Photo by Urban Sanden on Unsplash

There is a growing interest in quantifying the degree in which fiscal policy instruments, such as government spending or changes in our tax structure, shape the economy. Attention to these issues has gained momentum after the Great Recession of 2008-09.

Unfortunately, computing the effect of fiscal policy changes is associated with many challenges. In fact, in my recent research with colleagues Kerim Arin, Aysegul Corakci and Nicola Spagnolo, we identify yet another challenge: the way fiscal policy instruments are designed may unintentionally favor some gender/racial/ethnic groups over others.

I will start by providing a brief background on how our knowledge on fiscal policy instruments evolved in recent years.

Challenges in understanding the effects of fiscal policy shocks

One challenge involves the measurement of fiscal shocks. While there is no consensus, the narrative approach has grown increasingly popular. In 1998, Valerie Ramey and Matthew Shapiro pioneered this approach with respect to defense spending, where they constructed a measure of unexpected changes in defense spending on major events in US history such as the Vietnam War, the Korean War, and the Carter-Reagan military build-up, to identify shocks to fiscal policy using various media outlets. In the same vein, Christina Romer and David Romer used information from presidential speeches and Congressional reports to identify and quantify exogenous tax changes for the US in 2009.

Another challenge involves how findings are interpreted when people’s expectations are accounted for. For example, several researchers have found that an increase in defense spending leads to a decrease in consumption and an increase in the number of work hours because people presume that such a rise in government spending will be financed by an ensuing tax hike. Relatedly, there is a lag between the announcement of fiscal policy changes and their implementation. For example, let’s suppose policymakers announce a tax cut but do not immediately implement it.

Upon announcement, people start working fewer hours and investing less while waiting for tax cuts to materialize, whereas after implementation, people work more hours, invest more, and overall output increases. These findings are shown by Karel Mertens and Morten Ravn in 2012, and are consistent with the understanding that people are highly responsive to incentives and make optimal decisions such as increasing (decreasing) hours and income when tax cuts are implemented (anticipated). However, these findings also highlight that conflicting results may emerge depending on how the period, before and after the announcement, is chosen. In other words, the negative anticipated effects measured post-announcement and prior to the implementation of tax cuts are substantial and should be disentangled from the real economic effects once tax cuts are applied.

The last major difficulty involves how uniform all these effects are. One can easily imagine how some people might be more susceptible to tax hikes and increases in military spending than others, especially during economic downturns. In fact, one of the most contentious issues of the Great Recession was persistently high unemployment and underemployment rates, and recent evidence suggests that the impact varied significantly. For example, Owen Zidar shows that the positive relationship between tax cuts and employment growth can be primarily explained by tax cuts for low-income groups. Furthermore, Danny Yagan shows that local areas in the US that experienced disproportionally high unemployment rates in 2008-09 continued to have lower employment rates among working age individuals and the impact was more concentrated among older and low-wage earners.

Which groups are most likely to be reached?

Our research contributes to the last major difficulty, which involves understanding which groups are the most likely to be affected by the way fiscal policy instruments are designed. The novelty in our approach is that we incorporate the importance of gender, race and ethnicity in addition to specific information about an individual’s job such as the last sector (public/private) of employment, industry and occupation.

We find that the employment status (employed or unemployed) of Black Americans are the least responsive to changes in the tax structure, relative to non-Hispanic Whites and Hispanic-Whites, and that this can be primarily explained by the fact that Black women are over-represented in the public sector. In other words, once the sector of employment is accounted for, there are no racial/ethnic differences in how individuals respond to tax cuts or tax hikes.

We also show that the employment outcomes of non-Hispanic Whites react strongly to military spending shocks, while those of minority groups do not. Such results are related to the over-representation of non-Hispanic White males among the self-employed or the manufacturing industry.

In the case of shocks in taxes and defense spending, we find that the employment status of non-Hispanic Whites are the most likely to change while those of disadvantaged minority groups (Blacks, Hispanics) are mild. However, in the case of tax changes, the results are primarily explained by how women of different racial/ethnic subgroups sort into the labor market, while the effect of defense spending shocks is completely driven by men. Thus, labor market segregation by race, ethnicity and gender can explain how some groups have been favoured by fiscal policy innovations in the past.

Some benefit more than others do

In a nutshell, it is our contention that ethnic minorities are less reachable by fiscal policy instruments and gender plays a crucial role in exacerbating the relationship between changes in fiscal policy measures and the likelihood of employment. The research on this issue is nascent and subject to interpretation. Yet, we are in dire need to know whether fiscal policy instruments are poorly designed for certain subgroups.

When an economic downturn takes place, many policymakers have the benevolent intention to target the lowest income groups by cutting personal income tax rates (for the low and middle classes) to stimulate the economy and help families recover. But in reality, these benefits may not be shared equally by low/middle income households if the labor markets are segmented by gender, race and/or ethnicity and some market segments are financed better than others.


Further reading

Adnan, W, Arin, KP, Corakci, A & Spagnolo, N 2019, ‘A closer look at the employment effects of fiscal policy shocks: what have minorities got to do with it?’, CAMA Working Paper 66/2019, Centre for Applied Macroeconomic Analysis, Crawford School of Public Policy, Australian National University, Canberra.

Leave a comment

Your email address will not be published. Required fields are marked *