You need two things: First, you need to have a judgment about what will happen in the future; second, you also need to assess how likely your judgment is to be correct.
No asset is so good that it can be bought regardless of price; conversely, very few assets are so bad that they remain unattractive even when priced extremely low.
On March 17, at the Perella Weinberg Capital Markets Conference held in New York, Howard Marks, co-founder of Oaktree Capital, was interviewed by Bloomberg host Lisa Abramowicz.
In the current context, the timing of this dialogue is quite delicate. On one hand, artificial intelligence is transforming the business world and reshaping the boundaries of capital markets' imagination at an unprecedented pace; on the other hand, private credit, after years of fervor, has started to become a focal point for growing concerns over risk.
Of course, during the brief 20-plus-minute conversation, broader issues such as geopolitics and monetary policy were not explored in depth. Even so, Howard Marks candidly remarked that this might be the most challenging phase he has encountered in his lifetime.
Rather than settling on generic judgments like 'AI is important,' this discussion brought together two of the most closely watched topics today: the significant uncertainty brought by artificial intelligence and the rising anxiety in the private credit market following a prolonged period of prosperity.
Marks’ core argument is clear: lending money to businesses is not inherently wrong. The issue has never been the asset class itself but rather what happens when too many people rush in, driving down interest rates, eroding safety margins, and eventually exposing risks.
In the past, the market focused on whether private credit represented an opportunity; now, the question being asked is whether private credit will encounter problems. This shift reflects a classic turning point in the cycle.
Marks is neither a technological pessimist nor a rash technological optimist. He rationally reminds us that if you are betting on artificial intelligence, it may not be appropriate to buy bonds; instead, you should consider buying stocks. When facing fundamental changes at the business model level, fixed income may not adequately compensate for such uncertainties.
Moreover, in his view, even though AI can already help organize information and generate insights, it still cannot replace the most critical elements in investing: intuition, prudence, judgment of people, and the sense of timing needed to act decisively in extreme moments.
The core ideas presented in the dialogue are quite simple, focusing on how these basic principles influence our judgment across different scenarios: respect for cycles, respect for prices, respect for risks, and also respect for those elements of human judgment that machines cannot yet replace.
These insights allow us to remain composed even amidst chaos.
AI has made decision-making in investment harder than ever before.
Host: Many people are currently concerned about private credit. Do you think this concern is justified?
Marks: In my opinion, lending money to businesses is fundamentally a solid and legitimate activity.
I have been involved in lending to sub-investment grade enterprises for 48 years. In 1978, Citi invited me to launch their high-yield bond business, and fortunately, it was just when this market was beginning to take shape. Throughout this journey, our clients' overall performance has consistently been strong.
I believe that approximately 99% of the high-yield bonds we purchased were ultimately repaid.
So fundamentally, there is nothing wrong with this practice. The issue arises when too many people want to engage in this activity, becoming overly aggressive in pursuing deals, which drives down interest rates and erodes safety margins.
Subsequently, common problems in the investment industry emerge: these products end up being sold to individuals who should not buy them—those who are unprepared and do not fully understand them—and once issues arise, they are caught off guard.
Host: Do you think the emergence of artificial intelligence, along with its replacement of certain corporate functions, a form of 'disintermediation,' and even the potential obsolescence of some software, will alter your investment approach?
Max, I believe that the changes taking place now, especially the introduction of artificial intelligence, are making the world more unpredictable than ever before.
It is arguably the most difficult phase to assess that I have encountered in my lifetime.
In the investment industry, most people make decisions based on their predictions about what will happen in the future. However, I have always believed that this approach is insufficient.
You need two things. In a book I wrote in 2018 about cycles, I mentioned that first, you need to form an opinion about what will happen in the future; second, you also need to evaluate how likely your prediction is to be correct.
Not all predictions have the same probability of success.
And I think this era ushered in by artificial intelligence is one of the most unpredictable scenarios I can imagine. We truly do not know what will happen next.
Incidentally, precisely because its power is so immense, it becomes so unpredictable.
What can it do, and what can it not do? To what extent will it replace us? And to what degree will it cause unemployment? If a large portion of the population loses their jobs, what will society look like?
These are questions we have never truly faced before.
Host: This is why many people today are concerned that certain types of investments are becoming increasingly concentrated in the hands of private equity fund managers. Coupled with a lack of transparency, it naturally raises concerns.
