Thank you so much for this wonderful research. Questions- Does the outstanding volume of the basis trade force the Fed to overestimate the market signals of QT on repo as they approach LCLoR? As in without this basis trade related demand for repo, LCLoR would be lower. Also, in the era of plentiful reserves, if we meet this T-bill moneyness demand, do we enable fiscal largess at the expense of bank deposits -crowding out private sector credit creation? Once money is trapped in the Treasury debt loop, is it less likely than repo to return to bank deposits? Thanks, Tim
This is excellent, Steven, thanks. You highlight index replication needs, which rise when Treasury issuance rises, but there is also a endogenous demand for dollar assets. This would suggest the high yields currently offered by the Treasury short-end conflict with that demand. In the 2000s, we argued about a 'Savings Glut'. With ONRRP and T.Bill rates so high, perhaps we should start talking about a Return Glut for those savings, especially as it is clear other central banks cannot sustain high rates.
Your solution is logical, but I wonder how much incremental bill issuance would be required to move the needle on UST duration to truly reduce the need for asset managers to enter into these duration hedges. I also wonder whether the treasury should alter its funding plans to accommodate certain asset managers reaching for yield.
If you break this down this trade there are a couple of points. First, the fundamental demand is being driven by AMs demand for leverage (and duration-index mirroring). As the TBAC notes, these funds could simply pledge their own Treasuries and use proceeds to buy higher yielding credit. But this would make their leverage more obvious, and would create direct "interest expense" that the fund would report. So instead, they achieve this same economic structure via the futures derivatives by paying the HFs via the arb to hold leverage on their own balance sheet instead.
HF's have practically limitless ability to provide this leverage relative to their own capital because of their loss-sharing arrangement with the ultimate funds provider (repo lender), as dictated by initial and maintenance margin requirements.
It seems this step is the key point of concern - overly aggressive repo lending practices and (the lack of) appropriate and prudent margin requirements? And would it not make sense for regulators to simply attack that issue directly?
Thank you. Am I right in thinking that it’s solely by shortening benchmark duration that proportionally greater Treasury bill sales reduces the Treasury basis trade? You note that QE partially offset the duration-lengthening effect of Treasury deficits. Now that QT removed the offset, can we expect the trade to grow in size to a new record, it’s an arbitrage after all?
Thoughtful piece but there are several inaccuracies/misunderstandings.
The first issue is the overlooked error contained in Figure 27 of the recent TBAC report. The figure implies that drawdowns associated with a 20x levered basis trade would have exceeded the equity capital/margin of hedge funds on multiple occasions since 2018 and to a striking degree during the COVID pandemic. However, TBAC’s figure misconstrues the relationship between equity capital and basis trade drawdowns. The key error is that while returns to an unlevered trade that exceed 5% would indeed wipeout equity capital of a 20x levered trade (5% * 20x = 100%), the chart shows returns to the LEVERED trade. This is made clear by the fact that in Figure 21 the unlevered return in COVID is ~0.75% which corresponds to the levered trade drawdown in Figure 27 of ~15% (20 * 0.75% = 15%). The implication is that the levered equity capital drawdown 15% is very well short of a blowup! Either the chart should show a redline at -5% on and returns on an unlevered trade or a redline at -100% of the levered trade, but not a mix of the two!
The second issue pertains to this statement “But this trade, funded as it is in the repo market,1 tells the Treasury something else: There’s a ton of structural demand at the short end of the issuance curve.” The reasoning is backwards here. The basis trade returns are derived from the funding need required by asset managers long futures position. The demand for treasury futures creates a discrepancy between funding rates attainable in the repo market and implied funding rates on Treasury futures. The fact that hedge funds are able to obtain funding to extract the difference in funding costs does not imply that there is structural demand for short dated assets. In fact, the opposite takeaway could be true. Were Treasury to issue more bills, they may in fact “crowd out” repo funding supply and thereby exacerbate the dislocation between cash and futures. Said another way, the richness of short dated assets is largely irrelevant for basis trade returns. What matters is the relative richness of short-dated assets to implied funding. The presence of demand for the basis trade only indicates the existence of a positive funding/implied funding spread.
I also disagree with the conclusion that more issuance at the short-end will tend to reduce asset managers positions in treasury futures. As discussed at length in the TBAC presentation, index replication is only one source of demand for treasury futures. An additional source of demand is from the procurement of leverage afforded by futures contracts. For example, a common practice in pension investment schemes includes the purchase of futures to match liability duration with the left over cash being used to obtain equity exposure. Its precisely the levered component of futures that makes them attractive to asset managers!
All of this is not to say that there aren’t simple reforms that could be made. The fact that some mutual funds will purchase more costly treasury futures instead of repo-funding cash treasuries as a result of the treatment of repo in expense ratios is one example of a silly inefficiency that harms investors and does no meaningful good. That said, the overall thrust of this post needs to be reworked. If the target is to improve market resiliency of the treasury market by reducing the scale of the basis trade, the appropriate target should be the asset managers not the hedge funds.
