Tuesday, October 27, 2026

Update on Microsoft Copilot

I posted about my first impressions of Microsoft Copilot in Excel about three weeks ago. I exhausted the credits available for the standard home package in a week! So, I paid AUD 329 for the "Premium" tier. I am using it for two main things. As discussed in that post, I am debugging monthly accounts. The second use is building monthly spreadsheets from transaction level data. This is a huge time-saver.

A few years ago, I set up a spreadsheet that consolidated all the individual transactions from our bank and brokerage accounts into monthly closing values and various monthly aggregates: Investment income, non-investment income (salaries, taxes etc.), spending, and transfers. But the coverage of past months was sporadic, with only the last couple of years or so complete. It seemed like a project that might never get finished. The value of the workbook (dubbed Ledger) is that it makes compiling bug-free monthly accounts much easier and quicker. You can just copy last month's accounts and move all cell locations one cell forward. 

Copilot means that this project now might be completed–for all 30 years of data–relatively quickly. You can just ask it to fill the monthly cells in Ledger with various calculations on the individual transactions for that month from the relevant Excel workbook. By hand, you would need to adjust the range of cells to capture the relevant transactions individually for each month (or know very esoteric conditional formulae). So far, I have worked on the easier cases and will complete those first. Not all transactions spreadsheets are as nicely formatted as I would like. Copilot makes mistakes, but I can debug these using a worksheet that computes:

Residual(t) = Value(t)-Value(t-1)-Spending(t)-Transfers(t)-Income(t)

I then ask Copilot: "Why is there an error in this cell?" and it points me to the causes of the error.

Friday, October 16, 2026

Maas Group

Maas Group is closely connected to Firmus, which scrapped its planned IPO last week. MGH's share price more than halved as a result. But this selloff seems irrational to me, so I bought 20k shares. Firmus is not shutting down and Maas continues to work on their projects. The Firmus shares MGH owns were bought at a far lower price than the proposed float price. MGH recently sold a division for more than its current capitalization. Morningstar value it at $6.44 per share.

Found a $10k Error in Moominmama's Taxes

Throughout the year, I maintain a tax return spreadsheet for each of us, recording all relevant transactions etc. When tax time comes it doesn't take long to prepare our tax returns. I was updating Moominmama's spreadsheet today with her latest share transactions and found I made a massive error in this year's tax return (I start with the previous year's spreadsheet and modify it). I forgot to convert more than USD 20k of capital losses from Angellist into Australian Dollars. That means I overestimated her income by around AUD 5k. This is because only half of long-term capital gains are taxed under the current tax system. But the long-term capital gain is only halved after applying tax losses. Without this adjustment, it would be an AUD 10k over-statement.

Here is her restated tax return:

The submitted tax return resulted in an expected tax refund of $1,700. Now, I estimate we should get around twice as much back. I need to wait for the submitted return to be processed before filing an amendment.

Wednesday, October 14, 2026

L1 Global Long-Short Entitlement Offer

The L1 Global Long-Short Fund (GLS.AX) announced a 1 for 2 entitlement offer (known as a rights issue in the UK). You can buy one new share for every two you have for $1.76 a share. The current share price is $1.84, but they have traded as high as $2.17 in the recent past. You might think that issuing that many new shares will crash the share price. My recent experience is that isn't the case for funds like this, so it is likely to be possible to sell for a profit. 

I have 170,000 shares and am taking up the full 85,000 new shares. I am planning to hold onto 30,000 of them and to sell the rest. I am buying from my Commonwealth Bank offset account rather than trying to buy on my margin loan, which is a hassle. As the issuer has my CHESS HIN (Australian Stock Exchange ID) the shares will appear in my margin account. The highest you can raise your BPAY (an Australian payment system) limit to online is $100k. Until recently, $20k was the maximum you could raise it to yourself. So, I had to phone the bank and ask to raise it for the $149,600 payment. They then subject you to a bunch of questions about scams... They actually had to raise it to $2 million, as that is the next limit. I have now sent the money to L1 and then lowered the limit back to $20k. You don't need to submit a form, just pay via BPAY with the correct individual reference number. It's very convenient.

Tuesday, October 13, 2026

Ruffer Investment Company Update

In February 2024, I got out of Ruffer Investment Company. It was never meant to be an exciting investment, but even by that standard it was disappointing. We held for about 3 years with an IRR of -3.9% and a loss of £4,523. It was one of the investments I dropped in response to our disappointing 2023 performance.