One might wonder, how exactly are these loans created? Whether it’s leverage or contractual terms, is there any indication that we may be nearing the end of this cycle? The market is starting to heat up, and people are saying, “Maybe it’s time for a correction.” Is this assessment reasonable?
Marks Over the past few months, especially in the last six weeks, the question I’ve been asked most often is, “What do you think about private credit?”
A year ago, when people asked me about private credit, the tone was completely different. Back then, they were asking: Are we allocating enough? Should we allocate more? Is it the perfect thing that can make money without much risk?
I’ve been in the market long enough to have witnessed many similar cycles—not economic booms and busts, but rather the rise and fall of certain 'ideas.'
It’s always the same. First, something new appears. It usually has to be new because novelty is what ignites people’s imaginations the most and is easiest to sell. And precisely because it’s new, people haven’t had the chance to truly see how its flaws might emerge.
That’s why this kind of thing keeps happening. I don’t think mundane things can form bubbles.
Steel won’t have bubbles, and hamburgers won’t either. What truly tends to form bubbles are new technologies or financial innovations.
People buy into it because of the promising prospects it paints, but they don’t fully understand the downside risks.
When something happens to dash their expectations, they realize that they didn’t fully grasp what they were doing, nor did they truly understand its limitations. Naturally, disappointment follows.
The relative advantage in the private credit space is now gone.
Moderator: You once mentioned a viewpoint that the excessive enthusiasm in the private credit market might actually make the public market appear more valuable and attractive.
What about now? Do you think there is enough fear in the market? Is it attractive at this point, or does it still need more fear? Are people still relatively greedy? Is there still some overheating in the system that needs to gradually dissipate?
Marks: I really don’t know. I’m not sure about the pricing of private credit today.
In 2011, after the global financial crisis, banks were hit hard and faced stricter regulations, causing them to retreat from leveraged buyout financing. Subsequently, so-called non-bank lenders stepped in, starting to provide direct lending, specifically for mid-sized leveraged buyouts.
At the time, the private equity industry had a strong demand for capital, but the available capital from non-bank institutions was limited, so these institutions could naturally demand higher interest rates and secure better safeguards.
However, one of the most classic sayings in our industry is: What smart people initially do will eventually become something everyone rushes into foolishly.
Once profits are made, others see it and imitate, leading to an increasing number of new participants, and the original “uniqueness” gradually disappears.
So I think sufficient occurrences of this have already happened in private credit. Interest rates have been compressed, as has the level of security.
Perhaps just a few months ago, maybe two months ago, I would have said that the yield on public market credit was around 7%, while direct lending was approximately 8.25%.
In my view, a liquidity premium of 125 basis points is fair and generally sufficient, but by no means generous.
In my view, its original 'distinctiveness' has disappeared. Private credit and public markets have largely reached an equilibrium, where transactions are fair but no longer offer significant bargains.
So when people ask me, 'What do you think about private credit?' my response is: 'What do you think about credit as a whole?'
Of course, there is also a saying in our industry known as 'talking your own book,' meaning people tend to advocate for the business they are involved in. Oaktree Capital has been in the public market credit business for 48 years, so naturally, we hope the market pays more attention to this area.
But what I mean is, when the pricing of A and B is roughly in balance, instead of putting all your money on one, it may be better to allocate some to both.
Host: Another concern about private credit now is that it is increasingly extending itself to retail investors, hoping to include those who desire a certain degree of regular liquidity. Do you think that under these circumstances, the transparency of assets, valuation methods, and price marking need to be more standardized for this to truly make sense?
Marks: Clearly, I would say the answer is yes.
First, some institutions, including ours, cannot claim to be entirely uninvolved. We believe selling private credit products, such as BDCs, to retail or pension investors can be a viable business. So, everyone proceeds accordingly.
But I will take the risk of saying this: not everyone buying these products fully understands their nature, especially the fact that these assets inherently lack liquidity and there is no real-time market available for mark-to-market valuations.
They still buy them. I think it’s more due to a lack of understanding rather than making a fully informed decision and deciding it's not an issue.
I was also discussing this with another group of investors yesterday.
At the time, I said it reminded me of 'Casablanca.' Everyone has seen 'Casablanca,' right?
There is a very classic scene where the police chief walks into a casino bar that he had been taking bribes from for the past twenty years and pretends to be shocked, saying, “What? Gambling is happening here? I’m shocked, absolutely shocked.”
What I mean is, these things that investors are only now discovering and becoming uneasy about are not new at all.
These products have always been illiquid and have never been marked-to-market daily, among other things. But you know, when everyone is excited and eager to invest, these issues tend to be overlooked; once problems start to arise, they suddenly become a source of anxiety.
One reason private credit has come under scrutiny is that unseen things tend to make people uneasy.
Credit has experienced a 17-year low-default environment.
Host: Do you think the stress we're seeing in some private credit funds now, and even the discussions around them, reflect a broader sense of tension? Could this pressure continue to spread and eventually turn into a larger wave of defaults or more serious problems across the entire credit market?
Marks: Credit is inherently cyclical.
There are times when everything goes smoothly, and borrowing money is easy; there are other times when conditions deteriorate, and borrowing becomes difficult.
When financing is easy, everyone competes for deals, as I just mentioned, and interest rates naturally get pushed down.
When times get tough, lenders become less enthusiastic, and those few who are still willing to lend can demand higher interest rates.
This is the basic way in which the credit cycle operates.
Last autumn, when issues surrounding First Brands (an American automotive parts company) and Tricolor (a company related to car dealerships and subprime auto loans) came to light, their bankruptcies, along with likely fraud, shocked everyone.
Jamie Dimon once made a remark to the effect that you usually don’t see just one cockroach. I later wrote a memo titled 'Cockroaches in the Coal Mine,' combining several metaphors.
In that memo, I noted that as long as the music keeps playing, capital remains loose, prompting lenders to continually lower standards in order to secure deals. As a result, borrowers who should not have qualified for loans receive funding, and fraudsters find it easier to perpetuate their schemes.
There’s an old saying in banking: the worst loans are often made during the best of times. This is the reasoning behind it.
And we’ve now experienced a full 17 years of good times.
The stock market bottomed out in March 2009, probably on March 6, and it has been exactly 17 years as of this month.
Since then, financial markets have not truly experienced a particularly difficult phase. 2015 was not great, there were three very rough weeks during the pandemic in 2020, and 2022 was not a good year either.
However, if we calculate from around September 30, 2022, I estimate that the S&P 500 has roughly doubled. Enterprise value has not doubled, intrinsic value has not doubled, but prices have doubled.
Therefore, these have always been good times. Good times make people more eager to invest their money, but they also weaken analysis, discipline, high standards, and skepticism, replacing them with FOMO, or the fear of missing out.
When the market is driven by FOMO rather than skepticism, bad deals are bound to happen.
This is actually nothing new.
We have experienced 17 years of a low-default environment.
The Fed's operations at that time caused the global financial crisis, one of the most destructive environments I have experienced in my lifetime, to result in only one year of high defaults in the high-yield bond sector, whereas it would normally last for two years.
That was supposed to be the most severe crisis, but the number of defaults was actually lower than in many other crises I have experienced.
Later, in 2020, there was once speculation about whether the default rate would surge to 15%, but in the end, I remember it settled at around 5.5%.
Thus, people have become accustomed to the absence of defaults. But defaults are a normal phenomenon, especially after a large amount of credit has been extended, followed by a phase where problems gradually surface.
Buffett always has the best way of putting things. His most classic quote is: 'Only when the tide goes out do you discover who's been swimming naked.'
It is only during tough economic times that we see who lent too casually and to whom. And I think such a phase is still ahead of us.
Host: From the broader perspective of the corporate bond market, does the current pricing reflect, to some extent, the risks of a higher default cycle?
Marks: As I just mentioned, I do not know the specific pricing at this moment.
But if you ask me about the general trend, I would say no.
The most frequently asked question recently is why the yield spreads for sub-investment grade credit remain at the lower end of the normal range.
If the spreads are still at the lower end of the normal range, then the only conclusion is that the market has not fully priced in the concerns of higher defaults and is not offering additional compensation for increased default risks.
If betting on AI, one should buy equities rather than bonds.
Host: We have been discussing how artificial intelligence and technology will change many previously assumed premises. As you mentioned, this may be the hardest environment to predict that you have ever experienced.
From your perspective, does it make sense for companies with large scale and strong capabilities, such as Google, Microsoft, and Amazon, to issue 30-year, 40-year, or even 100-year bonds at a time when the outlook for the next five years remains unclear?
Marks: I believe this is exactly the key point I wanted to emphasize earlier.
To take your example, I remember Google issued a 100-year bond with a coupon rate of approximately 5.8%.
If you think carefully about this matter, I believe you can only say that what prevails now is not pessimism, but optimism; not skepticism, but credulity.
And when optimism and credulity prevail, it becomes very difficult to make investments that truly deliver excess returns.
Excess returns refer to the portion of returns above what the risk warrants.
Host: So, if a company is highly relevant to these new technologies, and the technology indeed has tremendous prospects, how would you invest? In such an uncertain situation, how do you identify the companies that will truly benefit?
Marks: First of all, it's not just artificial intelligence companies that are affected, or even just tech companies—many industries will be drawn in.
As for the question I mentioned in my memo on December 9 last year: should one lend money to those artificial intelligence companies, or to tech companies for investing in artificial intelligence?
At that time, I quoted a statement from my colleague Bob O'Leary.
His point was roughly this: if you really want to invest in a company betting on artificial intelligence, you should probably buy its stock rather than lend it money. Because if it ultimately succeeds, you should gain from the upside, not just receive a fixed return.
I think that’s correct.
You made a very apt point earlier. You’re lending money to a company for 30, 40, or even 100 years, yet we can’t even foresee what will happen in the next five years.
Since you are taking on the fundamental risks at the business model level of the enterprise, shouldn't you be compensated as an owner rather than as a fixed-income investor who only receives a minimal guaranteed return?
Host: Of course, artificial intelligence will affect almost every company.
Marks: Indeed, it will.
Precisely because its impact is so widespread and unpredictable, the world has become more challenging.
To be honest, I still prefer the world as it used to be. Do you feel the same way?
Host: You mean the era of your childhood when you walked to school barefoot and backwards?
Marks: Exactly. Back then, year after year, almost nothing changed.
Now, things change almost every day.
Of course, this constant change brings a certain vitality. In many ways, life today is much better than it was during my childhood.
But it also means that we can rely less and less on our expectations.
AI will not bring us investment intuition.
Host: How is artificial intelligence changing Oaktree Capital's daily operations now? Will it affect your staffing levels or how you allocate your teams? Has it already become a tool that everyone uses?
Marks: It has not yet changed our business model, nor has it altered how many people we hire or how we operate. For now, it remains more of an auxiliary tool, helping us organize data and mobilize information.
When I write these things, I usually ask my son for help. He is an excellent venture capitalist. After I finish writing, I ask him, 'Do you want to take a look?' He replies, 'No need, just send it to Claude.'
The problem is, I wrote this on my desktop computer. To send it to Claude, I have to copy it out first, paste it into an email, and then send the email to myself so I can open it on my phone. Then, on my phone, I have to copy it again and paste it into the AI to send it to Claude.
The entire process took about two minutes. But as soon as I hit send and scrolled down to the bottom, the answer was already there.
It was truly astonishing. I was completely taken aback at that moment.
So, AI can indeed do a lot of such tasks. It has read almost everything ever written, remembers what it has read, and can retrieve it again, which is quite different from us.
It can also identify which past patterns led to success and which led to failure, and then extend those patterns into the future.
Moreover, it is highly unlikely to make arithmetic errors or logical mistakes. It also won't make emotional errors. It won’t get overly excited at the peak or excessively discouraged at the trough.
So, this is certainly a significant improvement.
As for whether it performs better than 80% of people or 90% of people, I don't know. I just hope that it hasn't become strong enough to surpass 100% of people.
I hope there are still some things humans can do, and I believe there indeed are.
I don't think it has intuition. You know, sometimes when you read a prospectus, you get a chilling feeling at the back of your neck. AI doesn't have hairs on the back of its neck that stand up.
We have always believed that over the years we have saved our clients a lot of money, and one important reason is that we haven't dealt with unreliable people.
I think when it comes to identifying 'bad guys,' we are probably still better than AI at this. There are many similar tasks like this.
Therefore, I think AI understands history well and excels at recognizing patterns and conducting extrapolations. Ultimately, I increasingly feel that what AI does is prediction. It's not 'answering questions'; it's 'making predictions.'
Take the simplest example. When you pull out your phone to write an email, the bottom of the screen will automatically suggest the next word. Essentially, it's predicting what the next word in your sentence should be.
For instance, if you write: 'I hope you'll go with me to...' it might automatically suggest 'party.' Of course, it could also be 'ballgame,' perhaps listed a bit lower down.
The reason it can provide that word is because it has already processed millions of sentences. In phrases like 'I hope you'll go with me to the...', there may be a 73.7% probability that the next word is 'party.'
So it can indeed achieve that.
