One of the outstanding problems in general relativistic cosmology is that of the averaging. That is, how the lumpy universe that we observe at small scales averages out to a smooth Friedmann-Lemaitre-Robertson-Walker (FLRW) model. The root of the problem is that averaging does not commute with the Einstein equations that govern the dynamics of the model. This leads to the well-know question of backreaction in cosmology. In this work, we approach the problem using the covariant framework of Macroscopic Gravity (MG). We use its cosmological solution with a flat FLRW macroscopic background where the result of averaging cosmic inhomogeneities has been encapsulated into a backreaction density parameter denoted $\Omega_\mathcal{A}$. We constrain this averaged universe using available cosmological data sets of expansion and growth including, for the first time, a full CMB analysis from Planck temperature anisotropy and polarization data, the supernovae data from Union 2.1, the galaxy power spectrum from WiggleZ, the weak lensing tomography shear-shear cross correlations from the CFHTLenS survey and the baryonic acoustic oscillation data from 6Df, SDSS DR7 and BOSS DR9. We find that $-0.0155 \le \Omega_\mathcal{A} \le 0$ (at the 68\% CL) thus providing a tight upper-bound on the backreaction term. We also find that the term is strongly correlated with cosmological parameters such $\Omega_\Lambda$, $\sigma_8$ and $H_0$. While small, a backreaction density parameter of a few percent should be kept in consideration along with other systematics for precision cosmology.

We investigate the effect of averaging inhomogeneities on expansion and large-scale structure growth observables using the exact and covariant framework of Macroscopic Gravity (MG). It is well-known that applying the Einstein's equations and spatial averaging do not commute and lead to the averaging problem. For the MG formalism applied to the Friedmann-Lemaitre-Robertson-Walker (FLRW) metric, this gives an extra dynamical term encapsulated as an averaging density parameter denoted $\Omega_A$. An exact isotropic cosmological solution of MG for the flat FLRW metric is already known in the literature, we derive here an anisotropic exact solution. Using the isotropic solution, we compare the expansion history to current data of distances to supernovae, Baryon Acoustic Oscillations, CMB last scattering surface, and Hubble constant measurements, and find $-0.05 \le \Omega_A \le 0.07$ (at the 95% CL). For the flat metric case this reduces to $-0.03 \le \Omega_A \le 0.05$. We also find that the inclusion of this term in the fits can shift the values of the usual cosmological parameters by a few to several percents. Next, we derive an equation for the growth rate of large scale structure in MG that includes a term due to the averaging and compare it to that of the LCDM concordance model. We find that an $\Omega_A$ of an amplitude range within the bounds above leads to a relative deviation of the growth from that of the LCDM of up to 2-4% at late times. Thus, the shift in the growth could be of comparable amplitude to that caused by similar changes in cosmological parameters like the dark energy density parameter or its equation of state. The effect could also be comparable in amplitude to some systematic effects considered for future surveys. This indicates that the averaging term and its possible effect need to be tightly constrained in future precision cosmological studies. (Abridged)