Understanding Crypto 4: Prof. Tobin Hanspal: The Characteristics of Crypto Investors
Tobin Hanspal is an assistant professor of finance at Vienna University of Economics and Business (WU Vienna). His research focuses on how experiences, beliefs, and advances in financial technology affect the decision making process of households and individual investors, and has been published in leading finance and economics journals.
His current work focuses on the intersection between investment choices and consumption spending. Tobin is from Boston, has a PhD in Economics from Copenhagen Business School, and lives with his family in Upper Austria.
Welcome to another special episode of Rational Reminder Podcast, a show to help us learn about cryptocurrencies and their role in our current and future financial systems. In today's show, we speak to Tobin Hanspal, an Assistant Professor of Finance at the Vienna University of Economics and Business who has written several papers focused on household finances. Tobin's research area offers insights into the behaviours of retail investors in the crypto space and how this may affect household finances. In this episode, we take a deep dive into some of the papers that Tobin has authored and how his findings relate to the behaviours and biases of crypto adopters. We discuss the investment behaviours of early crypto adopters, the role of EFTs in reducing risk, the different types of investor groups, how past experience negatively affects investor confidence, how behaviours change after an initial crypto investment, the disposition effect, how cryptocurrencies are an extension of existing behaviours, and much more! Be sure not to miss out on this informative episode with expert, Tobin Hanspal!
Key Points From This Episode:
How Tobin investigated the investment behaviour of early crypto adopters. [0:04:24]
Whether indirect crypto investments are a good proxy for crypto investors. [0:08:10]
Why it is important to consider the different types of investor groups. [0:10:23]
The differences between individual characteristics of crypto adopters and non-adopters. [0:10:55]
Comparison of eventual crypto adopters and non-adopters portfolios. [0:12:37]
What kind of sector ETFs do crypto adopters choose to invest in. [0:13:48]
Differences between the crypto and non-crypto investors, in terms of typical investor behaviour biases. [0:15:01]
How cryptocurrencies are an extension of traditional high-risk investing. [0:16:39]
Whether the behaviour of investors changes after their first crypto investment. [0:17:37]
The differences in behaviour between early and late adopters. [0:19:15]
What insights Tobin has regarding the geographical location of crypto adopters. [0:20:36]
What percentage of their portfolios’ do adopters allocate to crypto. [0:21:11]
Find out if crypto investors buy lower-risk assets to make up for cryptocurrencies. [0:21:36]
What differences exist between crypto and non-crypto investors regarding efficiency. [0:22:51]
Description of the typical crypto investor characterized in their study. [0:23:39]
Tobin explains the disposition effect and how belief systems play a role. [0:25:56]
How risk appetite is related to the disposition effect. [0:28:05]
People’s beliefs: are expected returns affected by past experience in expected realized returns. [0:29:20]
Whether positive or negative realized past return experiences have the same effect on beliefs. [0:31:19]
How peoples’ beliefs affect investing in riskier assets. [0:32:07]
Changes in behaviour on a household level from past negative investment experiences. [0:33:23]
The role experiences of peers and/or relatives have on investment belief. [0:38:16]
Reasons for people reducing risk in their portfolios. [0:38:50]
Tobin shares if he thinks cryptocurrencies will have similar effects on peoples’ behaviours. [0:39:42]
How applicable the findings are from Tobin’s study to other parts of the world. [0:41:54]
What the ideal theoretical response is to losing money on an investment. [0:42:47]
Important takeaways that Tobin has for crypto investors. [0:43:23]
Read the Transcript:
Tobin, cryptocurrencies are anonymous or pseudo-anonymous. How did you manage to investigate the investment behavior of early crypto adopters?
Yeah, there's a number of studies and there's a lot of interest about asking what type of investors are trading cryptocurrencies and who are cryptocurrency investors. But obviously the major challenge is that it's basically impossible to look at the detailed investment characteristics of these individuals, or their actual trading behavior, because exactly as you said, most cryptocurrencies are basically anonymous or partially anonymous, at least. If you want to look at individual's portfolios and equities, you generally can partner with a bank or a brokerage and use anonymized trading histories to construct some sort of analysis about what investors are doing. With crypto, this is obviously a bit more challenging and I think researchers have to be a little bit creative. You might be interested in aggregate flows and so on, but this doesn't tell us much about detailed investor choices and behavior. We could also field surveys and ask subjects directly questions related to crypto, but of course this leaves us limited in other dimensions.