What has it done since then?

It has gained a bit, but vastly underperformed the FTSE Index. It has gained 5.4% p.a. over the last three years and 3.7% over the last year. So, this was a good decision.

Monday, October 12, 2026

Out of Hearts and Minds

I am finally totally out of Hearts and Minds (HM1.AX). It was a great concept but a disappointing result. The concept was to invest in the "best ideas" of a small number of fund managers plus each year fill 35% of the portfolio with the pitches of invited managers at the Sohn Hearts and Minds Conference. And they donated to medical research instead of collecting management fees.

Based on the idea that fund managers fill a lot of their portfolio with second rate ideas in order to diversify and reduce concentration risk, picking their best ideas should work out well, no? There shouldn't be too much concentration risk because there are multiple managers and 30-40 stocks in the portfolio. 

My realized internal rate of return was 7.3% p.a. for a total profit of AUD 63k. I first invested eight years ago at the IPO. Here is my profit graph:


I did trade in and out of it a bit. Did that affect my returns? They claim a 10.3% p.a. rate of return since inception, so the answer is yes. But 10.3% underperformed the MSCI World Index, which has returned 12.5% p.a. over the same period.

What will I do with the money? I could reduce my USD margin loan in the account or buy more crypto-related investments. 

Tuesday, October 06, 2026

Using AI to Debug My Monthly Accounts

My accounting system produces all the nice numbers, tables, and graphs you see on this blog as well as our tax returns etc. But it is a complicated web of interlinked Excel spreadsheets. Each investment has its own workbook as does each set of monthly accounts and then there are the main sheets that consolidate all those numbers. It gradually grew, though I have done some major reforms to make things work better now and then. But there are always errors in the monthly accounts, which I mostly can track down quite easily and fix, but some are very hard to find.

So, I just started a free one month trial of Microsoft Office to see whether using Copilot inside Excel can find the elusive bugs. My former employer where I am now an emeritus faculty member doesn't support this function in Office.

The first goal was to fix discrepancies of around AUD 4k in the September accounts. I had to open around 35 files simultaneously so Copilot could see everything relevant that was linked. And then a dialogue ensued that took an hour or two. It quickly began to understand quite well what was going on in the system. It then got convinced it was an error in my transfers to Angellist of USD 3k. We went through step by step why that wasn't the case. I had to give it a lot of explanations of why it was wrong. Then it found the first genuine error–a transfer of AUD 10k to USD 7k-ish was missing from the sheet that calculates intramonth forex gains or losses (i.e. due to conversions between currencies during the month). But that didn't explain many dollars.

But once that was out of the way it only took us another 11 minutes to find the main error. I had written down a transfer of almost USD 3k out of the SMSF's Interactive Brokers account that went nowhere. Originally, this was a conversion from USD to CAD to buy Defi Technologies shares in Canada. But as there were a lot of transactions, I later consolidated all of them into one line. But I forgot to delete that number!

There is still an error of around $5. It's not worth tracking that down, though it might be fun 😀. 

Maybe this is worth paying for (AUD 150 per year). Clearly though this kind of tool only can work as a "co-pilot" with an expert on the system under investigation. 

Also, I wonder why I still calculate the monthly reconciliation for non-retirement accounts alone, which was one point of confusion for the AI. One of the things it thought might be an error was my pension from Unisuper, which appears to come out of nowhere. It might be simpler to just do it for all accounts.

2025-26 Tax Returns

Here is a summary of my 2025-25 Australian tax return (all numbers in Australian dollars):

You can find previous years here. Salary was up massively due to the redundancy package. I only actually worked for 5 months of the year. Also, $100k of the redundancy package was totally tax free and is not included in my tax return. Also, I started to realise more capital gains–up 129%. Dividends increased more modestly and total income was a record $370k, up 41%. 

Deductions also increased this year, mainly because of the personal superannuation deduction of $13k. In theory you can contribute up to $32.5k per year in concessional superannuation contributions. But I had already used more than half the cap in employer superannuation contributions. I also loaded up on employment deductions before quitting :). Deductions totalled $43k, up 52%.

So, in summary taxable income was $327k, up 39%. My estimate of tax payable depends on me understanding the correct interaction of the tax on the redundancy package and the tax on my regular income. I could be wildly out on the $95k. But I have $14k in franking credits from Australian dividends and a disappointingly small venture capital offset to reduce that tax down to an estimated $80k. I estimate I will owe $2.5k in tax :(. 