What it provides us are predictions, which I interpret as hypotheses. However, I wouldn't invest clients' money solely based on its recommendations.
I believe it is still necessary for humans to review these hypotheses. Of course, this step reintroduces the possibility of human error. Nevertheless, I still consider it indispensable.
We are not ready to hand over the entire process to AI, but we are pleased to have its assistance.
It is far from the point where cautious investors should take action.
Host: In such a highly unpredictable era, do you think it becomes more important to maintain higher liquidity or hold more assets that can be quickly liquidated and readily reallocated?
Marks: You know, when I joined Citi in September 1969, Citi and most so-called 'money center banks' were buying the 'Nifty Fifty'.
These were the 50 best and fastest-growing companies in the U.S. at the time. People believed these companies were infallible, and no price was too high. If I recall correctly, I joined Citi on September 22, 1969. If you had bought those stocks on that day and held them tightly for five years, even if they were shares of the greatest American companies, you would have lost about 95% of your money.
Later, I was 'exiled' to the bond department, which felt like Siberia, and was asked to manage a high-yield bond fund.
Looking back, that was probably one of the luckiest things that ever happened to me. What was the result? I began to earn stable and secure returns from some of the worst-performing listed companies in the U.S.
This has made me realize that no asset is so good that it can be bought regardless of price; conversely, very few assets are so bad that they remain unattractive even when priced extremely low.
Therefore, despite our many concerns about AI and despite our lack of complete confidence in our own judgment, I still believe that certain assets will eventually fall to a level worth buying, even in such an environment.
As for how we will feel when that time comes, it is hard to say now because that moment has not yet arrived.
At present, we are far from the point where everyone is selling indiscriminately, as seen during previous major market crashes where people were 'throwing out everything.'
But I believe that such a moment will eventually come. I also want to believe that when it does, we will step up, just as we did in September and October 2008, when we took bold actions despite the majority believing the financial system was on the verge of collapse.
Host: Do you think we are heading toward another moment like that?
Marks: I have a very profound answer to this question: who knows?
Benjamin Graham famously said: In the short run, the market is a voting machine; in the long run, it is a weighing machine.
I don't know how everyone will vote next month. If they vote for a 'crash,' then it will crash. If they regain calm and decide 'it’s fine, we’re okay,' then it won’t crash.
It’s almost as simple as that.
Therefore, I believe that this matter itself is unpredictable. However, I am not in a rush to invest all my money just because 'a market crash may not happen.'
I will wait and see what happens next. If it really comes, I think we will become very aggressive.
But as for when it will come, I do not know. Even if it has already arrived, you may not be able to confirm it immediately. You will never know for sure whether prices have dropped to a level where they are worth buying.
You can only rely on a feeling.
By the way, I am also not sure whether AI has this kind of feeling.
I do not think that asking Claude, 'Please tell me, is it time to buy now?' will result in a particularly brilliant answer.
So, when you ask about the role of AI in the investment process, I think this is precisely a good example to show that there are still some things it cannot do.
What most people underestimate is still the impact of AI.
Host: If you were to describe your current attitude towards the market, would you say it is cautious?
Marks: I have always been cautious, unless I reach a point where I am sure it is time to act.
Because at my core, I am a cautious person, and this caution only loosens when reality gradually disproves it piece by piece.
Moreover, fundamentally speaking, lenders ought to be predominantly cautious. A lender who is not cautious enough or overly aggressive usually does not fare well.
This is because lending only has downside risk and no upside potential. What is referred to as 'upside' merely means that the contract was fulfilled and promises were kept.
Therefore, I believe that until we are confident that market disappointment and price declines have reached a sufficient depth, and it is time to be aggressive, we will remain cautious.
When that time truly comes, I hope that, as in the past few cycles, we will become one of the most aggressive and decisive groups in the world.
By the way, perhaps by then, there will be very few people left on Earth who still dare to be aggressive.
Host: We have about a minute left. I’m very curious—what do you think is the most underestimated thing today?
Marks: I believe what most people underestimate is the impact of AI.
About 18 days ago, on a Friday, a company called Block announced that out of 10,000 employees, 4,000 were laid off on the same day. Within a single day, 40% of the workforce was cut because AI could perform these jobs at a lower cost and faster speed.
Think about what this truly means. How many people in this world genuinely understand the weight behind it?
Looking to pick stocks or analyze them? Want to know the opportunities and risks in your portfolio? For all your investment-related questions,just ask Futubull AI!
Editor/joryn