What we actually did was we worked with a brokerage in Germany and we looked at trading data and records on cryptocurrency related structured retail products. We basically scanned their investors and picked out the ones that were holding these related products. And you can refer to these in many different names or different ways. These are generally structured retail products or tracker certificates, index certificates, or various types of derivatives. We refer to these as indirect cryptocurrency investments. And there's actually a couple of benefits for using these to proxy for cryptocurrency investments. On one hand, we observe the entire portfolio of the investor and we can investigate different investment biases and if these investors are using other types of asset classes and other types of financial innovations. And also these investors are already customers of this brokerage when these products were introduced. So we have less of a selection issue where investors may be joining specific brokerage houses to trade in specific assets.
These indirect crypto products they're basically issued by banks and offer retail clients... I mean, in our setting in Europe, of course, but these are available in a number of markets. They give investors the opportunity to basically track the price path of volatility of Bitcoin and other cryptocurrencies. These are not unique to the German market. I think we have Bitcoin future ETFs in various settings. And I think in the Toronto Stock Exchange, you actually have an actual Bitcoin ETF for the last couple of months now, right?
We had this partnership with this large online brokerage in Germany, and we accessed the transaction records and monthly holdings for a sample of investors for a relatively long horizon. We've had this ongoing relationship. So this is 2003 until 2017. These indirect cryptocurrency products were basically introduced as early as 2014, but really became traded actively in the price developments towards the end of 2017. So 2016 and '17 are basically the key timeframe of our setting here. And what we can do is we can look at the investors that hold such assets and compare them to the average retail investor that's not holding them. And since we have this time series, we can try to see what happens before investors start purchasing these assets and what happens after, basically.
It's a really smart way to solve the problem. Do you think that indirect crypto investments are a good proxy for crypto investors in general, who would be maybe accessing it right on the blockchain?
This is a nuanced question, right? A proxy is exactly that. There's obviously going to be some noise when we try to extrapolate a lot from these types of investors. And I think it depends, but in general, it's not a terrible proxy. I would argue that it's probably the best we have in terms of painting a picture of investors who are interested in cryptocurrencies. And I also think that there are different types of cryptocurrency investors. It's like, we can't really put all of them in one box and say, "This is a cryptocurrency investor." Right? We have the type of investors who are really into the crypto scene, they're well versed in crypto since the very early days, they use it, not only as an investment, but also as a way of an exchange, potentially. These are the investors or individuals who are probably really into things like Web3 and all of the blockchain related technology and innovation today. I don't think that tracker certificates or ETFs are a great proxy for these types of investors. Those individuals are almost surely holding cryptocurrencies directly, outside of traditional brokerage accounts.
But I do think it's a pretty good proxy for the type of investors who are not these advanced heavy users, but traditional retail investors who are interested in gaining some crypto exposure, without necessarily getting heavily involved learning a whole new set of technologies. Investing in a tracker certificate, for many, is an easy and relatively low effort or low risk way to test the waters with cryptos. You can get exposure to the price path and volatility without setting up and funding a wallet and learning a whole completely set new of platforms or mobile apps and so on. The last thing I want to say here is I also think that this is a very important and growing segment of the market, right? If you think about bringing Bitcoin or cryptocurrencies to the masses, I think these types of investors are actually the early adapters that we would precisely be interested in.
That's an interesting point, even on mass adoption, you wouldn't necessarily expect everybody to get to the point where they're comfortable setting up a wallet and going through all that and managing their keys.
If you think about retail investors, you obviously have a group of retail investors outside of cryptocurrencies that are doing heavy amounts of fundamental research and so on. And these are the tail shareholders. And then you have the ones that are not, right? They just want some exposure to the S&P 500 or Tesla, or something. And I think it's important to look at both these types of groups.
You have data pre and post crypto adoption. Pre-adoption, how do the individual characteristics of crypto adopters compare to the non adopters?
