Thursday, July 30, 2009

Inflation? What Inflation?

I've spent the last couple days sitting in a seminar at INTAN on 1Malaysia. So far its been an interesting experience, with some good thoughtful speakers (as well as some not so good ones). There are some takeaways I'd like to talk about here, not so much on 1Malaysia but some of the nuggets that got bandied about, particularly on distributional issues. But this post will be tangential to that.

The seminar was supposed to be opened by the Chief Secretary to the Government yesterday, but due to a scheduling conflict, INTAN replaced the morning session with an impromptu 1hr talk by Lee Heng Guie, head of economic research at CIMB Investment.

No big surprises in his MY economic outlook - weak recovery in developed economies, and Malaysia should see positive growth in the second half of this year. He gave me some ideas that I'd like to follow up on statistically (US PMI as a leading indicator for MY exports for instance), but he repeated a stylized fact that is unfortunately all too common - quantitative easing (aka printing money) in the US will lead to higher inflation and a depreciating USD going forward. I'll buy (with reservations) the depreciating USD story, but inflation is by no means a given and its source will not be due to the Fed's liquidity injections or printing presses.

I'm seeing this high inflation story a lot in both the press and the blogosphere - this BusinessWeek article is one mild example of inflation hysteria (on a sidenote - Arthur Laffer is a prominent economist? Prominent maybe, but calling him an economist of any standing outside conservative circles had me choking).

You can get even more extreme drivel if you do a Google search for "libertarian gold nut" or Ross Perot. If I sound caustic on this subject, it's because I have little patience for a seriously outmoded form of monetary system that is so obviously economically unbeneficial (you can find my thoughts on gold here and here). Zimbabwe and the Weimar Republic frequently appear as models of comparison - as if the situations are at all comparable. The Fed is very far from triggering increased inflation, much less hyperinflation.

But what's wrong with the popular picture of QE driving inflation expectations? On the face of it, higher inflation is what we should expect: basic economic theory says that an expansion of the monetary base for a given level of output will necessarily turn up as an equivalent increase in the price level. As the saying goes, too much money chasing too few goods.

There's no doubt that the Fed, along with other major central banks have indeed conducted liquidity operations and QE on an unprecedented scale during the depths of the crisis. And under ordinary circumstances, this should indeed put upward pressure on the price level. The problem with this narrative is that it is essentially a static equilibrium analysis, which ignores two things:

1. The level of output has fallen below potential
2. The demand for money has increased

With slack in the economy, increases in the money supply will not cause inflation as both firms and workers lack pricing power. Firms can't increase prices without suffering loss in demand (and therefore profits), while workers can't demand higher wages in the presence of high unemployment. Also, under the present circumstances, firms and consumers prefer holding money rather than spending or investing it, which raises the amount of money supply that can be supported at a given level of output - again non-inflationary, as money velocity has fallen.

Most economists intuitively understand these factors, so much of the serious discussion is concentrated on what could happen after a recovery. If output closes on its long term potential and velocity rises back to its normal range in the next couple of years, won't the increase in the monetary base push inflation expectations along?

This is where the central banks' exit strategy becomes important. I've become far more sanguine on the ability of the Fed and ECB to unwind their balance sheet expansion than I was just a few months ago. There are a few reasons for my thoughts on this:

1. The monetary transmission mechanism is and will continue to be broken
The Fed's liquidity operations are not having an impact on the broader economy. The "massive" increase in the monetary base is not turning up as a corresponding increase in the broader money supply - the money multiplier has fallen. In fact M2 growth is hovering around its long term average of around 7% despite double-digit M1 growth (log annual changes; 1960-2009):

There are a few things going on here, of which the first is the increase in the demand for money I mentioned above. Secondly, banks are still recovering from the shellacking of the past year, and aren't that eager to lend especially as house prices are still trying to find a stable bottom and unemployment continues to increase. This situation is something that will in all probability take a few years to resolve, which gives time for the Fed to reverse its purchases of securities. Third is that the ongoing deleveraging of the shadow banking industry will continue to exert downward pressure on liquidity. And finally, the Fed (and a few other central banks) are using some unorthodox methods, which leads to my next point.

