by Asset Allocation Committee

Beginning in the late 1990s and becoming an even bigger issue after Dodd-Frank regulations were imposed following the Great Recession, many market participants worried that bond market liquidity would be absent in the face of a crisis.  Strict regulations discouraged banks and brokerage firms from holding large inventories of corporate bonds, thereby precluding them from their former role as willing buyers in the face of a wave of selling by investors.  Compounding the concerns were the sheer size of flows into traditional open-end and ETF bond portfolios.  Over the 10-year period from 2009 through 2019, a total of $2.7 trillion flowed into these instruments, almost a third of which were in ETFs.  Entering the fray was the popularity of risk parity funds that levered their bond holdings.  Though each risk parity scheme has a nuanced approach, they all act in a similar fashion.  The concerns were that when risk parity funds and individual investors rushed to exit their bond holdings, market liquidity would be absent, particularly for corporate bonds.  Numerous studies and articles appeared over this time frame warning of dire consequences for bond investors, notably violent price movements due to the fact that banks and brokerage firms were no longer participating to a significant degree to assist in maintaining order to the bond market.

In March, the veracity of these concerns was tested as the bond market worked its way through the financial stress triggered by the pandemic.  During this episode, ETFs proved their endurance by moving into the void and providing the necessary stabilization of the market.  The use of creation/redemption units on the part of Authorized Purchasers [APs] mitigated the severity of the downturn, especially among investment-grade corporate bonds.  As the accompanying chart illustrates, spreads for BBB/Baa-rated corporates relative to the 10-Year Treasury gapped to their widest level since the Great Financial Crisis, yet quickly repaired.

An examination by Blackrock of the trading during this critical period in its largest corporate bond ETF, the iShares iBoxx Investment Grade Bond ETF [LQD], provides evidence that not only did the activities of the APs aid in maintaining market liquidity, but the market price of LQD actually led the price discovery process.  In other words, rather than the market price exacerbating the premium/discount to net asset value [NAV], the market price was the precursor of the daily NAV print.  Moreover, the direction of the market price of LQD often preceded the direction of LQD’s indicative value [IV], which is the real-time estimate of its fair value, based on the most recent prices of its underlying bonds.  Since LQD’s 2,169 underlying bonds trade over the counter, and therefore many may contain stale pricing, the price discovery proved invaluable to market functioning.  After the dust settled, it was evident that LQD and other bond ETFs provided market pricing that was at least as good as, and oftentimes better than, individual bonds.

Although the market function provided by LQD during the period is not necessarily indicative of every bond ETF in each bond sector, it does underscore the notion that ETFs can provide the necessary liquidity during a period of crisis.  The creation/redemption mechanism of ETFs allow for arbitrage opportunities and allow the supply/demand of the ETFs to achieve equilibrium with the value of the underlying bonds.  In essence, ETFs are now creating most (if not more) of the liquidity previously provided by banks and brokerage firms, often doing so with greater efficiency.

View the PDF

by Asset Allocation Committee

As the Federal Reserve expands its balance sheet and the federal deficit balloons, there is a legitimate concern about the potential for inflation.  In this week’s report, we will examine inflation from a theoretical perspective.

There are two simple tools we use to discuss inflation, the equation of exchange and the intersection of aggregate supply and demand.  To start, this is the equation of exchange:

MV=PQ

M is the money supply and V represents the speed at which it is spent; on the other side of the equation is P, the price level, and Q, output.  In the calculation of the variables, the right side of the equation is nominal GDP and M is one of the various formulations of the money supply.  However, the power of the equation, in our opinion, comes from what the variables represent.  For example, Q is best thought of as the productive capacity of the economy, the ability of an economy to procure goods and services.  It would include not just actual production, but also excess capacity.  And V isn’t just a residual; it can tell us the effectiveness of money policy.

The classical economists assumed that V and Q were fixed; V represented the institutional structure of money demand and thus only changed when spending and income patterns were adjusted.  Since prices were flexible, Q was always at full employment and thus didn’t change.  If these assumptions were true, any increase in M would lead to a proportional rise in P.  However, it turns out neither V nor Q were fixed; in some periods, increasing M led to higher price levels, but in others, it did not.

In the last episode of the Federal Reserve balance sheet expansion, velocity fell, leaving prices and quantity mostly unchanged.