Yeah, they are quite different. Cryptocurrency investors have significantly more assets under management at this brokerage than the non cryptocurrency investors in the sample. It's about 50% larger. We're looking at the average portfolios is 75,000 versus 50,000. They also report basically twice the average monthly income. These guys are generally well off. And I'm often referring to them as guys, which is not sexist, but by and large, these are men. So there's 90% men are cryptocurrency investors, and that's compared to the baseline share of non cryptocurrency investors of about 76% male. They're about the same age, tiny bit younger on average, it's about 47 years old versus 48 years. So it's not really a big difference there. Also quite important is in contrast to the average retail investor, where we tend to think about, in at least in academic studies, cryptocurrency investors are way more active traders. They're making about nine to 10 trades a month compared to about two from the non crypto sample. And they're actually actively checking their portfolio very often. They're logging into their brokerage account almost 90 times a month, which is pretty a heavy user and doesn't really sound like the average passive buy and hold investor.
I think the important part there is, this is before they've invested in crypto. So these are people who will go on to invest in crypto, but you're observing them before they do that. I think that's a... it's just fascinating. How do the portfolios of the eventual crypto adopters compare to the non adopters?
Even before they differ dramatically, right? Crypto currency investors are much more likely to hold single stocks rather than funds. They also hold funds, which generally are associated with a higher risk capacity, such as different types of equity derivatives and warrants. On average, they hold, I think, 15 or 16 unique securities, compared to about seven from the average sample investors. And they also tilt their portfolios from a values weighted sense more heavily into these riskier securities. They hold about 70% of their portfolios in some sort of risky asset relative to cash, compared to about 50% of the average portfolio. And when it comes down to the actual crypto investments, on average, these investors hold about 4,000 euros in cryptos across two to three unique assets. But there is a very large standard deviation here and some investors are holding obviously significantly more.
So they're holding pre crypto adoptions, they're holding more risky assets, what about other stuff like sector ETFs or thematic ETFs and that kind of thing?
Previously, they were more likely to invest in sector ETFs, about three times. They were one and a half times more likely to invest in commodity ETFs. And then these were obviously several years ago when these were introduced as well. We mapped this towards being a little bit eager to try out new financial products and also a bit tech savvy in the sense. Our hypothesis was that these crypto investors are exactly that, more interested in new financial products and tech savvy. So we found that these investors are about three times more likely to use the brokerage mobile app, as opposed to the online portal. In one area where they differ is when it comes down to robo advice. Robo advice, we also think about as kind of a new technology offering. Crypto investors are not using it however, which goes hand in hand, basically with our theory that these investors are actually super active and they're not these passive delegators, which may be more interested in using savings plans or different types of robo advice.
Was there a big difference between the crypto and non crypto investors in terms of the typical investor behavior biases?
This is a good question. We thought there might be, given how active these guys are. So we looked at a couple of different outcomes. We found that crypto investors are generally about three times more likely to chase trends and purchase stocks with recent high performance. These investors seem to be drawn to these recent peak performance products that are following a momentum strategy. At the same time, crypto investors seem to be trading single stocks, which carry also high media sentiment relative to other stocks. It seems like crypto investors are drawn to these trending stocks, which are talked about in the media, or maybe even among other active traders on different web portals and so on.
In general, they're also about three times more likely to trade penny stocks and twice as likely to buy stocks that have lottery like characteristics. These lottery like stocks, you're probably well familiar with them, but they're characterized by low prices and high historical volatility. We're generally interested in these types of securities in academic studies, not only because they resemble gambling or trading for entertainment, but also more recently, there's been some evidence that they carry a much higher risk for being featured in pump and dump schemes, which I think is very interesting and relevant when it comes to crypto and some of the tail end of coins, which were offered towards 2017, for sure.
It's unreal. It's like crypto is just an extension of what they were already doing.
Yeah. Exactly. Yeah. The last thing we wanted to check there in terms of biases, this isn't a bias per se, but we wanted to see if these investors would consider themselves as chartists or technical traders. Yeah, as I said, it's not exactly a bias and there is some evidence that technical trading for cryptocurrencies may have some value given the ambiguous nature of the fundamentals. But I think the evidence for equities is pretty negative on technical trading. It does seem that when they make purchases for single stocks, they're more likely at least to follow moving average indicators. This was another thing that we thought was a potential red flag.
