Tuesday, October 23, 2012

Is the US breaking out of its liquidity trap?

Out of all of the paradigm shifts that have occurred since 2007, the liquidity traps impeding advanced economies are among the most worrisome, as they render monetary policy impotent. The liquidity trap narrative is ubiquitous now: against the backdrop of impaired household and financial sector balance sheets, private sector credit demand becomes inelastic, and if policy rates are falling from a sufficiently low base, they hit the zero lower nominal bound before credit demand constraints are alleviated.

The characteristic drop in money velocity in liquidity traps reflects the weakened private sector credit demand, which is due to a combination of: liability defaults/charge-offs, diminished marginal propensity to borrow, and elevated lenders' creditworthiness standards. The twin deleveragings in the shadow banking and household sectors, both due to impaired balance sheets, impede the flow of dollars through the monetary circuit. This depresses aggregate demand, while monetary policy transmission mechanisms remain frozen.

However, the US is now five years into its private sector deleveraging, and the stock of imbalances built up during the bubble have made significant progress in unwinding. Even while monetary policy transmissions remained frozen, nonperforming loans became charge-offs, shadow banking liabilities contracted, and large fiscal deficits put a floor under aggregate demand. And all the while, the economy climbed out of its liquidity trap, slowly but stably. Although deleveragings and liquidity traps are inextricably linked, sufficient income growth and credit velocity can allow for the economy to leave the liquidity trap, even as it continues to deleverage (through the denominator's impact). This allows for a virtuous cycle to return, as traditional monetary policy regains efficacy and feeds into aggregate demand growth.

Policy rate models, such as the Taylor and Mankiw rules, illustrate the existence of a liquidity trap with negative implied nominal rates. The Taylor rule, a simplified version of which is calculated by rate = 1 + 1.5*inflation + 0.5*output gap, targeted significantly tighter policy than the Fed pursued. This is perhaps due to the calculations of inflation and output gap. Because of the replacement of housing prices with owners' equivalent rent in CPI calculations, the housing boom and bust were not proportionately accounted for in inflation gauges. Furthermore, the typical calculation of the output gap, (GDP - potential GDP)/potential GDP, falls victim to the inductive reasoning that characterizes an extrapolation of past trends into perpetuity. A better proxy for "too-hot inflationary pressures" is the total capacity utilization (TCU) rate, which tracked the output gap very closely, until diverging during the early 2000s recession (when, it can be argued, the US growth trajectory began decelerating relative to prior trend - see chart below).

Source: FRED

Substituting a housing-adjusted PCE deflator for the inflation input and a TCU-derived indicator for the output gap input, we derive a modified Taylor rule model. Given the relative hawkishness inherent to the traditional Taylor rule calculation, we reason that the inflation coefficient is too high relative to the output gap coefficient, and we thus modified the inflation coefficient to 1.25. This Taylor rule model suggests policy should have been looser going into the early 2000s recession (again, because of a deceleration in the US growth trajectory), and was very dovish going into the 2007-09 recession, implying -5% policy rates at the heart of the crisis. However, it has since turned back upward, and it recently broke back into positive territory for the first time since the recession hit, presently implying an almost 2% policy rate. This suggests that the US has broken out of its liquidity trap, and positive deviations of model rates from policy rates will, going forward, serve to offset the accumulated negative deviations during the liquidity trap, rather than add to them. The chart below presents the federal funds rate, the simplified Taylor rule, and our model Taylor rule.

Source: FRED

However, given that the Fed's dual mandate relates to price stability and employment, rather than aggregate demand shortfall, the Mankiw rule, which takes inflation and unemployment rate as its inputs, is worth investigating. The rule calculates implied policy rates by the relation: rate = 8.5 + 1.4*(inflation - unemployment). Making a similar inflation substitute as in the above Taylor rule model, our adjusted Mankiw rule model implies that, unlike in the above Taylor rule model, policy was too tight going into the early 2000s recession, but too loose coming out of it. More interestingly, this model also implies a positive nominal rate, just recently breaking back above zero. The divergence between this Mankiw rule model and the model Taylor rule above likely is due to the unemployment input replacing the TCU-derived output gap input. Both unemployment and the output gap are meant to represent proxies for underutilized resources, and so we derive a new Mankiw rule, using an average of total capacity non-utilization (100-TCU) and unemployment as the new "underutilization" input. After a bit of trendline fitting, coefficients are modified for this new adjusted Mankiw rule, and the result gives a similar signal as the Taylor rule model above, including the lower implied policy rates in the early 2000s recession and the recent return to positive implied nominal rates. The chart below presents the federal funds rate, the Mankiw rule, our model Mankiw rule, and our TCU-adjusted model Mankiw rule.

