RaboResearch - Economic Research

Dit artikel is ook beschikbaar in het Nederlands

Why the Trump phenomenon is here to stay

Economic Report

Share:
  • This year’s US presidential election process spawned Donald Trump as an anti-establishment, who made it to the Economist Intelligence Unit top ten risks
  • Statistics tell us that stagnant median wages and declining manufacturing employment correlate strongly with Trump’s primary results
  • In the short run, wages will remain stagnant, owing to persistent underemployment, while manufacturing jobs lost in the last decades will not be returning promptly
  • This suggests that the Trump phenomenon is likely to stay, even if Trump himself does not make it to the White House

The US is getting ready for the 2016 presidential election. Currently we are in the primary process, a phase in which both parties nominate their presidential candidate. Both the Democratic and Republican primaries feature candidates more extreme than we have seen in primaries in recent years. Senator Sanders has declared himself an economic socialist and refers to a rigged economy while Donald Trump has vowed to build a wall on the Mexican border to stop immigration and make Mexico pay for its construction, amongst other extreme proposals. Both candidate rail against the establishment to the extent that it has become a buzz word in this campaign, which bears some resemblance to the populist messages we have become so accustomed to in Europe in recent years.

Figure 1: Share of Trump votes in total votes per primary
Figure 1: Share of Trump votes in total votes per primary Source: Politico

In this report we focus on Donald Trump because he is at this point the most likely candidate to clinch the Republican nomination. Also, his (better than expected) results so far, combined with his extreme views and proposals have made him number 6 in the Economist intelligence unit top ten risks. As the success of Trump unfolds, many theories are being forwarded to explain his success with primary voters[1]. Instead of the usual campaign narrative, in this article, we use an economic approach to explaining their success. We analyze how the state of the economy affects people’s preference for his more radical and populist messages.

Explaining with the economics toolkit

Our hypothesis is that worsening economic circumstances since 2007 have fueled discontent among voters. In the light of Trump’s anti-establishment campaign, this discontent would help him to attract protest votes on the current state of the economy. The economic view would be that the establishment can ward off Trump-type attacks by addressing the real and perceived economic challenges of the wider population. To gauge economic circumstances, we look at four indicators that may explain voter discontent:

  1. (Real) median income
  2. Unemployment
  3. Manufacturing jobs
  4. House prices

Median income is an indicator of how the median household is doing, which may differ from average income. Unemployment shows what part of the working population is disintegrated with the economic system. Off-shoring of jobs particularly hurts low-skilled workers who have difficulty in transitioning to a new (service) profession, which could fuel discontent. In addition we also look at house prices as an indication of households financial wealth, as over half of the population lives in a house they own (U.S. Department of Commerce, 2016).

The macro picture…

Figure 2: Development of selected economic indicators
Figure 2: Development of selected economic indicatorsSource: Macrobond, Bureau of Labor statistics, U.S. Census Bureau, Federal Housing Finance Agency, Rabobank

Seven years into the economic recovery, we see that unemployment and housing prices have recovered to pre-crisis levels for the US as a whole (figure 1). Nationwide unemployment is at pre-crisis levels and house prices even exceed their pre-crisis peak.
However, manufacturing jobs remain substantially below their 2007 level and have been steadily dropping for most of the past one-and-half decade. Median income has been stagnant, both since the crisis and during the years before it. These macro-trends obscure regional differences, which we will explore below precisely to explain the success of Trump.

… versus the regional picture

Figures 3 to 6 plot the Primary results up to Tuesday 15 March against the various indicators per state (for underlying data see statistical annex). Without checking for significance, we see that most variables correlate with Trump’s share of the votes in the way we expect. The deeper the (remaining) scars in terms of median income in a state in 2014, the stronger the preliminaries support for Trump (figure 3). A larger decline in manufacturing jobs is also associated with Trump’s share of the vote[2] (figure 4) while a larger decline in house prices results in stronger support for Trump (figure 5). Only unemployment does not show a very strong correlation with Trump’s success in a state (figure 6). To check how these variables jointly affect Trump’s share of the vote, we perform a regression analysis below.

