by Rick Davis
We have often commented about the disparity between shrinking per capita disposable income (which is contracting at an annualized -0.22% rate) and the BEA’s reported growth in consumer spending (reportedly growing at an annualized +1.91% rate). Assuming for the moment that the consumer spending growth is real and not an artifact of under-recognized inflation, we have proposed the following possible sources for the cash flows being used for the increased spending:
— Household savings rates have plunged from 5.6% to 3.6% over the past 18 months, freeing up $200 billion a year for spending,
— Student loans have added an additional $100 billion annually to consumer spending,
— Strategic and involuntary mortgage defaults have added another $80 billion per year in consumer disposable cash.
In addition to those sources, Dick Lepre kindly sent us a link to a Wall Street Journal article that estimated the scale of the household cash-flow benefits that have come from mortgage refinancing activities over the past few years. The bottom line from that article is that refinancing over the past three years has provided nearly $50 billion per year in improved household cash flows.
The total of all of these non-income cash flows ($430 billion) could have boosted apparent headline GDP growth by 2.8% through increased consumer expenditures — all without households actually having any sustainable or recurring increase in disposable income.
(In the above list we have not included changing energy prices, which generally just shift the same level of consumer spending between BEA categories (e.g., discretionary vs non-discretionary, durable goods vs non-durable goods, and goods vs services) while not growing or shrinking total consumer spending significantly. Each penny of price movement in gasoline prices has the potential to move about $1.3 billion in annual spending between such categories. However, if household budgets are already under stress, rising gasoline prices could also temporarily suppress savings rates until consumers can reacclimate to new spending priorities. On the other hand, our data has consistently shown that dropping gasoline prices quickly move spending into discretionary durable goods — as seen in late 2008 and again in mid 2011, when about $100 billion in increased annualized pocket money propped up the 4Q-2011 durable goods numbers and prematurely raised expectations of an accelerating recovery).
But we have been repeatedly asked: to what extent is this growth sustainable? And how long will the unsustainable portion last?
— At first glance the drop in household savings is a cyclical reversion to norm from the intense deleveraging that occurred during 2007-2010. (In fact, during the 12 months from June 2007 to June 2008 the personal savings rate skyrocketed from 2.3% to 6.2%, sucking $450 billion annually out of consumer spending and dropping GDP by 3.1% all by itself — arguably the real consumer demand trigger for the “Great Recession.”) Presumably the savings rate will stabilize at some new “norm” and this contribution to spending growth will disappear.
But: what exactly is the “norm” for the savings rate? In fact the long term average (line 35 since 1947 in the BEA’s Table 2.1) is just over 7%, with a record low of 1.3% set during the third quarter of 2005 (the “good old days” of easy credit) and the record high of 12.5% set during the second quarter of 1975. Statistically, the current rate of 3.6% is more than a standard deviation below the long term average (and the 2005 low was a “2-sigma” event).
Unfortunately, the long term averages may be meaningless as a consequence of a serious paradigm shift in household economics. The last quarter when the savings rate was above the long term average was in 1992. One might argue that households have not been able to maintain historic savings rates for the past 20 years, and that the economic stress felt by households actually dates back to that era.
If we accept that there is a new post-1992 savings “paradigm,” then the average savings rate is somewhat closer to 4%. And if we look at the “post housing bubble” era (2007 to date), the number is slightly higher at 4.5%. In either case (and given U.S. demographics) the current rate of 3.6% has little plausible downside movement remaining without households tapping into significant new sources of credit (e.g., student loans). Without massive re-leveraging at 2005 rates this contribution to the growth of consumer spending has likely run its course, and it could very well reverse again if consumer confidence erodes from continued weak job numbers or fear contagion from Europe.
— Student loans have apparently soared over the past few years, at growth rates that are orders of magnitude faster than tuition costs have inflated. There are a number of factors involved, and some of them are as mundane as accounting changes in the Federal Reserve’s G.19 report. At least some of the increase is an artifact of loans moving “on book” (from a Fed perspective) to the “Direct Loans Program” as a consequence of the “Higher Education Opportunity Act of 2008” (i.e., they are now being issued from accounts that the Fed is obligated to include in the report) and the implementation of FAS 166/167 (which encouraged the reclassification of securitized funds).
Yet there is likely more involved in these numbers — as a result of real changes in student circumstances and behavior. We suspect that many students are staying in school (or returning to school) for what are actually “economic reasons” (i.e., lack of employment opportunities). We further suspect that a significant number of expiring unemployment benefits have defacto morphed into extended student loans. We also suspect that this source of non-dischargeable loans may be one of the few lines of long term credit available to a vast demographic pool.
(For a more complete discussion of U.S. student loan programs in the context of the possible drag caused by higher education on U.S. economic growth in general, please see our previous commentary on the subject.)
But can this contribution to the growth of consumer expenditures be sustained? As bad as the $100 billion year-over-year increase may seem, that is actually lower than growth rates seen over the prior two years. And the most recent quarter-to-quarter (and month-to-month) changes also show a moderation of the growth rates. We suspect that students have begun to concern themselves with the “end game” for those non-dischargeable loans, especially when the lifetime economic benefits (or even short term job prospects) of their formal education seems to be shrinking.
The bottom line on student loans is that although the cash flows from the loans will probably continue, the growth rate of those flows is likely to moderate over the next few years — lowering their contribution to the growth rate of the GDP’s consumer spending.