Very interesting. Again, that was in stocks. You mentioned that there's maybe some evidence that works a bit in crypto, it doesn't work in stocks, but these guys were doing it in stocks?
Exactly.
Does their behavior change after they've made their first crypto investment?
Yeah, we believe that it does. I mean, we have to be a little bit cautious here, but in order to examine this as clearly as possible, what we did was we took our cryptocurrency investors and then we did a matching exercise where we tried to reduce any bias which would come, at least from observable characteristics. We matched these investors to non crypto investors based on age, gender, tenure with the brokerage geographic location and so on and so forth. And then we basically, we just look at the time series at how these investors changed behavior or keep the same behavior from before to after their first crypto purchase.
We find that crypto investors, who again, are already very active in managing their portfolio, become even more active after their first purchase. They start to log into their account even more often. And the number of average monthly logins increases by about seven on average, an additional seven to 10 times more logins every month. They also trade more often. They increase the number of assets in their portfolio by adding more non crypto securities as well. And what's really interesting is that they seem to increase their portfolio concentration in these riskier assets. So they're buying more penny stocks and more stocks with high skew and high volatility. And actually they're buying less ETFs. We're not considering a crypto ETF in this case, but they're buying less passive ETFs.
Within the crypto adopters, are there differences between the early adopters and the later adopters?
Yeah, there are. We don't have a massive sample here. It's a relatively small segment of the investor population, but our study does allow us to say a little bit about this. What we can do is we compare the first and the last quartile of cryptocurrency investors in terms of their adoption time. We compare the earliest to the latest and we find that the earliest actually are even more active. They make about 13 trades on average per month, compared to about six. And they log in even more often, they log into their account 115 times compared to about 60. And they also have a higher participation rate in risky products, such as other types of derivatives. 61% of them are holding other derivatives, compared to about 40% for the later adopters. Even the later ones that we pick up here are significantly different from the average sample investor, let's not forget that. It could be that we're eventually converging to the average investor, but it doesn't seem like this is happening super quick. It does seem that the super early adopters are the ones that are the most different from the average investor.
You mentioned location earlier, did you learn anything of interest there?
Not really. We didn't plot these investors by specifically where they live. We're basically just using the geographic dimension as a control variable to make sure that we're not picking up big differences from, let's say, city folk and investors that live out in the country or something like that. But that's interesting. I mean, it's definitely a dimension that could be explored further.
You mentioned that they're adding crypto to their portfolios and adding other risky assets. What percentage of their portfolios are the crypto adopters allocating to crypto?
The average is about 10% to 15% of their portfolio, which is definitely not nothing. But this varies quite a bit. I think the top of the distribution looks more like 35% to 40% of their portfolio held in these assets. Definitely room to learn about cryptocurrencies for these investors. Yeah.
Did crypto investors reduce the riskiness and the rest of their portfolios to compensate for the addition of the crypto?
Yeah, I mean, we would hope so, but that's not the case. Investors don't do that at all in our setting. Instead of basically rebalancing their portfolios to compensate for this additional risk, they tend to do precisely the opposite. As I mentioned, based on after their first cryptocurrency investor, they seem to go all in and invest in other riskier assets as well, such as penny stocks and these high vol or other lottery like assets. So not really, not at all.
What do you think explains that? Like why would their risk seeking behavior increase after a crypto investment?
Yeah, it's a good question. I was thinking that maybe there's some mental accounting going on. These investors are potentially compartmentalizing these assets as a separate thing and not really thinking about the entire portfolio. It could also be that they're just drawn to the entertainment value of this high volatility or the gambling nature of these assets. And once they start, they just want to push it even further.
We know from other research that the individual investors tend to build inefficient portfolios relative to a benchmark index. Is there a difference between the crypto and non crypto investors in terms of efficiency?
Yeah, I mean, we looked at this a little bit early on and I think we ended up delegating it to an online appendix, because the paper became relatively bulky and this wasn't really our contribution. But if I recall, crypto investors, on average, did a tiny fraction of an amount better than the average investor. I would suggest that it probably all came from chance and potentially from being invested in a larger set of assets, and not at all coming from these cryptocurrency investments, for example. But I mean, I would definitely agree that these retail investors tend to be pretty far from the benchmark in general, in our setting as well.
Looking back on this work, how would you describe the typical crypto investor in the study?
I think that these are retail investors who are a little bit savvy and perhaps very eager, as I said, to test out new and innovative financial products. They're not necessarily experts, but they don't necessarily need to be or want to be, right? They learn about things like cryptocurrencies relatively early, and then they use these vehicles to gain some exposure to cryptos.
It is beneficial for those investors? That's a different story, right? So gaining some exposure to cryptocurrencies may be absolutely rational and completely right for these investors. They're not necessarily low wealth, they're not low income, they're not retirees. I also think about a number of recent other studies, which have shown things like structured retail products, yield enhancing products, thematic ETFs, and so on, are all very costly and on average underperform for the vast majority of investors. They're likely also even constructed to capitalize precisely on such a type of investor base, which is really eager to try these things out. I mean, I would say the welfare analysis isn't really solidified yet. Our data gives us an early look at who's buying these assets, but it would be great to do the next analysis, which tells us how do these investors end up over time? What happens if they experience volatility and large price drops and so on and so forth, right?
We covered thematic ETFs in a recent podcast episode where I think it was like the back test always had positive five factor alphas, but post launch of the live fund, they all had negative five factor alphas, big negative alphas.
Yeah. I mean, this is probably coming from this new working paper from Itzhak Ben-David and-
That's the one.
... Yeah, exactly. I'm a big fan of that paper and yeah, for sure.
Yeah. That's a good paper. Oh, I want to switch into a couple of the other papers, not directly related to cryptocurrency, but I think that they're still interesting on this topic. You've got a paper on the disposition effect, which is the tendency to sell winning stocks too early and hold on to losers too long. And I think that's really relevant in crypto where the market is exceptionally volatile. How do the beliefs that people hold about expected returns affect the incidents of the disposition effect?
This study is, this is actually from Denmark. What we do is we measure investors' risk attitudes and expectations on the stock market in a controlled experimental laboratory setting. And then we relate these measures to portfolio characteristics, portfolio choices, and eventually biases that these investors make in the real world. What they make in the field after the experiment takes place. And what we find in general is that investors who are more optimistic on the market as a whole seem to be more likely to exhibit the disposition effect in their trades in the future, relative to other active investors who are a bit more pessimistic. And when I talk about the market as a whole here, I'm talking about Denmark, so we're talking about the OMX Copenhagen 20.
But what we find is essentially that this relationship between individuals investors' beliefs and the incidents or their propensity to have the disposition effect in the future is fairly non-linear. The incidence of the disposition effect is increasing at low optimistic levels and begins to decrease as expected returns on the market increase to very high levels. Once that expected return on the market reaches 25% or so, then the incidence of the disposition effect starts to trend downward, because there's no reason to sell off stocks when you think that the market is going to return an additional 25%. What's interesting is that these empirical findings, which we document, are actually in line with previous theoretical models of the disposition effect, which focus heavily on prospect theory and other preference based explanations. I would say our paper fits in alongside these, in that what we're trying to do is we're trying to extend the belief side of the picture and not just look at preferences to understand how expectations may play a role in investors' choices, and ultimately in the biases that we observe.
Fascinating. Is risk appetite related to the disposition effect?
Yeah. We actually find a very limited correlation between the measures of risk, which we elicited among investors, and their incidents of the disposition effect in the field. On the other hand, it's very difficult to pin down precisely what is driving the incidence of the disposition effect. Most of the research to date, as I mentioned, has really focused on investor preferences. So differences in risk appetite, loss aversion, other risk stories, basically. And only recently have we started thinking about heterogeneous beliefs. What we did was we really spent a lot of time trying to ensure that we are adequately controlling for differences in risk attitudes. We're using different structural utility measures of different types of risk preferences and so on and so forth. Honestly, no matter what we did, we couldn't really break the correlation between beliefs and the disposition effect, and risk attitudes didn't really seem to add too much to that analysis. Basically we find the correlation between risk and the disposition effect is relatively minor.
