atans1

Heng needs AI to help him in making Budget forecasts

In Economy, Political governance, Public Administration on 20/02/2018 at 9:41 am

Because if my favourite fortune-teller had made the Budget surplus prediction of S$1.91bn that Heng made in 2017, she would lose all credibility. The 2017 surplus is S$9.6bn: 5 times or 503% bigger than projected last year. This is a miss of S$7.7bn, or, as Chris K points out, nearly 1.8% of GDP.

As usual the “blame” for the whooping error is put on stamp duty. And the next PM said this is a one-off. If I recall, this has happened more than a few times already. Still a one-off?

But Heng and the rest of MoF, and the entire PAP administration are not held accountable for getting the 2017 projected surplus horribly wrong.

Yesterday morning, in Budget: Consistently flawed/ Use more from Reserves meh?, I pointed out that the previous year’s Budget surplus is always bigger than predicted because

Consistently expenditures will be found to have been overestimated, and revenues underestimated

And that this tot was triggered by FT’s description of a Japanese mgt practice

[T]he pattern is too consistent for comfort, often strays into the deliberately deceptive, and is carried out as part of a habit of systemic conservatism

Let me be clear. I am not accusing anyone in MoF or the govt of being  “deliberately deceptive”. Here in S’pore, the pattern of underestimating revenue and overestimating expenditure “is too consistent for comfort and is carried out as part of a habit of systemic conservatism”).

Chris K spotted two more whopping misses in 2017 that are likely to be repeated based on the forecasts for 2018

Land sales revenue is estimated to be 12,2b for 2018 but for 2017, land sales revenues are revised from 8.2b to 12.9b. A revenue miss of 4.7b.

Investment income pertaining to interest and dividends only is estimated at 11.5b for 2018. But for 2017, it was revised from 10.5b to 17.5b, a whopping miss of 7b. Why I say whopping? Interest and dividends from an investment portfolio are fairly predictable, what is not predictable is the change in market value of investments. But the latter is not included so why such a large miss?

In total, both land sales revenues and investment income are 23.7b estimated for 2018 and revised upwards to 30.4b for 2017.

Facebook

Coming back to Heng and AI, maybe MoF should use IBM’s Watson cognitive computing innovation to help it improve its forecasting techniques.

After all in 2014,

DBS Bank and IBM today announced an agreement in which DBS will deploy IBM’s Watson cognitive computing innovation to deliver a next generation customer experience. This collaboration is part of an ongoing journey by DBS to shape the future of banking.

 

  1. so heng expected 2017 to be a bad year economically, with disappointing land sale prices/volume and poor investment returns; he was wrong, but I dont think he was the only one

    before trump was elected, people expected the market to crash if a surprise result occurs; since then the stock market rose 45%

    when janet yellan put up interest rate in december 2015, there was a slump in oil price, which brought down the stock market with it in first quarter 2016, and europe/japan both brought in negative interest rates; the fear turned out to be completely misguided; she continued to up the rate, but little damage to oil and share prices resulted in the rest of 2016 and whole of 2017

    is the world economy strong in 2018? who knows; I expect US stocks to crash because prices are just too high, but wont be held responsible for the prediction; please go after heng instead

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: