Blog

The measures announced by the US Treasury to support Fannie Mae / Freddie Mac and the US mortgage-backed securities market are comprehensive and should lower the cost and increase the availability of mortgage finance.

Two aspects of the plans are particularly significant. First, Fannie and Freddie will expand their portfolios of retained mortgages and mortgage-backed securities until the end of 2009 before embarking on a controlled reduction. This contrasts with the UK authorities’ policy following the rescue of Northern Rock of reducing the bank’s mortgage book by £20 billion a year in 2008 and 2009. As argued previously, this has unnecessarily aggravated mortgage and housing market weakness.

Secondly, the Treasury will buy mortgage-backed securities issued by Fannie and Freddie in the open market, with no predetermined limit on the scale of purchases. Again, there is a contrast with UK policy: the Bank of England’s special liquidity scheme allows mortgage-backed debt to be swapped for Treasury bills but has had little positive impact on the prices of such securities because credit risk remains with the banks.

As the chart shows, the woes of Fannie and Freddie have been reflected in a rise in the spread between conventional mortgage rates and interbank swap rates in recent months. A lowering of their funding costs should allow mortgage rates to fall significantly.

A major investment bank believes UK banks will have tapped the Bank of England’s special liquidity scheme to the tune of over £200 billion by the time the drawdown period ends in October. To put this into context, £200 billion is the equivalent of 14% of annual GDP or 6% of banks’ and building societies’ total sterling liabilities.

According to the interim Crosby report on mortgage finance, outstanding UK residential mortgage-backed securities and covered bonds totalled £257 billion at the end of last year. The Financial Times reported in May that banks had created almost £90 billion of additional securities for use under the scheme. So there is sufficient paper available for the £200 billion estimate to be plausible.

However, it is difficult to find corroborating evidence of activity on this scale from Bank of England data on banks’ assets and liabilities. Banks obtaining Treasury bills under the scheme would be expected to use these as collateral for increased repo borrowing. Yet official data show a fall in banks’ repo liabilities of £33 billion between March and July.

It is likely that banks are channelling their SLS activities through off-balance-sheet entities. Such entities were previously used to issue RMBS to the market, with the proceeds routed back to the related bank. They may now be borrowing in the repo market using Treasury bills obtained under the SLS, again on behalf the parent bank. Consistent with this hypothesis, “intermediate other financial companies” increased their deposits with UK banks by an estimated £38 billion in the second quarter. As argued previously, this has resulted in a major upward distortion to M4 money supply growth.

If confirmed, would SLS take-up of £200 billion imply a significant beneficial impact on UK banks? To the extent that the scheme allows banks to avoid a step-up in funding costs when existing wholesale borrowing matures, the effect is to prevent further damage rather than provide a positive benefit. However, it should also have allowed some banks to reduce their average cost of funds.

The scheme is generating significant profits for the authorities. The fee charged on borrowings of Treasury bills is the spread between three-month LIBOR and the three-month general collateral gilt repo rate – currently 70 bp. Assuming an average spread of this level in the first year, and borrowing of £200 billion, the Bank of England would earn £1.4 billion from operating the scheme. This is a multiple of the income the Bank generates annually from its main revenue source, the system of cash ratio deposits, under which banks are required to hold a proportion of their eligible liabilities in a non-interest-bearing deposit at the Bank. SLS profits will presumably be remitted to the Treasury.

Unsurprisingly, my MPC and ECB models signal no change in rates at today’s meetings but both are consistent with an easing bias.

The MPC-ometer suggests either a 5-4 vote for unchanged rates (four votes for a 25 bp cut) or 6-2-1 (two votes for 25 bp, one vote for 50 bp – no prizes). Key contributors to the dovish forecast are the downward revision to second-quarter GDP, a fall in price expectations in the EU consumer survey and lower short-term gilt yields. Partially offsetting these factors are the weak exchange rate and higher share prices.

One caveat: the actual vote has been less dovish than the model’s predictions recently. However, historically it has sometimes been early in picking up shifts in the MPC’s thinking.

For comparison, the Sunday Times Shadow MPC has voted 7-2 for unchanged rates (two votes for a 50 bp cut).

The ECB-ometer, which signalled July’s 25 bp rise, is now suggesting a one-third chance of a cut. Factors contributing to the dovish swing include weak business and consumer surveys, a slight decline in inflation, slowing M3 growth and a fall in short-term bond yields.

With policy tightened only two months ago, officials are understandably reluctant to concede that incoming data now warrant an easing bias. However, there may be some minor changes to the language in today’s policy statement to the effect that either upside inflation risks have diminished slightly or – more likely – downside growth risks have increased.

The first revision to second-quarter GDP, published on Friday 22 August, downgraded growth from the first quarter from 0.2% to zero. According to press reports, this was the weakest performance since the second quarter of 1992 – the tail-end of the last recession.

Such statements are wrong because they compare the first revision to the latest quarter with final GDP estimates for prior quarters. In fact, there were three quarters after 1992 for which the first revision indicated no growth in GDP – the first quarter of 1999, the fourth quarter of 2001 and the first quarter of 2002. The latest data shows GDP increases of 0.4%, 0.4% and 0.5% respectively for the three quarters. Clearly, there is a significant chance that the current second-quarter number will be revised higher.

The OECD yesterday published forecasts showing GDP declining at annualised rates of 0.3% and 0.4% in the third and fourth quarters, satisfying one definition of a recession. However, the OECD quotes a standard error of 1.2% around its projections, implying the forecast declines are not statistically significant. Moreover, according to a footnote, the UK model has been revised to include residential property prices, which are thought to be important currently, but the change causes a deterioration in historical tracking performance. In other words, the OECD has intervened judgementally in the statistical forecasting process in a way likely to have resulted in more negative projections.

The OECD’s forecasts are hardly authoritative. Back in March, the organisation projected no change in US GDP in the second quarter. The latest estimate shows growth of 3.3% annualised.

The modest improvement in services sector activity reported in today’s purchasing managers’ survey for August is a welcome antidote to current excessive economic gloom. As suggested in a previous post, business and consumer surveys earlier in the summer were artificially depressed by the temporary surge in oil prices. Taking into account manufacturing and construction results released earlier in the week, the PMI surveys are consistent with a stagnant rather than contracting economy.

In an article discussing economic prospects and policy options, former MPC member Charles Goodhart has argued that the Debt Management Office should reduce gilt issuance in favour of increased sales of short-term Treasury bills. This is an excellent idea: it would help to arrest the recent worrying slump in broad money supply growth and is a more appropriate response to current conditions than a cut in official rates.

A government running a budget deficit injects money into the economy. If the deficit is financed by selling gilts to the non-bank private sector, the cash injection is reversed, leaving the money supply unchanged. However, sales of Treasury bills are less likely to have this sterilising effect, because they are bought mainly by banks rather than non-banks. So the combination of the budget deficit with Treasury bill financing boosts the money supply.

Annual growth in adjusted broad money M4 – excluding deposits of intermediate “other financial companies” – fell from 13.6% to 6.5% between June 2007 and June 2008. In the second quarter alone adjusted M4 rose by only 3% annualised. I think M4 growth of 6-8% per annum is consistent with achievement of the inflation target over the medium term. A lower rate of expansion would risk unnecessary economic weakness and an inflation undershoot.

Goodhart’s proposal offers a way of boosting monetary growth without cutting official interest rates – risky against a backdrop of high inflation expectations, sterling depreciation and fiscal deterioration. Suppose the DMO switched half of the financing of a £50 billion annual budget deficit from gilt sales to the non-bank private sector to Treasury bill sales to banks. Ceteris paribus, this would boost annual broad money growth by 1.4 percentage points.

As discussed previously, the authorities could also support monetary expansion and the housing market by slowing down the rate of contraction of Northern Rock’s balance sheet.

The Goodhart suggestion is the mirror-image of proposals made in 2006/2007 for the DMO to curb then-rampant monetary expansion by “overfunding” the budget deficit through additional gilt issuance. This would have helped to limit credit and housing market excesses.

The Treasury and Bank of England ignored such proposals at the time. As Goodhart notes, his latest policy suggestion would require more gumption on the part of the authorities than has been shown to date.

Last autumn I argued the consensus was too pessimistic about US economic prospects while neglecting the risk of serious weakness in Europe.

According to the latest vintage of data, US GDP rose by 2.2% in the year to the second quarter. This compares with increases of 1.5% in the Eurozone and 1.4% in the UK.

US pessimists have now rolled their forecasts of weakness into the second half of the year. GDP will slow sharply, they argue, as tax rebate stimulus reverses and slumping foreign growth hits exports. Real personal consumption in July was 0.4% below the second-quarter average.

I agree that the growth contribution of consumer spending and net exports will fall. However, the July real consumption number was depressed by a temporary spike in energy prices. Nominal spending actually grew by 0.2% on the month. Real consumption should revive in August and September.

