Fiscal Realities

A couple of things were raised last week that I want to address:

Issue 1: The Difference between Operating and Developing Expenditure

I’ve had to explain this at least twice over the last few days, so I thought I might as well spell it out. Malaysia is one of the very few countries that actually subdivides spending between operating and development expenditure – actually, I think Singapore is the only other country that does this. MOF keeps these accounts entirely separate (I’ll touch on how they intersect in a bit), whereas most other countries consolidate the two.

Here’s the main distinction between the different accounts:

1. Operating expenditure (OpEx) is recurring expenditure related to the running of the government, and is funded by revenue. Salaries, pensions, interest payments, office supplies and the like fall under this broad category.
2. Development expenditure (DevEx) generally refers to capital expenditure, but also covers things that are expected to bring a long term economic return. Schools, roads, hospitals and such, fall under this category. Development expenditure is partly funded by revenue whenever there’s a surplus on the operating budget, but mainly funded by borrowing.

For you accounting types out there, OpEx is the stuff that goes on your income/expenditure or profit/loss statement, whereas DevEx causes changes to the balance sheet.

The difference in funding is important, and is termed the “golden rule” – you should only borrow for investment, or purchasing assets, and never for consumption. As far as the data goes back, this rule has only ever been broken three times – in 1972, and in 1986-87 (RM millions):

In other words, the Malaysian government has almost always run a surplus on the operating budget, and borrowing has only been used on development or capital expenditure. Of course from a legal standpoint this is only an administrative rule, but it also a highly sensible one and in keeping with fiscal policy prudence.

The reason I’m raising this is because, in the runup to GE14, both sides are accusing the other of promising too much, and potentially “bankrupting” the government and putting government debt growth on an unsustainable path.

I don’t think so, given the contraints I’ve outlined above. Virtually all the measures mentioned in either manifesto refers to OpEx, not DevEx. Neither spending plan is likely to result in an expansion in borrowing, because the development budget won’t actually be affected. Keeping to the Golden Rule means that funding will have to come from either revenue expansion (from raising taxes or from organic growth), or from reallocation or cuts to OpEx. Unfortunately, there’s very limited scope for reallocation under the operating budget, so cuts are the more likely if revenue can’t keep up.

The biggest difference – and inconsistency – is with PH’s proposal to revert back to the old Sales and Services Tax regime (SST) and drop the Goods and Services Tax (GST), presumably at the old rates, though this isn’t mentioned explicitly. The other issue is to implement more efficient procurement and review “mega projects”.

What’s the implication of this? First, that means the more or less RM20 billion reduction in tax revenue will have a negative impact on the operating budget. Second, all the expenditure “savings” will occur on the development budget. The upshot is that PH’s manifesto is likely to result in a relatively smaller deficit than BN’s, but at the price of sharper cuts in OpEx. The government footprint in the economy will be smaller overall, but the delivery of essential services such as healthcare and education could be threatened, as these costs are rising faster than economic growth. From a sector balance viewpoint, this will also likely result in higher household and corporate debt (as happened in the 1990s when the government ran a surplus), and/or slower economic growth.

Pick your poison.

Issue 2: Debt comparisons within the economy

For the second issue, someone raised the question of what is the share of government debt within the economy. Most people would use gross government debt as the primary indicator of national indebtedness, but doing comparisons with this number across the economy is hard, because there’s far less transparency on either household debt or corporate debt.

Let’s try anyway. First, overall government debt (RM millions):

Since this looks frightening, let’s contextualise it by stating the above as a ratio to GDP:

The mean of the time series above is 54.0%, while the median is 49.6%. In other words, government debt is currently roughly in line with historical norms.

Getting household debt numbers is a little harder. BNM reports overall household on an annual basis, comprising both bank and non-bank debt, but I’ve only got figures from 2002:

…and as a ratio to GDP:

Corporate debt is even harder to figure out. There’s three basic components to corporate debt:

1. Bank loans
2. Capital market debt instruments
3. Others (including accounts payable)

Bank loan data is easy to access (but the time series only starts in 2006, due to definition changes), while capital market debt requires a bit of detective work, and some value judgement. For example, what would be the treatment for Cagamas bonds, or for that matter, Khazanah bonds? Second, there’s no time series available, so building one requires more assumptions on categorisation of the available issuance data. As for the third category, frankly I throw my hands up. So let’s just go with the first two (RM millions):

I’ll admit that the PDS estimate in the series above is more than a little fuzzy – it looks about RM100 billion too high, but since this is just to give a flavour of the amount (you’ll see why in a minute), I’ll let that pass. Here’s the same data as a ratio to GDP:

Now, all three put together:

Total debt rose from approximately 200% of GDP in 2006, to a peak of 245% of GDP in 2015. Here’s the attribution table and chart:

	Govt	Hhold	Corp
2007	2.1%	2.9%	3.3%
2008	3.1%	3.3%	4.9%
2009	3.9%	3.1%	3.1%
2010	2.8%	4.9%	3.8%
2011	2.8%	5.7%	4.6%
2012	2.3%	4.4%	5.1%
2013	1.7%	4.3%	3.0%
2014	1.8%	3.4%	3.1%
2015	1.8%	2.7%	3.5%
2016	0.6%	1.7%	2.2%
2017	1.4%	2.0%	2.9%

From the table, most of the increase in total debt actual came from either the household sector or the corporate sector. Only in 2009, when government revenue was hit by the lagged effect of the global recession, did government debt contribute more than the other two sectors.

Here’s government debt as a ratio of total debt:

The share peaked in 2010 at about 23%, and has been slowly declining after, though not by much. The increase hasn’t been really that much either – just a 2.5 ppt increase in the share of debt.

Of course, most of this is really beside the point. There’s two methodological issues when doing this exercise.

The first is that a significant portion of this debt actually offsets. For example, EPF and KWAP are large holders of government debt, and these holdings should be netted off. For the same reason, retail bonds and corporate housing loans should also be netted off.

The second issue is we’re completely ignoring the other large sector in the economy, i.e. the banks. But bank balance sheets in this instance are an exercise in offsets – they hold as assets the loans that other sectors have to repay, but owe the same sectors the liability arising from the deposits placed with them. The net effect is close to zero (actually something like 90%). Insurers are another case in point – they are big holders of government securities, but their liabilities are the corporations and households who hold insurance policies, who in turn provide the resources to pay off those bonds. This kind of circularity is why aggregating debt (or for that matter, anything else) is an exercise fraught with pitfalls.

On a national level therefore, it makes sense to only use external liabilities as a measure of national indebtedness. But here the direct government share is almost a rounding error:

If we include foreign holdings of government securities (which makes sense in the context of an absence of the circularity argument above), it rises substantially, but not much more than the total debt exercise I’ve conducted above – somewhere around 23% on average. The sharp increase you see there is due to a higher level of foreign ownership post-2009.

Technical Notes:

Most of the data is from BNM’s MHS, except for

1. PDS data from BNM FAST
2. Household debt data from various issues of the Financial Stability and Payment Systems Report