2. Paying interest on reserves
It used to be that banking reserves kept at the central bank attracted no income. This meant that banks would immediately utilise any excess over statutory reserve requirements, as reserves still had to be funded - reserve money was a dead loss. Now the Fed is paying overnight rates on all reserve funds, which means that banks should be indifferent to keeping their money in their reserve account or the money market. The importance of this is that the Fed can fine tune their control over the timing of shrinking their balance sheet. They don't need to mass dump securities on the market to mop up excess liquidity and potentially take a loss - they can raise the interest rate on reserve funds instead.

3. The Fed isn't actually printing money (at least, not so much you'd notice)
QE is such a bad couple of words that any hint of it has inflation hawks yelling blue murder. But how much monetization of government debt is the Fed actually doing? They have on their books just under US$700 billion, or something like RM2 trillion, worth of treasuries. On an absolute scale that sounds like a lot of money - it's rather larger than for example twice the entire M3 money supply of Malaysia. But scaled against the actual US national debt, it's somewhat less than massive at under 9% of estimated treasuries outstanding (held by the public) for 2009. If you include official holdings of treasuries, the ratio is even less. More importantly, the Fed is buying T-bills off the open market rather than at auction, so the effect on the monetary base is actually net zero and also has the impact of capping long term interest rates.

Based on these factors, I don't think monetary expansion will support a rise in inflationary expectations over the medium term - inflation is more likely to come about from elevated commodity prices. What I think will happen however is a general rise in interest rates, especially at longer tenures as the Fed only directly controls the 1 month rate.

I don't think the Fed has that much appetite for more QE than it absolutely has to - they're committed to another US$300 billion by September, but there's lukewarm support on the FOMC for continuing that program if I'm reading the signs right. And higher interest rates will also provide some notional support for the USD - which makes the USD collapse story less likely as well.

Technical Notes:
US money supply data from the Federal Reserve, estimates of US national debt from the Congressional Budget Office. The Fed's current holdings of treasuries is available here.

Thursday, July 23, 2009

One Good Reason Why Central Banks Should Be Independent: Politicians Don't Understand Monetary Policy

This article on Bloomberg caught my eye yesterday. The relevant passage is:

"Representative Alan Grayson, a Florida Democrat, questioned what authority the Fed used to lend hundreds of billions of dollars through currency swaps to central banks around the world.

'One of the arrangements is $9 billion for New Zealand -- that works out to $3,000 for every single person who lives in New Zealand,' Grayson said. 'Wouldn’t it have been better to extend that kind of credit to Americans rather than New Zealanders?'"

Here's Bernanke's reply:

"Bernanke countered that 'we are lending to all U.S. financial institutions in exactly the same way' and that 'we have a longstanding legal authority to do swaps with other central banks.'"

...which is a nice way of saying Rep. Grayson doesn't have a clue of what he was talking about.

What happens when central banks do a swap? Using the NZ example, the Fed loaned the RBNZ US$9b. Can the RBNZ use this to increase credit in the NZ economy? How so when the USD is not legal tender in New Zealand? Getting that USD into the NZ banking system in NZD form implies a contraction of the domestic portion of the money supply, not an expansion, unless it's fully sterilised. So why do the swap?

Because NZ, just like everybody else in the aftermath of the Lehman collapse, faced a flight to safety of foreign investment and domestic capital which caused a spike in USD demand. There was a currency mismatch between the foreign currency assets and liabilities of the banking system. If the central bank's international USD reserves were also insufficient, then NZ banks would have failed to meet their international obligations. This would have an impact not only on NZ's credit standing, but also on the counterparties on the other side of the transactions.

Not so bad if you're domestic: all that happens is that your capital can't leave. But what about foreign creditors? And all those hedge funds and banks who played the forex carry trade (remember NZ's high deposit rates? Was it only last year?)? The swap lines the Fed engineered allowed US firms to call back their foreign-based capital and bolster both cash and capital reserves right when they needed it most.

Failure to meet USD obligations would have greatly exacerbated the liquidity and credit crunch of late-2008. The NZ swap line of US$9b was relatively small - the ECB got US$200b.

The net actual effect is that the Fed and participating central banks transformed private sector USD liabilities into official sector USD liabilities, with the corresponding increase in credit worthiness. It also meant that any USD the Fed actually lent out through the swap lines came right back to the US.

Having Congress overlooking the Fed's shoulders gives me the willies. If Rep. Grayson is any example of the average level of economic competence there, politicians and monetary policy shouldn't mix - and that goes double for BNM and Parliament.

Mark Thoma at Economist's View has a video and a nice discussion going on about the same subject of Bernanke's testimony. Plus a correction: RBNZ's swap line was $15b (of which none was actually utilised!), and the ECB got $300b not $200b.

Wednesday, July 22, 2009

Deflation? What Deflation?

Today's CPI report from DOS brings some positive news: inflation is back, albeit very mildly. Don't pay any attention to the year-on-year number (log annual change; 2005=100):

Disinflation was always on the cards after last year's runup in petrol prices. What matters now is that the price level appears to have stabilised:

And monthly price level changes are positive for the second month running, the first time that has happened since last August (log monthly change; 2005=100):

Why is inflation at this juncture positive news? Because it signals recovery in domestic demand. To underscore this point, rising prices in June are not due to food or petrol price increases but is rather more broadly based.

To check this I stripped the (volatile) food and transport categories out of the CPI, which leaves an approximate measure for "core" CPI. In retrospect, looking at the core inflation trend shows that Malaysia was in little danger of a deflationary spiral this year (log annual changes; 2005=100):

Mea culpa - I should've done this earlier. I'm going to try and extend the core measure back as far as I can, and it will probably force me to reassess BNM's monetary policy moves as well as real interest rate measures. The lower core inflation indicates that monetary policy is in fact tighter pre-2009, and looser this year, than I thought it was.

Technical Note:
I used an arithmetic average of the remaining weights after stripping out Food (31.4%) and Transport (15.9%). That approach is probably questionable - I prefer using geometric averages when I can, but time constraints did not permit. Plus stripping away what amounts to half the CPI is also grounds for caution - but the core measure does explain why BNM was relatively unmoved by the commodity bubble of 2006-2008, so I'm going to keep calculating this and see where it leads.

Thursday, July 9, 2009

May IPI: Still No Heartbeat

Today's industrial production release has both good news and bad news.

The good news is that the decline in production has essentially stopped (in levels; 2000=100):

The major component indices have all stabilised, which seems to confirm we've reached a bottom, or at least a plateau from which further declines are possible but unlikely.

The bad news is that there's precious little evidence of a recovery as yet (monthly log changes):

Looking at the breakdown of the manufacturing sub-sectors, it appears consumption and contruction related industries are starting to see some action, but electronics and electricals (which has the highest weighting) is still trending downwards. China's huge stimulus package, implemented since December last year, may explain some part of this improvement; but I'm wondering how sustainable that is, since some of it leaked out into asset markets (both house prices and the stock markets have staged impressive recoveries from their lows).

Anecdotal evidence suggests that there may have been a reversal in E&E production last month, so hopefully we'll see some improvement when the data actually comes out for June.

Tuesday, July 7, 2009

Banks: Leverage and the Interest Rate Spread

I've been railing against the slow adjustment in the interest rate margin between what Malaysian banks are charging and their cost of funds (proxied by the overnight interbank rate). Turns out they're not alone in doing this.

An article on VoxEU* examines the profit record of banks during the Great Depression, and makes some telling comparisons with current developments. To make a long story short, commercial banks are probably going to remain relatively healthy (at least, those that don't actually go bust), especially compared to their investment banking brethren. What really caught my eye though is the remarks about the interest rate margin - Euro area and US commercial banks are charging approximately between 2.5%-3.0% above their cost of funds, which is suffcient to handle an average loan portfolio default of about 5% over the next four years. That actually corresponds nicely with what Malaysian banks are charging:

If that's the case, the spread's likely to remain pretty much as it is until we get a handle on actual default rates later this year. Since we're looking at the bottom probably occuring in 2Q2009, that means defaults should peak some time in 1Q2010 assuming a sustainable recovery emerges in the second half of the year.

So we're at or close to the bottom as far as lending rates are concerned, though I still think banks could have cut faster in response to the cuts in the OPR.