Moreover, the stocks cycle and a diminishing drag from housing construction should compensate for smaller contributions from consumption and trade. The ratio of inventories to sales has fallen sharply, suggesting firms have been surprised by the resilience of final demand – see chart. Even a stabilisation of stock levels will give a sizeable lift to second-half GDP.

What could go wrong? A faster decline in employment could undermine my forecast of consumer resilience. August numbers released on Friday will be important but it would be surprising if firms stepped up job-shedding following solid second-quarter GDP expansion. Growth in withheld employment taxes – a coincident indicator of labour income – has strengthened in recent weeks, arguing against labour market deterioration.

Broad money supply and credit numbers have been boosted by the financial crisis and the Bank of England’s special liquidity scheme (SLS) has increased the distortion.

The money supply effect is the result of banks cutting back on traditional unsecured interbank lending in favour of various forms of secured loans. These secured loans are typically channelled through non-bank financial institutions, often bank subsidiaries, whose transactions are included in M4 and M4 lending.

The money holdings and bank borrowing of these “intermediate other financial companies” – to use the Bank of England’s terminology – have no implication for spending on goods and services and should be removed from the data for the purposes of monetary analysis. The effect is similar to “roundtripping”, which occurred in the 1980s when non-financial companies took advantage of interest rate anomalies to borrow from banks and redeposit the funds at a higher rate.

The distortion has increased since the introduction in April of the SLS, which has expanded the available pool of high-quality collateral against which banks can borrow and lend.

Headline M4 and M4 lending rose by 11.2% and 13.5% respectively in the 12 months to June. Stripping out intermediate OFC flows, the growth rates fall to 6.5% and 9.4%, according to the Bank of England.

In the second quarter alone, when the SLS was in operation, headline M4 grew at a 11.7% annualised rate but the adjusted measure is estimated to have risen by just 3%.

The level and pace of slowdown of adjusted M4 growth warrant consideration of an interest rate cut but the MPC is constrained by high and rising inflation expectations, sterling weakness and fiscal laxity, likely to be confirmed by the coming Pre-Budget Report.

MoneyMovesMarkets is taking a break next week. Please check back in September.

Is the economy now contracting?

A composite indicator based on activity and orders indices from the purchasing managers’ surveys is at a level suggesting a small decline in GDP in the third quarter. However, the surveys extend back only to the early to mid 1990s, so their history does not encompass a full recession, implying uncertainty about the relationship. Moreover, the latest results, for July, were probably depressed by a surge in energy prices, which has since reversed. An alternative coincident indicator with a long history and a good record of signalling economic contractions is the rate of change of job vacancies. Quarterly GDP falls have historically been associated with a three-month decline of over 10% in the stock of vacancies. The three-month change in July was -8%, suggesting a stagnant or very slowly growing economy.

What is the risk of a full-blown recession?

In the recessions of the mid 1970s, early 1980s and early 1990s, the annual change in GDP bottomed at -2.7%, -4.1% and -2.1% respectively. Forward-looking indicators have yet to signal comparable weakness. For example, a monetary forecasting model – with inputs including interbank rates, credit spreads, narrow and broad money supply growth and the effective exchange rate – projects a fall in annual GDP growth to about 0.5% by the first quarter of 2009 (see chart). Based on the model’s forecasting error, this implies a 30-40% probability of a recession – defined as an annual fall in GDP – in the first quarter. The MPC’s latest forecasts are similar but slightly more downbeat: annual GDP growth is projected at 0.1% and zero respectively in the first and second quarters of 2009 assuming unchanged interest rates, consistent with an evens chance of a recession.

How do current conditions differ from prior cyclical downturns?

In terms of the monetary model, there are three contrasts with prior pre-recessionary periods. First, the interest rate “shock” suffered by the economy has been smaller. In the two years before the onset of the last three recessions, the three-month interbank rate rose by an average of 680 basis points; in the last two years the increase has been 120 b.p. Of course, some households and firms seeking to raise new funds have experienced a larger rise but interest rate changes affect the economy partly via their impact on the debt servicing burden and disposable income of existing borrowers: the average rate paid on the stock of outstanding mortgages has increased by only 50 b.p. over the last two years. Secondly, narrow money typically contracts in real terms before recessions but is currently still growing: non-interest-bearing M1 – currency plus interest-free sight deposits – rose by an annual 7.3% in June. Thirdly, the effective exchange rate has fallen by 11% over the last year versus an average 1% rise in the year before the last three recessions.

