Pages

Wednesday 28 August 2013

Repo Volume and Asset Values - A Simple Balance Sheet Analysis



There's been a couple of news stories in the last week on the repo market, relating to the potential impact of the bank leverage ratio, and on a proposal for a Fed fixed rate reverse repo facility.  The latter, in particular, is of particular interest to those who have expressed concern over the impact of QE operations on the pool of eligible collateral in the repo market.

The ability of certain investors to fund their positions with repo has a significant impact on their desire to hold those positions.  The size of the repo market therefore has an important effect on the overall demand for investments and by implication for the market value of securities.  The consequences of this should be clear, as a contraction in the repo market was one important element in the financial crisis.

However, the relationship between the volume of repo finance and the price of securities is a complex one.  A reduction in repo does not necessarily go hand in hand with a fall in asset values - it depends what is driving it.  The balance sheet analysis that follows is intended to illustrate in a fairly simple manner how this relationship could go either way.

The table below shows an extract from the national balance sheet.  This involves four entities: leveraged funds (LF); banks (B), unleveraged funds (UF) and the central bank (CB).  Leveraged funds are taken here to represent the sorts of entities (broker/dealers, hedge funds) that might typically fund themselves with repo.   The asset shown are Treasuries, high grade mortgage-backed securities (MBS), repo, reverse repo and deposits and reserves.  Here, I am taking repo and reverse repo to be from the bank's perspective.  That means that a repo is where the bank has effectively taken in a deposit against collateral and a reverse repo is where the bank has effectively provided a loan against collateral.


LF
B
UF
CB
Total
Treasuries
50.00

200.00

250.00
MBS
50.00

200.00

250.00
Bank reverse repo
-80.00
80.00


0.00
Bank repo

-80.00
80.00

0.00
Deposits

-120.00
120.00

0.00
Reserves






I am assuming in the above that the leveraged fund has repoed out all of its Treasuries and 30 of its MBS.  The remaining 20 of the MBS are ineligible as collateral.  For simplicity, I have ignored the haircuts - the additional collateral that has to be provided over the amount of cash raised.

So, I now want to look at what might happen if the providers of cash get concerned about the quality of the collateral.  In particular, I want to see what happens if they decide that they will no longer accept some of the MBS currently being repoed.  This is one element of what happened in the financial crisis.

To do this, I need to make some assumptions about the portfolio decisions of the leveraged and unleveraged funds.  To make it really simple, I'm going to assume that the unleveraged fund just maintains an equal investment (by value) in Treasuries, MBS and monetary assets (repo plus deposits).  I'm going to be a bit vague about the investment criteria of the leveraged fund, in order to keep to round numbers, but I'm going to say that the amount of Treasuries and MBS they wish to hold depends on the price.  The greater the price, the lower the expected yield, so the less they wish to hold.


LF
B
UF
CB
Total
Treasuries
47.50

190.00

237.50
MBS
35.00

190.00

225.00
Bank reverse repo
-70.00
70.00


0.00
Bank repo

-70.00
70.00

0.00
Deposits

-120.00
120.00

0.00
Reserves






The table above shows the new equilibrium position.  Various things have happened here.  The trigger is an amount of repoed MBS ceasing to be acceptable collateral.  This leads to part of the repo and reverse repo being repaid.  The leveraged fund finances this by selling the MBS it can no longer post.  These are purchased by the leveraged fund with the cash freed up from reduced repo. 

These cash movements are accompanied by movements in the market price of both MBS and Treasuries.  The attempt by the leveraged fund to offload the MBS it can no longer post pushes down the price, enough to induce the unleveraged fund to buy.  As the price falls, the unleveraged fund attempts to rebalance its portfolio by selling Treasuries.  This pushes down the price of the Treasuries as well and the unleveraged fund increases its holding slightly.  The fall in the price of Treasuries and MBS can be seen by looking at the total value in the last column.  Both prices fall, but that for MBS falls further. (Note that we are assuming that this extract balance sheet represents a "closed" system.  There are no sales or purchases with other investors not shown, nor any new issuance.  As aggregate sales is zero, equilibrium can only be attained by price movements and the change in the value of holdings is entirely reflective of changes in price.)

So this scenario results in a decline in asset values, as well as a reduction in bank lending and the broad money supply (reverse repo is a form of bank lending and repo is included in the broad money measure).

The second scenario I wanted to look at was a central bank purchase of Treasuries.  The starting point for this will be the same base case as before.  The mere announcement of the purchase will lead to an increase in the price of Treasuries due to the extra demand.  The rise in price will induce the leveraged fund to reduce its holding, selling Treasuries which the central bank ultimately acquires.  The sale proceeds are used to reduce the reverse repo position.  


LF
B
UF
CB
Total
Treasuries
45.75

203.00
10.00
258.75
MBS
49.50

203.00

252.50
Bank reverse repo
-73.00
73.00


0.00
Bank repo

-73.00
73.00

0.00
Deposits

-130.00
130.00

0.00
Reserves

10.00

-10.00
0.00

The unleveraged fund also sells Treasuries, prompted by the rise in price.  The unleveraged fund simply aims to keep a balanced portfolio, so as the value of its Treasury holdings rises, it must sell some in order to rebalance.  It uses some of the sale proceeds to acquire more MBS and some to build up its monetary assets.  The increased demand for MBS pushes up the price of those securities inducing the leveraged fund to sell.  The final equilibrium is shown in the table above.

The bank now holds a greater reserve balance, but this is offset by greater deposits from the unleveraged fund.  Overall monetary assets have risen, as has the value of Treasuries, the value of MBS and overall net private sector financial assets.  Note that although the overall value of Treasuries has risen, along with the price, the value of private holdings of Treasuries has fallen.

One main purpose of this exercise was to highlight what has happened to debt levels.  Although broad money (which includes repo) has increased as a result of the central bank asset purchases, lending (including reverse repo) has fallen.  This (reflux) effect might appear to be contractionary in that we normally associate a fall in lending with a reduction in demand.  However, although the decline in lending may itself be contractionary, it is actually just a partial response to an expansionary measure.  It reduces the impact of the expansion, but it can never be so great as to negate it.

In terms of the real economy, this form of lending (i.e. for asset purchase) can have no expansionary effect beyond its impact on the price of assets.   The test of whether the central bank asset purchases have been expansionary is whether asset prices have increased, not whether net lending has.  How expansionary an increase in asset prices is for the real economy is another matter.

Tuesday 20 August 2013

Macro Model - Update



I've been fairly busy updating my UK macro model for the 2013 Blue Book data, which included significant revisions to the calculation of fixed capital formation.  I have added more details to the page including a full equation listing and list of variables.

I'm fairly pleased with the way the model is working, but the more I do, the more areas it seems to open up to explore.

The graphs below show the results of a simulation over the period 1997 Q2 to 2013 Q1.  The solid blue line shows the actual data; the red dotted line the simulation.

 
 
 
 

This is for a full long term simulation.  This means that all lagged variables relating to periods from 1997 Q2 onwards are those generated by the model.  This inevitably means some compounding of errors, but the strong stock-flow framework tends to pull the results back towards the actual outcomes.