trend?

We document that declining hours worked are the primary driver of widening inequality in the bottom half of the male labor earnings distribution in the United States over the past 52 years. This decline in hours is heavily concentrated in recessions: hours and earnings at the bottom fall sharply in recessions and do not fully recover in subsequent expansions. Motivated by this evidence, we build a structural model to explore the possibility that recessions cause persistent increases in inequality; that is, that the cycle drives the trend. The model features skill-biased technical change, which implies a trend decline in low-skill wages relative to the value of non-market activities. With this adverse trend in the background, recessions imply a potential double-whammy for low skilled men. This group is disproportionately likely to experience unemployment, which further reduces skills and potential earnings via a scarring effect. As unemployed low skilled men give up job search, recessions generate surges in non-participation. Because non-participation is highly persistent, earnings inequality remains elevated long after the recession ends.

The Rise of Us Earnings Inequality: Does the Cycle Drive the Trend?

We document that declining hours worked are the primary driver of widening inequality in the bottom half of the male labor earnings distribution in the United States over the past 52 years. This decline in hours is heavily concentrated in recessions: hours and earnings at the bottom fall sharply in recessions and do not fully recover in subsequent expansions. Motivated by this evidence, we build a structural model to explore the possibility that recessions cause persistent increases in inequality; that is, that the cycle drives the trend. The model features skill-biased technical change, which implies a trend decline in low-skill wages relative to the value of non-market activities. With this adverse trend in the background, recessions imply a potential double-whammy for low skilled men. This group is disproportionately likely to experience unemployment, which further reduces skills and potential earnings via a scarring effect. As unemployed low skilled men give up job search, recessions generate surges in non-participation. Because non-participation is highly persistent, earnings inequality remains elevated long after the recession ends. Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

Machine Learning for Recession Prediction and Dynamic Asset Allocation

The authors introduce a novel application of support vector machines (SVM), an important machine learning algorithm, to determine the beginning and end of recessions in real time. Nowcasting, forecasting a condition in the present time because the full information will not be available until later, is key for recessions, which are only determined months after the fact. The authors show that SVM has excellent predictive performance for this task, capturing all six recessions from 1973 to 2018 and providing the signal with minimal delay. The authors take advantage of the timeliness of SVM signals to test dynamic asset allocation between stocks and bonds. A dynamic risk budgeting approach using SVM outputs appears superior to an equal-risk contribution portfolio, improving the average returns by 85 bps per annum without increased tail risk. TOPICS:Big data/machine learning, financial crises and financial market history, portfolio construction, tail risks

A Great Year for the Industrial Laser Business in the USA

AbstractThe global industrial laser business in 2017 experienced a record year in terms of revenues, topping the 4.3 billion dollars mark. From the end of the Great Recession (officially June 2009), according to Industrial Laser Solutions (ILS), market revenues grew at a 14.95 % CAGR for a 250 % increase over end of recession values (Fig. 1). Leading this charge were fiber laser revenues which rose at an exceptional 37.38 % CAGR to more than $2 B in 2017.

Can stock market liquidity and volatility predict business cycles?

PurposeThe purpose of this paper is to predict real gross domestic product (GDP) growth and business cycles by using information from both liquidity and volatility measures.Design/methodology/approachThe paper estimates liquidity and volatility measures from over 5,000 NYSE rms and extracts a common factor, which the paper calls uncertainty. In-sample and out-of-sample forecasting tests are used to determine the ability of the uncertainty factor to predict growth in real GDP, industrial production, consumer price index, real consumption and changes in real investment.FindingsThe paper finds that on average, positive shocks to the uncertainty factor occur in the quarters preceding and at the beginning of a recession. During the quarters toward the end of recessions, there are negative shocks to uncertainty on average.Originality/valuePrevious research has explored using either liquidity or volatility to forecast economic activity. The paper bridges the two branches of research and finds a link to real GDP growth and business cycles.

