Last week's Michigan primaries - where voters chose among competing candidates for their party's 2016 presidential nomination - made headlines because of the victories they gave to Republican Donald Trump (as expected) and Democrat Bernie Sanders (in a "historic upset" over Mrs Hillary Clinton).
I have lived as a Singaporean in Michigan for most of the past 40 years. As a non-voter, I have usually not paid attention to primaries, though I have to the presidential election, since we are supposed to be a "swing state" (between Republicans and Democrats), and neighbouring Ohio, a half-hour away, has the distinction of "deciding" the presidency since Bush v Gore in 2000.
Electoral pundits and pollsters focus on these two states of the industrial Midwest because of their population size and, hence, electoral heft, and their political heterogeneity - roughly equally divided between conservative rural constituencies and more "liberal" urban-industrial ones. This means that a relatively small (1 per cent or 2 per cent) shift in voting patterns can "swing" the final state - and hence national - electoral outcome one way or the other.
In fact, unlike Ohio, Michigan has voted for the Democratic candidate in every presidential election since Mr Bill Clinton in 1992, though it has a Republican state legislature and switches between Republican and Democratic state governors. Last week's primaries saw record voter turnout for both parties, but Michigan is still expected to "lean Democratic" in the November presidential election.
Most commentators and exit polls have attributed the large Trump (11.4 per cent margin) and narrow Sanders (1.5 per cent) Michigan victories to those candidates' strong protectionist positions on trade policy, especially free trade agreements like the North American Free Trade Agreement (Nafta), passed in 1994 under President Bill Clinton. They argue that Michigan's largest industry, cars, and US manufacturing in general have been losing jobs due to import competition resulting from "unfair trade treaties" that Mr Sanders accuses Mrs Clinton of supporting.
Mr Sanders has especially attacked Mrs Clinton for supporting the Trans-Pacific Partnership (TPP) when she was President Barack Obama's Secretary of State, and she has now backed away from the deal. Mr Trump has promised to leave Nafta if he becomes president, and has (erroneously) blasted China for downward manipulation of its currency.
For economists, both Mr Trump and Mr Sanders are seen as major dangers for the economy, because of their proposed trade, fiscal and monetary policies, and the discouragement of business investment likely to result from either's election.
In Michigan, Mr Trump attacked Ford Motor for closing factories in the US and "outsourcing jobs" to Mexico. Several months ago, Ford itself mounted a vigorous campaign against the TPP, because of the inclusion of Japan; the company has since announced its withdrawal, after several unsuccessful decades, from the Japanese market.
All this has led pundits to pronounce the re-emergence of "Reagan Democrats" - mainly white, unionised blue-collar voters in Macomb County, a suburb of Detroit - who historically voted Democratic, but in 1980 swung to vote for Republican Ronald Reagan instead. They hypothesise that this group supported Mr Trump and Mr Sanders rather than Mrs Clinton, because they are anti-trade. But this argument is far too simplistic.
First, the US car industry has been booming, ratcheting up record sales and profits based on strong domestic market demand. Outsourcing the manufacture of labour-intensive parts to Mexico has helped the US industry remain competitive with imports, despite the burden of the strong dollar. At the same time, Ford and General Motors' heavy investments in manufacturing in China have given them significant market shares in the world's largest car market, which, until the recent slowdown, was a major source of profits. The Chinese government's efforts to strengthen (not weaken) the yuan in recent years also add to the dollar value of profits remitted to the US.
Second, high sales and profits have not translated into proportionate job creation, but this is due mainly to technological change, especially heavy automation and use of robotics. The jobs created are increasingly not in blue-collar manufacturing, but rather in white-collar research, engineering, design, software, marketing and so on, which account for an ever-higher ratio of the automotive value chain. The industry has also been experiencing a skills shortage, with thousands of vacancies remaining unfilled because of a lack of suitably qualified workers to fill them.
Third, both Japanese and European (and a few Chinese) competitors and suppliers of the US carmakers now have manufacturing facilities in the US itself, employing American workers - in Michigan and Ohio as well as lower-cost South Carolina or Texas. Some car workers thus appreciate the value of two-way trade and foreign investment. For example, during the 2008-09 financial crisis, Fiat of Italy purchased part of Chrysler, helping it to exit from bankruptcy, while in my town of Ann Arbor, the largest employers (outside of the University of Michigan) are Toyota's Technical Centre, followed by Hyundai-Kia's.
Fourth, unionised car workers' numbers and share of the Michigan electorate have shrunk dramatically since the Reagan years. The state's population is also more diverse today: as far back as 20 years ago, there were already a reported 10,000 Chinese-surnamed engineers working for the "Big Three" carmakers, and Dearborn, Michigan, the home of Ford Motor, is known as "the most Muslim city in America", with 40 per cent of its population being of Arab origin.
So, if it is not (mainly) a protectionist car industry, what explains the Michigan primary results, and how generalisable are they to the US as a whole, and to the presidential election?
One complication in interpreting the results arises from Michigan's "open" primary system, which allows registered voters to "cross party lines" to vote in the other party's primary. Some 7 per cent of Republican primary voters were registered Democrats, who could have voted for Mr Trump because they are "Reagan Democrats", or think Mr Trump is an easier candidate for the eventual Democratic nominee to beat (as polls predict both Mrs Clinton and Mr Sanders will). Or they could have voted for Mr John Kasich to "stop Trump". Similarly, Republicans could have "crossed over" to vote for Mr Sanders to "stop Clinton".
Primary votes do not necessarily predict votes in the presidential contest, and they can deliver some unexpected results. For example, Mr Trump's lead over his Republican rivals was greatest in areas with more college graduates, despite the belief that he does particularly well with "the poorly educated". Mrs Clinton's lead over Mr Sanders was also biggest in areas with more college graduates, despite his popularity among college students (Mr Sanders carried Washtenaw County, dominated by the college towns of Ann Arbor and Ypsilanti). And Mrs Clinton edged out Mr Sanders in Macomb County, the home of supposedly anti-trade "Reagan Democrats".
If trade policy was not the major differentiator among candidates driving voter behaviour (since Mr Trump, Mr Sanders and Mrs Clinton all expressed anti-trade sentiments), what was?
Many Trump voters were driven by the same issues that have carried him elsewhere, like immigration, nativism and, yes, racism (non-Muslims in Dearborn voted heavily for him). Both Trump and Sanders voters are also caught on the wrong side of the widening inequality divide, and object strongly to big banks and big business, with which Mrs Clinton is seen as collaborating. This is the reason for traditionally liberal college students' preference for Mr Sanders, the self-described "socialist", and for polls predicting a bigger victory for Mr Sanders than for Mrs Clinton over Mr Trump, since the two men partly compete for the same voter base.
For economists, however, both Mr Trump and Mr Sanders are seen as major dangers for the economy, because of their proposed trade, fiscal and monetary policies, probable inability to work with Congress, and the discouragement of business investment likely to result from either's election. This would slow the US' currently relatively healthy economic growth, hurting the already fragile world economy, and all of us.
- The writer is Professor of Strategy at the Stephen Ross School of Business, University of Michigan in the United States.