Over the past month economic data have been a confusing mixture ranging from lousy (e.g., first quarter GDP growth) to pretty good (e.g., April labor market data). The policy front has been equally confused. Our forecast has the first quarter growth shortfall reversing in the current quarter. But given the underlying confusion the probability of error is higher than usual.
The advance data for first quarter GDP raise more questions than they answer. On the pessimistic side, growth in total output was put at 0.7%, just one-third the fourth quarter value and the worst quarterly result since Q1 2014. This deceleration was due to a 1.7% decline in government expenditures and to consumer spending which registered growth of just 0.3%. Together these two components made a slightly negative contribution to overall growth. But there also are some optimistic signs in the first quarter data. First, final sales grew a more respectable 1.6%. This implies that reduced inventory accumulation cut 0.9% from overall growth. That will not continue. Second, unlike the situation in 2016, when consumption alone accounted entirely for growth, in the first quarter all of final sales growth (and then some) came from investment. [The first two sets of columns in Figure 3 on the next page depict these shifts.] Business investment in equipment, which declined at a 3.8% rate during 2016, increased at a 9.1% rate in Q1. Construction spending was even more dramatic: Investment in business structures surged 22.1% (compared to 1.9% in 2016); residential investment rose 13.7% (versus 1.1%).
In summary, the first quarter had nearly a total reversal of the patterns during the previous year. Consumption went from strength to weakness. Business investment and housing went from very weak to very strong. Trade went from negative to positive; government did the opposite. Looking ahead the questions are: Which of these shifts (if any) will persist? Which will prove to be transitory?
Recent monthly data give a few clues, but no definitive evidence, toward answering these questions. The labor market produced a strong April report. Payroll employment rose by 211 thousand, largely erasing the concerns raised by the meager 79 thousand March number. There were job gains across the economy although more so in service sectors than in good production. A BLS index of aggregate weekly payrolls increased by 0.7% (monthly rate). In the household survey unemployment declined another tick to just 4.4%, matching its lowest reading going back to 2001. The improvement was most dramatic for those at the bottom – unemployment for those with just a high school diploma (or less) was down 0.3% and the number of people working part time who would rather work full time declined by 281 thousand. Consistent with a healthy employment landscape real disposable income rose by 0.5% (monthly rate) in March. Consumer sentiment moderated slightly in April, but remained very strong.
All of this seems incompatible with a continuation of the dramatically weak consumer spending seen in the first quarter data. A return to the 3.1% growth experienced during 2016 is probably not in the cards, but a significant rebound from Q1 seems supportable.
On the investment front recent data suggest a little restraint for those inclined to break out the champagne. Housing starts fell back in March from a very strong February, when good weather may have accelerated construction activity. Industrial production rose 0.6% in March (monthly rate), but this was mainly due to weather related production by utilities. Manufacturing activity fell 0.4% for the month. The ISM manufacturing index which also declined in March had a larger drop in April, but at 54.8 it still indicates solid expansion. Their non-manufacturing index was very strong in April regaining the ground it lost in March.
Overall recent monthly data contain little that implies the extreme shifts in the GDP report on Q1 represent a permanent shift in the trajectory of the U.S. economy.
As can be seen in the chart, our new forecast has the economy bouncing back over the next two quarters from its Q1 swoon. The current quarter has growth at 3.3%, followed by 2.9% in the third quarter. After that, however, the economy settles back to “new normal” like growth averaging a little below 2.2%, which is slightly lower than in our February outlook.
At the disaggregate level our forecast for the rest of this year and 2018 strikes a median between the configuration of growth in 2016 and that in the first quarter. It has consumption far above the latter, but also below the unsustainable strength of 2016. Business investment and housing follow this pattern, except in reverse – below Q1, but better than last year. The return to growth in domestic spending (and the imports it induces) shift the trade balance back toward a rising deficit. Government spending resumes, and there is a small increase in inventory accumulation.
The most problematic of these shifts is the swing in business investment. Several comments on this: (1) The increase in investment is coming from the model, not from ad hoc adjustments. (2) Some improvement in Q1 came from the energy sector. For that to continue, oil prices probably need to stay around $50 per barrel (as they are assumed to do in the forecast). (3) The forecast investment growth rate (5.0%) is good, but far from unprecedented. Over the period 2010-2014, investment growth averaged 6.4%.
In the labor market, we have since February reduced our assumption for the long-run equilibrium unemployment rate from 4.5% to 4.3%. This has resulted in a lower trajectory for unemployment in the forecast.
A few additional aspects of our forecast scenario warrant mention:
- Our baseline forecast is based on a policy status quo. In particular this means that we have not introduced anything related to possible actions related to tax reform or to major changes on the spending side of the budget (for instance an infrastructure package).
- Inflation is a little higher in the current forecast than in our February outlook. Our outlook for the personal consumption expenditures price index, which was well above the Federal Reserve’s target of 2% in the first quarter, runs close to or above 2.5% for most of the forecast period.
- In our forecast we assume the Fed will implement three increases in the federal funds rate per year until early in 2020. The increase implemented in March is consistent with this assumption.
Usually our “baseline” forecast represents our best estimate about the course of the economy over the next couple of years, followed by a couple of years of our view of its long-run equilibrium trajectory. But this is not quite the case at present. The policy status quo assumption we have used for our baseline scenario is very unlikely to materialize in practice. The problem is that we have no idea how the actual policy environment will unfold. Our uncertainty exists on multiple dimensions. For one thing, at seven months past the election there is still very little concrete information about central Trump administration economic initiatives (tax policy and infrastructure at the top of the list). Even if there were, how it would emerge from the legislative process is imponderable. If it emerges at all. By the end of this year Congress will be absorbed in the 2018 election making any significant action much less likely.
In an optimistic scenario Trump et. al. implement a “successful” set of policies – reasonable tax reform, more cost efficient regulation, long-term infrastructure improvement, no trade war. This implies growth faster than our forecast, although not until next year and beyond.
But if the policy process stalls in Congress, that also amounts to deviation from the status quo. Even though policy itself might be status quo, perceptions would not. A lot of economic actors are sitting on the fence right now. A policy failure in the U.S. could cause some of those fence-sitters to move toward foreign projects.
This means that the probable error bands around our current forecast are larger than usual. The current recovery is two months short of eight years in age. With a successful policy outcome it could easily surpass the 1990s as the longest on record (10 full years). But if the Trump administration is not successful reaching that mark becomes unlikely.