Are the beliefs that people hold about expected returns affected by their past realized expected returns?
Yeah, this is important. Generally investors form beliefs, in our setting based, on various attributes of their past returns and portfolio experiences. We looked not only at the mean expected return, but also in the variance. In general, we found that the variance of expected returns didn't vary dramatically based on different past returns or different portfolio attributes, but the mean varied significantly. Expected returns was affected by individual's past returns, both realized gains as well as paper gains. Higher portfolio returns led to more optimistic beliefs. And worse returns led to more pessimistic beliefs.
When it comes to looking at paper gains versus paper losses, we found a pretty large effect of paper gains affecting future beliefs, but actually paper losses didn't really seem to have an effect. And what that suggests to us is if you have this unrealized loss, you might not be updating your beliefs based on it yet. You might be ignoring it, to some extent, and it doesn't have an effect yet on your expectations of the future. Which fits well into our study on the disposition effect, and actually share some similarities with previous experimental work, which has shown that when the outcome doesn't really line up with what you expected it to do, you're not going to adequately update for the future. And thus you have this tendency to then have this paper loss, which causes investors to ignore that information, which would then continue propagating the disposition effect, because you wouldn't be updating based on this paper loss.
These insights are absolutely fascinating. Do positive and negative realized past return experiences have the same effect on beliefs?
Yeah, I think we found that realized gains had a strong effect on optimism, but past returns had less of an effect. I think that's basically what we found. Realized gains were creating this optimism in individual's beliefs, whereas realized losses have, to some extent, this effect, but not as much, not nearly as much.
Interesting. Okay. Paper losses have pretty much no effect, realized losses have some effect, but still a weaker effect?
Exactly, exactly. Yeah.
Interesting. And I just want to reiterate for listeners, when we're talking about expected returns in this case, we're literally talking about the returns that the individual investors expect to earn on their investments.
Exactly.
We talk about expected returns a lot, but we're usually talking about discount rates.
Yeah, that's true. For sure. Yeah.
Do the beliefs that people hold effect how much they allocate to risky assets?
Yeah. They do. And actually we're happy to find that individuals' belief actually matter for how they allocate their wealth. If they didn't, we would question our measures of beliefs. But we indeed find a pretty strong, positive correlation. And then number of previous work has also used, I think, surveys to document another correlation, or a similar correlation. We find basically investors, they allocate about five percentage points more of their financial wealth to risky assets, if their expected returns increased by 10%. If you were to look at a sample of pessimist and optimist, you would see a pretty large wedge in the overall fraction of wealth they have devoted to risky assets. I should also note that the average risk share of the investor in our sample is about 35%. So a five percentage point effect is definitely economically important.
To switch to one other paper, which is another fantastic one, you did some really cool stuff at the household level. How do people that get burned by bad outcomes and risky investments change their future behavior?
Thanks for asking about this paper. I like this one a lot. I mean, I like all of them, but this one I think is very interesting. More importantly, experience effects and people getting burned and updating their future behavior, I think is just so relevant and so important for retail investors and households. A number of studies have previously shown that there are these very long run cohort level effects, which are drawn from formative experiences. The best known one is obviously investors from the era of The Great Depression, much later on in life, tend to hold the lower fraction of their wealth in risky assets and they tend to participate in financial markets at a much lower rate.
But one question which comes up from this is actually how important is the degree to which you experience something? And how important is that degree related to future financial risk taking? Are basically these large shocks enough to change behavior? Or is it actually that the ones that react and shy away from risk taking, the ones that were more closely affected during that time? To try to say something about this, we fielded this research study in Denmark. I did my PhD in Denmark, that's why I have all these studies from Denmark. But we have a very interesting setting in Denmark, that's why this is super interesting. But basically prior to the great financial crisis of 2007 to 2009, we saw many households in Denmark holding sizable investments in the shares of their own local bank stocks. These banks are regional, but publicly held. And households were actually encouraged to invest a relatively large amount of their long term savings directly in these assets to capitalize on, yeah, higher expected returns or the boom market at the time.
