Revisiting the Wage Phillips Curve in the Three Economic Periods from 1960 -2022

 

Abstract

This paper will examine the economic period from 1960 to 2022, which captures the role of monetary policies impact of the Phillips Curve on post war US economy. There is a consensus of three distinct economic periods - the post war recovery and golden age 1960-1972; Stagflation and neoliberal adjustment, 1973-1993; and, Great moderation, secular stagnation, and major crises, 1994-2023. The later economic period considered the dawn of the neoliberal stage of capitalism where collective bargaining power of workers began to diminish. Examining each period is significant for the empirical and theoretical work of Heterodox and Post Keynesian economists who in general believe thus conflict is essential in examining the flattening of the Phillips Curve. The paper builds and extends the work of these economists.

Introduction

 

The paper will examine the effect of monetary policies on the overall workers' compensation in manufacturing. Empirically, the paper expands on prior econometric Vector Autoregression (VAR) models and extends the analysis incorporating Vector Error Correction Methods (VECM) and Structural Vector Error Correction (SVEC) econometric techniques to account for economic shocks and structural shifts and includes Shaikh’s  (2013) work when incorporating unemployment intensity and real wage share in the empirical analysis.

The Phillips Curve is one of the most important relationships in macroeconomics, sparking volumes of papers on the subject for decades, especially questioning monetary policy’s impact in shaping the curve in the current economy. The original Phillips Curve (Phillips, 1958) was a purely empirical finding of a relationship between inflation and unemployment where money wages rose in a nonlinear manner when unemployment was below critical levels and fell in a comparable manner when unemployment was above that level.

Figure 1 below shows the original intent of the Phillips Curve using data from 1861-1957 the cyclically adjusted rate of change of money wages in the UK was positive when unemployment was below a certain critical level u and was negative when unemployment was higher.

Kaleckian Phillips Curve

 

Kalecki, along with lesser-known writers as Henri Aujac (1950), were  of the view that inflation is primarily an expression and the outcome of class conflict (or conflicting claims) over national output firms price their products (mark ups) over workers’ wage settings. In general, inflation in both Kaleckian and post-Keynesian models are determined by conflict over income distribution between capital and labor. However, Kaleckian theory tends to go beyond the various Post Keynesians class conflict theories by emphasizing the role of monetary policy in a Phillips Curve.

Post War Economic Periods and the Phillips Curve

 

The period examined includes the post war period (1960) up to the Post COVID period (2022). Based on the graph below although there are outliers, one can infer that there is a relationship between wage inflation and unemployment rates. Figure 4 shows the Wage Phillips Curve from 1960 – 2023, shows not only downward sloping Phillips curve but also clustering and outliers. 

For the sake of simplicity, the paper will use the time - the post war recovery and golden age 1960-1972; Stagflation and neoliberal adjustment, 1973-1993; and, Great moderation, secular stagnation, and major crises, 1994-2022: the neoliberal era.

Figure 5 above reflects the economic period where during the Golden Age there was high productivity coupled with wage growth. Satterfield (2022) period captures the gist of the argument that the Golden Age (1960 – 1972) era had a downward-sloping curve with low unemployment rates. The Stagflation and Neoliberal Adjustment period (1973-1993) saw a breakdown in the capital – labor “social bargain” during the Golden Age. This breakdown “created structural instability in the Phillips Curve that obscured the underlying inverse relationship between inflation and unemployment” (Setterfield 2022:12). Finally, the Great moderation, secular stagnation, and major crises, 1994-2023 saw rising labor productivity with flat real wage growth. During this period labor strength continued to weaken with decreases in unionization rates, increased globalizations, and major recessions. As a results, labor’s ability to ‘bid up” wages declined (pre- COVID)

Preliminary analysis, using two variable Granger Causality tests {results not shown) to capture Phillips Curve dynamics - unemployment (unemployment intensity) and wage inflation (real wage share inflation rate) does show causality varies based on economic periods. This is consistent with prior empirical work from Palley (1999). Setterfield calls for a “tripartite” relationship which encompasses inflation, unemployment, and distribution of income variables. Discussion and results of the four variable VAR model performed in the paper attempts to capture the tripartite relationship suggested by Satterfield while addressing the impact monetary policy on the Phillips Curve.

Concluding Remarks 

 

The paper revisits the Wage Phillips Curve and tries to keep the original intent of the model while also following the Heterodox/Post Keynesian tradition. By including variables consistent with heterodox   conflict inflation theory along with a monetary policy component   utilizing VAR, VECM, SVEC econometric techniques, the paper is reflective of this tradition. The empirical results show that there are structural breaks in the post war economy. Based on reported structural changes in the economy it appears that the Wage Phillips Curve can be downward sloping, vertical, flat, or backward bending (Palley,2008). Palley’s empirical results indicate that is a flat wage Phillips Curve especially during the Great moderation, secular stagnation, and major crises period where workers strength tends at its lowest point (pre COVID).

 

The empirical results were based primarily on real wage share inflation of workers in the manufacturing sector, however additional industry and subsectors were analyzed. The results were not presented in its entirety, but wage share in the goods producing industry needs to be explored further along with the respective subsets.

 

Not surprising but may contradict some standard economic empirical results, real wage share inflation does not supply significant shocks or impact on industrial commodity price inflation, unemployment intensity nor interest rates. Tests such as granger causality, counterfactuals, historical decomposition, IRF, FEVD analysis in manufacturing its sub sectors including non-durable and durable, goods. The results show minimal influence.

The yield curve had the most consistent impact. This can be due to the role it plays as a macroeconomic business cycle indicator and long-term interest rates (and can be a better indicator of monetary compared to the Fed Funds Rate).

