Kate Murphy

Effects of Productivity Shocks on Unemployment

 

         

Changes in productivity have various effects on the macro economy; one variable that productivity affects is unemployment.  This paper discusses the effect of a positive productivity shock on the unemployment rate.  Increases in productivity, such as newer and more efficient techniques and methods of production, allow firms to replace workers with machines.  Newer technologies also let firms teach certain workers to be more efficient, eliminating the need for as many employees.   The question is how firms react to changes in productivity: do they lay off workers after learning of a new technological advancement, or do they increase output?  The hypothesis is that when firms use productivity shocks to become more efficient, they choose to decrease their workforce.  Therefore, the hypothesis states that productivity is positively correlated to the unemployment rate; increases in productivity cause the unemployment rate to increase.  The data used to answer this question show that the hypothesis is wrong, but the results are in accordance with the aggregate supply and demand model which states that increases in productivity increase output which decreases unemployment. This answer tells us that firms use new technological advancements to increase their output rather than reduce their labor force.  If a productivity shock affects one field and a few firms initially choose to use the shock to increase output, then other firms that wish to stay competitive in the industry must also keep workers and increase output. 

          The aggregate supply and demand model is used to estimate the relationship between productivity and unemployment.  Although the unemployment rate is not included in this model, gross domestic product is included; when gross domestic product increases, as a result of changes in the various independent variables, then the unemployment rate decreases.  As changes in the independent variables cause the supply and demand curves to shift, new equilibrium points are created, indicating new levels of output, and in turn, new levels of unemployment.  The independent variables which cause shifts in the aggregate supply and demand curves are the variables which were included in the regression to estimate the relationship between productivity and unemployment.  The other independent variables used in addition to productivity include growth in the consumer price index, the budget deficit, the exchange rate for the US dollar, money supply growth and oil prices.

          Estimating the relationship between productivity and unemployment requires time series data from the US macro economy.  The data for the majority of the variables begins in 1959 and ends in 2002, data for CPI and money supply begin in 1960, while oil price data does not begin until 1974.  Productivity and oil prices are the variables which affect the supply side of the model.  The productivity variable is measured in output per hour, this measurement includes workers from all fields excluding the agricultural industry.   The differences between output per hour of all workers and output per hour excluding agricultural workers are very minimal.  The units for the oil price variable are merely the price, in US 1983 dollars, per barrel of oil.  The variables which affect the demand side of the model include CPI, money supply, budget deficit, and the exchange rate.  The CPI and money supply variables needed to be modified in order to account for inflation; rather than using the actual data points from year to year, new variables were created to calculate CPI growth and money supply growth from year to year.  The budget deficit variable quantifies the United States government budget deficit in billions of 1983 dollars.  The exchange rate for the US dollar is an index variable which measures the value of US dollar in 1983 dollars. 

          The results of the initial regression showed that both the CPI growth and exchange rate variables proved to be insignificant.  F-Tests were performed on the two restricted regressions and the tests showed that it was acceptable to drop the variables from the regression.  After dropping the CPI growth and exchange rate variables from the equation, all of the remaining variables appeared to be statistically significantly different from zero.  The final regression is as follows,

Unemployment = β1productivity+β2deficit+β3moneysupply+β4oilprice+ε

The results of this regression show that increases in two of the variables, productivity and budget deficit, cause decreases in the unemployment level.  Increases in the remaining two variables, money supply and oil prices, cause increases in the unemployment level.  The following table shows the estimated values and standard deviations of each of the variables.                     

 

Productivity

β1

Budget Deficit

β2

Money Supply

β3

Oil Price

β4

Estimated Value

-.032

-.502

.138

.048

Standard Deviation

.006

.035

.023

.008

          The results of the final regression show that the hypothesis that positive productivity shocks increase unemployment was incorrect.  Technological advancements actually increase the level of employment; these results are in accordance with the aggregate supply and demand model which states that increases in productivity shift the supply curve to the right, creating a higher level of output, which decreases the unemployment rate.  These results demonstrate that firms do not choose to lay off employees after learning of a new technological advancement, but rather use the knowledge to keep workers and increase productivity.