Figure 3: Change in median income
Figure 3: Change in median incomeSource: U.S. Census Bureau, Politico
Figure 4: Change in manufacturing jobs
Figure 4: Change in manufacturing jobsSource: Bureau of Labor statistics, Politico
Figure 5: Change in house prices
Figure 5: Change in house pricesSource: Federal Housing Finance Agency, Politico
Figure 6: Change in unemployment rate
Figure 6: Change in unemployment rateSource: Bureau of Labor statistics, Politico

If we formalize the pictures above into a simple linear model, we see that changes in house prices and changes in unemployment turn out insignificant (table 1). Omitting these variables from our model yields regression 2, which shows that both an 1%-point increase in median income and an increase in manufacturing jobs leads to a decrease in the share of votes for Trump of -1.6%-point and a -1.6%-point respectively. Both variables are highly significant and perhaps more strikingly, they explain around 86% of the variation between states in the share of votes for Trump in the primaries.

Table 1: Regression results
Table 1: Regression resultsSource: Macrobond, Bureau of Labor statistics, US Census Bureau, Federal Housing Finance Agency, Rabobank

Why Trump is here to stay

The results above show the economic situation of the US goes a long way in explaining the success of anti-establishment candidate Donald Trump in the Republican primary. Bearing in mind figure 1, we see that the trends towards declining manufacturing employment and stagnant median incomes are actually rather persistent throughout the last fifteen years. Analyzing those trends, we can only conclude that the Trump phenomenon –broad electoral support for anti-establishment and / or populist politicians– is here to stay, even if the person Donald Trump is unable to bring that phenomenon all the way into the White House in the present presidential elections race.

Figure 7: Substantial underemployment remains
Figure 7: Substantial underemployment remainsSource: U.S. Bureau of Labor Statistics

In the short run, (median) incomes will remain stagnant as labour market slack (figure 6) weighs on wage growth (Aaronson and Jordan, 2014). The declining trend in manufacturing employment can be attributed partly to the US’ trade policy vis-à-vis China (and probably other countries) as Pierce and Schott (2012) have found a link between the US opening to Chinese exports and the sharp drop in manufacturing employment. Although Trump may have successfully tapped into the anti-China sentiment, no policy will promptly bring back manufacturing jobs. This suggests that whoever succeeds Obama next year as the 45th President of the United States will have to address the issues of a stagnant median wage (figure 8) and declined manufacturing employment (figure 9), to neutralize the (populist) anti-establishment sentiment.

Figure 8: Change in median income since 2007
Figure 8: Change in median income since 2007Source: U.S. Census Bureau
Figure 9: Change in manufacturing employment (jobs) by state since 2007
Figure 9: Change in manufacturing employment (jobs) by state since 2007Source: U.S. Bureau of Labor Statistics

Footnotes

[1] Note that this article was finalised before the Wisconsin Republican primary. The outcome of that primary does not change our analysis as our out-of-sample prediction showed that Trump would lose (see Table I of the statistical annex).

[2] In the correlation, we excluded the District of Columbia (DC) as it suffered an enormous decline in its rather modest manufacturing base. Including it would give it unduly weight in the correlation.

Literature

Aaronson, D., & Jordan, A. (2014). Understanding the relationship between real wage growth and labor market conditions. Chicago Fed Letter, (Oct).

Pierce, J. R., & Schott, P. K. (2012). The surprisingly swift decline of US manufacturing employment (No. w18655). National Bureau of Economic Research.

U.S. Department of Commerce (2016). U.S. Census Bureau News.

Statistical annex

Tabel I: Model outcome with prediction
Tabel I: Model outcome with predictionSource: Macrobond, Bureau of Labor statistics, US Census Bureau, Federal Housing Finance Agency, Rabobank
Table II: Economic variables on state level
Table II: Economic variables on state levelSource: Macrobond, Bureau of Labor statistics, U.S. Census Bureau, Federal Housing Finance Agency, Rabobank
Share:
Author(s)
Jurriaan Kalf
RaboResearch Netherlands Rabobank KEO
+31 (0)30 21 62666

naar boven