Yangzijiang Financial Holding

[CY09]

Yep, in the upcoming AGM, I will be asking the question on how they provision for impairments based on the 02 statements we know so far, having track YZJFH results/presentations.

1) In the past YZJFH has claimed that the pledged collaterals are 2 times the value of the principal;
2) Despite the DI portfolio being reduced by 34% (980 million) 2.908 billion to 1.928 billion, YZJFH has increased the allowance for impairment by 3 million from 255 mil to 258 mil. This means the credit provision has rose from 8.8% to 13.3%

Why such a large delta in provision despite the company's claim in the past

[weijian]

YZJFH's allowance for impairment rose from 255.1 million to 258.6 million (2023 vs 2022); this was after setting aside an allowance of 39 million in its income statement. This year they redeemed about $0.980 billion and the actual impairment was about 36 million.

Comparing 2022 vs 2021, YZJFH's allowance fell from 398million to 255 million which meant when they redeemed their debts there was quite a substantial amount of losses. This was on top to the 130 million of allowance it recognised for 2022. For year 2022, there was a substantial amount of DI which turned sour and liqudated value was lower than the principal. That year YZJFH redeemed about $1.02 billion in debts. So the amount of realised impairment was [(398-255)+123] million. Approx 266 million was impaired when they redeemed $1.020 billion in DI

hi CY09,

About 6-9 months ago, I was educated by our VB accountant Yoyo on this forum about the movement of its DI. So I think when you calculated realized impairment is 36mil and 266mil for FY23 and FY22 respectively, I do not think it should be represented that way.

A few notes that I learnt based on Yoyo's education back then:

(1) NPL classification and final calculated allowances are all Mgt's estimates. I think they are not useful and so made the decision to ignore them. Allowances are not cash cost and the quantum is also dependent on how much total DI there is. So for the "398mil and 255mil" that you mentioned, are estimates - they are not cash but they are recognized on the B/S (a bit like FVOCI type of accounting)

(2) Realized impairment is a cash cost. FY22=123mil and FY23=31mil (you mentioned 39mil but it is for 2H23) is the net realized loss.

[weijian]

Below is my simplified model of this whole DI thing:

NPL classification --> estimated allowances by Mgt (non cash) --> realized losses (cash)

Besides parent YZJSB's realized losses (ie. 44%), what would be a robust estimation of actual "allowances" (independent of Mgt), or in other words, would be a better leading indicator for future realized losses?

Rather than looking at what Mgt stated, maybe we should look at what Mgt did. Since there is a mandate to reduce DI and the goal hasn't been reached so far, it might be prudent for the OPMI to fully adopt the spirit of "amend, extend and pretend". As such, any "additions" to the DI is construed as "bad debt":

FY22 (pg14 of FY22 ppt), all sums in mil SGD
Additions: 1,444.8
Redemptions: 1,638.6
Realized losses (cash cost): 121.5
Allowances estimated by me = (1,444.8+121.5)/(1,444.8+1,638.6+121.5) = 1,566.3/3,204.9 = 50%

1H23 (pg16 of 1H23 ppt), all sums in mil SGD
Additions: 448.7
Redemptions: 452.0
Realized losses (cash cost): 8.5 (reversal)
Allowances estimated by me = (448.7-8.5)/(448.7+ 452.0-8.5) = 440.2/892.2 = 49%

FY23 (pg16 of FY23 ppt), all sums in mil SGD
Additions: 283
Redemptions: 968
Realized losses (cash cost): 34
Allowances estimated by me = (283+34)/(283+968+34) = 317/1,285 = 25%

In summary, these estimated allowances look close to the "realized losses" by parent YZJSB, so close that I am amazed at my own estimation. I think OPMIs should also see that at the end of FY23, the estimated allowances have reduced, meaning something positive probably happened in 2H23. It is also reflected on YZJFH's balance sheet at the end of FY23.

[donmihaihai]

YZJFH's allowance for impairment rose from 255.1 million to 258.6 million (2023 vs 2022); this was after setting aside an allowance of 39 million in its income statement. This year they redeemed about $0.980 billion and the actual impairment was about 36 million.

Comparing 2022 vs 2021, YZJFH's allowance fell from 398million to 255 million which meant when they redeemed their debts there was quite a substantial amount of losses. This was on top to the 130 million of allowance it recognised for 2022. For year 2022, there was a substantial amount of DI which turned sour and liqudated value was lower than the principal. That year YZJFH redeemed about $1.02 billion in debts. So the amount of realised impairment was [(398-255)+123] million. Approx 266 million was impaired when they redeemed $1.020 billion in DI

hi CY09,

About 6-9 months ago, I was educated by our VB accountant Yoyo on this forum about the movement of its DI. So I think when you calculated realized impairment is 36mil and 266mil for FY23 and FY22 respectively, I do not think it should be represented that way.