Source: FRED

Finally, we present a compilation of all of the discussed rates, rules, and models in the chart below. As is evident, policy rate models are now all implying positive nominal rates.

Source: FRED

The debate about what constitutes optimal monetary policy is well beyond the scope of this article, and we acknowledge the crudeness and imprecision of our approach. However, three key insights present themselves:
  1. The assumption of extrapolated output growth trends may have caused monetary policy rules to systematically underestimate the fragility in growth in the early naughts, by overestimating overheating inflationary pressures.
  2. The substitution of owners' equivalent rent for primary residences, for house prices, in CPI and PCE inflation metrics may have caused policy to be excessively tight on the margin, due to underestimating inflationary pressures during the housing boom and overestimating inflationary pressures during the housing crash.
  3. There is a consensus among the various policy rate models tested and constructed that implied policy rates are now back into positive territory again for the first time since the zero bound was hit.
This presents interesting evidence that the US has escaped its liquidity trap for the time being and that credit demand will become more price elastic going forward as transmission mechanisms thaw. Certainly household and shadow banking balance sheets have a long way to go before being fully and sustainably rebalanced. But it is no coincidence that aggregate household liabilities (not their leverage ratio versus income) are finally approaching the zero break-even level, as well (see chart below).

Source: FRED

It is also no coincidence that the housing market -- the primary culprit behind the inability for private markets to clear and the asset whose corresponding liabilities' deleveraging have driven the household paradox of thrift -- is showing its first signs of life. The proportion of households with negative equity continues to decline, each new positive equity mortgage thawing the monetary transmission at the margin, and housing starts and homebuilder confidence are rising impressively. The chart below shows the relationship between housing starts and mortgage debt growth.

Source: FRED

Whether these housing "green shoots" are sufficient for a self-sustained recovery remains to be seen, but if these incremental upticks are the beginning of a new fixed residential investment cycle (as Conor Sen has been anticipating), they would reignite the virtuous process from higher private fixed investment to softened deleveraging to marginal employment growth to higher tax revenues to public fixed investment (a pathway described by Conor).

The Fed Senior Loan Office Survey's data supports this idea that the US is breaking free of its liquidity trap. The latter half of 2011 marked the beginning of a strong jump in the percentage of banks reporting greater demand for prime mortgage loans, while the same is true of auto, credit card, and consumer loan demand (see chart below). Liquidity traps are demand-facing credit constraints and the unlocking of this latent credit demand is a very supportive signal to the US economy's prospects for igniting a virtuous cycle within its private sector.

Source: FRED

One of the best characterizations of the impaired market clearing processes in the housing and labor markets is the charge-off rate on all commercial bank loans. The majority of the household's deleveraging (contraction in liabilities/disposable income) has been from defaults/discharges, more so than higher income growth, diminished net new origination, and forbearance. As such, a normalization in charge-off rates likely will represent the completion of the "involuntary" phase of household deleveraging, where capital structure impediments like negative equity prevent households from accessing credit. Charge-off rates have declined by about 2/3, suggesting they are about 80% through in the process of normalizing creditworthiness standards (see chart below).

Source: FRED

However, households are not the only sectors that are hoarding cash. Corporates have near-record margins, profits, and cash flows, but are using their cash to engage in capital structure arbitrage and other return streams resulting from Fed-induced borrowing cost manipulations (and their consequent impact on earnings yields). Substantial investment in labor and capital expenditures are still a far cry in this economy, but corporate propensity to deploy cash seems to be picking up on the margin, especially when the near-term negative prospective impact of the fiscal cliff is accounted for.

The Fed's financial repression has removed a substantially greater amount of net interest income from households than net interest expense, while massive federal deficits lower corporate interest burdens and repair corporate balance sheets. This acts as a subsidy from households (in need of net surpluses through which to deleverage) to corporates (successfully normalized to sustainable leverage and in possession of excess net surpluses). Corporations also take advantage of the cheap long-term financing by locking in favorable terms and repurchasing equity with some of the proceeds. However, going forward, corporations will have to actually deploy capital, whether for labor, LBOs, or M&As, as profit margins decline from record peaks, and organic sales and market share growth are required for continual earnings expansion. This is already starting to begin in every S&P 500 sector besides IT and financials, as the chart below shows:

Source: Ed Yardeni

This leads right to a question Matt Busigin investigated this week: is it time for wage growth to return? Busigin presents a wage growth model whose inputs are nominal US Treasury interest rates of different tenors. The calculation, (5yr-10yr) + (1yr-2yr) appears counterintuitive and arbitrary on the surface, but may better be represented as the inverted 2yr fwd 5s10s curve, a measure of expected curve steepness, accounting for inflation expectations. With wage growth tied to the hip to core CPI, it appears that interest rate markets are pricing in wage growth in the US by next year. Busigin's model and actual wage growth are presented below (keep in mind, the visual relationship appears even stronger when scaled for different times and y-axes).

