Economists argue that trade is good, essential in fact to growing an economy. If you excel at growing tomatoes, and I carrots, we will each have a healthier and more plentiful diet by bartering carrots for tomatoes. Once you add in far more people (across the globe no less), more goods (beef, cars, iPhones), services, and a monetary system that converts the value of any item to a price, gains in our simple local barter example grow exponentially. The world benefits when people trade, exporting what they are uniquely good at and importing what others are more efficient at making. We end up with more goods and services at a lower price when trade crosses city, state, regions and national boundaries.

But what is true in theory is not always true in fact. The theory of trade assumed scarcity of labor in developed countries.  Importing frees up workers to move into domestic industries needing labor. But we do not live in a world of full employment. In fact, we have globally deficient demand. So we need to think more carefully about trade.

Welcome to the “bad” of trade.

If one country artificially manages its currency to keep its exports cheap and its imports expensive, a trade deficit arises in the importing country. Because Gross Domestic Product depends on exports less imports, an increase in the US trade deficit reduces US growth. This pattern – of imports growing faster than exports until imports exceed exports – happened in the US with China for most of our trading years.

If one country imposes restrictions on trade, be they tariffs or other barriers to trade, trade is not fair, and we may import far more than we should be exporting under fair trade. The Chinese requirement to share IP and have a local partner as part owner of your business to sell in China is but one example. Dumping, a situation in which a country artificially under-prices its exports, creates a similar problem – you import far more than is justified by efficient markets.

Another problem-in-fact arises because labor is not fungible. The displaced manufacturing worker may not be able to switch occupations (to being a data analyst, for example) or move to locations where manufacturing jobs are growing. The slums of Milwaukee, Pittsburgh, Cleveland, and other Rust Belt cities reflect the loss of jobs initially to the South and then to other nations.

Another problem is that prices do not reflect total costs. The prices of Chinese goods do not incorporate the environmental damage caused by China’s weak environmental protection laws. When you read that the Beijing Airport is closed for three days due to smog, think about the unfair price advantage of the Chinese and the true cost (in Chinese lives and global warming) of our cheaper goods from China.

A final problem is a lack of economic resiliency when entire industries leave the US. Innovation emerges from the interaction of researchers, developers, and makers. When entire manufacturing output leaves a nation, as many parts of the electronics industry have done in the US, our innovation potential is at risk.

What about the ugly side of trade?

It emerges from our political system. Economists have always recognized that some people will benefit, and others will be hurt by trade. But, we argue, the net gains should be more than enough to compensate the losers. Unfortunately, we have not managed our unemployment system or labor policies to recognize this fact. Retraining due to trade-induced losses is neither uniform nor adequate.

Finally, today’s trade agreements like the Trans-Pacific Partnership are far more about managing trade than creating more trade. The agreement limits trade-based competition for pharmaceuticals, healthcare providers, Intellectual Property owners, and others while making lower-skilled markets easier for those wishing to export to the US. We already trade with most of the countries in the TPP, so there will be limited net gains in trade from this agreement.

Trade issues cannot totally account for the gap between our full-capacity Gross Domestic Product and today’s level and the resulting under-employment of US labor. We must take into account of forces such as automation, failures in our educational system, an archaic unemployment compensation system, mobility-reducing home ownership patterns, and growth in other nation’s capabilities. But there is no doubt that our continued trade imbalances mitigate growth and lower real wages in occupations and communities heavily influenced by trade. Estimates range around 3 million jobs or more lost due to trade inbalances.

I’m an economist and I am for trade. But Trump and Sanders are right to say we need fairer trade, although Trump’s solution – raising tariffs to 45% – is as ridiculous as his hair. Making trade fairer must be part of parcel of a smarter approach to creating middle-class jobs in a dynamic global economy.