A few notes that I learnt based on Yoyo's education back then:

(1) NPL classification and final calculated allowances are all Mgt's estimates. I think they are not useful and so made the decision to ignore them. Allowances are not cash cost and the quantum is also dependent on how much total DI there is. So for the "398mil and 255mil" that you mentioned, are estimates - they are not cash but they are recognized on the B/S (a bit like FVOCI type of accounting)

(2) Realized impairment is a cash cost. FY22=123mil and FY23=31mil (you mentioned 39mil but it is for 2H23) is the net realized loss.

Allowances are not considered cash costs. However they will be matched with actual amounts at redemption. Any actual losses at redemption not previously recognised as allowance in income statement will be recognised at this point. The reverse is true for DI that the company collected in full but has previously provided allowances for, resulting in the release of allowance.   

So, what you see in the income statement is what the company provided plus under/over of previous provision. Which mean that any under provision by the management now will need to "payback" in the future. 

There is no two different ways to recognise allowance and cash cost. Expenses has 3 place to hit in the financial statements. 1) income statement, 2) other comprehensive income and 3) equity movement under changes in equity statement. And this type of expense can only be in 1)

The figures on the B/S 258M and 255M are cumulative .

[donmihaihai]

Just to add,

DI under current assets - 1.1B
DI under non current assets - 0.55B

since current assets is anything under 12 months, it is very hard to play with allowances in a given financial year as any under provision will hit the next financial year.

[jin]

from the ppt,
https://links.sgx.com/FileOpen/20240228%...eID=790307

slide 17,
i noticed that project C and D have recovered amounts below principal but they are classified under "Fully Recovered NPL"

is this normal?

[weijian]

from the ppt,
https://links.sgx.com/FileOpen/20240228%...eID=790307

slide 17,
i noticed that project C and D have recovered amounts below principal but they are classified under "Fully Recovered NPL"

is this normal?

hi jin,

If one has participated in auctions (or at least look at one), there are many instances where individuals continue to stay bankrupt as they still owe the bank after their property has been auctioned off net of fees or the value of the collateral has simply deteriorated too much.

As such, full recovery does not refer to "the amount has been fully recovered in terms of amt", but been able to "complete the recovery process".

[jin]

from the ppt,
https://links.sgx.com/FileOpen/20240228%...eID=790307

slide 17,
i noticed that project C and D have recovered amounts below principal but they are classified under "Fully Recovered NPL"

is this normal?

hi jin,

If one has participated in auctions (or at least look at one), there are many instances where individuals continue to stay bankrupt as they still owe the bank after their property has been auctioned off net of fees or the value of the collateral has simply deteriorated too much.

As such, full recovery does not refer to "the amount has been fully recovered in terms of amt", but been able to "complete the recovery process".

ah, i see, thanks for the explanation.

so this "completed recovery process" would be the case for all NPLs eventually? 
basically saying, we did whatever we could and this is the end result?

[weijian]

As such, full recovery does not refer to "the amount has been fully recovered in terms of amt", but been able to "complete the recovery process".

ah, i see, thanks for the explanation.

so this "completed recovery process" would be the case for all NPLs eventually? 
basically saying, we did whatever we could and this is the end result?

hi jin,

You should read back this thread and estimate the end result yourself. For example, parent YZJSB has disposed the DI on their books (and it is supposed to be "problematic ones") in 1H23. That is a golden data point, IMO.

Also, things are dynamic and not static. So a few things to consider:

(1) When Mgt says collateral is "XX" of loan and well covered, does it means the valuation at the point of loan? Is it still the same when the loan comes due in the future? Has Mgt updated the valuation of their collateral continuously?

(2) Another point to consider is the difference between "theory" and "practice". Valuation by valuers is "theory". Actual sales proceeds is "practice". When the collateral is seized and exercised (ie. auctioned on open market for example), what price do they really fetch on current market? The old adage of "theory meets practice" is not theoretical but practical.

(3) Finally, while parent YZJSB's 1H23 disposal is "golden data point" but that was already in the past. What is the discount base rate moving forward? The intelligent investor needs to figure it out himself/herself.

[jin]

As such, full recovery does not refer to "the amount has been fully recovered in terms of amt", but been able to "complete the recovery process".

ah, i see, thanks for the explanation.

so this "completed recovery process" would be the case for all NPLs eventually? 
basically saying, we did whatever we could and this is the end result?

hi jin,

You should read back this thread and estimate the end result yourself. For example, parent YZJSB has disposed the DI on their books (and it is supposed to be "problematic ones") in 1H23. That is a golden data point, IMO.

Also, things are dynamic and not static. So a few things to consider:

(1) When Mgt says collateral is "XX" of loan and well covered, does it means the valuation at the point of loan? Is it still the same when the loan comes due in the future? Has Mgt updated the valuation of their collateral continuously?

(2) Another point to consider is the difference between "theory" and "practice". Valuation by valuers is "theory". Actual sales proceeds is "practice". When the collateral is seized and exercised (ie. auctioned on open market for example), what price do they really fetch on current market? The old adage of "theory meets practice" is not theoretical but practical.

(3) Finally, while parent YZJSB's 1H23 disposal is "golden data point" but that was already in the past. What is the discount base rate moving forward? The intelligent investor needs to figure it out himself/herself.

hi weijian, thanks for the points highlighted, will read and think through.