Fiscal Multipliers at the Zero Lower Bound: The Role of Policy Inertia

AbstractThe presence of the lagged shadow policy rate in the interest rate feedback rule reduces the government spending multiplier nontrivially when the policy rate is constrained at the zero lower bound (ZLB). In the economy with policy inertia, increased inflation and output due to higher government spending during a recession speed up the return of the policy rate to the steady state after the recession ends, which in turn damps the expansionary effects of the government spending during the recession via expectations. In our baseline experiment intended to capture the effectiveness of the American Recovery and Reinvestment Act of 2009, the output multiplier at the ZLB is 1.9 when the weight on the lagged shadow rate is zero, and 0.5 when the weight is 0.85.

Fiscal Multipliers at the Zero Lower Bound: The Role of Policy Inertia

The presence of the lagged shadow policy rate in the interest rate feedback rule reduces the government spending multiplier nontrivially when the policy rate is constrained at the zero lower bound (ZLB). In the economy with policy inertia, increased inflation and output due to higher government spending during a recession speed up the return of the policy rate to the steady state after the recession ends. This in turn dampens the expansionary effects of the government spending during the recession via expectations. In our baseline calibration, the output multiplier at the ZLB is 2.5 when the weight on the lagged shadow rate is zero, and 1.1 when the weight is 0.9.

Nowcasting Recessions Using the SVM Machine Learning Algorithm

We introduce a novel application of support vector machines (SVM), an important machine learning algorithm, to determine the beginning and end of recessions in real time. Nowcasting, forecasting a condition in the present time because the full information will not be available until later, is key for recessions, which are only determined months after the fact. We show that SVM has excellent predictive performance for this task, capturing all six recessions from 1973 to 2018 and providing the signal with minimal delay. We take advantage of the timeliness of SVM signals to test dynamic asset allocation between stocks and bonds. A dynamic risk budgeting approach using SVM outputs appears superior to an equal-risk contribution portfolio, improving the average returns by 85 bps per annum without increased tail risk.

World trade recovering but still reasons for caution

▀ World trade has picked up in recent months, expanding at the fastest pace in six years in the first quarter, with the rise fairly evenly split between advanced and emerging markets. Stronger activity in China and a broader upturn in global investment have been key factors. But there are still reasons for caution. Although the ‘cyclical’ element in world trade is improving, the ‘trend’ element is not thanks to changes in supply chains and a lack of trade liberalisation. ▀ World trade growth looks set to reach about a 4% annual rate in Q1 2017, the fastest pace since 2011. Alternative freight‐based indicators confirm the upturn. This suggests some modest near‐term upside risk to our world growth forecasts. ▀ Recent growth has been evenly split between advanced countries and emerging markets (EM). In EM, the end of deep recessions in Russia and Brazil and an upturn in China have been key factors. China directly added 0.5 percentage points to annual world trade growth over recent months and firmer growth there has also pushed up commodity prices and the spending power and imports of commodity exporters. ▀ Another important positive factor is an improvement in investment, which is a trade‐intensive element of world GDP. Rising capital goods imports across a range of countries suggest the drag on world trade from weak investment is fading. ▀ The decline in the ratio of world trade growth to world GDP growth over recent years has both cyclical and structural elements. But while the cyclical component now seems to be improving, there is little evidence that the structural part – responsible for between a half and two‐thirds of the recent decline – is doing likewise. ▀ Key factors behind the structural decline in world trade growth are changes in supply chains and a lack of trade liberalisation/protectionism. Both are likely to remain a drag over the coming years. Meanwhile, a levelling‐off of growth in China and drop back in commodity prices could curb the recent cyclical uptick.