— Strategic defaults and involuntary foreclosures have created “rent free living” that has freed up at least $80 billion per year in household cash flows (see our earlier commentary on strategic defaults for the details behind that estimate). There are a host of variables that can drive this number up or down, including the aggressiveness of lenders, the duration of foreclosure proceedings and the rate at which homeowners are initially going delinquent.
Our concern here is whether any of those variables are likely to change in the near future. In fact the percentage of residential loans in one form or another of “rent free living” has dropped by nearly a percent over the past year to 11.33% (per the latest release from the Mortgage Bankers Association). While this number is still extremely high by historical standards, it is lower than a year earlier — and the total number of households benefiting from “rent free living” has begun to drop. For this reason we suspect that any growth contributions to consumer spending from this source peaked last year.
Ultimately, for each such household there is a day of reckoning when they are forced to find a new place to live. Because of impaired credit most of these new places will be rented (if not the basements of close relatives). In such cases consumer spending will probably return to some intermediate level between pre-default rates and the more recent “rent free” rates — but in many cases with household savings rates boosted to offset the loss of the budgetary cushion provided previously by household credit. And presumably very few of those households would be anxious to rush back into the housing market, even if their credit was fully rehabilitated.
The bottom line on “rent free living” is that the growth it provided to consumer spending has probably already peaked, and the baseline spending behaviors of those households will take a long time to revert to pre-default levels.
— The good news with mortgage refinancing is that the raw household cash-flow benefits from refinancing are not only sustainable, they are “locked in” for the duration of the loans. Many of those households fortunate enough to have purchased homes before (or early into) the “housing boom” (and with conventional loan-to-value ratios) have actually seen their housing costs deflate — to the aggregate tune of $50 billion annually. (The flip side of that $50 billion is that the lending institutions have seen their revenues shrink accordingly — as well as the “net present value” of their truly healthy loan portfolios.)
That said, even though the cash flows persist, the spending growth contribution from refinancing will last only so long as refinancing transactions continue at previously unprecedented levels — they currently comprise nearly 80% of all mortgage applications (see our chart below and the latest Mortgage Bankers Association refinancing data for details):
As our chart indicates these numbers will fluctuate based with the public’s perception of interest rate direction — but at the present time the demand for refinancing continues to run at nearly double last year’s rate with no signs yet of a plunge back to more “normal” transaction rates. Clearly this source of “one time” cash flows still has legs, even if each round of refinancing has both diminished returns for homeowners and generally increased “term” deleveraging.
That said, at some point in time mortgage rates will no longer be able to fall enough to fundamentally change any homeowner’s refinancing cost/benefit calculations. Similarly, at some point in time the vast majority of eligible homeowners will have already refinanced at levels low enough that another round of refinancing would result in only marginal or no net gain. And even the procrastinators will have no new incentives to refinance as long as the Federal Reserve continues to assure people that the low rates will continue for some time.
(Once the vast majority of qualifying homeowners have already acted, the only potential sources for a renewed rush to refinance are: (a) the pool of existing loans that qualify for refinancing increase through dramatically rising home values, or (b) legislation radically changes the entire refinancing landscape.)
The bottom line on mortgage refinancing is that it is likely to continue to provide growth in consumer spending through at least the end of this year.
Of the total $430 billion (2.8% of GDP) in consumer spending growth identified above, only student loans and refinancing are likely to persist as sources of spending growth deep into this year. With student loan demand apparently moderating, we might expect the aggregate of those two sources to be around $100 billion this year (0.65% of GDP) — still providing growth, but at a rate too low to fully mask other structural weaknesses in the consumer portion of the economy.
Our inflation neutral Absolute Demand Index has continued to show consumer demand for discretionary durable goods running some 15% below the levels we saw in 2005:
To be fair, 2005 was arguably the peak year for the discretionary spending of the housing bubble — when households were spending on all aspects of housing, home improvements and appliances, even as they were leveraging themselves and their homes to the hilt. And when taking a longer view of the above data we still see reasons to be cautious, since last year’s summer rebound seems to be fading away:
One of our charts that best illustrates the general shape of what we have been seeing over the past year is our weekly Appliances chart:
We saw discretionary expenditures recover nicely last summer, in almost lock step with the drop in gasoline prices (in fact, the above curve forms something of a mirror image of gasoline price movements over the same time period). Other charts, however, show a more sustained spurt, including our weekly Clothing graph:
The key beneficiaries of the discretionary spending splurge included the major brand name department stores:
The surge in the second half of last year reflects more that just dropping gasoline prices. We feel that a significant amount of “frugality fatigue” had also set in by then, and households indulged in some degree of self prescribed “spending therapy” in the period leading up to the holidays — contributing to unreasonably high expectations for the holiday season itself.
Some of the classic discretionary spending targets have shown noisy but slowly eroding year-over-year growth, including branded travel destinations:
and Cruises, which have had even more noise than usual as a result of unwanted headlines:
Air Travel has perhaps shown the most consistent boost on a year over year basis, although even that seems to be reverting to more normal levels:
Meanwhile the previously moribund Game Console business recovered nicely in the late summer and fall of last year, before dropping back to neutral territory at year end, before again rising to double-digit percentage growth over the past quarter:
But the poster child for market saturation and the displacement power of the new phone platforms is in Cameras, which have shown year over year contraction now for over 40 months:
We are now entering the anniversary zone for the gasoline price drops that appear (at first approximation) to have triggered the growth in consumer discretionary spending last summer. We will continue to monitor our charts, but we expect that the recent highly favorable year-over-year comparisons will soon begin to lose their luster.