Source: FRED, Matt Busigin/Macrofugue

The following concurrent developments compel us to consider the prospect of the US escaping its liquidity trap:
  1. House price stabilization
  2. Federal funds rate models implying positive nominal rates
  3. Increased household loan demand
  4. Household liability growth approaching and breaking the zero YoY level
  5. Open-ended Fed credit easing (targeting MBS as opposed to USTs arguably qualifies this as credit easing) that overwhelms net supply issuance
As various others have noted, there are many signs pointing to the US beginning a transition from an underutilized, low-velocity economy plagued by the fallacy of composition, to a neutral, normalized-velocity environment in which labor and housing markets begin clearing in earnest and Adam Smith's invisible hand retakes control over aggregate income allocation. The most interesting point to us is that the Fed announced its open-ended MBS purchases against a backdrop of range high inflation expectations, rising housing prices, and slowing household deleveraging. This further reinforces the Fed's strategy of emphasizing the expectation channel, as the Fed is already beginning to follow through on its promise to remain reactive to employment as opposed to proactive to inflation.

Whether a suitably and sufficiently determined monetary sovereign central bank can generate inflation remains to be seen, but for the first time, there are significant tailwinds behind the Fed's back. Because of the household balance sheet issues at the heart of the broken monetary circuitry presently, the rise in housing prices directly increases the Fed's policy efficacy, as opposed to the quantity of the Fed's policy initiatives influencing lending decisions and house price appreciation (loans create deposits, not vice versa). The added kicker is that the Fed is engaging in these policies at the early stages of household liability stabilization. Where housing prices go from here will be vital to determining whether the US economy has exited the muddle-through or will be at persistent risk of slipping back into a liquidity trap.

But it is important to maintain perspective about the magnitude of the economic impediments plaguing the US. Specifically, the prospect of a fiscal cliff looms ominously, and corporate uncertainty about the fiscal cliff is feeding into conservative investment and project origination, at least through year end. Perhaps it will be the conclusion of the 2010-present Congressional economic deadlock that determines the difference between two binary outcomes:
  1. 2004/05 moment: Corporates focus on top-line growth, particularly domestically, invest in labor, M&A, LBOs, & cap ex, drive up wages, causing household incomes to accelerate, consumption to pick up, balance sheets to continue rebalancing, and credit accessibility increasing. Housing markets continue to clear more effectively and shadow supply continues to dwindle, while former negative equity mortgageholders successfully refinance as their assets rise in value. Employment growth picks up steam, alleviating state & local and federal deficit burdens and "crowding in" public investment (including for public sector labor demand, which has been one of the biggest headwinds to the recover).
  2. 1937/38 moment: Corporates brace for a demand drop off, while households garner incrementally smaller net surpluses and income, slowing and prolonging their deleveragings. Housing dips back down, unwinding any thawing of monetary transmissions, and US economic sensitivity to global output dynamics spikes. Labor markets face headwinds from the fall in aggregate demand, and the high base from which unemployment is starting would lead to excessive labor market inefficiencies and skills gaps. The trend since the recession of unsatisfied labor force members dropping out of the labor force to (re)enter school on student loans is likely to continue in this scenario. And with more than one fifth of student loans outstanding delinquent, such a high base of nonperforming loans from which to accelerate a boom in student loan demand could cause student loans to flip from net borrowing (mitigating the household mortgage deleveraging and shadow banking collaterealized lending deleveraging) to debt stock reduction (exacerbating the household deleveraging).
Either way, the association between economic performance and financial market returns will remain very nuanced and complex: after all, if money velocity picks up and corporates start investing in labour and M&A, employment growth will benefit, corporate margins will decline, labour share of income will rise, and equities will only benefit if there exists sufficient revenue growth to offset the margin contraction. US stocks and bonds both returned dramatic performances since 2009, while real GDP has hovered around 2% and unemployment around 8%. Could we be in for a stage in this economic cycle where real GDP accelerates to 3-3.5%, unemployment drops to a low 7-handle, and yet asset prices tread water? After all, earnings reports are starting to come in weak just as the median US households/employees/consumers appear to be bouncing back strongly.

I don't know the answer to these questions, but one thing I do hold conviction in is: if the US is exiting a liquidity trap and money velocity is about to pick up (albeit slowly -- these are not binary on/off things), then fighting the Fed will be a dangerous move, so long as the flow of newly-created reserve credits overwhelm net fixed income product issuance and crowds out private sector demand for risk-free securities, pushing investors in aggregate into more risk seeking. 

1 comment:

  1. InstantProxies.com offers private HTTP proxies with the most advanced proxy service on the web.