# Lamperti Walks

The theory of simple random walks on the integer lattice is a classical topic in probability theory. Polya proved in the 1920s that such a SRW on $\mathbb{Z}^d$ is recurrent only for d=1 or 2. The argument is essentially combinatorial. We count the number of possible paths from 0 back to itself and show that this grows fast enough that even with the probabilistic penalty of having a particular long path we will still repeatedly see this event happening. In larger dimensions there is essentially ‘more space’ at large distances, at least comparatively, so a typical walk is more likely to escape into this space.

As Kakutani (of the product martingale theorem) said, and was subsequently quoted as the dedication on every undergraduate pdf about random walks: “A drunk man will find his way home, whereas a drunk bird may get lost forever.”

But transience in some sense a long-distance property. We can fiddle with the transition rates near zero and, so long as we don’t make anything deterministic this shouldn’t affect transience properties. Obviously if we have a (space-)homogeneous nearest-neighbour random walk on the integers with non-zero drift the process will be transient: it drifts towards positive infinity if the drift is positive. But can we have a random walk with non-zero drift, but where the drift tends to zero at large distances fast enough, and the process is still recurrent? What is the correct scaling for the decay of the drift to see interesting effects?

The answers to these questions is seen in the so-called Lamperti random walks, which were a recurring theme of the meeting on Aspects of Random Walks held in Durham this week. Thanks to the organisers for putting on such an excellent meeting. I hadn’t known much about this topic before, so thought it might be worth writing a short note.

As explained above, we consider time-homogeneous random walks. It will turn out that the exact distributions of the increments is not hugely important. Most of the properties we might care about will be determined only by the first two moments, which we define as:

$\mu_1(x)=\mathbb{E}[X_{t+1}-X_t | X_t=x],$

$\mu_2=\mathbb{E}[(X_{t+1}-X_t)^2 | X_t=x].$

Note that because the drift will be asymptotically zero, the second term is asymptotically equal to the variance of the increment. It will also turn out that the correct scaling for $\mu_1$ to see a phase transition is $\mu_1(x)\sim \frac{c}{x}$.

We begin by seeing how this works in the simplest possible example, from Harris (1952). Let’s restrict attention to a random walk on the non-negative integers, and impose the further condition that increments are +1 or -1. In the notation of a birth-and-death process from a first course on Markov chains, we can set:

$p_j:=\mathbb{P}(X_{t+1}=j+1| X_t=j), \quad q_j=1-p_j.$

We will set $p_j=\frac12 + \frac{c}{2j}$. Then a condition for transience is that

$1+\frac{q_1}{p_1}+\frac{q_1q_2}{p_1p_2}+\ldots <\infty.$

In our special case:

$\frac{q_1\ldots q_r}{p_1\ldots p_r}\approx\frac{(r-2c)(r-1-2c)(r-2-2c)\ldots}{r!}\approx \frac{1}{r^{2c}}.$

So we can deduce that this sum converges if c>1/2, giving transience. A similar, but slightly more complicated calculation specifies the two regimes of recurrence. If -1/2<=c<=1/2 then the chain is null-recurrent, meaning that the expected time to return to any given state is infinite. If c<-1/2, then it is positive recurrent.

In general, we assume $\mu_1(x)\sim \frac{c}{x}$ and $\mu_2(x)\approx s^2$. In the case above, obviously $s^2=1$. The general result is that under mild assumptions on the increment distributions, for instance a $(2+\epsilon)$-moment, if we define $r=-\frac{2c}{s^2}$, then the RW is transient if r<-1, positive-recurrent if r>1, and null-recurrent otherwise. This is the main result of Lamperti.

To explain why we have parameterised exactly like this, it makes sense to talk about the more general proof methods, as obviously the direct Markov chain calculation won’t work in general. The motivating idea is that we can deal well with the situation where the drift is zero, so let’s transform the random walk so that the drift becomes zero. A function of a Markov chain that is more stable (in some sense) that the original MC, for analysis at least, is sometimes called a Lyapunov function. Here, the sensible thing is to consider $Y_t=X_t^\gamma$, for some exponent $\gamma>0$.

So long as our distributions are fairly well-behaved (eg a finite $2+\epsilon$-moment), we can calculate the drift of Y as

$\mathbb{E}[Y_{t+1}-Y_t| X_t=x]=\frac{\gamma}{2}x^{\gamma-2}(2c+(1-\gamma)s^2) +o(x^{\gamma-2}).$

In particular, taking $\gamma=1+r$ results in a random walk that is ‘almost’ a martingale. Note that the original RW was almost a martingale, in the sense that the drift is asymptotically zero, but now it is zero to second order as well.

To draw any rigorous conclusions, we need to be careful about exactly how precise this approximation is, but we won’t worry about that now. In particular, we need to know whether we can take this approximation over the optional stopping theorem, as this allows us to say:

$\mathbb{P}(X\text{ hits }x\text{ before 0})=\mathbb{P}(Y\text{ hits }x^\gamma\text{ before 0})\sim x^{-\gamma}.$

This is particularly useful for working out the expected excursion time away from 0, which precisely leads to the condition for null-recurrence.

In his talk, Ostap Hryniv showed that this Lyapunov function analysis can be taken much further, to derive much more precise results about excursions, maxima and ergodicity. Results of Menshikov and Popov from the 90s further specify the asymptotics for the invariant distribution, if it exists, in terms of r.