In another VoxEU article**, the Research Dept of the Bank of Italy looks at leverage ratios in the global banking industry over the last ten years or so. As you may know, leverage before the crisis, which is the multiple of assets over the underlying capital base (or alternatively your gearing level)***, was sky high in some of the more badly affected banks - the US investment banks were allowed from 2004 to leverage their balance sheet up to 30x, which to my mind was insane.

If you like big numbers, that's 3000% of their capital base. I was surprised to learn that the Euro area banks actually had much higher leverage ratios (one bank exceeded 60x), which may explain why the European banking system was as badly affected as that of the US and UK. In any case, it's clear that leverage was a contributory factor in the fragility of the international financial system.

What are the leverage ratios in Malaysia? Not that bad by comparison:

Of course, this doesn't include off-balance sheet assets, but since domestic banking institutions haven't been that active internationally, I don't think the risk element is any higher than what is implied by the "official" numbers. Having said that, this is one metric I'm going to pay attention to in future.

Technical Notes
*"Lessons from banking profits in the Great Depression", Daniel Gros

**"Financial sector pro-cyclicality: Lessons from the crisis, Part I", Columba, F; Cornacchia, W; and Salleo, C

***There's a difference between leverage as defined here and the risk weighted capital ratio (RWCR), which is a reciprocal of the leverage ratio but with a difference in calculation. The RWCR (min: 8% under Basle I, equivalent to approximately 12x leverage) is based on risk weighted assets, where certain asset classes attract lower risk weights, which effectively reduces the value of assets used in the calculation.

Monday, July 6, 2009

Trade in May: It's A Bumpy Ride At The Bottom

According to an economists' poll conducted by Bloomberg, May trade data was worse than expected. I'm left wondering how many of these guys actually run models, even ridiculously simple ones like I've used in this blog.

My two models rather neatly bracketed the actual results for May which came in as RM42946 for exports (-35.2% yoy log difference), with the forecast for the model based on seasonally adjusted data at RM42237, and the seasonal effect model forecasting RM43255. I'll have to agree however, that the results aren't that encouraging:

Log annual changes

Log monthly changes

Trade levels, seasonally adjusted

More or less what I expected - a bumpy ride on the bottom. While I won't go back on my April call for a turnaround, the other side of what I think will happen is also pretty clear - we're sitting on the bottom instead of rising to the surface. In short, any talk of recovery is still premature: I'm still thinking in terms of the past few months being more of an inventory adjustment than a sustainable uptick in economic activity. I'd say the odds are against a further worsening in conditions, but that doesn't necessarily equate to any meaningful closing of the output gap.

If my models' forecasts continue to be right, then trade is going to continue to be bumpy:

Seasonally Adjusted Model

Seasonal Effect Model

June forecasts from the models are as below:

Seasonally Adjusted Model:
Point forecast:RM39529, Range forecast:RM44427-RM34631

Seasonal Effect Model:
Point forecast:RM42499, Range forecast:RM47858-RM37140

Technical Notes:
June trade data from Matrade. Details on how the models were constructed are here.

Wednesday, July 1, 2009

May 2009 Monetary Policy Update

BNM at the latest MPC meeting kept the Official Policy Rate (OPR) at 2%, but average lending rates have thankfully kept coming down:

The latest May data shows average lending rates just a hair over 5%, which is the lowest on record (at least since 1980), although on a real basis, it's still fairly high at about 2.6%. The spread on lending (based on interbank overnight rates) is also nearer "normal" levels, but still too high for my comfort given the current circumstances:

Money supply growth has responded to lower deposit rates, continuing to decelerate (log annual changes):

But so have loans (log annual changes):

That's the first time loans have seen net repayments since Dec 2007. In the money market, the MGS yield curve has continued to steepen, particularly at the long end:

That's factoring in a lack of demand for longer tenures due to uncertainty, as well as a lack of supply. I'm a little unsure what to make of the evolution of the MGS yield curve. On the one hand, it approximates the spreads seen between 1999-2005, and as such might conceivably be seen as "normal" in one sense. On the other, the compression in spreads between 2005-2008 suggest that the lifting of capital controls and the float of the MYR allowed for greater interest and participation from foreign investors; which immediately suggests that the widening we're seeing now is a function of the pullout by foreign investors over the last year, and that we should see the yield curve flatten again once conditions stabilise. I prefer the latter explanation, but the other could equally be true.