 

Unemployment Intensity was a useful variable the addresses deficiencies as opposed to an unemployment rate variable, especially examining manufacturing real wage inflation and its subsectors. Data constraints, the earliest manufacturing unemployment data to my knowledge begins in the year 2000.

 

Therefore, there will be a tradeoff between a possible robust model and shorter time. More research is called for such as using econometric VAR, SVAR, VECM and SVEC models, and expanding the VAR model to six variables to reflect labor strength, and trade openness by industry. Also, a deeper dive into manufacturing subsectors should provide more clarity on manufacturing and the Wage Phillips Curve. Finally, since there are structural breaks (shifts) in the full sample, further research examining these models would be fruitful.

 

 

 The Impact of COVID 19 on Essential Workers  

in Selected Industries 

 

Summary of Findings 

    The purpose of the summary is to discuss the results of Hurst’s multivariate regression modes to determine which industries deemed essential workers were adversely affected with high mortality rates due to the initial COVID 19 outbreak. The industries include nursing facilities, agricultural, mining, construction, food manufacturing, transit, warehousing, ambulance service, hospital, food, and drink places. 

    It is important to note the International Association of Machinist and Aerospace Workers (IAM) represents workers in industries considered essential workers. Recently, the IAM has had an increasing presence in the health care industry including hospitals and nursing facilities. Therefore, it is in the IAM’s interest to see which industries were disproportionately impacted by COVID. 

     Chen et al (2022) shows that industries with the highest mortality rates include accommodation and food services (45.4 per 100,000); transportation and warehousing (43.4); agriculture, forestry, fishing, and hunting (42.3); mining (39.6); and construction (38.7)”. Transit and medical variable were also included based on the research by Heinzerling et al (2022. They found, “public transportation industries in California experienced cumulative COVID-19 outbreak incidence and mortality rates 1.5 times as high as that for all industries; outbreak incidence was 5.2 times as high, and mortality was 1.8 times as high in bus and urban transit industries as in all industries” (p.1055). Additionally, Heinzerling et al (2022) suggest an extension of their research should occupational risks (or industries) with race, ethnicity, and socioeconomic factor as it relates to COVID mortality rates.  

    Yu (2021) provides compelling evidence on the variation of COVID on disparate groups across the United States. Empirically, Yu used a straightforward linear regression cross-sectional model to find the predictors for the COVID-19 accumulated case and death rates. Yu ran seven multivariate regression models and used models 1-3 as the benchmark models to explain the regression findings. 

    Hurst (2022) extends the Yu’s model by using average annual pay (instead of employment) for selected NAICS code three-digit industries to test the premise of essential workers in the respective industries had high mortality rates due to COVID. Additionally, Hurst includes race and ethnicity, median income, poverty, unemployment, and state variables. The assumption is that these variables coupled with industry/occupation variables can help explain the variation and disparity in COVID cases and deaths across the county. Multiple regression models were performed using the variables mentioned above and like Yu, Hurst’s results (data and script available upon request) show how essential workers were overly exposed to COVID which helps explain positively correlated industry variables with mortality rates despite differences in average annual pay (proxies for compensation and wellbeing).  

    Further, workers in nursing facilities in the various models were statistically significant compared to the other industry variables. The results are aligned with prior research and reports showing nursing homes were the epicenters for coronavirus outbreaks in states such as Washington and New York 

References 

Amy Heinzerling, Alyssa Nguyen, Matt Frederick, Elena Chan, Kathryn Gibb, Andrea Rodriguez, Jessie Wong, Erin Epson, James Watt, Barbara Materna, and Seema Jain, 2022:Workplaces Most Affected by COVID-19 Outbreaks in California, January 2020–August 2021, American Journal of Public Health 112, 1180_1190 

 

Yea-Hung Chen, Ruijia Chen, Marie-Laure Charpignon, Mathew V Kiang, Alicia R Riley, M Maria Glymour, Kirsten Bibbins-Domingo, Andrew C Stokes. COVID-19 mortality among working-age Americans in 46 states, by industry and occupation. medRxiv. 2022 

 

Tazewell Hurst. (2022), The Impact of COVID 19 on Essential Workers in Selected Industries., Upper Marlboro MD: IAMAW Strategic Resources Department  

William Yu,  Health in America: What Explains the Variation in COVID-19 Mortality Rate Across the United States. The UCLA Anderson Forecast, March 2021. 

 

 

 

 

 

Is the Economy Nearing a Recession? It Depends

The Sahm Indicator Rule is an excellent measure for determining a recession. An official recession is based on measures by the National Bureau of Economic Research. However, the Sahm follows the beginning of recession seamlessly and shows their lagging impact of the ending of the official recession on the respective labor markets.  

The Sahm Rule was developed to flag the onset of an economic recession more quickly than other indicators. The Sahm Rule signals the start of a recession when the three-month moving average of the national unemployment rate rises or stays consistently above a minimum of 0.50 percentage points or more relative to its low during the previous 12 months. Once the indicator exceeds the 0.85 mark it is almost safe to assume a recession is occurring or forthcoming. 

Current unemployment data (May2024 US Jobs Report) shows the Sahm recession indicator was 0.4, a slight increase from the previous month. Nationally, the economy is not in a recession based on the Sahm Rule index but can be inching towards an economic downturn in the next couple of months. 

 

Figure 1, Historical Sahm Rule Indicator of the US Unemployment Rate 

   

The West census area, at 0,7 percentage point increase over a twelve-month period, is the only region where a recession may be present. The Northeast region (lower left panel) is at 0.4 percentage points which is slightly shy of the 0.5 recession threshold. 

The next couple of months of unemployment will have a significant impact not only on the state of the overall economy but also on the financial markets and the elections in November. 

 

Figure 2 Sahm Rule Indicator by Census Region