ICI Viewpoints

ICI Viewpoints
Share

Policymakers Say Bond Mutual Funds Contributed Significantly to Treasury Market Stress but…

By Shelly Antoniewicz and Sean Collins
22-view-bondfund-survey-3-ff.png

Policymakers continue to assert that bond mutual funds contributed significantly to the distress in the US Treasury bond market in March 2020. They claim that heavy selling by bond mutual funds helped cause the dislocation in the Treasury market. They are incorrect, according to ICI daily survey data. Our survey shows that the Treasury market became significantly dislocated several days before bond mutual funds began selling Treasury bonds in any appreciable magnitude. And in the financial markets of March 2020, a single day was a lifetime. Moreover, bond mutual funds’ small share of Treasury trading volumes indicates their net sales must have had only a minor impact on the Treasury market.

COVID-19 Banner for Bond Funds

During March 2020, as financial markets reeled from the economic impact of the COVID-19 global pandemic and government-mandated closures, yields on Treasury bonds behaved uncharacteristically. Typically, during periods of financial market stress, Treasury bond prices rise and, equivalently, their yields fall. This is because market participants across the globe presume that the US Treasury market is the deepest, safest, and most liquid financial market and, consequently, investors shift toward US Treasury securities as a safe haven.

But events didn’t play out that way in the Treasury market in March 2020. Yields on Treasury bonds rose as the market increasingly came under stress. Demand for liquidity—which has been characterized as a dash for cash—escalated rapidly as investors sought to protect or bolster their cash positions in the face of tremendous uncertainty.[1] Demand for liquidity may also have been intensified by a range of factors, such as investors selling assets to get cash to meet margin calls or to deleverage their balance sheets, or by banks seeking to preserve cash in anticipation of corporations drawing on their bank credit lines.[2]

These developments led to an exceptional desire by investors around the globe to sell Treasury bonds, challenging the capacity of the Treasury market infrastructure to manage these flows.[3] Early in March 2020, the Treasury market began showing signs of stress. One such signal—the bid-ask spread on the off-the-run 10-year Treasury bond—rose significantly.[4] This indicated that market participants were becoming unwilling to hold US Treasury bonds that are almost always viewed as safe. Normally, bid-ask spreads for off-the-run Treasury bonds are quite narrow, in the range of 1 to 4 basis points, indicating that market participants who wish to sell can do so at minimal cost. Over the course of the next few weeks, the bid-ask spread for the off-the-run 10-year Treasury widened to more than 40 basis points—indicating a market under severe strain.

ICI Survey Data Cast Doubt on Narrative That Bond Mutual Funds Significantly Contributed to Treasury Market Disruption

Policymakers have asserted that bond mutual funds contributed significantly to the stress in the US Treasury bond market in March 2020. As we discussed in an earlier ICI Viewpoints post, this assertion is based on quarterly data from the Federal Reserve’s Flow of Funds Accounts, which are obviously not nearly granular enough to show what bond mutual funds were doing day by day in March 2020.[5]   

Our daily survey data[6] show that the Treasury market had become significantly dislocated several days before bond mutual funds began selling Treasury bonds in any appreciable magnitude. And when bond mutual funds stepped up their net sales of Treasury bonds in mid-March 2020, those additional sales added only marginally to Treasury market trading volumes, challenging policymakers’ assertions that the size of bond mutual funds’ sales “were one of the significant contributors to the stress.”[7]

Figure 1 makes this clear. It shows bond mutual funds’ net daily purchases or sales of Treasury bonds (shown as the blue bars). Bond mutual funds are net buyers when the bar is above zero and net sellers when it is below zero. The solid blue line is the bid-ask spread on the off-the-run 10-year Treasury. If the narrative is correct that bond mutual funds added substantially to Treasury market distress, the bid-ask spread should increase the most on days when bond mutual funds are heavier sellers of Treasuries. In other words, changes in bond mutual funds’ net sales of Treasuries should be correlated with changes in the bid-ask spread.