This chart shows calculated velocity and the Fed’s balance sheet.  The gray shaded areas indicate periods of official QE.  Note that there is a strong inverse correlation between velocity and the balance sheet expansion, suggesting that households and businesses are not demanding the funds being provided to the banking system.  Thus, the inflationary impact of expansionary monetary policy has been reduced.  Given current risks in the economy and markets, we would expect the current balance sheet expansion to lead to a similar result in the short run—another decline in velocity.

The second tool is aggregate supply and aggregate demand.  Although neither of these can be calculated with any degree of confidence, we can use the tools for illustration.

The key point to this schematic lies in the supply curves, labeled S, S₁ and S₂.  The latter is what we believe is our current supply curve; demand has shifted from D to D₁.  One of the consequences of COVID-19 is that we expect the trend toward deglobalization to accelerate, which would likely mean a shift in the supply curve from S₂ to S₁.  As deglobalization is eventually tied to reregulation of the economy, we will shift to S.  Of course, the key to rising prices will be the path of demand.  The longer it takes for demand to recover, the less likely it is that inflation will return with any significance.  However, when demand returns, we will likely see upward price pressures; if rising demand coincides with the eventual shift to the terminal supply curve S, the markets could be in for a notable inflation surprise.  We don’t expect this terminal shift to occur in the next three years, but it is highly likely in the latter half of the decade.

Tying this back to the equation of exchange, the supply curves above are represented by Q.  So, as supply becomes increasingly constrained, velocity will need to fall further in order for prices to remain steady as the money supply rises.  If inflation expectations change, it would be reasonable to expect velocity to increase, which would tend to lead to higher inflation.  The key points from this analysis are that (a) Fed policy actions, in isolation, are not necessarily inflationary, and (b) constraining supply, which is an element of deglobalization, could lead to higher price levels once demand recovers.

View the PDF

by Asset Allocation Committee

The U.S. economy has been hit with shelter-in-place orders that have depressed consumption and reduced production.  The economic readings on Q2 will be historically bad, perhaps even worse than the Great Depression.  However, based on the assumptions that (a) aggressive policy support will continue, and (b) the pandemic will wane over time, either due to natural herd immunity or medical intervention, financial markets are assuming the impact of COVID-19 will be severe in impact but short in duration.  Although we mostly agree with that assessment, the rebound has much to do with how GDP is reported.  The common way is to annualize the quarterly change.  Thus, if we get a large decline in Q2, a very modest rise in GDP in Q3 will tend to look quite large.  This is not the only way to measure the change in GDP.  The year-on-year change will look less onerous in Q2 but won’t look as impressive in Q3.

Another way we like to look at the GDP data is against its long-term trend.  To forecast Q2 GDP, we are taking advantage of a new high-frequency index created by the New York FRB.  The index is designed to measure the yearly change in GDP—how much GDP has changed compared to Q2 2019.  It is currently forecasting about a 12% yearly change.[1]  We can then take that number and calculate what the level of GDP looks like compared to its long-term trend.

To generate this chart, we log-transform real annual GDP and regress a time trend through the data.  The important line is the lower deviation line.  There are two periods when GDP was well below trend; the Great Depression and the post-Great Financial Crisis to the present.  The Great Depression showed a massive drop in the level of GDP that took until 1942 to return to trend.  Although fiscal and monetary policy were expanding after 1932, it took war spending to finally bring the economy back to trend.  What has been disconcerting about the past expansion is that it was much slower than the long-term trend.  And so, the level of GDP continued to fall compared to trend; our estimate for GDP in the COVID-19 recession shows a downleg in growth that is rivaling levels seen in 1935.

The point of this analysis is that the trend should represent some degree of capacity.  It is possible the trajectory for GDP that held from 1901 to 2008 is no longer possible.  But, if that trendline does represent a normal level for GDP, it would suggest the economy can absorb aggressive fiscal and monetary policy expansion before inflation becomes a problem.  If inflation starts to rise before the trendline is achieved, it would confirm that the trendline is no longer valid; however, for the political class, it would seem natural to find out if the long-term trend line remains a measure of capacity.  In other words, we would expect even more deficit spending and easy monetary policy until inflation returns.  Given how far GDP remains below trend, the return of inflation may take a rather long time.

View the PDF

[1] Although this is not the focus of this report, the forecast level of decline is consistent with a 35% drop in annualized GDP in Q2.