As the crisis evolved and spread to Europe and to Denmark, a number of these banks actually became distressed and many then defaulted. For households, their savings and deposits as bank customers were guaranteed and insured, but any investments that they had in these risky assets, like in the bank stock directly, incurred major losses. So these investments mostly, if not all of them, effectively went to zero. And because of that, we can't really say how their financial wealth allocation to risky assets is affected, because you have this mechanical effect of losing this wealth shock.
So what we did was we looked at this sub sample of investors who are also making inheritance decisions at the same time. These investors held these securities at the bank and for completely unrelated reasons, are inheriting a portfolio from a deceased parent or relative. Of course, this sub-sample and this cross section of the number of investors who have this experience and the ones that are also inheriting is quite small, but the benefit of these types of studies from Denmark is you actually have the universe of Danish investors at your disposal. As a researcher, you work with statistics. Denmark and you can observe basically everyone. The sample size is definitely large enough.
Just getting to the main point, what we do here is we look at this sub-sample investors who are making inheritance decisions. And what we find is that those who have this firsthand experience, the ones that are actually getting burned and losing these investments in their bank stocks, are much more likely to take these inheritances of risky assets and liquidate them to cash. They are basically selling the risky assets off and holding the cash. What we observe is that these investors are basically reducing their risk taking. And the effect size is about 10 percentage points relative to other investors who are making inheritance decisions, who did not have this firsthand experience of losing investments in their bank stocks.
One thing that's really nice about the setting is that we can see what would happen to these investors had they inherited shortly before they made the experience. There's investors who haven't yet experienced the bank shock, but are still inheriting some months or a year before the financial crisis. Basically, these investors make absolutely no changes to the inherited portfolio. They're very comfortable holding the risky assets. It's really only the investors that make these firsthand personal experiences, which causes them to actively sell off these inheritances of risky assets. And they prefer to hold cash.
Wow. So once burned, twice shy?
Yeah, exactly.
Incredible. Do the experiences of peers or relatives have a similar effect?
Yeah, I think this is really important. We find that the degree of how closely you make this experience is critical for affecting your future risk taking. The peers or neighbors or relatives of those with the firsthand experiences, don't really change behavior at all. There's some marginal rebalancing here and there, but it really seems that the investors that put their hand in the fire are the ones that then really jump back and turn down their risk taking.
What do you think causes the reduced risk taking of the people who've had these bad experiences?
I think we can partially at least attribute the change in risk taking to basically a reduction in trust into financial markets. In our setting, financial advisors essentially violated fiduciary duty by encouraging these households or these novice investors, most likely, to take undiversified risks in their own banks or their bank's own stock. I can well imagine that these investors shy away from financial decisions because of this. And in fact, we also find that the largest active reduction in risk taking in the future comes specifically from avoiding stocks in the banking and financial sector. That suggests that there might be some mistaste for finance.
Do you think people who might get burned investing in crypto will see a similar effect?
Yeah. I've been thinking a lot about this and I think it's likely. But anecdotally I do know many who got involved with crypto towards mid to end of 2017. We had this big price run up for Bitcoin. And these were relatively young, novice investors who invested relatively heavy into a very wide spectrum of various coins, many which surely don't exist anymore. I know of several who lost quite a bit and stepped fully away from crypto assets. On one hand, making these types of experiences at a young age may be important for building up better habits in the future. If you lose a little bit early on, maybe that saves you from making a much bigger mistake later on, as long as you adequately learn from these experiences. On the other hand, experiencing a large loss later in life could be very detrimental and it could be a very large risk that future cost of nonparticipation is going to compound these experienced losses.
Just to get back to that question, I think it's definitely possible. My hope is that the crypto investors are at least a bit younger and maybe we'll learn from it for the future. Whereas opposed to the danger setting, where I think a lot of our households and sample investors were a bit older and this was serious, long term savings that they were losing.
I know we're just speculating at this point, but it's super interesting to think about in your sample, the investors shied away specifically from investing in bank stocks. In the crypto scenario, if they get scared off of investing in crypto, but not other financial markets, maybe that's not such a big loss, but if they get scared off of investing in general, then it's a big loss?
