Risk and Macroeconomic Activity

This article explores the macroeconomic role that risk plays using the BAA‐AAA spread as the measure of risk. First, it shows that meaningful upward movements in this spread are associated with recessions and their severity. Second, it includes the BAA‐AAA spread in a structural vector‐autoregression (VAR) to identify a shock‐to‐risk and finds that it causes a statistically significant and economically important decrease in output as well as increased holdings of real‐money balances. Third, it uses historical decompositions to show that the shock‐to‐risk explains an important part of the declines in output during four post‐1970 recessions. Notably, the shock‐to‐risk explains almost none of the decline in output during 2001 prior to the September 11, 2001, terrorist attacks but does clarify why the recovery was relatively weak afterwards, and it explains the bulk of the decline in output during 2008 and 2009.     

THE PROPERTIES OF CYCLES IN SOUTH AFRICAN FINANCIAL VARIABLES AND THEIR RELATION TO THE BUSINESS CYCLE

The paper describes cyclicality in a range of local and international financial variables and their relation to cyclical behaviour in the South African real economy. Cycles are derived using a dating algorithm similar to that used to determine business cycle turning points and falls within the Burns‐Mitchell tradition of business cycle analysis. Co‐movement between phases in financial variables and similar phases in the business cycle are described using the concordance statistic, instead of the correlation statistic (which requires stationarity). This is a preliminary step in identifying financial variables that can act as leading indicators of economic activity.     

A New Approach to the Study of Jobless Recoveries

This article is concerned with the measurement of jobless recoveries and the elements that may explain their emergence. We first introduce a measure that maps the various elements that define a jobless recovery into a single number that we label the jobless recovery depth. We then construct a database of 389 state‐level observations and study the cross‐sectional variations that emerge. We find that jobless recoveries in the United States are not a nation‐wide phenomena, but a local event confined within a cluster of states that expands slowly between 1975 and 2015. We find the state‐level evidence to be consistent with theories that link jobless recoveries to unusually long expansionary periods, less dynamic labor markets, and the advent of the great moderation. The evidence is not consistent with theories that link them to decreases in union power, increases in income inequality, or increases in health care costs.     

LONG-RUN M3 DEMAND IN SOUTH AFRICA: A COINTEGRATED VAR MODEL

This paper presents a system cointegration analysis of a long‐run demand for money (measured in terms of M3) in South Africa. In particular, the paper estimates a cointegrated vector autoregression model, consisting of real money, income and the opportunity cost of holding money. Using a variety of theory consistent identification schemes, the money demand function is identified along with other two cointegrating relations, namely, an IS‐type relationship and a relationship relating inflation to the spread between long‐ and short‐term interest rates. The model shows that of the variables used, only income and real money are error‐correcting to the money demand relation. The money demand relation is found to be relatively stable over the sample period, when short‐run fluctuations are corrected for. The model further shows that the long‐run link between money and inflation is rather weak.     

A SMALL-SCALE DSGE MODEL FOR FORECASTING THE SOUTH AFRICAN ECONOMY

This paper uses a version of Hansen's (1985) Dynamic Stochastic General Equilibrium (DSGE) model to forecast the South African economy. The calibrated model, based on annual data over the period of 1970‐2000, is used to generate one‐ to eight‐quarters‐ahead out‐of‐sample forecast errors for the period of 2001:1 to 2005:4. The forecast errors are then compared with the unrestricted versions of the Classical and Bayesian VARs. A Bayesian VAR with relatively loose priors outperforms both the classical VAR and the DSGE model.     

THE SOUTH AFRICAN PHILLIPS CURVE: HOW APPLICABLE IS THE GORDON MODEL?

Is there a Phillips curve relationship present in South Africa and if so, what form does it take? Traditionally the method to establish whether or not there is a relationship between the output gap and the change in inflation is merely to regress the latter on the former. This yields the well‐known augmented Phillips curve. However, Gordon has argued that this specification of the Phillips curve produces biased results. Instead, he puts forward and estimates successfully for several industrialised countries his so‐called triangular model that tests for hysteresis and inertia in the behaviour of inflation, as well as the impact on inflation of changes in the output level. This paper considers whether or not Gordon's triangle model is applicable to South Africa, i.e. are hysteresis and inertia present in South Africa? In addition, in an attempt to find a better estimation of the output gap, the paper also experiments with alternative ways to estimate the long‐run output level, including the standard HP‐filter, as well as a production function approach.     

Understanding the Cyclical Nature of Financial Intermediation Costs

Intermediation costs (i.e., all noninterest bank expenses) are counter‐cyclical and their changes have significant effects at the country‐level; however, the literature is silent on what drives their cyclicality. Previous studies have examined costs using cross‐sectional or low frequency data and thus cannot capture dynamic macroeconomic effects across time. We fill this gap by examining U.S. intermediation costs using a quarterly, bank‐level dataset from 1993 through 2012. This data set allows for the separation of microeconomic and macroeconomic factors affecting the cyclicality of costs. The analysis shows that house prices are the main driving source of this counter‐cyclicality. Because housing is used as collateral, a price decline leads banks to increase their operating costs for monitoring, screening and litigation costs which cause a credit crunch in the economy.     