Gender Differentials in Unemployment Ins and Outs during the Great Recession in Spain

The Great Recession had led to gender convergence in unemployment rates. In this paper we seek its sources to assess whether this convergence will remain once the Great Recession ends. We use Social Security records to study the determinants of unemployment ins and outs for men and women separately, over the course of a whole business cycle, i.e. 2000–2013. We focus on Spain—a country hit hard by unemployment increases in downturns. We find that unemployment outs are crucial in understanding changes in unemployment rates in Spain as well as to understand the gender convergence in unemployment rates. Among the determinants of the large drop in unemployment outs, lack of demand and negative state dependence emerge as key sources, which affect men more negatively than women. In a scenario of upcoming recovery, unemployment outs will increase for short-term unemployed, particularly for males. On the contrary, both male and female long-term unemployed workers will face enormous difficulties to access a job, as the job access rates for long-term unemployed is not sensitive to the economic cycle. Hence, we expect that the gender convergence in unemployment rates will persist only when considering the long-term unemployed.

The long-term interest rates correlations: a new indicator predicting recession

Purpose Existing research has theoretically modeled conditional correlations between the long-term interest rates as a function of macroeconomic variable. In line with it, the purpose of this paper is to explore whether conditional correlations can be a new signal to predict recessions. Furthermore, this paper also tries to investigate among the four factors – the time difference of the beginning and the end of recessions, financial integration (FI), and trade integration (TI) – which factors drive the direction of change in conditional correlations. Finally, this paper is to explain the implication for Korea trade. Design/methodology/approach This study uses a probit regression model for 33 country during the period from 1972 to 2015. To measure the time-varying interest rates conditional correlations, a VAR(1)-DBEKK-GARCH(1,1) model is adopted due to its statistical advantages. Furthermore, the authors also construct the four measures – time difference of the beginning of recessions (BEG), time difference of the end of recessions (END), FI, and TI. The authors first study the predictive power of correlations in both in and out-samples test, and study which factors determine the different behavior of interest rate co-movements using the four measures. Findings The empirical results show that the conditional correlations between the long-term interest rates of the USA and individual countries contain information about recessions a few quarters ahead which term spreads of neither individual countries nor the USA conveyed in. However, there is a heterogeneity of the significance and direction of interest rate correlations. A further research reveals that especially the heterogeneous degree of TI leads to the different overlapped recession period of individual countries with the USA, resulting in heterogeneous behavior of interest rates among countries. Research limitations/implications As a limitation of this paper, the forecasting power of interest rate correlations is not always significant in all countries. Despite this, the study has a profound implication that for those countries where the US accounts for the high proportion of trade, increase in conditional correlations can be a signal for future recessions. Especially, given a considerable portion of trade in GDP and the more sensitive trade activity of Korea to a contagious recession than a domestic recession, the conditional correlation measure is particularly useful for Korean policy makers. Originality/value Although many papers model interest rate co-movement as a function of macroeconomic condition, this paper provides the first evidence to show interest rate co-movement precede the macro shocks empirically. Furthermore, this paper determines the precise channel through which TI affects the time-varying behavior of interest rate co-movements before recessions.

Forward Guidance and Credible Monetary Policy

The effectiveness of forward guidance depends crucially on the credibility of the central bank. Instead of assuming full commitment or full discretion of the central bank, this paper sheds light on the best credible forward guidance policy a central bank can offer by solving for the whole set of sustainable sequential equilibria (SSE) in a standard New Keynesian model with the occasionally binding constraint of a nominal interest rate of zero. The inflation bias resulting from eliminating price distortion under the best SSE is much smaller than the bias under full discretion. In the presence of the zero lower bound (ZLB), while the full commitment solution implies that forward guidance has a longer duration of the nominal rate at zero bound followed by a quick revert-to-normal path, the best SSE features low but non-zero rates, and even a prolonged period of higher inflation after the recession ends. To make forward guidance credible and as stimulative as possible, the adjustment of policy rates, and in particular, the rise of policy rates, should be smooth and gradual. In other words, policy normalization does not mean non-accommodation. Ramsey equilibria are not generally implementable.

Fiscal Multipliers at the Zero Lower Bound: The Role of Policy Inertia

The presence of the lagged shadow policy rate in the interest rate feedback rule reduces the government spending multiplier nontrivially when the policy rate is constrained at the zero lower bound (ZLB). In the economy with policy inertia, increased inflation and output due to higher government spending during a recession speed up the return of the policy rate to the steady state after the recession ends. This in turn dampens the expansionary effects of the government spending during the recession via expectations. In our baseline calibration, the output multiplier at the ZLB is 2.5 when the weight on the lagged shadow rate is zero, and 1.1 when the weight is 0.9.

