What stage of deleveraging are we in?
August 17, 2011
The Litman Gregory research team has been assessing how long it will take for the U.S. economy to deleverage and when we can expect earnings to revert to the old trend line, which has been adjusted downward slightly after the great recession. The bottom line is that we think deleveraging started in late 2008 and that it will probably take roughly 10 years to complete the process. By December 2018, in our base-case scenario, we will remove the earnings reset we have currently baked in due to the ongoing deleveraging. What this means is that going forward we will roll earnings forward at a slightly higher rate than we have in the recent past and, as a result, our fair-value point for the S&P will also increase at a slightly higher rate.
For those interested in more detail, below are notes from the research-team's discussion.
- According to Carmen Reinhart, unwinding of debt takes seven years on average. Of the 15 post-WWII financial crisis' she looked at, seven involved a double dip.
- Unlike smaller countries that were able to devalue and utilize exports to grow their way out, the U.S. will have difficulty in doing so. It is a much larger country and it needs large markets to be growing at a healthy rate. This is not the case with most of the developed world in deleveraging mode. Emerging-markets are a bright spot but it does not appear they are yet large enough to bail the U.S. and other developed countries out of trouble (we do factor in the fact that they are helping more than they have in the past).
- Similarities and differences of present day deleveraging with that of the 1930s
- Both the recent great recession and the Great Depression of the 1930s were global in nature.
- The U.S. total debt level and total credit debt as a percentage of GDP is more than it was in the 1930s.
- In the 1930s, we partly de-levered through "sharp defaults" (1930 to 1932-33) and then via a "belt tightening mode" (1933 to 1937), where the economy grew quite robustly (double digit real rates on the back of fiscal spending) which combined with inflation served to lower our total debt/GDP ratio. According to the consulting firm McKinsey & Co., credit growth was positive (1% to 3%) but less than nominal GDP growth, which also helped the economy de-lever. This time around, we experienced some defaults but not nearly the magnitude we did in the early 1930s. But ultimately we have to pay the piper, i.e., reduce our debt to manageable levels. So, that argues for a longer "belt tightening" period than the four years in the 1930s.
- In 1937, we cut fiscal spending slightly (and it seems raised taxes too), and we did this when our fiscal deficits in the 1930s were around 5% to 6%, much below the 10% levels we have now. The economy went into a recession, implying that deleveraging still had a ways to go and/or the private sector was not ready to step in and that the economy needed government support. It was a short recession though and the economy recovered in 1938 as the government started spending again (this is just pre-WWII). The spending in the late 1930s and pre-WWII appears to be not as significant as the one that took place during the belt-tightening period (1933-37), but still was enough to get the economy going. Then WWII happened—fiscal deficits blew out and the economy continued to grow at a rapid pace.
- Another stark difference is that the unprecedented fiscal stimulus (it is greater and was implemented more quickly this time) is not leading to robust nominal GDP growth today, at least not yet, as it did in the 1930s. It raises the question of whether our fiscal spending has been as productive this time around. Partly as a result of this, our total debt to GDP ratio has barely started going down today whereas in the 1930s it declined in a steady fashion during the "belt-tightening" period. That is a concern and argues for a longer belt-tightening experience than the 1933 to 1937 period.
Today, the government has tried to shift private sector debt (mostly related to the financial sector) onto the public sector. We think that is preferable because the government has more degrees of freedom, i.e., it can abrogate debt through inflation, currency devaluation, and even default (though the last option comes with consequences and is least appealing, especially to a reserve-currency country). That said, we believe the total economy-wide debt—whether in the public or private sector—matters, and are factoring that into our decision making.
We looked at various charts to assess how much we have de-levered and how much further we may have to go. If we look at debt-servicing charts, it would appear we are close to being done with deleveraging. But that assumes interest rates remain low at current levels, and we think that is not a reasonable assumption given markets could force rates up at any time. Moreover, debt-servicing charts do not fully reflect the principal debt that needs to be repaid. Therefore, we relied more on charts that show household debt as a percentage of disposable income. Looking at this, it seems we are about 1/3 of the way there. We think this chart accounts for various risks—rising rates, rising taxes, anemic income growth—that the economy and households face. Assuming deleveraging started in late 2008, and if it took us 2.5 years to get to 1/3, then maybe we'd be done in 7-8 years, also in line with Reinhardt and Rogoff and McKinsey findings. However, the process might have been sped up because of some defaults upfront and deleveraging going forward might progress at a slower rate. Hence, we might be deleveraging for more than 7-8 years.
We projected U.S. nominal GDP growth and household debt using what we think are reasonable GDP and credit-growth assumptions. This analysis suggests we could be deleveraging for more than 10 years. This makes intuitive sense. We avoided the short-term pain of massive defaults (which is part of the healing process but very painful) and that has likely come with a trade off in that we will be in a deleveraging mode longer than observed in the 1930s or than observed in other deleveraging episodes on "average."
We discussed the positives—Emerging-markets growth being a greater factor in offsetting domestic deleveraging pressures than we expect, financial repression that may make debt servicing manageable for a much longer period than we expect and allowing households to fix their balance sheets, U.S. dollar reserve currency status allowing us the ability to keep rates lower and grow more robustly than we expect, housing recovers faster than we expect.
We discussed the negatives—total debt to GDP is still high, household deleveraging based on total debt metrics is only 1/3 of the way and likely will proceed at a slower rate than seen thus far (because most defaults are behind us), pressures for government to de-lever are growing (likely, there will be fiscal cuts at an inopportune time, similar to the late 1930s), European debt crisis may make markets less open to letting the U.S. get away with fiscal imprudence, and the poor quality of growth we see stemming from China poses risks to the view that emerging-market growth will offset developed-markets deleveraging.
Weighing the positives and negatives, the historical evidence, and our own analysis of the kind of deleveraging we are undertaking (likely, an extended period of belt tightening), we think that deleveraging could take about 10 years, plus or minus a few. We estimate our deleveraging, especially in the household sector, to have started in late 2008. By end of 2018, we expect S&P's normalized earnings power to be around its long-term trend level. Starting now and until the end of 2018 we think it makes sense to slowly remove the earnings reset we have been baking in on account of the deleveraging we believe we are and will be going through (after all, the healthier parts of the household sector will start borrowing and spending). What this means is that we roll earnings forward at a slightly higher rate than we have in the recent past and, as a result, our fair-value estimate for the S&P also increases at a slightly higher rate.
(c) Litman Gregory