Measuring Potential Output for the South African Economy: Embedding Information About the Financial Cycle

Borio et al. show that information embedded in the financial cycle can serve to improve measures of potential output and output gaps. They show that incorporating information on the financial cycle yields measures of potential output and output gaps for the United States, United Kingdom and Spain that are estimated more precisely and are more robust in real time. With its well‐developed financial markets and relatively open capital markets, the South African economy is potentially susceptible to the build‐up of the sort of financial imbalances that characterised the recent financial crisis. Using the framework developed in Borio et al., a finance‐neutral measure of the output gap is estimated for South Africa. The traditional Hodrick–Prescott filter is extended to incorporate information on credit and property prices. Including financial cycle proxies result in output gaps that are estimated more precisely and are more robust to data revisions and the arrival of new data points (i.e. estimated output gaps are more robust in real time), while also reflecting the impact of financial variables on economic activity. As such, the estimated finance‐neutral output gap seems to represent a more appropriate measure on which to base monetary policy decisions.     

A BVAR MODEL FOR THE SOUTH AFRICAN ECONOMY

The paper develops a Bayesian vector autoregressive (BVAR) model of the South African economy for the period of 1970:1‐2000:4 and forecasts GDP, consumption, investment, short‐term and long term interest rates, and the CPI. We find that a tight prior produces relatively more accurate forecasts than a loose one. The out‐of‐sample‐forecast accuracy resulting from the BVAR model is compared with the same generated from the univariate and unrestricted VAR models. The BVAR model is found to produce the most accurate out of sample forecasts. The same is also capable of correctly predicting the direction of change in the chosen macroeconomic indicators.     

INTEREST RATE PASS-THROUGH IN THE COMMON MONETARY AREA OF THE SACU COUNTRIES

We investigate the interest rate pass‐through in the four Common Monetary Area (CMA) countries of the South African Customs Union (SACU). We employ an empirical pass‐through model that allows for thresholds, asymmetric adjustment, and structural changes. We show that CMA bank lending markets exhibit quite some degree of homogenization as the pass‐through is often fast and complete. Deposit markets are somewhat more heterogeneous by showing differing degrees of interest rate stickiness and asymmetric adjustment. Policy makers should therefore be concerned about imperfect competition which may be at the heart of the remaining cross‐country differences in monetary transmission in the CMA.     

MEASURING THE WELFARE COST OF INFLATION IN SOUTH AFRICA

In this paper, we estimate the long‐run equilibrium relationship between money balance as a ratio of income and the Treasury bill rate for the period of 1965:02 to 2007:01, and in turn use the relationship to obtain welfare cost estimates of inflation. Using the Johansen technique, we estimate a log‐log specification and a semi‐log model of the above relationship. Based on the fits of the specifications, we decided to rely more on the welfare cost measure obtained under the log‐log money demand model. Our estimates suggest that the welfare cost of inflation for South Africa ranges between 0.34% and 0.67% of GDP, for a band of 3‐6% of inflation. Thus, it seems that the South African Reserve Bank's current inflation target band of 3‐6% is not too poorly designed in terms of welfare.     

FORECASTING THE SOUTH AFRICAN ECONOMY WITH GIBBS SAMPLED BVECMs

The paper uses the Gibbs sampling technique to estimate a heteroscedastic Bayesian Vector Error Correction Model (BVECM) of the South African economy for the period 1970:1‐2000:4, and then forecasts GDP, consumption, investment, short and long term interest rates, and the CPI over the period of 2001:1 to 2005:4. We find that a tight prior produces relatively more accurate forecasts than a loose one. The out‐of‐sample‐forecast accuracy resulting from the Gibbs sampled BVECM is compared with those generated from a Classical VECM and a homoscedastic BVECM. The homoscedastic BVECM is found to produce the most accurate out of sample forecasts.     

FORECASTING THE SOUTH AFRICAN ECONOMY WITH VARs AND VECMs

The paper develops a Bayesian Vector Error Correction Model (BVECM) of the South African economy for the period 1970:1‐2000:4 and forecasts GDP, consumption, investment, short and long term interest rates, and the CPI. We find that a tight prior produces relatively more accurate forecasts than a loose one. The out‐of‐sample‐forecast accuracy resulting from the BVECM is compared with those generated from the Classical variant of the VAR and VECM and the Bayesian VAR. The BVECM is found to produce the most accurate out of sample forecasts. It also correctly predicts the direction of change in the chosen macroeconomic indicators.     