Div 293 tax is an extra 15% tax on superannuation contributions if you earn more than $250k. I doubt I will ever need to pay that again.

Here is a summary of Moominmama's tax return:

She had big jumps in Australian dividends and foreign source income. The latter is mainly gains on Australian Dollar futures.

As a result gross income came in at $116k, up 6%. Deductions were down by 26% due to keeping more money in our offset account, which reduces mortgage interest. Yes, it is tax-deductible if you do debt recycling.

As a result, taxable income was up 39% to $75k. Gross tax of $15k is offset by $10k in tax credits to result in net tax of $5k. She should get a refund of c. $1,700. 

Separate to all this is the tax return for our self-managed super fund. I still need a tax statement from Assetora, but the deadline for SMSF's to submit tax returns is much later. 

Saturday, October 03, 2026

30 Years of Net Worth Data, Half a Decade at a Time

I started tracking expenditure, income, and net worth in Australian Dollar terms when I moved to Australia for the first time in October 1996. My accounts from before that date are in Pounds Sterling or US Dollars and I wasn't investing. So, the base net worth for rate of return calculations was September 1996. We now have 30 years of data. I thought it would be interesting to track progress five years at a time. I think I got the idea from some Berkshire Hathaway report that tracked progress in a similar way. 

All numbers are in Australian Dollars. Investment profit is post implicit superannuation taxes but the accumulation index is pre-tax. Net worth growth rate and annual rate of return are compound growth rates over the five years previous to the date on the left.

Overall, net worth increased by about 800 times and investment achieved a 6x multiple. But each half-decade was unique. September 2001 was before the worst of the tech wreck. Profit and net worth were already down from their peaks but I still had achieved an average 7.8% annual rate of return and a 67% net worth growth rate and had saved over $130k, which is pretty impressive on a salary of about $50-60k per year.

In September 2006, I was back in the US and markets were a year from their pre-GFC peak. Returns were weaker at 5.4% per year because of the losses in 2001-2003. I had added another $160k to saving. My salary was higher in the US, but I dissaved a lot in 2001-2002 when I was unemployed and travelling to find a new academic job. Net worth grew an average of 23% per year.

In September 2011, I was now married and back in Australia. I had recently got a continuing position as full professor. However, the previous five years included the GFC. So the average rate of return was negative at -7.9%. Moominmama was a post-doc fellow and I only had a job for part of the time. But we still managed to increase net worth by 2% per year by saving $223k! 

The next decade saw net worth grow by 28% per year and investment returns were above 10% per annum. In September 2016, we had reached millionaire status in both Australian and US Dollar terms by saving more than $600k and adding $1/2 million in investment returns. This was the best 5-year period for investing with an average return of 12.5% per annum.

By September 2021, we were in the post-COVID stock market rally. Net worth was over $5 million after saving another $400k, receiving the main inheritance, of $1.8 million, and adding close to $2 million in investment returns.

Finally this month, net worth is over $8.5 million. We can no longer grow net worth as rapidly. We continued to save after the inheritance–including the boost from the redundancy package and despite the beginning of retirement dissaving this year. We added another two and a half million of investment returns. Returns were lower at 8.6% per annum. They are still 1% per annum ahead of our target portfolio and 3% p.a. ahead of a Vanguard Australian Dollar 60/40 passive benchmark. It might come as a surprise, but major stock indices also grew more slowly in the last five years than in the previous five.

Friday, October 02, 2026

September 2026 Report

After five up months, September was a down month. The Australian Dollar fell from USD 0.7164 to USD 0.6983. So, our Australian Dollar returns are higher than our USD returns. 

Here is the performance of our benchmarks (total returns including dividends):

US Dollar Indices

MSCI World Index (gross): -1.08%

S&P 500: -0.35%

HFRI Hedge Fund Index: 0.08% (forecast)

Australian Dollar Benchmarks

ASX 200: -2.10%

Target Portfolio: 0.40% (forecast)

Australian 60/40 benchmark: -0.90%

In Australian Dollar terms we lost 1.52% and in US Dollar terms 4.01%. As a result we only outperformed the ASX 200 benchmark. The SMSF returned -2.64%, underperforming both Unisuper and PSS(AP) which lost 0.85% and 0.46%, respectively. Here is a report on the performance of investments by asset class:

The asset class returns are in currency neutral and gross asset terms and do not include investment expenses such as margin interest, and so the total differs from the Australian Dollar returns on net assets mentioned above. All asset classes apart from futures and commodities lost money. Within futures and commodities, cryptocurrency and managed futures made positive contributions and Australian Dollar futures detracted. Hedge funds had both the worst return and detracted the most. However, the Acadian Global Long-Short Fund was one of the best performers this month. There is a lot of variability within these asset classes.