Thank you so much for this wonderful research. Questions- Does the outstanding volume of the basis trade force the Fed to overestimate the market signals of QT on repo as they approach LCLoR? As in without this basis trade related demand for repo, LCLoR would be lower. Also, in the era of plentiful reserves, if we meet this T-bill moneyness demand, do we enable fiscal largess at the expense of bank deposits -crowding out private sector credit creation? Once money is trapped in the Treasury debt loop, is it less likely than repo to return to bank deposits? Thanks, Tim
This is excellent, Steven, thanks. You highlight index replication needs, which rise when Treasury issuance rises, but there is also a endogenous demand for dollar assets. This would suggest the high yields currently offered by the Treasury short-end conflict with that demand. In the 2000s, we argued about a 'Savings Glut'. With ONRRP and T.Bill rates so high, perhaps we should start talking about a Return Glut for those savings, especially as it is clear other central banks cannot sustain high rates.
Thanks so much for this! Your Substack is highly instructive and informative!
Steven, Thanks for this great note.
Your solution is logical, but I wonder how much incremental bill issuance would be required to move the needle on UST duration to truly reduce the need for asset managers to enter into these duration hedges. I also wonder whether the treasury should alter its funding plans to accommodate certain asset managers reaching for yield.
If you break this down this trade there are a couple of points. First, the fundamental demand is being driven by AMs demand for leverage (and duration-index mirroring). As the TBAC notes, these funds could simply pledge their own Treasuries and use proceeds to buy higher yielding credit. But this would make their leverage more obvious, and would create direct "interest expense" that the fund would report. So instead, they achieve this same economic structure via the futures derivatives by paying the HFs via the arb to hold leverage on their own balance sheet instead.
HF's have practically limitless ability to provide this leverage relative to their own capital because of their loss-sharing arrangement with the ultimate funds provider (repo lender), as dictated by initial and maintenance margin requirements.
It seems this step is the key point of concern - overly aggressive repo lending practices and (the lack of) appropriate and prudent margin requirements? And would it not make sense for regulators to simply attack that issue directly?
Thanks again!
Thank you. Am I right in thinking that it’s solely by shortening benchmark duration that proportionally greater Treasury bill sales reduces the Treasury basis trade? You note that QE partially offset the duration-lengthening effect of Treasury deficits. Now that QT removed the offset, can we expect the trade to grow in size to a new record, it’s an arbitrage after all?
Thoughtful piece but there are several inaccuracies/misunderstandings.
The first issue is the overlooked error contained in Figure 27 of the recent TBAC report. The figure implies that drawdowns associated with a 20x levered basis trade would have exceeded the equity capital/margin of hedge funds on multiple occasions since 2018 and to a striking degree during the COVID pandemic. However, TBAC’s figure misconstrues the relationship between equity capital and basis trade drawdowns. The key error is that while returns to an unlevered trade that exceed 5% would indeed wipeout equity capital of a 20x levered trade (5% * 20x = 100%), the chart shows returns to the LEVERED trade. This is made clear by the fact that in Figure 21 the unlevered return in COVID is ~0.75% which corresponds to the levered trade drawdown in Figure 27 of ~15% (20 * 0.75% = 15%). The implication is that the levered equity capital drawdown 15% is very well short of a blowup! Either the chart should show a redline at -5% on and returns on an unlevered trade or a redline at -100% of the levered trade, but not a mix of the two!
The second issue pertains to this statement “But this trade, funded as it is in the repo market,1 tells the Treasury something else: There’s a ton of structural demand at the short end of the issuance curve.” The reasoning is backwards here. The basis trade returns are derived from the funding need required by asset managers long futures position. The demand for treasury futures creates a discrepancy between funding rates attainable in the repo market and implied funding rates on Treasury futures. The fact that hedge funds are able to obtain funding to extract the difference in funding costs does not imply that there is structural demand for short dated assets. In fact, the opposite takeaway could be true. Were Treasury to issue more bills, they may in fact “crowd out” repo funding supply and thereby exacerbate the dislocation between cash and futures. Said another way, the richness of short dated assets is largely irrelevant for basis trade returns. What matters is the relative richness of short-dated assets to implied funding. The presence of demand for the basis trade only indicates the existence of a positive funding/implied funding spread.
I also disagree with the conclusion that more issuance at the short-end will tend to reduce asset managers positions in treasury futures. As discussed at length in the TBAC presentation, index replication is only one source of demand for treasury futures. An additional source of demand is from the procurement of leverage afforded by futures contracts. For example, a common practice in pension investment schemes includes the purchase of futures to match liability duration with the left over cash being used to obtain equity exposure. Its precisely the levered component of futures that makes them attractive to asset managers!
All of this is not to say that there aren’t simple reforms that could be made. The fact that some mutual funds will purchase more costly treasury futures instead of repo-funding cash treasuries as a result of the treatment of repo in expense ratios is one example of a silly inefficiency that harms investors and does no meaningful good. That said, the overall thrust of this post needs to be reworked. If the target is to improve market resiliency of the treasury market by reducing the scale of the basis trade, the appropriate target should be the asset managers not the hedge funds.