 How an Actuary Saved Elon Musk Hi everyone, it's MJ the Fellow Actuary and I received the story from the South African Actuarial Society during lockdown and was told that it would make for a great video. So let's jump into it. It's the year 2000 and while 9 year old me was starting my Pokemon collection, another group of South Africans were about to change the world. PayPal allowed for financial transactions to be carried out over the internet. Elon Musk was the CEO and rule of Bertha was the CFO. Now Musk needs no introduction and Bertha studied actuarial science at the University of Cape Town. Now have you ever wondered why PayPal became so dominant? Well it's because fraud wiped out most of the competition. Why didn't Wipeout PayPal? Well because they had an Actuary. Now what is an Actuary? An Actuary is someone who confronts the uncertainties of the future and financial institutions need them because of how severe the consequences of uncertainty are in investments, lending and insurance. Now these financial transactions have uncertainty because they are exposed to time and Actuaries love time. The first thing we learn in Actuarial Science is the time value of money which then leads to theory of interest and so on. So any tech company that wants to get involved in finance is wise to employ a couple of Actuaries. But what does an Actuary actually do? Well step one is to go around the risk control cycle. And risk is the combination of danger and opportunity. So think of a dragon guarding a pile of gold. The dragon is the danger. The gold is the opportunity and to gather that risk. So if we were to go through the risk control cycle for PayPal, the first thing you want to do is awareness. Make everyone in the organization aware that risks exist. Is there any dangers and opportunities? If you're just focusing on the dangers, you're not going to want to move forward and do business. But if you're just focusing on the opportunities, you're not going to be cautious and you're exposing yourself to ruin. The next thing we need to do is identify risks. Is there any opportunity for facilitating online transactions? And if the answer is yes, well are there any dangers there? Maybe we're going to be exposing ourselves to fraud. Once we've done that, we can then measure the risk. How much can we make from facilitating online transactions? And how much do we stand to lose from various other risks such as fraud? We then need to manage these risks and every management action has a cost. If the danger is severe, then the benefit of managing it outweighs the cost. If the danger isn't that severe, then maybe it's not worth implementing an expensive risk management technique. So this step depends heavily on the previous one. And we complete the risk control cycle by monitoring the risks. Basically we're comparing the actual results with that which is expected. Did we make as much money as we thought we would from operations? And if not, why not? Or did we lose more money than we thought we would from each of the dangers? And if yes, why? You then need to investigate. Now, actuaries run through the risk control cycle so that an organization's risk profile can be determined. This can be used to project the future position of an organization and help capital providers decide whether or not they should invest and how much money the organization should hold onto as a reserve in case the dangers manifest. Now, if we look at PayPal, we see that they were aware of risk. They had successfully identified fraud as a danger. However, their fraud risk model said that it wasn't that big of a deal. And so because the risk measure was said it was small, they didn't spend that much energy on managing this risk. Fortunately, the actuary picked up the issue during the monitoring stage. Because what we see is that the actual losses from fraud were a lot higher than the expected losses predicted by the model. Now, Roliff said the following. However, the fraud estimates kept coming in wrong and the prediction error kept growing. After some investigation, I discovered that we were using a non-matched approach. We were comparing a given month's losses to the same month's payment volume to the estimate the loss rate. As an actuary, I spotted the similarities to property and casualty insurance. There was a delay between when the payments occurred and when we had the true fraud numbers associated with the same transactions. Our rapid growth rate coupled with this inherent latency meant that we were significantly underestimating our loss rates. We had not set up the equivalent of an incurred but not reported reserve to cover the true extent of our cost from fraudulent transactions. Now, it's a mistake to think that actuaries only work in insurance. But we do know a lot about it and Bertha was able to configure a model that was originally designed by actuaries for insurance for PayPal. And this is what he had to say. After we had this insight, we implemented the chain ladder technique to build cumulative loss distribution functions to get a more precise estimate of the problem. The findings were astonishing. Fraud rates were four times higher than we had previously estimated in growing exponentially. During the four months between July and October 2000 alone, we incurred fraud losses due to totaling $5.7 million. If we didn't act, PayPal faced imminent bankruptcy. People always want to know why do actuaries get paid such large salaries? And as we've just seen above, it's because they can help businesses save millions of dollars through efficient risk management techniques. So let's return to this risk control cycle. Monitoring had indicated that there was an issue. Bertha discovered that the issue was with the current model. A new model was run and new results were found. Now that fraud was a serious issue, more energy was needed in its risk management. So returning to Bertha, we see him saying the following. We took quick action to curb the fraud. On the technical side, we developed tools to reduce it, including one of the first commercial's captures. On the risk management side, we built out a team designed scoring algorithms for every transaction and modeled the data using logistic regression and neural network models to help prevent fraudulent transactions. Of course, PayPal continued its rocket growth. Today it is the leading global online payment company with over $100 billion in annual payment volume and with annual reserves of over $4 billion. For me, the experience was a powerful reminder of the value of actuarial training. Now this year alone, PayPal could see over a trillion dollars in payment volume. The South African Actuary didn't just save Elon Musk, he saved the internet's economy. So no wonder Musk is seeking more Actuaries to join him at Tesla. After PayPal, Bertha became a venture capitalist and has been involved in a couple of tech companies. YouTube, Instagram and Unity to name just a few. Now before I end this video with Bertha's thoughts about actuarial science, I just want to mention two things. One, if you're interested in actuarial science, check out my free Udemy course on how to become an Actuary. And two, if you're into NFTs, I'm making a treasure hunt with actuarial icons and both Elon and Bertha are pieces in this internet puzzle. Now I will be providing links to both projects in the description. So let's now end off with Bertha's thoughts about actuarial science. My experience with PayPal is just one example of how actuarial skills can be applied across a wide range of non-traditional fields. A strong sense of fiduciary responsibility and the ability to solve problems, understand mathematics and above all think long term are useful in many contexts and provide us with tremendous opportunities. To shape our world for the better. Thanks so much for watching and I'll see you soon for another video. Cheers.