Absolutely. Yeah. Maybe it's not a bad thing at all if they learn that cryptocurrencies are maybe not where they should be first investing and maybe that's only a small fraction of their portfolio and so on and so forth. But yeah, it's an empirical question, I think.
You gathered this data in Denmark, Ben and I are in Canada. Is there any reason that all these findings wouldn't apply to other markets or parts of the world?
I think the experience effects would definitely apply broadly. There's plenty of studies. There's studies from Finland. There's studies from Germany. There's studies from the United States on these types of experience effects. And I can't imagine a scenario where one group of investors would basically make these types of experiences and then either not react to them or react in a perfectly efficient way. I think that this is inherently a feature of human behavior, that many people are going to experience a shock and unfortunately shy away from risk a little bit, at least in the short term, following those events.
What's the more theoretically appropriate response to losing money in a concentrated investment like the bank stocks in your sample?
It would be ideal if investors learn from their mistakes and rather than hold concentrated portfolios and then shy away from risk, they actually say, "Well, I should diversify." Would've been fantastic to see investors with stronger experiences then heavily tilt towards passive ETFs. But unfortunately that's just not what we're observing.
I want to come back quickly to the disposition effect paper. What are the lessons when we're talking about crypto specifically in this highly volatile market from the findings in the disposition effect paper? Are there takeaways that you think are important for crypto investors specifically?
This is a very new area. Because of the data availability, we don't know much about how individuals are actually trading cryptocurrencies. My guess is that we would find a lot of similarities in terms of biases and in terms of the disposition effect. It's really an empirical question and we haven't been able to open this black box. In our setting from the German data on indirect investments, we have really a pretty limited time series. So I can't really observe people that are holding losses and not realizing them and so on and so forth. My guess is that we would find a lot of similarities between the biases investors make on single stocks and equities, as the ones that they would in cryptocurrencies.
The disposition effect is very well documented and it's something that comes up in all different types of asset markets. We see them in funds, in stocks in real estate, even in art markets and so on and so forth. So I can't imagine cryptocurrencies would be immune to them. In fact, I think they're probably even stronger when you think about the movements of investors holding Bitcoin and so on and so forth, because they're hoping for very high returns in the long run. That being said, I think it's very much an expectations and belief story rather than a risk story. And that would be how I would link these two themes.
That's where my question was coming from is like, you found this relationship between the returns that investors expect when they buy an investment and the disposition effect. And then anecdotally, from people that I know investing in crypto, they are expecting exceptionally high returns, which is one of the main reasons they're investing in the asset class.
Exactly.
Cool. All right. Well, Tobin, this has been fantastic. It was great to talk about your paper on crypto investors, but these other two papers also, I think, add lots to the discussions. We really appreciate you coming on the podcast.
Great. I really appreciate that you guys had me on. It's been great. So thank you.
Tobin. Great to meet you. Great insights. And thanks everybody out there for listening.
Participate in our Community Discussion about this Episode:
Links From Today’s Episode:
Rational Reminder on iTunes — https://itunes.apple.com/ca/podcast/the-rational-reminder-podcast/id1426530582.
Rational Reminder Website — https://rationalreminder.ca/
Shop Merch — https://shop.rationalreminder.ca/
Join the Community — https://community.rationalreminder.ca/
Follow us on Twitter — https://twitter.com/RationalRemind
Follow us on Instagram — @rationalreminder
Benjamin on Twitter — https://twitter.com/benjaminwfelix
Cameron on Twitter — https://twitter.com/CameronPassmore
Tobin Hanspal on Twitter — https://twitter.com/tobinbridge
Tobin Hanspal — http://www.tobinhanspal.com
'The Characteristics and Portfolio Behavior of Bitcoin Investors: Evidence from Indirect Cryptocurrency Investments' — https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3501549
'Beliefs and the Disposition Effect' — https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3516567
'Once Bitten, Twice Shy: The Power of Personal Experiences in Risk Taking' — https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2506627
22 in 22 Reading Challenge — Join the Rational Reminder’s 22 in 22 reading challenge!
Ben’s Reading Code (22 in 22 Challenge): 7XWESMK
Cameron’s Reading Code (22 in 22 Challenge): N62IPTX