This correlation is far from evident. For example, the Treasury market starts to dislocate on March 6 with the bid-ask spread moving up out of its more normal range, jumping to 29 basis points by March 11. Over the same days, bond mutual funds, on net, sold only small amounts of Treasuries, $2 billion to $4 billion a day.[8]

It’s not until outflows from bond mutual funds pick up on March 12 that their sales of Treasuries become more substantial—$11.6 billion. But that day, the bid-ask spread rose only 1 basis point! The pattern is similar on March 13: bond mutual funds, on net, sold a sizable amount of Treasury bonds ($9.6 billion), but the bid-ask spread increased just 4 basis points. On March 13, bond mutual funds, on net, sold three times as many Treasuries ($9.6 billion) as they did on March 6 ($2.8 billion), yet the 10-year Treasury bid-ask spread rose by the same amount on both days—4 basis points.[9] In short, the bid-ask spread does not appear to rise the most on days when bond mutual funds are heavier sellers of Treasuries. This indicates that policymakers’ narrative that bond mutual funds added substantially to the distress in the US Treasury market in March 2020 is incorrect.

FIGURE 1
Treasury Market Dislocation Began Several Days Before Bond Mutual Funds Sold Treasuries in Earnest
Daily, February 28–March 31, 2020
22-view-bondfund-survey3-fig-1.png

Sources: Investment Company Institute and Bloomberg

The lack of a correlation between bond mutual funds’ sales of Treasuries and measures of stress in the Treasury market is underscored in Figure 2. The figure examines two-day increments for the six consecutive business days before the Federal Reserve announced on March 15 that it planned to inject liquidity into the Treasury market by buying $500 billion in Treasury securities. In the first two periods (March 6–9 and March 10–11), bond mutual funds had relatively small net sales of Treasury securities, yet the 10-year Treasury bid-ask spread rose sharply. In the third period (March 12–13), bond mutual funds had much larger net sales of Treasuries, but the bid-ask spread rose by a much smaller amount (5 basis points).

FIGURE 2
Bond Mutual Funds’ Treasury Sales Were Not Correlated with Treasury Market Stress
March 6–13, 2020
22-view-bondfund-survey3-fig-2.png

Sources: Investment Company Institute and Bloomberg

However, one might ask: “What about bond mutual funds’ stepped-up sales of Treasuries from March 12 to March 18 during the height of the Treasury market dislocation? Bond mutual funds sold, on net, $52 billion of Treasuries and the bid-ask spread rose 13 basis points.…doesn’t that prove bond mutual funds added significantly to stresses in the Treasury market?”

The answer is no. There were many other factors at work in the Treasury market at that time. For instance, a range of other market participants sought to sell Treasuries. As the Federal Reserve has indicated, investors outside the United States—which includes official institutions such as central banks—sold nearly $420 billion in Treasuries in March 2020.[10] Also, the Federal Reserve’s announcement on March 15 that it intended to undertake mass purchases of Treasuries may have facilitated sales of Treasuries—seemingly the objective the Federal Reserve hoped to achieve.[11] Finally, if bond mutual funds’ net sales from March 12 to 18 were creating significant impacts on the Treasury bond market, those sales presumably must have accounted for a large share of Treasury market trading volumes.

But they didn’t. In every week of March 2020, bond mutual funds’ net sales of Treasuries accounted for only a fraction of Treasury market trading volumes. As Figure 3 shows, even during the week of March 16–20, when bond mutual funds’ net sales of Treasury bonds were heaviest, they amounted to just 5 percent of Treasury market trading volume.[12]

It’s helpful to put this 5 percent figure in context. At the end of February 2020, right before stresses began appearing in the fixed-income markets, bond mutual funds held 6 percent of Treasury notes and bonds outstanding. Thus, during the week of March 16–20, their share of Treasury market trading volume was entirely consistent with their share of the Treasury bond market.

FIGURE 3
Bond Mutual Funds’ Small Share of Treasury Trading Suggests Minor Impact from Sales
Taxable bond mutual funds’ net sales of Treasury bonds as a percentage of Treasury market trading volumes, weekly, March 2020
22-view-bondfund-survey3-fig-3.png

*Calculated as the absolute value of taxable bond mutual funds’ gross purchases minus gross sales divided by one-half of total dealer-to-customer trading.