Will the economy recover later in 2009?

The MPC’s forecasts show annual GDP growth recovering from zero in the second quarter of 2009 to 1.0% by the fourth quarter, reflecting easier credit conditions, a diminishing drag from higher energy costs and a boost to net trade from the lower pound. The monetary model forecasts only to the first quarter based on current data but can be used to project further ahead assuming no change in the inputs; this also suggests a revival in annual growth. However, risks lie on the downside – the inputs may deteriorate, with monetary expansion, in particular, likely to slow further, possibly significantly.

Will inflation return to target within two years?

Base effects will be significantly favourable for headline inflation during 2009. If energy and food prices fall back, a return to 2% or even below is possible by late next year. However, the medium-term outlook will be shaped by “core” developments. Annual CPI inflation excluding fresh food and energy rose to an MPC-era high of 2.6% in July and is likely to climb further for three reasons: pass-through of recent cost increases, continuing sterling weakness and earlier monetary excess. The lower exchange rate is lifting core inflation partly via higher non-commodity import prices – manufactured import costs rose an annual 7% in June, having been unchanged in the previous 12 months – but also by reducing competitive restraints on domestic firms’ price and wage decisions. Monetary trends influence inflation with a variable lag averaging two years. Adjusted for financial distortions, broad money growth peaked in the second quarter of 2007 and is now slowing sharply, suggesting inflation relief in the second half of 2009. However, any decline may be insufficient to reverse the earlier increase and return core inflation to 2% in two years.

Will the MPC cut official rates significantly?

Aside from its own forecast uncertainties, the MPC is constrained by three factors. First, Bank Rate at 5.0% is low relative to current inflation and – more importantly – inflation expectations. According to the latest NOP / Barclays Basix survey, households expect inflation to stand at 4.8% in five years’ time. Cutting the policy rate below public inflation forecasts would risk damaging the MPC’s inflation-fighting credibility, with negative longer-term repercussions. Secondly, reductions in UK rates unmatched by a similar move overseas – as implied by current market expectations – could exacerbate sterling weakness, thereby extending the overshoot in core inflation. Thirdly, the coming Pre-Budget Report is expected to deliver further fiscal stimulus, which – together with cyclical deterioration – may push public net borrowing up to 4-5% of GDP next year. Fiscal slippage further raises the risks to credibility and sterling if monetary policy is loosened substantially.

One way of assessing the downside for house prices is to ask how much further they would need to fall to achieve either a “fair” level by historical standards or a given level of undervaluation – on the assumption that markets typically undershoot on the way down.

The discussion is usually couched in terms of the house price to earnings ratio but – as explained in an earlier post – the rental yield on housing is a superior measure of valuation.

The chart below shows historical National Accounts data on the rental yield together with a current estimate based on prices having fallen 11% from their peak late last year (as suggested by the Halifax index).

The current estimated yield of 3.4% compares with a long-term average of 3.6% – a reasonable estimate of “fair value”. Assuming no change in rents, prices would need to fall a further 6% to bring the yield up to the average. The RICS survey of letting agents released today indicates that rents are still rising so a smaller decline would be possible.

Now suppose the market undershoots to the same extent that it overshot in 2007. The yield got down to 2.9% last year – 70 basis points below the average. A rise to 70 b.p. above the average would take it to 4.3%. This would involve a further 21% fall in prices from current levels, assuming unchanged rents.

The bottom line? A further 10-20% fall in prices would bring them to an attractive level for a long-term owner or investor.

The consensus expects US GDP growth to slow sharply from a currently-reported 1.9% annualised gain in the second quarter (likely to be revised up significantly next week), reflecting an unwind of tax rebate stimulus. I am more optimistic based on the lagged impact of Fed easing, energy price relief, the stocks cycle, a diminishing drag from housing construction and ongoing trade improvement.

There were glimmers of hope in business and consumer surveys last week. The first of the regional manufacturing surveys released for August – from the New York Fed – reported a significant recovery in expectations for future activity – see first chart. If confirmed by this week’s Philadelphia Fed survey, this could presage an improvement in the national ISM survey over coming months.

Meanwhile, the early August national consumer survey conducted by the University of Michigan showed a further recovery in household perceptions of the housing market: the percentage balance of respondents judging the present to be a good time to buy a home rose to a three-year high – see second chart. However, an increase in mortgage rates following the financial crisis at Fannie Mae / Freddie Mac may delay any impact on activity.