Inflation Forecast Targeting: An Alternative Approach to Estimating the Inflation‐Output Variability Tradeoff

We suggest a new way of computing the inflation‐output variability tradeoff under inflation forecast targeting. Our approach is based on dynamic, stochastic simulations of the average inflation rate over a two‐year horizon using the moving average representation of a vector autoregressive (VAR) model. Using real‐time data over two samples, we estimate the inflation‐output variability tradeoff for the United States and show that it has shifted favorably over time. We analyze the policy interventions required to achieve target inflation in each sample and compare these interventions over time.     

A recession‐and‐state forecasting model

The state forecasting model (SFM) developed herein predicts the probability that a recession will begin or end within the next quarter. It is also important to forecast the state, as models prone to false alarms are not useful. The model combines parsimony, predictive accuracy, and ease of estimation using readily available data. No regressor is benchmark‐revised, and only one is seasonally adjusted. The SFM correctly forecasts, out‐of‐sample, the binary state in 163 of 164 quarters for an overall proportion of correct forecasts of 0.994. The results demonstrate the discriminating ability, longevity and striking forecasting ability of the model. A benchmark model based on the newly revised Leading Economic Index of the Conference Board has zero marginal predictive power in the presence the SFM.     

A Simple Empirical Measure of Central Banks' Conservatism

In this article we suggest a simple empirical and model‐independent measure of Central Banks' Conservatism based on the Taylor curve. This new indicator can easily be extended in time and space, whatever the underlying monetary regime of the countries considered. We demonstrate that our measure evolves in accordance with the monetary experiences of 32 OECD member countries from 1980 and is largely equivalent to the model‐based measure provided by 2005 for the post‐1985 period. We finally present the relevance of such an indicator for further empirical analysis addressing the preferences of Central Banks.     

MONEY,OUTPUT AND INFLATION IN AFRICAN ECONOMIES1

This paper examines the long‐run money‐inflation relation for 36 African countries using cross‐section and panel data analysis. The focus is on the recent claim by De Grauwe and Polan that the common finding in multi‐country studies of a strong positive link typically reflects the presence of high‐inflation countries in the sample and on Nelson's criticisms of the data and methodology employed in that study. Adjusting the De Grauwe and Polan methodology to take account of many of Nelson's criticisms, I confirm a weak long‐run relation between money growth and inflation for countries when money growth and inflation are below 10%, but a strong relation when money growth and inflation move much above that number. This result is not dependent on the inclusion of high inflation countries in the cross‐section and panel data samples.     

ON INFLATION

There is currently much more common sense in the South African inflation debate than a few decades ago. In particular, the South African Reserve Bank exhibits a pragmatic, eclectic approach to inflation (as reflected in its bi‐annual Monetary Policy Reports). This is in stark contrast to the narrow, monetarist‐type thinking that tended to dominate during the 1980s. This paper is an attempt to contribute to the debate by highlighting a few issues, including the widespread substitution of the CPI by the CPIX, the fact that inflation is a process, the need to combat inflation, the causes of the decline in inflation in South Africa and the essential features of an inflation‐targeting framework for monetary policy.     

Persistence in U.S. State Unemployment Rates: Errata and Extensions

Using panel unit root tests allowing for breaking deterministics, 2007 found that many U.S. state unemployment rates were stationary, a result at odds with the traditional view that unemployment rates are path‐dependent and subject to shocks that have permanent effects. 2009 extended the analysis to directly examine whether the series were fractionally integrated and reported that models with two breaks‐in‐mean do suggest many state unemployment rates were mean‐reverting. The purpose of this note is to correct an error contained in 2009 , which when modified, indicates that state unemployment rates were non‐stationary processes well‐characterized by hysteresis.     

INFLATION TARGETING AND THE FISHER EFFECT IN SOUTH AFRICA: AN EMPIRICAL INVESTIGATION

The paper analyses the relationship between expected inflation and nominal interest rates during a period of inflation targeting in South Africa, i.e. from 2000 to 2005. Specifically, it investigates the Fisher hypothesis that nominal interest rates move one‐to‐one with expected inflation, leaving the real interest rate unaffected. The analysis distinguishes between a short‐run Fisher effect and a long‐run Fisher effect. Using cointegration and error correction models (for monthly data for the period April 2000 to July 2005), it was found that the short‐run Fisher hypothesis did not hold during the relevant period under the inflation targeting monetary policy framework in South Africa. This is attributed to a combination of the South African Reserve Bank's (SARB) control over short‐term interest rates and the effects of the monetary transmission mechanism. The long‐run Fisher hypothesis could not be confirmed in its strictest form: while changes in inflation expectations move in the same direction as the nominal long‐term interest rate. This suggests that monetary policy has an influence on the real long‐term interest rate, which has positive implications for general economic activity, thus confirming the credibility of the inflation targeting framework.