Sources: Investment Company Institute and Financial Industry Regulatory Authority (FINRA)

Conclusions

Policymakers continue to assert that bond mutual funds contributed significantly to distress in the US Treasury bond market in March 2020. As a result, policymakers are discussing measures that could affect investors’ ability to redeem mutual fund shares. Such measures also could be very costly and extremely challenging to implement—all to the detriment of the nearly 103 million US residents that invest in mutual funds. ICI survey data on the daily activities of bond mutual funds during March 2020 indicate that this assertion is incorrect. Our survey data—by far and away the most detailed data brought to bear on this narrative—show that the Treasury market became significantly dislocated several days before bond mutual funds began selling Treasury bonds in any appreciable magnitude. And even when bond mutual funds’ sales did pick up thereafter, they were a small share of overall trading volume and therefore must have had only a minor impact on the Treasury market.

It’s time to hit the Family Feud “Survey Says” buzzer on policymakers’ ongoing narrative that bond mutual funds were a significant source of Treasury market distress in March 2020. Bzzzzt.

 

Notes

[1] Although the events have been described as a “dash for cash,” it would be more apt to describe them as a “dash for US dollar liquidity.”

[2] For more information on developments in the Treasury market during March 2020, see Investment Company Institute, “The Impact of COVID-19 on Economies and Financial Markets,” Report of the COVID-19 Market Impact Working Group (October 2020), 21–32.

 

[3] See Lorie K. Logan, “Treasury Market Liquidity and Early Lessons from the Pandemic Shock,” remarks at Brookings-Chicago Booth Task Force on Financial Stability (TFFS) meeting, panel on market liquidity (October 23, 2020).

[4] “Off-the-run” refers to a Treasury note or bond that was issued before the most recently issued Treasury note or bond of a particular maturity. The most recently issued Treasury note or bond of a particular maturity is commonly referred to as “on-the-run.” The bid-ask spread of the off-the run 10-year Treasury is calculated using the bid and ask prices of the next most recently issued 10-year Treasury security.

[5] See Shelly Antoniewicz and Sean Collins, “Setting the Record Straight on Bond Mutual Funds’ Sales of Treasuries,” ICI Viewpoints (February 2022), for a detailed explanation on why the Flow of Funds Accounts substantially overstate mutual funds’ net sales of Treasuries in 2020:Q1.

[6] See Shelly Antoniewicz and Sean Collins, “ICI Bond Mutual Fund Survey Brings Facts to the Debate,” ICI Viewpoints (February 2022), for a description of the survey.

[7] See Financial Stability Oversight Council Statement on Nonbank Financial Intermediation (February 4, 2022), which argues that “Open-end funds were not the sole or primary cause of market stress [in March 2020]—there was no single, primary cause—but the size of their asset liquidations indicates that they were one of the significant contributors to this stress” (emphasis added).

[8] We examine net sales of Treasuries because we think this is a good measure of bond mutual funds’ ultimate net demand for liquidity in the Treasury market. Our results remain much the same when we look at daily gross sales of Treasuries by bond mutual funds.

[9] From March 15 on, the Federal Reserve began intervening in the Treasury market, which no doubt heavily influenced the bid-ask spread, thus clouding the picture between indicators of Treasury market stress and bond mutual funds’ net sales.

[10] According to the Federal Reserve, non-US central banks and other non-US investors sold nearly $420 billion in Treasuries in March 2020—four times more than bond mutual funds did. See “Setting the Record Straight on Bond Mutual Funds’ Sales of Treasuries.”

[12] Although not presented here, bond mutual funds’ gross sales of Treasury bonds as a share of dealer-to-customer Treasury trading volume lead to essentially the same conclusion: bond mutual funds’ sales of Treasuries constituted a minor share of Treasury market trading volumes.

Shelly Antoniewicz is the Deputy Chief Economist at ICI.

Sean Collins is Chief Economist at ICI.