Return space it performance distribution. What if we cannot see all the portfolio holdings? So by some reason, manager does not reveal the insights of his portfolio. For example, he does not want you to know exactly what he's investing in because he doesn't want you to replicate his investment strategy. Can we actually hypothesize what is his investment style? What are sources of his investment success. In that case we need to have an information at least about price performance of his portfolio. Just about returns, if we have returns you already know that we actually can calculate drawdowns. You can see downside and tail risks. You can calculate different ratios and you can look at related risks. And now we will speak about factor based analysis. We will apply factor based analysis to time series of returns of some portfolio strategy or fund. And then we can attribute style using results for analysis and we can actually even look at style mismatch and style drift. Style drift is when we have manager with some period of time. Manager is producing one style and then he starts pushing the other style. In that case the data. Us that he has a style drift, so to perform return space it performance analysis we need first to make benchmark indices. So we have the universe of assets and then this universe should be divided to different styles. We have to build style indices and this index would represent some particular style for us. And then we will calculate exposures to that style indices, and we will look at the how exposures, different style indices evolve it overtime. Let's look how to do that in Bloomberg. Alright, let's perform return spaced analysis and Bloomberg. We're using FFT, a function to perform returns based analysis. So you have to first remember the name of some fun. So let it be SPY. I'm sorry, SBY. AUS equity. Which stays for Esped a index fund. So if I want to analyze this fund. This is a fund which actually track exchange traded fund, which tracks as and P500 index. So let's say that I want to perform. Return space analysis of this fun. Then I'm using the FSDA function style analysis. So this is what this function does, so I'm pushing the SPYUS equity ticker here. Then I select how often I want to re balance, re estimate my model. I want to re estimate my model daily, only window of length one year. Then I have several indices here. Every index represents some portion of a financial or financial market, so this indices would be my factors. So for example. I make Bloomberg estimate a model. Now I have the R square at 997., which means that 997 variability of returns of my fund SBYUS equity dale returns, on average 997 is X plane. It's by variation of returns of these eight indices. Let's look at this in this in more detail. So here we have family of indices so. These indices slice the total universe of stocks, United States stocks, and this indices do not enter set. So the here I have indices for US large growth sucks you large values stocks, mid growth mid value small growth, small value and I also have indices for three months library and for Canada stocks. So I see that these indices capture all the variability, in fact, all the variability of returns of my fund or fund or the style which I'm analyzing. So I see that approximately half. No variability is captured by US large growth stocks and the rest is captured by US large value stocks. And just a notch is captured by me growth Canada and LIBOR. Nothing is captured by mid value small growth small value. Now let's look at the this picture in Dynamics. Let's press rolling window tab. In rolling window I see how beetus evolve it How bid has evolved over time. So you see that most of the time, approximately half of exposure of my original fund is to US large value stocks. And the rest is on UOS large growth stocks. And you see how stable is this distribution? So I see that historically, the style was very persistent. There were no so-called style drift, right? So the style is this same. Now let's analyze which is actually quite understandable if we would look at description so the job of SPDR S&P 500 ETF TRUST is to track the index. And by the way here in Rolling window, we also can look at, for example, tracking error, you see that the tracking error is always very, very small, old point or 9%. So less than one person, this is tracking error, standard deviation of active, return. You remember that? And the R squared is always very high. It's almost 100 persons through all the time. Now let's look at the other funds. The other one which I have chosen for to analyze style of that fund is Berkshire Hathaway. So this is famous Fund, which is led by prominent United States investor, Warren Buffett. So this is how Berkshire Hathaway looks in style analysis function in lumber. again, the universe of indices which we would apply to this fund is the same as we have applied to the previous fund. Although we can actually choose factors so there are different factor models for different funds. We can select best fit. Which means that the Bloomberg Fit the whatever model have the biggest r squared. Or it may use some suggested model suggested by the very read the description actually all the funds. Or we can actually make our own model using custom factors, but I will use default. So default is the same. This is the standard model for United States equities universe. Again, I would estimate my model on a window of one year. I would re-estimate my model daily. So rolling window is analysis. Again, one year For the last seven years. Nice, nice. What do we have here? You see that clearly the style of the font differs from the style of the previous font. We know that actually because Berkshire Hathaway is an active fund. So this fund is not tracking any index but can we conclude something about this style of this fund? Definitely, we can see that majority of the time this fund Is largely dependent on return of us large value stocks. There is quite big contribution of so called other factors. So these factors means that they are not captured by LIBOR, Canada and so on and so forth. And r square is warm model is lower than the r squared of the previous model, which means that using these factors we manage it to explain only part of variability of returns of Berkshire Hathaway. Berkshire Hathaway have quite big alpha.It have big tracking error. And we see that this style of Berkshire Hathaway exhibits quite significant drift. So sometimes the Berkshire Hathaway is completely invested in large value stocks. Sometimes return of Berkshire Hathaway is best explained by some other factors. We did not know why. But we can actually attribute these factors to the skill of Mr. Warren Buffett. So basically this is how retorts basic style analysis is performed in Bloomberg. You can use this analysis to any Font and you can conclude whether there is or there is no style drift and or how what is what is actually the style of some particular fund.