One cautionary remark I should make is that earlier I implied that once we know the drift of such a random walk is zero, we have recurrence. This is true on $\mathbb{Z}$ with very mild restrictions, but is not necessarily true in higher dimensions. For example, consider the random walk on $\mathbb{R}^2$, where conditional on $X_t$, the increment is $X_{t+1}-X_t$ is of length 1 and perpendicular to the vector $X_t$. The two possible directions are equally likely. The drift is therefore 0 everything, and the second moment is also well-behaved, but note that $||X_t||^2=t^2$, just by considering Pythagoras. So in higher dimensions, we have to be a bit more careful, and put restrictions on the covariance structure of the increment distributions.

As a final comment, note that from Lamperti’s result, we can re-derive Polya’s result about SRW in higher dimensions. If we have $X_t$ an SRW on $\mathbb{Z}^d$, then consider $Y_t=||X_t||$. By considering a couple of examples in two-dimensions, it is clear that this is not Markov. But the methods we considered above for the Lamperti walks were really martingale methods rather than Markov chain methods. And indeed this process Y has asymptotically zero drift with the right scaling. Here,

$c=\frac{1}{2}(1-\frac{1}{d}),\quad s^2=\frac{1}{d},$

and so r=d-1, leading to exactly the result we know to be true, that the SRW is transient precisely in three dimensions and higher.

REFERENCES

Harris – First Passage and Recurrence Distributions (1952)

The slides from Ostap Hryniv’s talk, on which this was based, can be found here.

# Invariant Distributions of Markov Chains

My lecture course in Linyi was all about Markov chains, and we spent much of the final two sessions discussing the properties of invariant distributions. I was not surprised, however, that none of the class chose this topic as the subject for a presentation to give after the end of the teaching week. One of the main problems is that so many rather similar properties are introduced roughly simultaneously. As we did in the class, I thought it was worth making some sort of executive summary, as a mixture of revision and interest.

Definition: $\pi$ is an invariant measure if $\pi P=\pi$. If in addition $\sum_{i\in I}\pi_i=1$, then we say it is anÂ invariant distribution. Of course, if I is finite, then any invariant measure can be normalised to give an invariant distribution.

The key initial questions are about existence and uniqueness. First, if there are multiple communicating classes, then an invariant measure (resp. distribution) is a linear (resp. affine) combination of the invariant measures / distributions on each (closed) class. So we restrict attention to irreducible Markov chains.

In the finite case, P is a stochastic matrix so has a column eigenvector with eigenvalue 1, namely the vector with all entries equal to 1. Thus, by reference to general theory in linear algebra, P has a row eigenvector $\pi$ with eigenvalue 1. To paraphrase a remark made by one of my students, what is not clear is that this should be a measure. Demonstrating that this is true is rather non-trivial I think, normally done by reference to the rather more general Perron-Frobenius theorem, though on the flight home I came up with a short argument using Lagrangian duality. For now, we accept existence in the finite case, and note that we typically show existence by showing that the vector of expected time spent in each state between successive visits to a fixed reference state satisfies the properties of an invariant measure.

This is a good moment to note that recurrence is not a necessary condition for the existence of an invariant measure. For example, the random walk on $\mathbb{Z}^3$ is transient, but the uniform measure is invariant. However, it is not a sufficient condition for the existence of an invariant distribution either. (Of course, an irreducible finite chain is always recurrent, and always has an invariant distribution, so now we are considering only the infinite state space case.) The random walk on $\mathbb{Z}^2$ is recurrent, but the invariant measure is not normalisable.

The property we in fact need is positive recurrence. This says that the expected return time to each point is finite. Again, this is a class property. This is a common requirement in probabilistic arguments: almost surely finite is often not strong enough to show results if the expectation is infinite (see for example the various requirements for the optional stopping theorem). If this holds, then $\pi_i=\frac{1}{\mathbb{E}T_i}$, where $T_i$ is the the return time starting from some $i\in I$.

The final question is ‘Why are we interested?’ One of the best answers is to look at convergence properties. A simple suggestion is this: if we start in equilibrium, then $X_0,X_1,X_2,\ldots$ are all equal in distribution. Note that the dependence structure remains complicated, and much much more interesting than the individual distributions. Next, we observe that a calculation of n-step transition probabilities for a finite chain will typically involve a linear combination of nth powers of eigenvalues. One of the eigenvalues is 1, and the others lie strictly between -1 and 1. We observe in examples that the constant coefficient in $p_{ij}^{(n)}$ is generally a function of j alone, and so $p_{ij}^{(n)}\rightarrow\lambda_j$, some distribution on I. By considering $P^{n+1}=P\cdot P^n$, it is easy to see that if this converges, $(\lambda_j)_{j\in I}$ is an invariant distribution. The classic examples which do not work are

$P=\begin{pmatrix}0&1\\1&0\end{pmatrix}$ and $P=\begin{pmatrix}0&1&0\\ 0&0&1\\1&0&0\end{pmatrix}$,

as then the distribution of $X_n$ is a function of the remainder of n modulo 3 alone. With a little thought, we can give a precise classification of such chains which force you to be in particular proper subsets of the state space at regular times n. Chains without this property are calledÂ aperiodic, and we can show that distributions for such chains converge to the equilibrium distribution as $n\rightarrow\infty$.