How Different Is This Time? Investigating the Declines of the U.S. Real GDP During the Great Recession

This paper investigates to what extent the declines of the U.S. real GDP are permanent during the Great Recession. Our trend-cycle decomposition indicates that the Great Recession is associated with a substantial permanent loss in the real GDP manifested in decelerated mean growth rate and lowered trend level. However, the cyclical component played a relatively minor role comparing to other post-war recessions. If the potential output grew at its pre-crisis rate, it could have been 4.5% higher at the end of the Great Recession.

PERSONAL-BANKRUPTCY CYCLES

This paper estimates the dynamics of the personal-bankruptcy rate over the business cycle by exploiting large cross-state variation. We find that bankruptcy rates are significantly above trend during a recession and rise as a recession persists. After a recession ends, there is a hangover in which bankruptcy rates begin to fall but remain above trend for several more quarters. Recovery periods see a strong bounce-back effect, with bankruptcy rates significantly below trend for several quarters. Despite the significant increases in bankruptcies during recessions, the largest contributor to rising bankruptcies during these periods has tended to be the longstanding upward trend.

A Dynamic Factor Model of the Yield Curve as a Predictor of the Economy

In this paper, we propose an econometric model of the joint dynamic relationship between the yield curve and the economy to predict business cycles. We examine the predictive value of the yield curve to forecast future economic growth as well as the beginning and end of economic recessions at the monthly frequency. The proposed nonlinear multivariate dynamic factor model takes into account not only the popular term spread but also information extracted from the level and curvature of the yield curve and from macroeconomic variables. The nonlinear model is used to investigate the interrelationship between the phases of the bond market and of the business cycle. The results indicate a strong interrelation between these two sectors. The proposed factor model of the yield curve exhibits substantial incremental predictive value compared to several alternative specifications. This result holds in-sample and out-of-sample, using revised or real time unrevised data.

A Dynamic Factor Model of the Yield Curve as a Predictor of the Economy

In this paper, we propose an econometric model of the joint dynamic relationship between the yield curve and the economy to predict business cycles. We examine the predictive value of the yield curve to forecast future economic growth as well as the beginning and end of economic recessions at the monthly frequency. The proposed nonlinear multivariate dynamic factor model takes into account not only the popular term spread but also information extracted from the level and curvature of the yield curve and from macroeconomic variables. The nonlinear model is used to investigate the interrelationship between the phases of the bond market and of the business cycle. The results indicate a strong interrelation between these two sectors. The proposed factor model of the yield curve exhibits substantial incremental predictive value compared to several alternative specifications. This result holds in-sample and out-of sample, using revised or real time unrevised data.

Labor market dynamics over the business cycle: evidence from Markov switching models

This paper explores the evolution of the U.S. labor market across the business cycle and specifically the relationship between the unemployment rate and the average duration of unemployment. Labor market recoveries have long been thought of as lagging recoveries in broad economic activity. In particular, the unemployment rate peaks several months after official business cycle troughs and the average duration of unemployment lags further behind. Using estimates from Markov switching models of the unemployment rate, average duration of unemployment, jobless claims, and the exhaustion rate of regular unemployment insurance, this paper dates contractionary and expansionary phases of various aspects of the labor market and their relationship to the official phases of the business cycle. Evidence from these models suggests that inflows into unemployment recover almost contemporaneously with broad economic activity, while outflows recover almost a year after the end of official recessions. The differential timing in the recoveries of unemployment inflows and outflows, which is not a characteristic of most macro models of the labor market, accounts for the observed pattern between the unemployment rate and average duration of unemployment. Finally, when comparing the phases of the labor market to periods where Congress extends unemployment insurance benefits, it appears that policymakers target periods where the job finding rate is low, rather than periods where the stock of unemployed workers is high.