When it comes to the value of the dollar, there are clear trade offs between American export competitiveness and import prices. How these opposing forces balance out nationally and across regions is the subject of a great deal of debate but surprisingly little empirical work. Paul Krugman, in a recent post, predicts advantageous outcomes if the dollar depreciates but cites no convincing evidence. Dani Rodrik’s recent paper, which is the most solid and relevant econometric work I’ve seen on the subject, finds that an under-valued currency increases economic growth. His results are compelling; yet, since they focus on developing countries, I decided to take a look at how a weaker dollar would affect the United States.
Though the method, described below, is debatable, the results show that export growth, driven by a forecasted change in the exchange rate, increases GDP growth enough to offset the negative effect of price increases. In other words, having a weaker dollar would strengthen the U.S. economy.
The first step in performing this exercise was to forecast the exchange rate over the next 10 years. What we need is at least one variable that strongly affects the exchange rate, but is not significantly affected by it. I choose global GDP to play this role. As other economies grow, international investors recognize the benefits of purchasing assets in non-U.S. markets regardless of the dollar’s value. Why, for instance, should Chinese investors accept a pittance from U.S. treasuries or a modest return from U.S. stocks when they can earn triple the return for real estate development in Brazil, India, or Indonesia?
With the key variable selected, the econometric part of this involves regressing the quarterly exchange rate (the Federal Reserve’s trade-weighted index) on world GDP, the U.S. share of world GDP, the log of U.S. population, the number of countries in the WTO, and dummy variables for decades and the post Cold War era. The WTO variable signifies more international openness to trade which could affect U.S. exchange rates. I did this using data from 1973 to present. I was helped in this by using forecasted data from Moody’s Economy.com for the U.S. and World GDP figures, and I used the Census Bureau’s annual population forecasts. The results predict a 10 percent devaluation in the U.S. dollar index between 2009 and 2019.
The second step was to estimate the effect of the predicted exchange rate on U.S. exports and prices. First, using only historic data, I regressed U.S. exports of goods and services on the predicted exchange rate, and the other variables from above, except global GDP, which is forced to work only through the exchange rate. I then did the same thing to estimate a comprehensive price index, which combined (using factor analysis) the CPI, and the producer price index for intermediate goods, crude materials, finished metals, and farm products.
So what, then, is the relationship between the relative value of the dollar and export intensity (which is defined as the value exports as a percentage of all trade)? The relationship is strongly negative overall. This simply confirms the intuition: A weak dollar is good for exports. The major exception is the period from 1988 to 1993, which was marked by rising export intensity and export levels. This may be explained in part by the collapse of the U.S.S.R, which left a massive hole in the global economy that was filled, in part, by a surge in U.S. exports to Eastern Europe, Asia, and Latin America.
To conduct the final step, I plugged in the estimated level of exports and prices into a GDP growth rate model, adjusting for the variables mentioned above, with the exception of world GDP and the U.S. share of world GDP.
It turns out that the positive export effect was large enough to outweigh the negative price effect by an average of 1 percentage points of GDP for each quarter from the end of 2009 to 2019. That effect is implausibly large on its surface, but other factors in the model are pushing GDP growth down. The bottom line is that a depreciated dollar, on balance and at this point in history, looks good for economic growth.
In more detailed work, we will be investigating how these macroeconomic factors will play out across metropolitan areas and industries. Places dominated by the domestic-oriented service sectors, like health care, personal, and business services should remain relatively unaffected. Some metro economies will surely lose out, such as those reliant on imports for retail sales or inputs into production, and we should design policies to help them adjust, but other more export-oriented metros like Rochester, New York, San Jose, and Detroit stand to profit handsomely.