Things that worked well this month:

  • We had 5 investments that gained AUD 10k or more : Silvia (SVIA, 21k), Bitcoin (18k), Acadian Global Long-Short (13k), L1 Gold Fund (LGF.AX, 10k), and ZIM (10k).

What really didn't work:

  • We had 9 investments that lost AUD 10k or more: Gold (34k), Pershing Square Holdings (PSH.L, 27k), Tribeca Global Resources (TGF.AX, 26k), Regal Investment Fund (RF1.AX, 16k), 3i (III.L, 15k), Australian Dollar Futures (15k), L1 Global Long-Short Fund (GLS.AX, 14k), Pengana Private Equity (PE1.AX, 11k), and URF (10k).

Our distance from our new target allocation increased slightly. Our actual allocation currently looks like this:


Around 70% of our portfolio is in what are often considered to be alternative assets: real estate, art, hedge funds, private equity and credit, gold, and futures. A lot of these are listed investments or investments with daily liquidity, so our portfolio is not as illiquid as you might think.

Moominmama receives employer superannuation contributions every two weeks. There will still be capital calls from Aura Venture Fund II and III. I am receiving monthly pension payments from both Unisuper and our SMSF, totalling AUD 6,770 per month. I made the following moves:

  • I invested a total of USD 3,000 in two startups on Angellist.  
  • We continued to add to our "crypto basket" including 3,500 shares of IBTC.AX (= 0.35 bitcoin), around AUD 48k of QETH.AX (ethereum ETF), 3k shares of SVIA (fomerly BRR), and establishing a 50k share new position in Defi Technologies (DEFI.NE/DEFT), which we also invested in in the previous cycle. We now have 5.4% of net worth in this basket.
  • We added another 1,000 ZIM shares. 
  • We sold the following shares to fund this build up: 369 AMLP, 1,000 PBDC (whole position), 2,000 WCMQ.AX, 7,000 HM1.AX., and around 15k CDO.AX shares.
  • I also sold 50k L1 Gold Fund (LGF.AX) shares after receiving 30k in the rights issue. This brought our CommSec margin loan down to AUD 43k. Given its high interest rate, I plan to use it for trading rather than long-term investment going forward. This is part of the "restructure".
  • Continued to trade options on TWE.AX and did a couple of other stupid short-term trades that luckily didn't lose much money.

Here are the income and spending accounts * for this month ($ is Australian Dollar):

Other income includes Moominmama's salary and net employer superannuation contributions but also the tax paid by the SMSF. Spending was typical of non-school-fee months at AUD 8.2k. This number does not include our mortgage payments, which are regarded here as saving (principal) and investment costs (interest) but it does include depreciation on our car, which is a few hundred dollars a month now. We lost $111k investing. Retirement net contributions (Other income retirement) were $516. The transfer out of retirement is my pensions. Tax credits on non-superannuation accounts were $4k. As we report superannuation investment returns on an estimated pre-tax basis, the implicit tax ($386) needs to be deducted to get to the change in net worth.

Dissaving amounted to $4.3k, which is well within the 4% rule limit of AUD 23k.  

As a result of all this, net worth fell by AUD 119k to AUD 8.567 million.

* Results are shown separately for retirement and non-retirement accounts as well as housing, which nowadays doesn't have much activity. The grey shaded rows are additional notes. Total investment income is split into investment income before exchange rate moves and the contribution of exchange rates. Current other income is non-investment income including salaries and net income tax returns, while on the retirement side it includes employer superannuation contributions minus superannuation contribution tax and all SMSF tax payments to the ATO. Investment income is shown pre-tax. Tax credits include franking credits on Australian dividends etc. in non-retirement accounts and the SMSF and imputed tax on industry superannuation returns. These are taken away from investment income to get changes in actual net worth. Inheritances include gifts from relatives. Saving is from non-investment income, transfers, and inheritances not investment income.